Vous êtes sur la page 1sur 513

NASA

T!r
C.1

1
i

A.N. Eelikman, O.E. Krein, and G.V. Samsonov


I .

..

;
"
r

r-.
I

TRANSLATED FROM RUSSIAN

Published for the U.S. National Aeronautics and Space Administration and the National Science Foundation, Washington, D.C.

by the Israel Program for Scientific Translations

TECH LIBRARY KAFB, NM

1 l1 ll1 Il1 1 1 1 1lllllllllll1IlI 1 l 0068748

A. N. Zelikman, 0. E. Krein, and G. V. Samsonov

METALLURGY
OF RARE METALS

(Metallurgiya redkikh metallov) Second Edition, Revised and Enlarged Edited by L. V. Belyaevskaya Approved by the Ministry of Higher and Secondary Specialized Education of the USSR as a textbook for students of nonferrous metallurgy

Izdatel'stvo Metallurgiya Moskva 1964

Translated from Russian

Israel Program for Scientific Translations Jerusalem 1966

NASA TT F-359 TT 65-50137 Published Pursuant to an Agreement with THE NATIONAL AERONAUTICS AND SPACE ADMINISTRATION, U. and THE NATIONAL SCIENCE FOUNDATION, WASHINGTON, D. C.

s. A.

Copyright 01966 Israel Program for Scientific Translations Ltd. IPST Cat. No. 1455

Translated by A. Aladjem, Chem. Eng.

Printed in Jerusalem by S. Monson Binding: Wiener Bindery Ltd., Jerusalem

P r i c e : $ 8.12

Available from the U. S. DEPARTMENT OF COMMERCE


Clearinghouse f o r Federal Scientific and Technical Information
Springfield, Va. 22151

VI/18/3.5

- ..

-- -.

_.

._. .....

..

TABLE OF CONTENTS Page

........................................... FOREWORD ....................................................... INTRODUCTION ................................................... Classification of rare metals .......................... Certain features of the production of rare metals from
ores .............................................
LIST OF ABBREVIATIONS Development of the rare-metal industry in the USSR Part One

xiv

xvi

xvii
xxi
xxiii xxv

.....

. REFRACTORY METALS
Chapter I. TUNGSTEN ................................................ 1
1. GENERAL DATA ON TUNGSTEN ................................ 1

Brief Historical Note Properties of Tungsten ...................................... T h e Properties of Tungsten Compounds Uses of Tungsten .......................................... 2 MINERALS. ORES. A N D ORE CONCENTRATES .................... Tungsten Minerals Tungsten Ores and Deposits .................................. The Enrichment of Tungsten Ores 3 PROCESSING OF TUNGSTEN CONCENTRATES 4 . DECOMPOSITION OF WOLFRAMITE CONCENTRATES Fusion with Sodium Carbonate Decomposition with Sodium Hydroxide Solutions 5 PROCESSING OF SODIUM TUNGSTATE SOLUTIONS Removal of Impurities from t h e Solutions Isolation of Tungsten Compounds from Solution 6 PROCESSING OF SCHEELITE CONCENTRATES Fusion with Sodium Carbonate Decomposition in Autoclaves with Aqueous Solutions of
Sodium Carbonate Acid Decomposition

.......................................

........................

1
1

.........................................

3
7
9
9

10

. .

............................. 1 0
...................1 2
............1 3
................................ 13
................ 18

............... 19
...................... 1 9
................ 22
.................... 24
............................... 24

25

........................................ ........................................
...

27

lll

Page PURIFICATION OF TUNGSTIC ACID Dissolution of H2W04 in Ammonium Hydroxide Solution Isolation of Tungsten from Ammoniacal Solutions 8 PRODUCTION OF TUNGSTEN TRIOXIDE AND QUALITY
CONTROL 9 COSTS OF PRODUCTION OF TUNGSTIC ANHYDRIDE 10 PRODUCTION OF METALLIC TUNGSTEN 11 REDUCTION OF TUNGSTEN TRIOXIDE BY HYDROGEN Physicochemical Conditions of t h e Reduction Process Reduction Furnaces T h e Reduction Process Control of the Particle Size of Tungsten Powder 12 REDUCTION OF TUNGSTEN TRIOXIDE BY CARBON Physicochemical Conditions for Carbon Reduction Industrial Reduction of WO, by Carbon 13 PRODUCTION OF SOLID TUNGSTEN Pressing Sintering Sintering Mechanism Structure of the Bars Quality Control of t h e Sintered Bars Mechanical Working of Sintered Bars 7

........................ 35
............. 36
............. 36
......................................... 39
...................................... 41
................ 43
. ............... 44
............... 44
......................... 46
. ............................ 46
................................................... 47
.................................................. 49
......................................... 54
........................................ 55
............................ 56
.......................... 57
60
Chapter I1. MOLYBDENUM ............................................. 14. GENERAL DATA ON MOLYBDENUM ............................. 60
60
Brief Historical Note ......................................... Properties of Molybdenum .................................... 60
T h e Properties of Molybdenum Compounds ..................... 62
Uses of Molybdenum ........................................ 66
1 5. MINERALS. ORES. AND ORE CONCENTRATES .................... 68
Molybdenum Minerals ....................................... 68
Molybdenum Ores and Ore Deposits ............................ 69
70
Concentration of Molybdenum Ores ........................... 16. PROCESSING OF STANDARD MOLYBDENITE CONCENTRATES ....... 71
17. OXIDATIVE ROASTING OF MOLYBDENITE CONCENTRATES ....... 71
Roasting Procedure .......................................... 14
Roasting in Multiple- Hearth Furnaces .......................... 74
Fluidized-Bed Roasting ...................................... 77
1 8. PFODUCTION OF PURE MOLYBDENUM TRIOXIDE .................. 80
Distillation Method ......................................... 80
C h e m i c a l Methods ......................................... 81
Leaching ............................................. 81
Removal of copper and iron from the solutions .............. 83

.
iv

............................ 30
.......... 30
................ 31
..................................................... 33
.............. 34

Page Isolation of molybdenum from t h e ammoniacal solutions Extraction of molybdenum from the calcine-leaching
residues Extraction of molybdenum into the final product and
costing PROCESSING OF LOW-GRADE CONCENTRATES PRODUCTION OF METALLIC MOLYBDENUM Reduction of Molybdenum Trioxide with Hydrogen PRODUCTION OF SOLID. DUCTILE MOLYBDENUM Powder Metallurgy Method T h e Melting of Molybdenum

. 21 .

19 20

.......................................... 84
........................................... 86
............................................ 89
.................. 90
................... 92
............. 92
.............. 94
.................................. 94
................................. 95

................................ 99
................ 99
......................... 99
.......... 101
............................. 104
................. 106
.... 108
......................................... 109
..................................... 112

112

113

Chapter I11 TANTALUM AND NIOBIUM 22 GENERAL DATA ON TANTALUM AND NIOBIUM Properties of Tantalum and Niobium The Properties of Tantalum and Niobium Compounds Uses of Tantalum and Niobium 23 . MINERALS. ORES. AND ORE CONCENTRATES 24 . PROCESSING OF TANTALUM-NIOBIUM ORE CONCENTRATES 25 PROCESSING OF TANTALITE-COLUMBITE BY FUSION WITH
SODIUM HYDROXIDE 26 PROCESSING OF TANTALITE-COLUMBITE BY FUSION WITH
POTASSIUM HYDROXIDE 27 DECOMPOSITION OF TANTALITE-COLUMBITE WITH HYDRO-
FLUORIC ACID 28 . PROCESSING OF TITANIUM-NIOBIUM CONCENTRATES 29 SEPARATION OF TANTALUM AND NIOBIUM AND PURIFICATION OF THEIR COMPOUNDS Fractional Crystallization of the Complex Fluorides The Extraction Method Separation of Tantalum and Niobium b y the Chloride
Rectification Process Separation of Tantalum and Niobium by Selective Reduction
o f t h e Chlorides Separation of Tantalum and Niobium with the Aid of Ion
Exchange Resins 30 METHODS FOR THE PRODUCTION OF METALLIC TANTALUM
AND NIOBIUM 31 FUNDAMENTAL PRINCIPLES OF METALLOTHERMY 32 PRODUCTION OF TANTALUM AND NIOBIUM POWDERS BY
THERMAL REDUCTION WITH SODIUM 33 PRODUCTION OF TANTALUM AND NIOBIUM BY ELECTROLYSIS

.............................................. .........

....................................... 117
........... 118
.................................... 122

....................................

129

. .

......................................... 129
........................................ 130
.............................................. 131
.............. 131
......................... 135

137

Page Electrolytic Production of Tantalum Construction of t h e Electrolyzer and the Conditions of


Electrolysis Processing of the Cathodic Product 34 PRODUCTION OF NIOBIUM BY THE CARBIDE REDUCTION
METHOD 35 PRODUCTION OF TANTALUM AND NIOBIUM BY REDUCTION
OF THEIR CHLORIDES 36 PRODUCTION OF SCLID DUCTILE TANTALUM AND NIOBIUM The Powder Metallurgy Method T h e Melting of Niobium and Tantalum Vacuum Pumps and Vacuum Systems 37 RECOVERY OF METALLIC TANTALUM AND NIOBIUM WASTES

........................ .......................... . ..... 155


Chapter N . TITANIUM ............................................. 38. GENERAL DATA ON TITANIUM ................................ 155
Brief Historical Note ........................................ 155
Properties of Titanium ...................................... 155
T h e Properties of Titanium Compounds ........................ 158
Uses of Titanium ........................................... 162
39 . MINERALS. ORES. AND ORE CONCENTRATES .................. 164
40 . PRODUCTS OF THE PROCESSING OF TITANIUM CONCENTRATES.. 166
41 . REDUCTIVE SMELTING OF ILMENITE .......................... 168
42 . REDUCTION OF ILMENITE TO TITANIUM CARBIDE
170
(THE CARBJDIZATION PROCESS) ................................ 43 . PRODUCTION OF TITANIUM TETRACHLORIDE .................. 171
Reaction Mechanism ....................................... 171
Procedures ............................................... 173

Dust Collection an3 Condensation System PURIFICATION OF TECHNICAL GRADE TITANIUM CHLORIDE PRODUCTION OF TITANIUM DIOXIDE T h e Sulfuric Acid Method Production of Titanium Dioxide from Titanium Tetrachloride 46 PRODUCTION OF METALLIC TITANIUM 47 REDUCTION OF TITANIUM TETRACHLORIDE WITH MAGNESIUhl (THE K ROLL PROCESS) Physicochemical Principles of the Process Types of Reactors Used for Thermal Reduction with Magnesium T h e Reduction Process -Conditions and Control Processing of the Reduction Product T h e Technical and Economical Indexes of the Process 48 REDUCTION OF TITANIUM TETRACHLORIDE BY SODIUM Reduction 44 45

......................... ............................................. ........................... .................................................... ......................................... . ... .............................

137

138
140
142
144
145
145
149
152
154

. .

.... 176
........................ 180
................................. 180

.....................

175

. .

.......................

...

183
185
186 186 189 190 192 194 194 196






........................................ .................... .. ............... .......................... ......... ........ ...............................................

vi

Pages Processing of the Reaction Mixture Two-Stage Thermal Reduction with Sodium 49 . REDUCTION OF TITANIUM DIOXIDE Reduction by Calcium Reduction by Calcium Hydride 50 ELECTROLYTIC REFINING OF TITANIUM AND TITANIUM-
BASED ALLOYS 51 REFINING OF TITANIUM BY THE IODIDE PROCESS Equipment Procedure 52 PRODUCTION OF SOLID TITANIUM The Melting of Titanium T h e Melting of Titanium Alloys Production of Solid Titanium by Powder Metallurgy
Techniques

.
.

....................... 197
............... 198
....................... 199
................................... 199
........................... 201
........................................... 202
............ 204
............................................. 204
............................................. 206
....................... 208
............................... 208
.......................... 210
.......................................... 211

.......................................... ........................... ................................. .................................... ...................................... ........................................ ................


213
213
213
215
217
220
220
222
223
223
226

227
228
228
230

Chapter V ZIRCONIUM 53 GENERAL DATA ON ZIRCONIUM Properties of Zirconium Chemical Properties Uses of Zirconium Uses of Hafnium 54 MINERALS. ORES. AND ORE CONCENTRATES 55 PRODUCTS FORMED IN THE PROCESSING OF ZIRCON
CONCENTRATES 56 METHODS FOR THE DECOMPOSITION OF ZIRCON
CONCENTRATES 57 DECOMPOSITION OF ZIRCON BY FUSION WITH SODIUM
HYDROXIDE 58 DECOMPOSITION OF ZIRCON BY FUSION WITH LIME 59 ISOLATION OF ZIRCONIUM FROM HYDROCHLORIC AND SULFURIC ACID SOLUTIONS Isolation of Zirconium Oxychloride Hydrolytic Precipitation of Basic Sulfates Isolation of Zirconium as the Zirconylsulfuric Acid Hydrate 60 PROCESSING OF ZIRCON BY SINTERING WITH POTASSIUM
FLUOSILICATE 61 DECOMPOSITION OF ZIRCON BY REDUCTION WITH CARBON
TO CARBIDE OR CARBONITRIDE 62 PRODUCTION OF ZIRCONIUM TETRACHLORIDE Chlorination of Zircon Concentrates Chlorination o f Zirconium Carbonitride Chlorination of Zirconium Dioxide

. .

. .
. .

.......................................... .......................................... .............................................. ........


............................... ....................... ................. ...

. .
.

............................................. .............................. .............. ...................... ................... .......................

231
233
234
234
235
236

Page

............................................... 238
............. 238
................................ 240
.................................. 243
......................................... 244
.......................... 245
. ........................... 245
. .............................................. 246
................................... 247
................................. 249
................................... 250
. ... 252
. ........................................ 252
. ............ 253
. ............ 255
. ...................... 256
Part Two . T H E R A R E - E A R T H M E T A L S Chapter VI . THE RARE-EARTH METALS (LANTHANIDES) ............... 258
7 1. GENERAL DATA ON RARE-EARTH METALS .................. 258
The Place of Lanthanides i n the Periodic System and Their
Electron Structure ...................................... 258
Brief Historical Note on the Discovery of the Lanthanides ...... 261
Physical Properties ...................................... 261

Chemical Properties T h e Properties of Lanthanide Compounds Isolated from


Solutions Uses o 6 Rare-earth Metals OCCURRENCE PROCESSING OF MONAZITE CONCENTRATES T h e Sulfuric Acid Method T h e Alkaline Method SEPARATION OF LANTHANIDES Fractional Crystallization Fractional Precipitation Selective Oxidation Selective Reduction Separation with Ion Exchange Resins

METHODS FOR THE SEPARATION OF HAFNIUM AND ZIRCONIUM Fractional Crystallization of Complex Fluorides Separation by Extraction Ion Exchange Methods Rectification Selective Reduction of Chlorides 64 MANUFACTURE OF ZIRCONIUM 65 THERMAL REDUCTION OF ZIRCONIUM CHLORIDE WITH
MAGNESIUM T h e Separate Process The Combined Process T h e Vacuum-Thermal Distillation of Magnesium and
Magnesium Chloride 66 REDUCTION OF POTASSIUM FLUOZIRCONATE BY SODIUM 67 REDUCTION OF ZIRCONIUM DIOXIDE BY CALCIUM AND
CALCIUM HYDRIDE 68 PRODUCTION OF ZIRCONIUM BY ELECTROLYSIS 69 THE THERMAL DISSOCIATION (IODIDE) METHOD 70 PRODUCTION OF SOLID ZIRCONIUM 63

.....................................

263

72 73

. .

74

............................................. ............................... ........................................... .............. ................................ ................................... ........................... ................................ ................................. ..................................... .................................... .......................

265
267
270
272
273
275
278
278
280
280
281
282

Viii

Page Extraction Separation Scheme 75 CONTROL OF SEPARATION AND PURITY OF LANTHANIDE


COMPOUNDS 76 MANUFACTURE OF RARE-EARTH METALS Starting Compounds for t h e Manufacture of Metals Materials for the Smelting of the Rare-Earth Metals Electrolytic Production of Rare-Earth Metals Metallothermic Preparation of the Lanthanides. Reduction of halides by calcium Reduction of oxides with simultaneous distillation of
the m e t a l

.
.

........................................... ................................... ............................................. ................ ........ ........ ............... ............. ....................


.......................................

288
290
291
292
292
294
294
297
298 300

Chapter VI1 GERMANIUM 302 302 77 GENERAL DATA ON GERMANIUM 302 Properties of Germanium 306 The Properties of Germanium Compounds 307 Uses of Germanium 309 78 OCCURRENCE Behavior of Germanium During Processing of Sulfide
310 Raw Materials 312 Behavior of Germanium in the Processing of Coals 79 EXTRACTION OF GERMANIUM FROM VARIOUS RAW MATERIALS
312 312 Primary Processing of Germanium-Containing Products Production of Technical Grade Germanium Tetrachloride
313 from t h e Concentrates 80 EXAMPLES OF PROCESSES FOR THE EXTRACTION OF
314 GERMANIUM FROM RAW MATERIALS Extraction of Germanium from the Dusts formed in t h e
314 Agglomeration Roasting of Zinc Concentrates Extraction of Germanium from the Copper Concentrates of
314 Tsumeb Extraction of Germanium from the Dusts Formed in the Shaft
318 Smelting of Copper Concentrates from Katanga Extraction of Germanium from the Dust of Gas- Producing
319 Plants 81 PURIFICATION OF GERMANIUM TETRACHLORIDE AND 321 PRODUCTION OF GERMANIUM DIOXIDE 321 Purification by Rectification 322 Purification by Extraction with Hydrochloric Acid 323 Production of Germanium Dioxide

Part Three

THE SCATTERED METALS

.........................................
..........................
.................................
.................
....................................
...........................................
.......................................
..........
......

.................................

........................
.............
.............................................
............
..............................................
......................
.............................
...........
.......................

i x

Page

........................................... ........ .............................................. 341


............... 342
......................... 342
. .................... 345
..................... 346
............ 346
........... 347
. .............................. 347
........................
348 350 Chapter M . INDIUM ..............................................
350 90 . GENERAL DATA ON INDIUM ...............................
350 Properties of Indium ....................................
351 T h e Properties of Indium Compounds ......................
352 Uses of Indium .........................................
353 9 1. OCCURRENCE ............................................
92 . BEHVAIOR OF INDIUM IN THE PRODUCTION OF ZINC AND 354 LEAD .....................................................
354 Pyrometallurgical Production of Zinc ......................
355 Hydrometallurgical Production of Zinc ....................
356 Production of Lead ......................................
93 . EXTRACTION OF INDIUM FROM INDIUM-ENRICHED 357 PRODUCTS ...............................................
358 Production of Indium Concentrates ........................

OCCURRENCE Behavior of Gallium in t h e Production of AIuminum PREPARATION OF GALLIUM CONCENTRATE FROM ALUMINATE
SOLUTIONS Solutions Obtained by t h e Soda-Lime Method Solutions From t h e Baier Process 88 EXTRACTION OF GALLIUM FROM THE WASTES OF THE ELECTROLYTIC REFINING OF ALUMINUM Precipitation of Gallium Ferrocyanide Precipitation of Gallium with Organic Reagents Purification of Gallium Compounds by Extraction 89 PRODUCTION OF GALLIUM Production of Very Pure Gallium
86

............................................ ........................... ............................ ................ .................. ....................................... ..................... Chapter VI11. GALLIUM ........................................... 85 . GENERAL DATA ON GALLIUM ............................ Properties of Gallium .................................. T h e Properties of Gallium Compounds ..................... Uses of Gallium ........................................
. .

Precautions Used t o Obtain High-Purity Germanium


Dioxide 82 PRODUCTION OF GERMANIUM 83 PURIFICATION OF GERMANIUM AND PREPARATION OF GERMANIUM MONOCRYSTALS Purification by Fractional Crystallization Production of Germanium Monocrystals Quality Control 84. PROCESSING OF GERMANIUM WASTES

323
324
325
326
331
334
334
336
336
336
337
338
339
340

. 87.

...............

.- ................... ..,

Page

.............................. 360
.......................................... 361
.............................. 364
9 4. .................................. 364
................................ 364
..................................... 366
..................................... 367
.................. 368
Chapter X . THALLICM ............................................. 369

Production of Crude Indium Examples of Technological Processes for the Production of Crude Indium REFINING OF CRUDE INDIUM Selective Cementation Electrochemical Methods C h e m i c a l Methods Vacuum Distillation Zone Melting and Rod Drawing From Melt 95

..... ................. ................. ....................... 98. .......................... Chapter X I. RHENIUM ............................................. 382
99. GENERAL DATA ON RHENIUM .............................. 382
Brief Historical Note ..................................... 382
382
Properties of Rhenium .................................... 384
T h e Properties of Rhenium Compounds ..................... 386
Uses of Rhenium ........................................ 100. OCCURRENCE ............................................ 388
Behavior of Rhenium in the Processing of Molybdenite
Concentrates .......................................... 389

Behavior of Rhenium in t h e Production of Copper EXTRACTION OF RHENIUM FROM THE VARIOUS WASTES FORMED IN THE PROCESSING OF MOLYBDENITE CONCENTRATES Extraction of Rhenium from the Electrostatic Filter Dust Sorption of Rhenium from Solutions of Low Rhenium Content 102 PRODUCTION OF RHENIUM Reduction of Potassium Perrhenate with Hydrogen Reduction of Ammonium Perrhenate with Hydrogen Production of ammonium perrhenate Reduction Reduction of Rhenium Dioxide by Hydrogen
101

96 97

GENERAL DATA ON THALLIUM Properties of Thallium T h e Properties of Thallium Compounds Uses of Thallium OCCURRENCE EXTRACTION OF THALLIUM FROM INDUSTRIAL WASTES Pyrometallurgical Production of Thallium Hydrometallurgical Production of Thallium Examples of Technological Processes PRODUCTION OF PURE THALLIUM

............................ ................................... ..................... ....................................... ............................................

369
369
369
371
372
373
373
373
375
381

............. 390

. 390
...... 391

.............................................. 393
................................. 396
............ 396
......... 397
................. 397
....................................... 398
................. 399

xi

Page Reduction of Potassium Perrhenate i n a Solution by Hydrogen Under Pressure Production of Rhenium Powder by Electrolysis T h e r m a l Dissociation of the Halides PRODUCTION OF SOLID RHENIUM

103

................................ .............. ...................... ..........................

399
400
400
401

. THE LIGHT R A R E METALS Chapter XII. BERYLLIUM ........................................... 404


404
104. GENERAL DATA ON BERYLLIUM ............................ Properties of Beryllium ................................... 404
T h e Properties of Beryllium Compounds .................... 407
Uses of Beryllium ....................................... 410
105. BERYLLIUM MINERALS AND ORES ........................... 412
1 0 6. ENRICHMENT OF BERYLLIUM ORES .......................... 413
107. METHODS OF PROCESSING BERYL CONCENTRATES ............ 414
T h e Sulfate Process ..................................... 415
Decomposition of Beryl Concentrate by Sintering With Sodium Fluosilicate ( T h e Fluoride Process) ....................... 417
Production of Beryllium Chloride From Beryllium Oxide ...... 421
424
1 0 8. PRODUCTION OF METALLIC BERYLLIUM .................... Electrolysis of Beryllium Chloride ......................... 424
Thermal Reduction of Beryllium Fluoride With Magnesium .... 426
429
109. VACUUM MELTING OF BERYLLIUM .......................... 110. PRODUCTION OF SOLID BERYLLIUM BY POWDER METALLURGY
TECHNIQUES ......................................... 431
111. PURIFICATION OF TECHNICAL GRADE BERYLLIUM ........... 434
112. PRODUCTION OF BERYLLIUM ALLOYS ......................... 434
113 . TOXIC PROPERTIES AND SAFETY MEASURES .............. 431
Chapter XIIL LITHIUM ........................................... 439
114. GENERAL DATA ON LITHIUM ............................... 439
Properties of Lithium .................................... 439
T h e Properties of Lithium Compounds ...................... 441
Uses of Lithium ......................................... 445
447
115. LITHIUM MINERALS AND ORES ............................. 116. ENRICHMENT OF LITHIUM ORES ............................ 448
117. PROCESSING OF LITHIUM CONCENTRATES .................. 450
T h e Sulfuric Acid Method ................................ 450
453
T h e Lime Method ...................................... T h e Sulfate Method ...................................... 45.5
Chlorination Roasting .................................... 458
460
118. PRODUCTION OF LITHIUM .................................
Part Four

-.

.........................

............................. 'I

Page

........
460 ....
463 119. ............................... 465
BIBLIOGRAPHY ................................................... 467

Production of Lithium by Electrolysis of Molten Salts Production of Lithium by Thermal Reduction With Metals PURIFICATION OF LITHIUM

EXPLANATORY LIST OF ABBREVIATIONS OF U. S. S. R. INSTITUTIONS AND JOURNALS APPEARING IN THIS TEXT Abbreviation Full name (transliterated) Akademiya Nauk Kazakhskoi SSR Izdatel'stvo Gosudarstvennogo Komiteta Soveta Ministrov SSSR PO kpol'zovaniyu Atomnoi Energii Doklady Akademii Nauk SSSR Gosudarstvennoe Izdatel'stvo Fiziko- Matematicheskoi Literatury Germanskaya Demokratiches kaya Respublika Gosudarstvennyi Nauchno ksledovatel'skii i Proektnyi Institut Redko metallicheskoi Promysh lennost i Gosudarstvennoe NauchnoTekhnicheskoe Izdatel'stvo Khimicheskoi Literatury Izdatel'stvo Akademii Nauk SSSR Izvest iy a Kirgizskaya SSR Leningradskii Gosudarst vennyi Universitet Nauchno- Issledovatel'skii Institut Mekhanicheskoi Obrabotki Poleznykh Iskopaemykh Translation Academy of Sciences of t h e Kazakh SSR Publishing House of the State C o m m i t t e e of the Council of Ministers of t h e USSR on Uses of Atomic Energy Transactions of t h e Academy of Sciences of the USSR State Publishing House of Physical and Mathematical Literature German Democratic Republic State Scientific- Research and Planning Institute of Rare-Metals Industry

AN KazSSR
Atomizdat

DAN SSSR Fizma tgiz

GDR Giredmet

Goskhimizdat

State Scientific and Technical Publishing House of Chemical Literature Publishing House of the Academy of Science of the USSR Bulletin Kirgiz SSR Leningrad State University Scientific Research Institute for Mechanical Concentra ion of Minerals

Izd. A N SSSR

Izv.
Kirg SSR
LGU
Mekhanobr

XiV

__

.-.

.-

. -.

..

- .....

I I

.....

I...

.."

I . . . 1 . # . . # , 1 1 , ,

I,

I I I

Abbreviation MGU NTOTsM

Full n a m e (transliterated) Moskovskii Gosudarstvennyi Universitet Nauchno-tekhnicheskii Otdel Tsvetnoi Metallurgii Referativny i Zhurnal Sibirskoe Otdelenie Akademii Nauk SSSR Tsentral'nyi NauchnoIssledovatel'skii Institut Chernoi Metallurgii imeni J. P. Bardina Vsesoyuznyi Tsentral'nyi Nauchno- Issledovatel'skii Institut Tsvetnoi Metallurgii Ukrainskaya SSR Vsesoyuznyi Institut Nauchnoi i Tekhnicheskoi Informatsii Vysshee Uchebnoe Zavedenie Zhurnal Fizicheskoi Khimii Zhurnal Neorganichesko i Khimii Zhurnal Obshchei Khimii Zhurnal Prikladnoi Khimii

Translation Moscow State University Scientific and Technical Division of Nonferrous Meta1lu r gy Abstract Journal Siberian Department of t h e Academy of Sciences of t h e USSR Central Scientific-Research Institute of Ferrous Metallurgy im. J. P, Bardin All-Union Central Scientific Research Institute of Nonferrous Metallurgy Ukrainian SSR All-Union Institute for Scientific and Technical Information Higher Educational Institution Journal of Physical Chemistry Journal o f Inorganic Chernistry Journal of General Chemistry Journal of Applied Chemistry

RZh
SO AN SSSR

TsNIIChM

TsNIlTsM

UkrSSR VINITI

vu2
ZhFKh ZhNKh ZhOKh ZhPKh

xv

FOREWORD Rare metals play an important role in the development of m a j o r branches of industry, such a s vacuum equipment, semiconductor electronics, nuclear power and rocket production, a s well as in the production of special s t e e l s and hard, refractory and corrosion-resistant alloys. Rapid development and improvement in the production of r a r e metals took place in the ten y e a r s which have elapsed since the publication of the f i r s t edition of this book. These ten y e a r s have witnessed the beginning of l a r g e - s c a l e production of titanium, zirconium, and germanium, and a significant increase in the production volume; new, improved methods f o r the separation and purification of metals and compounds (ion-exchange, extraction, crystallization methods) a s well a s a r c and electron-beam melting p r o c e s s e s f o r metals were developed. This made it necessary t o rewrite most of this book. In view of the growing importance of the lanthanides and rhenium, chapters on these metals were also included. At the same time, we decided to dispense with the chapters on lead and antimony, sin.ce these a r e not usually listed a s r a r e metals. In describing the metallurgy of each metal, much attention was paid to i t s physicochemical nature and to the practical operations involved in the main technological p r o c e s s e s f o r the production of i t s chemical compounds and of the pure metal. This book is a textbook for students specializing in the metallurgy of the r a r e metals. It is assumed that the student is familiar with the physicochemical fundamentals .of metallurgy, o r e dressing, metallurgical furnaces, and p r o c e s s e s and apparatus used in extractive metallurgy. The description of standard equipment (leaching apparatus, thickeners, f i l t e r s , comminution installations, etc. ) has accordingly been omitted. The r e f e r e n c e s a r e grouped together at the end of the book. Chapters I, 111, IV, V, VI, VII, VIII, and XI1 were written by A. N. Zelikman, Chapters 1 1 , IX, X, and XI - by A. N. Zelikman and 0. E. Krein, and Chapters XIII and XIV - b y G. V. Samsonov. The authors wish to thank their many colleagues at the State Institute of Rare Metals (Giredmet). They a r e especially grateful to Professor G. A. Meerson f o r his valuable advice and help in the preparation of the manuscript and Associate P r o f e s s o r L. V. Belyaevskaya, the editor of this book.

xvi

INTRODUCTION At the beginning of the 20th century, certain chemical elements found important industrial applications f o r the first time. This group of elements is known a s " r a r e elements" or, in a m o r e narrow sense, " r a r e metals". As may be s e e n in Table 1 (in which the rare metals appear in bold face frames) the "rare metals" group comprises metals in all groups of the periodic table. Their physicochemical properties a r e accordingly very different, and the r a r e metals group comprises those metals which for a number of reasons have only recently found commercial use. The main reason for it is that most r a r e metals were only discovered at the end of the 18th and in the 19th centuries. At the same time, their scarcity and the fact that many of the r a r e metals a r e widely dispersed over the Earth's crust, a s well a s the difficulties encountered in the extraction and purification of some of them, also interfered with their industrial utilization. Table 2 (compiled by Academicians V. I. Vernadskii and A. E. Fersman) shows the use of chemical elements by man, in various historical periods. The use of a l l metals in the " r a r e metals" group dates from the 19th and 20th centuries. Thus, the concept " r a r e metal" originally r e f e r r e d to metals which had little o r no technological use. In our own time, however, many r a r e metals a r e extensively used in modern technology. The very existence of a number of branches of technology would be unthinkable without the use of r a r e metals. Of the 70 metals (listed in Table 3) in production at the present time, forty-one belong to the r a r e - m e t a l s groups. Thus, the "rare-metals" group i s not a scientific classification but has a historical meaning. It is a common e r r o r to associate the t e r m " r a r e metal" with a low abundance of the element in nature. To c o r r e c t this misconception, we shall give some recent data on the abundance of elements in the E a r t h ' s crust. Proceeding from analytical data, geochemists calculated the composition of the E a r t h ' s crust, down to a depth of 16 km. The f i r s t tables showing the composition of the Earth's c r u s t were compiled in 1889 by the American scientist F. Clark. Their accuracy was subsequently improved. An important contribution to the study of the abundance of elements in the E a r t h ' s c r u s t was made by the Soviet scientists, V. I. Vernadskii, A. E. Fersman, V. G. Khlopin, and A. P. Vinogradov. In accordance with the suggestion of A. E. Fersman, the average contents of the elements in the E a r t h ' s c r u s t a r e designated as "clarkes". The average contents of the elements, in weight percent, are shown in Table 4. The f i r s t row includes elements whose abundance is above lo%, the second - those with abundances between 1 and 10% the third from 0.1 to 170, etc. It is evident that the distribution of elements in

xvii

TABLE I

PERIODIC TABLE OF ELEMENTS

4.0026

39 TU2

4008

I 41.956 1

47.90

50942

51.996

54.jWr

55.847

I 9.9332

132.905

137.34

138.91

178. 1g

1 180.948

183.85

jl

1863

1 1m2

I192.2
LANTHANIDESIRME-EMTH ELEMENTS)

the E a r t h ' s c r u s t is v e r y nonuniform. The nine most abundant elements '0 ' of the total amount of m a t t e r in the E a r t h ' s crust, i. e., account for 98.13 7 the remaining 83 elements account f o r only 1.8770 of it.
TABLE 2 T h e increasing use of chemical elements (both in the free state and as compounds) Period Ancient times Elements N. Al. A L , K. C a , 0, Si. C l , Na, Pb.. Ae. C . C1. ~. . Zn.. Sb 'oral number ,f elements
~

s,
0.
I,

-I,

~~

2H &

19

U p to the 18th century

T h e 18th century
T h e 19th century

In addition t o the above: As, Mg. B i , Co. B . N i , P


I,

26

I*

"

H. Pt, Ir, J

30

I,

,,

I,

.I

"

Os, Pd,

@, Sr,

m, w, m, m,a,
Ba. Br.
F,

Cd, Mn. Cr,

62 I ,

and

lanthanides ( 1 5 elements) 20th century u p to 1915

Rh

20th century up to 1932

lanthanides (15 elements)

m,

69 Ar,

@,

m,

82

U p to 1957
N o t e : the metals used since ancient times are underlined; the rare metals are boxed in. TABLE 3 Metals produced today* ( t h e rare metals are underlined)
Group in the periodic table

Elements Li. Na. K . Rb, Cs. C u . Ag. Au

Tota 1 number
8

Rare metals
3

I
11 111

Be.
Mg.

--

Ca.Sr, Ea, Ra. Zn, Cd, Hg


T1

2
6
4

IV

---V, Nb. Ta. -Mn. R e

- - - - L L Ti. Zr. Hf. Ge. Sn, Pb

Al. Sc. Y. La, Ga. I,

7
6 6 6

VI
VI1 VI11

As. Sb. B i

3
5 1

Cr. Mo. W . Se, T e . Po

-----

Total.

2 9 14
3 70

Fe, Co. Ni. Ru, Rh. Pd, Os, It, Pt From No. 58 (Ce) to No.71 (Lu)

14

Lanthanides Actinides

--Th. U, Pu

3
41

Produced in the form of m e t a l , alloy or c h e m i c a l compound.

XiX

TABLE 4 Average ch em i cal composirion of the Earth's c r ~ s (according t to Vinogradov) (thickness 1 6 k m . excluding the atmosphere and the ocean) Decade Abundance limits, wt?
-

Chemical el em en t s and their occurrence in wt%

10-50 1-10

S i

47.2 27.6

A1

Fe

Ca

Na

Mg

8.80 5.10-~ 3.60 2.64 2.60 2.10

Ill

1 1 0 1

1v

Ti H C 0.60 (0.15) 0.10 __-Mn P S Ba C I Sr Rb 0.09 0.08 0.05 0.05 0.045 0.04 0.031 Zr Cr V Cu N 0.020 0.02 O.Oj5 0.01 0.01
~

F
0.027

lo-*-

10-3

N i

Li

Z n

ce

Sn

Co

8.10 6.5.10d3 5.10-3 4.5.10-3 4.10-3 3.10-3 Cia Pb Nd La Y 2.8.1(r3 2.5.10-3 1.8.10-3 1.6.10-3 1.5.10-3
Nb Gd

1.10-~
VI

1.1073

Cs 7.10-4 As

10-3-10-4

Th

Pr
7.10-4

6 . 1 0 ' Dy E r Yb U Sc . 1 0 ' 3.10-4 3.10-' 6 . 1 ~ - ~5.11)-~ 4.5.10-4 4 Br Ho Mo H f B T I 3.10-4 3.10-4 3.2.10-4 3.10-4 1.6.10-4 1.3.10-'

8.1Oy

Sm 7.10-4

Ge

Be

7.10-'

Eu
VI I
0-4-10-5

Lu

--VI11

IX

1.2.10-4 1.10-4 1.10-~ Tu Se Cd Sb I Bi 8.10-5 6.10-5 (5.lOP5) (4.10-5)(3.10-5) ( 2 . 1 0 ' ) Ag In (1.10-~) (1.10-5)
~~

Table 4 also shows that the abundance of most r a r e elements is low. However, many of the r a r e metals a r e much m o r e abundant than metals known to man f o r a long time. F o r instance, titanium occupies the ninth place in the o r d e r of abundance; zirconium, vanadium, lithium, cerium and some other r a r e metals a r e m o r e abundant than commonIy used metals such a s lead, arsenic, tin, mercury, s i l v e r and gold. Moreover, some metals a r e regarded a s very r a r e because of the manner in which they a r e dispersed, i. e . , because they do not form minerals or deposits. F o r example, the concentrationof gallium in the Earth's c r u s t is higher than the concentrations of antimony, arsenic and mercury.

xx

____

.-__..._ --..._.. .... ....


I

11.11.1.1.11

.....

I . . . .

11..."...11..1....

I . . .

1.11,

. I . . . I , ,

However, gallium does not occur a s minerals but is found in the lattices of other minerals, while antimony, arsenic and mercury form minerals and deposits, i. e . , occur in a m o r e concentrated form ar,d thus appear to be more abundant. A low abundance in the E a r t h ' s crust, therefore, is not a property of all the r a r e metals, but only of most of them.

Classification of r a r e metals Rare metals a r e usually classified in five groups (Table 5), the c l a s s i fication being based on similarities in physicochemical properties, methods of extraction and production of the metals, and some other characteristics.
TABLE 5 Technological classification of the rare metals Group in the periodic table I I1 IV
V

Elements Lithium, rubidium, cesium


Beryllium
Titanium ,zirconium, hafnium
Vanadium, niobium, tantalum
Molybdenum, tungsten
(Rhenium)*
Gallium, indium, thallium
Germanium
Selenium, tellurium
Rhenium
Scandium, yttrium. lanthanum and lanthani' (14 elemeiits from cerium to lutecium) Radium Actinium and actinides (thorium. protactinii uranium and transuranium elements) Polonium

Technological classification Light rare metals Refractory rare metals

VI VI1 I11 IV VI VI1 I11

Dispersed rare metals Rare earth rare metals

I1 I11

Radioactive rare metals

Rhenium is a typical dispersed e l e m e n t but because of its properties i t may also he classified as a refractory metal.

A description of these groups i s given below. Light rare metals. This group comprises the metals in Groups I and I1 of the periodic table (except radium). They have a low density (lithium 0.53, beryllium 1.85, rubidium 1.55, and cesium 1.87 g/cm3) and a r e very reactive. Their chemical compounds (oxides, chlorides) a r e very stable and a r e reduced to the metal only with difficulty.

Like nonferrous light metals (aluminum, magnesium, calcium) light r a r e metals are prepared by electrolysis of their molten salts o r by reduction with other metals. Refractory rare metals. The general physicochemical properties of these metals a r e determined by their place i n the periodic system [sic]. All are transition elements in Groups IV, V, and VI of the periodic table in which the d-electron sub-levels a r e being filled. This feature of their atomic s t r u c t u r e determines many of their physical and chemical properties. They are refractory (their melting points lie between 1660' for titanium and 3400" f o r tungsten), they are harder than other metals, and they have a high resistance to corrosion. They may display s e v e r a l valencies, which accounts for the wide variety of their compounds. All refractory metals form refractory, hard, chemically stable compounds with a number of nonmetals. Of these compounds, the carbides, nitrides, borides, and silicides have important commercial u s e s Because of the high melting point of the refractory metals, they a r e extensively produced by sintering methods; a r c and electron-beam melting methods have been recently developed for some of them. Because of the similarity in the properties of the refractory metals, they have many common uses. Thus, all refractory metals a r e used a s alloys with s t e e l ana a s carbides in hard alloys. Many refractory metals a r e used in electrical technology and in vacuum equipment manufacture. Dispersed r a r e m e t a l s . All these metals occur in a dispersed state in the E a r t h ' s crust. Most elements in this group do not form separate minerals ( o r else such minerals a r e r a r e ) . The dispersed elements a r e usually found a s isomorphous admixtures in very s m a l l concentrations in the lattices of other minerals, and a r e extracted as by-products from the wastes formed in metallurgical and chemical processing. Thus, gallium is found in aluminum minerals (bauxites and others) and is extracted from the intermediate and waste products of aluminum processing; indium, thallium, and germanium a r e often encountered in zinc blende and other sulfide minerals and a r e extracted as by-products during processing of sulfide ores. Germanium is often found in coals and is extracted from coal wastes. Rhenium occurs in molybdenum o r e s and is obtained a s a by-product of the processing o f molybdenum raw materials. Thus, scattered metals a r e produced from a wide variety of raw materials, including dust from furnaces used in calcination, s l u r r i e s from copper refining plants, dusts and cakes from zinc-lead industry, s l u r r i e s f r o m sulfuric acid plants, coal ash, etc. The production of these metals is closely related to the metallurgy of the common non f e r r o u s metals. The rare earths (lanthanides). The similarity of the physicochemical properties of the lanthanides (from cerium No. 58 to lutecium No. 71) is 1ttribute.d to the s i m i l a r structure of the outer electron shells of their atoms; the elements in the lanthanide s e r i e s correspond to the filling of the inner 4 f - s h e l l . The elements lanthanum, scandium, and yttrium which belong to Group I11 and which a r e usually included in the lanthanide

group have properties s i m i l a r to those of the lanthanides. The lanthanides always occur together in o r e s ; in the first stages of processing they a r e isolated a s a mixture of oxides or other compounds. Interest has recently been exhibited in the isolation of the individual lanthanides. The difficult problem of the separation of lanthanide metals has now been satisfactorily solved. Radioactive r a r e metals. This group comprises the naturally radio acti.ve elements: polonium, radium, actinium, and the actinides (thorium, protactinium, uranium, and the artificially produced transuranium elements - neptunium, plutonium, etc. ). The technology of these metals, the methods of their handling and their use are determined to a g r e a t extent by their radioactivity. In the actinide s e r i e s , the elements a r e formed by successive addition of electrons to the inner 5 f-shell. In this respect, the actinides resemble the lanthanides. The natural radioactive elements occur together in ores. They a r e often accompanied by lanthanides. Both natural and artificial radioactive elements play an important role in atomic power production.

Certain features of the production of r a r e metals from o r e s The technological p r o c e s s e s used in the metallurgy of r a r e metals depend on the nature of the raw material, the physicochemical properties of the metals, and the standards required in commercial production. 1. The o r e s containing r a r e metals a r e usually lean ores. Hence, the enrichment of the o r e s i s of particular importance. The r a r e metal o r e s often have a complex composition. The following o r e s may be taken a s an example: tungsten-molybdenum; titanium-niobium- tantalum lanthanide-containing o r e s ; uranium-vanadium; lithium-cesium; tungsten-tin; zirconium- niobium and other o r e s . Moreover, as has been mentioned above, some of the r a r e metals often occur at negligible concentrations in o r e s and in the wastes of f e r r o u s and nonferrous metal processing and of chemical industry. As a result, multiple processing of raw materials, with the recovery of all valuable components, becomes of considerable importance. 2. While the raw materials have a complex composition, industry s e t s very high requirements with respect to the purity of the metals. In the finished product, the maximum permissible concentrations of impurities a r e often of the o r d e r of a few hundredths, a few thousandths, and some times a few ten-thousandths of one percent. In some c a s e s the task becomes m o r e difficult because of the need to separate elements with very s i m i l a r properties (e. g . , hafnium from zirconium, niobium from tantalum, isolation of individual lanthanides in a pure state). As a result, the separation of impurities and the production of highpurity compounds plays an important p a r t in technological processing of the raw materials. 3. None of the r a r e metals can be smelted directly from the ore. The o r e concentrate must f i r s t be converted to pure chemical compounds which a r e used a s the raw m a t e r i a l for the production of the metal. The

processing of the o r e concentrate usually includes three main stages: a) decomposition of the concentrate; b) preparation of pure chemical compounds (oxides, salts) ; c) production of the metal from the chemical compound.
The object of the f i r s t stage is to decompose the mineral, to separate
the extracted metal from the bulk of the accompanying elements, and to concentrate it in solution o r in the precipitate. This is accomplished either by pyrometallurgical process e s (roasting, smelting, distillation, etc. ) or by hydrometallurgical methods (processing with acids, alkali solutions, etc. ). The second stage consists of the isolation and purification of chemical compounds and is characterized by the major p a r t played by chemical processes in aqueous solutions (precipitation, crystallization, extraction, ion exchange, etc. ). Pyrometallurgical processes (distillation of chlorides or oxides, fractionation) a r e occasionally used for the production of pure compounds. The third stage - preparation of pure r a r e metals - m a k e s use of various (mainly pyrometallurgical) processes for the reduction of the chemical compounds. In accordance with the reduction methods used, r a r e metals may be classified into three groups (Table 6 ) .

TABLE 6 Methods for the production of the rare metals Method of reduction Isolation from aqueous solutions by cementation or electrolysis Reduction of oxides or salts with hydrogen, carbon monoxide or carbon, at high temperatures Reduction of oxides or salts with m e t a l (metallothermy) or by electrolysis in molten media

Metal Gallium, rhenium. indium, thallium Tungsten, molybdenum, rhenium. germanium Tantalum, niobium, vanadium, titanium, zirconium, lithium. beryllium, lanthanides, thorium, uranium,

As is evident from the table, only four of the r a r e metals (gallium, indium, thallium and rhenium) can be isolated directly from aqueous solutions of their s a l t s while the remaining r a r e metals a r e produced by pyrometallurgical methods. In most c a s e s the refractory metals a r e f i r s t obtained a s a powder o r a spongy porous mass. The metal ingots a r e produced from these sponges by a r c o r electron-beam melting in a cooled copper crucible o r by powder-metallurgy techniques. The production of a number of r a r e metals in a pure state became possible due to the development of vacuum techniques.

xxiv

Development of the r a r e - m e t a l industry in the USSR


No rare-metal industry existed in Russia before the 1911 Revolution. Nevertheless, leading Russian scientists forecast the great potentialities of various rare-metal ore deposits in the enormous area of Russia. Important studies on the occurrence of scattered rare metals (indium. thallium, rubidium, and cesium) in Russia were conducted by V. I. Vernadskii in the years 1909-1915; they were a valuable contribution to the geochemistry of those elements. V. I. Vernadskii announced in 1916 that Russia had raw materials for t h e production of vanadium, lithium, lanthanum, cerium. thorium, boron. bismuth, cadmium. molybdenum, titanium, tin, radium, selenium, uranium. cesium, and zirconium. None of these metals was produced in Russia a t that time. Mendeleev's periodic table of elements played a most important part in the development of the chemistry of the rare metals. T h e periodic table m a d e i t possible to predict undiscovered elements and their properties and thus facilitated the prospecting work. T h e properties of gallium. germanium and scandium were brilliantly predicted by Mendeleev in 1811, before their discovery. A number of new metallurgical processes which were subsequenrly applied to in the production of rare metals were developed by Russian scientists before the revolution. These include the powder.metallurgy (sintering) methodused in the production of refractory metals (tungsten, molybdenum. tantalum. and niobium). T h e method was developed in 1826 by the Russian metallurgist P. G. Sobolevskii who used it in the manufacture of platinum products. In 1860-1865 Beketov laid the scientific foundations for a very important method of metal production thermal reduction. He discovered the aluminothermal method and was the first to use if in the production of barium, potassium, rubidium and cesium. Thermal reduction is used today i n the production of many rare metals and their alloys. Some attempts were made by Russian engineers and workers to organize small-scale production of certain rare metals and their compounds and alloys even before the Revolution. For example. molybdenum and tungsten steels w e r e produced in 1896 in the Putilov plant under the guidance of Prof. Lipin. and their properties were investigated, Further steps in the organization of the production of rare metals and their alloys were taken in Russia in the war years 1914-1918. During that period. a beginning was m a d e in the small-scale exploitation of tungsten ores from the Urals and the Transbaikalia and of molybdenum ores from the Chikoi deposits, and the production of ferrotungsten was organized a t the Obukhov plant in Petrograd and at the Motovilikha plant in Perm. T h e Izhora plant began the production of special molybdenum steels. The producrion of small amounts of selenium and tellurium compounds from the wastes of sulfuric acid plants was begun in 1916. T h e production technology of lithium compounds from lepidolite from Russian deposits was developed under the guidance of V . G. Khlopin. At the same t i m e , the "Elektrosila" plant in Petrograd began pilot-plant producrion of ferrotitanium and titanium carbide, and the conversion of the latter into titanium tetrachloride. However, these attempts were not continued. Geological surveys of rare metals were started as early as 1922. These surveys which became very intensive during the first f e w five-year planning periods resulted in a sufficient supply of local raw materials being made available to the Soviet rare-metals industry. In 1938 A. E. Fersman analyzed the surveys of raw materials in the USSR and noted that the occurrence of the following elements had not been reported: scandium, gallium. germanium. rhenium, and thallium. By now, deposits of these elements have also been discovered and investigated. Following the successful research work of the geologists, great advances were made in mining. ore dressing. chemistry and metallurgy. A rare-element office was set up in 1922. forthe investigation and development of new technological processes for the manufacture of rare metals and their compounds. T h e State Institute of Rare Metals was founded in 1931. In the same year the first chairs of technology and metallurgy of rare metals were created in higher schools of engineering in the Soviet Union. There was a rapid development in the production of rare metals in the USSR. Production of the following metals was organized: tungsten in 1927, molybdenum in 1928. hard alloys in 1929. alloys of iron with tungsten and molybdenum in 1931, beryllium in 1932, tantalum and lithium in 1933, and ferrovanadium in 1932-1935. A very rapid development in the production of rare metals has taken place in the last 15 years (the postwar period). This was caused by the need for materials with very different physicochemical properties in modem industries: high-speed and high-flying aircraft production, electro-vacuum technology, semi conductor electronics, and production of atomic power. Thus, the need of aviation for light and refractory alloys led t o the large-scale production of titanium -a m e t a l which was a laboratory curiosity only 15years ago. T h e production of germanium was organized as a result of the rapid development of semiconductor electronics. The birth of nuclear power industry necessitated the production of uranium and thorium -the

xxv

main nuclear fuels, as well as of other nuclear reactor materials, especially zirconium, beryllium, and lithium. T h e rare metals are of primary importance for the further increase in the production of special steels and of ultrahard, refractory and corrosion-resistant materials, electrical illumination lamps, radio valves, X-ray apparatus. radar sets, photoelectric devices, and various parts for cars, tractors, etc. At present, the Soviet Union produces on an industrial scale a l l techmlogically important rare metals. A continuously increasing consumption of the rare metals is due to the rapid development of the production of special steels, hard and refractory alloys, of the electrical and electronic industries and other branches of technology. As a result, i t is planned to increase t h e production of rare metals in order to satisfy the requirements of the national economy.

xxvi

P a r t One

REFRAC TORY METALS

Chapter Z
TUNGSTEN
1. GENERAL DATA ON TUNGSTEN Brief historical note The element tungsten was discovered in 1781 by the Swedish chemist K. V. Scheele during the decomposition of a tungsten mineral (heavy stone, subsequently known a s scheelite) with acids. Two y e a r s l a t e r , i. e . , in 1783, tungstic acid was isolated from another tungsten mineral wolframite. In the s a m e y e a r , powdered tungsten was produced f o r the f i r s t time by reducing tungsten trioxide with carbon. Tungsten became an important technological metal only about one hundred y e a r s after i t s discovery. The effect of tungsten on the properties of s t e e l was observed in the 18501 s , but tungsten s t e e l s found wide use only a t the end of the nineteenth and the beginning of the twentieth centuries. The invention of fast-cutting s t e e l s (demonstrated for the f i r s t time at the International Exposition of 1900) led to rapid technological advances which increased the efficieny of metalworking by cutting. Tungsten became one of the most important alloying metals. This was the s t a r t of the rapid development of tungsten production. The use of tungsten in tungsten filaments of incandescent electrical lamps was f i r s t proposed and implemented by the Russian inventor A. N. Ladygin in 1900. Such use of tungsten and i t s subsequent use in electronics became possible only after the development of a commercial method f o r the production of ductile tungsten by Coolidge, in 1909. Sintered hard alloys of tungsten carbide w e r e developed in 1927- 1928. These alloys, whose efficiency s u r p a s s e s that of the best tool alloys, play an important role in modern technology.

P r o p e r t i e s of tungsten Tungsten belongs to Group VI of the periodic table. The appearance of massive tungsten r e s e m b l e s that of steel. The melting point of tungsten i s higher than the melting point of any other element except carbon. The metal is characterized by a high boiling point, a low vaporization r a t e at high temperatures, and a low coefficient of t h e r m a l expansion. The electrical resistivity of tungsten is about t h r e e times that of copper.

The mechanical properties of tungsten depend to a g r e a t extent on its mechanical and thermal history Tungsten cannot be worked mechanically i n the cold; its forging. rolling. and drawing a r e c a r r i e d out when hot The physical and mechanical properties of tungsten a r e shown below

. .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . Melting point . 'C . . . . . . . . . . . . . . Boiling point. " C . . . . . . . . . . . . . Heat of fusion. c a V g . . . . . . . . . . . . Heat of vaporization. c a l / g . . . . . . . . . . Heat capacity . c a l / g . " C at:
Atomic number Atomic weight Density. g/cm3 Crystalline l a t t i c e type and parameter
18'

14 183.92 19.3

body-centered Fube a = 3.1641 A 3395 1 1 5


5930 61 1183 0.031 0.0365 0.043 0.048 0.31 0.28 0.24 4.98 X 5.5 27.14 40.0 66.0

. . . . . . . . . . . . . .

1000"
1400" 2100" 20" 821" 1121'

. . . . . . . . . . . . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . .

Thermal conductivity c a l / c m . sec "C. at:

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . . .
. . . . . . . . . . . .

Linear expansion coefficient ( 0 t o 500") of forged rods Electrical resistivity ohm c m . l o 6 at:
20" 300' 1200' 2000" 2430' 3030"

. . . . . .

Electron work function eV Energy radiated by the surface. w a t t / c m 2 . at: 800" 1600"
2200' 2700" 3030"

. . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

. . . . . . .

81.1
103.3 4.55

. . . . . . . . . . . . . .
. . . . . . . . . . . . . .
. . . . . . . . . . . . . .

0.9
18.0 64.0 153.0 255.0

. . . . . . . . . . . . . .

. . . . . . . . . . . . . .

Thermal neutron capture cross section. barn* Brinell hardness. kg/mmz: of a sintered rod of a forged rod . . . . . . . Tensile strength. kg/mmz: of a sintered rod of a forged rod of a nonannealed wire of an annealed wire . . . . . . Young's modulus (of a wire). kg/mmz

. . . .

19.2
200 -.230 350 .400

. . . . . . . . . . .
. . . .

. . . . . . . . . . . . . . . . . . . . . . . . . . . . . .

. . . . . . .

. . . .

13 35 .150 180 .415

110
35000

.38000

-____--------- * One barn = lo-" c m 2 .


Tungsten is stable in a i r

Noticeable oxidation s t a r t s between 400 and

500" (an irridescence appearing on the surface). and vigorous oxidation

to the yellow trioxide takes place at higher temperatures

. . . .

. . . . . .

Tungsten at t e m p e r a t u r e s below i t s melting point does not r e a c t with hydrogen. Reaction between tungsten and nitrogen, with formation of tungsten nitride WN2, takes place only above 2000". Tungsten is rapidly oxidized by water vapor at red heat (600 to 700") with formation of W Q : W

+ 3H,O 2 WO, + 3H,.

Solid carbon and carbon-containing g a s e s (CO, C H I , C2H2, etc.) r e a c t with tungsten at 800 to 1000" with formation of tungsten carbides (WC and W2C). The presence of even s m a l l amounts of carbide in the metal makes it brittle and strongly reduces i t s electrical conductivity. At ambient temperatures, tungsten is not attacked by hydrochloric, sulfuric, nitric, and hydrofluoric acids a t any concentration, o r by aqua regia. At 80 to 100" the m e t a l is not attacked by hydrofluoric acid; it is attacked v e r y slowly by hydrochloric and sulfuric acids, and m o r e noticeably s o by n i t r i c acid and aqua regia. Tungsten dissolves rapidly in a mixture of hydrofluoric and nitric acids. Tungsten is not affected by alkali solutions a t room temperature, but is oxidized by molten alkalies (in the presence of a i r ) with the formation of tungstates. This p r o c e s s is very rapid in the presence of oxidizing agents (NaNQ, NaN02, KClQ, PbO,).

The properties of tungsten compounds Oxides. T h e r e a r e four known oxides in the W - 0 system. They a r e tungsten trioxide ( o r tungstic anhydride) W q , the dioxide W Q , and the intermediate oxides W4OI1 ( o r WO2.75) and W~o&g (or WQ.90). The compo sition of the tungsten oxides may vary over a fixed range without change in crystalline structure. Thus, the composition of tungstic anhydride v a r i e s over the range WOz.95-3.0. The composition ranges of the inter mediate oxides a r e represented by .the formulas WO2.8o-2.9,3 and WO2.70-2-75 and that of the dioxide by W&.0-2.05. T u n g s t e n t r i o x i d e WO, is a yellow crystalline powder. Its density i s f r o m 7.2 to 7.4 [g/cm3]. Its melting point is about 1470". Its boiling point l i e s between 1700 and 2000". The heat of formation of W Q is 202.8 kcal/mole. Tungstic anhydride is sparingly soluble in water and in common inorganic acids (except hydrofluoric). Its solubility in water is 0.02 g/l. Tungstic anhydride dissolves in solutions of alkali hydroxides and sodium carbonate, with formation of tungstic acid s a l t s (tungstates) :
WO,

+ 2NaOH = Na,WO, + H,O.

Tungsten trioxide dissolves slowly in ammonia solutions, especially


if it has been ignited to a high temperature. At 800 to 900" W Q may be reduced to the metallic s t a t e by both hydrogen

and carbon monoxide. T u n g s t e n d i o x i d e W 0 2 is a chocolate-brown powder. Its density 1270" and it begins is between 10.9 and 11.1 [g/cm3]. Its melting point is 1700". The heat of formation to sublime a t 1000". The boiling point is of WO, is 134.0 kcal/mole.

l1l11ll11llIll l l I1 Il1

Tungsten dioxide is formed in the reduction of W Q with hydrogen at 575 to 600". The dioxide is insoluble in water, alkali solutions, hydrochloric acid and dilute sulfuric acid. It is oxidized by nitric acid to the higher oxide. T h e i n t e r m e d i a t e o x i d e s - Wl0Oz~ and W4O11 are formed in the reduction of W 0 3 with hydrogen at 300 to 550". They m a y also be prepared by ignition of a mixture of tungsten and W Q (or of W 0 3 and WOz) in an i n e r t atmosphere such as nitrogen, W l o Q 9 is a blue powder while W40l1 has a violet color. The density of WlOOz9is 7.1 to 7.2 while that of W4O11 is somewhat higher, 7.7 to 8.0 [g/cm3]. The heats of formation of the oxides a r e 193.1 and 180.3 to 163.1 kcal/mole respectively. The intermediate oxides a r e sparingly 23 soluble in water, in inorganic acids and in 20 dilute alkali solutions. 15 Tungstic acid. Tungstic acid exists in two 2 f o r m s : the yellow acid which is precipitated by acids f r o m hot tungstate solutions, and 3 LO the white colloidal form which is precipitated 45 in the cold. The yellow acid has the compoI# 200 3 m sition HzW04. The white form i s , apparently, Temperature, 'C hydrated tungsten trioxide since i t s dehydration curve has no breaks o r plateaus ( F i g u r e 1). FIGURE 1 . Dehydration curves of Water is detached from HzWO4 a t above 188", tungstic a c i d. with the formation of W Q . The white acid is converted into the yellow form by prolonged boiling. Tungstic acid is dissolved in solutions of alkali hydroxides, sodium carbonate and ammonia with formation of tungstates. Tungstic acid is capable of adding various numbers of W 0 3 molecules. This leads to the formation of polyacids whose composition corresponds to the general formula XHZO yW03 .nHzO where y > x. These acids have not been isolated in the f r e e s t a t e (except for the metatungstic acid HZW4Ol3.9HzO), but t h e i r s a l t s a r e known to exist. These s a l t s a r e known a s polytungstates. The general formula of the polytungstates is x M e 2 0 . yWOa. n HzO. When x = y = I, the s a l t corresponds to the normal tungstate, but polytungstates a r e formed when y : x > 1. The following types of poly tungstates a r e known: M e p O .2W03- ditungstates;

,s

3Me20 .7w03 - paratungstates; 5Me20'12w03 M e 2 0 .3W03- tritungstates; M e 2 0 .4W03 - metatungstates. Unlike tungstic acid, the metatungstic acid HZw4013 9HzO i s quite soluble in water and s o a r e mostof its s a l t s including the calcium, iron, and copper salts, etc. The density of metatungstic acid is 3.93 [g/cm3]. The m o r e important tungstates. T h e n o r m a l s o d i u m t u n g s t a t e Na2W04is one of the technically important s a l t s of tungstic acid, This s a l t i s formed when WO, is dissolved in solutions of alkali hydroxides or sodium carbonate. The s a l t crystallizes a s dihydrate from aqueous solutions a t above +6"; the decahydrate is stable a t lower temperatures. The melting point of anhydrous NaZWO4 i s 200", i t s density is 4.18 [g/cm3].

The solubility of sodium tungstate in water is:


Temperature."C Solubility of the anhydrous salt.%

. . . . -5 . . . . . . 30.6

0
35.4

5 41.0

6
41.8

10

20

40 43.8
' 2H,O

80 41.4

100 49.2

Na2W0,. 10H,O

---f

41.9

42.2 Na,WO,

S o d i u m p a r a t u n g s t a t e 5 N a 2 0 . 12W03.rtH10 crystallizes f r o m solutions of normal tungstates as a result of careful neutralization to a pH of 5.5 to 6. This is accompanied by hydrolysis:
I2Na,WO,

+ 7H20

5Na,O. 12W0,

+ 14NaOH.

The alkali hydroxide formed in the reaction is neutralized by the acid. The salt hydrates contain 28 and 25 water molecules after crystalliza tion in the cold and a t 60 to 80" respectively. The solubility of the paratungstate in water is strongly affected by the temperature :
Temperarure,'C , , . Solubility of the anhydrous s a l t , %

. .

. .

12.4
5.52

39.6 17.94

101.8 70 6

The melting point of the anhydrous salt i s 705.8" and i t s density is 5.49 [g/cm3]. S o d i u m m e t a t u n g s t a t e is formed by neutralizing a solution of sodium tungstate with acid to pH = 4. When its solution is boiled for a prolonged time in the presence of an alkali, the metatungstate is converted into the normal salt. The possibility of formation of metatungstates must be taken into account in production processes, since the presence of metatungstates in the solution may lead to incomplete precipitation of the tungstic acid. A m m o n i u m p a r a t u n g s t a t e 5(NH4)2012W03.nH2O may be prepared by neutralization or evaporation of a solution of the normal ammonium tungstate ( NH4)zW04*:
12 ( NH4)2 W 0, + 5 ( NH4)2
0*

12 W Os + 14NH3 + 7Hz 0.

Needle-shaped crystals containing 11 molecules of water a r e precipitated from cold solutions (below 50") ; the lamellar modification containing 5 water molecules is precipitated from heated solutions. The salt i s sparingly soluble in water. The solubility of ammonium paratungstate i s strongly affected by the temperature:
Temperature. " C Solubility of the anhydrous salt.% 17 1.064
29 2.014

45
3.467

49 4.341
d

52 3.280

IO
7.911

L____

n = llHZ0

n = SH,O

C a 1c i u m t u n g s t a t e CaWO, is a white, fine crystalline powder which is sparingly soluble in water and whose solubility d e c r e a s e s with increasing temperature: from 0.0064 g/1 at 15" to 0.0012 g / l at 100". Its density is 5.98 [g/cm3].

* The normal

ammonium tungstate exists only in aqueous solutions.

Calcium tungstate may b e prepared by precipitation on the addition of calcium chloride or lime to solutions of the alkali tungstate, or by d i r e c t solid-state interaction of calcium oxide with tungstic anhydride a t 600 to 800". Calcium tungstate is decomposed by acids, with formation of tungstic acid. Heteropolyacids and their salts. Tungsten has a number of known heteropolyacids which form s a l t s with phosphoric, arsenic, silicic and other acids. Those acids a r e formed through the substitution 'of (WOd2 and ( W ~ 0 7 ) ~ ions for oxygen ions in silicic, phosphoric, a r s e n i c and other acids. For instance, there a r e potassium and sodium s a l t s of silicotungstic an acid which does not exist in the free state. The acid H~[Si(W20~)6].nH~0, composition of the potassium s a l t is Ks[Si(W207)6].xHp0 ( where x is 12 o r 20) ~ )~ H z O (where x is 10 o r 18). Salts and of the sodium s a l t - Na4 H4[Si( W Z Oti] which a r e derivatives of phosphotungstic o r arsenotungstic acids (H7[R(W207),] where R = P or As) a r e also known to exist. The heteropolycompounds a r e formed on the acidification of sodium tungstate solutions containing s a l t s of silicic, arsenic or phosphoric acids. Tungsten bronzes. A peculiar type of compound, known as tungsten bronzes, i s formed on the reduction of sodium, potassium or lithium tungstates with hydrogen o r other reducing agents, during the electrolysis of m e l t s of these salts, or on the fusion of normal tungstates with W g in the absence of a i r . Those bronzes have a pleasant metallic luster, metallic-type conductivity (electrical) and considerable corrosion resistance. The composition of these tungsten bronzes corresponds to the formula MepO * W 0 2 - n W 0 3where n ranges from 1 to 4. The color of the compounds is gold-yellow where n = 1, red where n = 2, violet where n = 3 and blue where n = 4 .
e

It is now thought that sodium-tungsten bronzes are interstitial solid solutions of sodium in WO,. A fraction of the tungsten atoms. which corresponds t o the fraction of sodium atoms filling the vacant sites in the W03 lattice, is converted from the hexavalent t o the pentavalent state. A compound -NaWOs is formed when a l l vacant sites in the lattice are filled.

Tungsten forms a s e r i e s of chlorides: T u n g s t e n c h 1o r i d e s WCl,, WC15, WCl,, WCl2, and oxychlorides: WOCl, and W02C12. T h e h e x a c h 1o r i d e WCl, is formed when a s t r e a m of d r y chlorine i s passed over tungsten powder at 750 to 800" a s dark-violet crystals. The hexachloride melts at 2 7 2 and boils at 337". It exists in two crystalline modifications: a- WC16 and 6WCl,. The transition from the a- to the @-modificationoccurs a t about 230" and i s accompanied by a large change in volume. The hexachloride i s decomposed by water, with the formation of tunstic acid: WCI, 4H20 = H,WO, 6HCI.

The tunsten o x y c h l o r i d e s WOzClzand WOCl4 a r e formed as a result of the action of chlorine on tungstic anhydride or on a mixture of WO, and carbon. W0,Cl2 is pale yellow and melts at 266"; WOC1, is red-brown, melts at 209 and boils at 232". The lower tungsten chlorides a r e formed by the reduction or thermal dissociation of the higher chlorides. The pentachloride WC15 is a black crystalline substance which melts at 2 4 8 and boils at 276"; WCZ4 i s a graybrown nonvolatile substance.
1455

Tungsten sulfides. Tungsten forms two sulfides: WS2 which occurs in nature as the m i n e r a l tungstenite, and W%. The disulfide may be prepared by the interaction of sulfur vapor with powdered tungsten o r W Q a t 800 to 900". Its s t r u c t u r e and properties resemble those of molybdenum disulfide (see p. 105). The trisulfide WS, formed by bubbling hydrogen sulfide through hot acidified solutions of tungstates. WS, dissolves in alkali sulfides with formation of sulfur-containing salts. The dark-brown tungsten trisulfide is precipitated when a solution of the sulfur-containing s a l t is acidified:

is

Na,WS,

+ H,SO, = WS, + Na,SO, + H,S.

Tungsten carbides. Tungsten f o r m s two carbides: W C and W2C. Those carbides a r e hard, refractory substances. The melting points of W2C and W C a r e 2750 and 2900" respectively. The microhardness of WC is 1 7 6 0 kg/mm2. The m o s t common method for the production of the carbides is ignition of a mixture of powdered tungsten and carbon (carbon black) at 1000 to 1500".

U s e s of tungsten
Tungsten i s widely used in modern technology both in the pure form and in alloys, the most important of which a r e alloy steels, hard alloys based on tungsten carbide, and wear-resistant and refractory alloys. Tungsten compounds a r e used in various branches of technology. Tungsten in s t e e l s . More than half of the total amount of tungsten concentrates is used in the production of special steels. The most important of those s t e e l s a r e the high-speed steels, containing 8-2070 W, 2-7qG C r , 0-2.570 V, l-50/0Co, 0.5--1~010. The high-speed s t e e l s a r e characterized by their capacity for self-hardening in a i r and by the fact that their secondary hardening takes place at a high tempera ture (700 to 800"). As a result they maintain their high hardness and wear resistance up to 600 to 650". F o r the s a k e of comparison, it should be noted that the tempering of high-carbon tool s t e e l s takes place at 200 to 250". These properties, which a r e due to the presence of tungsten and chromium, have permitted an increase in cutting speed during mechanical working of s t e e l s from a few m e t e r s p e r minute to s e v e r a l dozens of m e t e r s . A g r e a t increase in productivity has resulted. In addition to the high-speed steels, other tungsten and chromiumtungsten s t e e l s have also found wide use. Different brands of C r -W s t e e l s containing 1 to 6 70W and 0.4 to 2 7 0 C r a r e used for the fabrication of tools: saws, cutters, dies, pneumatic tool p a r t s , e t c . Tungsten is used a s a component of magnetic s t e e l s . There a r e tungsten and tungsten-cobalt magnetic steels. The f i r s t of those contain 5 to 6 70W and 0.6 to 0.7570 C ; by comparison with nonalloyed magnetic steels, they have a higher intensity of magnetization and coercive force. Even stronger magnetic properties a r e possessed by tungsten-cobalt s t e e l s containing 5 to 9 % W and 30 to 4 0 % Co. These s t e e l s a r e character ized by t h e i r v e r y high coercive force (200 to 250 oersted).

Hard alloys based on tungsten carbide. Tungsten carbide WC p o s s e s s e s v e r y high hardness and wear resistance and is highly refractory. It has been used a s the b a s i s of the most productive hard tool alloys. These alloys contain 85 to 95% WC and 5 to 1 5 %Co. The Co s e r v e s a s a cementing component which i m p a r t s n e c e s s a r y strength to the alloy. Some types of alloys, used mainly in the working of steel, contain titanium, tantalum, and niobium carbides a s well as WC. A11 these alloys a r e produced by powder-metallurgy techniques. They maintain t h e i r high hardness and wear resistance up to 1000 to 1100". This has permitted a g r e a t increase in cutting speeds (to 150 to 250 m / m i n and higher) and a noticeable i n c r e a s e in the productivity of cutting p r o c e s s e s a s compared with the use of c u t t e r s made of the best tool steels. Hard alloys a r e used f o r the manufacture of working p a r t s of cutting and drilling tools, of dies for wire drawing and of other p a r t s requiring a high wear resistance and hardness. Cast tungsten carbides a r e used in addition to sintered hard alloys (which contain cementing additives such a s cobalt and nickel) in c a s e s where the alloys need not have a high strength but only a high wear resistance and hardness (i. e . , for some drilling tools and dies). Heat-resistant and wear-resistant alloys. As it is the metal with the highest melting point, tungsten is a component of many heat-resistant alloys. The alloys of tungsten with cobalt and chromium (the so-called stellites) a r e among the common wear- and heat-resistant alloys of tungsten, which have been known for a long time. Their composition i s : 3 to 1 5 %W, 25 to 3570C r , 45 to 6570Co, and 0.5 to 2.7570 C. The stellites have a high hardness and a high resistance to wear, corrosion, and high temperatures. Alloys of this type .are used mainly for coating (by the "surfacing" tech nique) machine p a r t s subjected to rapid wear (e. g . , valves of a i r c r a f t engines, working surfaces of s h e a r s for hot cutting, dies, turbine blades, excavation equipment, plowshares, etc. ). Contact alloys and alloys f o r radiation shielding. The tungsten-copper (10 to 4070 Cu) and tungsten-silver alloys (which a r e prepared by sintering) combine the high electrical and t h e r m a l conductivity of copper and s i l v e r with the wear-resistance of tungsten. As a result, these alloys a r e very effective m a t e r i a l s for the production of contacts in knife switches, cut out switches, spot welding electrodes, etc. The s a m e group of alloys includes the W-Ni-Cu alloys (85 to 9070 W, 3 to lO%Ni, 2.570 Cu) which have a high specific gravity and a r e used in radiation therapy for y-ray shielding. P u r e tungsten. Tungsten is used in the form of wires, s t r i p s and various forged products in the manufacture of electric bulbs, in electronics, and in X-ray technology. Tungsten is the best m a t e r i a l for filaments and s p i r a l s in incandescent lamps. Its high working temperature ( 2 2 0 0 to 2500") provides a high luminosity, while i t s low vaporization r a t e e n s u r e s long s e r v i c e life. Tungsten wire i s used in the production of directly heated cathodes and g r i d s for electronic transmitting tubes, cathodes for highvoltage r e c t i f i e r s , and heaters for indirectly-heated cathodes in various electronic devices. Tungsten i s used in the production of anticathodes and cathodes in X-ray and gaseous-discharge tubes, for the contacts in electric equipment and f o r the electrodes in atomic hydrogen torches. Tungsten wire and rods a r e used a s electric h e a t e r s in high-temperature furnaces (up to 3000"). The tungsten heating elements operate in an atmosphere of hydrogen or an inert gas, o r in vacuo.
8

I ,

I,,

1 . 1

I 1

I#

1111

1111

111

I 1

1111

1 1 1 1

Recently, alloys of tungsten with other r e f r a c t o r y metals (tantalum, niobium, molybdenum, rhenium) have found use a s r e f r a c t o r y m a t e r i a l s in a i r c r a f t and rocket technology and in other branches which require machine, engine, and tool p a r t s which a r e highly refractory. Chemical compounds of tungsten. Sodium tungstate i s used in the manufacture of s e v e r a l brands of light-fast lacquers and pigments. In addition, sodium tungstate is used in the textile industry for the weighting of t i s s u e s and a s a mixture with ammonium sulfate and phosphate for the production of heat-resistant and water-repellent tissues. Tungstic acid is used as an adsorbent, as a mordant and a pigment in the textile industry, and a s a catalyst in the production of high-octane gasoline in the chemical industry. The tungstates of lead, zinc, and barium a r e used a s f i l l e r s for white lead. Tungsten d{sulfide (WS2) is used a s a catalyst in the production of synthetic gasoline.

2..

MINERALS, ORES, AND ORE CONCENTRATES Tungsten m i n e r a l s

The abundance of tungsten in the E a r t h ' s c r u s t i s low. Its concentration The metal itself i s not found in in the E a r t h ' s c r u s t i s only 1 X nature in the native state. There a r e about 15 known tungsten minerals. Most of these a r e tungstates - s a l t s of tungstic acid. The only minerals of practical value a r e those belongfhg to the wolframite and scheelite groups. W o I f r a m i t e [(Fe, Mn)W04] is an isomorphous mixture (solid solution) of iron tungstate ( F e w 0 4 and manganese tungstate (MnW04). These s a l t s have crystalline lattices of the s a m e type, with s i m i l a r lattice p a r a m e t e r s , and thus crystallize together with F e and Mn atoms substituted for each other in the lattice sites. Wolframite containing l e s s than 20% manganese tungstate (i. e . , when the iron tungstate i s predominant) is known a s ferberite, while in c a s e s in which manganese tungstate i s the main component (above 8070)it is known as hubnerite. The mixtures between these compositions a r e known a s wolframites. The minerals belonging to the wolframite group have a black, brown, o r red-brown color and a high density, 7.1 to 7.9. Their hardness is 5 to 5.55 (Mohs scale). These minerals contain between 76.3 and 76.6 70 W 0 3 . Wolframite i s slightly magnetic. S c h e e l i t e CaW04 i s nearly pure calcium tungstate. Its color may be white, yellow, gray o r brown. Its density i s 5.9 to 6.1 and i t s hardness is 4.5 to 5. Scheelite often contains some powellite CaMo04. Under ultraviolet illumination scheelite exhibits a sky-blue o r blue fluorescence. The fluorscence becomes yellow if the molybdenum content i s m o r e than 1%. Scheelite is not magnetic. Other tungsten minerals of no commercial value a r e : tungstite o r tungsten ochre WQ H20, cuprotungstite CuW04. HzO, stolzite PbWO,, chillagite 3PbW0,. PbMoO,, ferritungstite Fe203. WQ.6Hz0, and tungstenite

ws2.

Tungsten o r e s and deposits The tungsten o r e s a r e usually of low tungsten content. The richest o r e s contain 0.5 to 2.0% W Q . In addition to the tungsten minerals, the o r e s contain molybdenite, cassiterite, pyrite, arsenopyrite, chalcopyrite and other minerals. Wolframite is most often associated with tin. With respect to their mineralogical compositions, there a r e two types of deposits -wolframite and scheelite; they exist in two forms - vein and contact types. The l a r g e s t fraction of the tungsten minerals in vein deposits l i e s in quartz veins 0.3 to 1.0m thick. The W Q concentration in o r e s of this type ranges from 0.4 to 2%. Deposits of the contact type a r e related.to the contact zones between granite rocks and limestone. Scheelite-bearing skarn (sileceous limestone) s t r a t a a r e characteristic of this type. Scheelite is often accompanied by molybdenum minerals - molybdenite (Mo&) and powellite (CaMoOJ. Ores of this type occur in the USSR, the USA, and Canada. During weathering of the vein deposits, the wolframite and scheelite (which a r e stable minerals) accumulate with the formation of placers. I n these, the wolframite is often assoc'iated with cassiterite. The tungsten deposits form a segmented ring round the shores of the Pacific Ocean: Burma, China, Korea, the USA, Mexico, Bolivia, Argentina, and Australia. Tungsten deposits exist and a r e exploited at a number of locations in the USSR - the U r a l s , Altai, the Caucasus, Central Asia, the Transbaikalia and the Far East. The world production of tungsten concentrates ( 6 0 % W q ) in 1 9 6 0 was about 55,000 tons (not including the USSR).

The enrichment of tungsten o r e s The task of enrichment is to produce o r e concentrates which a r e sub sequently used in the smelting of ferrotungsten o r a r e suitable for chemical or metallurgical processing. The standard concentrates contain 5 5 to 6 0 % W q and a Iimited amount of impurities (Table 7). The degree of enrichment, i. e . , the ratio of the tungsten concentration in the concentrate to that in the ore, ranges (depending on the tungsten content of the ore) from 30 to 120. The tungsten o r e s a r e enriched by different methods: gravitation, flotation, magnetic and electrostatic separation, and chemical enrichment methods. The sky-blue fluorescence of scheelite under ultraviolet light is used f o r preliminary o r e sorting in dark rooms. Because of their high densities, wolframite (7.1 to 7.9) and scheelite (5.9 to 6.1) can be separated from quartz (2.6) and other low-density minerals by wet jigging and concentration on tables and sluices. However, the gravitation method does not ensure the separation of cassiterite (density = 6.8 to 7) and sulfide minerals from wolframite and scheelite. In the case of minerals containing embedded particles of coarse grain structure, the separation of wolframite from cassiterite can be

10

successfully achieved by electromagnetic separation in a high-intensity magnetic field (wolframite is slightly magnetic while c a s s i t e r i t e is non magnetic). The magnetic separation is occasionally preceded by calcination, which converts the pyrite into the magnetic oxide Fe304 this being subsequently separated from the wolframite in a low-intensity magnetic field.
TABLE 7
Technological requirements for some types of tungsten concentrates (GOST 213-56) Chemical composition,qo

Concentrate brand

Type of concentrate

impurities (maximum)

Use

- -

Mo

As

Sn

cu

KVG

Wolframite hubnerite

0.04

0.5

0.2

0.2

0.2

'roduction of ferrorungsten

KVGT KShT, KShT,

Wolframite hubnerire Scheelite

0.10 0.20 0.20

0.8 0.8 0.8

0.10 0.10

0.10

1.00 0.20 0.20

0.15 0.20 0.20

0.06 0.04 0.30

Production of hard alloys and tungsten metal

The magnetic separation of wolframite from c a s s i t e r i t e i s hindered when the c a s s i t e r i t e i s coated with an iron oxide layer. In such a c a s e the iron oxide is dissolved in advance by treating the wolframite-cassiterite concentrates with hot solutions of sulfuric or hydrochloric acids. Unlike wolframite, scheelite i s non-magnetic and cannot be separated from c a s s i t e r i t e by magnetic separation. Scheelite i s separated from cassiterite by flotation o r by electrostatic separation, which i s based on differences in the electrical conductivity of the minerals. Chemical separation methods a r e used occasionally. Gravitation methods provide a satisfactory means for the isolation of tungsten from wolframite o r e s and up to now a r e the main way in which these o r e s a r e enriched. When scheelite o r e s a r e enriched by gravitation methods, the extraction of tungsten does not exceed 7070, because of the tendency of scheelite towards overdisintegration. This leads to the formation of fines and t o considerable l o s s e s of tungsten in the tailings. Recently, flotation methods have become the main means of enriching scheelite ores, especially if they a r e lean and finely impregnated. During flotation, sodium carbonate, water g l a s s and tannin a r e used a s d e p r e s s o r s and pH-control agents; oleic acid, sodium oleate, and liquid soap a r e used a s collectors; and pine oil, terpinol, technical grade c r e s o l and other reagents a r e used a s frothing agents. The flotation is c a r r i e d out in an alkaline medium at a pH of 9 to 10. The addition of copper and iron sulfates to the water g l a s s a s s i s t s the depression of calcite, fluorite and apatite.

A combined method, involving flotation and gravitation together with chemical treatment, is used occasionally f o r enriching scheelite o r e s . In many c a s e s the powellite (CaMo04)present in scheelite-type o r e s can be separated only if it is not bound isomorphously to the scheelite. As a result, scheelite concentrates often contain molybdenum, which is separated during the hydrometallurgical processing of the concentrates. The concentrations of various impurities a r e adjusted to the p r e d e t e r mined l i m i t s by various chemical p r o c e s s e s . Thus, the phosphorus content of scheelite is reduced by treating it in the cold with hydrochloric acid. Calcite and dolomite a r e removed at the s a m e time. To separate copper, arsenic, and bismuth the concentrates a r e f i r s t calcined and then treated with acid o r processed by other methods. The adjustment of the impurity content to a predetermined level is especially important in the case of concentrates used for smelting of ferrotungsten. Concentrates with g r e a t e r impurity content may be used occasionally in hydrometallurgical processing.

3.

PROCESSING OF TUNGSTEN CONCENTRATES

The main product formed in the direct processing of tungsten concen t r a t e s (in addition to ferrotungsten for f e r r o u s metallurgy) is tungsten trioxide. It is used a s the starting m a t e r i a l for the production of tungsten and tungsten carbide - the main component of hard alloys. There a r e s e v e r a l industrial processing methods of the concentrates. The selection of the method to be used depends on the type of raw m a t e r i a l (wolframite o r scheelite concentrate), the production volume, the requirements with respect to the purity of the tungsten trioxide and i t s physical properties (particle s i z e of the powder) and a number of conditions which determine the cost of processing. Each kind of scheme for the processing of tungsten concentrates c o m p r i s e s the following stages: 1) decomposition of the concentrate; 2) production of technical grade tungstic acid; 3) purification of the technical grade acid and manufacture of the required commercial product. The processing techniques of wolframite and scheelite concentrates differ mainly in t h e i r decomposition stage. The subsequent operations involving the separation of tungstic acid and i t s purification a r e s i m i l a r . The following methods f o r the decomposition of tungsten concentrates a r e used in industry. The wolframite and scheelite concentrates a r e calcined or fused with sodium carbonate, and the product is leached with water o r treated in an autoclave with an aqueous solution of sodium carbonate. Wolframite concentrates a r e occasionally treated with an aqueous solution of sodium hydroxide. Scheelite concentrates a r e decomposed with acids. Aqueous solutions of sodium tungstate a r e obtained whenever alkaline reagents (sodium carbonate, sodium hydroxide) a r e used for the decomposition; tungstic acid o r some other tungsten compound is then precipitated from such solutions. When acids a r e used for the decomposition of the concentrates, the product is a precipitate of technical grade tungstic acid contaminated with impurities. It i s purified in subsequent operations.

.. .

..

..

The individual techniques used in the processing of wolframite and scheelite concentrates are described below. Technological operations common to both types of concentrates (e. g., treatment of sodium tungstate solutions) a r e described in detail only in the section dealing with the production of tungsten trioxide from wolframite.

4.

DECOMPOSITION OF WOLFRAMITE CONCENTRATES Fusion with sodium carbonate / 1 - 3 /

This is the most common commercial method f o r the decomposition of wolframite. A technical flow sheet f o r the processing of wolframite concentrates by this method is shown in Figure 2. Fusion. The interaction of wolframite with sodium carbonate in the presence of oxygen may be described a s follows:
2FeW0, 3MnW0,

+ 2Na,C03 + l/gOp + 3Na,cO, + l/,O,

2Na,WO, 3Na,WO,

+ Fe,O, + 2CO,, + Mn30, + 3C0,.

The reactions a r e virtually i r r e v e r s i b l e since the COz i s removed from the reaction zone. The reaction takes place at 800 to 900". A 10 to 1 5 %excess of sodium carbonate above the stoichiometric amount required e n s u r e s a nearly quantitative (98 to 99.5700)decomposition of the concentrate. A n oxidant (saltpeter, in an amount corresponding to 1 to 4% by weight of the concentrate taken) is occasionally added in o r d e r to accelerate the oxidation of iron. During fusion with sodium carbonate, the common impurities in the wolframite concentrate (silicon, phosphorus, arsenic, molybdenum, and other compounds) form soluble sodium s a l t s :

+ Na,C03 = Na,Si03 + CO,;


+ 3Na,C03 = Na3P0, + 3CaC0,;
As,S, + 6NazC0, + 70, = 2Na3As0, 3Na,SO, + 6C0,; MoS, + 3Na,C03 + 4l/,O, = Na,MoO, + 2Na,S04 + 3C0,.
SiO, Ca, (POJ,
A

The tin-containing mineral c a s s i t e r i t e (SnO,) does not react with sodium carbonate t o any significant extent. There i s virtually no formation of sodium stannate. The excess sodium carbonate r e a c t s with the iron oxide, yielding sodium f e r r i t e :
Na,CO,

+ Fe,O,

= PNaFeO,

+ CO,.

When the fused m a s s is leached with water, the f e r r i t e decomposes with formation of sodium hydroxide:
PNaFeO,

+ H,O = 2NaOH + Fe,O,.

Depending on the temperature, the reaction mixture a t 800 to 950" i s either w semi-fused dough-like cake o r a liquid.

13

Wolframite concentrate

1
Comminution

Scheelite concentrate

1
Comminution

Fusion

Fusion

1
Cake (Na, WO,, FezOO, Mn,04 and impurities Water

> Leaching

1
1

Cake (Na,WO, , CaSiO, and impurities)

Filtration Tailings T o waste

s.

&
Solution (Na2W0,, Na2Si03 and other impurities) Purification from Si, P. As and Mo

CaC12 Precipitation of sodium tungstate

-1
I

Na2W0, solution

Wastes (caked impurities)

Precipitate CaWO, HC1\+ Decomposition

Mother liquor

J.

To storage container (trap)

Filtration and washing of the tungstic acid

1
To calcination or ammoniacal purification

Wash liquor

FIGURE 2.

Processing of tungsten concentrates by fusion with sodium carbonate.

The cake contains sodium tungstate, i r o n ( a s the oxides and f e r r i t e ) , manganese oxides, sodium s a l t s of the impurities, excess sodium carbonate, and unreacted wolframite, In industrial processing, the fusion of the wolframite concentrate is c a r r i e d out in batch or continuous-operation furnaces. The batch p r o c e s s is t o be p r e f e r r e d in s m a l l - s c a l e production. The charge is then fused in s m a l l reverberatory furnaces with a base a r e a of about 6 to 8 m2.

14

Continuous p r o c e s s e s a r e preferably used for l a r g e - s c a l e production. Tubular rotating furnaces lined with chamotte b r i c k s a r e employed in the continuous p r o c e s s . In a method developed in the USSR (by Bogomoltskaya, Semenovykh, Matusevich, e t al.), the composition of the charge is designed f o r it not to melt but to remain in the f o r m of particles that can be caked / 3 / . This protects the lining from corrosion by the melt and prevents deposit forma tion in various p a r t s of the furnace. In o r d e r to prevent melting of the charge the solid tailings from the leaching of the cake are added to i t in an amount calculated to reduce the W Q content to about 20 to 22%. The extent of decomposition of the concentrate in this c a s e is fairly high - 9 8 to 99 70. The disadvantage of the p r o c e s s l i e s in the fact that the concentrates a r e diluted with the tailings, which leads to a d e c r e a s e in the output of the furnace. However, this disadvantage is offset by the fact that standard rotating furances can be used and by the long s e r v i c e life of the furnace.

c ,
FIGURE 3.

I
I

Tubular rotating furnace.

1-furnace drum; 2-supporting bands; 3,4-supporting rollers: 5-electric motor; 6-reduction gear; 7-drive gear; 8-combustion chamber: 9-gas venting chamber; 10-feeder.

The rotating furnaces employed (Figure 3) a r e 2 0 to 2 5 % m e t e r s long, lined with wedge-shaped chamotte bricks. The internal diameter of the furnace is about 1.8 to 2.0 m. The rotating drum l i e s on r o l l e r s and is inclined some 2 to 3" towards the discharge end of the furnace. The upper and lower ends of the drum are accommodated in furnace heads. The furnace head is a chamber lined on the inside with r e f r a c t o r y brick. The lower (combustion) chamber is fitted with a chute to discharge the m a t e r i a l f r o m the furnace. The upper (gas-venting) chamber h a s a device for feeding the charge into the furnace (charging pipe or conveyor) and is connected through the gas-venting system with an exhaust fan which c r e a t e s a draft in the furnace. The drum is separated from the feed chamber by a s e a l which prevents leakage of a i r into the furnace. Occasionally, the furnace heads a r e detachable. The furnace is heated with powdered fuel (coal dust), fuel oil, or producer gas, supplied through b u r n e r s or j e t s

15

,!I

. I

II

. I

II

I .

111

111

II

..

1 1 1 .

1 1 1 .

II

1111

1 1 1 1

1 . 1

1.11.

11.11

located in the lower furnace head. The temperature is controlled by adjusting the r a t e of supply of the fuel or a i r to the b u r n e r or jet. The furnaces a r e rotated at 0.5 to 2 rpm. A furnace 20 m long and with an external d i a m e t e r of 2.2 m rotating a t a speed of 2.5 r p m and placed at a slope angle of 3", has a n output of 2 5 tonslday (for a charge containing 20 to 22'70 W Q ) . When a rotating furnace is used the charge must be fed in continously. The comminution of the concentrate is usually c a r r i e d out in ball mills with a peripheral discharge (through screening cloths), which operate in a closed cycle with an a i r separator. The charge components (concentrate, soda, and saltpeter) a r e fed from the s t o r a g e bins to the mixer-conveyor with the aid of dosing devices (automatic balance) and a conveyor system. When prepared, the charge is fed into the hopper of the furnace. At the exit of the furnace, the fused p a r t i c l e s a r e passed through a r o l l e r c r u s h e r (the crushed particle s i z e is 2 to 5 cm) into a continuous wet mill and the resulting pulp is fed into a leaching apparatus fitted with a s t i r r e r (Figure 5). Leaching. The following components of the cake a r e dissolved when i t is leached with water: sodium tungstate, soluble s a l t s of impurities such a s sodium silicate NazSiQ, sodium monohydrogen phosphate, and monohydrogen arsenate, NazHP04and NazHAsO,t:, sodium molybdate NazMo04, sodium sulfate NaZSO4, and unreacted sodium carbonate. The solid residue contains the oxides and hydroxides of iron and manganese, undecomposed concentrate, and other insoluble compounds. The leaching i s c a r r i e d out at 80 to 90" in batch or continuous equipment ( r e a c t o r s with s t i r r e r s or drum leachers) made of iron. As a rule, twoor three-stage leaching is c a r r i e d out in o r d e r to ensure quantitative extraction of tungsten. Regular or superheated steam is used for heating. The use of superheated steam is l e s s convenient, since this leads to dilution of the solution by the condensate. In the continuous process, the leaching is c a r r i e d out in rotary drum l e a c h e r s ; each leacher consists of a s t e e l drum (Figure 4) with s t e e l lids (fitted with hollow pivots fastened to the bearings) bolted to i t s ends. The inner p a r t of the casing and the marginal walls of the lids a r e protected by removable a r m o r plates. The drum is rotated by a motor, through a reduction g e a r which t r a n s m i t s the motion to a toothed wheel fastened on the end lids. The leacher i s held in a horizontal position on two supports. In o r d e r to accelerate leaching and, at the s a m e time, to c r u s h the cake particles, the leaching apparatus is filled with crushing rods 40 to 80 m m in diameter. The hot fused m a s s is continuously charged through the hollow pivot together with hot water (or a weak solution of sodium tungstate). The pulp is continuously discharged through the second hollow pivot. A f a i r l y high degree of extraction of the tungsten (about 98 to 99%) is obtained during the leaching. The concentrated sodium tungstate solutions have a density of 1.26 to 1.40 (depending on the p r o c e s s conditions.), i. e . , contain 190 to 270g W Q in one l i t e r of solution. The dry residues f r o m leaching and filtration amount to about 30 to 407' by weight of the starting concentrate (without taking into account the residues

* In

weakly a l ka l i ne solutions, most of the phosphorus a n d arsenic is present as the HP0:- and HASO: ions rather than a s the PO:- a n d As0:- ions.

FIGURE 4. Drum-type leacher.

1 -end

lid; 2-steel

casing; 3-armor

plates; 4-hOllOw

pivots; 5-bearings;

6-reduction

gear.

added to the charge in the continuous firing process). The W Q content of the tailings ranges from 1.5 to 2%. Residues containing over 2 % W Q are recycled (to the charge-mixing stage).

FIGURE 5 .
1 -elevator;

Diagram of installation for the continuous fusion and leaching process

2-ball mill; 3-conveyor (screw); 4-elevator; J - a i r separator; 6-auromatic dosing devices; I-conveyor; 8-mixing conveyor; 9-charge bin; 10-feeder; 1 1 -furnace; 12-roller crusher; 13-drum leacher; 14-reactor with stirrer; 15-filter.

In batch leaching processes, the filtration is c a r r i e d out with suction filters, f i l t e r p r e s s e s (with cast iron o r wood f r a m e s ) , o r other filters. Vacuum i l t e r s (disc, drum, band, e t c . ) a r e used in the continuous process. Disc f i l t e r s a r e the most convenient to use (because of the s m a l l e r energy consumption and the easy and fast replacement of the cloths and individual filtration cells). The equipment used for the continuous fusion and leaching p r o c e s s i s shown in Figure 5.

Decomposition with sodium hydroxide solutions [l,2 J A double decomposition reaction, with the formation of sodiuni tungstate and f e r r o u s and manganous hydroxides, takes place when finely divided wolframite is brought into contact with a solution of sodium hydroxide.
18

MnWO, + 2NaOH Z Na,W04

FeWO,

+ 2NaOH

Na,W04

+ Fe (OH),, + Mn (OH),.

As compared with fusion with sodium carbonate, the method has the economic advantage of not requiring the use of a furnace. It h a s the drawback that only very finely divided concentrate may be used (i. e . , investment in comminution equipment) and that it requires a l a r g e e x c e s s of sodium hydroxide, which is much m o r e expensive than sodium carbonate. If the particle s i z e employed is 0.03 -0.04mm, the temperature i s 1 0 0 110", and the excess of sodium hydroxide is 50% o r m o r e above the theoretical amount, a high extent of decomposition (98 - 9 9 % ) is obtained. The oxidation of the f e r r o u s and manganous hydroxides to f e r r i c and manganic kiydroxides is accelerated by bubbling a i r through the solution. The decomposition is c a r r i e d out in s t e e l vats fitted with s t i r r e r s and heated with regular steam. Soviet workers (Meerson and Nadoltskii 1 5 1 ) have shown that the consumption of alkali may be reduced and the decomposition may be accelerated by conducting the alkali processing in heated ball mills. The increase in the reaction r a t e is attributed to the abrasive action of the balls, which remove the iron and manganese hydroxide l a y e r s from the surface of the concentrate particles.

5.

PROCESSING OF SODIUM TUNGSTATE SOLUTIONS Removal of impurities from the solutions

Sodium tungstate solutions a r e contaminated with silicon, phosphorus, arsenic, molybdenum, and sulfur (in the form of sodium s a l t s ) and these may cause contamination of the tungstic acid. Preliminary purification of the solutions from contaminants often leads to the production of tungstic acid of a degree of purity which p e r m i t s the production of hard alloys without the need for additional purification. Some contaminants, such a s phosphorus and a r s e n i c compounds, interfere with the sedimentation of tungstic acid. They may also cause l o s s e s of tungsten probably owing to formation of sodium heteropoly tungstates (silico-, arseno- and phosphotungstates) in solution. Removal of silicon, Preliminary purification from silicon i s necessary if the SiO, concentration in the solution i s m o r e than 0.1 7 ' of the WOs concentration. The most common purification method is based on the precipitation of silicic acid in a given pH range. The process i s based on the fact that sodium silicate hydrolyzes when the solution is neutralized to a pH between 8 and 9 :
Na,SiO,

+ H,O 2 H,SiOs + 2NaOH.

When the solution is boiled, the silicic acid coagulates and s e p a r a t e s out as a voluminous precipitate. The lyes a r e usually neutralized with hydrochloric acid. The precipita tion is c a r r i e d out in solutions containing 130 to 150 g / l WO, (solution

19

1 l1 1 l1 l1 1 1 1 1 1 l1 l1 llIll l l I I1 Il 1 Il lIllIllIl1

density 1.16 to 1.20). Hydrochloric acid is added to the sodium tungstate solution after the l a t t e r has been heated to boiling. In o r d e r to prevent localized over-acidification, (which may lead t o the formation of silico and metatungstates which reduce the degree of extraction of tungsten from the solutions in subsequent operations) the acid is added slowly, in a thin s t r e a m , with continuous s t i r r i n g . The neutralization is controlled by titration of aliquots of the solution, using phenolphthalein a s indicator. Ammonium chloride, which hydrolyzes in aqueous solutions with the formation of hydrochloric acid (which neutralizes any alkali in the solution) is used occasionally a s the neutralizing agent, instead of hydrochloric acid:
NHJI + H,O NH.,OH HCI, NaOH HCI 2 NaCl H,O.

+ +

The use of ammonium chloride eliminates the hazard of localized 0ve.r acidification. Moreover, the u s e of ammonium chloride is compatible with the subsequent removal of phosphorus and arsenic. Removal of phosphorus and arsenic. Arsenic and phosphorus may be precipitated from the solutions a s magnesium phosphate and arsenate. The ammonia-magnesium p r o c e s s , based on precipitation of the sparingly soluble magnesium ammonium phosphate and arsenate (Mg(NH4)P04.6Hz0 and Mg(NH4)As046HzO) f r o m the solution is the best purification method. The solubility of these s a l t s in water a t 20" is 0.053 and 0.038 7'0 respectively. The solubility is further reduced by the presence of Mg2+and N Q ions. The precipitation reactions may be expressed by means of the following equations:
Na,H!W4 Na,HAsO,

+ MgCI, + NH,OH = Mg (NH,) PO., + 2NaCI + H,O, + MgCI, + NH40H = Mg(NH,) AsO, + 2NaCI + H,O.

The magnesium ammonium s a l t s may hydrolyze with the formation of m o r e soluble acid phosphates and a r s e n a t e s :
Mg (NHJ PO, H 2 0 MgHPO,

+ NH4OH.

As i s evident from the above equation, a certain excess of ammonia must be present in solution in o r d e r to prevent hydrolysis. The presence of ammonium chloride is also n e c e s s a r y since i t prevents precipitation of magnesium hydroxide by reducing the OH--ion concentration to the extent where the solubility product of Mg(OH), is not reached. The precipitation of magnesium ammonium phosphates and arsenates i s c a r r i e d out in the cold. The required concentration of ammonia and ammonium chloride in the solution is established and an excess of a magnesium chloride solution is added. At a certain reagent r a t i o a crystalline precipitate consisting of the magnesium ammonium s a l t s is formed in the solution, on prolonged (up to 48 hours) standing. The gelatinous normal phosphate and arsenate (Mg,(P04), and Mg3(As04)z)a r e partially precipitated together with the ammonium salts.

20

Removal of molybdenum. Molybdenum must be eliminated from the sodium tungstate solution if i t s concentration exceeds 0.3 g/l*.
Sodium tungstate solution (contaminated with Si, P . As. ivlo)

Sedimentation and decantation Solution

4 . 1

MgCIz

H,SiO, precipitate (contaminated with W)

soEton

Separation of arsenic and phosphorus

Reprocessing (in the fusion section)

1
Sedimentation and decantation

.
Precipitate
T o waste disposal

Solut ion

blolybdenum ;eparation Filtration Pure solution of sodium tungstate


T o tungsten separation from solution

. 1 -

.1
bios, Precipitate

FIGURE ti.

Purification of sodium tungstate solurions.

The most satisfactory commercial method for the separation of molybdenum from tungsten is based on the precipitation of molybdenum a s M o S ~ . Separation may be brought about by this method owing to the fact that the sulfides of molybdenum and tungsten form under different conditions. When a s m a l l amount of NazS (sufficient f o r the formation of molybdenum sulfide alone) i s added to the solution, molybdenum can be precipitated quantitatively ( a s MOSS) by adjusting the p H to between 2.5 and3. The method is based on the reaction:
.______________

* I f the niolyhdcnum concentration is lower, i t s .iihsequent purification.

it

m a y be removed during isolation of the tungstic acid and

21

Na,MoO, 4Na,S 4H,O = Na,MoS, 8NaOH, Na,MoS, + 2HC1= MoS, 2NaCl+ H,S.

Oxysulfides, such as Moos2, a r e partially coprecipitated with the MoS~, since the solution contains some oxysulfide s a l t s (Na2MoOS3and others). Sodium sulfide is added to the solution is in accordance with the above equation, The solution is then neutralized to pH z 3 (using Congo Red paper a s the indicator). A brown precipitate consisting of molybdenum trisulfide (and oxysulfides) is formed after heating the solution for 1 to 2 hours. Not m o r e than 1% of the tungsten present in the solution i s coprecipitated with the molybdenum. The weakly acid solution remaining a f t e r the molybdenum separation stage contains sodium metatungstates, which may cause a non-quantitative precipitation of tungsten from solution in subsequent operations. In o r d e r to decompose the m e t a s a l t s , sodium hydroxide i s added and the solution is boiled. The m e t a s a l t s decompose in an alkaline medium. The molybdenum separation stage usually follows the a r s e n i c and phosphorus separation stage. The purification of sodium tungstate solutions is diagramatically shown in Figure 6. The separation of silicon, phosphorus, and arsenic from solution is c a r r i e d out in iron o r wooden vats fitted with s t i r r e r s , while the separation of molybdenum is c a r r i e d out in iron vats coated with rubber o r with some acid-resistant lining for protection against corrosion. The vats must be fitted with efficient exhausts to remove the hydrogen sulfide formed in the process.

Isolation of tungsten compounds from solution The tungsten may be isolated from sodium tungstate solutions in the form of various compounds. Three different methods a r e used in practice. 1. Direct isolation a s tungstic acid. 2. Precipitation of calcium tungstate and i t s subsequent decomposition with acids. 3 . Crystallization of sodium tungstate o r paratungstate. The direct precipitation of tungstic acid would appear to be the simplest method. However, i t is associated with many difficulties because of the formation of fine precipitates which tend t o become colloidal. This method is occasionally used for the production of finely divided tungstic acid. The second method produces c o a r s e r precipitates of tungstic acid, which a r e e a s i e r to wash. This is an important advantage which explains i t s m o r e common use in industry. The crystallization of sodium tungstates is used only i f these s a l t s a r e the required final product. Precipitation of tungstic acid. Tungstic acid is usually precipitated from the sodium tungstate solution with the aid of hydrochloric acid:
Na,WO, f 2HC1= H,WO, f 2NaCI.

The nature of the resulting precipitate depends on the concentration of the starting solurion, the temperature, and the method of precipitation. A colloidal precipitate of white tungstic acid i s formed during precipitation from cold dilute solutions. C o a r s e r precipitates of yellow tungstic acid,

22

which may be washed with relative ease, are formed when a hot concentrat ed solution of sodium tungstate is poured into boiling hydrochloric acid. The r a t e of introduction of the solution into the hydrochloric acid affects the particle s i z e of the precipitate. This may be due to the fact that different r a t e s of pouring bring about formation of different numbers of crystallization c e n t e r s which, a t a given r a t e of growth of the crystals, determine the particle s i z e of the precipitate. The tungstic acid is precipitated in earthenware r e a c t o r s o r in i r o n r e a c t o r s lined with rubber o r with acid-resistant material. A sodium tungstate scilution containing 110 to 120 g W 0 3 / l is heated to 80 to 90" and added at a predetermined r a t e to a boiling 25 t o 30% solution of hydrochloric acid. The r a t e of addition is determined empirically, depending on the required particle s i z e of the acid. The purity and the particle s i z e of the acid also depend on the final acidity of the solution, which is maintained in the range between 7 and 12 70. The precipitated tungstic acid must be thoroughly rinsed to remove sodium chloride and other soluble contaminants. The multiple washing is c a r r i e d out f i r s t by decantation with hot distilled water and then in suction-filters which a r e lined with rubber o r some other acid-resistant material. The washing is repeated 6 to 8 times; HC1 o r NH4Cl is added to the water used in the third rinsing to a concentration of 1%in o r d e r to p e r m i t better sedimentation of the tungstic acid. Glass or chlorinated PVC cloth is used a s the filter material. F u r t h e r amounts of water a r e removed from the filtered acid by centrifuging. The total yield at the precipitation stage is 98 to 9970, and the l o s s e s during washing amount to 0.3 to 0.4%. Precipitation of calcium tungstate. The most widely used method for the isolation of tungsten compounds from solution is the precipitation of calcium tungstate (artificial scheelite). A fairly complete precipitation may be obtained because of the low solubility of calcium tungstate, which d e c r e a s e s a s the temperature is increased ( s e e p. 5). At the s a m e time, sodium ions remain in the mother liquor, and the acid obtained has a low content of sodium-containing contaminants. To effect the precipitation, an aqueous solution of calcium chloride is usually added to the sodium tungstate solution. A crystalline precipitate, which s e t t l e s readily, is formed:
Na,WO,

+ CaCI, = CaWO, + 2NaCI.

The completeness of the precipitation of calcium tungstate depends on the alkalinity and concentration of the sodium tungstate solution. It is recommended that the precipitation be c a r r i e d out f r o m solutions heated to boiling, and having a density of 1.14 to 1.16 (120 to 130 g WO3/l) and an alkali hydroxide content of 0.3 to 0.70/0. Incomplete precipitation is observed if the hydroxide concentration is below 0.3 70, while the precipitate formed in the presence of m o r e than 0.770 hydroxide is voluminous, settles slowly, and entrains impurities. Other sparingly soluble calcium s a l t s , CaS04, C a C Q , CaSiQ, Ca3(P0.),, and CaMo04, a r e coprecipitated with calcium tungstate. The concentrations of silicon, phosphorus, and molybdenum in the CaW04 precipitate a r e considerably reduced by the preliminary purification. The calcium sulfate concentration in the precipitate depends on the initial concentration of

23

sulfate ions in the solution and on the excess of the calcium chloride precipitant used. The l a r g e r the excess taken, the higher the concentration of CaS0, in the precipitate. The considerable solubility of CaSO, (2g/l a t 20") p e r m i t s its removal by washing with hot water. Calcium hydroxide a s slaked l i m e may a l s o be used to precipitate calcium tungstate; however, the precipitate then contains m o r e impurities. The precipitati,on of calcium tungstate is c a r r i e d out in iron or wooden vats fitted with stirrers. The extent of precipitation of tungsten from solution is 99 to 99.5%. The concentration of W Q in the mother liquor should not exceed 0.05 to 0.07 g/l. The washed precipitate, as a s l u r r y or a paste, is decomposed by heating with hydrochloric acid:
CaWO.,

+ 2HC1

S,WO,

+ CaCI,.

A high final acidity of the s l u r r y (90 to 1 0 0 g HCl/1) is maintained during the decomposition stage; this e n s u r e s that the phosphorus, a r s e n i c and p a r t of the molybdenum a r e separated from the tungstic acid precipitate (molybdenic acid dissolves in hydrochloric acid). The acid decomposition of calcium tungstate is c a r r i e d out in i r o n r e a c t o r s lined with rubber o r some other acid-resistant coating and fitted with rubber-lined mechanical s t i r r e r s . Hydrochloric acid is heated to 60 to 65" and the aqueous s l u r r y o r paste containing the calcium tungstate is added with constant s t i r r i n g . The spent acid usually contains 0.3 to 0.5 g W Q / 1 , which is recovered a s calcium tungstate by precipitation with lime. The tungstic acid is washed by the method described above. The total degree of extraction is 98 to 9 9 % . The filtered and dried acid usually contains 0.2 to 0.370 of impurities.

6.

PROCESSING OF SCHEELITE CONCENTRATES Fusion with sodium carbonate

In contrast with the fusion of wolframite concentrates, the fusion of scheelite concentrates with sodium carbonate does not e n s u r e a satisfactory leaching of tungsten even if a large excess of sodium carbonate is added to the charge. At the fusion temperature (800 to 900") the reaction equilibrium is shifted to the right, viz., the formation of Na2W04. However, the r e v e r s e reaction, corresponding to the equilibrium:
CaWO,(solid)

+ Na,CQ(solution)

Na2W04(solid) + CaCQ(so1id).

may take place during the leaching of the cake. This is confirmed by the fact that the tungsten concentration in the precipitate i n c r e a s e s as the leaching time of the cake i s increased. According to the data of Zelikman and Ryabova, the equilibrium constant of the above reaction, expressed a s the ratio of the concentrations by weight in the solution a t 90, i. e . , [Na,C031, equals 0.78
INa2WOII

121.

24

The value of the equilibrium constant may be used to calculate the minimum amount of sodium carbonate (over the stoichiometric amount) which must be taken in o r d e r to bring about quantitative dissolution of the tungsten. The amount of Na2CQ required f o r the formation of 294g of Na2W04is 1 0 6 g. In o r d e r to retain the sodium tungstate in solution, it is n e c e s s a r y that the solution contain a certain amount of sodium carbonate, which may be calculated from the equilibrium constant:
[Na2c03
294

= 0.78; [Na,COs] = 294.0.78 = 230

g.

The minimum amount of sodium carbonate theoretically required is therefore 230g + 1 0 6 g = 336g, which is 3 . 1 7 times the stoichiometric amount. In practice, even l a r g e r amounts a r e required. In o r d e r to make the scheelite decomposition p r o c e s s i r r e v e r s i b l e , quartz sand i s added to the charge in an amount calculated to bind the calcium a s the insoluble silicate:
CaWO,

+ Na,CO, + SiO, = Na,WO, + CaSiO, + CO,. + + CO,,

Sodium carbonate most probably f i r s t r e a c t s with the sand, yielding sodium silicate, which subsequently r e a c t s with the calcium tungstate:
Na,CO, SiO, CaWO, Na,SiO,

= Na,SiO,

= CaSiO,

+ Na,WO,.

Thus, the charge used in the processing of scheelite concentrates consists of the finely divided concentrate, sodium concentrate (in a 50 to 1 0 0 % excess over the theoretical amount), quartz sand and waste. The furnaces employed f o r the fusion a r e of the type used in the fusion of the wolframite - sodium carbonate charge. The cake contains sodium tungstate, calcium silicate, sodium silicate, sodium s a l t s of contaminants (molybdenum, arsenic, phosphcrus, etc. ) and the undecomposed fraction of the scheelite. It is processed by the method used to p r o c e s s the cake from wolframite concentrates. The processing of both types of concen t r a t e s is diagramatically shown in Figure 2.

Decomposition in autoclaves with aqueous solutions of sodium carbonate* f 4, 20, 2 6 1 The p r o c e s s is based on the reaction: CaW04(solid) + Na2CQ(solution) e Na2W04( solution)

+ CaCQ(so1id).

A soda leaching process of wolframite concentrates with sodium carbonate solutions in autoclaves has been recently developed / 2 0 / . In order to obtain an effective decomposition it is necessary that the CO, formed in the reaction
FeWO,

+ Na,CO. + H,O

Fe(0H).

+ Ne.WO, + COB

be continuously removed from the system.

25

The studies of Maslenitskii have shown that a t 1 8 0 - - 2 0 0 " the reaction proceeds a t a satisfactory r a t e and to a satisfactory degree of completion. If the solution temperature must be increased above the normal boiling point, the heating is done in autoclaves (hermetically sealed containers) and not i n open vessels. When an aqueous solution is heated to 200" in a n autoclave, its vapor p r e s s u r e reaches 1 5 atm. Syrokomskii and Maslenitskii / 4 / w e r e the f i r s t to develop a method for the decomposition of scheelite concentrates with sodium carbonate solutions in an autoclave. This method gives good r e s u l t s i n the processing of both lean (containing up to 2 0 % W03) scheelite Concentrates and r i c h tailings (containing up to 5 7 0 W03) from enrichment plants, as well as ordinary concentrates. As compared with the fusion of scheelite concentrate with sodium carbonate and sand, the decomposition with sodium carbonate solution has the advantage of not requiring a furnace or the associated preparatory stages. The disadvantage of the autoclave p r o c e s s is that the consumption of sodium carbonate is high. The amount of sodium carbonate required for the quantitative decomposition of the scheelite concentrate depends on its W 0 3 concentration. An amount 2 . 5 to 3 t i m e s the stoichiometric is required f o r the decomposition of rich concentrates. An amount 4 to 4.5 t i m e s the stoichiometric is required f o r the decomposi tion of lean concentrates (15 to 2 0 F WO3). The processing of the con centrates with sodium carbonate solutions is c a r r i e d out in autoclaves of various types : a) fitted with a mechanical s t i r r e r and heated by means of spent s t e a m ; b) heated and s t i r r e d by means of live steam; the mixing is accom plished by bubbling the s t e a m through the s l u r r y ; c) horizontal autoclaves, containing balls rotating with the charge. Heating is c a r r i e d out by means of live steam (Figure 7). Recent investigations have shown that the use of autoclaves containing milling equipment r e s u l t s in an i n c r e a s e in the r a t e of decomposition and a reduction in the sodium carbonate consumption (to 1.75 to 2 . 0 equivalents) in the processing of standard scheelite concentrates. This can be attributed to the abrasive action of the balls, which remove the CaC03 l a y e r f r o m the surface of the scheelite particles. However, the design of ball autoclaves is m o r e complicated than that of other types. The autoclaves a r e built of special nickel steel and ordinary carbon steels. Batch p r o c e s s is generally used in autoclave operation. An aqueous s l u r r y containing the sodium carbonate a n d the scheelite concen t r a t e (particle s i z e 0 . 5 to 0.6") is run from the mixer to the autoclave through a gravity-flow line ( s e e Figure 7). When leaching is complete, the s l u r r y is discharged from the autocalve into the evaporator, in which the p r e s s u r e is lower than in the autoclave (i. e . , about 1 . 5 to 2 atm) and in which the s l u r r y boils vigorously and is thus rapidly cooled. The evaporator (Figure 7) consists of a cylindrical s t e e l vessel with conical upper and lower ends. The s l u r r y is fed from the autoclave into the center of the evaporator through a pipe. The steam formed passed through a splash-head and is discharged through a nozzle. The s l u r r y is run-off through the lower nozzle. The lower end of the evaporator is protected against the thrust of the inflowing s l u r r y s t r e a m by means of a guard made of armor-plated steel. The s l u r r y from the evaporator p a s s e s into a collector and, from there, to the filtering and washing stages.

26

Auxiliary steam

- 1 . .

FIGURE I.

Autoclave processing of scheelite concentrates.

autoclave; 2-feed pipe (used also for the supply of steam); 3-discharge pipe; 4-perforated partition (if balls are used); 5-evaporator; 6-guard (armor-plated steel); I-splash-head; 8-inlet of slurry; 9-discharge nozzle; 10-slurry collector; 11 -reactor for the preparation of the slurry: 12-manometer.
1 -horizonral

A s an illustrative example we have shown below the processing of standard scheelite concentrates containing 4 to 5 y o molybdenum (Figure 8) / 2 9 / . The p r o c e s s i s designed to produce tungstic anhydride and . 5 7 ' 0 WOs. molybdenum concentrate containing 57 to 61% Mo and about 2

Acid decomposition The d i r e c t decomposition of scheelite concentrates with hydrochloric acid according to the equation:
CaW04

+ 2HCl

H,WO,

+ CaCI,.

is widely used in industrial processing of these concentrates ( s e e Figure 9). A s a result of the decomposition, calcium chloride and other soluble impurities p a s s into the solution, while tungstic acid, together with s i l i c a and undecomposed scheelite remains in the residue. The technical grade tungstic acid thus obtained is purified by the ammoniacal method. As compared with alkaline decomposition, decomposition with acids involves a s m a l l e r number of operations. Moreover, when employing the acid decomposition process, a substantial fraction of the impurities is removed already during the decomposition with hydrochloric acid.

21

Scheelite concentrate (contaminated with molybdenum)

Sodium carbonate Preparation of slurry

.1
Leaching in autoclave Filtration

$ Residue

Solution

4
Washing Wash water To waste

7 Separation of silica
Filtration

HCl

4
Silica-containing residue Precipitation of MoS, Filtration and washing

.1

Solution Moss precipitate

3.
Evaporation CaCb \

-1
Conversion into pulp and filtration

.1

Precipitation

.I4
Precipitate Drying and partial calcining Solution Second precipitation

Filtration and washing CaWO, precipitate HCl

A Solution

To waste disposal.

Molybdenum concentrate

Decomposition

Washing and filtration

To waste disposal

7
Solution to waste disposal

H,WO, precipitate Drying Calcination

J.

J.

FIGURE 8.

Processing of scheelite concentrates by the autoclave process.

28

Scheelite concentrate Size reduction Decomposition

Hydrochloric acid + O.Z%niuic

Washing and filtration

. I
Solution (CaC12, FeC13, e t c . )

_____1
Residue (HzWO4. SiO,, undecomposed mineral, etc.) NH40H Dissolution

I
Neutralization

. I
TO

__ -

collecting vessel

1I

.-

----yl

1
Filtration and washing
-

I-- Ammonium tungstate solution Hydrochloric acid

_____I

I
I
I

1
I

Residues (SiOz and undecomposed mi nera 1)

Neutralization (or evaporation) Crystallization Filtration

$--Mother liquor HC1 >ation

7
Ammonium paratungstate crystals Calcination

J.
of tungstic acid

Filtration

.L
Precipitate (containing impurities)

.L Solution

1
To collecting vessel

2
I

FIGURE 9.

Processing of scheelite concentrates by the acid process.

Hydrogen sulfide liberated during the decomposition of sulfide impurities may bring about p a r t i a l reduction of the tungstic acid to a bluecolored, lower-valency compound. This may be prevented by the addition of 0.2 to 0.570 nitric acid.

29

The p r o c e s s is c a r r i e d out in an iron r e a c t o r lined with rubber o r some other acid-resistant m a t e r i a l ; the v e s s e l is filled with concentrated hydrochloric acid (- 250% of the stoichiometric amount) and n i t r i c acid is added. The concentrate is then fed into the reactor. The p r o c e s s is c a r r i e d out f o r 6 t o 8 hours at 70 to 80" with stirring. The residue (which c o n s i s t s of tungstic acid, silica, and undecomposed mineral) is rinsed s e v e r a l t i m e s with hot water (as described on p. 23). The washed technical grade tungstic acid precipitate containing 2 to 3 % impurities is forwarded to the purification s t a t e (see Figure 9). One of the drawbacks of the acid decomposition method is the l a r g e consumption of.hydrochloric acid. This may be counteracted by working in heated ball mills, in which the decomposition and the milling a r e conducted simultaneously, while the tungstic acid layer is removed from the surface of the scheelite p a r t i c l e s f 51. This r e s u l t s in an increased r a t e of decomposition and a reduction of the acid consumption to 15070 of the stoichiometric amount.

7.

PURIFICATION OF TUNGSTIC ACID

The technical grade tungstic acid produced by one of the above methods may contain 0.2 to 3 70 impurities in the f o r m of calcium and sodium s a l t s , silica, molybdic acid, adsorbed iron, manganese and aluminum salts, phosphorus, a r s e n i c and other compounds, and undecomposed scheelite (in the case of tungstic acid produced by the decomposition of scheelite concentrate). The total concentration of SiO, and alkali and alkaline-earth metal s a l t s (which together account for the bulk of the impurities) is expressed as the "chlorination residue". This is the residue obtained on chlorinating an accurately weighed amount of tungstic acid ( o r tungstic anhydride) with d r y chlorine o r HCI a t 800". The chlorination l i b e r a t e s the volatile chlorides of tungsten, iron, manganese, aluminum, phosphorus, etc. The residue contains Si& and the alkali and alkaline-earth metal chlorides. The maximum permissible "chlorination residue'' in tungstic acid o r anhydride used for the production of tungsten wire is 0.1 70; in the anhydride usedfortheproduction of tungsten carbide i t is 0.1 to 0.15% (seeTable 10). Tungstic acid i s usually purified by the ammonia process. Tungstic acid dissolves readily in ammonium hydroxide solutions with formation of a solution of ammonium tungstate. Most of the impurities, silica, iron and manganesd hydroxides, and calcium in the form of CaW04 remain in the residue. The tungsten is precipitated from the ammoniacal solution a s tungstic acid o r ammonium paratungstate.

Dissolution of HzW04in ammonium hydroxide solution When tungstic acid is dried at temperatures not exceeding 170, it readily dissolves in aqueous solutions of ammonium hydroxide, When dried a t higher temperatures, the acid l o s e s i t s chemically combined

30

water and becomes l e s s readily soluble in ammonium hydroxide. Ignition of HzW04at 500" c a u s e s a complete l o s s of water andleads to the formation of tungstic anhydride, which is only sparingly soluble in ammonium hydroxide. Tungstic acid is s e n t to the dissolution stage in the f o r m of a suspension (1kg of WO, p e r 1.5 l i t e r of water) preheated to 80 to 85". The suspension is poured into a r e a c t o r containing a 25 To solution of ammonium hydroxide (the consumption of ammonium hydroxide ( d = 0 . 9 1 ) is 115 l i t e r s p e r lOOkg of WO,). The ammoniacal solution is left to settle f o r 8 to 1 2 h r s and is separated f r o m the residue (SiOz, CaO, etc. ) by decantation. The density of the solution is 1 . 2 9 to 1.30, which corresponds to a W q content of 320 to 330 g / l . The ammoniacal solution is f r e e of the bulk of the impurities originally present in the technical grade tungstic acid. However, it may contain molybdenum, sodium, magnesium and iron salts.

Isolation of tungsten from ammoniacal solutions Precipitation as HzW04. This is c a r r i e d out in the s a m e way a s the precipitation of the acid from Na2W04solutions (p. 22). To obtain high purity H2W04 the product is redissolved in ammonium hydroxide and reprecipitated by acid. Isolation of the paratungstate. 'The paratungstate i s isolated from ammoniacal solution either by evaporation o r by neutralization of the solution. T h e e v a p o r a t i o n m e t h o d . When an ammoniacal solution is evaporated, a p a r t of the ammonia is liberated and ammonium paratungstate is formed. The platelet-shaped modifi cation corresponding to the formula I 0 0 5(N&)20* 12W0,. 5Hz0 crystallizes 3 80 out of solution on cooling. The c r y s t a l s m
a r e separated f i r s t by filtration and n 60 then by centrifuging, and a r e then 4 0 rinsed with cold water and dried. The . . 0 solution i s evaporated to a definite z IO residual volume (calculated to result u in the separation of about 80% of the u zo uo 60 eu MU tungsten) in continuous or batch Precipitation of Mo and W from solution evaporators. The crystallization of a (as the para salts), '70 l a r g e r fraction of the tungsten is not recommended, since it would r e s u l t FIGURE 10. Variation of the amount of in contamination of the Crystals With crystallized ammonium paramolybdate ( 1 ) and paratungstate (2) with the degree of impurities. The degree of purification evaporation of the solution. obtained by the ammoniacal method is shown in Table 8. The tungsten in the mother liquor (containing an increased amount of impurities) is precipitated a s CaW04 o r H2W04and these are appropriately recycled f o r re-processing.

.z -

31

If the s t a r t i n g solution of ammonium tungstate contains molybdenum, ammonium paramolybdate crystallizes out together with ammonium paratungstate. The paratungstate is l e s s soluble than the paramolybdate; hence, molybdenum may be separated from tungsten by fractional crystallization.
TABLE 8 Efficiency of the ammoniacal purification of tungstic acid (according to Smithells j 1/) Product Initial tungstic acid (technical grade) . . Initial solution of H2W0, in NH40H . H,W04 precipitated from the initial solution

I
. . . . . . . . . . . . . . . . . . . . . . .

Total impurity content,%of WO,


3.2 0.2 0.25 (small increase, caused by the presence of iron in the hydro chloric acid) 0.07
0.04

Second solution of H,W04 in NH40H . Ammonium paratungstate crystals after evaporation and crystallization

The variation of the amount of crystallized ammonium paramolybdate and paratungstate with the degree of evaporation of the solution i s shown in Figure 10 / l / . F o r instance, the evaporation of 6 0 % of the liquid r e s u l t s in the precipitation of 5 5 yo of the tungsten and only 1 2 % of the molybdenum originally present in the solution. Thus, the f i r s t fractions of ammonium paratungstate c r y s t a l s may contain only t r a c e s of molyb denum and may be used f o r the production of pure tungsten. The subsequent fractions crystallized from the mother solution have a l a r g e r molybdenum content. T h e n e u t r a 1i z a t i o n m e t h o d Needle-shaped ammonium tungstate (with 11 molecules of water of crystallization) is formed when a cold ammoniacal solution is carefully neutralized; the overall reaction i s :

l2(NHJ,WO,

+ 14HC1+ 4Hz0 = 5(NH,),O.12WO3.11H,O + 14NH4CI

The solution is neutralized by a slow addition of HC1; continuous s t i r r i n g is required in o r d e r to prevent localized overacidification, which may lead to the formation of the metatungstate. The degree of precipitation is strongly affected by the pH, the optimum pH value being between 7 . 3 and 7.4. The solution is allowed to stand f o r a prolonged time (up to 24 hours) when 8 5 to 9 0 % of the tungsten precipitate as the needle-shaped p a r a tungstate. The s a l t is usually very pure. F o r the sake of additional purification i t is sometimes treated with hydrochloric acid to yield tungstic acid: 5 (NHJ20. 12W03 1 lHzO lOHCl = 12H2WO4 + 10NH4CI 4Hz0.

For this purpose, the s a l t is introduced in s m a l l amounts into a r e a c t o r containing concentrated HC1 and the resultant solution is heated to boiling.

32

8. PRODUCTION OF TUNGSTEN TRIOXIDE AND

QUALITY CONTROL Tungsten trioxide is produced by ignition of tungstic acid o r ammonium paratungstate :
HZWO,
+

5 (NH,),. 12WO3.nHZO

--+

WOs lZWO3

+ HZO,
+ IONH, + (n + 5) HZO.

Tungstic acid is completely dehydrated at 500" while ammonium tungstate decomposes quantitatively above 250".

The tungstic acid i s ignited to 750 to 850" in a rotary electric furnace. The furnace (Figure 11) consists of a stainless steel tube located in a heating chamber lined with refractory plates with grooves containing nichrome heaters. The chamber is thermally insulated and enclosed in an iron jacket. The tube i s rotated by means of a motor and reduction gear. The tungstic acid is charged continuously from the hopper into the furnace by means of screw feeders. The discharge end of the tube i s fitted with a drum sieve with 1 . 2 to 3 mm holes. The slope of the furnace may be changed with the aid of a special device. The vapors generated in the furnace a r e drawn o u t through a dust-collecting device (bag filters). Not only the purity of the tungstic anhydride, but also i t s particle size (which depends on the conditions under which the tungstic acid is formed and on the ignition temperature) is of g r e a t importance for the production of tungsten o r tungsten carbide. The particle size of any powder may be described a s i t s bulk weight. This t e r m is defined as the weight of a unit volume of the loose (uncompressed) powder; i t is usually expressed in g r a m s per cubic centimeter. The bulk weight of fine powders is lower than that of coarse powders. This is attributed to the fact that fine powders have a l a r g e r total surface and the action of cohesion (friction)

33

f o r c e s between the particles is stronger than in c o a r s e powders, and prevents the powder from becoming m o r e densely packed. A new and m o r e accurate method f o r controlling the physical s t r u c t u r e of the powder has been recently introduced; it is based on the m e a s u r e ment of the amount of adsorbed methaxiol vapors (the larger the specific s u r f a c e of the powder, the higher the degree of adsorption). The device shown in Figure 1 7 h a s been developed in the USSR f o r this purpose. As is evident from Table 9, the methanol adsorption method is m o r e sensitive than the bulk density method.
TABLE 9 Variation of the physical properties of tungstic anhydride powder with the temperature of ignition of tunstic acid
_.

_ _

sorbed, mg/g
750 800 850

Methanol adsorbed, m g g 0.14


0.06

Bulk density, g/cm3


0.61

0.73
0.19

650 700

0.59 0.47

0.62 0.63

0.04
. . .

TABLE 1 0 Approximate specification for technical grade tungsten trioxide Component Concentration, T o
~~ ~

for hard alloys Tungsten trioxide. . . . . Molybdenum. Arsenic Phosphorus Sulfur Chlorination residue Calcium o x i d e . Sesquioxides Weight losses upon ignition

for metallic tungsten


2 99.95
5
5

. . . . . . . . .......... . . . . . . . . ........... . . . . ...... . . . . . . . .

99.8 0.015 0.025

c 0.1
5 5 5

0.02 0.02

5
5

0.015 0.10 - 0.15

5 5

0.04 0.5

0.01 0.02 0.10 0.01 0.01


0.05

The particle size of tungstic anhydride prepared from ammonium paratungstate is usually l a r g e r than that of the anhydride prepared from tungstic acid. The purity specifications f o r technical grade tungsten trioxide a r e shown in Table 10.

9.

COSTS OF PRODUCTION OF TUNGSTIC ANHYDRIDE

An analysis of the costs of production makes it possible to reduce the expenditure involved.

34

.........

I -

. . . . I

...................

The approximate c o s t of producing one ton of tunstic anhydride from scheelite concentrates by the sodium carbonate method (see Figure 2) is shown below, as% of the total cost f . 0.b. factory:
Raw material -scheelite concentrate Reagents-sodium carbonate, quartz sand, hydrochloric acid. calcium chloride, etc. Auxiliary materials -filter cloth, etc. Energy consumption -electrical, fuel. steam, water Wages with benefits. Overheads

.............. .................. .............. .... ......................... ................................


Total

83.0

9.1 0.2 1.7 2.0


3.4
~

... 100.0

The main item in the cost price (83.0%) is the raw material. Hence, the cost of tungstic anhydride will be reduced in the first place by reducing the consumption of scheelite concentrate required to produce one ton of tungstic anhydride by reducing the l o s s e s occurring in individual processing stages. The total degree of extraction of WO, from the concentrate into the finished product (tungstic anhydride) is now 93 to 9570. The approximate l o s s e s a t each stage a r e shown below in %:
Size reduction and charge mixing. . . . . . . . . . . . . . . . . 0.5 Fusion (loss by dust entrainment) 0.5 Leaching and filtration. . . . . . . . . . . . . . . . . . . . . . . . 1.0 to 2.0 Purification of solution. 0.5 Precipitation of artificial scheelite 0.5 to 1.0 Production of tungstic acid and washing . . . . . . . . . . . . . 1.0 to 2.0 Drying and ignition (dust entrainment) 0.5

.................

....................... ................
.............

In o r d e r to i n c r e a s e the degree of extraction s t r i c t control of all production stages should be maintained to reduce mechanical l o s s e s caused by comminution of the concentrate and transport of the solutions, efficient dust trapping during ignition of the tungstic acid should be ensured and the tungsten should be recovered from the wash liquids from the calcium tungstate decomposition stage and from other wastes. Although the relative contribution of the wages to the cost price is not large, labor should nevertheless be saved by further mechanization and automation of the production process. This w i l l also result in decreased l o s s e s and improvement of the quality of the tungstic anhydride.

10.

PRODUCTION O F METALLIC TUNGSTEN

Because of i t s high melting point tungsten is produced from its compounds a s powder which is converted into metal p a r t s by using the methods of powder metallurgy. The possibility of producing tungsten p a r t s by modern vacuum melting methods (electron beam o r a r c melting) has been investigated over recent y e a r s . Tungsten powder may be prepared by the reduction of tungsten compounds (tungsten trioxide, tungstic acid, o r i t s s a l t s ) with different

35

reducing agents a t relatively low temperatures ( 8 0 0 to 1200"). Powders of various degrees of purity are obtained depending on the purity of the starting m a t e r i a l and the reducing agent used. The highest purity standards m u s t be m e t by powders used in the production of ductile tungsten. Reduction may be c a r r i e d out with the aid of hydrogen, carbon, metals (aluminum, silicon, sodium, e t c . ) or by an electrolytic method. The p r o c e s s e s used in industry a r e based on the reduction of tungsten trioxide by hydrogen or carbon; hydrogen is the only reducing agent used in the production of ductile tungsten. The standards f o r tungsten powder specify not only the degree of chemical purity, but also the physical s t r u c t u r e of the powder (particle s i z e and shape, size distribution, degree of conglomeration). These characteristics determine the behavior of tungsten when the powder is subsequently compacted and the behavior of the resulting metal p a r t s . The physical structure of the metallic powder depends on the structure of the starting m a t e r i a l and the conditions of the reduction.

11. REDUCTION O F TUNGSTEN TRIOXIDE BY HYDROGEN Physicochemical conditions of the reduction p r o c e s s The reduction of tungstic anhydride by hydrogen is a four-stage process / 6 / corresponding to the stepwise formation of four tungsten oxides, W 0 3 , W 0 2 . , (or Wl0029), WOz.75 (or W4011) and W 0 2 according to the following reactions:
10W0,

2W,00,, 3H, 5w,ol1 4H,O 3H, Z 4W0, -1- 3HzO; W,O,, WOz 42H2Z W 2HzO;

+ +

+ H z Z WioO2, + H@; +
+

(1)
(2)

(3) ( 4)

The overall reaction is:


WO,

+ 3 H , z W + 3H,O.

These reactions a r e reversible and their course is determined by the equilibrium constants

where p H Z O and p ~ a,r e the equilibrium partial p r e s s u r e s of hydrogen and water vapor. The effect of temperature on the equilibrium constants of the above reactions is expressed by the following equations / 6 f :
lg KD,= - ~

3266.9

T 4508 Ig Kpl = - T

+ 4.0667;

+ 5.1086;

36

Ig KP, = - 904

+ 0.9054; T

2325 lg K,,, - T

+ 1.650.

The log Kp--

relationship is l i n e a r a s shown graphically in Figure 12.

The graph showing the variation of K, with temperature indicates the optimum conditions f o r the reduction of tungsten oxides. Each value of log K Pon the ordinate corresponds to a fixed water-vapor to-hydrogen ratio in the gas mixture. Let u s assume that a given tem p e r a t u r e (e. g . , 850") the g a s composition corresponds to point A . No reduction of WO, takes place a t this composition a s the water vapor concentration in the gas is above the equilibrium value. In o r d e r to effect the reduction, i t is necessary to reduce the concentration of water vapor by drying the hydrogen. At 850' the reduction of W Q to tungsten could proceed only if the humidity of the hydrogen (i. e . , the below the value corresponding to point B .

-)
PH*

ratio

is

-041

FIGURE 12. Variation of the equilibrium constant of the reduction of tungsten oxides by hydrogen with temperature.

Consequently, either tungsten or some of i t s oxides may be the stable f o r m a t a given temperature, depending on the composition of the gaseous phase (water-vapor-to-hydrogen ratio). Thus, W 0 3 i s the stable form a t 700" and a t a H2 concentration below 23 % in the gaseous phase (WO, cannot be reduced under these conditions); the oxide W4OI1 is the stable form when the H2 concentration in the gas is between 23 and 45%; WO, is the stable form a t an Hz concentration of 45 to 73% and metallic tungsten at a hydrogen concentration above 73 %. The reduction temperature d e c r e a s e s a s the moisture content of the hydrogen is reduced. Thoroughly dried hydrogen must be used in o r d e r to ensure a sufficiently high r a t e of the reduction process.

37

I;
b

FIGURE 13. An eleven-rube hydrogen reduction furnace.

1-rubes; 2-heaters; 3-pushers; 4-hydrogen inlet pipe; 5-hydrogen discharge pipe.

Reduction furnaces The reduction of tungstic anhydride by hydrogen i s c a r r i e d out in stationary multitube furnaces (with continuous o r batchwise transport of the m a t e r i a l undergoing reduction along the length of the furnace) o r in r o t a r y tube furnaces. Multitube electrical furnaces a r e the ones m o s t commonly used. The construction of such a furnace is shown in Figure 13 131. The furnace comprises 11 tubes in two horizontal rows (6 in the lower and 5 in the upper row); the tubes a r e made of stainless chromium-nickel steel. The tube diameter is 50 to 70mm, and the tube length is 5 to 7 m . The tubes a r e enclosed in an iron jacket which is lined on the inside with a thermally insulating m a t e r i a l ; asbestos sheet, diatomaceous o r chamotte bricks o r asbestos cement plates. Directly below and above the tubes there is a layer of molded earthenware tiles with grooves accommodating s p i r a l nichrome-wire heaters, 4 . 5 to 5 mm in diameter.

FIGURE 14.

Rotary-tube furnace for t h e reduction of tungstic anhydride with hydrogen.

1 -furnace casing; 2-lining; 3-shaped cer am i c lining; 4-steel tube; 5-diaphragms; 6 longitudinal shelves; I-motor; 8-pulley; 9-reduction gear; 10-chain: 11-charging hopper; 12-loosening device; 13-screw feeder; 14-discharge hopper; 15-discharging device: 16 e l e c t r i c motor; 11-coupling Joint; 18-reduction gear; 19-base.

There a r e s e v e r a l ( t h r e e o r five) temperature zones within the furnace, whose total length i s about 4 m . The furnace power ranges from 3 0 to 50 kw. The boats containing the tungsten trioxide a r e moved along the tube with the aid of a mechanical pusher. The pusher mechanism, consisting of a g e a r box, a screw, a c a r r i a g e , and an electric motor, is mounted on a special table next to the entry of the furnace. The r a t e of movement of the boats in the furnace may be varied from about 5 to 30mmfminute. The discharge end of the furnace i s fitted with a cooling device in o r d e r to cool boats emerging from the hot zone of the furnace. Hydrogen is fed from a collector into the furnace tubes. The spent hydrogen from the tubes i s collected at the charging end of the furnace and is purified, regenerated, and recycled.

39

In addition to the multitube furnaces described above, continuous r o t a r y d r u m electrical furnaces are beginning to be used in the USSR 1 3 1 . Such furnaces have some advantages -. high productivity, the elimination of manual labor, and lower electrical energy consumption p e r unit weight of product. The electrical furnace designed by Babich (Figure 14) consists of a stainless s t e e l tube (400" in diameter and 4 m long) supported by two r o l l e r s and rotated by means of e l e c t r i c motor and a transmission system. The tube is placed in a heating chamber. The tube contains diaphragms 5 with an opening in their center and longitudinal fins 6 made of angle iron and welded to the tube. The diaphragms prevent the material, which is of a r a t h e r high density from slipping along the tube. The m a t e r i a l is thus kept in a reducing atmos phere in the heated zone for a longer time. The longitudinal fins d i s p e r s e the m a t e r i a l and thus promote better contact between the solid p a r t i c l e s and the hydrogen. The forward p a r t of the furnace accommodates the charging hopper 11 with a loosening device 1 2 and a s c r e w feeder 13. The end of the furnace is connected to the discharge hopper 14 with a s c r e w type discharging device 15. The furnace casing i s mounted on a base 1 9 made of angle iron, sloping a t 2 to 4" in the discharge direction. The length of the furnace is divided into four thermal zones; the temperature of each zone is regulated automatically. Hydrogen e n t e r s the tube through the discharge hopper of the furnace and the direction of i t s flow through the furnace is opposite to that of the material; the hydrogen leaves the furnace a t the upper end and is sub sequently regenerated. The furnace t u r n s a t about 2 r p m . The hydrogen s t r e a m moves a t a r a t e of 18 to 20 m3/hr. Supply of hydrogen to the furnace. The hydrogen fed to the furnace must be thoroughly dried and purified to remove oxygen and water vapor. The use of electrolytic hydrogen (i. e . , hydrogen prepared by electrolyz ing water) is preferable when pure tungsten powder is required. A little sodium hydroxide is added to the water in o r d e r to increase i t s conductivity. The hydrogen produced in electrolytic cells is collected in gas holders from which it is fed into the furnace through a system of pipes after passing through a purification system. The oxygen is removed by passing the gas through a vertical furnace packed with a catalyst (copper turnings or a special Fe-Cu-Ni catalyst) heated to 600 to 650". On the surface of the catalyst the oxygen combines with hydrogen, yieIding water. The water vapor is removed when the hydrogen is passed through a drying system consisting of columns packed with a moisture-absorbing substance (NaOH, CaC12, silica gel, PzO5). The drying is c a r r i e d out most conveniently with the aid of s i l i c a gel (dried and granulated silicic acid) which has a high sorption capacity. The silica g e l is regenerated by heating to 180". There is usually a heater on the column containing the s i l i c a gel. After a certain period, the column i s closed off and the s i l i c a gel i s regenerated. Dried hydrogen is supplied to all tubes of the furnace from the hydrogen delivery line through nozzles cut in the discharge ends of the tubes. In o r d e r to reduce the hydrogen l o s s e s during reduction, the spent hydrogen leaving the furnace is regenerated in a special regeneration

40

setup (Figure 15). At the exit from the furnace 1 the spent hydrogen e n t e r s the collector 2 , and p a s s e s through the t r a p 3 into condenser 4 . The trap is partially filled with water and the hydrogen delivery tube from the collector i s i m m e r s e d in the water. A fraction of the water vapor condenses in the trap. The condensation is completed in the water-cooled condenser tube 4 . The condensate is discarded. The hydrogen leaving the condenser is mixed with f r e s h oxygen-free hydrogen from the supply line and the mixture is passed through the drying tower 5 which contains one of the desiccants mentioned above. The moisture-absorbing particles entrained with the g a s s t r e a m a r e trapped in the bag filter 6 .

The dried hydrogen enters the compressor 7 where the required o-Jerpressure is produced (the required p r e s s u r e s in the system a r e : :00mm water before entering the c o m p r e s s o r and a maximum of 1000" water in the compressor). From the c o m p r e s s o r the hydrogen p a s s e s into the drying tower 8 (with a water supply tank 9 ) for a m o r e thorough drying. From the tower 8 the hydrogen i s fed into the furnace through the explosion extinguishers 1 0 . The explosion extinguisher consists of two metallic cylinders filled with thin tungsten wire; the cylinders s e r v e to extinguish explosions and to disconnect the furnace from the regeneration setup and the hydrogen supply line. The f r e s h hydrogen delivered from the supply line replenishes the volume lost in the reduction and during charging and discharging of the boats. The reduction process In practice, the reduction of tungstic anhydride is a two-stage process. This is due to the volume d e c r e a s e (by a factor of about 3) of the charge in the boats during the reduction of W 0 3 to W (the density of W 0 3 is 7.2 while that of W is 19.3). Because of this volume d e c r e a s e the boats would travel most of their path only partially filled in a one-stage process.

41

The f i r s t stage consisting of the reduction to W 0 2 is c a r r i e d out in one group of furnaces while the second stage (the reduction from W O , to W) in another. In the production of a special type of tungsten, which is used to manufacture non-sagging wire for lamp filaments, the two-stage reduction p r o c e s s may be used f o r the preparation of tungsten powder with a m o r e diversified particle s i z e . For this purpose, the boats entering the second reduction stage a r e loaded with a mixture of tungsten dioxide and tungstic anhydride a t various ratios (25 to 50% W q ) . The reduction conditions selected in accordance with the required particle s i z e of the tungsten powder. Some two-stage reduction conditions a r e listed in Table 11.
TABLE 11 Conditions during the reduction of tungstic anhydride by hydrogen

I
Reduction stage

Conditions rate of motion of the boat, cm/hr hydrogen consumption mrj hr


0.5-0.6 2-3
200 -300
~

I
Purpose

First. Second..

Second.. . .. .

.. . . . . . .... First. . .. ....

Production of fine tungsten powder Production of coarser

0.6-0.8 1.7-2.0

The fine tungsten powder (brand VCh) prepared under these conditions has the following granular composition: 56 to 60% below 0.6 1 ; 3 8 to '$070 in the 0.6 to 1 . 2 1 range; 3 to 5% in the 1. 2 to 1.81 range. The bulk density i s 2.2 to 2.5g/cm3'
The particle size of tungsten powder depends on a number of factors: the most important of them are the reduction temperature. t h e particle size of the starting oxides, and the rate of flow of hydrogen. The effects of the temperature and the particle size of the starting oxides are interdependent. Coarse tungstic anhydride cannot be used for the production of fine powder, but fine powder m a y he used for the production of coarse tungsten powder /28/. This is due t o the growth of oxide particles, whose rate increases with the temperature. The stronger such growth is exhibited by WO, particles, and the weakest by WO, particles. Moreover, the finely divided oxides grow more rapidly than the coarse ones. For instance. WO, particles with a size below 1 p show a noticeable growth rate a t 400", and grow rapidly a t 850. When the particle size is increased from 1 to 5 p, the growth of the WO, particles only starts a t 600 to 800".

It has been established (by observations made during the manufacturing process) that the formation of c o a r s e p a r t i c l e s of tungsten powder during the reduction of tungstic anhydride is mainly favored by: 1) high reduction temperature; 2) steep temperature gradient along the tube; 3) high r a t e of advance of the W 0 3 boats in the tube; 4) thick W 0 3 l a y e r in the boats; 5) low hydrogen flow r a t e ; 6 ) high moisture content in the hydrogen fed to the furnace. Meerson showed that the effect of all these factors on the i n c r e a s e i n the particle s i z e of the tungsten powder may be attributed to the considerable vapor p r e s s u r e of tungsten oxides at the reduction tempera tures, taking into account that W 0 3 is the most volatile 121.
42

The vapor p r e s s u r e over fine particles is higher than over c o a r s e particles; this can b e attributed to the accumulation of surface energy in fine particles. Hence, at high temperatures (- 800") the W Q particles grow because of the vaporization of fine particles and the condensation of the vapors on the l a r g e r particles. The tungstic anhydride not reduced in the low-temperature zone e n t e r s the high-temperature zone of the furnace. At temperatures above 600 to 700" the WO, is partially vaporized and reduced to W Q on the surface of the previously formed c o a r s e W Q crystals, thus assisting the further growth of those crystals. The f i r s t three factors favor the entry of unreduced W Q into the hightemperature zone. The remaining three factors (the increased height of the layer, the low flow r a t e of hydrogen, and the increase in i t s moisture content) have the s a m e effect since they diminish the reduction rate and thus favor the entry of a l a r g e fraction of the W Q into the hightemperature zone. By varying the factors mentioned above, it is possible to vary both the average particle size of the tungsten powder and the size distribution of the particles.

Control of the particle size of tungsten powder


A routine control of the particle size of tungsten powder i s necessary to s e e that it meets standard specifications. Direct and indirect particlesize determination methods a r e used in the control of the particle size of fine powders (below 1 0 , ~ ) .

Pump Particle size, p


FIGURE 16. Curves showing the size distribution of tungsten powder.

vacuum

FIGUW 11. Apparatus for the measurement of the adsorption of methanol.


1-flask, "generator" of methanol vapor; 2-mano meter for measurement of the methanol vapor pressure; 3-manometer for accurate measurement of the pressure within the system; 4-test tubes containing the powder; S-glass container -receiver.

43

Direct determination of the particle s i z e is c a r r i e d out by means of a statistical microscopic method, a s follows: a sample of the powder is mixed and triturated with a solution of turpentine in turpentine oil and a drop of the mixture is placed on a slide on a microscope fitted with an eyepiece s c a l e (or an eyepiece hairline). The r e s u l t s of several measurements (the total number of measured p a r t i c l e s must be not l e s s than 2 0 0 to 300) a r e used to calculate the distribution (in percent) of the p a r t i c l e s by s i z e (in microns) (Figure 16). The indirect methods for determining the granule s i z e composition of the powder include the determination of the bulk density and the determina tion of the relative specific surface of the powder. The bulk density is the weight of a unit volume of the loose (uncompressed) powder. It depends not only on the particle s i z e and the s i z e distribution of the particles, but also on their shape and the degree of roughness of their surface. Nevertheless, the particle s i z e remains the most important factor a s a rule, the l a r g e r the particles, the higher the bulk density of the powder. Another method which is used in the Soviet Union f o r the quality control of tungsten powder is the measurement of the relative specific surface of the powder. The relative specific surface i s obtained from measurements of the adsorption of methanol vapor (CH,OH) or the gas permeability. The apparatus used f o r measuring the adsorption of methanol is shown in Figure 17. The amount of adsorbed methanol i s calculated from the difference in the methanol vapor p r e s s u r e before and after adsorption. The amount of adsorbed methanol (in m g / g powder) is used a s a m e a s u r e of the relative specific surface of the powder. The finer the powder, the, l a r g e r i s i t s specific surface and the higher the amount of the adsorbed methanol vapor. The method of g a s permeability, developed by Deryagin, is based on the measurement of the r a t e of passage of a i r through a powder l a y e r of a given thickness.

12.

REDUCTION OF TUNGSTEN TRIOXIDE B Y CARBON

Tungstic anhydride may be reduced by carbon where the presence of carbon in tungsten is permissible, e. g . , in the production of hard alloys.

Physicochemical conditions for carbon reduction


T h e overall reaction of the reduction of tungstic anhydride by carbon is given by the equation:

W03+X=

W+3CO.

In reality, the reduction proceeds through intermediate stages in which lower oxides are formed (as in the reduction with hydrogen). The main reducing agent is carbon monoxide (CO) rather than carbon; it reduces tungsten oxide as follows*:

4W03

+ CO 2 W,OI~+ COZB

(1)

__--__--------

It is probable that there are actually four reduction stages, as in the reduction with hydrogen.

44

\'101l+ 300 2 4W02

+ XO*.

(2)
( 3)

WOI

+ 2co 2 w + 2c0,.
co, + c 2 2co.

T h e carbon dioxide formed in the reaction reacts ( a t the reaction temperature) with the carbon in the charge as follows:
(4)

The equilibrium constant of these four reactions is the ratio of the partial pressures of CO and CO, :

T h e variation of log

K p with

temperature is given by the following equations:

log Kp, = '555.5

7 - 1.1427,
321.6

log

K,

__

- 0.0647.

Figure 18 shows the equilibrium composition of the gaseous phase ( t h e CO content is in percent) as a function of the temperature, for the four reactions written above. The CO equilibrium pressure curve for the reaction W0,-W intersects the curve of CO equilibrium pressure over carbon a t about 750'. At that temperature. the CO has the s a m e equilibrium pressure in both reactions.

0 V

0" V +
e,

Temperature. ' C
FIGURE 18. Equilibrium composition of the gaseous phase (CO/CO,) as a function of temperature.

Below 750" the W 0 2 cannot be reduced to W since the equilibrium pressure of CO over C is lower rhan the equilibrium pressure in the reduction of WO,. In order to create a high concentration of CO in the gaseous phase and to accelerate the reduction process it is desirable t o carry out the process a t high temperatures (above 1000") where the gaseous phase consists almost exclusively o f carbon monoxide. In contrast to the reduction of W 0 3 by hydrogen. the reduction by carbon produces a fine tungsten powder even when high temperatures are employed (1400 to 1500') and the charge is introduced rapidly into the high-temperature zone of the furnace. T h e charge must not remain in the high-temperature zone for a longer t i m e than that required for a quantitative reduction, or sintering of the powder particles will occur. T h e growth of tungsten particles during the reduction with carbon is inhibited by the presence of a thin layer of carbon black on the surface of the oxide particles. Because of the adsorption capacity of the carbon black, the W 0 3 and lower oxide vapors which are formed a t high temperatures in the pores of the charge are reduced mainly on the surface of the carbon black particles rather than on the surface of the primarily formed crystals of reduced tungsten. Such reduction is favored by the fact that even if the average con centration o f CO, in the reaction space a t high temperatures is quite low, it is still lower on the surface of the carbon black particles and the concentration of the reducing agent (gaseous CO) is a t maximum.

45

Industrial reduction of W 0 3 by carbon


T h e tungstic anhydride used for the production of hard alloys must contain 99.85 to 99.9% WO,.
T h e reduction of WO, is carried out by using pure brands of carbon black, such as l a m p black, or gas
black, whose ash content is only a few hundredths of o n e percent. Usually. t h e amount of carbon added to the charge is equal to the stoichiometric (13.9%). This ensures that the C content in the reduced powder is low (i. e. , does not exceed 0.5%).

,a

Asbestos

81 I
4

8 1

s ! .I

9
FIGURE 19.
1 -casing; 5-cooler;

Carbon tube furnace. 2-graphite (or carbon) heating tube; 3-tubular screen; 4-carbon-black packing; 6-contact cones; 7-current leads; 8-hatch for carbon black; 9- bus bars.

T h e WO, and the l a m p black are mixed in ball mills, pug mills, or some other type of mixers for granulated substances. T h e mixture is swaged in carbon dies or compressed into cylindrical bricks. T h e reduction is carried ouf in carbon tube furnaces (Figure 19). A graphite or carbon tube serves in diamerer. both as the heater and as the furnace space. The tube is 1000 to 1 5 0 0 m m long and I S to 100" As a result of the low mechanical strength of graphite, i t is necessary t o use thick-walled tubes; hence, the electrical resistance of the heater is quite low. For this reason. the furnace requires a high current at a low voltage (15 t o 25V). T h e current is supplied through copper busbars to water-cooled contacts with conical openings in the center. Intimate contact between the carbon tube of the furnace and the current input is produced with the aid of graphite. copper, or cast-iron c l a m p cones of various types. Either carbon black or crushed c o a l is used as the thermal insulator. T h e furnace is contained within a n iron casing fitted with openings for the introduction and discharge of the thermal insulation packing. In order to facilitate the replacement of a burnt-out heater, some types of furnaces have the heater tube installed within a carbon tube screen (Figure 19). A cooler is connected directly to the graphite tube. Using a furnace of the above type. a temperature of more than 2000" m a y be obtained with a 20 to 50 kw heater. The reduction is carried out a t 1450 to 1550'. T h e molds containing the charge or the pressed briquettes move continuously along the furnace tube and remain in the hot zone for 30 minutes or one hour. T h e reduction yields friable gray blocks which a r e crushed and sifted. T h e tungsten contains 0.1 to 0.5%carbon and 0.1 t o 0.157aoxygen. The total concentration of other impurities (Si, Ca. Fe) does not exceed 0.1 to 0.2%.

13. PRODUCTION O F SOLID TUNGSTEN

As mentioned e a r l i e r , solid tungsten is prepared by powder metallurgy techniques, which were developed for the f i r s t time in 1826 by Sobolevskii for the production of platinum ware. The method comprises the following stages:
46

1) pressing of powder (into bricks, b a r s ) ; 2) sintering (heating to a given temperature) ; 3) working into the final product (by forging, drawing o r rolling).

Pressing When the powder is p r e s s e d , the product is an aggregate of a c,ertain shape and p o s s e s s e s a certain mechanical strength. The mechanical strength i n c r e a s e s because the particles have been brought close together and the contact a r e a between them has increased, which r e s u l t s in an i n c r e a s e in the cohesive f o r c e s (which a r e inversely proportional to the distance between the particles) and in the mechanical binding forces between the particles. The freely-flowing powder p o s s e s s e s a certain strength since i t s particles a r e in contact with each other although the contact a r e a is quite small. The strength of the freely-flowing powder is characterized by the angle of repose a, the l a r g e r the angle of repose, the higher the strength of the powder. Fine powders have a l a r g e surface a r e a ; hence, the cohesive force in such powders is higher than in c o a r s e powders; this i s displayed a s a l a r g e r value of the angle of repose and a lower bulk density of the powder. The cohesion and mechanical binding f o r c e s of the powder interefere with i t s compaction since a certain effort must be exerted in o r d e r to overcome these forces. Hence, powders with a lower bulk density a r e compressed with m o r e difficulty.

FIGURE 20. 1 -face

Dies for pressing tungsten powder into bars. pins.

plates; 2 -joint

FIGURE 21. The distribution of powder layers in a pressed

briquette.

Powders with a nonuniform particle s i z e usually have a higher bulk density and produce denser p r e s s e d compacts than powders of uniform particle size. However, the selection of the optimum particle s i z e of the powder is not governed by compressibility alone. The effect of the particle s i z e on p r o c e s s e s occurring during sintering must also be taken into account. The p r e s s i n g of tungsten powder into b a r s is c a r r i e d out in detachable s t e e l dies ( F i g u r e 20).

47

The p r e s s u r e exerted upon the powder during p r e s s i n g is not uniformly distributed within the compact. This is due to the friction between the powder p a r t i c l e s and the walls of the die. As a result, in any vertical section through the compact, the upper l a y e r s (i. e . , the l a y e r s lying n e a r e r to the r a m ) a r e d e n s e r than the lower l a y e r s ; in the upper l a y e r s the density i n c r e a s e s from the center t o the periphery, while in the lower l a y e r s it i n c r e a s e s from the periphery to the center (Figure 21). The nonuniform p r e s s i n g is exhibited especially strongly in the c a s e of compacts of considerable height. Hence, the shape of the dies used in the p r e s s i n g of tungsten powder m u s t be such that the compression takes place along the s m a l l e s t dimension. A lubricating substance - a solution of glycerol in alcohol (in a ratio of 1.5 : 1) o r of paraffin in gasoline (4 to 570 paraffin) - is a l s o added to the powder before the pressing, in o r d e r to prevent stratification and to obtain a compact of uniform thickness. During p r e s s i n g the lubricating solution is squeezed out onto the walls of the die and reduces the friction between the wall and the powder particles.

FIGURE 22.

A 520-ton hydraulic press.

1-press bed; 2 and 3-vertical and horizontal cylinders; 4 and 5-cylinders serving to return cylinders 2 and 3 to the starting position; 6-rings joining the cover to the press bed; 7 -piston rods; 8-directing wedge; 9-plunger; 10-replaceable gib (of the plunger).

The p r e s s i n g of tungsten powder is c a r r i e d out with hydraulic p r e s s e s of 270 to 2000 tons. The p r e s s most commonly used in the Soviet plants is a 520-ton hydraulic p r e s s , P-801 model (Figure 22). The p r e s s e d tungsten b a r s have a c r o s s section of 10 X 1 0 to 40 X 4 0 mm, The p r e s s u r e applied during pressing and a length of 500 to 650". ranges f r o m 2.5 to 5 tons/cm2, depending on the nature of the powder and

48

the s i z e of the bar*. The density of the b a r s produced under those conditions is 1 2 to 13g/cm3, corresponding to a porosity of 3 0 to 40%. The porosity of the pressed tungsten b a r cannot be reduced since the tungsten c r y s t a l s have a very high strength in the cold and a r e not deformed during pressing. When the tungsten particles have been compacted to a degree such that their displacement ceases, any further increase in the p r e s s u r e causes chipping o r lamination of the bar, which under these conditions behaves as a compact body. The limiting p r e s s u r e (above which lamination takes place) i s known as the critical p r e s s u r e and i t s value f o r tungsten is 4 to 6 tons/cm2.

Sintering The sintering of tungsten b a r s is a two-stage process: the first stage is a low- temperature presintering intended to increase the strength and the electrical conductivity of the b a r , while the second stage i s hightemperature sintering. Low-temperature presintering. The presintering is c a r r i e d out at 1150 to 1300" in a hydrogen atmosphere. The b a r s remain in the hot zone of the furnace for 3 0 to 120 minutes (depending on the s i z e of the b a r ) . Occasionally, the presintering is c a r r i e d out in two s t a g e s : first at 850 to 900" (this removes volatile substances such a s glycerol, alcohol) and then at 1150 to 1300". The compressed b a r s have open porosity and therefore the hydrogen diffuses into the p o r e s during sintering and reduces the thin oxide l a y e r s formed on the crystal surfaces. The reduction produces a true metallic contact between the c r y s t a l s in the b a r . The fine metallic c r y s t a l s produced in the reduction of the oxide films a r e located between the l a r g e r p r i m a r y tungsten crystals. They promote sintering of the particles because of their high surface activity. Presintering results in a noticeable strengthening of the b a r s . These a r e subject to linear contraction which may be as high a s 2 to 3 % of the initial length. Low-temperature sintering is c a r r i e d out in electric tube or muffle furnaces (Figure 2 3 ) . The muffle ( o r tube) of the furnace i s made of alundum ( A l z Q ) which does not soften up to 1600"; a molybdenum-wire heater is used. In o r d e r to protect the molybdenum heater against oxidation, the furnace casing (which is insulated with chamotte bricks) is made gas-tight and hydrogen is supplied continuously through a nozzle in the furnace cover. Since the alundum is porous, the hydrogen p a s s e s

_-_--_----___

Hence, t h e m i n i m u m :ated power of a press for the smallest hars would he:

10

500

x2.5

= 125

tons,

IO0

while for t he largest bars i t would he:

49

from the casing into the furnace space and mixes with the main s t r e a m of hydrogen which e n t e r s the furnace f r o m the discharge end through the cooler of the furnace. The p r e s s e d b a r s a r e mounted in nickel boats whose bottoms a r e covered with a thin layer of tungsten powder. Each boat contains from 3 to 3 0 b a r s , (depending on t h e i r size).
C B

FIGURE 23. Muffle furnace used in bar sintering.


1 -iron casing: 2-alundumtube with molybdenum heater; 3-thermal insulation (alundum powder packing); 4-hydrogen supply tube to the casing; 5-cooler; 6-tube for the supply of hydrogen to the furnace space; 7-water entry and outlet tubes.

High-temperature sintering. The preparation of massive metal having the most favorable s t r u c t u r e for subsequent mechanical working r e q u i r e s that the b a r be heated to about 3000" during the si,ntering. Such a high temperature is achieved by the d i r e c t passage of an electric c u r r e n t through the presintered b a r . In practice, this operation is known as "welding" and is c a r r i e d out in a special "welding" apparatus (Figure 24). The b a r is clamped (in a vertical position) between two contacts consisting of copper blocks containing two tungsten plates connected by a spring to form a clamping device. The blocks have grooves for cold water circulation. The water-cooled copper tube which supplies c u r r e n t to the lower contact p a s s e s through an opening in a s t e e l plate fitted with a c i r c u l a r groove. The groove contains a rubber ring which supports a water-cooled copper cap. During sintering, dry hydrogen is fed continuously from below the cap a t a r a t e of 0.8 to 1.0m3/hr. The lower contact must be movable a s a noticeable linear contraction of the b a r takes place during sintering (the length of the b a r is reduced by 1 5 to 1 7 % ) and f i r m clamping may r e s u l t in destruction of the b a r . The mobility of the lower contact is ensured by supplying the c u r r e n t by means of flexible busbars. A counterweight i s used to tighten the bar. The power needed f o r sintering the b a r is dissipated mainly a s radiation from the incandescent surface of the b a r and the clamps. Only a s m a l l fraction of the heat (1 to 2%) i s used in heating the bar. Hence, the maximum power required for sintering depends on the b a r surface and the temperature. The radiation intensity from a unit surface of the incandescent metal is proportional to the fourth power of the absolute temperature of i t s s u r f a c e

50

where E is the radiation intensity (watt/cm2), and c is the Stefan-Boltzmani? constant.

FIGURE 24.
sintering.

Apparatus for high-temperature bar

1 -steel plate; 2-cover; 3-busbar: 4-upper stationary contact: 5-lower movable contact: 6 -clamps; 7-current-supply busbars; 8-counter weight: 9-tungsten bar.

The radiation intensity emitted by tungsten at various temperatures is shown on page 2. At a sintering temperature of 3030" (3300K) the radiation intensity i s 245 watt/cm2. Making use of these data, it is possible to calculate the energy radiated by the entire surface of the b a r (i. e . , the approximate power required) by means of the equation:

= S.245,

where W i s the total radiation energy in watt and S is the incandescent surface a r e a (cm'). F o r the smallest b a r (10 X 1OX 500") the required power would be W = 1 x 5 0 x 4 x 2 4 5 = 4 9 0 0 0 watt = 4 9 k w , while for the largest b a r it would be (40 X 40 X 650") W~ 4 x 6x 5 4 x 245 =254800 watt = 254.8 kw.

51

These r e s u l t s a r e somewhat low since the energy radiated by the end s u r f a c e s of the b a r and the tungsten plates holding the b a r ends has been neglected. Moreover, a p a r t of the heat is l o s t b y convection through the hydrogen atmos phere. The various heat l o s s e s a r e tabulated in Table 12.
TABLE 12
Distribution of heat in the "fusion" of tungsten bars,% Item (heat losses) Bar size, m m
~

32x32~450

Heat lost through radiation Heat lost through radiation Heat lost through radiation clamps Heat lost through hydrogen Other losses

... .... ................................. convection .........


...............................

from the bar surface from the bar ends from the holding

I
76.0 0.5 13.0 0.5 10.0

I
76.0 5.5 8.0 0.5 10.0

The external view of a group of sintering devices is shown in Figure 25.

FIGURE 25. External view of a group of sintering furnaces.

Since the tungsten b a r has a low electrical conductance ( a t a relatively l a r g e c r o s s section - 1 cm2 and up, and a length of 500 to 650 mm), it requires a l a r g e low-voltage c u r r e n t to heat i t to 3000". Thus, a c u r r e n t of 2500 amp is required f o r a b a r with a c r o s s section of 1 0 X 1 0 mm, while 10,000 to 12,000 amp is required for l a r g e b a r s . The voltage applied to the end of the b a r is usually 10 to 20V. Hence, the c u r r e n t is supplied to the sintering furnace through a stepdown transformer, while an autotransformer (connected to the highvoltage side of the step-down transfbrmer) is used for continuous adjust ment of the c u r r e n t strength. The consumption of electrical energy may be reduced through the use of s e v e r a l thin cylindrical molybdenum s c r e e n s which a r e placed concentrically round the b a r . Such s c r e e n s reflect a fraction of the radiated energy and reduce the consumption of electrical energy by about 20 to 3 0 % . However, heat l o s s e s by convection through the hydrogen reduce the effectiveness of screening.

52

The simultaneous sintering of s e v e r a l b a r s , which is a recent p r o c e s s , p e r m i t s a noticeable reduction in the consumption of electric energy and increase in the output 1 3 0 1 . Six to eight b a r s a r e sintered simultaneously in one unit (Figure 26). The b a r s a r e placed in a single row. As a result, energy l o s s e s through radiation a r e lower than is the c a s e with the s a m e number of b a r s sintered separately since t h e r e is mutual screening of the inner surfaces of the b a r s . The c u r r e n t p a s s e s successively through a l l the b a r s , thus increasing the electrical resistance by a factor of 6 to 8 and permitting the u s e of lower c u r r e n t s and higher voltages for heating the b a r s to the sintering temperature. The b a r temperature is determined by the current intensity, which is proportional to the power supplied. Hence, a t a constant resistance of the b a r s (constant s i z e and powder structure) the sintering conditions a r e established and controlled by adjusting the c u r r e n t intensity. Accordingly, one o r two sample b a r s a r e used to find the c u r r e n t required to fuse the b a r (the fusion current) before the sintering of a group of b a r s is c a r r i e d out. The b a r is fastened between the FIGURE 2 6 . Sintering apparatus for a group of contacts, the cover is s e t in place, bars. a hydrogen atmosphere i s produced, and I --base; 2-cover: 3-contact heads: 4- conthe current is increased continuously nector: 5- riiolybdenuiii screen: 6-bars. until fusion of the bar occurs. Sintering conditions for the whole group of b a r s a r e thus established. During sintering, the c u r r e n t is increased within 8 to 10 minutes to the maximum value, i. e . , to 88 to 95% of the fusion current. The current is maintained at the maximum level for 1 2 to 15 minutes, and i s then switched off. The cover is removed 4 to 5 minutes after the interruption of the current, and the b a r s a r e taken out':<. The b a r contracts during sintering, and i t s density i n c r e a s e s from 1 2 to 17.5 to 18.5g/cm3. The residual porosity is 10 to 1570. The external appearance of p r e s s e d and sintered b a r s i s seen in Figure 27. A recent practice i s to use automatic control instruments,which a r e programmed to adjust the intensity of the c u r r e n t passing through the b a r 1 3 1 . Such control yields homogeneous b a r s and makes for better working efficiency.

* In

the case of some brands of tungsten sintering is carried out in two stages. In the first stage the ciirrent is increased to 40 t o 507oof the fusion current and is maintained a t that level for 5 t o I minutes. This produces vaporization of most of the additives (see p. 56) which are then deposited on the inner parts of the apparatus. T h e second stage of sintering is carried out in another vessel.

53

FIGURE 27.

Tungsten bars. b-sintered.

a-compressed;

Sintering mechanism

/ 7/

The sintered tungsten has the polyhedral s t r u c t u r e characteristic of compact metals (Figure 28). However, in contrast to the metal prepared by melting, the tungsten prepared by sintering still p o s s e s s e s some porosity which i s eliminated only through mechanical working of the b a r (forging, drawing).

FIGURE 28. Structure of a sintered tungsten bar.

FIGURE 29. Degree of saturation of at o m i c force fields in crystals.

The p r o c e s s e s of contraction and particle growth taking place during sintering a r e associated with the increased mobility of atoms at the sintering temperature. It is well known that the atoms in the crystal lattice of a solid have a certain vibration amplitude which i n c r e a s e s with increasing temperature. In contrast to the atoms inside the crystals, the atoms on i t s s u r f a c e have unsaturated force fields (Figure 29), which r e s u l t s in an excess of f r e e energy which i s exhibited a s surface tension and tends to reduce the f r e e a r e a of the c r y s t a l s . As the temperature i s increased, there is an i n c r e a s e in the mobility of surface atoms which move (migrate) from lattice points ( l e s s stable s i t e s ) into cavities and particle-contact points a t which the excess of f r e e energy is l e s s . The atomic configuration thus becomes m o r e stable. The surface migration of the atoms r e s u l t s in an increased smoothness of the f r e e s u r f a c e s and spheroidization of the pores. The total a r e a of contact between the p a r t i c l e s increases, a s is evident from Figure 3 0 . However, the surface migration of atoms cannot be the cause of the contraction of the sintered b a r s , since the total pore volume remains the same.
54

The contraction i s caused by surface tension which tends to reduce the total f r e e surface in the sintered b a r . The p a r t i c l e s a r e deformed and "flow" into the pores. Such deformation i s made possible by the fact that when the temperature i s increased, the strength of the c r y s t a l s d e c r e a s e s m o r e rapidly than does the surface tension.

FIGURE 30. The increase in contact areas a n d the spheroidi zation of pores as a result of the surface migration of atoms.

a-before

sintering; b-after

sintering.

Recrystallization (growth of some particles at the expense of o t h e r s ) , which begins at a given temperature, promotes rapid movement of the atoms and activates the forces cf surface tension causing the contraction. Noticeable contraction usually also takes place during sintering a t tempera t u r e s below the crystallization temperature. It i s associated with the volume deformation of the particles which has been mentioned above. Structure of the b a r s The sintered b a r s should have a fine-grained uniform s t r u c t u r e . The formation of such s t r u c t u r e depends on s e v e r a l factors: temperature changes during sintering, grain s i z e of the initial tungsten powder, and the presence of small amounts of special additives (Si@, KzO, Na20, ThOz, AkQ, etc. ) . During high-temperature sintering, the tungsten b a r i s heated in a non uniform manner. There is a temperature drop between the center and the periphery, a s well a s between the middle of the b a r and i t s ends. The b a r surface is cooled by heat radiation, while the b a r ends a r e a t a lower temperature, being in contact with the cool clamping device. The tempera ture difference may cause a f a s t e r growth of the c r y s t a l s . The s t r u c t u r e of the b a r may be controlled by changing the r a t e of heating: a b a r with a fine-grained s t r u c t u r e is obtained by a rapid i n c r e a s e of the temperature in the range of accelerated c r y s t a l growth ( 2 6 0 0 to 2800"). The s t r u c t u r e of the b a r is also affected by the grain s i z e of the initial powder. Coarse powders with an average particle s i z e of 8 to l o p a r e not suitable for the production of compact metal, and neither a r e powders which a r e too fine. The l a t t e r have a m o r e pronounced tendency towards grain growth than the c o a r s e powders. Powders with a particle s i z e of 0.5 to 5 to 6 f i (average particle s i z e = 2 to 3 p ) a r e used f o r the production of metal with a finegrained uniform structure.

55

1 1 1

I,

The b e s t workability during the subsequent hot-forging operation is obtained in the case of sintered tungsten b a r s with a uniform s t r u c t u r e and an average grain s i z e of about 2 2 p (i. e . , 2000 g r a i n s p e r mm'). The typical s t r u c t u r e of a b a r sintered a t a high temperature is shown i n F i g u r e 28. In o r d e r to control the recrystallization taking place in incandescent tungsten w i r e s used in e l e c t r i c lamps and electronic devices, and in o r d e r to obtain w i r e s with a predetermined s t r u c t u r e , a s m a l l amount of an additive, thorium oxide, aluminum oxide, o r a silica- containing mixture (SiOz+Na20+Kz0) is added t o the tungsten intended for use in wire manufacturing; The additives are usually added as a solution of the respective s a l t s to a suspension of tungstic acid in water. The components are thoroughly mixed, the mixture is evaporated, dried, and forwarded to the reduction stage. The composition of the additive and the amount of i t introduced into the tungsten determine the brand of the finished product. Most of the silica and the alkali metal oxides a r e vaporized during sintering and entrain with them other admixtures (CaO, MgO, e t c . ) . This cleanses the surfaces of the s e p a r a t e c r y s t a l s (the admixtures a r e present as films on the intercrystalline boundaries) and thus promote the rapid sintering and i n c r e a s e in density of the b a r s . The thorium and aluminum oxides a r e not expelled during sintering and affect the s t r u c t u r e of the sintered b a r , depending on the amount added. As a rule, thorium oxide at concentrations of 0.75 to 1.1 qo i n t e r f e r e s with c r y s t a l growth during sintering.

Quality control of the sintered b a r s The p a r a m e t e r s checked in the quality control of the b a r s a r e : external appearance, physical properties, chemical composition, and micro structure. The b a r s must have a uniform, somewhat lustrous surface, and m u s t not absorb ink lines traced on their surface. This is an indication of the required porosity. The curvature deflection must not exceed 4 mm for a b a r length of 300". The difference between the thicknesses of the ends m u s t not exceed 0.7mm, and the difference between the lengths of the b a r f a c e s (over i t s c r o s s section) must not exceed 0.4". The b a r s must contain a t l e a s t 99.870 W, and the maximum permissible impurity contents in 70 a r e :

...........0.02 ..............0.005 CaO. ............0.015 SiO, ............. 0.01 M o . . ............ 0.04


Ln,O,. Ni

The density of the b a r s , a s measured by hydrostatic weighing on a technicalbalance, m u s t b e within 17.5 to 18.5g/cm3. As mentioned above, the b a r must have a uniform structure, which is controlled by examination under the microscope. The number of g r a i n s p e r mm2 must be between 800 and 2000, corresponding to an average grain s i z e of 35 to 2 2 ~ .

56

111

I I, I

I 1

I111

In addition, the quality of a tungsten b a r is determined by i t s workability; to test this, 7 or 1 0 sample b a r s a r e taken out of each batch and subjected t o all stages of mechanical working.

Mechanical working of sintered b a r s The sintered tungsten b a r s a r e b r i t t l e and cannot be forged a t room temperature. They have a very low tensile strength and their elongation is virtually zero. In contrast to the usual behavior of other metals, fine-grained crystalline tungsten b a r s (10,000 to 12,000 grains/"') a r e m o r e brittle than coarse-grained b a r s (1000 to 5000 grains/"'). At high temperatures the b a r s may be forged and then drawn. If subjected to gradual deformation, with intermediate annealing periods, t h e i r ductility i n c r e a s e s . Ultimately, wires with a diameter of 0.01 to 0.015" may be drawn. The changes in tungsten s t r u c t u r e brought about by mechanical working a r e shown in Figure 31. The polyhedral s t r u c t u r e becomes fibrous.

FIGURE 31. Changes in the structure of a tungsten bar as a result of forging. a-sintered bar; b-forged rod, 2 m m i n diameter.

Forging i s c a r r i e d out on a special rotational forging machine (Figure 32). The heated b a r i s subjected to a l a r g e number of s t r o k e s (10,000 to 12,000 p e r minute) by two forging dies which rotate very rapidly around the tungsten rod a s axis. The forging machine consists of a c a s t iron stationary base 1; the base c a r r i e d r o l l e r s 3 mounted s o a s to form a c i r c l e . A shaft 2 rotates in the space between the r o l l e r s ; the shaft has a central channel to accommodate the bar and a groove containing two loosely held forging dies. As the shaft i s rotated, the dies move in alternation from the center to the periphery by centrifugal force and from the periphery to the center (by the projecting r o l l e r ) striking the b a r . The working p a r t of the forging dies has grooves whose s i z e corresponds to the diameter of the rod being worked. The number of s t r o k e s p e r minute depends on the number of r o l l e r s and the rotational velocity of the shaft: at a velocity of 1 0 0 0 rpm with 10 r o l l e r s the number of s t r o k e s is 10,000 p e r minute. The dies a r e made of fast-cutting steel. Before forging, the b a r s a r e heated in a molybdenum-coil furnace, in a hydrogen atmosphere. A s the diameter of the rod d e c r e a s e s , the forging

57

temperature i s reduced from 1350 to 1200". In forging from a diameter of 9 to a diameter of 5 m m , the b a r s a r e manually fed to the forging machine. When the diameter i s reduced the b a r s a r e fed to the next machine in which the dies have a s m a l l e r diameter. In the next forging stage (reduction of the diameter from 5 to 2 ")the rods a r e fed mechanically to the forging machine by means of a special conveyor. The rod f i r s t p a s s e s through a gas-heated furnace in which i t s temperature is increased to 1300" (Figure 3 3 ) . In o r d e r to prevent oxidation and to reduce wear of the dies, the rod is preliminarily coated with a layer of aquadag ( a mixture of fine graphite and ammonia water containing s u g a r ) .

FIGURE 32.

Design of a rotational forging machine. 4-steel ring; 5-forging dies; 6

1-base; 2-forging ram; 3-rollers; steel rings t o c l a m p the [large] ring.

The hot b a r s a r e drawn f i r s t on l a r g e drawing machines fitted with hard-alloy draw plates (here the diameter is reduced from 2 to 0.5") and then on medium and fine-drawing machines (from 0.5 to 0.01 mm). Diamond dies a r e used for drawing wires thinner than 0.3 mm.

FIGURE 33. Positioning of installation in the forging of tungsten.


1-bar to be forged; 2- furnace; 3- forging machine; 4-transporting device; 5 -guiding table.

FIGURE 34.

Setup for the drawing of tungsten wire. furnace; 4-Iubiicating box;

1-guiding drum; 2-draw plates; 3-gas 5-release drum.

58

I
The drawing of tungsten wire is schematically shown i n Figure 34. The wire p a s s e s successively through a lubricating box containing aquadag, and then through a g a s furnace where it is heated. It is then squeezed through the hot drawing plates and is wound on the guiding drum. Depending on the w i r e diameter, the drawing temperature v a r i e s from 800 to 550".

59

Chupter II
MOLYBDENUM
14. GENERAL DATA ON MOLYBDENUM
Brief historical note The t e r m molybdenum comes from the Greek word "molybdos". Up to the eighteenth century the t e r m molybdenum was used to designate lead, galena, many lead-like minerals and graphite. The most common molybdenum mineral, molybdenite, was f o r many centuries believed to be a modification of graphite, which it r e s e m b l e s in appearance. The element molybdenum was discovered in 1778 by the Swedish chemist Scheele. He isolated molybdic acid by decomposing molybdenite with nitric acid and he prepared a number of molybdic acid s a l t s . Metallic molybdenum was obtained for the f i r s t time in 1781 by Hjelm, a compatriot of Scheele, by reducing molybdenum trioxide with carbon. The metal, in a p u r e r form, was prepared a t the beginning of the nineteenth century by Berzelius, who used hydrogen a s the reducing agent. At the end of the nineteenth century it was found that addition of molybdenum imparts to steel a high strength and a capacity for s e l f quenching. Extensive production of molybdenum s t e e l s began in 1910, the y e a r in which it was found that molybdenum i m p a r t s special properties to gun s t e e l s . Later molybdenum became a most important alloying element in various steels. Commercial production of metallic molybdenum and i t s use in electrical technology began at about the s a m e time a s the production of tungsten in 1 9 0 9 to 1 9 1 0 when a powder-metallurgy method was developed for the production of those metal's in a solid form.

P r o p e r t i e s of molybdenum The physical, mechanical, and chemical properties of molybdenum resemble t.hose of tungsten, but t h e r e a r e certain differences. The physical properties of molybdenum a r e listed below.
Atomic n u m b e r . . .................................... Atomic weight ...................................... Density. gl c m 3 ......................................
42

95.95 10.2

60

Lattice type and parameter

...........................

Melting point, " C Boiling point, ' C ........... Superconductivity transition


Heat of fusion, cal/g ................................. Heat of sublimation, c a l l g Heat capacity, cal/g. " C (20 to 100") Heat conductivity, c a l l c m .set. ' C ( a t 20' Coefficient of [thermal] expansion (25 to 700')

.......

Body-cen!ered a = 3.14 A
2620

cubic

..........

4800 0.9 to 0.98

50
1620

10

...........................

....

0.065

0.35 5.8 t o 6.2. IO-'

..........

Resistivity, ohm.cm.10' a t temperatures ( " C ) :


20 130

1130

................ ... ....................................... ........................... ...............................

5.11 23.9 41.1

53.1

Radiation intensity, watt/cm2 a t temperatures (OC):


730 1330 1730 2330

................ ........ .......................................


....................................... ..... ..................

0.55 6.3 19.2 70

Electron work function, eV ........................... 4.37 2.6 Thermal neutron capture cross section, barns . . . . . . . . . . . Brinell hardness. k g / m m 2 : sintered bar ............................... 150 to 160 2 mm sheet .............................. 240 to 250 annealed wire .............................. 140 to 185 Tensile strength of wire, k g / m m z : unannealed (depending on the diameter). . . . . . . . . 140 to 260 annealed (elongation 20 to 257"). . . . . . . . . . . . . . . . 80 to 120 Modulus o f elasticity o f wire. kg/mmz .................. 28,500 to 30.000

Molybdenum is one of the highest melting metals. Only tungsten, rhenium, and tantalum have melting points higher than molybdenum. Noteworthy physical properties of molybdenum a r e i t s high boiling point and electrical conductance (which is lower than that of copper but higher than that of i r o n o r nickel) and i t s relatively low l i n e a r expansion coefficient (about 3 0 % of that of copper). The hardness and the tensile strength of molybdenum a r e lower than those of tungsten. Molybdenum i s readily worked under p r e s s u r e . The mechanical p r o p e r t i e s a r e strongly affected by the degree of purity of the metal and i t s mechanical and thermal history. One of the important properties of molybdenum is i t s low thermal neutron capture cross section (about one seventh of that of tungsten), which p e r m i t s i t s use a s a s t r u c t u r a l m a t e r i a l in atomic r e a c t o r s . Molybdenum is stable in a i r a t ambient t e m p e r a t u r e s . Slight oxidation (to an i r r i d e s c e n t color) is observed a t 400". The metal is rapidly oxidized above 600, with the formation of IVIoQ. Above 700" molybdenum i s rapidly oxidized by s t e a m to the dioxide, MoOz: Mo + 2H,O

MOO,

+ 2H,.

Hydrogen does not r e a c t with molybdenum up to the melting point of the metal. However, some gas is absorbed, with the formation of a solid

61

111

1 1 1 1

1111

111.11111111111

111111111111III1111111111111111111111 I

1111III I

I 11111111 11111 1.11111l11

Ill IIII I

II

solution, when the metal is heated in a hydrogen atmosphere. At 1000 the solubility of hydrogen is 0.5 cm3 p e r 100 g of the metal. Molybdenum r e a c t s with nitrogen at temperatures above 1500" with formation of the nitride. At low nitrogen p r e s s u r e s ( - 0.01 m m Hg) no such reaction is observed up to 2400". The metal i n t e r a c t s with solid carbon, hydrocarbons, and carbon monoxide at 1100 to 1200" with the formation of the carbide MozC. Fluorine attacks the metal at ambient temperatures, while chlorine attacks i t above 250" when the volatile halides, MoF, and MoC15 respectively, a r e formed, Iodine vapor does not r e a c t with molybdenum. Molybdenum r e a c t s with bromine at high temperatures. Molybdenum disulfide, MoSz, is formed in the reaction of molybdenum with s u l f u r vapor above 440" or with hydrogen sulfide above 800". Sulfur dioxide oxidizes the metal at 700 to 800". Silicon r e a c t s with molybdenum above 1200". The disilicide MoSO, formed is very stable in the a i r up to 1500". Molybdenum is resistant t o hydrochloric and sulfuric acids at room temperature, but dissolves to a certain extent in these acids at 80 to 100". Molybdenum is dissolved slowly by nitric acid or aqua r e g i a in the cold, and rapidly at elevated temperatures. Molybdenum is not attacked by hydrofluoric acid, but dissolves rapidly in a mixture of nitric and hydrofluoric acids. A mixture of five volumes of nitric acid, three volumes of sulfuric acid, and two volumes of water is a useful solvent for molybdenum. Molybdenum is not attacked by cold solutions of alkali hydroxides. Some etching takes place in hot solutions. The metal i s rapidly oxidized by fused alkalies, especially in the presence of oxidants, with formation of molybdates. The properties of molybdenum compounds Molybdenum belongs to group VIB of the periodic table. Its most common valency i s six. Its most important compounds a r e molybdenum trioxide, molybdic acid, and the molybdic acid salts, molybdates. There a r e also compounds of bi-, t r i - , t e t r a - , and pentavalent molybdenum. Oxides. Molybdenum f o r m s a s e r i e s of oxides of which the trioxide MOO, and the dioxide MOO, a r e the most stable. There a r e also some intermediate oxides, including a relatively stable oxide, Mo,Ol,. M o l y b d e n u m t r i o x i d e MOO,i s formed in the oxidation of molybdenum or its lower oxides, and in the roasting of the mineral molybdenite, MoS2. Molybdic anhydride is a white powder with a. faint greenish hue. Upon heating Moo3 a s s u m e s a bright-yellow color.
The density of MOO, is 4.69. Its melting and boiling points a r e 795 and 1155" respectively.
At 800 to 1000" molybdenum trioxide vapor exists a s the associate (MOO,), i 231. Noticeable vaporization of molybdenum trioxide takes place above 600 to 650" (see Table 16). The heat of formation of MOO, i s 178f 1.5kcal/ mole. Hydrogen at 800 to 900" reduces molybdenum trioxide to the metal. The approximate solubility of M o Q in water at 20" is 0.4 to 2 g / l . The resulting solution is acidic (pH = 4 to 4.5).

62

Molybdenum trioxide dissolves in hydrochloric and sulfuric acids. M o Q dissolves in aqueous alkalies and ammonia, with formation of molybdates. M o l y b d e n u m d i o x i d e M o o z is a dark-brown powder produced by reducing M o Q with hydrogen a t 450 to 470". Its density is 6.34 and i t s heat of formation is 141 kcal/mole. Molybdenum dioxide is virtually insoluble in water, aqueous solutions of alkali hydroxides and non-oxidizing acids. Nitric acid oxidizes Moo2 to MOO,. T h e i n t e r m e d i a t e o x i d e Mo4Oll is formed by reduction of M o Q with hydrogen, by careful oxidation of M o Q , and by heating a mixture of MoQ and M o Q or M o Q and molybdenum powder in an i n e r t atmosphere ( e . g. , in nitrogen). Mo4Oll has a blue-violet color. The oxide is sparingly soluble in water, sulfuric and hydrochloric acids and dilute solutions of alkali hydroxides. Molybdic acid and molybdates. White voluminous precipitates of hydrated molybdenum trioxide a r e formed when molybdate solutions a r e treated with acids. The dihydrate MoQ-ZHzO ( o r HzMoO4.H20) is stable up to 61; the monohydrate MoQ.Hz0 ( o r HzMoO,) is stable between 6 1 and 120". Elimination of water, with formation of MOO,, takes place above 120". Molybdic acid i s sparingly soluble in water, i t s solubility at 15 and 80" being 2 . 1 2 6 and 5.185 g/1 respectively. It is, however, readily soluble in strong inorganic acids. Molybdic acid s a l t s a r e known as normal molybdates. Their composition corresponds to the formula Me2O.MoQ o r MezMoO,, where Me20 is a monovalent metal oxide. Molybdic acid can add on various numbers of M o Q molecules with formation of polyacids whose composition corresponds to the general formula x H p O . yMoO,, where y > x . The s a l t s of these acids are known a s polymolybdates. In contrast with the normal molybdates, the MezO: M o Q ratio in the polymolybdates is s m a l l e r than unity and v a r i e s over a wide range. Thus, f o r instance, the following types of molybdates are known:
Me,O. 2Mo0,5Me20. 12Mo0,
Me20'3M003)

dimolybdates -paramolybdates -metamolybdates decamolybdates and many others.

Me2,0.4M00, Me,O. lOMo0,-

Polymolybdates a r e formed in the neutralization of a solution of an alkali metal molybdate, o r when M o Q is dissolved in a molybdate solution. Solutions at pH > 6.5 contain only molybdate anions (Moo:-). Polymeriza tion with the formation of polyanions (Mo4GSi Mo,O& and others) takes 6 . 5 and 2.5. Cations (e. g . , MOO? and place in the pH range between m o r e complex ones) a r e formed at pH below 2.5, and at pH below 1 the cations become the predominant form. The normal alkali molybdates a r e readily luble i n water; the molybdates of the alkaline e a r t h metals, of lead, iron, copper, zinc, and other metals a r e sparingly soluble in water.

63

The properties of some molybdates and polymolybdates a r e reviewed below. N o r m a l s o d i u m m o l y b d a t e NazMo04. This s a l t c r y s t a l i z e s f r o m solutions in which the NazO:M o Q r a t i o is higher than unity. The dihydrate, NazMo0,.2Hz0, is formed between 10 and IOO", while the decahydrate is formed below 10". The melting point of anhydrous Na2Mo04 is 627", i t s density is 3.28. The solubility of sodium molybdate in water at 15.5 and 100" is 39.27 and 45.57'70 respectively. S o d i u m p a r a m o l y b d a t e , 5Naz0.12MoQ.38Hz0, crystallizes from solutions of the normal sodium molybdate when these a r e neutralized to pH 5. The solubility of the crystalline hydrate a t 30" is 1 5 7 g p e r 1 0 0 g water. The anhydrous salt is sparingly soluble in water. z 0 crystallizes A m m o n i u m p a r a m o 1y b d a t e 3 ( NH4)zO 7M o a 4 H out from ammoniacal solutions in which the molecular r a t i o NH, : Moo3 is 6 : 7 or slightly higher. This r a t i o is produced by evaporation of the solution, the ammonia being expelled, o r by neutralizing p a r t of the ammonia. Ammonium paramolybdate is stable in a i r . Its aqueous solutions a r e slightly acid. Its solubility in water a t 20" is about 300 and a t 80 to 90" about 500 g / l . Ammonium paramolybdate begins to decompose a t 150" with evoIution of ammonia and formation of the tetramolybdate (N&)zO. 4MoQ. T h e , ammonia i s expelled quantitatively a t 350" and molybdenum trioxide remains behind. Ammonium paramolybdate is a common commercial product and an intermediate in the production of pure molybdenum trioxide. C a 1c i u m m o 1y b d a t e , CaMoO,, occurs in nature a s the mineral powellite. The s a l t i s white. It i s prepared from aqueous solutions of molybdates by the addition of calcium chloride. Calcium molybdate may be prepared by d i r e c t interaction of calcium oxide and molybdic anhydride a t t e m p e r a t u r e s above 450". The density of the s a l t i s 4.28. Its melting point is 1520". The solubility of calcium molybdate in water a t 20 and 100" is 0.0058 and 0.0235 g/lOOg solution respectively 1 2 5 1 . Calcium molybdate is an important technological product, used a s an additive in steel. It is also used in the smelting of ferromolybdenum. The normal ferrimolybdate, Fe2(Mo0,)3. n H 2 0 I r o n m o 1y b d a t e s s e p a r a t e s out a s a yellow precipitate when f e r r i c chloride o r sulfate is added to a solution of sodium molybdate. Precipitates, whose composition corresponds to the above formula, a r e formed only within a certain pH range (pH =: 3.5). The precipitate formed at higher pH contains iron hydroxide and has a brown color, while the precipitate formed at lower pH contains molybdic acid. When heated above 600, the ferrimolybdate decomposes to yield Fez03 and MoQ. F e r r o u s molybdate, FeMo04, i s not precipitated from molybdate solutions, since Fez+ions reduce (Moo4)'- ions. However, FeMoO, is formed when a mixture of FeO and M o Q is heated at 500 to 600" in the absence of a i r . L e a d m o l y b d a t e , PbMo04, is a white, sparingly soluble salt. It o c c u r s in nature as the mineral wulfenite. The s a l t may be prepared

64

by precipitation from solutions of alkali metal molybdates, o r by heating a mixture of PbO and MOO, at 500 to 600". The density of lead molybdate is 6.92. Its melting point is 1065". C o p p e r m o 1y b d a t e The anhydrous copper molybdate, C u M o 0 4 , is a green-yellow powder which is prepared by heating a mixture of CuO and M o 4 at 500 to 700". The s a l t m e l t s with decomposition a t 820". The basic, bright-green copper molybdates a r e precipitated from aqueous solutions of sodium molybdate by the addition of copper salts. Depending on the conditions of precipitation, the composition of the precipitate either corresponds to the formula C u O . 3 C u M o 0 4 - 5 H 2 0 o r r e s e m b l e s the composition of the mineral lindgrenite, ZCuMo04 C u ( o J 3 ) ~ . Heteropolyacids and their salts. Molybdenum, like tungsten, has a tendency to form complex compounds with phosphoric, arsenic, silicic, and boric acids ( s e e p. 6). Ammonium phosphomolybdate - (NH4)3P04. 1 2 M o Q 6HzO o r (NH4)3H4[P(M~2(&)6 4H20-is . one of the well known s a l t s belonging to this type of compounds. It is a sparingly soluble s a l t which is precipitated when a solution of ammonium molybdate in nitric acid is poured into a phosphate solution containing HN03. This reaction is used extensively for the detection and determination of phosphoric acid. Molybdenum blue. When a solution of molybdic acid o r an acid molybdate solution i s treated with a reducing agent such a s SQ, HzS, Zn, glucose, e t c . , the solution a s s u m e s a deep-blue color which is associated with the formation of the so-called molybdenum blue. Molybdenum blue i s a compound whose composition corresponds approximately to the formula Mo5014 . x H 2 0 ( o r according to other data - M0~023 . x H 2 0 ). The composition i s variable. In solution molybdenum blue exists in a colloidal state and is readily adsorbed by surface-active substances, e. g . , by plant o r animal tissues, to which it i m p a r t s a blue color. The formation of molybdenum blue is a widely used analytical reaction. Molybdenum sulfides. Molybdenum f o r m s four sulfides: MoS,, Mo2S5, MoS2, and Mo2S3. Only MoSz and MoS, a r e of industrial importance. MoSz exists in nature a s the mineral molybdenite, the main source of molybdenum ( s e e p. 6 9 ) . Synthetic molybdenum disulfide i s produced by heating a higher sulfide in the absence of a i r , by the action of sulfur vapor on molybdenum powder o r by fusing molybdenum trioxide with soda and sulfur. The higher sulfide, MOSS, is precipitated when H2S i s passed through hot acidified solutions of molybdates. Molybdenum trisulfide dissolves in ammonium and sodium sulfide solutions, with the formation of thiomolybdates: MoS, (NH,),S = (NH4),MoS4.

The thiomolybdates a r e readily soluble in water. When their solution i s acidified, they decompose with separation.of molybdenum trisulfide:

The precipitation of molybdenum trisulfide is used in analytical chemistry f o r the determination of molybdenum and in industry f o r the extraction of molybdenum from solutions.

65

TABLE 1 3 Some properties of molybdenum chlorides and oxychlorides / 19/ Compound


~ ~~

Color Violet- black

Behavior at various temperatures Melts at 194. boils at 268". In the gaseous phase dissociates with formation of MoCl, (gas) T h e solid decomposes above 130' into MoCll (solid) and MoC15 (gas). Between 330 and 1630" if is the main component of the gaseous phase T h e solid decomposes above 530' into MoCl, (solid) and MoCl, (gas) T h e solid decomposes above 730" into molybdenum and MoC1, (gas) Melts at 170' at a pressure of 1 . 5 a t m . A t 156" the vapor pressure of the solid oxychloride is 1 aim. Melts a t 104, boils a t 180'
. - -.. -

<eat of formation. kcal/ mole

MoC15

126.5

MoCl,

Brown

114.6 94.0
69 0

MoCls MoCl, MOO, C1,

Red-brown Yellow Yellowish-white

MoOCI,

Green

173.0 153.5

Molybdenum chlorides. Molybdenum f o r m s a number of chlorides and oxychlorides. The properties of some of them a r e shown in Table 13. Molybdenum pentachloride is formed when the metal or the disulfide, MoS2, is treated with chlorine above 500". The lower chlorides may be prepared by reduction of M0C15 with hydrogen, or by thermal dissociation of the chloride. MoC15 hydrolyzes in humid a i r and in water with the formation of the oxychlorides, Mo02C12 and MoOC13. Chlorine r e a c t s with M o Q above 500" with formation of the volatile oxychloride, Mo02C12. This oxychloride may also be prepared by heating a mixture of M o Q and NaCl at 500 to 600" / l l / .

Uses of molybdenum F e r r o u s metallurgy. More than 7 5 % of the total production of molyb denum is used in f e r r o u s metallurgy f o r the production of alloy s t e e l s . Molybdenum forms a solid solution in steel. The solubility of molybdenum in iron is 8%. A fraction of the molybdenum exists a s complex iron-molybdenum carbide. Usually, molybdenum is introduced into the s t e e l together with other alloying additives such a s chromium, nickel, and vanadium; the molybdenum concentration in s t r u c t u r a l s t e e l s does not exceed 0.57'0, while in high-speed s t e e l in which i t replaces tungsten i t s concentration reaches 7.5 to 8.5%. Molybdenum noticeably improves the properties of steel by producing a uniform and fine-grained structure. By reducing the eutectoid decomposition temperature of steel, molybdenum widens the temperature range of hardening and tempering and affects the hardenability of steel. Molybdenum improves the following mechanical properties of steel: the elastic limit, the wear resistance, and the impact strength. Alloying

66

o f molybdenum with chromium-nickel s t e e l s eliminates tempering brittleness. This is one of the most valuable properties of molybdenum. Molybdenum is usually added to s t e e l s in the form of ferromolybdenum (50 to 7 0 % Mo). Calcium molybdate, which is cheaper than f e r r o molybdenum, is the form used when it is only n e c e s s a r y to introduce s m a l l amounts of molybdenum. During smelting the calcium molybdate is reduced by the iron. The molybdenum formed dissolves i n the s t e e l while the calcium oxide remains in the slag. Molybdenum is a l s o used in alloys of c a s t iron. Molybdenum reduces the grain s i z e of g r a y i r o n and improves i t s high-temperature properties and wear resistance. C a s t i r o n containing silicon and molybdenum is used f o r the production of acid-resistant equipment. Heat- and acid-resistant alloys. Molybdenum is a component of many acid-resistant and heat-resistant alloys, the other components being mainly nickel, cobalt, and chromium. The m a j o r components of heat-resistant alloys a r e nickel and cobalt, Most of the heattheir concentration being a s high a s 50 to 6 0 % l r r e s i s t a n t alloys, which a r e also acid-resistant, contain 20 to 30% C and 1 t o 7 % Mo. Alloys having maximum resistance to acids (i. e . , which r e s i s t the effect of all mineral acids except hydrofluoric) contain up to 15 to 2 0 % molybdenum. The other components a r e nickel, cobalt, chromium, and iron. Molybdenum and molybdenum-base alloys. The high melting point, high-temperature strength, and high electrical conductivity of molybdenum a r e the reasons f o r i t s extensive use in the electrical bulb and radio industries. Molybdenum wire i s used in the manufacture of filament supports in electrical bulbs and g r i d s for electronic tubes. Molybdenum rods sealed into special g l a s s a r e used a s c u r r e n t terminals in electrical vacuum equipment. thick) which Molybdenum is readily rolled into thin foils (0.1 to 0.2" a r e used in the manufacture of anodes for oscillator tubes and kenotrons. Molybdenum foil is also used in the manufacture of X-ray tubes. Molybdenum wire and bands a r e used a s h e a t e r s for high-temperature e l e c t r i c furnaces. Molybdenum and tungsten w i r e s a r e joined in thermocouples used for measuring temperatures from 1 2 0 0 to 2000" in i n e r t or reducing atmospheres, The recently achieved production of l a r g e molybdenum ingots, weighing lOOOkg or m o r e , has extended the possible u s e s of molybdenum. Molyb denum is now used either pure or a s an alloy to ensure the maintenance of high-temperature strength, e. g . , in the manufacture of turbine blades and other p a r t s of jet engines and rockets. In o r d e r to prevent high- temperature oxidation of molybdenum p a r t s , they a r e coated with molybdenum silicide, a nickel-chromium alloy, o r some other protective coating. Molybdenum may be used a s a s t r u c t u r a l m a t e r i a l in nuclear power plants since it combines high- temperature strength with relatively s m a l l thermal-neutron capture c r o s s section. Large molybdenum rods (1m long and 30 to 40" in diameter) have been lately used a s heating elements in glass-melting furnaces. Molyb denum is virtually unaffected by molten glass. Molybdenum is also used in the manufacture of s t i r r e r s and other p a r t s f o r glass-producing equipment.

67

1,

I ,

, I

..

. , I

111

. I

I .

. . I

1 .

I .

I . ,

I .

I.

I.,

. I

.I*

I , .

I .

Uses of molybdenum compounds. Molybdenum disulfide (either pure molybdenite or synthetic MoS,) may be used as a lubricant. Molybdenite lubricants are superior to graphite and may be used between - 4 5 and -1400". Sodium molybdate is extensively used in the manufacture of pigments and lacquers. Molybdenum compounds a r e used f o r dyeing silk, wool, cotton fabrics, and f u r s ; advantage is taken of the fact that molybdates a r e readily reduced with formation of molybdenum blue. The molybdenum oxides M o Q and M o a a r e used a s catalysts in the chemical industry. The fact that microamounts of molybdenum in the soil stimulate growth in vegetables, especially the leguminous species, h a s been recently discovered. This h a s resulted in an extended u s e of molybdenum compounds (mainly ammonium molybdate) a s fertilizers.

15.

MINERALS, ORES, AND ORE CONCENTRATES Molybdenum minerals

There a r e about twenty known molybdenum minerals, of which only four a r e of industrial value: molybdenite MoS2, powellite CaMoO,, molybdite Fez(Mo04), 79 HzO and wulfenite PbMoO4. Except f o r molybdenite, all these a r e secondary minerals formed in the weathering of the p r i m a r y mineral - molybdenite. Dispersed molybdenum is present a s an impurity in limonite and in some sulfide minerals, mainly chalcopyrite. M o l y b d e n i t e MoSz is the most common and industrially most important molybdenum mineral. Molybdenite is a soft mineral with a metallic l u s t e r and a lead-gray color. In appearance it r e s e m b l e s graphite. The density of MoSz is 4.7 to 4.8, its Mohs hardness is 1 to 1.5. The mineral has a hexagonal lattice with a lamellar s t r u c t u r e (Figure 35). The molybdenumlayers lie between two l a y e r s of s u l f u r ions, forming three layer-packages. The s t r u c t u r e of molybdenite s a t i s factorily explains the perfect cleavage of i t s crystals. It is due to the weak bonds between the three-layer S-Mo-Sstrata. P a r t i a l dissociation of molyb denite takes place uponheating to 1300 to 1350"in the absence of air. Molybdenite melts with decomposi tion at 1650 to 1700". It is readily oxidized to Moo, when heated in a i r at 500 to 600". Molybdenite is oxidized by nitric acid and aqua regia. Molybdenum is most often found in quartz veins, and is frequently associated with scheelite, wolframite, cassiterite, pyrite FeS,, chalcopyrite CuFeS,, arsenopyrite FeAsS, bismuthinite Bi2S, and other minerals. Molybdenite often contains dispersed r a r e metal rhenium (an isomorphous impurity, in concentrations of 0.0001 to 0.04%). The upper s t r a t a of molybdenite veins a r e 0 1 2 3 4 5; OM0 oxidized as a result of weathering with the forma ' os tion of molybdenum ochers, molybdite, powellite FIGURE 35. The crystal lattice of molybdenite. and wulfenite.

68

PO w e 11i t e , CaMo04, a product of oxidation of molybdenite, is frequently found a s a secondary m i n e r a l in the form of a thin coating on molybdenite; it is r a r e l y found as a p r i m a r y mineral. The color of powellite ranges f r o m white to gray, i t s density is 4.35 t o 4.52; and its Mohs hardness is 3.5. The m i n e r a l often contains some tungsten since powellite and scheelite a r e isomorphous. M o l y b d i t e Fez( MOO^)^ 7 i H 2 0 is formed in the weathering of molybdenite and is frequently found together with the l a t t e r in molybdenite oxidation zones. Molybdite h a s a variable composition, which is some times described by the general formula nFepOs-yMoOs.zH20.Molybdite may be an important s o u r c e of molybdenum. F o r instance, about 25% of the molybdenum in the upper zones of the l a r g e s t deposits outside the Soviet Union (in Climax, Colorado) is in the form of molybdite. W u 1f e n i t e , PbMo04, is found in the oxidation zones of lead o r e deposits. Depending on the concentration of impurities, the color of the m i n e r a l may be yellow, bright-red, olive-green, o r grayish. Its density is 6.7 to 7.0, i t s Mohs hardness is 2.5 to 3. This m i n e r a l is not of industrial importance at present.

Molybdenum o r e s and o r e deposits Molybdenum is one of the l e s s common elements. Its average concentra tion in the Earth' s c r u s t is about 0.001%. The molybdenum concentration in o r e s is quite small. It is extracted from o r e s containing a few tenths o r even a few hundredths of one percent of molybdenum. T h e r e a r e s e v e r a l types .of molybdenum o r e s : 1. simple quartz-molybdenum o r e s , in which molybdenite is found in quartz veins; 2. quartz-molybdenum-tungsten o r e s , which contain both molybdenite and wolframite; 3 . skarn o r e s , in which molybdenite (often together with scheelite and some sulfides such a s pyrite and chalcopyrite) is found in quartz veins which fill c r a c k s in s k a r n s (siliceous limestones); 4 . copper-molybdenum o r e s in which molybdenite is associated with copper and iron sulfides. Some o r e s have a very low copper content, but in others copper is predominant and the molybdenum content is v e r y s m a l l ( a few hundredths o r a few thousandths of one percent). However, the total r e s e r v e s of molybdenum in these o r e s a r e l a r g e and they a r e an important source of the metal. The l a r g e s t molybdenum o r e deposits outside the USSR a r e found in the western USA, Mexico, Chile, southeast Canada, southern Norway, and the e a s t e r n s t a t e s of Australia. The y e a r l y production of molybdenum in the western countries is 27,000 to 30,000 tons of the metal ( a s concentrates). The Soviet Union has numerous molybdenum o r e deposits which yield enough molybdenum to satisfy the needs of the industry.

69

Concentration of molybdenum o r e s Molybdenum-containing o r e s a r e concentrated almost exclusively by flotation, which r e s u l t s in a fully satisfactory separation of molybdenite f r o m the gangue and the accompanying minerals. Gravitational concentra tion has limited use. Magnetic concentration is occasionally used a s an additional operation for the separation of iron. Molybdenite is an easily floated mineral. The simple quartz-molyb denum o r e s can be readily concentrated. Ores containing a few tenths or a f e w hundredths of one percent of molybdenite yield concentrates containing 85 to 95% MoS2. The degree of extraction is 90%. Usually, the f i r s t stage involves the collective flotation of sulfides. The molybdenite is then floated, using sodium sulfide or sodium or potassium cyanide in an alkaline medium (pH 8 to 8.4) a s a d e p r e s s o r for the copper and iron sulfides. Standard molybdenite concentrate is obtained after s e v e r a l purification stages. Pine oil is used a s the frothing agent, and kerosene, t r a n s f o r m e r oil, etc. a r e used a s the collectors. In the concentration of copper-molybdenum sulfide ores, the f i r s t product is a mixed copper-molybdenum concentrate. Different methods a r e used to separate the molybdenum from the copper: a) flotation of the molybdenite while the copper sulfide is depressed by sodium sulfide or sodium or potassium cyanide; b) flotation of the copper sulfide, while the molybdenite is depressed by starch. A combination of gravitational methods (for the production of wolframite concentrates) and flotation (for the production of molybdenite concentrate) is used to concentrate quartz-tungsten-molybdenum ores. In the concentration of s k a r n scheelite-molybdenite o r e s , the molyb denite is extracted f i r s t by flotation and the scheelite is then separated (by flotation, using fatty acids and water glass) from the tailings of the f i r s t flotation. Powellite accompanies the scheelite in the scheelite concentrate. Concentration of the oxidized molybdenum minerals, powellite and molybdite, involves g r e a t difficulties. Methods have recently been developed f o r the processing of powellite ores, but the problem of concen trating molybdite-containing o r e s has not yet been solved.
TABLE 14 Technological requirements for molybdenite concentrates Molybdenum concentration, minimum,%
~

Brand

I
P
0.01 0.07 0.15

Maximum impurity conrents.70 As


0.07 0.07 0.07

cu
0.5 1.0
2.0

sio,
5.0 7.0 9.0

Sn

KMl KM2 KM3

50
48

41

0.07 0.07 0.01

The concentration of some copper-molybdenum and, especially, oxidized o r e s yields low-grade (with respect to molybdenum) concentrates, which a r e subsequently treated by hydrometallurgical methods to yield a "syn thetic concentrate'! The technologic a1 r e quir e m ents for molybdenite concentrates produced in the USSR a r e listed in Table 14.
1455

70

16. PROCESSING OF STANDARD MOLYBDFNITE CONCENTRATES

Molybdenite concentrdtes a r e the p r i m a r y raw m a t e r i a l for the production of ferromolybdenum and molybdenum compounds of various degrees of purity, molybdenum trioxide, ammonium paramolybdate, sodium molybdate, and calcium molybdate. Irrespective of the ultimate product into which the molybdate con centrate is to be converted, the f i r s t industrial operation to which the concentrate is subjected is roasting which yields a calcine consisting of moIybdenum trioxide contaminated with a number of impurities. The calcines a r e used f o r the smelting of ferromolybdenum o r for the production of pure molybdenum compounds, the most important of which is molybdenum trioxide. It is prepared either by volatilization o r by f the c alcine. hyd rom etallurgic a1 (chemical) processing o Molybdenite concentrates may be processed directly by hydrometallurgical methods, omitting the preliminary roasting. Such methods include: treatment with nitric acid / 2 1 , oxidation of molybdenite in alkaline solution by oxygen under p r e s s u r e 1 2 1 1 , treatment of the concentrate with an alkaline solution of sodium hypochlorite 1 2 2 1 . These p r o c e s s e s a r e under investigation.
17. OXIDATIVE ROASTING OF MOLYBDENITE C ONC ENTRATES

A number of chemical reactions a r e involved in the roasting of molybdenite concentrates. These reactions may be divided into four groups f 181 : 1) oxidation of molybdenite with the formation of molybdenum trioxide; 2) interaction between molybdenum trioxide and molybdenite; 3 ) oxidation of the sulfide minerals of the accompanying elements (copper, iron, etc. ) with formation of molybdates. 4) reaction between molybdenum trioxide and the oxygenated compounds of the admixtures (oxides, sulfates, carbonates) with the formation of molybdates. The conditions under which these reactions take place will be discussed below. Oxidation of molybdenite. The m i n e r a l molybdenite is rapidly oxidized by atmospheric oxygen above 500", yielding molybdenum trioxide in an exothermic reaction:
MoS,

+ 31/,0,

MOO,

+ 2S0, + 266 kcal

This reaction is virtually i r r e v e r s i b l e and occurs even a t very low concentrations of oxygen in the gaseous phase. During the oxidation the molybdenite p a r t i c l e s a r e coated by a film of the trioxide formed. Hence, the reaction r a t e is determined by the s t r u c t u r e of the oxide film since the oxygen and the SOz must diffuse through the film in opposite directions ( s e e Figure 3 6 ) . It has been experimentally shown that a dense oxide film is formed at 400, and the r a t e of oxidation is determined by the r a t e of diffusion of the g a s e s through the solid film. The oxide film formed at 550 to 600" is porous (friable) and does not interfere with the course of the reaction. The r a t e of oxidation of the mineral a t 600" is about 0.009 m m / m i n / 101.
71

The high exothermic effect accompanying the oxidation of molybdenite p e r m i t s the roasting of molybdenite concentrates to be accomplished at the expense of the heat of reaction. It is well known that sulfides undergo spontaneous oxidation at tempera t u r e s above their flash point. The data in Table 1 5 show that molybdenite has a relatively low flash point, close to the flash points of pyrite and chalcopyrite.

FIGURE 36. Oxidation of molybdenite particles.

TABLE 1 5 Flash point and heat of the oxidation reaction of some sulfides Reaction MoS,
2Cu2S , Heat of reaction, kcal/mole S

Flash point for a particle size <0.063mm."C 365-465* 465** 665 615 360

MOO,

266.3
253.8

-+ 4CuO

ZNiS + 2 N i 0 2ZnS
2Zn0

217.8
212.6

FeS, + 1/3 Fes04

189.2

* **

T h e first figure is for a particle size


0.127 m m

<

0.063 m m . the second for particle sizes between 0.09 and

For particle sizes of 0.09 to 0.127mm.

Interaction between M o Q and MoS2. In the absence of a i r (e. g . , inside the particles formed a s a result of overheating during roasting), the formation of molybdenum dioxide takes place (in the calcine) at 6 0 0 to 700, through the reaction: MoS,

+ 6Mo0,

-f

7Mo0,

+ 2S02.

As is evident from Figure 37, 4 5 % of the MoS2 in a mixture with Moo3 reacts within 6 0 minutes (in an i n e r t atmosphere at 600"), while at 700" about 90% of its r e a c t s in the s a m e time / l l / . Since M o a is virtually insoluble in ammonia water, roasting must be carried out at temperatures not exceeding 600" in o r d e r to prevent sintering and interaction between MoSz and M o Q .

72

Oxidation of the sulfides of other metals. Roasting of molybdenite concentrates, c a r r i e d out at 550 to 600, causes the oxidation of iron, copper, and zinc sulfides which yield oxides (and partly sulfates) by the following reactions:
MeS

+ 0, z2s0,; MeO + SOs2 M e w , .


2s0,

+ ll/zO,Me0 + SOz;
+

It must be remembered that iron sulfates dissociate to a considerable extent above 450 to 500, while copper sulfates dissociate above 6 0 0 to 650" and zinc sulfate reacted above 700".

T i m e, minutes
FIGURE 31. Rate of interaction of MoS, with MOO, a t various temperatures. 1-700';
2-600";

/U MOO3

20 30 40 CUO, w t %

FIGURE 38. Phase diagram M o Q -CuO.

3-500".

In addition to copper, iron, and zinc sulfates, calcium sulfate may also be formed during roasting if the concentrate contains calcium carbonate:
Cam,

+ SO, + CaSO., + CO,.

Interaction of M o Q with oxide, carbonate and sulfate impurities. At


500 to 600" MOO, r e a c t s with the oxides, carbonates and sulfates of a number of elements, yielding molybdates / 11/ :

CaCO, MOO, + CaMOO,, CO,;


CuO + MOO, + CuMoO,;

CUSO,

+ MOO3
ZnO PbO

+ MOO,
+

CuMoO, so3(SO,, 02); + ZnMOO,;


MOO, + PbMOO,.

The formation of FeMoO, may also take place in the absence of a i r . Upon heating in air, however, FeMoO, oxidizes and decomposes into F e z Q and Mag. Copper molybdate and M o Q f o r m a low-melting eutectic (560, s e e Figure 38), which explains the lower sintering temperature of calcines

73

with an increased copper content 1181. Of the above molybdates, those of calcium and lead are sparingly soluble in ammonia water. T h e i r presence in the calcines c a u s e s a s h a r p d e c r e a s e in the degree of extraction of molybdenum. Copper and zinc molybdates are readily soluble in ammonia solutions, while iron molybdate is slowly decomposed by ammonia water. Molybdenite concentrates always contain silica. However, t h e r e is virtually no interaction between s i l i c a and M o q .

Roasting procedure Until recently, the roasting of molybdenite concentrates was c a r r i e d out in flame or muffle furnaces with manual raking of the m a t e r i a l , in rotating tube furnaces and in multiple-hearth furnaces with mechanical raking. The f i r s t two types of furnaces have s e r i o u s drawbacks. They do not utilize the heat of the oxidation reaction; the roasting is c a r r i e d out with a continuous supply of heat since there is no full countercurrent motion of the g a s e s and the m a t e r i a l being roasted. Moreover, temperature control in such furnaces is difficult. Overheating r e s u l t s in the sintering of the material, while the formation of lower molybdenum oxides and molybdates within the sintered particles reduces the degree of extraction of molybdenum from the calcine into the solution. The multiple-hearth mechanical furnaces, in which the gas counterc u r r e n t penetrates a l l the m a t e r i a l a s i t p a s s e s from one hearth to the other, a r e of m o r e advanced design. As a result, there is b e t t e r utiliza tion of the heat and the roasting takes place mainly at the expense of the heat of reaction. Fluidized-bed furnaces a r e used at present f o r the roasting of molybdenite concentrates in s o m e plants in the USSR.

Roasting in multiple-hearth furnaces Such furnaces have been used for a long time for the roasting of pyrites and of copper and zinc sulfide concentrates. These furnaces ensure good mixing, countercurrent flow of solid and gas, and rapid oxidation of the suspended m a t e r i a l passing f r o m one hearth to another. A vertical c r o s s section through an eight-hearth furnace, of which seven hearths a r e for roasting and the eighth, upper hearth is for drying, is shown in Figure 39. The furnace consists of a vertical cylinder made of s t e e l sheets and lined with r e f r a c t o r y (chamotte) bricks. The inner diameter of the furnace is 5.4m. The cylinder is divided into seven s t o r i e s by refractory brick crowns, which also s e r v e a s the eight hearths The distance between the c e n t e r s of the hearths is 0.894m. A rotating s t e e l shaft, 1.29 to 1.52m in diameter and lined on the outside with refractory bricks p a s s e s through the c e n t e r of the furnace. The center shaft c a r r i e s two r a k e s a t each hearth, which a r e fastened to special

74

sockets within the shaft. The rake is fitted with paddles, made of chromium-containing cast iron or a special alloy, to move the material. The center shaft and the r a k e s a r e air-cooled on the inside.

Sulfide concentrate

FIGURE 39.

A mechanical multiple-hearth furnace.

1-crowns; 2-central shaft; 3-raker; 4-paddles; 5-openings in the crowns, for the passage of material; 6-storage bin for the concentrate; I-feeder; 8-feed tray.

At the periphery or a t the center of each hearth there a r e openings for the passage of solids and gases. The openings in the f i r s t , third, fifth and seventh hearths a r e made in the periphery, while in the remaining hearths they a r e made in the shaft. Such an arrangement

75

e n s u r e s that the m a t e r i a l p a s s e s through the entire surface of the hearths. Since the paddles over each hearth a r e placed at opposite angles with respect to the shaft, the materia1 moves f i r s t from the periphery to the center and then from the center to the periphery and thus drops from hearth to hearth.

FIGURE 40. Temperature distribution on the hearths of a multiple-hearth mechanical furnace for the roasting of molybdenite. The air is introduced at each hearth and gases are led out through a common gas vent.

In o r d e r to ensure quantitative combustion of the sulfur, the lower hearths a r e fitted with burners. In addition, some of the hearths (e. g. , the even-numbered ones) a r e fitted with b u r n e r s to preheat the furnace. These b u r n e r s a r e extinguished when the furnace becomes sufficiently hot. The hearth temperature must not exceed 600". Temperatures a s high a s 750 to 800" must be avoided - Moo3 melts at such temperatures, with consequent clogging of the openings and a rapid wear of the paddles even though i t is difficult to maintain the temperature at 580 to 600" in furnaces of this type. The output of the furnace is 6 0 to 80 kg of concentrate p e r m 2 of hearth surface p e r day, depending on the operating conditions. The dust entrainment with the exhaust g a s e s reaches 8 to 1 0 % . The SO2 content in the g a s e s is 3 to 5%. In the Climax plant in the USA roasting is c a r r i e d out in 8, 12, and 16-hearth furnaces. The temperature is controlled by a s e p a r a t e supply of a i r to each hearth. The g a s e s p a s s out from s e p a r a t e outlets at each

16

.....

.. .-.

..

hearth to a common g a s vent. This p e r m i t s the maintenance of the required temperature a t each hearth. The resulting temperature distribution in the hearths is shown in Figure 40. The dust entrainment is a s high a s 1870, the dust being mostly unroasted concentrate. The dust is trapped and returned for roasting in the furnace.

Fluidized-bed roasting During the l a s t 1 0 t o 1 2 y e a r s the roasting of sulfide concentrates in chemical and metallurgical industries has also been c a r r i e d out by p r o c e s s e s in which the p a r t i c l e s of the m a t e r i a l a r e supported by an upward flow of gas in a "boiling" or pseudo-fluid state. The g r a n u l a r m a t e r i a l reaches such s t a t e at certain gas velocities. The particle bed remains stationary a t low gas velocities (below a certain c r i t i c a l value) ( s e e Figure 41). As the c r i t i c a l velocity ([hi,,)i s reached, the bed begins to expand and the m a t e r i a l then p a s s e s into the "fluidized" state, which is characterized by a rapid motion of the particles in the gas s t r e a m , and the appearance of the bed resembles that of a boiling liquid. As the gas velocity is increased further to a new critical value (znmaX), the whole bed becomes fluidized and is entrained with the gas s t r e a m . The advantages of fluidized-bed roasting a r e associated with the following properties of such beds. 1. Good contact is ensured between the particles and the gas, so that chemical reactions in a fluidized bed take place at a high r a t e , 2. The mobility of the l a y e r (which resembles the mobility of a liquid) p e r m i t s an easy, continuous gravity discharge ("efflux") of the m a t e r i a l from the furnace through the discharge tube. 3. Fluidized beds a r e characterized by a high thermal conductivity and high heat-transfer coefficients. This p e r m i t s the required temperature to be maintained throughout the volume of the bed, even when the heat of reaction i s high. The removal of excess heat from the bed is easily accomplished with the aid of coolers (for instance, water-cooled tubes fitted inside the bed). O D
The chemical nature of the roasting of 0 0 o o molybdenite concentrates discussed above shows that in o r d e r to produce calcines with a high content of extractable a b C molybdenum it is necessary that the roasting be carried out at closely FIGURE 41. T he three states of granulated led temperatures in o r d e r to avoid material in a n upward stream o f gas. sintering. The contact between the a -stationary (filtering) bed; b-fluidized particles must also be kept t o a minimum bed; c-suspended state. in o r d e r to eliminate, a s f a r as possible, formation of molybdates. T h e s e conditions a r e best satisfied during fluidized-bed roasting.

77

The fluidized-bed roasting of molybdenite concentrates was f i r s t studied in the USSR. It is now used on an industrial s c a l e 1121. The design of one industrial furnace and of the whole installation a r e shown in Figure 42. The furnace consists of a refractory chamber with a rectangular c r o s s section, whose lower p a r t is fitted with an a i r distributing screen. The s c r e e n consists of a s e r i e s of s t e e l nozzles with mushroom-shaped removable caps ( F i g u r e 4 3 ) . The space between the nozzles is filled with heat-resistant concrete. The a i r inlet holes a r e beneath the mushroom-shaped caps. This prevents leakage of the solid m a t e r i a l below the screen.

FIGURE 42.

Fluidized-bed inscallation for roasting molybdenite concentrates

1 -air blower; 2-burner; 3-pressurized fire-box; 4-water-cooling system for the removal of excess hear; 5-fuel oil storage; 6-tray-type feeder; I-bin: 8-gate valve; 9-furnace chamber; 10-cyclone: 11-screen (grate): 12-discharge baffle; 13-snioke exhaust; 14-electrostatic filter.

Uniform feeding r a t e of concentrate to the furnace i s most important in maintaining the required roasting conditions. Nonuniform feeding of the concentrate causes s h a r p changes in the temperature of the bed, since the oxidation reaction is accompanied by the evolution of large amounts of heat. The feed of the concentrate to the furnace is controlled by an automatic feeding device which consists of a cylindrical bin (the con centrate "sticks" to a conical bin) and a tray-type feeder with an adjustable rotational velocity beneath the bin. The feeder blade moves the concentrate into the bin of a hermetically closed gate from which i t flows by gravity through a spout into the fluidized bed. At a height of 1 0 0 0 to 1500" above the furnace hearth t h e r e is a discharge hole (discharge baffle) from which the calcine is discharged continuously and fed into the collectors through a gate valve. The gases, together with the entrained fines, a r e let out through gas vents at the top of the furnace chamber. The g a s e s p a s s through dust t r a p s (cyclones, electrostatic filters) and a r e discharged into the atmosphere.

78

The bed temperature is regulated automatically. The temperature control s y s t e m is based on variations in the concentrate supply rate. When the temperature becomes too high o r too low, a s compared with the required temperature, there is an automatic d e c r e a s e or i n c r e a s e of the r a t e of feed of the concentrate. This is accomplished by varying the r a t e of rotation of the feeder trays. P r a c t i c e has shown that the system p e r m i t s a reliable temperature control to within f 2 . 5 O of the average p r e s e t temperature If the average r a t e of feeding of the concentrate is such that too much heat is evolved, the excess heat is removed by water flowing through tubes I- \ 1 accommodated inside the fluidized bed. FIGURE 43. Design of a n o zzl e in t h e ai r In o r d e r to s t a r t the operation of the distributing screen. furnace, calcines heated to the flash l - p i p e : 2 - r e m o v a b k c a p with holes: 3 point of the molybdenite concentrate heat-resistant concrete. (500 to 510") a r e fluidized by means of hot a i r , and the concentrate is then fed in (at a r a t e of about 50 to 6 0 k g / h r . m 2 of furnace hearth). As the concentrate f a l l s onto the bed, it is ignited and the temperature i n c r e a s e s . After 15 to 30 minutes i t reaches the optimum roasting temperature of 560 to 570" which is then maintained by the automatic control system. The height of the bed i n c r e a s e s as m o r e concentrate is fed in; continuous discharge of the calcines s t a r t s when the height reaches the discharge baffle. A fraction of the concentrate fines ( 2 0 to 4070, depending on the granulometric composition of the concentrate) is entrained f r o m the bed with the gases. Quantitative trapping of the dust is effected by a system of cyclones and an electrostatic f i l t e r ( 9 0 to 957'0 of the dust is trapped in the cyclones). Under the roasting conditions described above, the dust is not quantitatively oxidized (the degree of oxid.ation is 70 to 85%) and it contains 8 to 1 0 % sulfur. The dust may be returned to roasting after preliminary granulation (particle consolidation) in a ladle-type granulator. Both furnace construction and roasting conditions a r e still being improved, but the experience already gained with the f i r s t industrial furnace models has shown that the p r o c e s s has some definite advantages: a) the furnace output is 1200 to 1300 kg/m2 of furnace hearth, which is 15 to 20 times higher than that of conventional roasting furnaces; b) the p r o c e s s is completely automatic (the roasting takes place by the heat of reaction a t closely controlled temperature) ; c) the quality of the calcine produced (the content of extractable molybdenum) is higher than that of calcines produced in hearth furnaces; this is attributed t o the suppression of reactions leading to the formation of molybdates and M o Q .

79

18.

PRODUCTION OF P U R E MOLYBDENUM TRIOXIDE Distillation method

The vapor p r e s s u r e of M o Q a s a function of the temperature is shown in Table 16. Molybdic anhydride begins to volatilize below its melting point. However, a noticeable i n c r e a s e in the vapor p r e s s u r e o c c u r s only above the melting point (795"). Volatilization at an industrially useful rate takes place only at 900 to 1100O. In addition t o M O ~ the , calcines may contain impurities such a s molybdates and oxides of contaminant metals. Iron, copper, and silicon oxides a r e not volatilized a t 900 to 1100". Calcium molybdate is thermally stable up to 1200 to 1300" and is not volatilized. -:&*.. . Copper molybdate, CuMo O,, decompose s a t 820" into CuO and MoQ. Lead molybdate, PbMoO,, volatilizes 4 noticeably at i t s melting point (1050"). Hence, if lead molybdate is present the distillation must be c a r r i e d out below 1000" in o r d e r to prevent contamination FIGURE 44. Distillation Of molybdenum miof the condensate with lead. There is oxide. practically no interaction between MoQ 1-quartz crucible; '2-jacket; 3-electric and quartz, and the volati,lization of heater; 4-compressed air supply; 5-exhaust ~~gis not affected by the presence of hood: 6-fan; 'I-bag filter. quartz. A diagram of one of the possible ways of c a r r y i n g out the distillation is shown in Figure 44. The roasted molybdenite concentrate is placed in the quartz crucible of a rotating electrical furnace. The furnace has a slope of 35". The sloping position of the furnace r e s u l t s in an i n c r e a s e in the surface of volatilization of the MoQ. The crucible is heated to 900 to 1000" with the aid of an electric coil heater. At that temperature the calcines melt. Air is fed continuously to the crucible, and the a i r s t r e a m entrains the M o Q vapor. The trioxide vapor p a s s e s from the furnace into a bag f i l t e r through a hood, fitted over the furnace, and a fan system. Distillation may a l s o be c a r r i e d out in a continuous-action furnace with a rotating hearth 1241. Of the total amount of molybdenum in the initial calcine, 6 0 to 70% is thus volatilized. The remaining fraction is extracted from the residue by hydrometallurgical methods. The trioxide produced by the distillation p r o c e s s contains 99.95% M o Q . It is in a highly dispersed s t a t e (its bulk weight is 0.5g/cm3). Such "light" and bulky powders cannot be conveniently transported. Before packing, the molybdenum trioxide is moistened with a s m a l l amount of distilled water and is compressed and dried. This r e s u l t s in an i n c r e a s e in the bulk density to 3.5 g/cm3.

80

TABLE 16 Vapor pressure of MOO, Temperature. " C


600 610 625 650 720 750 800

Pressure m m Hg
0.000 0.009 0.018 0.05 0.60 1.75

Temperature, 'C
850 900 950 1000 1050 1100 1155

Pressure m m Hg
23.4 53.9 105.1 179.8 288.3 476.2 760.0

10.1

Chemical methods /2,3/ The ammoniacal method f o r the processing of calcines is diagram matically shown in Figure 45. This is the most extensively used method. The calcines a r e leached with a solution of ammonia. The ammoniacal solution is treated to remove impurities and the molybdenum is isolated a s ammonium paramolybdate or other polymolybdates ( s e e below). Molybdenum trioxide is prepared by thermal decomposition of ammonium molybdate.

Leaching
On leaching with a solution of ammonia, molybdenum trioxide p a s s e s from the calcines into solution:
Ma,

+ 2NH40H = (NH,),MoO, + H20

The degree of extraction of the molybdenum depends on the composition of the calcine. In addition t o M o Q the calcine may contain calcium, copper, zinc, and f e r r o u s molybdate, Mooz, non-oxidized molybdenite, copper and calcium sulfates, iron oxide, silica, alkali metal s a l t s , tungsten minerals and other impurities. Calcium molybdate, M o a , and molybdenite a r e virtually insoluble in ammonia solutions. The molybdenum in these compounds remains behind in the leaching tailings. The copper, zinc, and nickel molybdates and sulfates a r e readily soluble in ammonia solutions, yielding ammoniates 12 6 1 .
MeMoO,
MeSO,

+ 6NH,OH + [ M e (NH,),] (OH), + (NH,),MoO, + 4H,O, + 6NH40H [ M e (NH,),] (OH), + (NH,),SO, + 4H,O.
+

Ferromolybdate is decomposed by ammonia solutions, but the reaction is slow because of the formation of the virtually insoluble f e r r i c and f e r r o u s hydroxides, which coat the molybdate particles 1 2 6 1 . A fraction of the f e r r o u s iron is dissolved by the ammonia solution in the f o r m of an ammoniate complex:
Fe (OH),

+ 6NH,OH

IFe (NH,),I (OH),

+ 6H20.

81

concentrate

r
Second extraction of molybdenum

Leaching

I,

NH,OH solution

Residues

Solution

Purification <

("3,S
Sulfide c a k e

Purified solution

/-----1
To waste
Method I1 Evaporation

Method I HC1 Acidic moiher liquor Molybdenum regeneration


+

Neutralization

A
Ammonium paramolybdate crystals

4-

.1

Polymolybdate precipitate Recr~srallization

-I

I
liquor Periodically drain off in order to remove impurities

P
Crystallization

.1

Mother 1iqu or

3.

1
Calcination

FIGURE 45.

Production of pure molybdenum trioxide from molybdenum calcines.

Depending on the composition of the calcines, the degree of extraction of molybdenum into the ammoniacal solution ranges from 80 to 9570. The amount of residue [tailings] (i. e . , the weight of the residues with respect to the weight of the initial calcines) ranges from 1 0 t o 30%, and the molybdenum concentration in the residue ranges from 5 t o 2570. As a rule, a second extraction of molybdenum from the residue is required. In o r d e r to reduce the degree of extraction of impurities into ammonia c a l solution, the calcines a r e occasionally washed with water before processing with ammonia solution. The water dissolves copper sulfate a n d alkali metal salts. However, it must be taken into account that some molybdenum is lost because of the higher solubility of M o Q in water in the presence of impurities / 2 6 / . The M o Q concentration in the aqueous solution is 3 to 5 g/l, and may occasionally be higher. The l o s s e s of molybdenum with the wash waters reach 4 to 5%.

82

. ..

-. - .

The calcines are leached with an 8 to 10% solution of ammonium hydroxide in the cold o r at 70"; the leaching is c a r r i e d out in i r o n vessels. Rotating-drum e x t r a c t o r s o r vats fitted with mechanical s t i r r e r s a r e used. The d r u m e x t r a c t o r s have the advantages of being hermetically sealed and of ensuring good mixing as well a s the possibility, when filled with balls, of effecting additional grinding. The s t i r r e r - f i t t e d e x t r a c t o r s a r e not hermetically sealed. As a result, a p a r t of the ammonia evaporates and the consumption of ammonia is increased. Depending on the type of apparatus used, the consumption of ammonia is 115 to 140% of the stoichiometric amount. T h r e e or four successive processing stages are used in o r d e r t o provide a higher degree of leaching of molybdenum from the calcines. The solutions f r o m the first two processing stages a r e collected together and are treated to remove impurities. The weak solutions from sub sequent leachings a r e returned to the initial leaching stage. The s l u r r i e s from each leaching stage a r e filtered on f i l t e r p r e s s e s or suction filters. The residues a r e rinsed with hot water and the wash waters a r e recycled t o the f i r s t leaching stage. The residue is subjected to a second molybdenum leaching stage (see p. 87). The strong ammonia solutions from the leaching contain 140 to 19Og M o Q / I . Their density i s 1.10 to 1 . 1 6 .

Removal of copper and iron f r o m the solutions The ammoniacal solutions a r e contaminated with copper, iron, occasionally zinc and nickel, alkali metal and sulfate ions. Copper and iron a r e removed a s sulfides by the addition of ammonium sulfide. Copper and iron sulfides a r e practically insoluble, the solubility of CuS being 9.1 X l o m z 3 moles/l. In spite of the fact that copper is bound a s a stable complex, the solution always contains some copper ions whose amount corresponds to the equilibrium dissociation reaction:
[Cu (NH3),l2+Z Cu2+

+ 4NH3.

As the Cu2+ions r e a c t t o form the sulfide, the equilibrium continuously shifts until quantitative precipitation of the copper takes place:

[Cu (NH,),] (OH),

+ NH,HS + 3H,O

CuS

+ 5NH,OH.

Iron sulfide is precipitated in the s a m e way.


The amount of ammonium sulfide added must be closely controlled since an excess of it in solution leads t o the formation of thiomolybdates which contaminate the ultimate product:
(NH,),MoO,

+ 4NH,HS Z (NH,),MoS, + 4NH,OH.

Ammonium sulfide solution is poured into the ammoniacal solution in s m a l l portions. An aliquot is taken after each addition and the degree of precipitation determined. The solution is then tested f o r sulfide ions by the addition of lead nitrate to a s e p a r a t e aliquot of the solution. A black precipitate (lead sulfide) is formed in the presence of an excess of

83

ammonium sulfide, while a white precipitate (lead molybdate) f o r m s in the absence of such an excess. The e x c e s s ammonium sulfide is bound by the addition of a certain amount of f r e s h ammoniacal solution from the calcine-leaching operation. The sulfides a r e precipitated in wood o r rubber-lined i r o n vats with stirring.

Isolation of molybdenum f r o m the ammoniacal solutions Separation of ammonium paramolybdate by evaporation. The normal ammonium molybdate (NH4)2Mo04is stable only in solutions containing an excess of ammonia. A p a r t of the ammonia is expelled during evaporation of the solution and ammonium paramolybdate is formed:

7 (NHJzM004

+ 4HxO

-+

3 (NHJe0.7M00,

+ BNH40H.

In o r d e r 00 prevent the formation of m o r e acidic molybdates (of lower ammonia content), it is necessary that an excess of f r e e ammonia (4 to 6 g / l ) be maintained during the evaporation, and that the solution be s t i r r e d in o r d e r to prevent local overheating. Molybdates m o r e acidic than the paramolybdate a r e precipitated a s finely crystalline, sparingly soluble precipitates. Solutions from which the heavy metals have been removed a r e taken for evaporation. These solutions contain 120 to 140 g M o Q / l . Their density is 1.09 to 1.12. The solutions a r e evaporated in stainless-steel kettles fitted with jackets, in a two-stage process. Advance evaporation is c a r r i e d out to a density of 1.20 t o 1.23. The solution is allowed t o stand and is filtered to remove the s m a l l amount of precipitate formed (copper and iron sulfides, as well a s iron hydroxides which a r e p r e cipitated a s a result of the coagulation of finely-dispersed suspended particles). The main evaporation stage is then c a r r i e d out and the density of the solution i n c r e a s e s to 1.38 to 1.40, i. e . , to a M o Q concentration of about 400 g/l. The hot solution is filtered and collected in crystallizers. Crystallization is generally c a r r i e d out in stainless s t e e l c r y s t a l l i z e r s containing s t i r r e r s and a cooling system. A finely-crystalline precipitate consisting of ammonium paramolybdate 3(NH4)20 7 M o Q 4H20 s e p a r a t e s out a s the solution cools. The paramolybdate c r y s t a l s a r e separated from the mother liquor by centrifuging, and a r e washed with cold distilled water directly in the centrifuge. The mother liquor is again evaporated and an additional amount of paramolybdate crystallizes out. Several successive crystallizations a r e c a r r i e d out. About 50 to 60% of the s a l t present in the solution s e p a r a t e s out during the f i r s t crystallization. The paramolybdate from the f i r s t two crystallization stages is of high purity. The residual mother liquor, which contains the concentrated impurities, is evaporated to dryness and the residue is ignited at 350 to 400". The resulting contaminated M o g is returned to the calcine-leaching stage. This method of production of the paramolybdate has s e v e r a l dis advantages : 1) it is time-consuming, since it requires s e v e r a l successive evaporations and crystallizations; 2) the c r y s t a l s from the second and subsequent crystallization stages #@en contain an increased amount of contaminants, above the technologically
84

permissible concentration, so that i t is n e c e s s a r y to c a r r y out additional recrystallizations. For these reasons, the polymolybdate separation method is now extensively used in the Soviet Union. Separation of polymolybdate by neutralization J 141 Polymolybdates of various compositions a r e precipitated during the neutralization of ammonium molybdate solutions by hydrochloric acid ; thzir composition depends on the ultimate pH of the solution and the temperature. In manufacturing practice, the precipitation is c a r r i e d out from solutions containing 280 to 300 g M o Q / l . When necessary, the solutions are preliminarily concentrated by evaporation. Up to 9 6 to 9 7 % of the molybdenum is contained in the precipitate in the form of tetramolybdate dihydrate if the hot (55 to 65") solution is carefully neutralized with hydrochloric acid (to a pH between 2 and 3 ) . The hydrolysis reaction is given by the equation:
4 (NHJ,MoO,

+ 5HtO -+ (NH,,&O.4Mo03.2HzO + 6NH40H.

The crystalline precipitate formed must be rapidly separated by filtration, since the dihydrate is converted into the anhydrous t e t r a molybdate (NH,J2O. 4MoOj on prolonged contact with the mother liquor. The conversion i s accompanied by a d e c r e a s e in c r y s t a l size. A highly dispersed precipitate which i s difficult to f i l t e r is formed a s a result. After centrifuging and washing with water, the polymolybdate precipitates s t i l l contain a s m a l l amount of contaminants. Most of the contaminants (Zn, Cu, Ni, Sb, As, Mg, P, and S ions) remain in the slightly acidic mother liquor. Tungsten, however, i s an exception, a s most of i t is coprecipitated with the polymolybdate. The precipitates also contain an increased amount of chloride ions (0.2 to 0.4%) which a r e difficult to remove by washing with water. Recrystallization of the polymolybdate is c a r r i e d out in o r d e r to remove the chloride ions. The s a l t is dissolved in a 3 to 57' solution of ammonia, at 7 0 to 80, in an amount such that a saturated solution i s formed (density 1.41 to 1.42). The solution is cooled to 15 to 20" and 50 to 6 0 % of the molybdenum is precipitated a s crystalline ammonium paramolybdate 3 ( N ~ ) 2 0 . 7 M O4Hz0. ~. The mother liquor i s used f o r the successive recrystallization of about ten portions of polymolybdate precipitates. Impurities gradually accumulate in the mother liquor and it has to be purified. The combined precipitation of polymolybdates through neutralization and subsequent recrystallization from an ammoniacal solution yields ammonium paramolybdate of a higher degree of purity than that obtained by evaporation and crystallization of the p a r a - s a l t . The ammonium paramolybdate prepared by the method contains the following impurities (in% of Mo) :
Sn, Pb, Bi, Cd . . . . . . . <0.0001 Zn, Mg, As, P. S. Ni. Cr. Ca <0.001 Si. A1 . . . . . . . . . . . <0.003 Fe 10.005

.............

The isolation of the paramolybdate is c a r r i e d out in c a s t - i r o n r e a c t o r s coated with acid-resistant enamel and fitted with enamelled anchor-shaped

85

s t i r r e r s . The r e a c t o r s a r e fitted with steam jackets. The addition of acid to the required pH between 2 and 3 is controlled with the aid of a g l a s s electrode pH-meter. The pH is approximately determined by the appearance of a violet color on Congo Red indicator paper. The acidic mother solution which remains after separation of the c r y s t a l s s t i l l contains 3 to 4% of the initial molybdenum content. This corresponds to a Mo concentration of 6 to 10 g/l. The solution is allowed to stand f o r a prolonged time, the pH is adjusted to 2, and amorphous polymolybdate precipitates of variable composition a r e separated and returned to the purification stage. The residual mother liquors contain about 1g Moil. Molybdenum may be isolated from them by sorption on ion-exchange r e s i n s (see p. 92).

Extraction of molybdenum from the calcine-leaching residues Depending on the composition of the concentrate and the method of roasting, the leaching residues contain from 5 to 25% molybdenum. The molybdenum is in the form of calcium and iron molybdates, M o Q and non-oxidized molybdenite. In addition, p a r t of the molybdenum is in the f o r m of MOO:- ions sorbed on the iron hydroxide and difficult to wash out with water. The remaining components of the residues a r e : iron oxides and hydroxides, silica, and occasionally tungsten (1 to 5%) a s scheelite or wolframite. The following methods a r e used for the extraction of molybdenum from the residues: fusion with soda ash, direct leaching with soda a s h solutions in autoclaves and digestion with acids. Fusion with soda ash. The moist residues a r e mixed with soda a s h and the paste-like mixture formed is heated in a furnace f o r 6 to 8 hours at 700 to 750". During the fusion all the molybdenum compounds react with the soda ash and a r e converted into sodium molybdate:

The cake is leached with hot water. Iron molybdate is precipitated f r o m the solution which contains sodium molybdate. The iron molybdate is decomposed by aqueous ammonia and the ammonium molybdate solution is recycled. The iron molybdate is precipitated by the addition of a solution of iron chloride at pH = 3.4 to 5. The precipitates contain a variable amount of Fe& and M o Q , and usually do not correspond to the composition of the molybdate Fe2(Mo0&. The mixture is filtered in a filter p r e s s and the moist cake is leached with ammonia water:
Fez(Ma,),

+ 6NH,OH

3 (NH,),Mo04

+ 2Fe (OH),.

. 86

The processing of molybdenum-containing calcines, including the stage of extraction of molybdenum from the residues by the soda-ash fusion method, is diagrammatically shown in Figure 46. The total extraction of molybdenum into the final product, including the residue processing stage, is 93 to 94%. The discarded residues contain 1 to 1.570 molybdenum.
Roasted concentrate (calcines)

I
Ammonia water

F
Leaching

3
Cake

'

Solution

1
Drying Soda ash
P

1
Cake

Mixing
.L

I
I

m
Solution

Removal of impurities

1
To separation
of ammonium
paramolybdate

Sinrering

+
Residues

Leaching

.L Solution

To waste

Precipitation of iron molybdate Solution


. I To trap

*
NHdOH

Precipitate Leaching

I I
HC1

Precipitate (Fe(OH), Dissolution

J.

. I

Solution

\ J

. 1

FeC1, solution

FIGURE 46. Processing of molybdenum calcines including sintering of leaching residues with soda ash.

Leaching with soda ash solutions in autoclaves. Calcines in which molybdenum occurs as CaMoO, and other molybdates (and which have a very low MOO, and MoS, content) may be leached directly with soda-ash solutions instead o f being fused with soda; the leaching is carried out in autoclaves at 180 to 200' which corresponds t o a pressure of about 12 to 15 a t m .

Digestion of the cinders with acids f 27 f Fusion with soda ash i s not a convenient method for processing calcines with an increased tungsten content ( 3 to 5% W) since all the tungsten p a s s e s into solution in the form

87

30% HC1

I
-

Drying and roasting Calcines

NHS solution

Leaching

Residues

Ammonium molybdate solution

Decomposition
Y N Hs :Neutralization

Washing and filtration Residues Roasting Ammonia water


/

; Wash waters
t o waste

1 1

Leaching

Filtration

.1
Residues to waste

. 1
Ammonium molybdate solution (NH&
b

Solution HCl

Sulfide waste

.1

Precipitation Ammonium polymolybdate precipitate Recrystallization

P
solution

1
Centrifuging

FIGURE 41. Processing of calcines, including acid decomposition of the leaching


residues.

88

of sodium tungstate. No methods f o r the separation of tungsten from molybdenum have a s yet been developed. In such a c a s e , the calcines a r e digested with hydrochloric acid (Figure 47). When the calcines a r e digested with hot 20 to 30% acid, the molybdates a r e decomposed quantitatively with formation of molybdic acid which readily dissolves in hydrochloric acid. Iron, calcium, copper, and other impurities a r e also dissolved. The tungsten minerals present in the calcines (scheelite, wolframite) a r e decomposed only partially by the acid. In addition to these minerals, the residue contains silica, M o a and MoS,. Neutralization of the acid pulp with ammonia water to pH 2.5 to 3 r e s u l t s in the precipitation of i r o n molybdate in a mixture with molybdic acid and polymolybdates. Calcium, p a r t of the iron, the copper, and other impurities remain in solution. The precipitate is separated by filtration. It is then ignited at 580 to 600" in o r d e r to oxidize M o Q and MoS, and is le?ched with ammonia water. The degree of extraction of molybdenum from the calcines by this method is 80 to 8 5 % .

Extraction of molybdenum into the final product and costing In the processingof standard molybdenite concentrates, the total extraction

of molybdenum into ammonium paramolybdate depends on the impurity


content in the initial raw material, the technological p r o c e s s used, and the maintenance of optimum technological conditions. The degree of extraction in manufacturing plants ranges from 93 to 95%. About 1 to 1.570 i s lost by roasting, 2 to 2.570 during leaching (including the reprocessing of residues) and 2 to 370a r e lost during the purification of solutions and the separation of ammonium paramolybdate. An analysis of the cost price of ammonium paramolybdate ( s e e below) shows that the main cost item i s the raw material (molybdenite concentrate). Hence, it i s important to i n c r e a s e the degree of extraction of molybdenum f r o m the concentrate. This requires a reduction of mechanical l o s s e s , more efficient trapping of dust, increased extraction of molybdenum from the wash waters and mother liquors and the use of fluidized-bed roasting which yields calcines containing a higher concentration of extractable m olybd enum . The approximate s t r u c t u r e of the cost price of one ton of ammonium paramolybdate (in '70 of the total cost f. 0.b. factory) is shown below:
Raw material: molybdenite concentrate.

Chemicals: ammonia water. hydrochloric a c i d , etc.

...................... ............

91.8 2.0 0.2 1.1 1.12 3.18


100.0

Auxiliary materials: filter cloth, bronze screen. wrapping paper, plywood,etc. .................................... Power consumption: electrical, fuel, steam, water. ............ Wages and benefits ......................................... Overheads ................................................ Cost f.
0.b.

factory..

...

89

19.

PROCESSING OF LOW-GRADE CONCENTRATES

The beneficiation of polymetallic molybdenum o r e s (e. g., coppermolybdenum ores) yields low-grade concentrates and intermediate products which contain, in addition to molybdenum, l a r g e amounts of iron, copper, and other impurities. In such materials molybdenum is present both as molybdenite and as the oxidized minerals, powellite and molybdite. The molybdenum concentration in the low-grade concentrates ranges from 5 to 20% i n the presence of l a r g e amounts of copper, iron and other impurities. For instance, the low-grade molybdenite concentrates obtained in the concentration of copper-molybdenum o r e s from a certain deposit had the following approximate composition: 1 5 to 20% Mo, 3 to 5 % Cu, 1 2 to 1 5 % S i Q , 1 0 to 1 5 % Fe, and 2 0 to 2 5 % S. Low-grade con centrates containing 5 to 6 % molybdenum a r e obtained by concentration of oxidized o r e s containing powellite.

- -

Roasting Trapping of dust < Gases t o the atmosphere Dust t o the ex traction of R e

.1
Calcine

First leaching

Soda ash

solutions Filrration
I _

CaC1,

&\ Filtrate

Returned cake Decantate to the exrrac ion of hlo. Re

Precipitation Soda ash

7
Water Water Soda ash

Calcium molybdate

Fifth leaching

I 7 Washing

7
\

Wash

- 5 Filtration
.1 Waste residues

Filtration Calcium molybdate

Filtrate to waste;

I
Drying

5
Packing FIGURE 48. Production of calcium molybdate from low-grade molybdenum Concentrates.

The processing of intermediate copper-molybdenum products and l o w grade powellite concentrates is shown in Figure 48. The processing

90

includes roasting of sulfide concentrates, leaching with soda a s h solutions and precipitation of technical grade calcium molybdate from solution / 15/. In o r d e r to ensure a high degree of extraction of the molybdenum, the calcines from the roasting of low-grade sulfide concentrates and powellite concentrates a r e leached with soda solutions which differ from ammonia water in that they readily decompose all molybdates. Leaching. The following reactions take place during leaching with soda solutions:
MoO, Na,COs = Na,Mo04 CO,;
PCuMoO, 2Na,COs H20 = 2Na,Mo04 CuCO,.Cu (OH), CO,;
FeMoO, Na,COs H,O = Na,MoO., Fe (OH), CO,;
CaCOs.
CaMoO, + Na,CO, Na,MoO,

+ +

+ + +

A p a r t of the siltca, phosphorus, arsenic, and a fraction of the copper in the concentrates a r e dissolved together with the molybdenum. Copper is probably dissolved in the form of r a t h e r unstable complexes of the type xCuC03-yNa2CO3. However, if the solution a t the end of the leaching is neutral or slightly alkaline, copper carbonate complexes a r e de composed and the copper is precipitated a s basic carbonates. Most of the silica i s also precipiated under the s a m e conditions (hydrolysis of NazSiO, with precipitation of HzSiQ). An 8 to 1 0 % soda solution is used for leaching, which is c a r r i e d out in 4 to 5 countercurrent stages ( s e e Figure 48). This r e s u l t s in a b e t t e r utilization of the soda and p e r m i t s the removal of solutions which a r e neutralized to a pH of 8 to 8.7 by f r e s h portions of the m a t e r i a l to be leached. The leaching i s c a r r i e d out in i r o n r e a c t o r s fitted with s t i r r e r s and heated by dry steam which p a s s e s through a steam jacket or coil. The filtered solutions containing 50 to 70 g of molybdenum p e r l i t e r a r e sent to precipitation of calcium molybdate. Precipitation of calcium molybdate. The addition of calcium chloride to the sodium molybdate solution precipitates calcium molybdate. The precipitation is c a r r i e d out in wood or rubber-lined iron vats at 80 to 90". The degree of precipitation depends on the pH of the solution, the amount of precipitant added and the initial molybdenum concentration. In o r d e r to precipitate at l e a s t 97 to 98% of the molybdenum, the p r o c e s s must be c a r r i e d out in neutral or weakly-alkaline solution, using a small (10 to 1 5 % ) excess of calcium chloride which prevents contamination of the precipitate by calcium sulfate. The white, fine crystalline calcium molybdate precipitate is washed with water t o remove sulfates and is then separated by filtration and ignited at 600 to 700" in muffle furnaces.
TABLE 1 7

Chemical composition of standard calcium molybdate Minimum molybdenum

Maximum impurity contents,%

Brand

MDKl MDKz
44.0

91

The approximate chemical composition of calcium molybdate used in f e r r o u s metallurgy for the introduction of molybdenum into s t e e l and in the smelting of ferromolybdenum is shown in Table 17. The concentration of molybdenum in the mother liquor from the precipitation of the calcium molybdate is about 1 g/l. Molybdenum is extracted f r o m these solutions by sorption on ion-exchange resins. The sorption is c a r r i e d out from weakly-acidic solutions (pH 2 to 3), using the anionic r e s i n of AN-1 brand. The sorption capacity of this r e s i n for molybdenum is 20 to 22%. The molybdenum is eluted with a solution of ammonium hydroxide, and the elution yields a solution containing 50 to 7 0 g of molybdenum p e r liter.

20.

PRODUCTION OF METALLIC MOLYBDENUM

A11 the methods for the production of metallic tungsten ( s e e previous chapter) can also be applied to the production of molybdenum. Molybdenum trioxide may be reduced to the metal by hydrogen, carbon and carboncontaining gases, a s well a s by thermal reduction with aluminum and silicon. The production of pure molybdenum powder (which is then converted to the solid metal) is based on the reduction of M o Q with hydrogen. P u r e molybdenum trioxide, which i s the starting m a t e r i a l for the production of molybdenum powder, is prepared by ignition of ammonium paramolybdate at 450 to 500" in muffle or r o t a r y tube furnace ( s e e Figure 11).

Reduction of molybdenum trioxide with hydrogen The reduction of molybdenum trioxide with hydrogen is analogous to the reduction of W Q described above. T h e r e a r e , however, some differences in the conditions of the reduction. The reduction is a three-stage process, corresponding to the existence of three molybdenum oxides:

+ H, 2 Mo,O,, + H,O; Mo.,Oll + 3H, 4Mo0, + 3H,O; MOO, + 2H, Mo + 2H,O.


4Mo0,

(1)

(2)

(3)

The equilibrium conditions for these reactions a r e slightly different f r o m those f o r tungsten. The equilibrium constant of all these reactions is expressed by the r a t i o of the equilibrium partial p r e s s u r e s of water vapor and hydrogen: The equilibrium constants for the reduction of MOO, a t 400 to 700" according to reactions (1) and (2) a r e higher than those for the reduction of WO,. This means that for an equal moisture content in the hydrogen, the f i r s t reduction stage ( M o Q MoG) takes place at temperatures lower than those required for the reduction of W Q to WO,.
--f

92

TABLE 18

Equilibrium constants of t h e reduction of MOO, and WO, by hydrogen

Temperature. ' C
MOO,
--f

Mo

wo,

. b

700

7 60 850 950 1100

0.38 0.40 0.57 0.72 1.13

0.40

0.51 0.72 1.0 1.44

The equilibrium constant f o r the reduction of MOO, by hydrogen (reaction (3)) is lower than that for the reduction of W 4 ; therefore at a given hydrogen : w a t e r vapor ratio, the reduction of M o a to the metal occurs at higher temperatures than that of W O , ( s e e Table 18). In accordance with the above, the f i r s t stage in the reduction of Moo3 is c a r r i e d out at low temperatures (about 450 to 550") while the third stage is c a r r i e d out at 1000 to 1100". In industrial practice the reduction of molybdenum trioxide i s c a r r i e d out in two or three stages. In the f i r s t stage ( M o o 3 Mooz) the temperature along the furnace tube r i s e s from 450 to 650"; the formation of M o a must be virtually completed before the temperature reaches 550" since the intermediate molybdenum oxide and M o Q form a eutectic which melts at 550 to 600". In the second reduction stage ( M o a Mo) the temperature along the furnace tube changes from 6 5 0 to 950". The molybdenum powder from the second reduction usually contains 0.7 to 2 % oxygen. To eliminate this oxygen a third reduction is c a r r i e d out at 1000 to 1 1 0 0 ~ The capacity of the boats during the second reduction is about twice and in the third reduction about five times that of the f i r s t reduction. This is due to the difference between the bulk density of M o Q (0.4 to 0.5 g/cm3), Mooz (1 to 1 . 5 g / c m 3 ) , and Mo ( - 2.5g/cm3). The f i r s t and second reductions a r e c a r r i e d out in furnaces with 9 to 11 chromium-nickel steel tubes. These furnaces have been described in the chapter on tungsten ( s e e Figure 13). At 1000 to 1100" the resistance o f the chromium-nickel s t e e l tubes and the nichrome h e a t e r s to the effect of a i r is noticeably lower than at lower temperatures. The third reduction is therefore c a r r i e d out in tube furnaces with a hermetically-sealed jacket which is filled with hydrogen in o r d e r to protect the tubes and h e a t e r s from oxidation. The molybdenum powder obtained after the third reduction contains about 0.25 to 0.370 oxygen. Its approximate granulometric composition is a s follows:

Particle size, p Fraction. %

...... > 0.6 .... . .... . 40-60

0.6 -1.2 20 -30

1.2 -1.8 10 -20

1.8-2.4 1-3

The average particle s i z e of the powder is 0.5 to'2 E * . This is l e s s than the particle size of tungsten powder. This is due to the low temperature of the f i r s t reduction stage and the consequent absence of noticeable vaporization of the oxides ( s e e p. 43).
93

I 11111111111 I l l I I

The f i r s t reduction s t a g e s have lately been c a r r i e d out in continuous rotary-drum furnaces, resembling those described in the chapter on tungsten.

21.

PRODUCTION OF SOLID, DUCTILE MOLYBDENUM

Solid molybdenum is prepared by powder metallurgy techniques o r by the recently developed a r c -melting method.

Powder metallurgy method The nature of the powder metallurgy method has been described in the chapter on tungsten. We shall deal h e r e only with some special features of this p r o c e s s a s applied to the production of ductile molybdenum. In the production of relatively s m a l l b a r s , with a c r o s s section between 2 and 1 6 cm2 and a length of 4 5 0 to 600 mm, the molybdenum powder is p r e s s e d in s t e e l dies a t 3 tons/cm2. The p r e s s e d molybdenum b a r s a r e s t r o n g e r than s i m i l a r tungsten b a r s , owing to the finer particle s i z e of the molybdenum powders and also to the fact that they p o s s e s s some plasticity. The presintering and high-temperature sintering of the b a r s is c a r r i e d out in the s a m e equipment a s that used in the production of tungsten. After a presintering in hydrogen at 1100 to 1200", the b a r s a r e sintered at high temperatures, the maximum temperature being 2200 to
2400".

In the c a s e of b a r s with a c r o s s section of 1 8 X 18 mm the maximum c u r r e n t applied ( - 90% of the fusion current) is about 4500 amp. As a result of the fine particle s i z e , the b a r s contract rapidly. Contraction i s complete within 1 0 to 15 minutes a t the maximum temperature. The porosity of the b a r is reduced from 4 0 % f o r the p r e s s e d b a r to 10 to 15%. The sintered b a r s have a g r a i n s i z e corresponding to 2500 to 5000 grains/ mm2. Large b a r s with a c r o s s section of 4 0 X 4 0 and 60x60" a r e sintered in installations operating under m o r e d r a s t i c conditions, viz., at c u r r e n t s of 8000 to 12,000 amp. The temperature is increased gradually, over 6 0 to 80 minutes, to the maximum temperature. The sintering time at this temperature is 50 to 70 minutes. The method cannot be used f o r the production of l a r g e r b a r s since direct heating of l a r g e r b a r s by the passage of a c u r r e n t is accompanied by a l a r g e temperature drop from the center to the surface of the bar, resulting in nonuniform properties of the b a r a c r o s s i t s section. It is difficult, moreover, to obtain a uniform density on p r e s s i n g l a r g e b a r s in s t e e l dies. Uniform l a r g e b a r s , weighing 3 0 to IOOkg, a r e produced by hydrostatic p r e s s i n g and such b a r s a r e sintered by indirect radiation heating. In hydrostatic p r e s s i n g the metal powder is contained in an elastic ( e . g . , rubber) shell which is placed in a hermetically-sealed compression chamber and water (or some other liquid) under p r e s s u r e is applied 1500kg/cm2 is created in the hydrostatic ( F i g u r e 49). A p r e s s u r e of chamber and the powder is subjected to uniform compression in all directions.

94

In contrast to conventional p r e s s i n g in a s t e e l die, hydrostatic compres sion is distinguished by the absence of friction between the powder and the die walls. It thus eliminates nonuniformity in the p r e s s e d b a r . In o r d e r to obtain regular-shaped b a r s , e. g . , a cylindrical o r rectangular c r o s s section, the rubber shell filled with the powder is placed in a perforated s t e e l c a s e in the shape of a tube or a rectangular c r o s s section, the rubber shell filled with the powder is placed in a perforated s t e e l c a s e in the shape of a tube or a rectangular box 1161. Large b a r s a r e sintered in muffle furnaces with molybdenum h e a t e r s ( s e e Figure 50).

FIGI'ItE - P I .

Iiydrostatic pressing iriachine.

1 -high-pressure pump; 2- hai in her; 3 - 1 n i n o mt'ter; I -closing device (lid); i-pressure


release valve: t i -pressed bar i n elastic shell.

FIGUItI: 5U. hluffle furnace with m o l y b denum heaters for sintering large inolyb

1- molybdenum heaters. 2 - refractory


lining: 3 - opening for the measurement of the temperature. 4 - thermal insulation. 5 - molybdenum bar. 6 - water jacket.

Because of the high sintering temperature, the furnace i s lined with alumina and zirconia refractories. The sintering temperature may be reduced to 1 6 0 0 to 1700" by increasing the duration of sintering and conducting it in an atmosphere of moist hydrogen obtained by bubbling hydrogen through water at 20 to 40". Sintered b a r s with a density of up to 1 0 g / c m 3 a r e obtained by holding for several hours at 1600 to 1700". In addition to muffle furnaces, induction furnaces may also be used for sintering l a r g e b a r s .

The melting of molybdenum

/ 8, 171

Owing to the need f o r l a r g e b a r s for the drawing of tubes, rolling of l a r g e sheets and for other products, an arc-melting method which p e r m i t s the production of ingots weighing 500 to 2000 kg h a s been developed. In all modern a r c furnaces the melting is c a r r i e d out in vacuo, using a consumable molybdenum electrode and cooled copper crucibles (molds). Because of the high t h e r m a l conductance of copper and the rapid heat removal, the liquid m e t a l coming into contact with the crucible walls solidifies and thus no interaction takes place between the copper and the molybdenum. The m e t a l is melted in the flame of an a r c created between

95

the upper electrode (consumable molybdenum electrode) and the lower electrode which consists of the molten metal in the copper mold. A direct c u r r e n t is used for melting, the consumable electrode being used a s the cathode and the melt a s the anode. The c u r r e n t is rectified with the aid of g e n e r a t o r s or l a r g e selenium or germanium r e c t i f i e r s (8000 a m p e r e s o r more).

FIGURE 51. Vacuum arc furnace melting of molybdenum.


1-window; 2 -melting chamber; 3-hopper for the supply of alloying additives; 4-consumable molybdenum electrode; 5- tube; +toller feed mechanism; I-peephole; 8-valve; 9-vacuum block; 1 0 -vacuum-measuring lamp (LT-2); 11 -vacuutii valves; 1 2 -liquid nitrogen trap; 13-rotary (preliminary) vacuum pump: 14-diffusion pump; 15-ceramic ring; 16-bottom plate; 1 7 cooled bottom of crystallizer; 18-cooling-water connections; 19-solenoid; 20 -cooled copper crystallizer.

In the most extensively used a r c furnaces the electrode is prepared in advance from sintered molybdenum bars. The b a r s a r e connected by butt welding to electrodes 1 to 2.5m long. The electrodes a r e then joined into packs, 4 t o 1 6 b a r s o r even m o r e being joined together depending on the s i z e of the crystallizers. The consumable electrode is moved during melting with the aid of a rod (which also s e r v e s a s the c u r r e n t c a r r i e r ) a t whose end the electrode is fastened. It may also be moved by a r o l l e r drive mechanism (Figure 51) / 1 7 / . A molybdenum disc is placed on the bottom of the c r y s t a l l i z e r before the melting. The furnace is evacuated to a p r e s s u r e of to Hg and the electrode i s lowered to s t r i k e the a r c . In some furnaces the bottom

96

I
plate of the crystallizer descends together with the ingot a s the melting proceeds (Figure 52) / 8 / . A constant a r c voltage of 30 to 40 V is maintained by means of automatic control of the distance between the electrodes (i. e., a r c length), which ranges from 1 0 to 25" depending on the melting conditions.

FIGURE 52.

Furnace with moving ingot for melting molybdenum.

1-ingot electrode; 2 -water-cooled copper mold; 3- ingot depression device; 4 -device for extracting the ingot from the mold: 5-consumable electrode; 6-connection to the vacuum line; I-booster pump; 8-diffusion pump; 9-rotary pump.

A magnetic coil (solenoid) which is placed off-center with respect to the crystallizer is used to control the shape of the a r c and to prevent the formation of side a r c s [e. g . , between the electrode and the crucible walls). The magnetic field of the solenoid also causes mixing of the metal. In o r d e r to obtain ductile metal, the molybdenum must be deoxidized during melting. The oxygen content of the molten molybdenum must not

97

exceed 0 . 0 0 2 ~ 0 , Carbon, which is added to the consumable electrode during i t s manufacture s e r v e s as the deoxidizer. Zirconium and titanium are also used as deoxidizers. The residual carbon which constitutes a f the metal is in the form o f isolated few hundredths of one percent o carbide occlusions and does not interfere with hot forging of molybdenum ingots. The alloying additives may be introduced a s components of the consum able electrode or may be fed into the a r c space ( a s shown in Figure 51). Since only a s m a l l fraction of the metal is in the molten s t a t e during a r c melting, the distribution of alloying additives is nonuniform. A second melting is usually c a r r i e d out in o r d e r to obtain ingots of uniform composition. The melting r a t e depends on the c r o s s section of the consumable electrode and on the melting conditions (current strength, the number of ampere-turns in the solenoid, the a r c voltage). Under optimum conditions ( a c u r r e n t of 4300 amp, a solenoid with 2500 to 3000 ampere-turns, and a voltage of 38 to 39 V) the melting r a t e of an electrode with a cross section of 30.7 cm2 is 75mm/minute. The consumption of electrical energy in melting is about 1 kilowatt-hour p e r kg metal. The ingots formed have a c o a r s e crystalline s t r u c t u r e (their grain size is 0.05 to 0.1 mm) which interferes with p r e s s u r e molding. However, processes have been developed to produce sheets, wires, tubes, and other molybdenum products from these ingots. In addition to a r c melting, electron-beam melting h a s been recently used f o r melting refractory metals including molybdenum. The nature of the electron-beam melting p r o c e s s will be discussed in the chapter on tantalum and niobium. Electron-beam melting increases the purity of the molybdenum from 99.8 to 9 9 . 9 9 % . A decrease in the concentration of many impurities occurs: 0, N, C, Si, P, As, Fe, Cu, Ni, Mn, Co, and Ti.

98

Chapter III
TANTALUM AND NIOBIUM
22. GENERAL DATA ON TANTALUM AND NIOBIUM

In 1801 an English chemist, Hatchett, discovered the element niobium (columbium) in a mineral from Columbia. He called it "columbite". Tantalum was discovered in two minerals, one from Sweden and the other from Finland, by a Swedish chemist, Ekeberg, in 1802. The name tantalum was chosen because of the "tantalizing difficulties" encountered when attempting to dissolve the oxide of the new element in acids. Later, the two new elements were considered identical. In 1844 Rose, a German chemist, proved that the mineral columbite contained two different elements: niobium, deriving i t s name from Niobe, symbol of maternal sorrow and the daughter of Tantalus, and tantalum. In 1865 Marignac showed that tantalum and niobium could be separated utilizing the difference in the solubilities of the fluoride complexes of these s a l t s , a method which found industrial application and is in use even today. For a long time all attempts to produce pure compact metallic tantalum and niobium m e t with failure. P u r e tantalum was produced f o r the f i r s t time in 1903 and niobium in 1907 by von Bolton. The commercial production of tantalum dates from 1922, that of niobium from 1928.

P r o p e r t i e s of tantalum and niobium Tantalum and niobium belong to Group V of the periodic table. The chemical and physical properties of these two elements a r e similar. They almost always occur together in o r e s . Consequently, their metallurgy is usually discussed jointly. Tantalum and niobium a r e steel-gray metals; tantalum has a faint blue hue. The pure metals a r e ductile and may be rolled into thin foils of 0.01 m m thickness without heating o r intermediate annealing. Some of the physical properties of tantalum and niobium a r e listed in Table 19. It should be noted that their melting points a r e higher and their electron work functions lower than those of other refractory m e t a l s The l a t t e r property c h a r a c t e r i z e s the electron (tungsten and molybdenum) emission capacity of the metals. It is of importance for the u s e of tantalum and niobium in vacuum tube industries. Both metals, especially niobium, have a high superconductance transition temperature. Tantalum and niobium a r e not affected by a i r a t room temperature. Oxidation begins with the formation of an iridescent film when the m e t a l s

99

are heated to 200 to 300". Rapid oxidation with the formation of Nb205 and Ta& takes place above 500".

TABLE 1 9
Physical properties of tantalum and niobium
Properties Atomic number Atomic weight. Density, g/cm3.. Lattice type Tantalum 73 180.88 16.65 b.c. c. a = 3.296 A
3000

__

Niobium 41 92.91 8.57 b.c.c. a = 3.294 A 2470 1 1 0 -4930 0.065 0.083 0.13 7.1 ' 13.2. 9.17 0.00395 4.01 6.3 19 30 70

..................................

.. ........ . ....... .. . ..... ., ... ........................... ,... .... .. , .. ............. .., .....
.... .... .... .. ..... ... .... . ..... ...... ...............

Temperature. 'C melting point , boiling point Specific heat, cal/g " C at: 0- 100 1600 Thermal conductivity ( at 20-loo"), cal/cm*sec, ' C Linear expansion c o e f f i c i w t ( 0 to 1 0 0 C ) . Electrical resistivity (20" C) ohm. c m , Superconductance transition temperature, ' K Temperature coefficient of electrical resistivity ( 0 to lOO"C).. * Electron work function, e V Total radiation (watt/cm2) at: 1330 C 1730 C 1930C 2330C 2530-c Thermal neutron cross section, barns Tensile strength of annealed sheet, k g / m m 2 Elongation of annealed sheet,% . Modulus of elasticity, k g / m m z Microhardness of annealed sheet ( 3 0 g load), kg/mm2

50

5300

................................ ................................. ........ . .. .. ..... ...... .... . .... ..... . . .. ... .. ..... . .... ....... ....... ................................ ................................ ................................ ................................ ................................ . .......... ..... ... ... ... ...... . .... ... ..... ..... ... . ..... . .... . . .. ....... . .. ... ..
. .

0.034 0.041 0.13 6.5-10-' 12.5' 10-6 4.38 0.00382 4.12 7.2 21.2 33
75

108 21 35-50 25-45 19000 108


. .

1.15 30-35 u p to 49 10600

Tantalum and niobium a r e characterized by their ability to absorb g a s e s such a s hydrogen, nitrogen, and oxygen. The presence of smaIl amounts of these elements strongly affects the mechanical and electrical properties of the metals. The absorption of hydrogen a t low temperatures i s slow; it occurs at the maximum r a t e a t about 500" in tantalum and 360" in niobium. At these temperatures both absorption and the formation of chemical compounds such a s NbH and TaH take place. Absorbed hydrogen makes the metal brittle. It may be expelled almost quantitatively by heating in vacuum above 600", the original mechanical properties thus being restored.

100

Tantalum and niobium absorb nitrogen a t temperatures a s low a s 600". At higher temperatures the nitrides NbN and TaN, with melting points of 2300 and 3087" respectively, a r e formed. Carbon and carbon-containing g a s e s interact with the metals a t high temperatures (1200 to 1400") with the formation of the hard, refractory carbides, TaC, melting at 3880, and NbC, melting at 3500" Tantalum and niobium combine with boron and silicon yielding the hard, refractory borides and silicides, NbBz (mp 2900"), T a & (mp 3000") and TaSiz (mp 2400"). Tantalum and niobium a r e not attacked by hydrochloric, sulfuric, nitric, phosphoric o r organic acids a t any concentration at temperatures below 100 to 150". The resistance of tantalum to hot hydrochloric and sulfuric acids is higher than that of niobium.
The metals a r e soluble in hydrofluoric acid, and dissolve very rapidly in a mixture of hydrofluoric
and nitric acids. Tantalum and niobium a r e l e s s resistant to alkalies. Hot solutions of alkali hydroxides cause noticeable etching of the metals, which a r e oxidzed rapidly by fused alkali hydroxides and soda ash with the formation of sodium tantalates and niobates. The properties of tantalum and niobium compounds The chemical compounds of pentavalent niobium and tantalum a r e the most stable and useful. Compourids in which the m e t a l s a r e in the di-, t r i - and tetravalent s t a t e a r e also known. Oxides. The higher oxides N b z 0 5 and TazO5 a r e usually obtained a s intermediate products in the commercial processing of o r e s . Both oxides a r e white; Nbz05 becomes yellow upon heating, but the yellow color disappears upon cooling. Some properties of the oxides a r e shown below:
Nb2.05 TazO5

Density Melting point, ' C Heat of formation, kcal/mole

............................ 4.55 .................... - 1 5 1 0 ......... 455.2

- 1620
488.8

8.71

The l a r g e difference between the densities of the oxides p e r m i t s an approximate evaluation of the chemical composition of a mixture of the oxides from density measurements. The oxides Nb205 and T a 2 0 5have distinctly acid properties. When fused with alkalies they yield niobates and tantalates, i. e . , s a l t s of niobic and tantalic acids. Precipitates consisting of the hydrated oxides Taz06 -xHzO and Nb205.xHZ0 a r e formed when acid solutions containing tantalum and niobium a r e neutralized o r when the niobates and tantalates of alkali metals a r e treated with acids. The hydrated oxides dissolve in concentrated hydrochloric and sulfuric acids, but a r e precipitated due to hydrolysis when the solutions a r e diluted. Twolower oxides of niobium a r e known: N b a (black) and NbO (gray powder); they a r e formed when Nb205is reduced by hydrogen or carbon. The lower tantalum oxides T a Q and TazO have been mentioned in the literature, but their existence has not been confirmed in m o r e recent studies.

Niobates and tantalates. Niobic and tantalic acids have not been isolated i n the f r e e state, but s a l t s of the m e t a and ortho acids, HROB and H3R04, where R is Nb or Ta, are known t o exist. In addition t o the ortho- and metaniobates (-tantalates), there also exist polyniobates and polytantalates whose composition corresponds to the general formulaxMepO yR205where MelO is an alkali metal oxide and R is T a or Nb. Depending on the amount of the alkali hydroxide used, the fusion of niobium or tantalum pentoxide with alkali hydroxides yields either the ortho salts Me3R04 or the m e t a s a l t s MeROs. The m e t a s a l t s are also formed in the interaction of tantalum or niobium pentoxide with solutions of sodium or potassium hydroxide at 150 t o 200" 120, 2 1 / . The anhydrous m e t a s a l t s a r e virtually insoluble i n water. The ortho s a l t s dissolve in water with hydrolytic decomposition which r e s u l t s in the formation of complex hydrated polyniobates (tantalates) of various composition / 21/ :

4Me@.3R,0j .nH,O( o r Me,R,O,,.nH,O) ; 7Me,0.5R,O,.nHZO ( or Me7R5OlO~nH,O); 7Me,0.6R,O,~nHZO (or Me1,R,,O9,.nH2O).

Thus, f o r instance, the following polyniobates (polytantalates) a r e formed when N b 2 0 5 or Ta205 is fused with an alkali hydroxide and the cake is then treated with water: 4 ~ ~3Nb205.16H20; 0 .
4K,0.3Ta2O5. 16H20; 7Na,0. 6Nb,05.32H,O; 4Na20.3Ta,0,.25H20. The potassium s a l t s dissolve readily in water, while the sodium s a l t s a r e sparingly soluble. According to Spitsyn and Lapitskii, the solubility of sodium niobate (the 7 : 6 salt) in water is 1 5 g / l at 20" and 2 6 g / l a t SO". The presence of an alkali in the solution causes a s h a r p drop in the solubility. The solubility of the tantalum s a l t in water a t 14" is about 0.2g/lOO m l 181. Tantalum and niobium fluorides. The higher niobium and tantalum oxides dissolve in hydrofluoric acid with the formation of solutions of the fluorides NbF5 and TaF5; in the presence of an e x c e s s of hydrofluoric acid these fluorides yield complex acids:
TaF, NbF,

+ 2HF 2 H, [TaF,], + 2HF H, [NbF,].

In a dilute acid solution (at HF concentrations below 7 % ) NbF5 hydrolyzes with formation of an oxyfluoride NbOF3 and the corresponding oxyfluoro niobic acid H2[NbOF5]. The addition of potassium fluoride to a solution containing tantalum or niobium fluoride r e s u l t s in the formation of complex fluorine-containing s a l t s (K2TaF7, K2NbF7, KzNbOF5) which play an important role in the manufacture of tantalum and niobium. Potassium oxyfluoroniobate crystallizes from solutions with an HF concentration below 770 as platelets of the monohydrate (K2NbOF5. HzO). The compounds K2TaF7 and K2NbFG crystallize from solutions a s isomorphous needleshaped c r y s t a l s . The solubility of K2TaF7is much lower than that of KzNbOF5. HzO. This fact is utilized in the separation of niobium from tantalum.
1455

I02

The anhydrous tantalum and niobium fluorides NbF5and TaF5a r e very hy groscopic, deliquescent substances which fume in a i r . They a r e formed when metallic tantalum o r niobium o r their pentoxides a r e treated with fluorine o r hydrogen fluoride; the fluorides have low melting points.
NbF,

TaFS 95 229.2

Melting point, 'C Boiling point, C

................... 80 ................... 234.9

Niobium and tantalum chlorides. The chlorides NbCl, and TaC15 a r e formed when the m e t a l s o r the oxides a r e treated with chlorine o r other chlorinating agents. Niobium oxychloride NbOCl, is preferentially formed when N b z 0 5 is treated with chlorine in the presence of carbon. The chlorides have low melting and boiling points:

Melting point, C Boiling point, C Density

................... ................... .............................

. .

NbC1, 204.7 248.3 2.75

TaCIS 216.5 234.0 3.68

Niobium oxychloride volatilizes a t about 400". The vapor p r e s s u r e of NbOC1, a t 330" is 482mmHg. The chlorides a r e very hygroscopic and tend to hydrolyze. Precipitates of white hydrated oxides a r e formed in aqueous solutions a s a result of hydrolysis:
2TaCI,

+ ( x +- 5)H,O = Ta,O,.xH,O + IOHCI.

Lower niobium and tantalum chlorides a r e also known: NbC14, NbCl,, TaC14, TaC13, and TaC12. They can be prepared by the reduction of the higher chlorides with hydrogen o r metais such a s aluminum. NbC15 is reduced m o r e readily than TaC15; this is the principle of some methods f o r the separation of niobium and tantalum. P e r a c i d s of niobium and tantalum. The freshly-precipitated hydrated oxides of niobium and tantalum dissolve in a mixture of dilute sulfuric acid and hydrogen peroxide yielding peracids whose composition i s either HRO, o r H3R08, depending on the number of peroxide groups. There exist s a l t s of peracids with a l a r g e r number of peroxide groups, niobate and tantalate peroxides of the type K3Nb08and K,Ta%. Solutions of the peroxide compounds a r e colorless. Complexes with organic acids. Soluble complex acids and their s a l t s a r e formed by +he reaction of niobic o r tantalic acid with oxalic, t a r t a r i c , and other s i m i l a r acids. In these complexes, p a r t of the oxygen in the niobic o r tantalic acid is replaced by the organic acid radical. F o r instance, the complex is formed with oxalic acid:
H, [RO (C20J31, where R-Nb

or Ta

Niobic and tantalic acids r e a c t with tannin (an e s t e r of glucose and digallic acid) to form an adsorption complex resulting from the neutraliza tion of oppositely charged colloidal particles. Tannin h a s a positive charge while the niobium and tantalum compounds a r e negatively charged, The

103

lemon-yellow tantalum complex is precipitated after boiling in a weakly acid solution at pH 3 to 4. The orange-colored niobium complex is precipitated from neutral o r very weakly acid solutions in the presence of an e x c e s s of tannin. The difference in the conditions of precipitation of the tantalum and niobium complexes p e r m i t s the separation of tantalum f r o m niobium, a fact which finds use in quantitative chemical analysis. The complexes of fluoroniobic or fluorotantalic acid with some organic solvents, e. g . , tributyl phosphate and methyl isobutyl ketone, a r e of technological importance (for their composition s e e p. 125).

Uses of tantalum and niobium Vacuum-tube technology and electrical technology. Tantalum was used for the f i r s t time in 1900 to 1903 for the production of incandescent filaments in electric bulbs, but was subsequently replaced by tungsten in 1 9 0 9 - 1910. The extensive use of tantalum i s associated with the development of vacuum-tube technology, including the production of radio, r a d a r and X-ray equipment. Tantalum and niobium a r e characterized by a combination of very valuable properties (high melting point, high emissivity, and the ability to absorb gases) which p e r m i t their use in the production of vacuum-tube p a r t s . Their ability to absorb g a s e s i s utilized in o r d e r to maintain high vacuum in radio and other vacuum tubes. Tantalum and niobium sheets and b a r s a r e used for the production of heated p a r t s - anodes, grids, indirectly heated cathodes, and other p a r t s of vacuum tubes and especially of high-power generator tubes. Tantalumniobium alloys a r e used for s i m i l a r purposes. Recently, tantalum and niobium have found important use in the manufacture of e l e c t r o l y t i c c o n d e n s e r s a n d r e c t i f i e r s . Advantage i s taken of the capacity of tantalum and niobium to form stable oxide f i l m s during anodic oxidation. The anodic films a r e stable in acid electrolytes and p e r m i t the flow of current in one direction only, from the electrolyte to the metal. The resistivity of Ta205films in the non conducting direction i s very high ( 7 . 5 10" ohm. cm) and i t s dielectric constant i s 1 1 . 6 . Solid-electrolyte condensers made of tantalum and niobium a r e characterized by a high capacitance and s m a l l size, by a high insulation resistance, 2 to 3 times higher than that of aluminum condensers, and by the stability of the oxide film. They can be used over a wide temperature range from -80 to +200". Miniature tantalum condensers a r e extensively used in radio transmitters, r a d a r setups, and other types of equipment. The tantalum o r niobium components of the condensers a r e prepared f r o m thin foils o r porous plates which a r e produced by p r e s s i n g the metallic powder. Construction of chemical apparatus. The corrosion resistance of niobium and especially of tantalum in acids and other media, together with their high electrical conductivity and ductility, make these metals valuable s t r u c t u r a l m a t e r i a l s for chemical and metallurgical equipment.
e

I04

Tantalum and niobium a r e used f o r the production of condensers, heaters, lining of vessels, s t i r r e r s , and other parts. Tantalum s e r v e s instead of platinum in the production of spinnerets f o r the drawing of synthetic silk threads. Atomic energy. Niobium i s characterized by a combination of properties that satisfy the requirements of atomic energy with respect to s t r u c t u r a l materials: high melting point, good workability, corrosion resistance, and a relatively low thermal neutron capture c r o s s section of about 1.15 barn. Below 900" t h e r e is almost no interaction between niobium and uranium, so that niobium is used to make the protective shells of uranium fuel elements for nuclear reactors. This p e r m i t s the use of liquid-metal coolants - sodium o r sodium-potassium alloys with which niobium does not r e a c t below 600". In o r d e r to i n c r e a s e the s e r v i c e life of uranium fuel elements, the uranium is alloyed with about 7% niobium. The addition of niobium stabilizes the protective oxide film on the uranium and thus i n c r e a s e s i t s resistance to water vapor. The production of refractory and hard alloys. Niobium and tantalum a r e components of various refractory alloys used in the turbines of jet engines. The properties of molybdenum, titanium, zirconium, aluminum, copper and their alloys a r e markedly improved by the addition of niobium o r tantalum. Niobium-based refractory alloys have been proposed a s a structural material for various p a r t s of jet engines and rockets, specifically for the production of turbine blades, the leading edges of wings, the foreparts of planes and rockets, and the lining of rockets. Niobium and niobium-based alloys may be used a t service temperatures of 1 0 0 0 to 1200". Niobium and tantalum carbides a r e components of some types of tungsten carbide hard c e r m e t s for metal cutting. The production of steels. Tantalum and especially niobium a r e extensively used a s alloying additives in steels. Niobium became of i n t e r e s t in 1933 to 1 9 3 4 in connection with the discovery of i t s effect on chromium-nickel stainless steels. The addition of niobium i n an amount 6 to 1 0 times higher than that of carbon in the steel eliminates intercrystalline corrosion of stainless steel and prevents damage to welded joints. In addition, niobium and tantalum a r e added to various refractory s t e e l s (especially those used for gas turbines) a s well a s to tool and magnetic steels. Niobium i s added to s t e e l s a s an alloy with iron (ferroniobium) containing up to 6 0 % Nb. Ferrotantaloniobium with various tantalum : niobium ratios in the alloy i s also used. Other uses. Tantalum wire and sheets a r e used in bone and plastic surgery, notably in the setting of bones, "patching" s k u l l damage, application of sutures, etc. The metal has no irritating effect on live tissues and does not impede the vitaL activity of the organism. Some tantalum and niobium compounds such a s fluoride complexes and oxides a r e used a s catalysts in organic synthesis.

I05

23.

MINERALS, ORES, AND ORE CONCENTRATES

The content of niobium in the E a r t h ' s c r u s t is 1 * by weight, whiIe by weight. Niobium and tantalum almost that of tantalum is 2 . always occur together in nature. They are components of about 100 dif ferent minerals, a l a r g e fraction of which consists of very complex s a l t s of niobic and tantalic acids. The minerals contain, in different ratios, iron, manganese, alkali, and alkaline-earth metals a s well as many r a r e elements : lanthanides, titanium, zirconium, throium, uranium, tin, antimony, bismuth, tungsten, and others. The most important tantalum and niobium minerals may be divided into two groups: 1. Tantaloniobates, which are s a l t s of niobic and tantalic acids. The main minerals in this group a r e tantalite and columbite. 2. Titano(tanta1o)-niobates, which a r e complex s a l t s of titanic and niobic (tantalic) acids. Almost all minerals in this group contain lanthanides. As in the tantaloniobates, the niobium :tantalum ratio v a r i e s within wide limits, but in most c a s e s niobium is predominant. Titanium is the predominant component of some minerals. Among the most important minerals belonging to this group a r e pyrochlore, loparite, koppite, betafite, and some others. The compositions and properties of the most important tantalum and niobium minerals a r e described below. T a n t a l i t e a n d c o l u m b i t e . Their composition can be expressed by the genera? formula (Fe, Mn) [(Ta, Nb)03I2. Accordingly, the minerals consist of an isomorphous mixture of four s a l t s : Fe (TaO,),, Mn (TaO,),, Fe (NbO,), and Mn ( N b O , ) , . Minerals with strongly varying ratios of niobium to tantalum and iron to manganese a r e encountered. The mineral is known a s columbite when the predominant component is niobium and tantalite when it is tantalum. The is total niobium and tantalum content of the mineral ( a s Nbz05+Taz05) 82 to 86%. Tantalite and columbite almost always contain titanium, tin, tungsten, and other impurity elements. The minerals a r e brown-black. Their density is strongly affected by the niobium and tantalum contents. Tantalum-free columbite has a density of 5. The density of tantalite is The density increases regularly between these limits with increasing 8.2. tantalum content of the mineral, and the tantalum and niobium contents of the mineral can be approximately evaluated from the density. The l e s s abundant minerals belonging to this group a r e : f e r g u s o n i t e tantaloniobate of yttrium, erbium and the elements of the cerium group; s a m a r s k i t e tantaloniobate of iron, calcium, yttrium, and cerium; s t i b i o t a n t a1 i t e - tantaloniobate of antimony, and others. P y r o c h 1o r e is a mineral with a very complex chemical composition. It is a s a l t of niobic and titanic acids (titanoniobate), containing sodium, calcium, lanthanides, and a number of other r a r e elements, a s cations. In addition, the mineral contains fluorine. Its general formula is (Na, Ca. ) a (Nb, Ti)z06[F,OH]. The color of pyrochlore is dark-brown, red-brown, or yellowish-green. Its density is 4.03 to 4.36.

..

106

M i c r o 1i t e is a mineral whose composition is analogous to that of pyrochlore, but with tantalum in place of niobium. Niobium and tantalum usually occur together in these m i n e r a l s since pyrochlore and microlite f o r m a s e r i e s of isomorphous mixtures. The composition of minerals in the pyrochlore-microlite s e r i e s is v e r y complex and variable:
0 % 0-70%
NbaOs, Ta,Os,
1-6% Na,O, 0-1.4% KaO, 4-18.1% CaO. 2-13.3% (Ln),O, 0-5.1% YaO,, 6 4 . 0 % SnO,, 0-5.794 ZrO,. 6 0 . 3 % WO,.

2-13.5% TiO,, 6 5 % ThOa. 0-11.4% UO , 0-15.5% Ud,

0-7.7%
Q10%

0-9.7%

MnO FeO F%O,

0-6.0%

HaO

The mineral koppite ( a calcium, sodium, and c e r i u m niobate) and a s e r i e s of other m i n e r a l s also belong to the pyrochlore-type group. L o p a r i t e is a titanoniobate of sodium, calcium, and the lanthanides. Its general formula is (Na, Ca, Ce.. .)z(Ti, Nb)2Q. The variations in i t s composition a r e relatively small. The mineral contains 39.2 t o 40% Ti&, 32 to 3 4 7 0 Ln&, 8 to 10% (Nb, Ta)zO5, 4.2 to 5.270 CaO, 7.8 to 9 % Na20, 2.0 to 3.4% SrO, 0.2 to 0.7% KzO, and 0.2 to 0.770 T h Q . The tantalum content i s lower (by a factor of 15) than that of niobium. The mineral is black o r gray-black and i t s density is between 4.75 and 4.89. The tantalum and niobium deposits a r e associated mainly with the socalled pegmatites. Pegmatite veins a r e characterized by a g r e a t variety of embedded m i n e r a l s and by their coarsely-crystalline structure. T h e r e a r e two types of pegmatites : t h e granite pegmatites and the alkaline nepheline syenite pegmatites. The minerals tantalite and columbite preferentially accumulate in the f i r s t type, while m i n e r a l s of the titano niobium group (pyrochlore, loparite, etc. ) a r e concentrated in the second type of pegmatites. Tantalite and columbite a r e concentrated not only in veined deposits but also in some placers. Up to 80% of all the tantalumniobium concentrates produced outside the USSR originate from placers. The pyrochlore o r e s from carbonatite deposits have recently become (1955 to 1956) an important source of niobium. The main non-metalliferous minerals in these o r e s a r e calcite, dolomite, ankerite, some silicates (phlogopite, forsterite) apatite and magnetite. In addition to pure niobium pyrochlores, the carbonatite also contains tantalum-containing pyrochlores (hatchettolite) a s well a s lanthanide m i n e r a l s - fluoro carbonates (bastnasite, parisite) 1231. The niobium and tantalum contents of o r e s a r e usually quite low, the total content of the pentoxides being between 0.003 to 0.170. In view of the complex nature of most of the deposits, the niobium and tantalum minerals a r e obtained a s by-products in the extraction of other valuable minerals, such a s c a s s i t e r i t e , wolframite, lithium minerals, beryl, zircon, etc. The main method for the beneficiation of o r e s containing columbite and tantalite is gravitational enrichment (wet jigging, tabling). The usual product i s an overall concentrate containing tantalite and columbite, a s well a s cassiterite, wolframite and some other minerals. F u r t h e r enrichment is produced by flotation and electromagnetic separation 1221. The approximate composition of the concentrates is shown in Table 20. In accordance with the official standards in f o r c e in the USSR, f i r s t grade tantalum concentrates must contain at l e a s t 6 0 to 6 5 % Ta205and

107

not m o r e than 10% NbzO5, while second-grade concentrates m u s t contain a t l e a s t 4 0 % Taz05; f i r s t - g r a d e columbite concentrates must contain at l e a s t 60% NbzQ and second-grade concentrates at l e a s t 50% Nb205 1 2 2 , 2 3 1 .

Composition

Tantalite 1 2 -30 53-59.6 2.09-15.7 2.95-17.2 0.250.136.4

Columbite

50-58
12-18 10.5-19.0 2.44-

8.3

6.8-11.6
0.12.7

0.9

Gravitational methods (jigging, tabling) a r e also the main means of enriching pyrochlore and loparite o r e s . The resulting "crude" concentrates a r e further enriched to the required concentration by flotation, electro magnetic and electrostatic methods 1 2 2 1 . According to Soviet standards, the pyrochlore concentrates, which a r e used mainly for the smelting of ferroniobium, must contain at l e a s t 3 7 % (Nb, Ta)205, while loparite concentrates must contain at l e a s t 870 The main western raw m a t e r i a l producing countries a r e : Brazil, Nigeria, Congo (Leopoldville), Mozambique, Australia, USA, Malaya, Norway and Western Germany.

(m,

am 1 2 2 , 2 3 1 .

24. PROCESSING OF TANTALUM-NIOBIUM

ORE CONCENTRATES The o r e concentrates a r e usually processed to yield three types of chemical compounds: oxides (Ta2O5, N'bzO5), fluoride complexes (K2TaF7 and KzNbF7), and chlorides. The selection of the technological p r o c e s s f o r the processing of concentrates is determined by the mineralogical and chemical composition of the concentrate, by the final product required, and by the degree of i t s purity. Tantalite and columbite a r e stable minerals which a r e virtually not attacked a t all by inorganic acids, except hydrofluoric acid. Thus, tantalite (or columbite) concentrates a r e decomposed by fusion with alkalies (NaOH, KOH) or by digestion with hydrofluoric acid. The m i n e r a l s belonging to the titanium-tantalum-niobium group (loparite, pyrochlore) a r e l e s s stable than tantalite and columbite, but their composition is m o r e complex. Their processing must provide for the separate extraction of all valuable components - niobium, tantalum,

I 0 8

. . I

I .

11.11

.,M.

111

1111.1

1111.1111111

I,.

.... .....,,...,,
I . ,

titanium, and lanthanides. The concentrates of these m i n e r a l s are decomposed by digestion with acids (H2S04, HF) o r by chlorination.

25. PROCESSING O F TANTALITE-COLUMBITE BY FUSION WITH SODIUM HYDROXIDE* / 2 /

Fusion with sodium hydroxide is a widely used method of decomposition of tantalite (columbite) concentrates. A flow s h e e t of the production of a mixture of tantalum and niobium oxides by this method is shown in Figure 53.

I
. 1

Tantalite-columbite (concentrate) Size-reduction

.1 .1

1
.1
Solution: Na2SiOs, Na,SnOs, Na,WO,, NaA10, and other impurities, and excess hydroxide To waste or to the precipitation stage in the regeneration of wastes

Fusion with NaOH

1
Aqueous processing of the m e l t Precipitate: sodium tantalate and niobate, Fe,Os, Mn,O,. NazTiO, and other impurities

-1
HC1 Solution: FeCl,, MnCl,

Decomposition Precipitate: (Ta. Nb),Os x H,O and impurities (Si, T i , Fe. etc. )

.1

. 1

Washing of precipitate and filtration Mixture of oxides ( T a , Nb),O,. x H,O

T o the separation of tantalum and niobium and salt production

FIGURE 53. Production of a mixture of tantalum and niobium oxides by fusion of the concentrate with sodium hydroxide.

_-_____________
A recently developed method for the decomposition of tantalite concentrates involves treatment with NaOH solutions a t 200' in autoclaves; this provides a high degree of decomposition and reduces the hydroxide consumption by a factor of three as compared with the fusion method 1241.

109

Fusion and aqueous processing of the melt, Assuming that ortho s a l t s are the only s a l t s formed, the p r o c e s s e s qccurring during the fusion may be approximately described by the following equations:
Fe (TaO& Fe (Nbo,),

+ 6NaOH = 2Na3Ta04+ FeO + 3H,O + 6NaOH = 2Na,Nb04 + FeO + 3H20.

Similar equations can be written f o r f i ( T a Q J z and f i ( N l ~ Q ) ~ . The ortho s a l t s undergo hydrolytic decomposition during the aqueous processing of the m e l t :
Na3TaOl 3OH,O = 4Na20.3Ta20,.25H20 1ONaOH; 12Na3Nb0, + 43H20 = 7Na20~6Nb20,~32H,0 22NaOH.

+ +

It must be taken into account that in addition to FeO and MnO, F e z 0 3and
a r e a l s o formed during fusion, since the oxygen present in the melt causes partial oxidation of the divalent i r o n and manganese to higher valency compounds. The f e r r i c i r o n is present in the melt a s f e r r i t e , NazO. FezQ. F e r r i t e decomposes during aqueous treatment according to the equation: Na,0.Fe20, H20 = 2NaOH F%OP
It'hI304

During fusion, most of the manganese may be oxidized to the manganate NazMn04, which is dissolved during aqueous processing, the manganate coloring the solution green. The impurity elements, silicon ( S i Q ) , tin (SnQ), tungsten (as wolframite), aluminum ( A l z 4 and aluminosilicates) form the soluble sodium s a l t s , NazSiQ, NazSnQ, NazW04, and NaAlQ during fusion with sodium hydroxide. The titanium, present a s an impurity in the f o r m of the m i n e r a l s rutile o r ilmenite, f o r m s sodium titanate N a z T i Q o r possibly NaZTi2O5 during the fusion, which is insoluble in water and which remains in the precipitate together with the niobium and tantalum. The solubility of sodium tantalate and niobate in water drops sharply in the presence of f r e e alkali or, rather, in the presence of an excess of sodium ions in the solutions. Thus, the solubility of sodium niobate (7Naz0. 6 N b z Q . 32Hz0) in a 1%solution of NaOH a t 90" is 1.1 g / l , while in water i t is 26 g / l , i. e . , 2 0 times higher. Hence, a certain level of f r e e alkali in the solution must be maintained in o r d e r to reduce to a minimum dissolution of niobium and tantalum during aqueous processing of the melt. In practice, aqueous treatment is c a r r i e d out at NaOH concentrations of a t l e a s t 5%. The fusion of tantalite o r columbite concentrates with sodium hydroxide is c a r r i e d out in i r o n crucibles which a r e placed in gas-fired o r electrical (shaft) furnaces. The approximate consumption of alkali is 300 kg p e r 100 kg of concentrate. The f i r s t operation involves fusion of the total amount of sodiGm hydroxide, ( o r sodium hydroxide containing 10% soda ash which is added in o r d e r to reduce the melting point and the viscosity of the melt) at about 400". The melt is then s t i r r e d continuously with s t e e l rods, and the concentrate, crushed to a particle s i z e of 0.1 mm, ( a s m a l l e r particle s i z e r e s u l t s in increased dust-entrainment) is gradually added. As new

110

portions of the concentrate a r e added, the temperature of the furnace is gradually increased to 8 0 0 ' . This is necessary in o r d e r to keep the m e l t liquid. The melt is held f o r 20 t o 30 minutes a t the maximum temperature and is then poured directly into water; safety m e a s u r e s must be taken, as contact of the alkali with the skin may cause s e r i o u s burns. The aqueous processing is c a r r i e d out at a solid :liquid r a t i o of 1 : 3 o r somewhat higher. A concentrated alkaline solution is formed in the f i r s t stage of the processing; this solution contains the bulk of the impurities - silicon, tin, tungsten, aluminum, sulfur, and phosphorus in the form of water-soluble sodium salts. The dark-brown precipitate contains sodium niobates and tantalates as well as i r o n and manganese oxides and hydroxides and sodium titanate. The precipitate is allowed to settle, the solution is separated by decantation, and the precipitate is washed s e v e r a l times with hot water containing about 5% NaOH, which is added in o r d e r to prevent the dissolution of niobium and tantalum. Digestion of the precipitbte with acid. The precipitate is decomposed by heating with hydrochloric acid. Sodium niobate and tantalate a r e converted into the hydrated oxides:

4Na20. 3Ta,05 + 8HC1+ ( x -4) H,O 7Na20.6Nb,0, 14HC1+ (x-7)H20

= 3Ta,05. xH,O

+ 8NaCl;
I4NaCI

= 6Nb20,.xH,0+

Iron and manganese oxides dissolve in the acid, yielding the respective chlorides. Chlorine is evolved a s a result of the reduction of manganese from higher valency s t a t e s to manganous compounds:
Mn,O,

+ 8HC1= 3MnCI, + C1, + 4H,O.

When the cake is treated with hydrochloric acid, sodium titanate is converted to titanic acid HzTiQ, which is partially dissolved. The bulk of the titanium remains in the precipitate together with niobium and tantalum oxides. Digestion with 20% hydrochloric acid is c a r r i e d out in rubber-lined iron vessels o r earthenware tanks heated by superheated steam. When the decomposition i s completed, the color of the dark-brown precipitate turns white, which i s the color of the niobium and tantalum oxides, while the color of the solution turns dark-yellow as a result of the presence of iron and manganese chlorides. The finely dispersed precipitate of hydrated pentoxides i s difficult to separate by filtration. The precipitate is rinsed with hot water, acidified with hydrochloric acid, in o r d e r to remove the iron and manganese salts, and the solution is separated by decantation. The rinsed precipitate is separated by filtration while s t i l l hot. The precipitate, consisting of the hydrated pentoxides, is dried at 100 to 120". The powder obtained after drying still contains 2 0 to 25% HzO. The approximate composition of the oxide mixture obtained by the above method i s : 96 to 99% (Ta, IW I)~ O ~ 0.1 , to 0.5% SnO2, up to 0.570 SiQ, 0.5 to 1.070 Ti&, up to 1% FeO+MnO, on the dry, water-free product. The oxide mixture is subsequently converted into pure compounds.

2 6 . PROCESSING OF TANTALITE-COLUMBITE B Y FUSION WITH POTASSIUM HYDROXIDE i 4 /


Fusion with potassium hydroxide is expedient if i t is desired t o produce a purer mixture of tantalum and niobium oxides. When the concentrate is fused with potassium hydroxide (in the same way as the fusion with sodium hydroxide described above). the minerals tantalite and columbite a r e decomposed, and water-soluble potassium tantalate and niobate, as well as iron and manganese oxides, are formed. T h e main impurities (silica, cassiterite, wolframite. and titanium minerals) react to yield the potassium salts of silicic, stannic, tungstic, and titanic acids. When t h e m e l t is leached wirb water. K2Si03, KzWO,, KzSn03, and KA102 are dissolved together with niobium and tantalum. Although K2Ti0, is a sparingly soluble salt, a fraction of the titanium dissolves in the presence of tantalum and especially niobium; this is attributed t o the formation of complex titanium-tantalum or titanium-niobium compounds. T h e alkaline solution is then treated with sodium chloride or sodium hydroxide, and the sparingly soluble sodium tantalate and niobate are precipitated. The silicon, tungsten, tin, and aluminum present as impurities remain in the solution. T h e precipitate, consisting of sodium tantalate and niobate is digested with acids, and a mixture of tantalum and niobium oxides with a very low impurity content is obtained. When t h e above method was used a concentrate containing 44%Ta20,, 16.1% Nb,Os, 6.5% TiO,, 9.9% SiO,. 6.1% FeO, 5.3% MnO, 4.4% SnO,, 3.7% A1,0, and 0.2% WO, yielded a mixture of tantalum and niobium pentoxides containing 99.22% Ta2O5+Nb2O5. 0.36% TiO,, 0 . 0 1 6 Fe203 and traces of SnO,. The compounds SiO, and MnO were not found in the mixture. A pentoxide mixture of the above composition c a n be used for the production of tantalum-niobium alloys by thermal reduction with carbon or by electrolysis (see below).

2 7.

DECOMPOSITION OF TANTALITE - COLUMBITE WITH HYDROFLUORIC ACID / I , 2/

The finely-divided tantalite (columbite) i s decomposed with hot concentrated hydrofluoric acid. The main decomposition reactions a r e :
Fe (RO,), 16HF + 2H,R%, Mn (R0J2 16HF 2H,RF,
-f

+ +

+ FeF, + 6H,O; + MnF, + 6H,O.

All elements (tin, titanium, silicon, tungsten, etc. ) originally p r e s e n t in the associated minerals a r e dissolved, together with niobium, tantalum, iron, and manganese. Thus, the decomposition of concentrates by hydrofluoric acid yields solutions which a r e contaminated with l a r g e amounts of impurity elements, which hamper the separation of pure tantalum and niobium compounds. This is the main reason for the limited use of the above method in industrial practice. However, recently developed methods for the separation of tantalum from niobium and their isolation from other elements have shown that the decomposition of tantalite (columbite) with hydrofluoric acid as a technological p r o c e s s has the advantage of being much s h o r t e r ( t h e r e a r e fewer stages) than the fusion of concentrates with alkalies. The digestion of concentrates with hydrofluoric acid is c a r r i e d out in s t e e l tanks covered with lead and lined with graphite blocks. The s t i r r e r s may be prepared from monel metal, an alloy of nickel containing 27 to 29% copper.

I12

The o r e concentrate, which is crushed to a particle s i z e of about 0.074mm, is decomposed with hot 70 to 72% HF. The acid consumption depends on the composition of the concentrate and the original mineral (the NbzO5: T a 2 a and F e O : MnO ratios), and on the processing conditions. The consumption of 70% H F is 1.2 t o 2.5 toniton concentrate. In s o m e industries, a mixture of hydrofluoric and sulfuric acids is used in o r d e r to i n c r e a s e the degree of decomposition of the concentrate. In most c a s e s the pulp is heated with superheated steam, which is injected through a rubber hose. The pulp may be heated by means of electric graphite heaters. P a r t of the hydrofluoric acid evaporates during the decomposition process. This r e q u i r e s that the apparatus be fitted with good exhausts and that the H F be absorbed in a s c r u b b e r irrigated with l i m e water. Twofold processing of the concentrate is occasionally required in o r d e r to provide quantitative decomposition. When the decomposition is completed, the pulp is diluted with water. The solutions a r e filtered using cotton or polyvinylchloride filter t i s s u e s in rubber-lined filterpresses. Until recently, tantalum and niobium were isolated from the hydrofluoric acid solutions in the form of the fluoride complexes, K2TaF7and KzNbOF5. HzO, by a method which is described below. At present tantalum and niobium a r e extracted from these solutions by selective extraction with organic solvents. The nature of the extraction p r o c e s s is discussed in Section 29. 28. PROCESSING OF TITANIUM-NIOBIUM CONCENTRATES 15,7,18j The c a s e of loparite concentrates will be examined a s an illustrative example of the processing technology of complex titanium-tantalum niobium raw m a t e r i a l s containing lanthanides. Loparite concentrates contain 8 to 9% ( N b + T a ) z Q (with a NbzO5:TazO5 ratio of about 1 4 : I ) , 32 to 35% Ti&, and 26 to Z8% (Ce, La.. )zq. Loparite may be decomposed with inorganic acids (concentrated sulfuric acid, hydrofluoric acid). However, the resulting solutions have a complex composition and the isolation of individual valuable components i s complicated. P a r t i c u l a r difficulties a r e encountered in the separation of titanium from niobium and tantalum. Chlorination is the simplest method for the extraction of all valuable components from loparite. This method involves treatment of the o r e concentrate with chlorine g a s at 750 to 850". Differences in the volatility of the chlorides formed p e r m i t separation of the main components of the concentrate. The tantalum, niobium, and titanium chlorides which have relatively low boiling points ( s e e Table 21) are volatilized and entrained with the gaseous phase and trapped in special condensers; the lanthanide, sodium, and calcium chlorides, which have high boiling points, remain in the furnace residue. A flow sheet of the chlorination of loparite concentrate is shown in Figure 54. The chlorination of the stable oxygen-containing compounds is usually c a r r i e d out in the presence of a reducing agent (coke), which p e r m i t s processing a t 750 to 850O.

113

TABLE 21 Melting and boiling points of certain chlorides Chloride Temperature, 'C bo i1ing NbCl, NbOCl, TaC15 FeCl,
204.7

Chloride

Temperature, 'C melting Sublimes boiling


180 58 1730 1750 2027 1465

- 67

(sublimes)
216.5 234 136 319

CeCIS CaClz
-

800 850 7 82 800

The main chlorination reactions a r e a s follows:


Nb,O, + 3C1, + 3C = 2NbOC13 3CO; Nb,O, + 5C1, + 5C = SNbCI, + 5CO Ta,O, + 5CIz + 5C = 2TaC1, 5CO; TiO, + 2C1, + 2C = TiCI, 2CO; Ce,O, + 3C1, f 3C = 2CeC13 3CO.

These reactions also yield COz and s m a l l e r amounts of phosgene (COCI2). The phosgene may chlorinate the oxides.
Crushed concentrate Mixing with coke and binder

1
Briquetting

Drying and coking of briquettes Chlorination

1
Residue: Ce, La Distillate chlorides, CaCI2. NaCl I Solid chlorides: NbOCls , NbC15, TaC1, , FeClS and others Liquid chlorides: TiCl, contaminated by SiCl,. and others
~

~.

1
To the extraction
lanthanides
Of

J
To t he production of
pure tantalum and niobium compounds

J
To purification

FIGURE 54.

Ptocessing o f loparite concentrates by t h e chlorination method.

114

The C O : C 0 2 ratio in the g a s e s depends on the temperature and the chlorination conditions. Carbon monoxide is the main component of the gaseous phase a t temperatures between 750 and 850". The chlorination is c a r r i e d out in shaft furnaces, and the charge is briquetted. Mixing, briquetting, and coking. Petroleum coke is used a s the reducing agent; the amount taken is determined experimentally, and fluctuates between 20 and 300/0. The crushed concentrate and coke (particle s i z e 0.177 to 0.15") are mixed in the required ratio in screw-type, paddle o r other type blenders. A binder is added to the mixture before briquetting; petroleum o r coal pitch, resins, and some other m a t e r i a l s a r e used as binders, Roller p r e s s e s a r e the most convenient briquetting equipment. They consist of two r o l l e r s rotating in opposite directions and fitted with belttype f o r m e r s with egg-shaped o r pillow-like cavities (Figure 55).

Chl ori

3
4

FIGURE 56. Elecrrical shafr furnace for the chlorination process. 1 - ~ a c k e t m a d e of steel sheets; 2-lining; 3-carbon electrodes; 4-feed device; 5 sream-gas mixture outlet; 6 -1uyeres (for chlorine supply); 7 -packing (carbon cylinders); 8-briquerring charge.

The charge on the rotating r o l l e r s is fed from the top and fills the f o r m e r cavities. P r e s s i n g i s done at about 150 to 300kg/cm2. When pitch o r resin a r e used a s binders the mixture is heated by means of steam fed to the steam jacket of the blender, to 80 to 1 2 0 " before pressing in o r d e r to soften the binder. The approximate s i z e of the pellet-shaped briquetts is 50X40X35mm. The briquettes a r e coked, i. e . , heated to 700 to 800" in the absence of air in o r d e r to i n c r e a s e their strength, to expel1 the volatile components and to i n c r e a s e their porosity (gas permeability).

I15

The coking is c a r r i e d out in batch or continuous r e t o r t furnaces. Chlorination. The chlorination of the briquettes is usually c a r r i e d out in electrical shaft furnaces (Figure 56) with an internal diameter of the shaft of about 4.5 to 6 m and a height of up to 8 m. The furnace is lined with dense chamotte o r Dinas bricks. The lower p a r t of the furnace contains two rows of carbon electrodes. There a r e t h r e e electrodes in each row, a t an angle of 120". The space between the electrodes is packed with carbon rods, which s e r v e a s an electrical r e s i s t o r which maintains the required temperature in the furnance. A voltage of 6 0 to 100 V is applied to the electodes, the voltage being reduced a s the packing heats up. The briquettes a r e charged into the furnace through a hermeticallysealed charging device. They f i l l a p a r t of the furnace space over the packing. By controlling the level of the raw m a t e r i a l in the furnace and the r a t e of supply of chlorine, the process may be c a r r i e d out at the expense of the heat of the chlorination and the supply of heat may be shut off o r considerably reduced. The chlorine coming from storage tanks, which a r e usually located in a special basement, p a s s e s through a chlorine-supply line and i s fed to the furnace through tuyeres placed below the upper row of electrodes. The variables which must be controlled during the operation of the furnace a r e : the r a t e of supply of chlorine, the g a s p r e s s u r e in the furnace, the temperature, and composition of the exhaust gases. The vapor-gas mixture containing volatile niobium, tantalum, titanium, iron, silicon and aluminum chlorides p a s s e s through a tube a t the top of the furnace into a s e r i e s of condensers (Figure 57). Solid niobium and tantalum chlorides (NbOC13, NbC15, TaC15) condense together with iron and aluminum chlorides in the f i r s t condenser which i s maintained at 1 6 0 to 180". The second condenser s e r v e s for the condensation of liquid titanium chloride. Spray condensers containing precooled l i q u i d TiC14 a r e most often used f o r this purpose. A fraction of the niobium, tantalum and iron chlorides, entrained with the gas s t r e a m , may condense in the second condenser, Moreover, SiC14, which is soluble in liquid titanium chloride, condenses together with TiC14 (for m o r e details s e e chapter on titanium). The g a s e s from the second condenser, which may occasionally contain some chlorine and HC1 vapors, a r e passed through a sanitary scrubber which is irrigated with l i m e water in o r d e r to bind the chlorine. The nonvolatile lanthanide, calcium and sodium chlorides formed in the chlorination a r e fused a t the p r o c e s s temperature (800 to 900") and flow through the p o r e s in the carbon packing onto the furnace bottom, from which they a r e periodically discharged. Thus, the chlorination of loparite yields t h r e e products: a melt containing lanthanide chlorides, a condensate consisting of tantalum and niobium chlorides contaminated by iron, and technical titanium t e t r a chloride. The purification of titanium tetrachloride will be described in the chapter I ' Titanium". However, we must note h e r e that tantalum and niobium chlorides a r e sparingly soluble in TiC14. At 18" the solubility of TaC15 and NbC15 in Tic14 is 0.4 and 0.4570by weight respectively, and i t i n c r e a s e s in the presence of aluminum chloride. The solubility of Nb0Cl3 and TiC14 does not exceed a few hundredths of one percent 1 2 5 1 . In o r d e r to obtain a mixture of tantalum and niobium oxides, the chloride condensate is subjected to hydrolytic decomposition in a hydro chloric acid solution:
1 I6

2 N W 1 3 ( X 3) HZO + N ~ , O ~ . X H ~ 6HC1; O 2TaC1, (x 5) HzO + TazO,. xHsO IOHCI.

+ + + +

+ +

Under certain conditions, which must be determined in each case, the hydrolysis can be c a r r i e d out so a s to ensure that the concentration of iron in the tantalum and niobium hydroxides is reduced to a negligible value. The hydrolysis also c a u s e s partial separation of titanium, which is p r e s e n t a s an impurity. The resulting mixture of hydroxides (up to 80% (Nb,Ta)z05, 1.5 to 370 Ti&, 3 to 7% Fez% and 0.5 to 2% S i Q ) is then forwarded to the purification stage and the separation of tantalum and niobium i s then c a r r i e d out.

To the sanitary scrubber

--

U
FIGURE SI. Installation for the chlorination of ritanium-tantalum niobium concentrates.
1 -furnace (chlorinator); 2-condenser for trapping niobium and tantalum chlorides; 3-container for solid chlorides; 4-spray condenser; 5-cooler for the titanium chloride which is returned for spraying; 6-condenser; 7 - ~ o n t a i n e r for titanium chloride; 8 - p u m p ; 3-thickener; 10-conveyer for the supply of slurries; I I -filter; 12-container for technical grade titanium chloride.

2 9 . SEPARATION OF TANTALUM AND NIOBIUM AND

PURIFICATION O F THEIR COMPOUNDS Because of the s i m i l a r properties of their chemical compounds, the separation of tantalum from niobium i s very difficult, Two separation methods a r e used in industry: 1) fractional crystallization of complex fluoride s a l t s (the Marignac process) ;

I I7

2) extraction with organic solvents.

O f m a j o r i n t e r e s t from the industrial point of view a r e also separation


methods based on differences in the volatilities of the chlorides (rectifica tion methods) and the selective reduction of niobium pentachloride to lower chlorides. Ion-exchange methods for the separation of tantalum from niobium have been developed, but t h e i r output is low and they may be used only for relatively s m a l l - s c a l e production of pure tantalum and niobium compounds. Removal of accompanying elements and contaminants from pure compounds usually takes place during the separation of niobium from tantalum.

Fractional crystallization of the complex fluorides

A method for the separation of tantalum and niobium, which until recently was the industrial method mainly employed for the separation of these elements, was proposed in 1865 by a Swiss chemist, Marignac. The method is based on differences in the solubility of potassium fluotantalate K2TaF7and potassium niobium oxyfluoride K2NbOF5. H20. The separation i s further facilitated by the different crystaIline s t r u c t u r e of these s a l t s . The separation of tantalum and niobium w a s greatly simplified a s a r e s u l t of studies c a r r i e d out in the Soviet Union, which permitted the processing mechanism to be scientifically understood.

Temperature, " C

HF concentration,%

FIGURE 58. Solubility of potassium niobium oxyfluoride and potassium fluotantalate in 1% HF.

FIGURE 59. Solubility of the complex tantalum and niobium fluorides in hydrofluoric acid at 25' as a function of the acid concentration.

Curves, based on the data of Meerson, Zverev, and Zubkova on the solubilities of potassium fluotantalate and potassium niobium oxyfluoride in 1% HF at 20 to 75" a r e presented in Figure 58. Within the temperature range mentioned, the solubility of the niobium s a l t is 1 0 to 1 2 times that of the tantalum s a l t / l o / .

I18

Curves based on the data of Savchenko and Tananaev 1111 on the solubilities of the complex tantalum and niobium fluorides a t 25" as a function of the H F concentration a r e presented i n Figure 59. The solubility isotherm of the complex niobium salts is composed of two sections. The f i r s t region of increasing solubility corresponds to the equilibrium between the solution and K2NbOF5.H20 in the solid, crystalline phase. The solubility of the s a l t i n c r e a s e s with increasing H F concentration, t o a maximum at 770HF. This i n c r e a s e in solubility is attributed to the increased stability of [NbF7I2- ions with increasing HF concentration. As a result, p a r t of the potassium niobium oxyfluoride is converted into potassium fluoniobate:

KZNbOF,

+ 2HF Z KzNbF7 + HZO

The solution becomes unsaturated with r e s p e c t to K2NbOF5 HzO and this c a u s e s dissolution of an additional amount of the crystals. When the HF concentration reaches 7 % , the solution becomes saturated with respect to both salts. This "transition point'' corresponds to the common solubility of K2NbOF5.H 2 0 and K2NbF7. The second region in the isotherm corresponds to the equilibrium be tween the solution and KzNbF7 crystals. The solubility of the s a l t f i r s t d e c r e a s e s ( t o 26 to 27% HF) and then increases. The d e c r e a s e in solubility is attributed to the suppression of the hydrolysis of K2NbF7with increasing H F concentration. The subsequent increase in solubility at higher H F concentrations is attributed to the formation of a new ionic species [NbF,]- according to the equation:
KsNbF7

+ HF

KNbF,

+ KHF,.

In contrast to potassium fluoniobate, potassium fluotantalate is not hydrolyzed in solutions containing a s m a l l amount of hydrofluoric acid. A s a result, the normal fluotantalate K2TaF7i s stable a t all H F concentra tions up to 45%. The solubility of KzTaF7 i n c r e a s e s with increasing HF concent r atibn A comparison between the solubility curves of the tantalum and niobium s a l t s (Figure 59) shows that the fractional crystallization of tantalum and niobium is conducted most conveniently a t H F concentrations between 1 and 7%. This i s the concentration range in which KzNbOF5. HzO is the stable f o r m of niobium, and the difference in the solubilities of the tantalum and niobium s a l t s is the greatest. The solubility of the complex niobium and especially tantalum s a l t is greatly affected by excess potassium fluoride added to the soiution. The solubility of KzTaF7drops by a factor of 10, 15 to 20, and 35 a s the excess concentration of KF reaches 1,2, and 5% respectively (Figure 60). In the case of the niobium salt, increasing the KF concentration to 570 causes a drop in the solubility of the s a l t by a factor of 2 to 3 I l l / . The solubility of tantalum s a l t s drops sharply in the presence of niobium s a l t s (Figure 61). The presence of the niobium s a l t leads to the creation of an excess of potassium and fluoride ions, which reduces the solubility of the tantalum salt. Thus, the solubility of K2TaF, drops to 0.05% at a niobium s a l t concentration of 1.5 to 1.770 (at 20") o r 3.5 to 3.7% (at 60").

119

KzNbOFSconcentrarion, 70 FIGURE 60. Variation of t h e solubility of K2TaF, with KF concentration. Upper curve -solution con
taining 1% HF; lower curve - 5 %
HF.
FIGURE 61. Variation of t h e solubility of
K2TaF7 with KZNbOF5concentration.

A flow sheet of the productipn of the complex tantalum and niobium s a l t s is shown in Figure 62. Moist tantalum and niobium oxides, o r tantalum and niobium oxides dried a t 100 to 120, a r e dissolved in concentrated (35 to 400/) hydrofluoric acid; an excess of H F is taken over the stoichiometric amount required for the reactions:
Ta,O, Nb,O,

+ lOHF = 2H, [NbOF,] + 3H,O.

+ 14HF = 2H, [TaF,] + 5H,O;

The dissolution is c a r r i e d out in rubber-lined or lead-lined v e s s e l s a t 70 to 80". The liquid is allowed to settle and the c l e a r solution is separated by decantation and filtered through cotton or polyvinylchloride filters. The volume of the solution is s o adjusted that after the subsequent addition of the potassium s a l t the concentration of KzNbOF5 reaches 3 to 6 % . This is below the solubility limit for that salt. Simultaneously, the concentration of f r e e H F is reduced to 1 to 270, which facilitates subsequent operations. A potassium salt (KCI and occasionally KzC03) is added to the hot dilute solution, in the amount required for the formation of KzNbOF5 and KzTaF7when complex fluorides a r e formed in the reactions:

H, [TaF,I H, [NbOFJ

+ 2KC1+ Ka [TaF,] + 2HC1;


+ 2KCI = K, [NbOF,] + 2HCl.

120

3 0 4 0 % HF

Dissolution

Settling, decantation, filtration

. 1
Solution of H2TaFl and HzNbOF5
7 -

I 5.
Insoluble residue To waste Dilution of the solution and heating

KCl Precipitate KZTaFl

Precipitation of the tantalum salt

Filtration and washing

1
Mother solution containing K,NbOF, and impurities ( T a . Ti, Fe)

Recrystallization from 1-270 HF

- 1

Evaporation Crystallization

Mother solution Drying a t 120-150'

I
Mother solution

I
KZNbOF5. HZO crystals
I

1
Mother solution is discarded after 6-10 recrysrallizations

. 1

Recrysta 11i za cion from 1-270 HF

3
Mother solution

5.

Drying

J.

FIGURE 62. Flow sheet of the separation of tantalum and niobium by fractional crystallization of their complex fluoride salts.

The hydrochloric acid formed in the reaction has little effect on the ratio of the solubilities of potassium fluotantalate and potassium niobium oxyfluoride. Potassium fluoride i s not commonly used in practice since it is expensive and difficult to handle (very hygroscopic).

121

The tantalum s a l t s e p a r a t e s out of solution in the form of needle-shaped crystals, which are separated by filtration after allowing the solution to cool. Silicon in the form of KzSiF, is precipitated quantitatively with the tantalum salt. The precipitate a l s o contains up to 50% of the titanium and a s m a l l fraction of the niobium (5 to 6 % of the amount initially present in the solution) in the f o r m of complex fluoride s a l t s . An additional purification stage involves the recrystallization of KzTaF, f r o m 1 to 2 7 ' 0 HF. The potassium fluotantalate produced contains 0.1 to 0.370 niobium. The silicon content reaches 0.370, that of iron 0 . 2 7 ' 0 , and that of titanium 0.01 to 0.02%. The mother solution is evaporated. The compound KzNbOF5' HzO crystallizes after the solution is cooled, and is then purified by recrystallization. Titanium, which is always p r e s e n t a s an impurity in the mixture of tantalum and niobium oxides, f o r m s the s a l t KzTiF6.HzO, which is iso morphous with the niobium s a l t KzNbOF5- HzO. At 20" the solubility of the titanium s a l t is about 1 2 g / l , which is lower than the solubility of the niobium s a l t (about lOOg/I). Since the s a l t s a r e isomorphous, the titanium s a l t crystallizes out together with the niobium salt. A s a result, the niobium s a l t always contains titanium, which is removed only partially in the recrystallization of the niobium salt. The niobium s a l t produced by crystallization of potassium niobium oxyfluoride a l s o contains some tantalum (0.5-2'%). The method yields a pure tantalum salt, but it is difficult to produce a niobium s a l t of sufficient purity. The niobium s a l t is usually contaminated with titanium, silicon, and tantalum. This i s the main disadvantage of the Marignac p r o c e s s .

The extraction method The fundamentals of extraction. Extraction p r o c e s s e s have lately found extensive use in the hydrometallurgy of the nonferrous and r a r e metals. These p r o c e s s e s a r e based on the selective extraction of the compounds of a given element from an aqueous solution into an organic solvent which i s immiscible with water. When the aqueous solution and the organic solvent a r e brought into contact, the substances p r e s e n t in the aqueous phase a r e distributed between the two phases. An equilibrium s t a t e is established after a certain time. The equilibrium distribution of a substance is characterized by the d i s t r i b u t i o n c o e f f i c i e n t , i . e . , b y the ratio of the equilibrium con centrations in the organic and aqueous phases:
C Y

c org .
caq

The distribution coefficient depends on the properties of the solvent, the concentration of the extracted substance in the aqueous solution, the presence of other substances, the acidity of the solution, and the temperature. Undissociated molecules a r e usually extracted from the aqueous phase into the organic solvent. Hence, weak electrolytes a r e extracted to a higher degree. In o r d e r to depress the dissociation of s a l t s in the aqueous solution, a l a r g e excess of an acid or a s a l t with the s a m e anion as that of the extracted s a l t i s added to the aqueous solution.

122

'I

When a solution contains simultaneously the compounds of two elements, the distribution of each element between the aqueous and organic phase will correspond to the values of their distribution coefficients. The r a t i o of the distribution coefficients of two elements is known as the s e p a r a t i o n c o e f f i c i e n t : /3 = % . This coefficient determines the
5%

effectiveness of the separation of a given p a i r of elements by extractive methods. It shows the change in the ratio of the concentrations of elements undergoing separation in the e x t r a c t (the organic solvent) a s compared with the initial ratio in the aqueous solution. Various organic solvents - e s t e r s , ketones, alcohols, amines, hydro carbons - are used a s extractants. The extractant must possess a high extractive capacity with respect to the substance to be extracted, and must be easily regenerated. Its solubility in water must be small to reduce losses. In o r d e r to obtain good phase separation after.the mixing, i t is necessary that the density of the extractant be s m a l l e r than that of water, and that the viscosity of the extractant be low. Moreover, for safety reasons, the extractant should be of low volatility and flammability, and should be non-toxic. Its stability in inorganic acids is also of importance. The cost of the extractant must also be considered. Extraction processes a r e mainly c a r r i e d out in two types of equipment: columns and m i x e r - s e t t l e r s .
2
Starting aqueous

Aqueous ohase

3
FIGURE 63. Diagram of a packed extraction column. 1-packed column; 2-input of the starting solution; 3-input of the organic solvent; 4 -siphon for the discharge of t he aqueous solution: 5-extract efflux.

FIGURE 64. Diagram of a pulsating tray column for extraction processes.


1-pulse-generator; 2 -perforated-tray column; 3 -phase-separation section; 4-organic phase level; 5-interphase area: 6 -perforated may.

123

1 l Il l l l l Il1

In extraction columns there is a countercurrent motion of the aqueous solution and the organic solvent. In o r d e r to i n c r e a s e the phase-contact a r e a , the column is packed (e. g . , with rings) or else contains a number of perforated t r a y s ( t r a y columns). A diagram of a packed extraction column is shown in Figure 63. Since the organic phase is lighter, i t is usually fed in a t the bottom of the column, while the aqueous phase is fed in a t the top. The upper p a r t of the column is not packed and s e r v e s for phase separation and f o r withdrawal of the organic extract. The aqueous phase is discharged from the bottom of the column through a siphon, which s e r v e s to maintain a constant level of the aqueous phase in the column. The countercurrent motion of the liquids in packed or t r a y columns i s a r e s u l t of gravity forces alone. Thus, such columns may be used only in c a s e s in which t h e r e is a sufficiently l a r g e difference between the densities of the aqueous and organic phases. The effectiveness of packed or perforated-tray columns is considerably increased if the liquid c u r r e n t s in the column a r e pulsated by means of piston pumps or other devices. In perforated-tray (hole diameter = 2 - 3 mm) pulsating columns (Figure 64) an upward pulse f o r c e s the organic phase through the perfora tions and makes i t r i s e towards the upper tray. A downward p u k e f o r c e s the aqueous phase through the perforations in a downward direction. Thus, the pulsation provides a countercurrent motion of the phases along the column and their dispersion. The pulsations in packed and t r a y columns increase their output approximately by a factor of three.

3
.- m
c 0 O .v u

2 0 2
Starting solution
O C Q

z-!- , t
-I n
I
1

outlet Aqueous phase Organic phase

-Mixer 4

, ;
I
I

!-Settler

'2 v

&.io

;g

-#
1

..
1-mixing chamber; 8-turbine; 3- hollow shaft; 4-partition; 5-settling chamber. current extraction in a cascade of mixer-settler cells (view from the top).

Mixer-settlers a r e extensively used in extraction p r o c e s s e s . They consist of a chamber for the mixing of phases with the aid of s t i r r e r s ( o r mixing pumps) and a settling chamber.
124

A cell of a well-known m i x e r - s e t t l e r extractor with a mixing pump is shown in Figure 6 5 . The cell consists of mixing and a settling chambers separated by a partition. The cells a r e connected together forming a cascade of m i x e r - s e t t l e r s (Figure 6 6 ) . The mixing chamber is fitted with a s m a l l turbine rotating on a shaft, which e n s u r e s a rapid mixing of the aqueous and organic phases, the t r a n s f e r of phases between adjacent cells, and level control of the liquid phases. The mixer is separated into two sections by a horizontal partition. The hollow shaft of the turbine p a s s e s through an opening in the horizontal partition. The turbine pumps the aqueous solution (i. e . , the heavy phase) through the hollow shaft from the lower to the upper section, dispersing the liquid through openings between the blades, where it is mixed with the organic solvent (the light phase) which is fed to the upper section of the mixing chamber. A constant level between the aqueous (heavy) phase and the phase mixture in the lower section is maintained automatically. This is explained by the fact that the liquid i s aspired by the turbine at a certain level. Thus, when the output of the turbine exceeds the r a t e of flowof liquid, the phase mixture is pumped from the upper to the lower chamber until a regular flow of the liquid is established. The level of the phase mixture in the upper chamber (i. e . , the level of the interface between the phase mixture and the a i r ) gradually d e c r e a s e s from cell to cell in the cascade, in the streaming direction of the light phase. The phase mixture flows by gravity through the partition and into the settling chamber in which the phases separate. The light phase p a s s e s through an opening in the upper p a r t of the settling chamber into the mixing chamber of the preceding cell while the heavy phase p a s s e s through an opening n e a r the bottom of the chamber into the lower section of the mixing chamber of the following cell. Thus, the flow of the heavy phase in one direction of the cascade i s forced by the turbine, while the flow of the light phase in the opposite direction is caused by the gradual d e c r e a s e in the liquid level in the upper section of the m i x e r s . In continuous countercurrenf extraction the starting aqueous solution i s fed to the middle of a system of e x t r a c t o r s of the m i x e r - s e t t l e r type, a s shown in Figure 6 6 . In this c a s e the system has extraction and washing sections. The organic solution entering the wash sections comes in contact with the aqueous washing solution which extracts the impurities previously removed by the solvent in the extraction section. The extractive separation of tantalum and niobium, with simultaneous removal of other (impurity) elements i s mostly c a r r i e d out from solutions of tantalum and niobium fluorides, containing hydrofluoric and sulfuric acids, The solutions a r e obtained a s a r e s u l t of the decomposition of o r e concentrates by hydrofluoric acid, a s described above, o r by dissolution of the oxides in HF. The best extractants of those tried out a r e methyl isobutyl ketone and tributyl phosphate (Table 2 2 ) . Tributyl phosphate (TBP) is superior to methyl isobutyl ketone because of i t s relatively high boiling and flash points. It has the disadvantages Of high density and viscosity. For this reason T B P is often diluted with some light i n e r t diluent, e. g. , kerosene o r xylene [C,H4(CH,)2 1. Tributyl

I25

phosphate f o r m s stoichiometric complexes with the extracted compounds. Thus, the complex T B P HF is formed with hydrofluoric acid. Depending on the hydrofluoric acid concentration, tantalum and niobium are present in solution i n the f o r m of various complex acids: HzTaF7, HzNbF,, HTaF6, HNbF6, and HzNbOF5.

Property
I

Methyl isobutyl ketone


7 I

Boiling point, " C

.......

115.9
0.80 27 1.7 -2.2

177-118 (decomposes at
289 'C) 0.97

Density Viscosity, centipoises Flash point, 'C Solubility in water. $70 ...

............... ..

........

0.546

3.45 145
0.6

These acids are extracted by T B P in the form of trisolvates: HzRF7. STBP, HRFs. 3TBP (where R i s Ta o r ,Nb) and H2NbOF5.3TBP /19/. Methyl isobutyl ketone and the complex fluotantalic and fluoniobic acids f o r m complexes of the oxonium s a l t type, in which the hydrogen of the acid is bound to the oxygen of the ketone:

[>-0-H]

TaF;;

Z-O-H]
2

TaF:-.

Such oxonium s a l t s a r e readily soluble in excess organic solvent. In the extraction of tantalum and niobium from solutions of their fluoride complexes by TBP o r methyl isobutyl ketone, the distribution coefficients a r e strongly affected by the hydrofluoric acid concentration. This is especially true in the case of niobium. As is evident from Figure 67, niobium is virtually not extracted from solutions containing less than 4 moles HF p e r l i t e r , while the tantalum is readily extracted. This can probably be attributed to the fact that in dilute HF solutions the niobium is present as H2NbOF5which is extracted to a l e s s e r extent than is HzNbF7. Tantalum and niobium may be extracted together from concentrated HF solutions, and at the same time these two metals may be separated f r o m many accompanying elements (e. g . , Fe, Mn, Ti, Sn) whose distribution coefficients a r e very low, especially in the presence of sulfuric acid (Table 23). The extractive separation of tantalum and niobium usually involves three stages: 1) joint extraction of tantalum and niobium, as a means of separating these metals from the accompanying elements (Fe, Mn, Ti, Sn, Si, and others) ;
126

2) re-extraction of niobium f r o m the organic solvent by water; 3) re-extraction of tantalum f r o m the solvent by water of by aqueous solutions of s a l t s such a s ammonium fluoride.

C I 1 0 12 1 4 HF concentration. N

1 6

FIGURE 61. Degree of extraction by methyl isobutyl ketone of tantalum and niobium from solutions with different HF concentrations.

TABLE 23 Distribution coefficients of the ions of some elements between TBP and an aqueous solution containing 10 moles HF and 12 moles H,SO, per liter
I

Ion and its valence

Concentration in the starting solution, g t1

a=

Corg caq

2 2 10 2
10

183 119 0.001 0.017 0.02 0.05

A simplified flow sheet of the production of pure tantalum and niobium oxides from columbite concentrate, based on the extraction p r o c e s s used in a factory in the USA, is shown in Figure 68. The o r e concentrate, crushed to a particle s i z e of -0.074mm, is decomposed by 70% H F with heating and stirring. The pulp is diluted to an H F content of 15 moles p e r l i t e r and is filtered in a rubber-lined f i l t e r p r e s s . The c l e a r solution is fed to the upper p a r t of a pulsating column with perforated trays. The column and all tube joints a r e made of polyethylene. Methyl isobutyl ketone is fed to the lower p a r t of the column by means of a diaphragm pump which c r e a t e s a pulsating s t r e a m of the solvent. In the column, the tantalum and the niobium a r e extracted by the solvent while the impurities remain in the aqueous phase. Dilute sulfuric acid is fed to the upper p a r t of the column together with the starting solution, in o r d e r to reduce the degree of extraction of the impurities.

127

Methyl isobutyl ketone

AqJeous
solution

'

I4

1 5

I,?

FIGURE 68. Flow sheet of the processing of columbite concentrate, including the ex tractive separation of tantalum and niobium: 1-teactor for the decomposition of the concentrate; 2-tank for the dilution of the pulp; 3-filter press; 4 -container for the filtrate; 5-column for the joint extraction of tantalum and niobium; 6-column for the reextraction of niobium; I-column for reextraction o f tanralum; 8 and 9-apparatus for the steam distillation of rhe solvent; 10 -collector for the tantalum solution; 11-vessel for the precipitation of T a 2 0 s ; 12 -filters; 13 -ignition furnaces; 14-collector for the niobium solution; 15 vessel for the precipitation of Nb,05.

TABLE 24 Composition of the oxides Element Concentration, 70


Nb205

ta205
< 0.05

Zr Ta Nh F Fe Ti Si

< 0.05 < 0.03

< 0.03

0.06
< 0.03

w
Ni cu A1 Mg, Co, Zn, Cr. Mn, Sn, V , Mo Cd B

< 0.015 0.01 < 0.01 0.005 < 0.004 0.002 < 0.002

0.06 < 0.01 < 0.015 0.01 < 0.01 0.005 < 0.004 0.002 < 0.002

< 0.0005 < 0.0001

< 0.0005 < 0.0001

The organic solvent containing the niobium and the tantalum is fed to the middle of a pulsating column serving for the reextraction of niobium with water which is fed into the upper section of the pulsating column.

128

__

._. .

.-.-.

-_... ....

. I

.. .

I,

. I

1 1 . 1

I,

I,

I . .

Most of the H F from the solvent is extracted into the aqueous phase together with the niobium. At the s a m e time, only a small fraction of the tantalum passes into the aqueous phase from which i t is separated in the lower p a r t of the column where the aqueous solution comes into contact with the upward s t r e a m of p u r e methyl isobutyl ketone. The aqueous solution coming from the column contains virtually only niobium. In the third column, the tantalum is extracted from the solvent by water. After the reextraction of niobium and tantalum the solvent is purified by steam distillation and returned to the extraction cycle. The pure tantalum and niobium hydroxides a r e precipitated from the aqueous solutions by the addition of ammonia water. The precipitates a r e separated by filtration in stainless-steel vacuum f i l t e r s (of the drum type) and a r e ignited to the oxides in drum furnaces. The concentration of impurities in the pentoxides produced by the above method is shown in Table 24.

Separation of tantalum and niobium by the chloride r e ctif i cation pro ce s s *;


Rectification may be successfully employed in the separation of the mixture of tantalum and niobium pentachlorides. as the boiling points of the chlorides differ by 14.3' (the boiling points of TaC15 and NbC1, are 234.0 and 248.3' respectively). T h e use of the rectification process for the separation of tantalum and niobium is expedient when the ore concentrates are processed by the chlorination method, yielding a condensare of tanralum and niobium chlorides, e. g., in the processing of loparite concentrates. The chlorination of the concentrate usually yields a condensate in which a large fraction of the niobium is present as the oxychloride NbOCI,. Thus. additional chlorination is required in order to convert the oxychloride to the chloride. T h e Soviet scientific literature contains descriptions of laboratory experiments on the rectification of tantalum and niobium chlorides in sieve-tray glass columns with a diameter of 32 t o 34" and 25 trays, as well as larger-scale experiments in stainless-steel columns /15/. The starting mixture of chlorides contained 80.2% Nb. 12.3% T a , 0.3% T i , 7.2% Fe. and 0.22% W; after rectification the yield of niobium chloride was about 6 0 % . and it contained 0.008% Ta (on the Nb), 0.005% T i , 0.00270 W . and < 0.001% Fe. About 80% T a was obtained in the form of a concentrate containing 8070Ta (on the total amount of T a + N b ) . This concentrate, as well as other intermediate fractions. m a y be subjected to a second rectification in order to obtain the pure compounds. The rectification method is characterized by a high output and a high degree of separation. It may b e carried out in continuous operation columns. T h e pure tantalum and niobium chlorides produced may be used as the starting material for the production of tantalum and niobium (see Section 35).

Separation of tantalum and niobium by selective reduction of the chlorides


As compared with tantalum pentachloride. niobium pentachloride is more readily reduced by hydrogen (or by metals such as aluminum) t o the lower chlorides. This is used as the basis of a process for the separation of tantalum and niobium. NbC1, is reduced by hydrogen at 450 -550" t o the nonvolatile trichloride: NbCl, H2+ NbCI, ZHCI.

TaC15 is not reduced under those conditions.

_______-----__ * The nature of the rectification process will b e discussed in Chapter IV.
129

In one variation of the method. tantalum and niobium are separated by volatilizing the NbCI, + TaCI, mixture a t 180-200 and passing the gaseous chlorides together with hydrogen through a long tube heated to 500-550. T h e NbCI, formed condenses on the walls of the tube, while the TaCI, remains in the gaseous phase and is collected in the condenser. T h e yield of pure NbCl3 from the pentachloride has been reported to be about 1 0 % / 1 / . T h e difficulty of effecting the reaction as a continuous process and the necessity for repeating the process on the condensate in order to obtain pure tantalum pentachloride are among the disadvantages of the process. Metallic niobium can be produced from rhe trichloride by reducrion wirh hydrogen.

Separation of tantalum and niobium with the aid of ion exchange resins* 1 1 2 1 Investigations c a r r i e d out in the USSR showed that the ion-exchange separation of tantalum and niobium may be successfully effected from hydrofluoric acid s o h t i o n s . It is recommended that the anion exchange r e s i n EDE-lOP, in which the active groups a r e (=NH) and (ZN), be used for the separation. The dynamic exchange capacity of this r e s i n (in the C1-form) with r e s p e c t to tantalum, niobium, and titanium is 10, 7, and 5 mg-equiv/g respectively. Titanium i s usually present as impurity in the solutions.
lrNHC1+OL75=M HF 3-W HCL J.w HC,! ~75-u HF

.l-

- P * - 1

J2

v ,

.* 2
Y

C U

u
Volume of solution passed through

FIGURE 69. Elution curves of niobium. titanium. and tantalum sorbed on anion-exchange resin.

The affinity for the r e s i n d e c r e a s e s in the o r d e r TaF:-> TiFi- > N b F : : Hence, in the elution of the niobium, tantalum and titanium ions sorbed in the upper layer of the r e s i n bed, niobium is the f i r s t to be eluted, followed by titanium and tantaIum. The separation effect may be enhanced by using acid solutions of various concentrations f o r the elution of niobium, titanium, and tantalum. F o r instance, niobium is eluted with a mixture of 3 5 g / l HC1 + 1 g / l HF, titanium with HCI at a concentration of 84 g/I, and tantalum with a mixture of 1 8 0 g/l HC1 + 1Og/l HF. The f i r s t fraction of the eluate contains up to 8070 niobium (titanium- and tantalum-free), the

-_ _ _ _ _ _ _ _ - - _ _ - _ -

T h e fundamentals of ion-exchange chromatography method of separation of elements with similar properties will be discussed in Chapter VI.

I30

second fraction contains the r e s t of the niobium and all the titanium, and the third fraction contains pure tantalum (see Figure 69).

30. METHODS FOR THE PRODUCTION O F METALLIC TANTALUM AND NIOBIUM

Tantalum and niobium a r e produced by reduction of very pure compounds: oxides, complex fluoride salts, chlorides. The industrial methods for the reduction of these compounds (which are characterized by a high stability) may be divided into three groups: 1) reduction of halides by active metals - sodium, magnesium, ca1cium ; 2) reduction of oxides by carbon (thermal reduction with carbon); 3) electrolysis of molten salts. Because of the high melting points of tantalum ( - 3000") and niobium ( - 2500"), these metals a r e obtained a s powders in the reduction processes. The subsequent conversion of the powders into solid, ductile metal is complicated by the fact that tantalum and niobium actively absorb gases (hydrogen, nitrogen, oxygen) which render the metals brittle. For this reason, the sintering or melting of the powders must be c a r r i e d out under high vacuum, which complicates the design of the production equipment. Of the production methods listed above, the methods based on reduction with metals a r e of g r e a t importance. Such methods a r e known as' "metallothermic". They a r e extensively used in the production of other rare metals (titanium, zirconium, beryllium). F o r this reason, we shall f i r s t discuss their theoretical principles. FUNDAMENTAL PRINCIPLES OF METALLOTHERMY

31.

Metallothermic production of metals i s based on the reduction of chemical compounds (oxides, salts) by metals whose affinity for the non metal in the compound is higher than that of the metal to be produced. The f i r s t studies on the displacement of metals from their compounds by other metals were c a r r i e d out by the Russian chemist Reketov in 1850 to 1860. These studies laid the foundations for the most common thermal reduction process - aluminothermy, which plays an important p a r t in the modern production of ferroalloys, and in the production of a number of metals (barium, strontium, calcium) and nonferrous and r a r e metal alloys of technological importance. Subsequently, silicothermy as well as magnesiothermy (mainly the production of rare metals) and thermal reduction with calcium and sodium also found wide use. The reaction of thermal reduction by metal may be generally expressed as follows: MeX + Me' 2 Me'X + Me Q ,

where MeX is the compound to be reduced (oxide, halide) ;

131

Me'is the reducing agent; Q is the heat of reaction. The r a t e and the extent of the reduction depend on the affinity of the metal acting a s reducing agent and the metal in the compound to be reduced for oxygen, chlorine, fluorine, and other elements.
-158

coo

-14d

-130

- 120

2 -I00 " \9

.
Y

g
m

-??0

.-

cm

-90

E
L

-80

2 -70

Y
Y

L ; I

-60

-50

-40

-30
-20

FIGURE 70.

The affinity of metals for oxygen.

In many cases, the affinity of a metal for a given element may be evaluated f r o m the heat of formation of a chemical compound from the metal and the element. The affinity is characterized m o r e accurately a s the decrease in the f r e e energy AF accompanying the formation of the compound from i t s elements.

I32

Temperature, ' K a
FIGURE 11. Affinity of metals for chlorine and fluorine:

a-chlorine:

b -fluorine.

The dependence of the f r e e energy of formation of oxides and halides on the temperature is shown in F i g u r e s 70 and 71. To facilitate comparison, the values of AF a r e related to one g-atom of oxygen, chlorine, and fluorine. A comparison of those values shows that the most active reducing agent f o r the r a r e metal oxides is calcium, followed by magnesium and aluminum; for the chlorides - sodium, calcium, magnesium; and f o r the fluorides - calcium, sodium, magnesium. If one of the participants in the t h e r m a l reduction reaction displays a high vapor p r e s s u r e a t the temperature of the reaction, the changes in p r e s s u r e may have an appreciable effect on the course of the reaction and may even r e v e r s e its direction. Thus f o r instance, the standard affinity of calcium for oxygen (144.3 kcal) is much higher than that of aluminum (123.8 kcal) and aluminum oxide should, therefore be reduced by calcium. However, the reduction reaction is reversible: A1,0,

+ 3Ca

3Ca0 + 2A1.

Depending on the external p r e s s u r e , it may proceed in the direction of the formation of aluminum or the reduction of calcium oxide by the aluminum. The boiling points of calcium and aluminum a r e 1480 and 2330" respectively. When the reaction is c a r r i e d out under reduced p r e s s u r e , calcium (which has a lower boiling point) is expelled from the reaction zone. Under such conditions, calcium oxide is reduced by aluminum (this p r o c e s s is used in the production of metallic calcium). In a sealed reactor, in which vapor p r e s s u r e of calcium may build up, the reaction would r e s u l t in the forma tion of elementary aluminum. The heat evolved in the t h e r m a l reduction by metals is often sufficient to support a spontaneous p r o c e s s a t a high r a t e (the reaction must first be initiated somewhere in the mixture). In other c a s e s the heat evolved is not sufficient to support the p r o c e s s and unless external heating is supplied, the reaction c e a s e s before i t can spread over the entire mixture. In o r d e r to find out if the reduction i s spontaneous o r must be supported by external heating, we must know the specific heat effect q, i. e . , the f the charge. amount of heat required p e r g o The specific heat effect of the reaction M e X + Me' = M e + Me'X + Q i s :

where MMcx i s the molecular weight of the compound and AMe is the atomic weight of the reducing agent. The maximum heat effect of the reaction is obtained a t a stoichiometric ratio of the components in the charge. An excess or deficiency of the reducing agent diminishes the specific heat effect since a p a r t of the m a t e r i a l absorbs heat (heating-up, melting) without taking p a r t in the reaction. Nevertheless, in most c a s e s an excess of the reducing agent i s required in o r d e r to obtain a higher degree of completion of the reduction. A heating additive (activator) is added to the charge if the specific heat effect of the reaction is not sufficient to support the reaction. The activator r e a c t s with the reducing agent with the evolution of a l a r g e amount of heat, which i n c r e a s e s the specific heat effect.

1455

134

In other c a s e s the heat of reaction is s o high that the reaction is extremely vigorous and partial ejection of the charge (caused by the rapid expansion of g a s e s entrapped in the p o r e s of the charge) a s well a s evaporation of the reducing agent and the compounds participating in the reaction take place. In such c a s e s the reaction i s slowed down by the addition to the charge of fluxes which absorb a p a r t of the heat, (this being consumed in the heating and melting of the flux). Fluxes a r e also added when a lower-melting, lower-viscosity s l a g is required. The reducing agents used in thermal reduction by m e t a l s must satisfy the following requirements : 1) the reducing agent (metal) must provide for the highest possible degree of completion of the reduction a t the lowest possible supply of external heat; 2) the slags formed and the excess reducing agent must be readily separated from the metallic product (by rinsing, vacuum distillation, removal of slag); 3) the reducing agent must be of high purity, in o r d e r to prevent contamination of the metal product; 4) the reducing agent must have a low solubility in the metal product and must not form chemical compounds with i t ; 5) the reducing agent must be readily available and reasonably cheap. The melting and boiling points of the most common metal reducing agents and the products of thermal reduction reactions a r e listed in Table 25.
TABLE 25 Melting and boiling points of metallic reducing agents and their compounds

-1Element Na Ca

- -

-_

Mg
A1
.

97.7 851 650 659


.

1480

1126 2330

2510 2800 2045

182 714 192.4

2021 1418 180

1418 1262 1270

1704 2501 2221 1272

3 2 . PRODUCTION O F TANTALUM AND NIOBIUM POWDERS B Y THERMAL REDUCTION WITH SODIUM

The thermal reduction of fluoride complexes of tantalum and niobium by sodium was the f i r s t method f o r the industrial production of these metals, and is still in u s e today. The fluoride complexes were selected as the starting compounds f o r this p r o c e s s since they a r e the end products in the separation of tantalum and niobium by the Marignac process. Tantalum will be used in describing this p r o c e s s ; until recently thermal reduction with sodium was the main p r o c e s s by which tantalum was produced f 2 9f Because of their affinity for fluorine; sodium, calcium, and magnesium a r e suitable reducing agents for tantalum and niobium fluorides ( s e e Figure 71, b). However, sodium is the only one used since the NaF formed is soluble in water and may be separated from tantalum powder

by washing. Calcium and magnesium fluorides a r e sparingly soluble in water and acids. The reduction reaction
K2TaF,

+ 5Na

-+ Ta

+ 5NaF + 2KF

is accompanied by the evolution of a l a r g e amount of heat ( - 713 kcal p e r kg s a l t and reducing agent), which is sufficient to support a self-sustained p r o c e s s ; the charge is locally heated to 450 to 500" and the reaction s p r e a d s rapdily through the bulk of the charge. The reduction is c a r r i e d out in s t e e l crucibles in which KzTaF7and Na a r e charged in alternate layers. The charge is covered with a layer of NaCl which f o r m s a lowmelting mixture with KF and NaF. The molten s a l t l a y e r protects tantalum p a r t i c l e s against oxidation. The s a l t s and the crucible must be thoroughly dried before the charging, since water vapor r e a c t s with sodium, yielding hydrogen, which may form an explosive mixture with a i r in the p o r e s of the charge. The ejection of the charge in this p r o c e s s is especially dangerous since i t contains droplet of molten sodium. As a safeguard, the crucible is placed in a metallic casing. The sodium b a r s (which a r e usually kept under kerosene) a r e cut into s m a l l pieces (special s c i s s o r s a r e used), washed with benzene and held in the a i r until the benzene evaporates. The amount of sodium added to the charge i s 125 - 1.50% of the stoichiometric. To initiate the reaction the bottom of the crucible is heated with a blowtorch until the appearance of a red spot. The reaction spreads rapidly throughout the charge and is completed within 1 to 2 minutes. During the reaction the temperature of the charge reaches about 800 to 900". The metallic tantalum obtained by this p r o c e s s is disseminated a s a fine powder within the fluoride-chloride s l a g containing the excess sodium metal. The molten salt is allowed to solidify and is knocked out of the crucible. The upper slag layer, which contains no metallic tantalum, is separated, the bulk of the slag is crushed into s m a l l particles in a jaw c r u s h e r and the p a r t i c l e s a r e charged in s m a l l portions, with constant stirring, into the water-filled reactor. This r e s u l t s in the "quenching" of the metallic sodium, which r e a c t s with water to form the hydroxide with the evolution of hydrogen. The bulk of the s a l t s i s washed out f i r s t with cold and then with hot water and the tantalum powder is washed in porcelain tanks with dilute hydro chloric acid. The acid is used in o r d e r to obtain a m o r e complete removal of the alkali s a l t s and in order to dissolve the iron. To reduce the concentra tion of oxides, the washing with hydrochloric acid is sometimes followed by additional wash with dilute cold hydrofluoric acid. The powder is then rinsed with distilled water, isolated by filtration, and dried a t 110 to 120". The resulting powders a r e of small particle s i z e : 8570 of the grains a r e s m a l l e r than 1p and the r e s t have a particle s i z e of 1 to 2 p . The powders contain 2 to 3 % oxygen, 0.1 to 0.15% (Na+K), and about 0.1570 hydrogen. The extraction of tantalum into the powder is 92 to 947'0. The l o s s e s a r e caused mainly by entrainment of the fine fraction of the powder with the wash water. The s a m e method may be used for the production of niobium metal from potassium fluoniobate / 2 8 / . To p r e p a r e KzNbF7 from potassium niobium oxyfluoride K2NbOFS.HzO (which is usually isolated from solution) the s a l t is recrystallized from solution containing 15 to 2 0 % HF.

136

33. PRODUCTION O F TANTALUM AND NIOBIUM BY ELECTROLYSIS

The production of metals by electrolysis is based on the cathodic reduction of the metal ions during the passage of an electric c u r r e n t through the electrolyte. Tantalum and niobium cannot be electrolytically separated from aqueous solutions. Like aluminum and other active metals, they can only be produced by electrolysis of molten salts. The production of tantalum powder by electrolysis i s increasingly used in industrial practice, and is about to displace the thermal reduction with sodium. This can be attributed to a number of advantages which will be discussed below. Electrolytic methods for the production of niobium and niobium- tantalum alloys have recently been developed; in general, these methods a r e s i m i l a r to those used in the electrolytic production of tantalum, but the efficiency o f the electrolytic production of niobium is not satisfactory.

Electrolytic production of tantalum The electrolyte used in the production of tantalum i s a melt consisting of KC1, KF, and potassium fluotantalate (K2TaF7). Tantalum pentoxide is dissolved in the molten s a l t s and is electrolyzed. Though it would appear more convenient to conduct the electrolysis using K2TaF7alone, difficulties a r e caused by the s o called "anode effect". It consists in a s h a r p r i s e in the bath voltage at a certain "critical" anodic c u r r e n t density; this is accompanied by a drop in c u r r e n t intensity and the appearance of a c h a r a c t e r i s t i c spark discharge on the anode. Belyaev e t a l . , studied the causes of the anodic effect in the electrolytic production of aluminum from cryolite-alumina melts (which a r e in many r e s p e c t s s i m i l a r to the m e l t s containing K2TaF7and Ta205)and established that the effect is due to insufficient wetting of the graphite by the electrolyte and, in particular, by molten fluorides 1311. The electrolyte is then separated from the surface of the anode by a film of the gases formed on the anode. The anode effect is eliminated by the addition to the electrolyte of surface active substances which improve the wetting of the graphite by the electrolyte (i. e . , reduce the surface tension a t the graphite-electrolyte interface). In fact, the addition of tantalum pentoxide to the electrolyte improves the wetting of the graphite electrode by the melt and i n c r e a s e s the critical c u r r e n t density associated with the appearance of the anode effect. One recommended electrolyte composition is :
55.0% KCI, 27.5% KF, 17.5%K2TaF,.

The presence of potassium fluotantalate in the electrolyte is necessary in o r d e r to ensure a satisfactory solubility of tantalum pentoxide. The concentration of Ta205in the melt commonly used in the electrolysis i s 2.5 to 3 % . Whatever the t r u e mechanism of the electrolysis, i t s end r e s u l t is the electrolytic decomposition of tantalum pentoxide, with the separation of tantalum on the cathode and oxygen on the anode:

137

4Ta5+ + 20e + 4Ta; loo*-- 20e 50,.


--f

The theoretical value of the decomposition potential of Ta205a t the electrolysis temperature (750") is 1.65 V. It is lower than the decomposition potentials of KzTaF7, KC1, and KF. When using a graphite anode, the decomposition potential of TazO5 is reduced because of a secondary p r o c e s s occurring on the anode - the combination of oxygen and carbon with the formation of CO and COz, which is accompanied by a liberation of energy. Thus, the only component consumed during the electrolysis is Ta205, and the electrolyte composition remains unchanged. The true mechanism of the electrolytic production of tantalum has not yet been finally clarified. Investigations c a r r i e d out in the USSR showed that the TazO5 added to the electrolyte r e a c t s with K2TaF7yielding the oxyfluorotantalate K3Ta02F4. The e x c e s s K,TaF7 combines with potassium chloride yielding a complex s a l t - K,TaF,, KC1. As a r e s u l t of the partial dissociation of (Ta@FJ3 and (TaF7)'- anions, the melt contains Ta5+cations /30/, Discharge of the T a 5 +cations takes place on the cathode:
Ta5+

+ 5e
-f

+ Ta.

The following p r o c e s s e s may occur on the anode:


ITaO2F4I3-- 4e [TaF41+ [TaF,]+

+ 0,;

+ 3F-

.+ [TaF,]Z-.

The TaF;- anions again r e a c t with TazO5, yielding T a Q F i - . The oxygen liberated in the p r o c e s s r e a c t s with the graphite anode yielding CO and CO,. According to the above mechanism, the oxygen migrates to the anode a s ( T a q F 4 ) 3 -ions. The Ta205content of the bath d e c r e a s e s in the course of the electrolysis, and the anodic effect i s observed upon reaching a certain lower concentra tion limit; this s e r v e s a s indication that a further amount of T a z 0 5must be added to the electrolyte.

Construction of the electrolyzer and the conditions of electrolysis

Of the electrolyzer designs which have been tried out best r e s u l t s were obtained when a nickel (or nickel-chromium alloy) crucible was used a s the cathode, with a graphite anode placed a t the center of the crucible 1141. A diagram of such an electrolyzer is shown in Figure 72. It consists of a nichrome crucible, appliances f o r supporting and lifting the anode, a feeder and a heat-insulating jacket. The lower, cone-shaped p a r t of the crucible has an opening which is closed by a stopper with a rod serving a s c u r r e n t terminal. A hollow perforated graphite anode is placed a t the center of the crucible. Tantalum pentoxide is fed into the bath through the hollow anode by means of an automatic batch feeder. This method of feeding prevents contamination of the cathodic deposit with the undissolved

138

pentoxide. A flange exhauster s e r v e s to evacuate the gaseous products. The electrolyzer is supplied with a controlled d i r e c t c u r r e n t .

FIGURE 1 2 .

Electrolyzer for the production of tantalum.

1 -hollow graphite anode with perforated walls: 2-nichrome crucible serving as the cathode: 3 -heat-insulating jacket; 4-stopper with a rod serving as

current terminal; 5-control wheel for lifting the anode; 6-cover: I-bracket holding the anode; 8 -feed bin for the supply of Ta,O,; 9-electromagnetic vibrator for the feed bin.

5 G amp/dm2 and an anodic c u r r e n t density of 1 2 0 to 160 amp/dm2.

The electrolysis is c a r r i e d out a t a cathodic c u r r e n t density of about The

I39

residual Ta205 concentration in the electrolyte, at which the anodic effect t a k e s place, depends on the anodic c u r r e n t density. As the anodic density is increased, the anodic effect is observed at higher Ta2O5 concentrations i n the electrolyte. The temperature of the electrolyte is maintained at 6 8 0 - '720'. The average bath potential is determined by means of the equation:

is the decomposition potential of TazO5 which, allowing f o r the oxidation of the graphite anode, is 1.41 V; E e l is the potential drop in the electrolyte, which i s equal to pD1 ( p is the resistivity of the electrolyte, D is the c u r r e n t density in the electrolyte, and I is the distance between the electrodes; the value of D is taken a s D where d, and d c a r e the anodic and cathodic c u r r e n t densities respectively) ; E , is the potential drop in the bus b a r s and contacts, equal to ?ZR where ? is the current, and ZR is the resistance of the conductors and the contacts. In the c a s e of a s m a l l electrolyzer designed for a c u r r e n t of 1 0 0 0 amp and having a total crucible (cathode)volume of about 1 0 l i t e r s , the bath voltage is 6.5 t o 7 V. In such an electrolyzer about 75 70of the dissipated* electrical energy is l o s t as heat (heat l o s s e s through the crucible walls, from the bath and anode s u r f a c e s , and heat l o s t in the evaporation of the salts). The fraction of energy l o s t in the form of heat should d e c r e a s e a s the s i z e of the bath is increased. The initial heating and melting of the electrolyte is accomplished by shortcircuiting the electrodes (anode and cathode) with the aid of a graphite tube. When the salts are molten, the tube is taken out, tantalum pentoxide is added to the melt (to a concentration of 2.5 to 3%) and the electrolysis begun; m o r e T a z 0 5is added in batches when the anode effect is observed. Tantalum is deposited on the walls and the bottom of the crucible in the course of the electrolysis. The electrolysis i s continued until about 213 of the useful volume of the crucible is filled with the cathodic deposit. Since a p a r t of the electrolyte evaporates during the electrolysis, a constant level of the melt in the bath is maintained by periodic addition of a s a l t mixture having the s a m e composition a s the initial charge in the crucible. When the electrolysis is completed, the anode is taken out and the electrolyte together with the cathodic deposit a r e cooled. Under these electrolysis conditions, the c u r r e n t efficiency is about 800/0,and the consumption of electrical power is about 2300 kwh p e r ton of tantalum.
where
Edp

=vG

Processing of the cathodic product T h e r e a r e two methods f o r processing the cathodic product, their aim being to s e p a r a t e the electrolyte from the tantalum powder particles: s i z e reduction with a i r separation and thermal purification in vacuo.

* [In

Russian: "total electrical energy", but this would be in contradiction with the statement in the last paragraph of the section (80% current efficiency). I

140

The air separation method. Because of the g r e a t difference between the densities of tantalum ( 1 6 . 6 g/cm3) and the components of the electrolyte, the l a t t e r a r e readily separated when the cathodic product is milled in a ball m i l l operating in a closed cycle with an a i r separator. This method p e r m i t s the separation of tantalum from the bulk of the s a l t s , which m a y be reused a s the electrolyte. The tantalum powder is then treated with water on beneficiation tables, which r e s u l t s in the removal of the electrolyte residues, and is then treated in porcelain vats with a hot mixture of hydrochloric and nitric acids in o r d e r to remove the impurities. The powder is then rinsed with water and dried. Purification by thermal processing in vacuo. This method was developed in the USSR. The bulk of the s a l t s is separated f r o m the tantalum by fusion (melting) in an atmosphere of argon, and the residual s a l t s a r e separated by vacuum evaporation. The apparatus used is shown in Figure 73 1141. The fusion and the vacuum evaporation a r e c a r r i e d out in a thick-walled r e t o r t 1 made of r e f r a c t o r y steel o r nichrome. The crucible 7 with the cathodic deposit is placed in the upper p a r t of the r e t o r t ; the stopper closing the opening in the bottom of the crucible is replaced by a flange with a screen. A water-cooled condenser 3 is connected through flanges to the lower p a r t of the r e t o r t ; container 4, which i s used to collect the molten electrolyte, is connected to the condenser. The electric furnace 2 is lowered over the r e t o r t . The r e t o r t is connected to a system of vacuum t o 1 0 4 m m H g , and with an argon pumps producing a vacuum of cylinder. The r e t o r t is evacuated, filled with argon, and heated to 1000" when the electrolyte melts. Vacuum is then created in the r e t o r t and the residual s a l t i s removed by evaporation. The imperfectly sintered tantalum dendrites remain on the walls of the crucible; they a r e removed and easily powdered by milling. The molten electrolyte i s returned to e l e c t r o l y s i s . The particle s i z e of electrolytic tantalum powder is much l a r g e r than that of the powder prepared by thermal reduction with sodium. The average particle s i z e of the cathodic product after milling and a i r separation is 30 to 7 0 p , while after thermal processing in vacuum it is 100 to 1201* The powders have the following approximate impurity contents: 0 . 1 to 0.2% 02, 0.03 to 0.2% C, 0.03 to 0.170 F e + N i , 0.0170 F, and up to 0.1% Si. The oxygen content i s about one-tenth that of the powder produced by thermal reduction with sodium invacuum, due to the l a r g e r particle s i z e of the electrolytic powder and the correspondingly s m a l l e r amount of oxide film. Powders with a high carbon content (above 0.5%) may be treated with a mixture of concentrated sulfuric and nitric acids to oxidize the graphite particles. The most important advantage of the electrolytic p r o c e s s is the production of better quality powders. The c o a r s e electrolytic powders have a lower oxygen content which simplifies their conversion into b a r s by pressing and sintering. Moreover, the specific consumption of electrical energy in the electrolytic p r o c e s s is lower than in the t h e r m a l reduction with sodium; if it i s taken into account that the sodium i s also produced by electrolysis. The production of tantalum and niobium by electrolysis of their chlorides i s of g r e a t interest. Tantalum and niobium pentachlorides dissolve in molten alkali chlorides yielding the complex s a l t s MeNbC16 and n/leTaCls
141

(where M e - Na or K). Coarsely crystalline deposits of niobium and tantalum a r e formed on the cathode in the electrolytic decomposition of these complexes, while chlorine is evolved a t the graphite anode.

FIGURE 73. Installation for the vacuum-thermal processing of cathodic deposits. 1-retort; 2-electric furnace; 3-condenser; 4-container used to collect the electrolyte; 5-diffusion pump: 6- trolley for inserting the crucible cathode; 7 -crucible cathode.

3 4 . PRODUCTION O F NIOBIUM BY T H E C A R B I D E R E D U C T I O N METHOD / 13/

The carbide reduction method is based on the interaction between niobium pentoxide and niobium carbide in vacuo a t 1800 to 1900":
Nb,O,

+ 5NbC + 7Nb + 5CO.

Niobium carbide is made by heating a mixture of Nb205 with pure lamp black: Nb,O, 7C PNbC 5CO.

--f

The pentoxide-lamp black mixture is pelletized and the pellets a r e heated to 1800 to 1900" in a tubular graphite furnace, in a hydrogen or argon atmosphere, or to 1600" in a vacuum furnace. The l e s s expensive method

I42

is to produce the niobium carbide in the tubular graphite furnace, with continuous motion of the pellets over the length of the furnace. The pulverized niobium carbide is mixed (in a ball mill) with a s m a l l ( 3 to 5 % ) excess of niobium pentoxide, This provides f o r a m o r e complete removal of the carbon ( a s CO) in the subsequent vacuum sintering of the compacts p r e s s e d from the powder.

To the'pump

"

FIGURE 14. Diagram of a vacuum furnace with a graphite heating element. 1-water-cooled cover; 2-water-cooled base plate: 3- tubular graphite heating element: 4 -current terminals; 5-screens; 6-support; I -pelletized charge; 8-sleeve connector to the pump; 9 -rubber seals; 10-washer (elec trical insulator); 11-window.

The pentoxide-carbide mixture is pressed into b a r s under a p r e s s u r e of 1 t o 1.5 tons/cm2. The b a r s a r e heated in vacuum furnaces with graphite heating elements ( r e s i s t o r s ) (Figure 74) o r in vacuum induction furnaces with a graphite tube. 0.01 to 0.001mmHg. The reduction is c a r r i e d out at 1800 to 1900", a t The reduction takes place through intermediate stages involving the formation of lower oxides (NbO, and NbO) and a carbide, Nb2C. Since the oxides NbO, and NbO a r e f a i r l y volatile a t 1800 to 1900" in vacuo, the gaseous oxides r e a c t with niobium carbide, a fact which explains the high r a t e of the reduction. The bulk of carbon monoxide is evolved rapidly, and a l a r g e proportion is evacuated by the pump within a few minutes. The p r o c e s s is completed when the vacuum attains lo-, m m Hg.

143

The product consists of sintered porous niobium pellets, which retain the initial shape of the compacts but with a certain reduction in size. The pellets a r e forwarded to electron-beam melting. They may a l s o be sintered in vacuo a t 2300 to 2350, yielding a ductile metal. In another technological p r o c e s s , the pellets a r e pulverized and the powder is p r e s s e d into b a r s and sintered in vacuo. To facilitate the comminution the pellets a r e hydrogenated (heated in hydrogen a t 360"). The metal absorbs hydrogen and becomes brittle. The carbide reduction method h a s the advantage of a high d i r e c t yield of the useful metal (at l e a s t 9670) and the use of a cheap reducing agent. The method may also be used f o r the production of tantalum and tantalumniobium alloys. It has the disadvantage of requiring the use of hightemperature vacuum furnaces of complex design.

35. PRODUCTION OF TANTALUM AND NIOBIUM B Y REDUCTION OF THEIR CHLORIDES


T h e production of tantalum and niobium from their chlorides has recently attracted attention: this method is especially suitable when the processing of concentrates results in chlorides a s the end product. Niobium and tantalum pentachlorides m a y b e reduced by metals (sodium, magnesium) or hydrogen. These methods are not industrially used as yet, and they will be described only briefly on the basis of laboratory data / l / . Reduction of chlorides by magnesium is an exothermic reaction: 2TaC1,

+ 5Mg + 2Ta + 5MgCI5 + 38 kcal.

T h e heat of reaction is sufficient for a sklf-sustained reaction. T h e reduction is carried out in a steel crucible containing a charge consisting of T a C l S , magnesium turnings and potassium and sodium chlorides which act as flux (reducing the reaction rate by decreasing the specific heat effect of the process). T h e crucible is slowly lowered into a furnace heated t o about 750'. The reaction is rapid. T h e tantalum particles formed are protected against oxidation by the molten salt mixture consisting of MgCl,, NaCl and KCl. T h e powder is freed from chlorides by washing with water and dilute hydrochloric a c i d , and is then rinsed with alcohol and dried in vacuum. T h e yield of tantalum powder is high (about 98%). T h e particle size of the powder is between 1 and l o p . T h e powder contains 0.1-0.5% Mg, 0.3% Fe. and -0.1% Ti. T h e same method may b e used to produce niobium or niobium-tantalum alloys, by simultaneous reduction of their chlorides. In order to produce tantalum and niobium powders of higher purity, it is preferable t o reduce the gaseous chlorides by liquid magnesium, as in the production of zirconium. Reduction of tantalum and niobium chlorides t o the m e t a l by hydrogen is carried out at temperatures above 600'. T h e production of niobium has been studied more thoroughly, and the process may be effected with simultaneous removal of iron chloride from niobium chloride and the separation of tantalum and niobium. According to one method which has been described, niobium pentachloride is produced by chlorination of ferroniobium, containing 70% Nb*Ta, 24.7%Fe, 1 . 7 % Mn, 1.5% Al. 1% Ti, 0.3% W , A mixture of NbCI,, TaCI,, and FeCIa was condensed after chlorination a t about 0.2% Sn. and 0.2% Cr. 1000". In order to separate the iron, the chlorides were distilled in hydrogen at 350". Under these conditions ferric chloride was reduced to the nonvolatile FeCI,. T h e tungsten present as an impurity was separated at the same t i m e . Tantalum was separated from niobium by passing a mixture of the gaseous chlorides and hydrogen through a rube heated to 500-551)". Under these conditions niobium pentachloride was reduced t o the nonvolatile NbC1, , while TaCI, was not reduced and was collected by condensation a t low temperatures. Niobium trichloride is stable in air. It is reduced t o the metal by hydrogen a t 800-1000"; the reduction is carried out in molybdenum boats placed in a tubular quartz furnace. The powder produced by the above method is characterized by low oxygen and nitrogen contents ( - 0.03 and 0.05'%respectively). T h e disadvantage of the process is the low reduction rate and the difficulties caused by the side reaction -the disproportionation of NbCI, with the formation of niobium and the volatile higher chlorides (NbCI, and NbCI,).

144

3 6 . PRODUCTION OF SOLID, DUCTILE TANTALUM AND NIOBIUM

Solid tantalum and niobium were previously produced solely by powder metallurgy techniques. Melting techniques, which p e r m i t the production of l a r g e ingots, have been recently developed. Both metals can be melted in a r c furnaces with cooled copper crucibles, o r in electron-beam furnaces. In contrast to tungsten and molybdenum, both the sintering and melting of p r e s s e d tantalum and niobium b a r s are c a r r i e d out in high vacuum. As a result, the m e t a l s a r e freed f r o m adsorbed gases, and f r o m oxygen, carbon, silicon and many other elements p r e s e n t as impurities.

The powder metallurgy method The fundamentals of the method have been discussed in the chapter "Tungsten". The p r o c e s s c o m p r i s e s two stages: p r e s s i n g of the powder into b a r s and sintering. Pressing. Tantalum and niobium powders a r e p r e s s e d into b a r s with rectangular (when intended for rolling into sheets) or s q u a r e (when intended for wire drawing) c r o s s sections. The length of the b a r s is from 600 to 750 mm, their c r o s s section - from 4 to 20 cm'. The pressing conditions depend on the nature of the starting powder. The fine tantalum and niobium powders produced by thermal reduction with sodium have a l a r g e specific surface and a r e thus p r e s s e d a t lower p r e s s u r e s than the electrolytic powders. The p r e s s u r e needed for tantalum powder is 3 to 5 tons/cm2, and for niobium powder about 1 ton/ om'. Electrolytic powders have a c o a r s e r particle size, and 'are p r e s s e d under 7 to 8 tons/cm'. The porosity of the pressed b a r s i s 2 5 to 30% by volume for powders 0 7 ' 0 by volume for produced by thermal reduction with sodium-, and about 2 electrolytic powders. Hydraulic p r e s s e s s i m i l a r to those used in the pressing of tungsten powder a r e used. Sintering. A number of complex physical and chemical p r o c e s s e s occur when tantalum and niobium b a r s a r e heated in vacuo; each p r o c e s s takes place within a certain temperature range following the i n c r e a s e in temperature. The occurrence of these p r o c e s s e s is evidenced by the changes in certain properties of the b a r s : electrical resistivity, density, weight, microstructure, and by the impurity content of the metal. The changes in electrical resistivity a r e especially indicative since the resistivity is strongly affected by the presence of impurities and the porosity of the b a r . Figure 75 gives curves showing variations in the specific electrical resistivity, density and weight of a tantalum b a r ( p e r cm2 of i t s surface) a s a function of the sintering temperature. The s h a r p d e c r e a s e in electrical resistivity occurring when the b a r is heated in vacuum to 600 to 650" is due to the escape of the hydrogen from the metal. The electrical resistivity does not drop significantly between 600 and 1100". Above 1100" there is a d e c r e a s e in the electrical

145

I I

resistivity up to 1600". The resistivity increases somewhat between 1600 and 1900" and drops to a minimum a s the temperature is further increased. The d e c r e a s e in resistivity at llC0 to 1600" is associated with incipient sintering and further removal (evaporation) of certain impurities (alkali metals - sodium and potassium, and their s a l t s ) , a s well as with incipient elimination of carbon as carbon monoxide, formed by the inter action of carbon with the oxide film. A slight i n c r e a s e in the electrical resistivity between 1600 and 1900", which is accompanied by an increase in the weight of the bar, is probably caused by some slight absorption of gases (nitrogen, oxygen) which a r e present in the interior of the vacuum vessel used for the sintering.

Temperature, ' C

FIGURE 75. in vacuum.


1-electrical

Changes in t h e properties of pressed tantalum bars during sintering resistivity; 2-density: 3-changes in weight.

A d e c r e a s e in the electrical resistivity and a considerable decrease in weight take place above 1900". Intensive elimination of carbon and silicon ( a s CO and the lower oxide S O ) , sublimation of lower tantalum oxides (e. g. , T a G ) a s well a s vaporization of impurities (titanium, iron, nickel) take place at 1900 to 23 00". The removal of impurities is accompanied by a simultaneous increase in the density(contraction) of the b a r s , Vaporizationof a p a r t of the tantalum and a f u r t h e r contractionof the b a r s is observed at 2400 to 2600". The s a m e p r o c e s s e s take place during the sintering of niobium bars, but the temperature ranges a r e somewhat lower. It w i l l be seen that the sintering in vacuo is accompanied by the removal of a number of impurities from the bars. The oxygen is removed a s CO, SiO, and lower tantalum o r niobium oxides. A considerable proportion (up to 5 to 10%by weight) of the metal may be lost as the lower niobium or tantalum oxides. Thus, it i s expedient to remove most of the oxygen a s CO. Hence, a calculated amount of carbon ( a s pure carbon black) is added to the starting powder if it contains a l a r g e amount of oxide films.

I46

When the powders contain a l a r g e amount of carbon and a s m a l l amount of oxide film, it is advisable to add tantalum ( o r niobium) oxide to the s t a r t i n g powder in o r d e r to remove the carbon during the sintering of the bars. Because of the considerable amount of g a s e s evolved during the sintering, the temperature must not be raised too quickly a s the rapid sintering and contraction of the b a r c a u s e s the open p o r e s to close up and thus interferes with the removal of the volatile impurities. This may result in high p r e s s u r e s being produced within the closed p o r e s and in swelling and formation of l a r g e p o r e s and cavities in the b a r s . The sintering conditions depend on the composition and particle s i z e of the powders. However, i n all c a s e s the temperature must b e raised stepwise and the b a r must be held f o r some time a t a constant temperature a t each successive stage. A s an illustrative example, one set of conditions recommended for the sintering of b a r s p r e s s e d from electrolytic tantalum powder is shown in Table 2 6 .

TABLE 26 Sintering parameters of tantalum bars ( 2 0 X 2 0 x 600 mm) pressed from electrolytic powder Temperature range. ' C u p to 1750 1750-2000 2000 2000 -2600 Mode of raising of the temperature and holding times After e a c h 100" hold for 4 minutes After each 20' hold for 4 minutes Hold for 60 minutes After each 25" hold for 4 minutes Total sintering t i m e Total time (raising and holding)
1 hour 10 minutes 50 minutes 1 hour 1 hour 40 minutes
4 hours 40 minutes

The total sintering time depends on the amount of impurities in the starting powder. The sintering time for b a r s of tantalum powders niade by thermal reduction with sodium i s 8 to 1 2 hours, that of electrolytic powder b a r s is 4 to 6 hours. The only difference between the sintering of b a r s made of tantalum and niobium powders is that in the case of niobium the maximum sintering temperature is 2200 to 2300". In o r d e r to achieve satisfactory sintering and removal of impurities, i t i s necessary not only to maintain the required temperature conditions, but also to provide f o r a rapid evacuation of the gases. The residual to 10-5mmHg. p r e s s u r e in the sintering furnace must be about The residual porosity of the sintered b a r s is usually 1 0 to 15%. A virtually pore-free metal is produced by forging of the sintered b a r followed by vacuum annealing a t a high temperature. Twice-repeated forging and annealing produces pore-free b a r s which are used for sheet rolling o r w i r e drawing. In most c a s e s tantalum and niobium b a r s are heated directly by passing through an electric c u r r e n t (Figure 76) a s is done in the c a s e of tungsten

147

b a r s + ; the only difference being that the space under the cover, which contains the bar, is connected to a vacuum pump. In addition, graphite and molybdenum sheet s c r e e n s are placed around the b a r in o r d e r to reduce heat l o s s e s and to provide a m o r e uniform heating of the b a r .

FIGURE 76. Diagram of a vacuum apparatus for hightemperature sintering of tantalum and niobium bars.
1 -cooled steel cover; 2-upper stationary contacr: 3-copper supports feeding the current t o the upper contact; 4 -bar to be sintered; 5-lower movable contact; 6-steel plate; '?-rubber rings; 8-steel plate; 9-flexible bus barserving ascurrent terminal; 10 -rubber hose for the feed and discharge of water to the lower contact; 11-bus bars; 12-weight; 13 -lower cover; 14 -connection to the vacuum

Pump.

Large pressed b a r s and products having a complex shape a r e heated indirectly in induction o r resistance-heated furnaces.

-___-_----_____
The electrical conductivity of bars pressed from tantalum powders made by thermal reductionwith sodium isoften too low to permit direct heating by the current. To increase the conductivity, the bars are presintered a t 1000-1200" i n a horizontal tubular vacuum furnace with an alundum tube heated by a molybdenum wire coil. Bars pressed from niobium powder and bars pressed from electrolytic tantalum powder usually do not require presintering.

148

FIGURE 11. Diagram of induction vacuum


furnace. 1-window; 2-water-cooled lid: 3-tungsten cylinder heated by induction current; 4-molyb denum screens; 5 -quartz tube; 6-inductor (water-cooled copper tube); I -water-cooled cylinder: 8 -bar t o b e sintered; 9-molyb denum screens; 10-graphite support; 11 waler-cooled plate; 12-bus bars (current leads to the inductor); 13-vacuum seals and elec trical insulation; 14-rubber rings.

A diagram of a vacuum induction furnace used for sintering is shown in Figure 77. The b a r to be sintered is suspended in a tungsten tube (made of tungsten rings) heated by an inductor mounted on a quartz tube. The heat is transferred by radiation from the tungsten tube to the bar. To reduce heat l o s s e s the tungsten tube is surrounded by cylindrical molybdenum s c r e e n s which a r e cut s o that the formation of closed-circuit induction c u r r e n t s is eliminated. It is recommended that the inductor be fed with a frequency of 2 to 4 kilocycles/ s e c / 11.

The melting of niobium and tantalum

/ 171

The arc-melting method. Niobium and tantalum may be melted in a r c furnaces with a consumable electrode and a cooled copper crucible; the construction of these furnaces has been described in the chapter "Molyb denum". Sintered b a r s a r e used as consumable electrodes. They a r e melted in o r d e r to produce l a r g e r b a r s which a r e used f o r sheet rolling. Since in the arc-melting p r o c e s s the metal r e m a i n s in the molten s t a t e only f o r a s h o r t time, purification by vaporization of impurities is

149

insufficient, This is especially true in the case of niobium, whose melting point is about 500" lower than that of tantalum. Because of the higher melting point of tantalum, a r c melting of this metal results in a m o r e complete removal of impurities. Electron-beam melting, This p r o c e s s , which was developed only recently, p e r m i t s v e r y high-purity niobium and tantalum ingots to be produced. Electron-beam heating of metals i s based on the conversion of most of the kinetic energy of the electrons into heat upon impact against the metal surface. A s m a l l e r proportion of the kinetic energy i s converted into X-ray energy. The impact of an electron moving with a velocity D, with the body to be heated r e s u l t s in an energy distribution obeying the equation:

where m is the m a s s of the electron; Q is the fraction of kinetic energy converted into heat; h C h is the fraction of kinetic energy converted into X-ray energy. Here h is the Planck constant, C is the velocity of light and i. is the wavelength of the X-rays. An increase in the velocity of the electrons is accompanied by an increase in the proportion of the energy emitted as X-rays since the X-rays become h a r d e r (i. e . , their wavelength becomes shorter). Thus, protective m e a s u r e s against radiation must be taken. The electron velocities in electron-beam melting equipment a r e theref o r e limited to acceleration voltages not exceeding 30 to 35 kV. The equipment consists of the following p a r t s : 1) an electron gun, in which a directed electron beam i s produced; 2) a melting chamber; 3 ) a high-vacuum system; 4) a source of high-voltage d i r e c t c u r r e n t , The maintenance of a high vacuum in the system is required in o r d e r to reduce to a minimum the l o s s e s of energy caused by the impact of the electron beam on the gas atoms and molecules in i t s path to the body to be heated, and to extract impurities from the molten metal which a r e volatilized during the melting. A simplified diagram of electron-beam melting i s shown in Figure 7 8 . The electron gun (which c r e a t e s a focused, intense electron beam) consists of a cathode, an anode, a control grid, and focusing and deflecting coils. An incandescent tungsten coil may be used a s the cathode, but indirectly heated cathodes made of pure tungsten and tantalum a r e m o r e frequently used. A high negative potential is applied to the cathode. The electron beam p a s s e s through a grounded hollow anode. A cylindrical control electrode (known a s a modulator or a grid) with a negative potential is placed between the cathode and the anode. The shape of the electron beam and the density of the space charge (around the cathode) a r e controlled by varying the potential of the control grid. The focusing and the deflection of the electron beam a r e accomplished by means of magnetic and an electrostatic field The design of a furnace used f o r electron-beam melting is shown in Figure 79. The furnace has a single electron gun, with an independent

evacuation system. The electron beam penetrates into the working chamber of the furnace through a special diaphragm which produces considerable resistance to g a s flow and thus ensures that the vacuum in the gun space is higher than that in the melting chamber. The electron beam is directed onto the end of the metal to be melted (a pressed, sintered bar). The electron beam also impinges on a water-cooled crystallizer, in which the metal r e s i d e s for a certain time in the molten state.

FIGURE 78. Diagram of electron-beam m e l t i n g . 1-cathode (under high voltage); 2 -control grid; 3-anodic tube; 4-focus ing coil; 5-deflecting coil; 6-metal to be melted.

FIGURE 19. Diagram of an electron-beam melting furnace. I-electron gun cathode; 2 -electron gun anode; 3-sleeve connection to the vacuum system; 4-electromagnetic coil (lens); 5-diaphragm separating the melting chamber of the furnace from the electron gun; 6-gate; I-melting chamber; 8 -electron b e a m ; 9-sleeve con nection to the vacuum system; 10-melted ingot; 11-ingot drawing mechanism: 12 water-cooled copper crystallizer; 13- sintered bar, to b e melted.

to lO-'mmHg) which The melting is c a r r i e d out in high vacuum ensures the removal of impurities which a r e vaporized a t the melting temperature. Thus, the melting of niobium is accompanied by the removal of oxygen ( a s C O and lower oxides), nitrogen, hydrogen, iron, nickel, chromium, and aluminum. However, refractory metals such a s molyb denum, tungsten, and tantalum a r e not removed. The approximate power of electron-beam equipment used for the melting of niobium and tantalum ingots of various diameters 1271 is given below:
151

Niobium : ingot diameter, m m power, kw Tantalum : ingot diameter, m m power, kw

50 30

75
60

100 120

35
30

50 60

75 120

100 240

IJp to 130kg niobium may be melted at 1 2 0 kw p e r hour. The furnace just described is charged with sintered b a r s . However, powder, sponge or turnings may also be melted in the crucible. Electron-beam melting has the advantages of producing a high-purity metal (because the molten metal may be held in high vacuum f o r a prolonged time), of low cost (the efficiency of the system is m o r e than 95%), and of the fact that the metal to be melted may be in any form ( b a r s , powder, turnings, waste lumps) while the use of a r c melting involves the preparation of consumable electrodes.

Vacuum pumps and vacuum systems The vacuum of the order of to lO-'mmmHg o r higher, which is required f o r the melting, is usually created by two types of pumps operating in conjunction - a mechanical oil pump (preliminary vacuum pump) and a diffusion pump. The mechanical pumps c r e a t e a vacuum of 0.05 to 0.5mmHg, which is required for the operation of the diffusion pump that c r e a t e s a higher vacuum and ejects the evaucated gas into the mechanical pump.

. )

FIGURE 80. Diagram of a mechanical (rotary) o i l pump.

FIGURE 81.

Diagram of a diffusion

Pump.

Mechanical pump. A diagram of one of the most commonly used mechanical pumps - the rotary oil pump - is shown in Figure 80. The

152

pump consists of a steel cylinder 1 with an excentric rotor 2 inside it. The r o t o r c a r r i e s blades 3 s e t within grooves; the blades p r e s s (with the aid of springs) against the inner s u r f a c e of the cylinder. When the r o t o r is s e t in motion, the blades act as pistons. When the r o t o r is rotated in a counterclockwise direction, the a i r is pumped through sleeve connection 5 and expelled through the escape valve 4 into the space above the oil. The cylinder is i m m e r s e d in oil which s e r v e s a s a lubricant and provides a hermetic s e a l on the pump. One- and two-stage pumps of this type a r e manufactured. The two-stage pump consists of two pumps connected in s e r i e s and mounted on a common shaft. Diffusion pumps. Diffusion pumps operating on m e r c u r y o r oil vapor, but based on the s a m e principle, a r e used in industry. A diagram of a diffusion pump is shown in Figure 81. The pump consists of a water-cooled cylindrical s t e e l vessel 1 containing m e r c u r y or oil (which have a low vapor p r e s s u r e a t ambient temperatures) on the bottom. A funnel-shaped tube 2 with a cap 3 suspended over it (thus creating a ring-shaped nozzle 4 in the upper p a r t of the FIGURE 82. Two-stage oil diffusion cylinder) is mounted inside the vessel, The Pump.
bottom of the pump is heated. The mercury 1-heater; 2-boiler; 3-cooler;
4-and 5-ring-shaped nozzles;
or the oil evaporates, the vapors r i s e within 6-connection to preliminary vacuum the tube, s t r i k e the cap and a r e thrown pump; ?-connection to the system a t a high velocity from the nozzle onto the to be evacuated; 8-copper rod cooled walls of the body. The gas from the (serving as heat-transfer medium). sleeve connection 5 is entrained downwards with the vapor s t r e a m and is pumped out through the sleeve connection 6 by the mechanical pump (described above). The body walls of the pump a r e cooled by the water jacket 7. Two- and three-stage diffusion pumps with a number of nozzles mounted over each other (Figure 82) are produced commercially. Oil-diffusion pumps a r e being increasingly used. Purified oils (n-dibutyl phthalate, butylbenzene phthalate) which have a very low vapor p r e s s u r e a t 20" a r e used f o r the purpose. To prevent the penetration of oil vapor into the furnace, the vacuum system is fitted with "freezing" t r a p s (cooled by liquid a i r o r nitrogen). A typical high-vacuum system is shown in Figure 83. The evacuation is effected a t first with the mechanical pump, the diffusion pump being disconnected. The diffusion pump is connected when the vacuum in the system becomes a s low a s 0.05 to 0.5mmHg; i t expels the evacuated gas into the preliminary-vacuum cylinder. In each c a s e the vacuum system is computed and the pumps selected f r o m the furnace size, the amount of g a s e s to be evaucated, and the r a t e of evaucation required.

I53

FIGURE 83.

A typical high-vacuum system.

1-volume to b e evacuated; 2-diffusion pump; 3-vacuum valve; 4 -vacuum faucets; 5-preliminary vacuum cylinder; 6-mechanical (rotary) pump.

The high-vacuum oil-diffusion pumps made in the USSR have an evacuation rate of 100 to 8000 l/sec o r m o r e in the p r e s s u r e range 2 . (10-4-10-5)mmHg.

37. RECOVERY OF METALLIC TANTALUM AND NIOBIUM WASTES


Two types of wastes accumulate in the production of the solid metals. B a r f r agments These consist mainly of the trimmed bar ends after high-temperature sintering since the b a r ends clamped in the contacts do not reach the maximum sintering temperature. Fragments of discarded b a r s (sintering wastes) also belong to this type of wastes. However, in a well-designed production such waste may be altogether eliminated. W a s t e s f r o m m e c h a n i c a l p r o c e s s i n g (cutting, rolling wastes). This type of waste also comprises processed tantalum and niobium-anodes, grids, and other p a r t s from discarded radio valves and rectifiers. The two types of wastes mentioned above consist of the pure metals. They a r e recovered (reconverted into useful metal), thus increasing the degree of extraction of the metal and reducing its price. Tantalum and niobium wastes are pulverized and the powders a r e sent to the bar-pressing stage Since tantalum and niobium a r e ductile, their crushing i s difficult. In the production of powders, use is made of the embrittlement of tantalum and niobium resulting from the absorption of hydrogen. Tantalum and niobium wastes are heated in hydrogen (niobium to 360" and tantalum to 500"). The hydrides produced a r e readily crushed to a fine powder in a steel ball mill. The powder is treated with hydrochloric acid to f r e e i t from iron (acquired contamination by friction during milling). The powder is then rinsed with water, dried, and combined with the powder sent to pressing. The hydrogen is removed from the metal during low- and hightemperature vacuum sintering. Wastes may also be reprocessed by melting in arc o r electron-beam furnaces.

154

-. . . .

..

I,

Chapter IV
TITANIUM
38.
GENERAL DATA ON TITANIUM Brief historical note The element titanium was discovered in 1791 by Gregor (a British amateur mineralogist) in the black magnetic (iron-bearing) sands of Menaccan in Cornwall. The new element was called menaccanite. In 1795 the German chemist Klaproth investigated the mineral rutile and found that i t was the oxide of a new element which he called titanium. It was subsequently established that titanium is identical with menaccanite. P r i o r to 1849 i t was believed that the metal-like titanium carbonitride found in the slags of blast furnaces was metallic titanium. Relatively pure titanium was obtained by Hunter in 1910 - 120 y e a r s after i t s discovery. The industrial utilization of titanium both a s chemical compounds and a s an alloy additive dates from the f i r s t decades of the twentieth century. Titanium attracted i n t e r e s t a s a potential s t r u c t u r a l m a t e r i a l during the Second World War. This resulted in the development of industrial methods for the production of the ductile metal and to the introduction of l a r g e - s c a l e production of the metal and its alloys in a number of countries.

Properties of titanium Titanium has the appearance of steel. The pure m e t a l is ductile and is easily worked by p r e s s u r e . Two crystalline modifications of titanium a r e known. The form which is stable up to 882" is @-titaniumwhish has a hexagonalzlose-packed lattice with the p a r a m e t e r s a = 2.951 A and c = 4.692 A. Above 882" a-titanium is converted into p-titanium, with a body-centered cubic lattice. The lattice p a r a m e t e r of p-titanium is a = 3.3065 1 (at 900"). Some physical properties of titanium a r e reported below, while the mechanical properties of titanium a r e listed in Table 27.
Atomic n u m b e r . . ............................. Atomic weight ............................... Density, g/cmS Melting point, 'C Boiling point, " C Superconductance transition temperature. K
22 41.90 4.51 (a-titanium) 1668i 4 3300 0.53

.............................. ............................ ............................ ......

I55

Vapor pressure, mm Hg. a t temperatures, 'C

1227

1442 1727 2477 Heat of fusion, cal/g Specific heat (in the range 0 to 100') c a l / g , "C ... Thermal conductivity (in the range 0 t o 200') cal/cm.sec, "C Linear expansion coefficient a t 20 t o 300'C Specific electrical resistance, ohm. cm. IO6 at: 20'C 800C Electron work function, eV T h e r m a l neutron capture cross section, barns

.................................. ................................... ................................... ................................... ........................


...........................

10 1 10 -104 0.127 0.04 8.2. 42 180 4.09 5

.....

lo-'

.................................. ..................... .....


TABLE 27

..................................

T h e mechanical properties of titanium

Method of producti6n and processing

3
M
*" U E
I E
-,A

. e
Y

823

Gi

00

High-purity tiranium (iodide method) after vacuum annealing a t 8OO'C ....... T h e same, after deformation (50 01. reduction) Prepared by reduction of titanium chloride with magnesium and melted in an arc furnace with a cooled copper crucible, after annealing T h e same, after deformation (50 70 reduction)

27-34
68-75

40-55
5-11

9850 - 10900
11300

13

...........

................ ...........

U--60

20-25 12-14

11200-11600

20-185 .50-280

75-83

The mechanical properties a r e strongly affected by the purity of the metal and i t s mechanical and thermal processing history. Titanium I s characterized by i t s ability to absorb g a s e s - oxygen, nitrogen, and hydrogen. These g a s e s a r e soluble in titanium up to a certain limit. The presence of even a s m a l l amount of oxygen and nitrogen reduces the ductility of titanium ( s e e Figure 84). Very s m a l l amounts of hydrogen (0.01 t o 0.005'%) noticeably increase the brittleness of titanium, which is measured by means of the impact strength (Figure 85). Titanium is r e s i s t a n t to the effect of a i r a t ambient temperatures. When heated to 400 to 550" the metal is coated with an oxide-nitride film which adheres strongly to the metal and prevents further oxidation. At higher temperatures there is an i n c r e a s e in the r a t e of oxidation and in the solubility of oxygen in titanium.
156

Titanium r e a c t s with nitrogen at temperatures above 600" with the formation of nitride (TiN) films and solid solutions of nitrogen in titanium. Titanium nitride is very hard and m e l t s at 2950". Titanium absorbs hydrogen with the formation of solid solutions and hydrides (TiH and TiHz). The solubility of hydrogen in a-titanium is about 1% by weight ( o r 3 3 % (atomic)). Unlike oxygen and nitrogen, the absorbed hydrogen may be expelled almost quantitatively from titanium by heating in a vacuum at 1000 to 1200". Titanium r e a c t s with carbon and carbon-containing g a s e s (CO, CHd at high temperatures (above lOOO"), yielding a hard, refractory carbide ( T i c , mp 3140"). The presence of carbon (as an impurity) has a strong effect on the mechanical properties of titanium (see Figure 84).

" E E

" 3 B
#+

.
M

f?

z 2
DG

ai az ~3

L V

~6

m U

$ 0

Concentration, %

om5 a010 a m m o
Hydrogen content.%

FIGURE 84. Effect of nitrogen. oxygen, and carbon on the hardness of titanium after melting.
1 --metal containing nitrogen; 2 metal containing oxygen; 3-metal containing carbon.

FIGURE 85. Effect of hydrogen on the impact strength of titanium.

Titanium r e a c t s with fluorine, chlorine, bromine, and iodine at relatively low temperatures (100 to 200"), yielding the volatile titanium halides. F r o m the standpoint of i t s corrosion resistance titanium resembles the chromium-nickel stainless steels. The metal is not corroded by cold o r boiling water. Titanium is virtually unaffected by cold or hot nitric acid of any concentration, due to the formation of an adherent, protective oxide film. Titanium r e s i s t s the effect of dilute (up to 5'70) sulfuric acid at room temperature. However, the r a t e of corrosion increases as the acid concentration is increased, reaching a maximum at 40% H2S04, then dropping to a minimum at 60% HzS04, increasing to a second maximum at 80% HzSO,, and decreasing again at higher acid concentrations (Figure 86). Titanium displays a satisfactory resistance to dilute (5 to 1 0 % ) hydrocMoric acid at room temperature. However, the r a t e of corrosion increases rapidly when the acid concentration o r the temperature is increased. The corrosion of titanium in hydrochloric acid may be considerably reduced by the addition of s m a l l amounts of an oxidizing substance (HN03,KMn04, KzCr04, and copper o r i r o n salts) to the solution.

157

Titanium dissolves readily in hydrofluoric acid. Titanium is resistant to cold and hot solutions of alkali hydroxides (at concentrations of up to 20%).

HzSO, concentration. 70by weight FIGURE 86. Effect of the acid concentration on the corrosion f titanium in UzSO,. rate o 1-at 30'C; 2-at 40C; 3 - a t 50C; 4-at 1OO'C.

The high resistance of titanium to corrosion in s e a water and in solutions of chlorides of a number of metals a r e of great importance.

The properties of titanium compounds Titanium belongs to Group IV of the periodic system. In its most stable and important compounds titanium is tetravalent: There a r e a l s o compounds of di- and trivalent titainium. The lower valency compounds are unstable in aqueous solutions, in which they act as strong reducing agents.
TABLE 28 Properties of titanium oxides

1
Oxide Melting point, " C Density 1850 1900 -2130 1750-2020

4.18-4.25 4.6 4.93

Heat of Formation, :cal/mole 225.5 362.9 123.9

Crystalline structure and l a t t i c e parameters Tetragonal(ruti1e type), a=4.48 i, c=?.95 Trigonal (corundum type), a = 5.14 A , c = 13.61 A Face-ceptered cubic. The composition fluctuates between TiO, .a, and TiOl.3,.

1
158

The l a t t i c e parameter fluctuates from


a=4.182 ton=4.152 A

I
Oxides. Titanium f o r m s three simple oxides in which it is t e t r a - , t r i - and divalent: T i Q , Ti&, and TiO. In addition, there are oxides of intermediate composition containing titanium atoms in various valency states. F o r instance, Ti,Ob ( o r Ti& Ti%) is one of these oxides. The higher oxides a r e amphoteric, but the lower oxides a r e basic. Some of their properties a r e listed in Table 28. T i t a n i u m d i o x i d e is one of the most important products obtained by processing of titanium-containing raw m a t e r i a l s . The pure oxide is white. It o c c u r s in nature in the form of three m i n e r a l s - rutile, anatase, and brookite, which a r e allotropic modifications of titanium dioxide. Rutile is the most stable of these modifications. Titanium dioxide is virtually insoluble in water and dilute acids. It dissolves in hot concentrated sulfuric, hydrochloric, and nitric acids, a s well a s in hydrofluoric acid. In acid solutions titanium is present either a s the TiG cation o r a s a divalent cation - TiO% (titanyl ion). Titanic acid s a l t s (titanates) a r e formed when titanium dioxide is fused with alkalies. T i t a n i u m s e s q u i o x i d e (Ti,Q) is a dark-violet powder. It may be produced by reduction of Ti02 with hydrogen a t 1100 to 1200" o r with carbon at 1350 to 1400". T i 2 Q is sparingly soluble in water. It dissolves in hot sulfuric acid, yielding violet-colored solutions of titanium sulfate (in which titanium is trivalent). T i t a n i u m m o n o x i d e (TiO) is a golden-yellow substance with a metallic luster. Titanium monoxide is produced by heating a mixture of titanium dioxide and titanium powder in vacuo a t 1550" o r by reduction of TiOz with magnesium, aluminum o r carbon. Titanium monoxide is insoluble in water. It r e a c t s with sulfuric and hydrochloric acids with the evolution of hydrogen and the formation of trivalent titanium solutions. A gelatinous precipitate consisting of orthotitanic Ti t an ic a c idE (or a-titanic) acid (H4Ti04)is formed in the cold when sulfuric o r hydro chloric acid solutions a r e neutralized. The orthotitanic acid is converted into a m o r e stable substance - metatitanic ( o r 8-titanic) acid (HzTi03)when the solution is boiled o r allowed to stand for a prolonged time. In contrast to orthotitanic acid, metatitanic acid does not dissolve in cold dilute inorganic acids and alkali solutions. Metatitanic acid dissolves in hot, concentrated. sulfuric acid (at concentrations of 6 0 to 70% and higher) and in hydrofluoric acid. Titanium is characterized by the formation of pertitanic acids, which a r e produced by the effect of hydrogen peroxide on acid o r neutral solutions of tetravalent titanium compounds. Pertitanic acids having the compositions H4Ti05 and H4Ti08, a s we11 a s their salts, a r e known to exist. Solutions containing pertitanic compounds a r e yellow ( a t low concentrations) o r orange-yellow ( a t high concentrations). This is used in the analytical determination of titanium. Titanic acid s a l t s , titanates, a r e formed when titanium Titanates dioxide is fused with alkalies. T h e r e a r e meta- and orthotitanates (NazTiQ,Na4Ti04), a s well a s polytitanates in which the molar ratio TiOz : NazO > 1, e. g . , Na2Tiz05, NazTi3Q. Alkali titanates a r e sparingly soluble in water. O f the other metal titanates we may mention calcium titanate C a T i Q which is found in nature a s the mineral perovskite; iron titanate F e T i Q , found in nature a s the mineral ilmenite; calcium silicotitanate CaO. T i Q S i Q , found in nature a s the mineral sphene; and barium titanate B a T i Q

159

which has an extremely high dielectric constant and which is used for the production of high-capacity electrical condensers. T i t a n i u m s u 1f a t e s Solutions containing mainly titanyl sulfate TiOSOI a r e formed when titanium dioxide or metatitanic acid a r e dissolved in sulfuric acid. The normal sulfate Ti(S0J2 has not been isolated from aqueous solutions, although i t is possible that in strongly acid solutions p a r t of the titanium is present as Ti& ions. Titanyl sulfate, which crystallizes from solutions as the dihydrate TiOSO4. 2Hz0, is the only stable sulfate known. The stability of titanium sulfate solutions v a r i e s depending on the concentrations of the titanium and the sulfuric acid. When the solutions a r e diluted and heated, the sulfates hydrolyze with the precipitation of the sparingly soluble metatitanic acid:

TiOSO,

+ 2H,O

-f

H,TiO,

+ H,W4.

This reaction is used in the manufacture of titanium dioxide. The sulfates of tetravalent titanium a r e readily reduced in solution (for instance, by zinc o r iron) to the trivalent titanium sulfate Ti2(S0J3 o r the acid salt Ti3H(S04),. Ti3+ions a r e violet. Halides. Titanium combines with halogens yielding volatile higher halides of the type TiR4 (where R is F, C1, Br, o r I). In addition, there a r e halides in which titanium is trivalent o r divalent. The properties of some of the halides a r e listed in Table 29.
TABLE 29 Some properties of titanium halides Melting point, " C
421 -23 921 1025 38 150 1030 630

Halide

Color

Boiling point,

'c

Heat of 'ormation, kcallmole


310 190.4 110.0 121.5 155 101 102 16

TiF4 TiC14 TiCl, TiC1, TiBr4

Colorless Violet Black Yellow Dark-red Dark-violet Black- brown

284 136

Decomposes below bp
1500 230

-.

311 Decomposes below bp


-1030

Titanium tetrachloride TiCld is the main T i t a n i u m c h 1o r i d e s starting compound for the production of titanium metal. It is a colorless liquid with a density of 1.727. The chloride is produced by the effect of chlorine on titanium dioxide (usually in the presence of carbon) at 700 to 9 0 0 ' . Titanium chloride hydrolyzes in water o r in moist air, with the precipitation of metatitanic acid:
TiCI,

+ 3H,O

H,TiO,

+ 4HCI.

This explains the appearance of a dense white smoke when titanium chloride is brought into contact with humid air. Titanium chloride is a

160

volatile substance which boils without decomposing at 136". The heat of evaporation of TiC1, is - 9 kcal/mole. Two lower chlorides a r e known: TiC13 and TiC12. They a r e produced by reduction of TiC1, with hydrogen, titanium, sodium, and other metals, a t 600 to 800". The lower chlorides tend to disproportionate, i. e . , to decompose into compounds in which the valency of titanium is lower and higher than i t s valency in the starting compound:
2TiCI, 2 TiClz
above

+ Tic!,, 2TiC12ZTiCI, + Ti.


4600

above

4 7 6 0 ~

Both lower chlorides a r e hygroscopic - TiC13 dissolves in water, while TiClz r e a c t s with water with the evolution of hydrogen and the formation of a solution of trivalent titanium. TiClz is oxidized in a i r , sometimes spontaneously bursting into flame. Titanium f o r m s three iodides - Ti14, Ti13 and T it an iu m i odide s TiIz ( s e e Table 29). The higher iodide i s produced by the action of iodine vapors on titanium powder a t 1 0 0 to 200; the lower iodides a r e formed by thermal dissociation of TiI, at 400 to 1100". Above 1100" all the iodides decompose into elemental titanium and iodine. This is one of the methods for the production of high-purity titanium. Titanium fluoride TiF4 is of no importance, T i t a n i u m f 1u o r i d e s but the complex potassium and sodium fluotitanates K2TiF6 and Na,TiF6 a r e used in the production of titanium. They a r e produced by the addition of sodium o r potassium s a l t s to solutions of titanium dioxide in hydro fluoric acid. The solubility of KzTiF6in water at 20" is about 2.2 %, that of NazTiF6 6 . 5 7 0 . Both s a l t s crystallize readily f r o m solutions. Titanium carbides and nitrides (Table 30). Titanium carbide T i c and nitride TiN a r e very hard, refractory substances with some metallic properties (metallic luster, electrical conductivity). They have the s a m e crystalline s t r u c t u r e (face-centered cubic) and form with each other a continuous s e r i e s of solid solutions.

TABLE 30 The properties of titanium carbide and nitride


Co111 pou n d Tic TIN

4.31 4.23

3140

57.25 80.3

2850 2160

Titanium carbide is formed by heating a mixture of titanium dioxide and carbon a t 1900 to 2000": TiOz 3C = TIC 2CO.

The pure nitride may be produced by the action of nitrogen on titanium powder a t 8 0 0 to 1000" o r by heating TiCl., in a s t r e a m of ammonia:
STiCI,

+ 16NH, = 3TiN + 12NH,CI + 1/*Np


161

Uses of titanium Titanium and titanium-based alloys. The interest in titanium, and the resulting development of the titanium industry in 1948 to 1950 were related to the necessity for s t r u c t u r a l m a t e r i a l s combining minimum weight with great strength. The main consumer of such materials is the a i r c r a f t industry, in which the saving of one kg on an engine is equivalent to a saving of 8 to lOkg in the overall weight. A s is evident from Figure 87, titanium-based alloys possess a high specific strength (strength-to-specific gravity ratio). The best modern titanium-based alloys (containing Al, Cr,V, Mo, Sn) have a tensile strength of 1 2 0 to 150 kg/mm2 after heat treatment, which is equivalent to a specific strength of 2 7 to 3 3 . Alloy s t e e l s of the same tensile strength have a specific strength of 15.5 to 19. At room temperature, titanium-based alloys a r e only slightly superior to high-strength alloys based on aluminum and magnesium. However, aluminum alloys rapidly lose their strength at 150 to 430" while titanium alloys conserve it. At these temperatures titanium is superior to stainless s t e e l a s well.

20''

l o o ' <

FIGURE 87. Comparison between the strengths of titanium alloys, steels, and nonferrous m e t a l alloys (per unit weight).

In the a i r c r a f t industry, titanium alloys a r e used for the manufacture of the fuselage, pistons, connecting rods, valves, and jet-engine p a r t s (compressor discs, turbine blades, rotors, etc. ).
of parts

162

The use of titanium in shipbuilding and the automobile and r a i l r o a d - c a r industries is very promising. In addition to t h e i r high strength, titanium and titanium-based alloys p o s s e s s an increased resistance to corrosion. A s a result, they find increasing use in the manufacture of chemical equipment, medical instruments, etc. P u r e titanium (in the f o r m of forged p a r t s , ribbons, and wire), is used in the vacuum-tube industry for the production of anodes, grids, anticathodes, X-ray tubes, and other p a r t s . Titanium powder is used in the s a m e field a s a g e t t e r (gas absorber). At present, the leading countries in the production of titanium and titanium-basedalloys a r e the USA and the USSR. Titanium is aloo commercially produced in Canada, United Kingdom, Japan, and other countries. The titanium production in the USA in 1960 to 1 9 6 2 amounted to about 7000 to 8000 tons p e r year. A s the c o s t of titanium d e c r e a s e s , there will be an i n c r e a s e in the demand and production volume of this metal. n s t e e l s . Because of i t s g r e a t affinity f o r oxygen and Titanium i nitrogen, titanium is used to deoxidize steel and to f r e e i t from dissolved nitrogen (titanium f o r m s with nitrogen a nitride insoluble in the steel). The removal of oxygen and nitrogen from the steel leads to the creation of a fine, dense s t r u c t u r e of improved mechanical properties. In addition to the oxygen and nitrogen, titanium also binds sulfur with the formation of a stable sulfide. This prevents the formation of l a y e r s of the low-melting eutectic F e - F e S on the grain boundaries of the steel, which causes hotbrittleness of the steel. Common brands of s t e e l a r e deoxidized and freed from nitrogen by the addition of 0.03 to 0,2%titanium, in the form of ferrotitanium (an irontitanium alloy containing 18 to 2570 Ti). Titanium is used a s an additive to manganese, chromium, chromiummolybdenum, and chromium-nickel steels. The addition of titanium to chromium-nickel stainless s t e e l s prevents intercrystalline corrosion, which is caused by the precipitation of chromium carbides on the grain boundaries (during welding or tempering of the steel). Titanium in nonferrous alloys. The addition of titanium to copper or to copper- or aluminum-based alloys improves their physical and mechanical properties and their resistance to corrosion. Cuprotitanium ( a copper-titanium alloy containing 6 to 1270 Ti) is usdd for the deoxidation of copper. The strength of aluminum bronze ( a copper-aluminum alloy) is increased by the addi+:?n of 0.5 to 1.5570 Ti. The titanium i s added in the form of "alutite">I:, containing 40% Al, 22 to 5070 Ti, 40% Cu. Titanium in r e f r a c t o r y alloys. Titanium carbide, which is distinguished by i t s high hardness and refractoriness, is a component of hard titaniumtungsten tool steels. These alloys contain 1 0 to 40% T i c , 50 to 85% WC, while the remainder is cobalt. These alloys a r e effectively employed in the production of steel-cutting tools, and a r e of g r e a t economic importance. Titanium carbide is also a component of s c a l e - r e s i s t a n t refractory alloys which a r e used f o r the production of gas-turbine p a r t s f o r jet engines Titanium dioxide. The most important use of titanium dioxide is the production of a white pigment - titanium white. Titanium white has good

_______________

[Russian trade name for a titanium master alloy. 1

163

covering power and is nonpoisonous, which makes it superior to lead whites. Titanium dioxide is used in painting machine p a r t s and ships, is added to the rubber compositions which it is d e s i r e d to color white, is used f o r imparting a dull finish to silk, is added a s an opacifier to paper pulp, and is used in the production of glazes and enamels. The total production of titanium pigments in the Western countries was 400,000 and about 1,000,000 tons in 1955 and 1960 respectively. The high dielectric constants of the various crystalline f o r m s of titanium dioxide (anatase 78, rutile 173 to 180) and of barium titanate is the reason f o r the use of these compounds in the production of solid di e l e c t r i c s for condensers, radio p a r t s , and high-frequency furnaces. Natural titanium dioxide (rutile) or the technical grade dioxide a r e used a s components of the coating of electrical welding electrodes.

39.

MINERALS, ORES, AND ORE CONCENTRATES

Titanium is the tenth element in o r d e r of abundance. Its concentration in the E a r t h ' s c r u s t is 0.61% by weight. It is one of the most abundant metals, following aluminum, iron, and magnesium. Titanium i s encountered in all igneous and sedimentary rocks, as well a s in the soil, peat, coal, many plants, and the blood and bones of animals. The position of titanium in the periodic table explains, to a certain extent, i t s geochemical link to other elements in nature. Figure 88 shows the so-called "geochemical s t a r " of titanium, i n which the a r r o w s indicate the attraction of m H P various elements (which a r e neighbors of m d P titanium in the periodic table) in the o r e towards the titanium. L e s s strong bonds a r e designated by dotted lines. h\ In the E a r t h ' s c r u s t titanium is present Se tt % J --+V as the dioxide or a s titanates - s a l t s of titanic acid. The element tends to form complex anions, comprising niobium, silicon, and Y p zirconium. Thus, in addition to the simple titanates, there a r e complex titanoniobates, m f ! TO silicotitanates, etc. T h e r e a r e about 70 known titanium minerals. The ones of g r e a t e s t commercial importance a r e rutile, ilmenite, perovskite, 77, and sphene, which a r e true titanium minerals. FIGURE 88. The "geochemical R u t i l e , a n a t a s e a n d b r o o k i t e are star" of titanium, according t o allotropic modifications of titanium dioxide. Fersm an. F e r r o u s iron is present a s an impurity. Ofthe three minerals, rutile is of com.mercia1 i m portance. Rutile has adiamond-metallic luster, is transparent, and its color is red-brown or, occasionally, yellowish, bluish, violet, or black. Its density is 4.18 t o 4.28. Large rutile deposits a r e r a r e . Rutile deposits have been found in Canada, Brazil, and Australia. The l a r g e s t amounts of rutile (up to

' 0

I \ " I

164

70 to 80% of the total o r about 70,000 tons) is mined in Australia. Rutile concentrates contain 90 to 95% Ti%. I l m e n i t e (iron metatitanate F e T i Q ) is the most common titanium mineral. It was first found in the Ural mountains, in the llmenskie mountains. The mineral is brown o r black-brown, its density is 4.56 to 5.21. Its chemical formula indicates a T i 4 content of 52.6670. However, some ilmenites contain 60 to 6570TiG. This is attributed to the oxidation of divalent to the trivalent iron as a result of prolonged weathering.

TABLE 3 1 Approximate composition of some ilmenite concentrates. in 70 Chemical composition TiO, FeO Fez03 SiO, 4 O 3 Cr, 0, ME0 MnO CaO zrn,
v25

USSR
Ural
44.0 31.4 16.9 1.8 2.5

Irsha
48.15

Samotkanskii
60.3*

India, Travancore deposits


54.3 26.0 15.5 1.4

Norway

{ 45.9

---

3.3 0.5

2.16 0.12 1.3

1.3 1.1

25.88 1.20 2.0 1.36 0.9 0.85 0.11 0.33

1.1

42.3 33.9 12.9 3.5 1.8

0.85 0.40 0.08 2.18 0.20

1.6 0.85 0.20 0.03 0.40

* The high

TiO, content of Samotkanskii concentrates is attributed to the fact that they contain the transformation production of ilmenite which are enriched in TiO, (leucoxenes).

In addition to iron oxides, ilmenite may also contain ( a s impurities) magnesium and manganese since F e T i Q , MgTiQ and MnTiQ a r e isomorphous. Ilmenite is often closely associated with magnetite (Fe304). Such o r e s a r e known a s t i t a n o m a g n e t i t e s . In addition, there a r e o r e s known a s h e m a t i t o i l m e n i t e s , which consist of mixtures of ilmenite and hematite ( F e 2 Q ) . Up to 4 0 % of the ilmenite is produced from r i v e r and coastal (sea) placers which a r e formed a s a result of the weathering of various ilmenite containing rocks. The l a r g e s t ilmenite-bearing placers a r e found in India, (Travancore). Such p l a c e r s a r e also found on the coast lands of Australia, Indonesia, Africa, South America, and the USA.. In the USSR, titano magnetite-bearing sands a r e found in the a r e a s of Dnepr-Donets and the Black Sea depressions. The p l a c e r s now being worked contain from one to s e v e r a l dozen kilograms of ilmenite p e r cubic m e t e r of sand. The sands are often of a complex composition. Thus, sands from Samotkanskii Si02) and contain rutile, ilmenite, zircon, as well a s disthene ( A l & staurolite [Fe(O m 2 2A1Si05]which a r e valuable s o u r c e s of aluminum 1271 Titanomagnetite deposits a r e an important source of ilmenite. The largest such deposits a r e found in Canada, the USSR, the Scandinavian countries and Brazil. P e r o v s k i t e - calcium titanate, C a T i Q - contains 58.7% T i Q and 41.3% CaO. In some deposits it also contains niobium, yttrium, manganese,

165

and magnesium as impurities. Its color is black, red-brown, red, or yellow. Its density is 3.95 to 4.04. Large deposits of perovskite o r e s have been found in the USSR. The beneficiation of these o r e s is simple. With increasing consumption of titanium raw materials, perovskite o r e s will become one of the most important s o u r c e s of titanium. Sp h e n e o r t i t a n i t e (calcium titanosilicate CaO. Ti@ S i Q ) contains 38.870 Ti&. A fraction of the CaO may be replaced by F e O and MnO. The m i n e r a l has a yellow color, its density is 3.4 to 3.56. Sphene deposits have been found in the USA, Canada, Madagascar, and in many regions in the USSR. The titanium content in sphene is lower than in ilmenite, rutile o r pervoskite. However, i t s utilization may be profitable under certain conditions (combined utilization with other o r e s - apatite and nepheline). In addition to the raw m a t e r i a l s listed above, complex minerals may also be used a s a source of titanium. Loparite and some other titaniumniobium minerals belong to this class. The T i 4 content in the o r e s processed v a r i e s within wide limits - from 6 to 357'0. In the beneficiation of ilmenite sands, the heavy minerals, the so-called schlichs (magnetite, ilmenite, rutile, zircon, and others), a r e extracted f i r s t , by gravity methods (wet jigging, screw-type separators, tables). The main methods used for the separation of the schlichs a r e electrostatic and electromagnetic. If the magnetic permeability of iron i s assigned a value of 100, that of magnetite would be 40.2, of ilmenite 24.7, of rutile 0.4, and of silicates l e s s than 0.2. Magnetic separation is used to separate magnetite from ilmenite, and ilmenite f r o m rutile and nonmagnetic heavy minerals. In the case of the collective concentrate of heavy minerals obtained from the sands of the Samotkanskii deposit, electrostatic separationin conjunction with magnetic separation is successfully employed to produce rutile concentrates (containingup to 957'0 TiO,), ilmenite concentrates (containing60 to 6270 TiO,) and zircon concentrates (up to 67.5% ZrG) /27/. Flotation methods have been recently developed f o r the extraction of ilmenite. Magnetic separation is used with success for the beneficiation of titanomagnetite o r e s impregnated with c o a r s e ilmenite particles (in the By using weak magnetic form of c r y s t a l s with a size of 0.1 to 0.2"). fields an iron concentrate (magnetite), an ilmenite concentrate (the weakly: magnetic fraction), and tailings (the nonmagnetic fraction) a r e obtained. The approximate composition of ilmenite concentrations is shown in Table 3 1. Some titanomagnetites cannot be enriched by mechanical processes, a s they consist of a solid solution of TiO, in Fe304. Metallurgical beneficiation - smelting in electrical furnaces with the production of pig iron and Tirich slags - may be used in such cases.

40. PRODUCTS O F THE PROCESSING OF TITANIUM CONCENTRATES

Three types of products a r e obtained directly from the titanium concentrates : titanium tetrachloride, titanium dioxide, and ferrotitanium. T i t a n i u m t e t r a c h l o r i d e is the main s t a r t i n g m a t e r i a l in the production of metallic titanium. High-purity tetrachloride is required for
1455

166

.. .

this purpose. Hence, the technical grade titanium chloride obtained in the chlorination of titanium raw m a t e r i a l s is subjected t o further purifica tion. The purified titanium chloride is a transparent, c o l o r l e s s o r light y e l l o q liquid. T i t a n i u m d i o x i d e is produced in different g r a d e s of purity and s t r u c t u r e , depending on its intended use.

1
. I

n m e n i t e concentrate

I
J-

Sulfuric acid method for

Reductive smelting

Pig iron By-product FeSO, * IHZO Chlorine

7 Chlorination

1
Technical TiC1, Purification

Hydrolysis or "burning" T o the production of titanium FIGURE 89. General flow sheet of the production of titanium tetiachloride and titanium dioxide from ilmenite concentrates.

Pigment grade titanium dioxide (titanium white) contains from 94 to 98.570 TiO, and some oxide admixtures (ZnO, A l z Q , Si& and occasionally Sbz03), which a r e introduced in o r d e r to obtain the required s t r u c t u r e and physico chemical properties. Some pigment brands have the s t r u c t u r e of rutile, others of anatase. The particle s i z e of pigment-grade titanium dioxide must be < ID. The specifications for titanium dioxide used in metallurgy a r e only concerned with i t s chemical purity; i t s s t r u c t u r e is of no importance. The p u r e s t titanium dioxide (contahing a t l e a s t 95.570 Ti-) is used in the production of carbide-type hard alloys and for the production of titanium metal. Dioxide containing at l e a s t 9970 Ti& is used f o r the production of alloys with nonferrous m e t a l s (nickel, copper, aluminum), while the dioxide containing a t l e a s t 97.570 Ti- is used for the production of weldingelectrode coatings.

167

F e r r o t i t a n i u m is produced from ilmenite concentrates by thermal reduction with aluminum in electrical furnaces. The alloys contain 25 to 30% Ti, 5 to 870Al, 3 to 4.570 Si; the remainder is iron. The industrial methods f o r the production of titanium tetrachlor' e and titanium dioxide from the main raw material, ilmenite concentrates, are described below. A flow sheet of the processing of ilmenite concentrates is shown in Figure 89. It will be seen that titanium chloride is produced by chlorination of titanium slags (75 to 8570Ti%) formed in the smelting of ilmenite concentrates. Two methods a r e used f o r the production of titanium dioxide: a) direct decomposition of ilmenite concentrates (or titanium slags) with sulfuric acid followed by precipitation of metatitanic acid from the sulfate solution; b) hydrolytic decomposition (or "burning") of titanium tetrachloride.

41. REDUCTIVE SMELTING OF ILMENITE /5, 15, 25, 28/

Ilmenite concentrates contain 42 to 60% Ti& and 40 to 48'70 F e O + F e 2 Q (see Table 3 1 ). The high concentration of iron interferes with the production of titanium chloride f r o m ilmenite. A l a r g e amount of chlorine is consumed in the formation of iron chloride during direct chlorination of the concentrates. Moreover, there a r e difficulties in finding use for the large amounts of the iron chloride formed. The iron is separated by selective reductive smelting of ilmenite, with the formation of pig iron and titanium slags (80 to 87% Ti0.J. In this process, the separation of titanium and iron is based on the large difference between the stabilities of their oxides. Titanium slags have a high melting point (above 1500") and a considerable viscosity. For this reason, iron o r e s with a high titanium content ( e . g . , titanomagnetites) a r e usually not smelted in blast furnaces. However, electric a r c furnaces may be used f o r the reductive smelting of ilmenite concentrates. The following main reactions take place during the reduction of ilmenite by carbon at various temperature ranges:
FeTi03+ C = Fe + TiO, CO up to 1240" 3Ti0, C = Ti30, CO at 1270 to 1400" 2Ti30,+C=3Ti,03+C0 Ti,03+C=2Ti0+C0 at 1400 to 1600"

The intermediate oxide Ti305and the oxide Ti2% may dissolve the f e r r o u s oxide and the ilmenite, and form solutions with Ti02 and the monoxide TiO. These reactions lead to the formation of complex compounds which crystallize when the slags solidify. The most important of these substances is a n o s o v i t e , whose composition can be expressed by the The formula general formula m[(Mg, Fe, Ti)O 2TiO&z[(Al,Fe, Ti)203.Ti02]. shows that titanium exists in anosovite in various valencies - Tik, Ti3+, and Ti2+.
168

In addition to anosovite, the lower titanium oxides and the solid solution of ilmenite in Ti203, the slags may also contain titanium oxycarbonitride Ti(C, 0, N). This is a solid solution of titanium carbide, nitride, and monoxide (Tic-TiN-TiO). The oxycarbonitride i s formed a t tempera t u r e s above 1600" in the presence of an e x c e s s of carbon. The lower oxides, and especially titanium oxy carbonitride, cause an i n c r e a s e in the melting point and the viscosity of the slags. The melting point and viscosity of the slags may be reduced by the addition of fluxes such a s calcium, magnesium, and aluminum oxides. However, this r e s u l t s in a .~ ." d e c r e a s e of the TiOz content of the FIGURE 90. Diagram of an electric ar c furnace s l a g s and causes increased consump for the smelting of slags. tion of chlorine, because of the formation of calcium, magnesium, 1 - electrodes ( t he third electrode fs not shown): 2 -charge; 3-steel jacket: 4-asbestos-cement and aluminum chlorides. Hence, the sheet: 5 --magnesite grit; 6 --magnesite re tendency i s to c a r r y out the smelting fractory bricks; I -tapping hole; 8-pig iron; with the addition of a s little flux a s 9 -slag hardened on the walls: 10-slag. possible, o r altogether without it. The reductive smelting i s c a r r i e d out in three-phase electric a r c furnaces (5000 to 10,000 kVa), s i m i l a r to the ones used in the production of ferroalloys such a s ferrochromium. A diagram of a 5000 kVa furnace is shown in Figure 90 / 251. The bath is a cylindrical jacket lined with magnesite r e f r a c t o r i e s . The bath walls have descending steps. The tapping hole is a t a height of 310" above the lower point of the hearth. In the working zone of the furnace the walls have a thickness of 690" and the hearth 1380 mm. Above the bath there a r e three graphitized electrodes (500" in diameter) which a r e held in cooled clamps and a r e connected to a displacing device. A l a y e r of slag (about 0.5m thick) is allowed to harden preliminarily on the lining of the furnace in o r d e r to protect i t against rapid attack by the corrosive titanium slags. A thick l a y e r of pig iron ( 0 . 5 m o r more) i s always left on the hearth in o r d e r to prevent attack by the slag. The smelting i s c a r r i e d out with a covered charge hole, i. e . , with the a r c covered by the charge (the electrodes a r e immersed deep inside the charge). Coke or anthracite, crushed to a particle s i z e of 3 to 4 m m , is used a s the reducing agent. The components of the charge (concentrate, coal) p a s s from the storage bins through metering trolleys into three furnace bins which a r e positioned above the furnace. F r o m the bins the charge on moving trays i s charged into the space between the electrodes and along the furnace walls. For the best utilization of the reducing agent and in o r d e r to reduce dust formation and the consumption of electrical energy, it is most convenient to use a pelletized charge. However, this causes difficulties because of the sintering of the pellets and their cementation by the boiling melt, which reduces the gas-permeability of the charge. A s a result, in industrial practice the smelting is c a r r i e d out on a mixed charge, i. e . , a

_-

169

111

111

11.

I1

III1111111111111

I I

11111

I 1

II.

111111II111111111111

1111

mixture of pelletized and pulverized charge. The ratio of pelletized to pulverized charge depends on the type of titanium concentrate taken. The pulverized charge content v a r i e s between 20 and 50%. The pellets are prepared in r o l l e r p r e s s e s , using sulfite-pulping liquor as the binder. Various smelting conditions a r e used. One of these, used in a batch p r o c e s s , is described below 15, 251. The smelting cycle is divided into three stages: fusion of the charge, rapid reduction, and finishing the s l a g s before tapping. The amount of carbon i s calculated to produce a slag with a f e r r o u s oxide content of 5 to 6 % . Slags with a lower F e O content a r e fused with difficulty without the addition of fluxes, a s they have a high melting point and rapidly solidify. This is attributed to the formation in these slags of titanium carbide o r oxycarbide. In slags with a sufficiently high FeO content the oxycarbide r e a c t s with it according to the reaction: Tic. Ti0 3Fe0 = Ti,O, CO + 3Fe.

The smelting i s facilitated by adding to the charge at f i r s t only a fraction of the carbon, calculated to produce a low-melting s l a g (up to 20% FeO). The r e s t of the carbon is then added gradually, and the reduction is continued in the liquid phase until the F e O content reaches 5 to 6 % . The l a s t stage of the smelting - the finishing - has the purpose of melting the s l a g frozen on the walls and the crowns formed by it. In this stage the electrodes a r e withdrawn and the furnace operates with an exposed arc. The s l a g s and the pig iron a r e tapped periodically through the tapping hole into a common pan lined with chamotte bricks. The temperature of the s l a g during the tapping is 1570 to 1650". The pig iron and the s l a g s a r e allowed to s e p a r a t e into l a y e r s , to solidify, and a r e then separated. The smelting i s c a r r i e d out a t a voltage of 130 to 135 V. The total time f o r the smelting of a charge of 1 2 tons is about 4 hours. The consumption of electrical energy is about 3000 kwh p e r ton of slags. The smelting of I r s h a ilmenite concentrates ( s e e Table 31) by the above method yielded slags with the following composition: 85.770 T i Q , 6.570 FeO, 5.65% Si&, 1.9% AlzQ, 3.070 MgO, and 0.75% CaO. The degree of extraction of titanium from the concentrate into the slags was about 95% 151. 42. REDUCTION OF ILMENITE TO TITANIUM CARBIDE (THE CARBIDIZATION PROCESS)
When a mixture of ilmenite concentrate and carbon is heated t o 1900-2000' yielding titanium carbide and iron (pig iron): FeTiO, 4C -+ Tic Fe XO. the ilmenite is reduced

+ +

As the pores of the charge contain some nitrogen, the product is, in fact, not the pure carbide but the

carbonitride Ti(C,N) (a solid solution of titanium carbide and nitride) or the oxycarbonitride Ti(Ci,N,O). T h e formation of such products is associated with the fact that the carbide T i c , the nitride TiN, and the monoxide T i 0 have crystal lattices of the same type with close values of lattice parameters. These compounds form a continuous series of solid solutions. After the cake has been ground, most of the iron is extracted from titanium carbide by electromagnetic separation. Titanium carbide (or, more accurately, titanium oxycarbonitride) obtained is forwarded to chlorination.

170

T h e reaction of the carbide (or oxycarbonitride) with chlorine is a rapid reaction taking place a t a relatively low temperature (300-400") and is accompanied by the evolution of a considerable amount of heat. which is sufficient t o support the Chlorination process. This is the main advantage of the chlorination of titanium carbide over t h e chlorination of rutile or titanium slags.

/4

/3

/'
b

a FIGURE 91. Diagram of a core furnace:

a-longitudinal section; b -cross section: 1-refractory brick lining; 2 -graphite electrodes; supplying the current to the core: 3-granulated coke core; 4 -charge positioned around the core; 5-reacted part of the charge. It has been reported that in some plants the carbidization of ilmenite is a stage preliminary to the chlorination. According to o n e report, the carbidization is carried out i n resistance furnaces resembling the ones used for the production of silicon carbide (carborundum). Such furnace is shown schematically in Figure 91. It is pan-shaped, and its end walls are connected to current terminal electrodes. A core consisting of coke particles, which serves as the heating e l e m e n t , is positioned between the electrodes. T h e charge is poured around the core. T h e charge layers adjacent to the core are heated to the required temperature (1900 to 2000"). T h e temperature decreases with increasing distance from the core, and no carbidization rakes place in the outer layers of the charge adjacent to the walls. These layers serve as thermal insulation, and are returned to be mixed with the fresh charge. T h e consumption of electrical energy is about 5000 kwh per ton of cake. After grinding to a particle size of -0.30 m m and magnetic separation, oxycarbonitride of the following composition is obtained: 68.1% T i , 2% Fe, 9.3'70 C . and 0.470 N.

43. PRODUCTION OF TITANIUM TETRACHLORIDE


Reaction mechanism Titanium dioxide r e a c t s with chlorine; the reaction is endothermic:
Ti0,+2Cll =TiCI,+O,-45

kcal.

This reaction is v e r y slow even a t 800 to 1000". A reaction r a t e high enough to be of practical importance is obtained a t 700 to 900" in the presence of a reducing agent - carbon. The carbon combines with the oxygen, yielding CO, CO, and some phosgene C0Cl2, and all the chlorina tion reactions a r e exothermic:
TiOl 2C12 C TiCI, COz 49 kcal; TiOa + 2C1, 2C ZTiCI, -+ 2CO + 78 kcal; TiO, 4C1, 2C TIC], f2COC1, 62 kcal.

+ + z + + +

(1)
( 2) (3)

The equilibrium p r e s s u r e of phosgene (COC12) at 600 to 800" is extremely low, and reactions (1) and (2) a r e predominant (see Table 32). The relative amount of CO in'the gas mixture i n c r e a s e s with increasing temperature. At 800" and above the chlorination (under equilibrium

171

conditions) should proceed mainly according to reaction (2), i. e . , with the formation of carbon monoxide. Since reaction (2) yields two molecules of CO while reaction (1) yields one molecule of CQ, increasing the chlorina tion temperature i n c r e a s e s the consumption of carbon p e r mole of TiC1, and increases the total volume of the gaseous phase (and thus reduces the TiC1, concentration in the g a s mixture).
TABLE 32 Equilibrium composition of the gaseous phase i n t h e chlorination of titanium dioxide by chlorine in t h e presence of carbon (According to t o d n e v and Pamfilov) TemperaPartial pressure. (equilibrium), a t m .

0.600

0.370 0.047

4.9.10-7

In practice, no equilibrium is reached in the chlorination. F o r this reason, even a t 900" the gases contain CO,. The C O :C & ratio in the g a s mixture depends on the method and conditions of chlorination.

3 0 0

400 500 600 Temperarure. C

FIGURE 92. Degree of chlorination (by chlorine) after one hour at various tempera tures.
1 -TiO;

2 - T i O+ C ;

3-Ti0,

+C.

A t any given temperature, the r a t e of chlorination of titanium slags is higher than that of rutile concentrates. This is attributed to the fact that titanium salgs contain lower titanium oxides and occasionally
titanium oxycarbide, whose reaction with chlorine is m o r e vigorous than
the reaction between chlorine and titanium dioxide. Thus, titanium
monoxide r e a c t s with chlorine at an appreciable r a t e at 300" in the absence
of carbon, according to the equation 1171

2Ti0 + 2C1, = TiCI,

+ Ti02.

At 500, in the presence of carbon, the chlorination of T i 0 is 2 to 2.5 times f a s t e r than that of Ti% (Figure 92). This is explained by the fact
172

that the active titanium dioxide formed in the reaction is rapidly chlorinated in the presence of carbon. It is probable that the formation of active titanium dioxide is the f i r s t stage in the chlorination of T i 2 Q and Ti305 as well. In addition to titanium oxides, the m a t e r i a l to be chlorinated contains various amounts of the oxides of iron, manganese, calcium, magnesium, aluminum, silicon, vanadium, and some other elements. Of these elements, Fe, Al,Si, V, Cr, Ta, and Nb form volatile chlorides ( s e e p. 177). F r e e silicon r e a c t s slowly with chlorine in the presence of carbon a t 800 to goOD, but the s i l i c a t e s a r e chlorinated rapidly.

Procedures T h r e e methods of chlorination a r e used a t present in industry: chlorination of pelletized charge in a stationary layer, chlorination in molten s a l t s , and chlorination in a fluidized bed. Chlorination in a stationary layer. The p r o c e s s is c a r r i e d out in electrical shaft furnaces. The chlorination is preceded by preliminary operations - grinding of charge components, mixing, briquetting, and coking of the briquet.tes 11, 3 0 1 . The amount of carbon added to the charge depends on the composition of the m a t e r i a l to be chlorinated and the p r o c e s s temperature. Assuming that at 800 to 900" the chlorination proceeds with the preferential formation of CO, the theoretical consumption of carbon by a charge containing 8070 Ti02 is 24 kg p e r 100kg slag. This does not allow f o r the fact that a p a r t of the titanium in the slags is in the form of lower oxides. In practice, the charge contains 2 0 to 2570of pulverized petroleum coke, Sulfite lyes::;, coal pitch, and other m a t e r i a l s a r e used a s binders to produce strong pellets. The briquetting technology is the s a m e a s for niobium ( s e e Chapter III); the type of shaft furnace used for the chlorination and i t s operation a r e also the s a m e ( s e e Figure 56). The stoichiometric amount of chlorine needed for the production of one ton of TiC14 is 0.75 ton. In practice, 0.85 to 0.90 ton of chlorine is spent in the chlorination of slags p e r ton of chloride, because of the chlorine lost through the formation of chlorides of other elements. Sodium, calcium, magnesium and manganese chlorides, which have high boiling points, f o r m a liquid melt which flows downwards through the p o r e s of the charge and the coal packing on the hearth and is periodically tapped into the ladle. The vapor-gas mixture, containing carbon monoxide and dioxide, TiC14, volatile chloride impurities ( FeCl,, AlCl,, SiC14, V0Cl3, CrQC12, e t c . ) and a s m a l l amount of e x c e s s chlorine is vented through a tube in the upper p a r t of the furnace into a dust-.collecting and condensation system.

* These a r e t h e waste liquors of cellulose pulping.


lyes have binder propertes.

They contain organic substances.

T h e concentrated

173

The specific output of chlorinators with a stationary layer of pelletized charge (in the chlorination of titanium slags) is about 7 0 to 100 kg TiC14 p e r s q u a r e m e t e r of furnace c r o s s section p e r hour. The degree of extraction of titanium f r o m the slags to form the technical grade TiCll is about 83 to 85%. The CO : C Q weight r a t i o in the vapor-gas mixture at the exit from the shaft furnace ranges from 5 : 1 t o 8 : 1 depending on the temperature conditions and theheight of the charge layer. In the presence of an inflow of a i r , the gas mixture becomes explosive. To prevent the inflow of a i r , an over p r e s s u r e of 3 to 5 m m H g i s maintained a t the outlet of the shaft furnace. In shaft furnaces with a productivity of 25 tons TiCI4 p e r day, most of the heat (62.070) comes from the chlorination reaction. The heat furnished by the electrical elements amounts to 38% , corresponding to a consumption of 7.84 kwh p e r ton TiC14. The vapor-gas mixture entrains 66.2% of the heat from the furnace. Heat l o s s e s through the lining and the jacket of the furnace account f o r 33.8% of the heat, which i s nearly equal to the amount supplied by electrical heating / 3 2 1 . Chlorination in the melt 1301. This method f o r the chlorination of titanium s l a g s was proposed in the USSR by an engineer, Solyakov. The chlorination i s c a r r i e d out in a bath of molten potassium and sodium chlorides, in which the charge ( a mixture of ground slag and coke) is introduced. The charge may be introduced through a charging tube in the roof of the chlorinator, with the aid of dry compressed a i r o r nitrogen. The chlorine i s fed to FIGURE 93. Diagram of a three-chamber the lower p a r t of the chlorinator and is fluidized-bed chlorinator. dispersed through a grid fitted in the interior, or by means of special nozzles. 1-stack: 2-overflow, for the passage of At 700 to 800" the chlorination in the material from higher t o lower chamber; 3 -dispersing grating; 4-fluidized bed; melt proceeds rapidly. The titanium 5 -storage bin; 6 -collector for residues chloride together with the volatile chloride (nonchlorinated material and nonvolatile impurities (SiC14,AlC13, FeCl3) p a s s e s chlorides). into a condensation system, while the nonvolatile chlorides (MgC12,CaC12,etc. ) . - -. -. remain in the melt. After considerable amounts of the nonvolatile chlorides have accumulated, there i s a change in the properties of the , melt, and especially in i t s viscosity, which i n t e r f e r e s with its saturation by chlorine. For this reason, the melt is periodically tapped out of the chlorinator and a f r e s h batch of the molten s a l t s is added.

174

In the chlorination, the required melt temperature is maintained by the heat evolved in the exothermic reactions. The excess heat is removed by means of cooled ducts (made of graphite plates) mounted on the walls of the chlorinator. A s compared with the chlorination of pelletized charges in electrical shaft furnaces, chlorination in m e l t s has the advantages of not requiring a pelletizing operation, of a higher output of the chlorination apparatus, and of a lower dust content in the vapor-gas mixture. Anodic chlorine (the gas from the electrolytic production of magnesium, 2570 Nz, and 6% consisting of a mixture of chlorine and a i r - 7070 Clz, &) may be used for the chlorination, instead of pure chlorine. This reduces considerably the cost of titanium chloride. Chlorination in a fluidized bed 1 3 0 1 . Short communications have been published on the chlorination of titanium slags or rutile concentrates in a fluidized bed. Some idea of fluidized bed p r o c e s s e s was given in the chapter "Molybdenum". Fluidized-bed chlorination equipment has the important advantage of permitting a continuous p r o c e s s to be c a r r i e d out; the output is high. The chlorinators a r e cylindrical stacks lined with dense Dinas bricks; in the lower p a r t of the stack there is a grating (hearth) serving to d i s p e r s e the chlorine fed to the apparatus. The charge, which i s a mixture of crushed slag (or rutile) and coke is fed into the fluidized bed with the aid of an air-tight feeder o r with the gas s t r e a m passed over the surface of the grating. F o r best utilization of chlorine, it i s advisable to use a chlorinator consisting of s e v e r a l chambers placed one over the other (i. e . , to mount s e v e r a l gratings in the stack, with a "fluidized bed" of the m a t e r i a l on each of them). The f r e s h charge i s fed to the upper chamber, in which it is partially chlorinated by the chlorine coming from below, and from there i t p a s s e s through an overflow tube into the chamber beneath (Figure 93). It is very important to select the optimum particle s i z e of the slag ( o r rutile) to be chlorinated and of the coke, in o r d e r to prevent preferen tial entrainment of one of the components from the charge bed. In the chlorination of titanium slags, difficulties a r e caused by the presence of unusually high amounts of calcium, magnesium and sodium oxides which yield relatively low-melting nonvolatile chlorides (CaC12, MgC12, NaC1); these collect in the fluidized bed and may cause sintering of the particles and settling of the bed. In such c a s e s , the working temperature of the l a y e r must be maintained a t 600", which reduces the output of the installation. The chlorination of rutile concentrates may be c a r r i e d out a t 900 to 1000". The output of fluidized-bed chlorinators ranges from 200 to 500 kg TiC1, p e r square m e t e r of hearth p e r hours, depending on the chlorination temperature; this is much higher than the output of electrical shaft furnaces.

Dust collection and condensation system The dust collection and condensation system usually consists of d u s t c h a m b e r s (serving to s e p a r a t e the dust particles mechanically entrained from the furnace, and to settle the iron chloride), c o n d e n s e r s

175

for the TiC14, and s a n i t a t i o n s c r u b b e r s , f o r the absorption of the excess chlorine. A d i a g r a m of the condensation system was shown in Chapter I11 (Figure 57). The bulk of the dust and the iron chloride a r e trapped in a dust chamber, with walls cooled by a s t r e a m of a i r passed through the jacket. The gas temperature at the exit of the dust chamber is maintained a t 160 to 180". The vapor-gas mixture coming out of the dust chamber s t i l l contains a substantial amount of dust. A.ccordingly, the condensation of titanium chloride i s c a r r i e d out m o s t conveniently in s p r a y condensers, in which cooled titanium tetrachloride is used a s the s p r a y liquid ( s e e Figure 57). The s l u r r y from the irrigated condenser p a s s e s through an intermediate collector into a thickener f o r the separation of s l i m e s . P a r t of the chloride is withdrawn from the intermediate collector and is used for irrigation. The clarified chloride from the thickener flows into collector tank f o r technical grade chloride. The thickened pulp is sent to an evaporator, to remove titanium tetrachloride / 3 0 1 . The g a s e s issuing from the s p r a y condenser s t i l l contain s o m e titanium chloride. It may be trapped in tubular condensers cooled by brine (CaC12 solution) a t -10 to 15'. F r o m there the g a s e s p a s s through a sanitation scrubber sprayed with l i m e water (for the absorption of chlorine and hydrochloric acid vapors) and a r e vented to the atmosphere. The pipelines of the condensation system may be occasionally clogged by a mixture of iron chloride, titanium oxychloride and dioxide. The l a s t two compounds a r e formed in the reaction of TiC14 with water vapor:
TiCI, t :H,O TiOCI, TiCI, 2H2O -+ TiOz
--f

+ 2HC1, + 4HC1.

Hydrochloric acid is also formed in these reactions, with resultant corrosion of s t e e l condensers and pipelines. The moisture may enter the furnace o r the condensation system together with the a i r o r with the pellets if they a r e fed to the furnace while cold and with moisture adsorbed on their surface. The condensation system may be simplified i f the iron chloride and the other solid chloride a r e previously separated by passing the vapor-gas mixture issuing from the dust chamber through a column packed with granular sodium chloride, in which the temperature is maintained a t 350 to 450". The iron and aluminum chlorides form with NaCl eutectic mixtures of low melting point which flow downwards and a r e periodically tapped 1311. The solid particles may also be separated from the vapor-gas mixture by installing fiber-glass bag filters, in which the temperature is maintained a t 200 to 250". The total d e g r e e of extraction of the titanium with the technical chloride in the chlorination of titanium slags is 90 to 9370.

44. PURIFICATION OF TECHNICAL GRADE

TITANIUM CHLORIDE Technical titanium tetrachloride contains a number of impurities, both dissolved and a s a fine mechanical suspension.
176

Gases (N, C1) and s e v e r a l chlorides S i c 4 , SnC14, VOCIs, CrQC12, AICI3, titanium oxychloride TiOC12,NbC15, and TaC15 are present in a dissolved state. The main impurities suspended in titanium chloride are iron chloride, titanium dioxide, and fine carbon particles. The solubility of FeC13 in TiCll is very low: 0.04370 at 110" and 0.00370 at 40". The solubility of A1C13, TaC15, and NbC15 is substantial, as is evident from Table 33.
TABLE 33 Solubility i n titanium tetrachloride,%/ 19/ Temperature. 'C TaCIS
0.40

NbC15
0.45 2.4

I
I

AlClS
0.26
1.1 3.8

18
55 80 106

10.0

12.2
I

Silicon tetrachloride and titanium chloride form a continuous s e r i e s of liquid solutions (Figure 94). The solubility of titanium oxychloride TiOClz in TiC14 at various temperatures is shown in Figure 95.

O W 70 W W I N fill, 1 0 20 30 40 S SiC1, , mol. 7 0


FIGURE 94. TiCl,-SiCl,. Phase diagram of the system FIGURE 95. Solubility of titanium oxychloride TiOC1, in TiC1,. as a function of the temperature.

The mechanical impurities a r e separated from the titanium chloride by filtration through porous earthenware f i l t e r s o r through a l a y e r of activated carbon. In the l a t t e r case, p a r t of the dissolved chlorine i s also adsorbed from the solution. After filtration, technical grade titanium chloride has the following composition: 9 7 to 99% TiCI4, 0.2 to 0.570 Si, 0.02 to 0.0020Jo Fe, 0.1 to 0.03% C1, 0.06 to 0.2% V, and some Al, Cr, and other elements. Most of the impurities a r e removed from the chlorides by the rectifica tion method, which is based on the different boiling points of the chlorides. The boiling points of the chlorides are tabulated below (in "C) :

177

...... ...... ...... CrOaC12. . . . . . VCI, . . . . . . . FeC18. . . . . . .


T i C I , VOCI, S i C I ,

136

TaCI,.

127 NbCI, 58 NbOCI, 116,7 164 Al2C1, 319 S n C l , .

...... ...... .....

239
254 400 (sublimates) 180

...... ......

113

Some impurities, e. g., vanadium oxychloride, a r e difficult to remove by rectification since VOC13 and TiC14 have close boiling points. Therefore, VOCls is preliminarily reduced to the less volatile lower chloride. Copper powder is used a s the reducing agent. When titanium tetrachloride i s mixed with copper powder, the oxychloride is reduced in the reaction:

VOCI, f c u

VOCI, + CUCI.

Some VO, is also formed. It is most probable that copper powder a l s o reduces the chromium oxychloride and tin chloride (present a s impurities) to the lower chlorides. The presence of aluminum chloride (which i s v e r y corrosive) in titanium chloride before rectification is undesirable. To separate aluminum chloride, a s m a l l amount of water equal to the amount required for the formation of aluminum oxychloride: A1C13+HzO + AlOCl +HC1, is added to the titanium chloride. The precipitate formed (AIOCI) is separated by filtration. In some plants, vanadium is removed by reduction with hydrogen sulfide f I / . A s a result, vanadium is precipitated a s the sulfide. In one procedure, H2S is passed for 4 to 6 hours through a vat containing 2000 kg of TiC14 a t 90". Hydrogen sulfide is fed in slowly (at a r a t e of 0.45 kg/hr) f r o m stee1 cylinders, as i t s solubility in the titanium t e t r a chloride is low. The precipitates formed readily settle and a r e easily separated by filtration. Purification by rectification. Rectification is a p r o c e s s in which the components of a homogeneous liquid mixture a r e separated in columns; it is based on the interaction of countercurrent flows of vapors and liquid (formed by the condensation of the vapor) along the column. We shall discuss h e r e the rectification p r o c e s s in the c a s e of a mixture consisting of two components, one with a lower boiling point (LB) and the other with a h i g h e r boiling point (HB). Let u s visualize a packed column, with the vapor mixture from the pot entering the bottom of the column and a liquid consisting of the almost pure LB fed to the upper part. A s the vapors come into contact with the descending liquid they t r a n s f e r to it a p a r t of their latent heat and a fraction of the vapors condenses, while a fraction of the liquid is vaporized. The HB condenses perferentially from the vapor phase while the L B evaporates preferentially f r o m the liquid. Hence, the descending liquid becomes enriched with HB while the ascending vapor becomes enriched with LB. As a result, the vapors a t the i s s u e from the column consist of almost pure LB. The LB component e n t e r s a condenser (the so-called dephlegmator). A fraction of the condensate is used to i r r i g a t e the column (this fraction is known as the phlegm); the r e s t is collected in the distillate collector. The HB remains in the pot. In o r d e r to obtain a satisfactory separation of the components in the column, it is n e c e s s a r y to provide for the l a r g e s t possible contact a r e a

178

between the vapors and the descending phlegm. This is achieved by packing the column, or by mounting a number of bubble-cap, sieve, or meshed t r a y s along i t s height. Each t r a y is covered by a layer of liquid; the vapors bubble through the opening in the column and through the liquid on it. P a r t of the liquid flows down through the openings ( o r overflow tubes) of one t r a y onto the t r a y beneath it. The rectification may be c a r r i e d out either batchwise or continuously. A diagram of a continuous rectification column is shown in Figure 96. Continuous rectification columns have two sections a lower (stripping) and an upper (rectify ing) section. The starting mixture is fed to the upper p a r t of the stripping U section. T h e r e it comes into contact with the vapors whose initial composition is FIGURE 96. Diagra'm of a continuous rectific.i:ion installation. s i m i l a r to that of the pot residue (HB). A s a result the L B is extracted (stripped) 1 -pressure tank with TiCI,; 2-heater; from the mixture. In the upper section 3 -stripping section of the rectification column: 4-rectifying section of the the vapors coming from the stripping column; 5-dephlegmator; 6-boiler; section of the column come into contact I-condenser; 8-pot residue with the descending liquid (the phlegm) collector; 9 -distillate collector; whose initial composition is the same 10-flow-rate adjustment valve for the a s that of the distillate (pure LB). A s a distillate. result the vapors a r e enriched (rectified) with LB. The vapors used to feed the column a r e produced by evapora - repeated tion of the pot residue (HB) in the evaporator, while the phlegm is produced by condensation of the vapors in the reflux condenser. The ratio of the amount of phlegm (the distillate returned to i r r i g a t e the column) to that of the distillate taken out of the column is known a s the reflux ratio. It indicates how many kg-moles of distillate m u s t be returned to the column in o r d e r to collect one kg-mole of distillate. The purification of titanium tetrachloride by rectification is c a r r i e d out in stainless s t e e l columns with perforated t r a y s (or t r a y s with slitshaped perforations) in two stages. In the f i r s t stage silicon tetrachloride is removed by maintaining a temperature of 58" in the upper p a r t of the column (the boiling point of SiC14). The pot residue (Tic14 containing the high-boiling chlorides and oxychlorides as impurities) e n t e r s the second column in the top section of which a temperature of 136" (the boiling point of TiC14) is maintained. The distillate consists of pure titanium chloride. Impurities such a s chlorides with a higher boiling point (AlOCl, FeC13, NbC15, e t c . , a s well a s titanium oxychloride TiOC12) a r e separated out. Purified titanium chloride is a transparent, c o l o r l e s s or slightly yellowish liquid in which the impurity content is close to the sensitivity limit of spectroscopic analysis. The concentrations of the most common impurities such as Al, V, C r , Cu, Si, Mn, Ta, N b , and Zr range from lo-' to 10-37~ (of each impurity).

Jrl

--

179

The degree of extraction of titanium (yield of purified chloride from the technical tetrachloride) purified chloride is 9570.

45.

PRODUCTION OF TITANIUM DIOXIDE The sulfuric acid method / 4 /

The bulk of the titanium dioxide produced from ilmenite concentrates is prepared by the sulfuric acid method which c o m p r i s e s the following s tages : 1) decomposition of the concentrate with sulfuric acid; 2) removal of i r o n from the solutions; 3) hydrolytic precipitation of metatitanic acid f r o m the sulfate solutions; 4) ignition of the precipitate accompanied by the formation of titanium dioxide. This method e n s u r e s a high degree of extraction of titanium into the final product, through the use of a single reagent - sulfu.ric acid. Decomposition of the concentrate. The following reactions take place in the decomposition of ilmenite concentrates with sulfuric acid:
FeTiO, FeTiO,

+ 3H,SO, + 2H,SO.,

= Ti

(SO,),

= TiOSO,

+ Few, + 3H,O; + FeSO, + 2H20.

(1)
(2)

Hence, the solution may contain both Ti4+and Ti@ ions, but Ti@ i m s a r e usually predominant. F e r r i c oxide, which is always present in the concentrates, dissolves with the formation of Fez(S0d3. A l l the reactions involved in the dissolution a r e exothermic. In industrial practice technical grade concentrated sulfuric acid (either oil of vitriol - 9 2 to 9470HzSO4, o r oleum - concentrated H2S04containing up to 20% dissolvedS03) is used f o r the decomposition. Since the decomposition reactjons a r e exothermic, it is sufficient to heat the acid ground concentrate mixture to the temperature at which rapid decomposi tion begins (125 to 135') and the reaction then proceeds vigorousIy, the temperature of the mixture r i s e s to 180 to 200", and the reaction is completed within 5 to 1 0 minutes. The decomposition product is the socalled "melt" consisting mainly of fine crystalline titanyl sulfate (TiOSO,. HzO), f e r r i c sulfate, and some sulfuric acid. The decomposition is c a r r i e d out either a s a batch o r a s a continuous process. The apparatus for the batch p r o c e s s is a steel tank with a conical bottom, lined with two l a y e r s of acid-resistant (diabase) plates. As a r e s u l t the apparatus may be used both for the decomposition (which does not require a protective lining) and for the subsequent leaching of the melt with water. The continuous decomposition of ilmenite may be c a r r i e d out in an apparatus consisting of a heated screw-conveyer mixer into which i s fed the concentrate mixed with sulfuric acid. Removal of iron f r o m the solution. The decomposition of ilmenite by sulfuric acid yields solutions containing 110 to 1 2 0 g/1 of T i @ ( a s titanyl sulfate TiOS04 and the normal titanium sulfate) and iron sulfates (FeS04 and Fez(S04)3),and 200 to 240 g/1 of active H2S04*.

* T h e active

H,SO, is the total amount of che free acid and the acid bound to the titanium sulfates.

180

The bulk of the iron must be removed froin the solution prior to the hydrolytic precipitation of the titanic acid, in order to prevent contamina tion of the precipitate. In o r d e r to remove the iron from the solution, the trivalent iron i s reduced to the divalent state, and f e r r o u s sulfate i s crystallized ( a s FeSO4- 7Hz0) by utilizing the considerable d e c r e a s e in the solubility of FeS04 accompanying the cooling of the solution. The solubility of FeSO, 7 H z 0 in a solution containing 1 2 0 to 1 4 0 g/1 TiOz and 240 g / l i-IzSO, i s (the values a r e converted to F e concentration in the solution):
Temperature.'C. Fe content,g/l

...........

........

30 88

20 70

14 48.5

10

5
35

43

25

-6 14

The reduction is effected by introducing iron turnings into the solution: Fez(S04)3 + F e = 3FeS04. The reduction of Fe3+to Fez+is completed when the solution turns violet; this is caused by the reduction of p a r t of t e t r a valent titanium to the trivalent titanium sulfate: 2Ti0(SO4)

+ Fe + 2H2S0, = Ti, (SO,), + FeSO, + H,O

This reaction begins only after the quantitative reduction of the trivalent iron. During the decomposition and leaching in batch equipment, the iron turnings a r e placed in a perforated stainless-steel basket which is immersed in the solution after the leaching, and i s allowed to r e s t there with slow s t i r r i n g by a s t r e a m of air. The reduction i s terminated when the Ti3+content reaches about 3 to 5 g / l . In the continuous p r o c e s s the iron is added to the tanks of the leaching system. The pulp is allowed to settle in continuous rake thickeners. Coagulants, such a s j o i n e r ' s glue, a r e added to the solution in o r d e r to accelerate the settling of fine dispersions. The product from the bottom of the thickener (the slimes) i s diluted with cold water and is filtered in drum filters. The dilute solutions obtained a r e returned to the leaching. The c l e a r solution from the thickeners is forwarded to the crystallization of f e r r o u s sulfate F e S 0 4 - 7 H z 0 . To this end, the solution i s cooled to 1 0 to 15". The f e r r o u s sulfate c r y s t a l s have a tendency to adhere strongly to the heat-transfer s u r f a c e s (the apparatus walls, which a r e cooled by water flowing within a jacket, the cooling coil surface), which r e s u l t s in a d e c r e a s e in the r a t e of heat removal and reduces the output of the apparatus. Thus, i t is most convenient to use continuous vacuum c r y s t a l l i z e r s in which the precipitate cannot adhere to the walls. After the crystallization of the FeS04. 7 H z 0 there i s a d e c r e a s e in the volume of the solution on account of the water of crystallization in the iron sulfate, The resulting solution contains about 140 to 150 g/1 TiO,, 280 to 300 g / l active H2S04, 30 to 35 g / l F e ( a s f e r r o u s sulfate), a s well as some aluminum, magnesium, and manganese sulfates. Large amounts of the by-product - f e r r o u s sulfate - may be used in the production of sulfuric acid. For this purpose the f e r r o u s sulfate i s ignited to 800 to 900" in the presence of pyrite o r coal. The sulfuric anhydride formed is forwarded to the production of sulfuric acid by the contact process. In addition, f e r r o u s sulfate is used to combat plant pests in agriculture.

181

Hydrolysis. Metatitanic acid is precipitated from titanium sulfate solutions by hydrolytic decomposition. The main hydrolysis reaction i s :
TiOSO, f H , O
. +

H2Ti03

+ H2S0,.

The composition of the solution and the method of hydrolysis used influence strongly the composition and s t r u c t u r e of the resulting precipitates. F o r instance, the production of fine titanium dioxide (used a s a pigment) r e q u i r e s that the hydrolysis be c a r r i e d out in solutions containing 180 to 2OOgJl Ti& and having an "acid factor" not higher than 2%. F r o m solutions containing up to 1 2 0 to 1 5 0 g / l Ti& metatitanic acid is obtained which upon ignition yields c o a r s e titanium dioxide. The specifications with regard to the physical properties of titanium dioxide used in metallurgy (for the production of hard alloys, nonferrous metal alloys and titanium metal) are not very s t r i c t , and the main requirement is that the product be of the required degree of purity. However, h e r e too (and especially f o r the production of hard alloys) the product must be of a standard particle size. Because of i t s l a r g e specific surface, the precipitated metatitanic acid usually contains adsorbed S O : - ions which a r e strongly bound to the precipitate. The sulfate groups a r e removed completely only by igniting the precipitate to 850 to 900". T h e r e a r e two commercial methods f o r the hydrolysis - the dilution method, and the nucleation method. T h e d i l u t i o n m e t h o d . The starting solution is concentrated by evaporation to a TiO, content of 240 to 260 g / l and an active H2S04content of 480 to 520 g / l . The concentrated solution i s then diluted under s t r i c t l y controlled conditions, by the addition of water to an active H2S04content of 3 8 0 to 400 g/l. Crystallization nuclei a r e formed in the solution during dilution, and metatitanic acid then precipitates out. In o r d e r to prevent the premature s t a r t of the hydrolysis the evaporation is c a r r i e d out at 70 to 75" under reduced p r e s s u r e (60mmHg) in vacuum evaporators. The dilution conditions must be s t r i c t l y controlled in o r d e r to obtain a precipitate with the required structure. T h e n u c l e a t i o n m e t h o d . In o r d e r to effect the hydrolysis, nuclei, which a r e prepared in advance in the form of a colloidal solution of hydrated titanium oxide, a r e introduced into the titanium sulfate solution. The colloidal solution is obtained by partial neutralization of the titanium sulfate solution with a sodium hydroxide solution (about 1 0 0 g j l ) to pH = 3 . The Ti@ content of the solution containing the nuclei i s about 5 0 g j l . The nuclei a r e added in an amount equal to 1 %of the T i 4 content, the solution is s t i r r e d , heated to boiling, and allowed to stay at that temperature for 2 to 4 hours. The precipitate contains 95 to 9 6 % of TiO,. Metallurgical titanium dioxide is usually produced by the nuclei in troduction method, which is less expensive since it p e r m i t s the use of the sulfate solutions directly after the filtration (without advance concentration by evaporation). The metatitanic acid precipitate produced by one of the above methods i s separated by filtration, and washed on drum or sheet vacuum filters.

_---------_--- * The "acid factor" is the ratio of

the concentration of active H,SO, to the concentration of TiO, in the solution (see footnote. p. 180).

182

The ignition of metatitanic acid. The ignition of the metatitanic acid precipitate r e s u l t s in the removal of S Q and water and in the formation of crystalline titanium dioxide. The water is removed at 200 to 300", the SO, at 500 to 950". Depending on the temperature, the ignition yields T i Q with the s t r u c t u r e of anatase ( a t temperatures up to 950") or rutile (above 950"). In the production of metallurgical titanium dioxide the ignition is c a r r i e d The product out at 1000 to llOOo, which e n s u r e s complete removal of SQ. is a c o a r s e dioxide with the s t r u c t u r e of rutile. When using the above technique, the m a t e r i a l s consumed in the production of one ton of titanium dioxide a r e :
ilmenite (42 % T i 0 3 sulfuric acid (as the monohydrate) iron turnings

............... 3.1 tons ..... 4.7 tons ....................... 0.24 tons

The sulfuric acid method f o r the production of titanium dioxide has the disadvantage of being expensive, because of the consumption of l a r g e amounts of sulfuric acid to dissolve the iron in the ilmenite (1.76 kg of H2S04 is theoretically needed to dissolve one kg of iron).

Production of titanium dioxide from titanium tetrachloride

/ 111

The production of titanium dioxide f r o m titanium tetrachloride has recently become of interest. The economic advantages of such a p r o c e s s a r e evident, a s a single p r o c e s s would yield two main products - TiC14 and T i Q . T h e r e a r e three known methods for the production of titanium dioxide from titanium tetrachloride: 1) hydrolysis of aqueous solutions of titanium chloride; 2) decomposition of gaseous titanium chloride by water vapor (gaseous-phase hydrolysis) ; 3) "combustion" of the chloride in a i r or oxygen at high temperatures.
Each of these methods is briefly discussed below.
Hydrolysis of aqueous solutions of titanium chloride. In this method
aqueous solutions of titanium chloride a r e prepared in advance. For this purpose, TiC14 is introduced into cold water o r dilute hydrochloric acid. The solution f i r s t becomes hot and turbid, which is associated with the hydrolytic precipitation of titanium oxychlorides and hydrated titanium dioxide. However, when further amounts of TiC14 a r e bubbled through,the precipitated compound dissolves and c l e a r solutions containing amounts of titanium equivalent to 550 g / l T i Q and 600 g / l HC1 may be obtained. Evaporation of HC1 takes place because of the high temperature of the solution. F o r this reason the molar ratio T i : C1 in the solution usually differs f r o m the stoichiometric and v a r i e s from 1 to 2.7. The titanium tetrachloride is introduced through a tube i m m e r s e d in the aqueous solution. In o r d e r to prevent clogging of the tube, dry a i r o r an inert gas is passed together with the chloride. The solution is s t i r r e d to prevent localized overheating. The concentration of T i Q in the starting solutions fluctuates f r o m 150 to 350 g/l.

183

The chloride solutions are hydrolyzed by nucleation (seeding) o r by the dilution method. The hydrolysis reaction i s :
TiCl, f 3H20 2 H,TiOs

+ 4HC1.

When using the seeding method f o r the hydrolysis, the nuclei a r e precipitated from a p a r t of the solution by neutralization with NaOH to pH 2 to 3 and heating to 80". The precipitate is then added to the main solution which is heated to 100". The hydrolysis is rapid and is completed within 1 0 minutes with the precipitation of 95 to 96% of the titanium from the solution. The hydrolysis may be successfully c a r r i e d out without seeding by pouring the concentrated solution of titanium chloride into boiling water and boiling the mixture. The precipitates formed in the hydrolysis are washed and ignited at 850 to 900" yielding titanium dioxide having the c r y s t a l lattice of rutile. High-purity titanium dioxide is obtained by using purified titanium chloride. Hydrolysis in the gaseous phase. The interaction of water vapor and gaseous TiC14 a t 300 to 400" yields titanium dioxide:

TiCI,

+ 2H20

Ti02

+ 4HCI.

The hydrolysis is c a r r i e d out a s a continuous p r o c e s s by feeding a s t r e a m of a i r saturated with water vapor and a s t r e a m of a i r saturated with titanium tetrachloride vapor into a reaction chamber preheated to 300 to 400". Before entering the chamber each s t r e a m is preheated to 300 to 400", The g a s e s from the hydrolysis chamber a r e fed to a dustseparation chamber (which is heated to 2 0 0 to 400" in o r d e r to prevent condensation of the HCl), f o r the separation of the T i Q . The T i 4 may be entrapped in porous c e r a m i c filters. The main difficulty in the p r o c e s s l i e s in the selection of a m a t e r i a l resisting the effect of hydrogen chloride in the presence of water vapor. Combustion of Tiel4. F r o m the standpoint of chlorine regeneration, the most expedient method f o r the production of TiOz is by the action of oxygen on Tiel4: TiCI, 0, + TiOz + X I , .

The reaction takes place a t 900 to 1100" and may be c a r r i e d out a s a continuous process. In one variant of the method, the a i r s t r e a m and the T i e l 4 vapors (which a r e usually diluted with nitrogen) a r e preheated to 1000 to 1100" and fed to reaction chambers in which the temperature is maintained at 750". The titanium tetrachloride r e a c t s with the air a t the exit of the g a s feeding tubes, with the formation of a yellow-green flame. The gas s t r e a m entrains the titanium dioxide particles (as smoke) to the dust- separation chamber. The amounts of g a s introduced p e r l i t e r of reaction-chamber volume a r e : 0.5 1 TiC14, 0.5 1 nitrogen, and 1 1 air. The selection of [structural] m a t e r i a l s f o r the above p r o c e s s is e a s i e r than in the c a s e of gaseous-phase hydrolysis.

184

. .. ..

... . ..

_ .. _.

46.

PRODUCTION OF METALLIC TITANIUM

The high chemical activity of titanium makes the production of the pure metal from i t s compounds difficult. At the s a m e time, the modern specifications f o r titanium require a high-purity metal. A s mentioned above, oxygen, nitrogen, and carbon present a s i m purities increase the strength of titanium and reduce i t s ductility. The presence of hydrogen has little effect on the hardness and strength of titanium, but causes a s h a r p d e c r e a s e in impact strength. This is attributed to the precipitation of titanium hydrides on the grain boundaries.
TABLE 34
Free energy changes in the interaction of titanium with various gases Reaction Ti 0, = TiO, . . . . . 2Ti CO, = TiO. Tic i Ti Ne = TiN . . .

250

8000

+ ......... + + . . . . . . . . . + -= . . . . . . . . . L 3Ti + 2CO = TiO, + 2TiC . . . . . . . . . Ti + 2H.O = TiO, + 2Hz . . . . . . . . . Ti + 2C0, TiO, + 2CO . . . . . . . . . Ti + C = TIC. . . . . . . . . . . . . .

-212
-i75 -73.4 -258.2 -51 .5 -44.7 -57.2

-186.5 -152
-69

-217 -48.2 -48.2 -60

The reactions of titanium with oxygen, nitrogen, carbon, carbon oxides, and water vapor a r e accompanied by a l a r g e decrease in f r e e energy ( s e e Table 3 4 ) . Hence, v e r y s m a l l amounts of the g a s e s mentioned above and carbon r e a c t with the titanium yielding titanium oxides, carbide, and nitride. The tendency of titanium to dissolve oxygen and nitrogen must be allowed for. The above data show that a t any stage of the production of titanium m e a s u r e s must be taken to prevent it from reacting with oxygen, nitrogen, water vapor, carbon, and carbon-containing gases. This is accomplished by conducting the reduction and melting of the metals in hermetically sealed equipment under an inert-gas atmosphere (argon, helium) o r in vacuo. One of the important conditions is that the starting titanium compounds and the reducing agents used be of a high purity. The production methods of metallic titanium may be divided a s follows: 1) reduction of titanium tetrachloride by magnesium o r sodium; 2) reduction of titanium dioxide by calcium o r calcium hydride; 3) electrolytic methods; 4) t h e r m a l dissociation of titanium halides. The largest p a r t of the titanium produced today is made by reduction of titanium chloride with magnesium o r sodium. A flow sheet of these p r o c e s s e s is shown in Figure 97 and Figure 103. Of the two reducing agents used, magnesium was the f i r s t to be used in the development of industrial titanium production. It was assumed that the reduction of Tic14 with sodium would be m o r e difficult to accomplish, a s sodium and T i e l 4 r e a c t with the evolution of l a r g e amounts

185

of heat, and the reaction is explosive and difficult t o control. Moreover, the high chemical activity of sodium necessitates special preventive m e a s u r e s when dealing with it. Investigations have shown, however, that the difficulties accompanying the reduction of titanium chloride by sodium have been exaggerated to a g r e a t extent. A l s o , thermal reduction with sodium has some advantages over the thermal reduction with magnesium ( s e e below, Section 48).

Reduction with magnesium

1 I
3
Condensate MgC12, Mg

Vacuum distillation

Electrolysis Chlorine Magnesium

electrodes

To the chlorinator, for the production


F

Arc melting

Pressure working

_.^.

J
Wastes

Electrolytic refining

FIGURE 91. Flow sheet of commercial methods for the production of titanium from titanium tetrachloride.

Depending on the reduction method employed, titanium is obtained either a s a spongy m a t e r i a l o r a s a powder. Subsequently, i t is converted into the solid metal (after separation of the slag and the excess reducing agent) by fusion o r occasionally by powder metallurgy methods.

47. REDUCTION OF TITANIUM TETRACHLORIDE WITH MAGNESIUM (THE KROLL PROCESS)

Physicochemical principles of the process The reduction of titanium chloride with magnesium is c a r r i e d out in hermetically sealed steel r e a c t o r s filled with an i n e r t gas (argon o r
186

helium). A pool of molten magnesium lies at the bottom of the reactor, and titanium tetrachloride is fed to the apparatus (from a p r e s s u r e tank) at a controlled r a t e ; the T i C b vapor r e a c t s with the magnesium as follows:, TiC14 (gas) + 2Mg (liquid) -+ Ti (solid) + 2MgC12 (liquid). The reaction is exothermic. The heat evolved (122 kcal p e r mole of T i c 4 o r 2545 kcal p e r kilo of titanium) suffices to maintain the reaction without a supply of heat from the outside. It would seem a t f i r s t sight that the reduction with magnesium should proceed without difficulties up to the quantitative utilization of the magnesium, since a t the p r o c e s s temperature (800 to 900O) the p h a s e s should be separated into l a y e r s : the upper l a y e r consisting of liquid magnesium (density 1.47g/cm3), the lower of MgC1, (density 1.67g/cm3), and on the bottom heavier p a r t i c l e s consisting of titanium sponge agglomerates (density 4.5g/cms). Thus, the liquid magnesium upper l a y e r should be accessible f o r reaction with the TiC14 vapors throughout the reduction process. The true mechanism of the reduction is m o r e complex and has not been adequately studied. Since this is a batch process, i t s r a t e v a r i e s with the amount of titanium chloride added into the r e a c t o r and consequently, with the amount of magnesium consumed and the accumulated reaction products - titanium sponge and magnesium chloride (slag). At the s t a r t , when l e s s than 6070 of the TiC1, have been introduced into the apparatus, the reduction takes place very rapidly. After about 50% of the magnesium have been consumed the remaining magnesium disappears a s a one-phase melt, since i t is completely absorbed in the p o r e s of the sponge; this leads to a d e c r e a s e in the reduction rate. The accumulating slag (liquid magnesium chloride) also reduces the p r o c e s s rate. One l i t e r of magnesium yields 0.354 1 of titanium and 3 . 6 8 1 of magnesium chloride. The level of the magnesium chloride gradually r i s e s , above the sponge level, when the reaction practically ceases. This is shown by a d e c r e a s e in the temperature and an i n c r e a s e in the p r e s s u r e in the r e a c t o r (since the titanium tetrachloride supplied to the r e a c t o r is not consumed). Magnesium chloride is discharged periodically to b a r e the surface of the sponge and to utilize m o r e fully the working volume of the reactor. This results in an i n c r e a s e in the reaction rate. Rather than charging the apparatus with the whole amount of magnesium a t the s t a r t , i t is b e t t e r to add the magnesium periodically in o r d e r to have a uniform-rate process. The periodic addition of magnesium r e s u l t s in a higher degree of utilization of the magnesium and in the production of titanium sponge of m o r e uniform composition. The lower temperature l i m i t of the reduction is the melting point of MgCl, (714") while the upper limit is 975". Above that temperature the titanium is fused together with the iron (the melting point of the titaniumiron eutectic i s 1085"). Moreover, the iron r e a c t s with the Tic14 yielding lower titanium chlorides and iron titanide (FeTi):
Fe + PFiCI, = FeCl, 2TiC&; Fe TiCI, = FeCI, TiCI,; 3Ti 2FeC1, = 2FeTi TiCI,;

+ +

+ + +

187

Fe

Fe + PTiCl, = FeTi + TiCI,; + 4TiC1, = FeTi + BTiCl,.

This may r e s u l t in the contamination of titanium with iron, and to a rapid failure of the equipment, The temperature range in which the reduction may be c a r r i e d out is fairly wide (720 to 975'). In practice the p r o c e s s is c a r r i e d out a t 800 to 900". The temperature may be maintained by controlling the rate of supply of titanium tetrachloride. In o r d e r to i n c r e a s e the output of the apparatus, a p a r t of the excess heat is removed by cooling the outer walls of the r e a c t o r with cold air. Lower titanium chlorides a r e formed towards the end of the reduction when most of the magnesium has been consilmed; some of the reactions involved a r e :
PTiCI, TiCI,

+ Mg = TiCI, + MgCI,; 3TiC1, + Ti = 4TiC1,;

+ Mg = PTiCI, + MgCI,;

TiCl, +Ti = PTiCI,.

The lower titanium chlorides dissolve in the liquid magnesium chloride. A s the melt flows through the p o r e s of the sponge, the lower chlorides a r e partly reduced by the magnesium in the pores. An excess of magnesium (about 15 to 25% of the theoretical amount needed) is used in o r d e r to ensure quantitative reduction, and a t the end of the reduction (i. e . , after the interruption of the supply of TiC14) the r e a c t o r is held a t a high temperature ( 9 0 0 to 920") to complete the reduction of the lower chlorides. The mechanism of formation of the sponge in the apparatus has not yet been finally clarified. In
vestigations with industrial equipment have shown that the main factor governing the reduction p r o c e s s is the r a t e of supply of TiC14. This factor governs the growth and s t r u c t u r e of the titanium sponge, and the temperature and p r e s s u r e inside the equipment. It has been established that the reduction of the t e t r a FIGURE 98. Approximate chloride by magnesium i s autocatalytic*<. The disposition Of the titanium reaction r a t e i n c r e a s e s with the formation and the sponge (the reaction mass) growth of the titanium sponge. In other words, the towards the end of the reaction r a t e depends on the development of the reduction. surface of the titanium sponge formed in the reaction, 1 -reaction vessel; 2 This is governed mainly by the r a t e of supply of sponge ("bloom"): 3 TiC14 to the apparatus 1 2 6 1 . sponge lining. Figure 98 shows the approximate disposition of the titanium sponge in an industrial. vessel operating without periodical addition of magnesium. The bulk of the sponge filling the apparatus (about 75 to 80%of it) is the so-called bloom.

* Autocatalyric reactions proceed


reaction products.

_____________-_

at a gradually increasing rate because of the catalytic effect of the

188

Of the total amount of magnesium chloride formed, 75 to 85% is drained during the reduction and forwarded to electrolysis.

Types of r e a c t o r s used f o r t h e r m a l reduction with magnesium / 8 /

In titanium plants the reduction of titanium tetrachloride by magnesium is c a r r i e d out in steel r e a c t o r s designed to produce from 500 to 6 0 0 to 1000 to 1500 kg of titanium sponge in one operation. Most of the body of the equipment ( r e t o r t ) is made of stainless s t e e l (18% Cr, 8% Ni, 2 % Mo) which r e s i s t s oxidation and has a satisfactory strength at 950 to 1000". Reactors made of such steel may be withdrawn from the furnace at 700 to 800 and directly cooled with water, with consequent increase in the turnover. The r e t o r t i s covered with a lid, which is fastened to the flange with bolts*. Hermetic sealing is accomplished with the aid of rubber-ring gaskets cooied by water flowing through a jacket. Sleeves in the lid s e r v e to introduce solid o r liquid magnesium, to feed titanium tetrachloride from the p r e s s u r e tank, to join the installation to the vacuum system, and for the supply of argon. Two types of r e a c t o r s a r e used: r e a c t o r s with an i n s e r t pot, and without an i n s e r t pot. The main difference between various types of r e a c t o r s with an i n s e r t pot is the accessory serving to discharge magnesium chloride. The simplest technique of slag evacuation i s employed in r e a c t o r s with an i n s e r t pot with a perforated bottom (Figure 99). At the beginning of the p r o c e s s the liquid magnesium level i s slightly above the perforated bottom of the pot. The titanium sponge gradually grows and fills the pot. The liquid magnesium is absorbed into the sponge p o r e s from below. Magnesium chloride FIGURE 99. Diagram of an flows through the perforations into the r e t o r t space insert pot reactor with a below the pot, and i s discharged. perforated bottom. Reactors with an i n s e r t pot have the following 1-body; 2 -insert pot with disadvantages: a low coefficient of utilization of perforated bottom; 3-device the working volume of the r e t o r t (0.4 to 0.45); a for discharging slag; 4 ring-shaped gap between the pot and the r e t o r t lid with sleeves for the in which i n t e r f e r e s with heat t r a n s f e r and reduces troduction of magnesium, chloride, and protective gas. the r a t e of the p r o c e s s ; the complex design of slag discharge mechanism. A.s a r e s u l t extensive use is made of r e a c t o r s without an i n s e r t pot, in which the reaction products (the reaction m a s s ) accumulates inside the retort. In these r e a c t o r s the slag is discharged by a simple technique (through a sleeve fitted to the bottom of the r e t o r t ) ; the volume of the

w3 1

T h e lid is occasionally welded to the flange of the retorr. is cut on a l a t h e and the lid is removed.

At the end of the operation the welded joinr

189

r e t o r t is utilized m o r e fully (the utilization coefficient is 0.5 to 0.6); the heat is led off m o r e efficiently through the r e t o r t walls which a r e cooled by a s t r e a m of a i r , s o that the p r o c e s s r a t e can be increased. The temperature in r e a c t o r s of this type may be controlled by simply taking the temperature of the e x t e r i o r of the walls at s e v e r a l height levels of the retort, without the need f o r inserting thermocouples into the reactor, The r e a c t o r s a r e heated in electrical or gas-fired furnaces. It must, however, be borne in mind that heating is required only a t the beginning of the reaction (heating the installation fusion of the f i r s t batch of magnesium charged) and a t its end.

FIGURE 100. Installation used for industrial-scale reduction of T i C 1 , by


magnesium. 1 -reactors; 2 -furnace: 3-pressure vessel with TiCI,: 4-fan: unit: 6 -vacuum pumps; 7-overhead hoist: 8-platform. 5-distillation

F i g u r e 100 shows one of the setups used for the l a r g e - s c a l e reduction.

The reduction p r o c e s s - conditions and control The p r o c e s s conditions depend on the type of apparatus used, i t s size, the manner in which magnesium is introduced (liquid or a solid, with or without periodical replenishment) and the r a t e of removal of heat from the walls. A s an example, the reduction p r o c e s s in an apparatus without an i n s e r t pot and without replenishment of magnesium is described below. The installation is assembled and checked in a special room. The lid is fitted to the r e t o r t and the tightness checked. For this purpose, the a i r is evacuated and the leak r a t e measured (by measuring the i n c r e a s e in p r e s s u r e ) . The assembled installation is inserted into the furnace with

190

. . .

..

. ....

...

the aid of a hoist, and filled with argon ( o r helium). The apparatus is heated to 700 t o 750" and is charged with liquid magnesium*; compressed TiC14 is then supplied f r o m a tank. The furnace is disconnected and the tempera ture maintained a t 850 t o 900" by controlling the r a t e of supply of titanium chloride. In o r d e r t o i n c r e a s e the output of the installation, p a r t of the e x c e s s heat is removed by cooling the r e t o r t with a i r , which is blown in the c i r c u l a r gap between the furnace and the retort. The s l a g is discharged for the first time after about 60% of the magnesium has been consumed. The argon p r e s s u r e in the apparatus is i n c r e a s e d somewhat during the discharge. Subsequently the slag is discharged a f t e r about 80% and 95% of the TiC1, have been consumed. After the whole amount of the titanium FIGURE 101. Diagram of t h e automatic chloride has been introduced into the control of reduction of TiC14 with mag reactor, the slag is discharged f o r the nesium. l a s t time and the r e a c t o r is held for 1 -multiple-junction thermocouple; 2 3 0 to 6 0 minutes, at 900 to 920", in maximum-temperature selector; 3 - m ea o r d e r to reduce the lower chlorides. suring device (electronic potentiometer); The furnace is then shut off and when the 4 -isodromic [sic] regulator controlling temperature reaches 600 to 700" the the actuating mechanism; 5-actuating reactor is withdrawn f r o m the furnace mechanism connected to t h e TiCI, -supply and placed in a special nest in which valve; 6-valve: 7 -body of t h e apparatus: water-cooling takes place. The cooled 8 -tube for t he supply of TiC1,. installation is then dismantled. During the reduction p r o c e s s the required temperature is maintained automatically in the r e a c t o r by control ling the r a t e of supply of the TiC14. The automatic control of the tempera ture is complicated by the fact that the temperature field of the r e a c t o r is nonuniform. A s the reaction proceeds the maximum temperature shifts upwards, together with the reaction zone. The r e t o r t walls a r e a t t h e i r highest temperature within the ring-shaped region in which the reaction takes place a t the most rapid r a t e a t the given moment. Hence, the temperature is measured with s e v e r a l thermocouples fitted a t different height levels in the retort. A s the reaction zone shifts, a special device ("the maximumtemperature selector") automatically connects the TiC4-supply regulator to the thermocouple which is measuring the highest temperature a t that instant. A diagram of the temperature control system is shown in Figure 101 / 8 / .
T h e above sc he me does not ensure t h e maximum process rate or the maintenance of standard working conditions. This is because t h e determining effectof the rate of supply of TiCl, on t h e process ra t e is not taken into account.

_-------------
* When using solid
magnesium as feed, t h e magnesium bars a re first etched in nydrochloric acid. then rinsed. dried, and charged into t h e apparatus before t h e assembly.

191

As a result, a new au t o m at i c control system has been recently developed; i t consists in t h e o p t i m u m p r o g r a m m i n g of t h e supply of TiC1, t6 t h e installation. T h e temperature and pressure are used only as check, but they do not determine t h e r at e of supply of TiCI, t o t h e installation. T h e reduction t i m e be comes shorter and enables t h e process t o b e carried out under standard conditions /26/

Processing of the reduction product The reduction product (the reaction m a s s ) contains 55 to 65'7'0 titanium, 20 to 3070 residual MgC12, and 1 0 to 20% excess magnesium. There a r e two known methods for the purification of titanium sponge leaching and vacuum distillation. The leaching method was used in the f i r s t titanium-producing plants. The reaction m a s s (which was drilled out of the r e t o r t or the insert pot) was treated with dilute (about 1% ) cold hydrochloric acid, with the purpose of removing the bulk of magnesium and magnesium chloride. The sponge is then subjected to wet milling in a ball mill, and is leached with 10% hydrochloric acid at 45' with the purpose of removing the residual magnesium and magnesium chloride. The powder is then rinsed with water, dried, and passed through an electromagnetic s e p a r a t o r to remove iron particles which may have contaminated the sponge during the drilling. The titanium powder produced contains 0.1 to 0.3 70hydrogen (which i s evolved when magnesium i s dissolved in the acid) and up to 0.1% oxygen, whose presence is caused by the oxidation of titanium during the wet processing. The oxidation is m o r e rapid if heat is applied during the dissolution of magnesium in the acid. The hydrogen is removed (and its concentration reduced to a permissible level) during a r c melting in vacuum, but the oxygen remains and the metal has an increased hardness. A method employing vacuum distillation to remove magnesium and magnesium chloride from the reaction mixture, developed with the purpose of improving the quality of the metal, has now superseded the leaching method. The vacuum distillation method. The purification of the sponge by this method utilizes the relatively high vapor p r e s s u r e of magnesium and magnesium chloride at 850 to 950", and the very low vapor p r e s s u r e of titanium at that temperature:
Temperature, 'C Vapor pressure, m m Hg: of magnesium of magnesium chloride 700
5 E00 25 2.2

960 EO 8.0

1000

250 EO

Prolonged heating of titanium sponge in high vacuum ( 2 . -3 . mmHg) in steel retorts, at 900 to 925", causes vaporization of magnesium and magnesium chloride, which condense on the surface of a water-cooled condenser. The condensate is electrolyzed in o r d e r to recover the magnesium. In industrial practice the distillation i s carried out by one of two methods: 1) in the r e t o r t (or insert pot) which was used for the reduction, with out removing the reaction mixture.

192

2) the reaction mixture is extracted from the r e t o r t (or the i n s e r t pot) with the aid of a pneumatic chisel or by drilling - on a lathe and is charged in a granulated form into a perforated s t e e l basket which is fitted into the distillation retort. The second method p e r m i t s a n i n c r e a s e in the output of the distillation unit, because of the m o r e complete filling of the retort. How ever, the removal of the reaction m a s s and i t s t r a n s f e r to the basket is accompanied by the absorption of moisture (by magnesium chloride), and formation of the nonvolatile hydroxychloride MgOHCI, which r e s u l t s in an i n c r e a s e in the oxygen content of the sponge. For this reason i t is recommended that the removal and charging of the m a s s be c a r r i e d out in special d r y rooms, with an atmospheric moisture content not higher than 0.5% (dew point -40"). The need for dry rooms i n c r e a s e s the capital investment. Moreover, the working conditions in such rooms a r e unfavorable to the health of the presonnel. In o r d e r to prevent the r e t o r t walls (which a r e heated to 900 to 1000") f r o m caving in under the effect of the atmospheric p r e s s u r e , FIGURE 102. Thermal purification of titanium sponge in vacuo, with the furnace jacket i s hermetically sealed and a out transfer from the reduction "countervacuum" of about 0.2 mm Hg is main vessel. tained in it. A vacuum of the o r d e r of 1 -reduction vessel (retort); 2 Hg i s created inside the r e t o r t with the aid of a titanium sponge; 3-electrical vacuum system consisting of a r o t a r y and furnace with a sealed jacket; a diffusion pump ( s e e Figure 83). 4 -sleeve connection t o vacuum A water-cooled condenser i s fitted to the system (to produce counterflange of the r e t o r t i f the distillation i s c a r r i e d vacuum in the furnace); 5 out in the r e t o r t used for the reduction, with rubber gaskets; 6-screens; 1 out transferring the reaction mixture 1 3 3 1 . water-cooled condenser; 8 The r e t o r t is then t r a n s f e r r e d to the distillation condensate - Mg and MgC1,; 9-sleeve connection to the furnace (Figure 1 0 2 ) . The r e t o r t is allowed vacuum system. to remain in the furnace for a prolonged time ( 2 0 to 25 hours) a t 900 to 925', and i s then withdrawn, cooled, t r a n s f e r r e d to a special room, dismaitled, and the sponge is excavated with pneumatic chisels. Depending on the degree of adhesion of the bulk of the titanium sponge (the bloom) i t is withdrawn either in one piece or in a granulated f o r m . The fraction of the sponge removed from the walls i s collected separately since it contains l a r g e r amounts of impurities. The sponge particles a r e crushed to the required particle s i z e f i r s t with a h a m m e r or in a p r e s s , and then in a mill. The refined sponge is characterized by the following approximate impurity contents: 0.0016% Hz, 0.05% Q, 0.002% Nz, 0.07% C1, 0.08% Mg, 0.13% Fe, 0.03% Si, 0.1% C. After fusion in an a r c furnace, the hardness of the titanium i s 1 0 0 to 150 kg/mm2.

193

The condensate produced contains magnesium and magnesium chloride, and is returned to the magnesium production plant. The vacuum distillation method has the disadvantages of low output, high consumption of electrical energy (18 to 22 kwh/kg sponge), and the need f o r a complex equipment to c r e a t e the high vacuum.

The technical and economical indexes of the p r o c e s s A well-designed titanium plant must include electrolysis of magnesium chloride, in o r d e r to produce magnesium and chlorine as by-products from the recycled m a t e r i a l ( s e e Figure 97). When using titanium s l a g s containing 80 to 8570 Ti& as the starting r a w material, the total degree of extraction of titanium from the s l a g s and the sponge is about 7270 (without taking into account the recovery of titanium from the wastes). The approximate degrees of extraction of titanium in the separate stages are, i n % :
From From From From slags into pellets pellets into technical chloride technical into purified chloride purified chloride into commercial T i sponge

.............................. 97-98 .................. 85- 86 ................ -95


.......
90-91

The degree of extraction may be increased by reducing the amount of rejected sponge and by its processing (e. g., by electrolytic refining), by improving the chlorination and chloride-purification processes, by recovering the titanium from the chlorination and purification wastes, and by reducing the mechanical l o s s e s in the pelletizing stage. In the total cost of titanium sponge, about 76.270 a r e accounted f o r by the cost of titanium tetrachloride and magnesium. The power consumption accounts f o r 2.470, wages f o r 5.570, and overheads for 15.670 of the cost. This shows that the cost of the sponge may be reduced mainly by reducing the p r i c e of titanium tetrachloride.

48. REDUCTION OF TITANIUM TETRACHLORIDE BY SODIUM / 9 /


A s mentioned above, many plants u s e sodium a s the reducing agent. There a r e s e v e r a l advantages of sodium over magnesium. 1. Sodium has a lower melting point (98"), which facilitates i t s piping and feeding into the reactor. Oxide films and certain impurities a r e readily removed from sodium by filtration. 2. The reduction of TiC14 by sodium is more rapid, and the sodium utilization coefficient is as high as 100% (in the reduction with magnesium it ranges between 70 and 90% ). The quantitative utilization of the reducing agents p e r m i t s the p r o c e s s to be c a r r i e d out without tapping off the sodium chloride, which simplifies the construction and maintenance of the equipment.

194

Sodium

Filtration

1 1
Fusion

I
1
Metering

Titanium terrachloride

I Metering

TLIArgon

Reduction Removal of the reaction mass

1
Crushing

Water

I
Washing the sponge Solution to waste Drying the sponge Fractionation by screening

Hydrochloric acid
I

Leaching :

Solution for neutralization

. 1
Pelletizing

I
Ingot casting

FIGURE 103. Flow sheet of the production of titanium by thermal reduction w i t h sodium.

3 . Unlike magnesium chloride, sodium chloride i s only slightly hygroscopic and does not hydrolyze in aqueous solutions. This, a s well a s the negligible concentration of elemental sodium in the reaction mixture, makes i t possible to s e p a r a t e slag from titanium by simply leaching with water, thus avoiding vacuum distillation, which is a complex and energy-consuming process. 4. Titanium powder produced by reduction i s e a s i e r to use than titanium sponge for the production of alloys since it i s readily mixed with the alloying additives. This r e s u l t s in the formation of uniform alloys. Some of the disadvantages of the method a r e : the volume of the reducing vessel and s l a g p e r kg of titanium a r e l a r g e r than when magnesium is used, which r e q u i r e s the use of l a r g e sized equipment; the reduction is accompanied by the evolution of l a r g e quantities of heat (about 70% m o r e than when magnesium is used), which causes difficulties in the removal of the heat; the safety precautions required a r e more stringent than in the c a s e of magnesium, because of the high chemical reactivity of sodium.

195

However, i t was found in practice that these disadvantages and difficul ties can be overcome. The overall flow sheet of thermal reduction with sodium is shown in Figure 103.

Reduction The reduction of TiC14 by sodium may b e c a r r i e d out in t h r e e tempera t u r e ranges: at temperatures below the melting point of sodium chloride (i. e . , below 801"); a t 801 to 883", i. e., between the melting point of sodium chloride and the boiling point of sodium; and a t temperatures above the boiling point of sodium. In industrial p r a c t i c e the reduction is c a r r i e d out in the range 801 to 883". In this c a s e the reaction occurs mainly between gaseous TiC14 and liquid sodium: TiC14(gas) + 4Na(liquid) = Ti(so1id) + 4NaC1 (liquid)

+ 172.8 kcal.

However, a noticeable p a r t of the reaction takes place in the gaseous phase above the melt since at 801" the vapor p r e s s u r e of sodium is about 340 m m Hg: TiC14(gas) + 4Na(gas) = Ti(so1id) + 4NaC1 (liquid)

+ 262.0

kcal.

The thermal effect of the reduction by sodium vapor is 84 kcal higher than in the reduction by liquid sodium. The above equations r e p r e s e n t the overall reactions. I n fact, the reduction takes place through s t a g e s involving the formation of lower titanium chlorides, TiC13 and TiC12, which dissolve in sodium chloride+. The lower chlorides a r e formed through the interaction of gaseous TiC14 with titanium particles: 3TiC14(gas) + T i (solid) 4TiC13 (liquid), TiC14(gas) + Ti (solid) P 2TiC12(liquid). The lower chlorides then diffuse into the bulk of the melt. Here they may undergo disproportionation by the reactions written above (which in this c a s e proceed from right to left) yielding finely-dispersed titanium powder. The following reactions are also possible: 2TiC13(liquid) Ft T i e l z(liquid) + TiC14(gas); 3TiC12(li'quid) + 2TiC13(liquid) + T i (solid). Sodium chloride and TiClz form a eutectic mixture ( a t a TiClz concentra tion of 33%by weight, which corresponds to the composition 2NaCl+TiC12) with a melting point of 605". In the system TiC13 - NaCl t h e r e is a eutectic mixture with the composition of 437'0by weight T i e l 3 and 577'0 by weight NaC1, which m e l t s at 554O. In addition, t h e r e is a chemical compound with the composition TiC13 * 3NaCI. A t e r n a r y compound having the composition 13 NaCl. 3TiC13. 2TiC12 may be formed in the presence of both chlorides. Sodium is dissolved by molten NaC1, as shown by the data below:
?'emperarure,'C ........... 810 Solubility o f s o d i u m . % ....... 1.12
825 1.99 890 3.88 930 9.63 950 11.02

____________-__
*
The properties of the lower chlorides have been described on p. 161.

196

The reduction is c a r r i e d out in stainless-steel r e a c t o r s , into which measured amounts of T i Q and liquid N a are charged simultaneously a t a ratio close to the stoichiometric (i. e . , 2.06 kgTiCI4 p e r kg Na). The apparatus is f i r s t evacuated, then filled with argon and heated in a furnace to about 650 to 700". The reaction is allowed to proceed for a certain time, during which a liquid melt accumulates in the r e t o r t ; the furnace is then shut off and the p r o c e s s continues at the expense of the heat of reaction, which maintains the temperature a t 850 to 880". The e x c e s s heat is removed by blowing air round the retort. In the first stage of the reaction, the reduction takes place to a con siderable extent in the gaseous phase, with the formation of finelydispersed titanium and the lower chlorides. The lower chlorides dissolve in the sodium chloride, and a r e reduced by the sodium present in the melt, yielding titanium c r y s t a l s which a r e heavier and thus settle to the bottom. At the end of the p r o c e s s the furnace is turned on and the r e t o r t is maintained a t 950 to 970" in o r d e r to effect quantitative reduction of the lower chlorides and to cause consolidation of the metal particles. In one plant in the USA (in the city of Ashtabula) the p r o c e s s i s c a r r i e d out in r e a c t o r s made of two steel l a y e r s (the inner l a y e r i s high-carbon steel, the outer is stainless steel). The r e t o r t h a s a diameter of 1.5 m, a height of 4.2 m, and a wall thickness of 2 5 mm. Electrolytic sodium i s used a s the reducing agent. It is supplied in metallic drums, under a protective layer of paraffin, o r in tank c a r s in an argon atmosphere. In o r d e r to pump the sodium out of the tank c a r i t is melted with the aid of oil (heated to 120") circulating within a coil fitted in the tank c a r . The sodium i s purified by filtration through 2 0 l a y e r s of a 0.044" nickel gauze and a porous metallic f i l t e r with 2 0 b holes. The purified sodium contains not m o r e than 0.0570 impurities. The sodium and titanium chlorides a r e continuously metered into the r e a c t o r for 5 to 7 hours. In o r d e r to complete the reaction the furnace is turned on and the r e a c t o r is allowed to remain at about 1000" for 4 to 6 hours, the r e a c t o r i s then cooled f o r 2 to 4 hours and transferred to a "nest" in which it i s cooled by a forced flow of a i r . The r a t e of supply of TiC1, to the r e a c t o r is on the average 400 kg/m2. h r , which is twice that used in the thermal reduction by magnesium. Each reduction cycle yields about one ton of titanium.

Processing of the reaction mixture The reaction mixture containing 1 7 % Ti, 8370NaCl and very s m a l l amounts of impurities (sodium and the lower chlorides) i s taken out of the r e t o r t by cutting with a special d r i l l and is then passed through a s c r e e n with 10" holes. The c o a r s e fraction is crushed in a hammer mill. The crushed m a t e r i a l is leached with water containing up to 1% HC1 in rubber-lined s t e e l vats fitted with s t i r r e r s . The m a t e r i a l is slowly added and the hydrogen evolved in the reaction between the sodium and the water is driven off with the aid of an exhaust system. The mixture i s s t i r r e d for about 30 minutes and the titanium powder is separated from the sdlution by centrifuging. The powder is then dried in a vacuum d r i e r under a p r e s s u r e of 50 mm Hg.

I97

Titanium powders have a particle s i z e ranging f r o m +2,4 to -0.07". The bulk of the p a r t i c l e s have a s i z e between 2.4 and 0.2". The approximate concentration of impurities is: 0.01 to 0.7% C, 0.04 to 0.15% 0, 0.001 to 0.02% N, and 0.005 to 0.019 H . The titanium is c a s t into ingots with a Brinell hardness of 110 to 160kg/ mm2 in an a r c furnace. Its mechanical properties a r e not inferior to the titanium produced by t h e r m a l reduction with magnesium.

Two-stage t h e r m a l reduction with sodium


An improved version of the reduction of titanium chloride by sodium is the recently developed two-stage reduction process. In the f i r s t stage the reduction is c a r r i e d out inti1 a eutectic melt whose composition corresponds to 2NaCl + TiClz is formed. In the second stage the dichloride is reduced to titanium by the sodium present in the melt: F i r s t stage

TiC14(gas) + ZNa(1iquid) = 2NaCl(liquid) + TiC12(liquid) + 111.3 kcal. Second stage ZNaCl(1iquid) + TiC12(liquid) + 2Na (liquid) = 4NaCl(liquid) + Ti(so1id) + 63.5 kcal.

Of the total amount of heat, 6470 is evolved in the f i r s t and 36% in the second stage. This r e s u l t s in a noticeable alleviation of the problem of heat removal from the reactor. The second important advantage of the two-stage p r o c e s s is that favorable conditions are created f o r the growth of l a r g e titanium c r y s t a l s in the second stage, in which the reduction takes place i n the melt. The large-sized crystalline dendrites formed (50 mm or l a r g e r ) a r e distinguished by their high purity. The two-stage reduction can easily be c a r r i e d out as a semicontinuous process. In the f i r s t stage sodium and titanium chloride a r e simultaneously charged into the s t e e l reactor, a t a m o l a r ratio of 2 : 1. The reduction i s c a r r i e d out at 700 to 750" in an argon atmosphere. After the accumulation of a liquid melt(of the composition 2NaCL + TiC12) in the apparatus, the melt is conveyed by argon p r e s s u r e through a heated s t e e l tube into the second-stage reactor. The reduction is completed in the second reactor, a t temperatures between 650 and 900". The temperature is controlled by varying the r a t e of supply of liquid sodium to the reactor. A t the end of the reaction the r e a c t o r is held at 950" and is then cooled. The withdrawal of the cake and its leaching are c a r r i e d out in the s a m e manner a s that described above f o r the one-stage process. Casting the coarsely crystalline titanium yielded ingots with a Brinell hardness of 85.5 kg/mm2, a tensile strength of 25 kg/mm2, and an elongation of 567'0 ; these values a r e close to the values obtained with extremely pure (iodide) titanium.

1455

198

49. R E D U C T I O N O F TITANIUM DIOXIDE

Titanium is also produced by the reduction of titanium dioxide. When selecting a reducing agent for titanium dioxide, consideration must be given to the fact that the reduction proceeds through an intermediate stage involving the formation of titanium monoxide TiO, which is a v e r y stable chemical compound.
TABLE 35 Free energy of formation o f some m e t a l l i c oxides
A F , kca'

[-atom 0
1000C

Oxide

25'C
-146

............... MgO ............... 1 - AlZOs ............ 3


CaO

-121 -112

-131.2 -125 -116

-100

Ti0

NaZO

................. ...............

88.0 55.1

89.2

Of the available reducing agents, calcium has the highest affinity f o r oxygen and is able to reduce titanium dioxide to the metal (Table 3 5 ) . However, even the use of so strong a reducing agent does not yield titanium with an oxygen content below 0.2%. This is due to the solubility of oxygen in titanium (12.5% by weight of 30 at.%), and the fact that the bond between oxygen and titanium is strengthened as the oxygen concentra tion in the solid solution is reduced. The f r e e energy of formation of solid solutions of oxygen in titanium at 1000" is listed below:
Oxygen concentration, at.%

4F. kcal/ g-atom 0

0.2 -122.0

1.6 -112.5

10 -108.7

20 -100.2

30

-92.5

In a solid solution containing 0.2 at.% ( o r 0.07'70 by weight) of oxygen, the affinity of oxygen for titanium is equal to the affin.ity of oxygen f o r calcium at 1000" (see above). Under these conditions, quantitative removal of oxygen is impossible. Calcium hydride is also used a s a reducing agent f o r Ti&. A s a reducing agent, magnesium is inferior to calcium. When TiOz is reduced with magnesium the reaction proceeds preferentially until the lower oxide T i 0 o r solid solutions of oxygen in titanium a r e formed. Reduction by calcium The reduction involves the following exothermic reaction: Ti&

+ 2Ca

= Ti

+ 2Ca0 + 85.4 kcal.

The heat evolved (about 530 kcal p e r kg of stoichiometric mixture) is not sufficient to maintain a spontaneous reaction, and external heating is necessary.
199

... ..........................-.-,...,. ..... I!. .................................


.I._ 1 1 1 . . . 1 . . . . . 1

I ,

................

The reduction is c a r r i e d out in argon atmosphere at 1000 t o 1100". At that temperature the calcium is liquid, and a fraction of it is in the gaseous stage (the vapor p r e s s u r e of calcium at 1000" is l l m m H g ) , which e n s u r e s good contact between the calcium and the titanium dioxide. The calcium used in the reduction must be preliminarily purified by distillation and f r e e f r o m nitrogen and carbon (in o r d e r to prevent contamination of titanium by t h e s e elements). The reduction yields fine grained titanium (particle s i z e 2 to 3 ~ since ) the presence of s t r a t a of r e f r a c t o r y calcium oxide i n t e r f e r e s with the growth of titanium particles. The growth of the p a r t i c l e s is favored by the addition of CaClz (about one mole of CaClz p e r 2 moles of CaO), which f o r m s a liquid phase.

FIGURE 104. Reactor for the reduction of titanium dioxide with


calcium.
1-body (retort); 2 -steel crucible; 3 -pelletized charge:
4-lid; 5-thermal insulation packing; 6-electrical furnace;
I -seals; 8 -sleeve connections t o the pump or the argon source.

The reduction is c a r r i e d out in hermetically sealed r e a c t o r s made of r e f r a c t o r y steel, which a r e charged with a pelletized mixture of titanium dioxide, calcium chloride, and calcium ( F i g u r e 104). The calcium used is in the f o r m of fine particles or turnings. The r e a c t o r is evacuated, then filled with argon, heated to 1000 to l l O O o and held at that temperature f o r about one hour. In o r d e r to prevent partial removal of calcium from the reaction zone by condensation on the colder reactor lid, the r e a c t o r is placed in a furnace in which a uniform temperature is maintained.

200

The reduction product is crushed, treated f i r s t with a l a r g e volume of water and then with dilute acids (acetic, nitric, or hydrochloric), washed with water, and vacuum-dried a t 40 to 50". In o r d e r to reduce the oxygen concentration in the powder to about 0.270 it is n e c e s s a r y to c a r r y out a second reduction of the powder by calcium. Even after the second reduction the metal is characterized by g r e a t e r h a r d n e s s and lower ductility a s compared with titanium produced by the reduction of TiC1,; this is a result of i t s higher oxygen and nitrogen contents. Titanium produced by t h e r m a l reduction with calcium has the following composition: 98.5 to 99.070 Ti, 0.03 to 0.1570 N, 0.2 to 0.4% 0, 0.01 to 0.03% H, 0.1 to 0.270 Si, 0.01 to 0.0570 C, 0.10 to 0.2570 Fe, 0.05 to 0.1570 Al, 0.1 t o 0.370Ca, < 0.0370 Mg, 0.01 t o 0.1%. Cu. Titanium powders produced by the reduction of T i Q with calcium a r e used f o r the manufacturing of solid titanium b a r s and products by powder metallurgy methods and for the production of alloys of titanium and other metals. Moreover, they may be used a s the starting m a t e r i a l for the production of high-purity titanium by t h e r m a l dissociation of titanium iodide or by electrolytic refining ( s e e below).

Reduction by calcium hydride The reduction by calcium hydride (CaH2) is a variant of the t h e r m a l reduction of TiOz with calcium. Calcium hydride is produced by the action of dry hydrogen on calcium a t 400 to 600". The hydride decomposes above 800", yielding calcium and hydrogen. The nascent atomic hydrogen instantaneously combines into hydrogen molecules. At 970" the equilibrium hydrogen p r e s s u r e over the hydride is almost one atmosphere. The reduction of Ti02 by calcium hydride involves the following over-all reaction: TiO, + 2CaH, + TiH, + 2Ca0 + H,. The change in the f r e e energy of the reaction a t 1000" is AFlooo0=-66.1 kcal. Meerson and Kolchin, who studied the chemistry of the process, showed that the main reducing agent is calcium r a t h e r than atomic hydrogen ( a s was previously assumed). In contrast with the thermal reduction by calcium, the reduction yields titanium hydride or a solid solution of hydrogen in titanium. This is an advantage since in the leaching of the calcium oxide the titanium hydride powder i s oxidized to a l e s s e r extent than is titanium powder. When pellets made of compressed titanium hydride a r e sintered in vacuo, the hydride decomposes with the evolution of atomic hydrogen which causes partial reduction of the oxide films and thus reduces the oxygen content of the metal. Calcium hydride is readily decomposed by atmospheric moisture:
CaH,

+ 2H,O

-+

Ca (OH),

+ 2H,.

Hence, it must be stored as lumps in hermetically sealed containers, and the crushing and charge mixing (TiOz + CaH2) must be c a r r i e d out in sealed m i l l s and mixers.

20 1

The reduction p r o c e s s is c a r r i e d out in equipment made of refractory steel, which is charged with the pelletized titanium dioxide - calcium hydride mixture. The air in the r e a c t o r is evacuated and the reactor is filled with dried hydrogen and heated to 900 to 1100". The powder produced is washed with water and dilute acids (hydrochloric or acetic) to remove the excess calcium and calcium oxide. Powdered titanium hydride is mainly used in the powder metallurgy of titanium.

50. ELECTROLYTIC REFINING OF TITANIUM AND TITANIUM-BASED ALLOYS

A l a r g e proportion of the total cost price of titanium produced by thermal reduction methods is constituted by the cost of the reducing agents (magnesium or sodium) which a r e manufactured by electrolysis of molten chlorides. The cost p r i c e of titanium may be substantially cut by replacing the thermal reduction methods by direct electrolytic processes. Extensive r e s e a r c h is being c a r r i e d out on the subject, but there is a s yet no electrolytic method f o r the production of titanium which could compete with the manufacturing processes currently used. On the other hand, electrolytic refining of titanium (e. g . , of low-quality sponge, or titanium melting wastes) and titanium-based alloys can be c a r r i e d out. The electrolytic refining of titanium i s already used in some industrial plants. In the electrolytic refining, the impure titanium is used a s the anode and is inserted into the molten electrolyte. During the electrolysis, the titanium anode dissolves in the melt and is then deposited on the steel cathode. Molten alkali metal chlorides (NaC1 or a NaCl - KC1 mixture) containing 1.5 to 5'7'0 of dissolved lower titanium chlorides (Tiel2, Tiel3) a r e used a s the electrolyte. The electrolyte is prepared by the reduction of T i e l 4 with sodium or titanium s c r a p in molten sodium and potassium chlorides. During the anodic dissolution the titanium p a s s e s into the electrolyte in the form of Ti2+and, to a certain extent, TiS+ions:
Ti - 2e-+Tiz+, Ti2+ -e --f Ti3+.

The r e v e r s e reactions take place on the cathode:


Ti3+ e + Tizf, Tiz+ + 2e + Ti.

The main impurities in the crude titanium a r e oxygen, nitrogen, carbon, iron, and silicon. Titanium-based alloys also contain various alloying additives - Al, Cr, Mn, V, Mo, and Sn. The electrolytic refining of titanium is based on the difference between the electrode potentials of titanium and the contaminants. During the anodic

202

dissolution of titanium the oxygen present a s impurity remains in the anodic slag a s titanium oxides (Ti@, TiZQ). The carbon in a f r e e state remains on the surface of the electrolyte o r in the form of carbide in the anodic residue; the nitrogen remains in the anodic residue a s nitride, or is evolved as gas at the anode; silicon is evolved with the g a s e s a s SiCI,; the iron and the nobler metals (Ni, Cu, Sn) accumulate in the anodic residue. The anodic dissolution potentials of s e v e r a l alloying additives (AI, Cr, Mn, V) are close to that of titanium. Thus, they p a s s into the molten chloride and, when their concentration reaches a certain level, they may be deposited on the cathode together with the titanium. The refining conditions depend on the content of impurities. When various elements a r e present together, they have a combined effect on the anodic dissolution. The removal of impurities such a s 0, N, C, Fe, and Si and of the alloying additives Mo and Sn is easiest. It is m o r e difficult to remove V, Al, and Mn present a s impurities. These elements may be removed by twice-repeated refining. Electrolyzers with packed anodes a r e the ones most often used in electrolytic refining. A diagram of such electrolyzer is shown in Figure 105. The metal to be refined (in the form of turnings or granules, 2 to 3 m m in size) is placed in a perforated steel basket which s e r v e s as the anode. When a deposit of titanium accumulates on the cathode, the cathode is raised into the upper chamber, a discharge pallet is placed underneath and the deposit i s cut out with a special knife.

FIGURE 105. Diagram of electrolyzer used for electrolytic refining


of titanium.
1-stainless steel cathode; 8-chamber for the removal of the cathodic deposit; 3-body of the electrolyzer (anode); 4-basket for the packed anode: 5 -anodic titanium; 6-electrolyte: I-discharging device: 8 -pallet; 9 -cathodic deposit.

203

The electrolysis is c a r r i e d out a t 850" in an argon atmosphere, a t an anodic c u r r e n t density of 0.1 to 0.5 amp/cm2 (depending on the composition of the titanium to be refined) and a starting cathodic c u r r e n t density of 0.5 to 1.5 amp/cm2. A c o a r s e crystalline deposit of titanium f o r m s on the cathode. The c u r r e n t efficiency is 90% and higher. The use of electrolytic refining for the purification of crude titanium produced by d i r e c t reduction of titanium slags (e. g., by aluminum o r magnesium) is of g r e a t interest. Research on the subject is now in progress

51.

REFINING OF TITANIUM BY THE IODIDE PROCESS

Sma1.1 amounts of extremely pure ductile titanium were produced a s e a r l y a s 1925 by van Arkel and de Boer. Their method was based on the thermal decomposition of gaseous Ti14 on a surface (e. g . , an incandescent tungsten or titanium wire) heated to a high temperature (1300 to 1500), which resulted in the deposition of the titanium on that surface. At present, the t h e r m a l dissociation of titanium iodide is used f o r s m a l l s c a l e production of high-purity titanium. The iodide p r o c e s s f o r the refining of titanium may be represented by the following scheme: Ti
(contaminated)

+ 212(vapor) >t

LOO--2000

Tih1300-L5000

Ti

+ 212 (vapor).
I

(vapor)

(pure)

Titanium r e a c t s with iodine a t a low temperature (100 to 200"). The gaseous titanium iodide is then made to dissociate on the surface of a wire heated to 1300 to 1500". The iodine freed in the dissociation again r e a c t s with the contaminated titanium, at a l o w e r temperature, which is present in the vessel. A l a y e r of p u r e titanium is gradually deposited on the surface of the wire. Dense r o d s or coarsely crystalline, l e s s dense deposits a r e obtained, depending on the p r o c e s s conditions. The elimination of oxygen and nitrogen during the refining of titanium by the iodide p r o c e s s is due to the fact that titanium oxides and nitrides do not r e a c t with iodine at the low temperatures required for the formation of TiI,.

Equipment The iodide p r o c e s s was originally c a r r i e d out in g l a s s containers with tungsten terminals (which do not r e a c t with the iodine) sealed into the glass. A titanium filament was suspended between the leads, and heated by electrical c u r r e n t supplied to the terminals. The purified titanium was deposited on the bottom and walls of the container. Industrial s c a l e refining of titanium by the iodide p r o c e s s is now c a r r i e d out in metal installations. One of the designs used is shown in Figure 106 1 2 2 1 . The body of the apparatus is made of chromium-nickel alloy (80% Ni, 2 0 % Cr) which is resistant to iodine and Ti14. Crude titanium powder or turnings a r e placed along the inner walls of the apparatus in the c i r c u l a r gap formed by a cylindrical s c r e e n of molybdenum gauze.

204

Titanium wire 3 to 4 m m in diameter (in the form of U-shaped loops) is stretched with the aid of tungsten hooks fastened to steatite insulators. The total length of the filament is about 11 meters. The ends of the filament a r e fastened t o molybdenum terminals. The airtight lid is fitted with a socket holder f o r the g l a s s ampoule containing the iodine, and with sleeve connections to the vacuum pump. The r e t o r t is placed in a thermostat which maintains the temperature of the titanium undergoing purification a t the desired level (between 1 0 0 and 200"). The r e a c t o r is first evacuated to a residual p r e s s u r e of 2 5 The r e a c t o r is then disconnected from the evacuation Hg. system and the iodine is introduced. In o r d e r to introduce the iodine, a special device is used to break the drawn-out tip of the ampoule.

FIGURE 106. Apparatus for the production of titanium by thermal dissociation of titanium iodide.
1 -body of the apparatus: 2-molybdenum screen; 3-sup port of titanium filament: 4 -terminals; 5-thermostat. with glass ampoule containing iodine; 6-vacuum lock; 7 --sleeve connection t o the vacuum system: 8-lid: 9 molybdenum hooks; 10 -titanium filaments; 11-titanium sponge.

205

Electric c u r r e n t is then applied to the filament. The amount of iodine introduced into the r e a c t o r is determined by the optimum p r e s s u r e of the iodine vapor. In practice, the amount of iodine added is 7 to 10% of the amount of titanium to be refined. The installation is designed to produce 24 kg of refined titanium p e r working cycle, or 1 0 kg p e r day.

Procedure Refining titanium by the iodide p r o c e s s comprises the following stages : 1) interaction of titanium and iodine at low temperatures (100 to 200") with the formation of Ti&: 2) transport of titanium iodide vapors to the surface of the heated filament ; 3) decomposition of the iodide on the surface of the filament, at 1300 to 1500"; 4) transport of iodine vapors from the filament to the crude titanium.
24
22
20 18 16
It

I4
>
a M l
u 0

I O
8 6
0

I W

Z W

3m

4 c M

5 1 1 0

T i 4 pressure, m m Hg FIGURE 107. Rateof deposition of titanium a t a filament temperature of 1500", as a function of the vapor pressure of Tire

Current. a m p FIGURE 108. Voltage-current curves for the de position of iodide titanium a t various deposition temperatures.

Investigations have shown that the process r a t e is determined by stages 2 and 4, i. e . , by the r a t e of transport of the iodide to the filament and of the iodine to the raw titanium. The r a t e of t r a n s f e r of the iodide depends mainly on the vapor p r e s s u r e Ti14 in the vessel, determined by the temperature of the reactor walls.
A s is evident from Figure 107, at a filament temperature of 1500" the r a t e
of deposition is at the maximum at iodide p r e s s u r e s from 2 to 20 m m Hg,
which corresponds to a reactor-wall temperature between 140 and 200".
The r a t e of deposition of titanium d e c r e a s e s at higher iodide p r e s s u r e s
since the degree of dissociation of iodide then becomes s m a l l e r 1341.
Although the maximum r a t e of deposition corresponds to a wall tempera t u r e of 140 to 200, the p r o c e s s is also fairly rapid at 100". The main tenance of a constant temperature is then accomplished simply by im mersing the r e a c t o r in boiling water. The temperature of the filament is the second factor affecting the rate of the deposition. The r a t e of deposition is highest at 1500". Vacuum

206

vaporization of titanium is observed at higher temperatures. The filament temperature is usually maintained at 1300 to 1400O in o r d e r to prevent the fusion of the filament. One s e r i o u s difficulty is to maintain a constant temperature on the filament surface during the accumulation of the deposit. Neglecting the heat l o s s e s by heat conductance through the electrodes and by convection (which are relatively small), we may assume that the electrical energy supplied s e r v e s to compensate f o r the heat lost by radiation emitted by the surface of the filament on which the titanium is deposited. In o r d e r to maintain the temperature of the filament constant a s its diameter grows, i t is necessary that the c u r r e n t intensity and the voltage be s o adjusted that the power radiating from unit incandescent surface remains constant. The following equation must then be satisfied: where
I.!?
= Y = const,

I is the c u r r e n t intensity; E is the voltage,


T h e above equation is easily derived. T h e maintenance of a constant temperature requires that

W --C=const.
where W=I E is the radiating energy; S = n D L is the filament surface (D-diameter. L-length); C = e a' T is the energy radiated by unit surface a t the given temperature; e is a coefficient showing the ratio of the true to t h e radiation temperature; a is the Boltzmann constant. T h e total power radiated by a filament of a length L is:

W = I E = CKDL.
T h e resistance of the filament is:

(1)

R=-= E& ,

I n D '
d

where p is the specific resistance squaring equation (1) and combining it with equation (2) we obtain
IEJ = 4spCBL3

At a given temperature and filament length, the right-hand side of the equation is constant. Hence

IEa = K

= const.

The value of K = 4np C2L3 may be determined if the specific radiation power C and the specific resistance p a r e known. F o r titanium at 1350" C = 21.8watt/cm2 and p = 240. ohm. cm. The effect of the voltage on the c u r r e n t is plotted a s a curve (the voltage-current curve) from the calculated value of Y; this curve i s used to control the process. The voltage-current curve for 1350" is shown in Figure 108 / 2 2 / . Under conditions adopted in industrial practice, the growth r a t e of the titanium rod is about 1 0 to 20mm/day. A comparison between the concentrations of impurities in titanium made by the iodide p r o c e s s and in titanium made by thermal reduction with magnesium is shown in Table 3 6 . As compared with titanium made by thermal reduction with magnesium, iodide- refined titanium contains lower amounts of oxygen, nitrogen, iron,

207

l1ll1l1Il l l

magnesium, and manganese. The mechanical p r o p e r t i e s of both types of titanium a r e listed in Table 27. The titanium refined by the'iodide p r o c e s s is much m o r e ductile.
TABLE 36 Composition of titanium made by the iodide process and by thermal reduction with magnesium,in%. Titanium sponge (reduced with magnesium)
0.01-0.03 0.05-0.15 0.01 -0.05 < 0.005 < 0.03 0.03-0.2

Element

Iodide titanium

Element

Iodide titanium

Titanium ,ponge (reduced with magnesium)

Carbon Oxygen Nitrogen Aluminum Copper Iron

0.01 0.03 0.005-0.01 0.001 -0.004 0.013- 0.05 0.0015- 0.002 0.0035- 0.025

Magnesium Manganese Molybdenum Nickel Silicon Tin

0.0015- 0.002 0.005-0.013 0.0015 0.003 0.03 0.001-0.01

0.04- 0.12
0.03-0.06

< 0.01 < 0.03

52.

PRODUCTION OF SOLID TITANIUM

The titanium sponge o r powder prepared by one of the methods described above i s converted into solid, ductile titanium by melting in an a r c furnace o r by powder-metallurgy techniques. The melting p r o c e s s is the most widely used; l a r g e ingots, weighing f r o m one to f o u r tons may be produced in this way. Powder metallurgy techniques a r e used for the manufacture of s m a l l b a r s o r parts.

The melting of titanium /1, 3, 6, 23, 24/ The melting of titanium is made difficult by the rapid reaction of titanium with g a s e s and by the interaction of the molten metal with all known r e f r a c t o r y materials: In the e a r l y development of titanium production, the titanium sponge was melted in induction furnaces with graphite crucibles, in argon atmosphere or in vacuum. However, even the relatively s m a l l amount of carbon absorbed by the titanium during the melting (0.25 to O.80/) i m p a i r s i t s physical and mechanical p r o p e r t i e s to a considerable extent, Hence, a titanium-melting method which was subsequently developed and which is now extensively used consists of arc-melting in a cooled copper crucible. The melting is c a r r i e d out in vacuum arc-melting furnaces with consumable electrodes, whose operation and design have been described in Chapter 11. The vacuum melting resulted in a reduction of the hydrogen concentra tion in titanium ingots to 0.001 to O.O02%(the hydrogen concentration in titanium melted in argon was about 0.020/0), with consequent improvement of the mechanical properties of titanium, and especially of i t s impact strength.
208

Various types of furnaces with consumable electrodes a r e used f o r the melting of titanium. In the most widely used @pes of furnaces the consumable electrode of the desired length is prepared outside the furnace and is then inserted in the furnace which contains a long, water-cooled copper mold (Figure 109). The consumable electrode is welded to the electrode holder ( a r m ) which s e r v e s to supply the c u r r e n t to the electrode and to move it during the melting. As a safety precaution, a r c furnaces a r e mounted in a chamber with thick reinforced concrete walls and ceilings. The control panel is located behind a concrete wall. The consumable electrodes a r e prepared by p r e s s i n g titanium sponge (particle s i z e 5 to 3 0 " ) o r powder in hydraulic p r e s s e s fitted with s t e e l dies, under p r e s s u r e s of 2 to 4 tons/cm2. In most c a s e s the titanium i s p r e s s e d into cylindrical pellets of s m a l l height (diameter 500") which a r e then welded together in an argon a r c into electrodes of the desired length. In some c a s e s the pellets a r e joined together to form the FIGURE 109. Diagram of electricconsumable electrode by sintering with the arc furnace with consumable elec aid of contact heating in the melting furnace trodes. itself. To accomplish this, the pellets a r e 1-current leads t o the mold: 2 put in the furnace one over the other. The sleeve connection to vacuum: 3 furnace is evacuated, the electrode holder flexible bus bar supplying the is welded to the upper pellet and a p r e s s u r e current to the electrode: 4-de of about one ton i s applied to the pellet column vice for lifting the electrode: 5-electrode holder (arm): 6 by means of the feed mechanism. The current vacuum seals: I-lid: 8-furnace from the generators used f o r the melting i s body: +water inlet and outlet then turned on. The contact s i t e s between the tubes: 10-rubber seals: l l - c o n pellets, which have a high resistance, a r e sumable titanium electrode: 12 rapidly heated to about 800". At this tempera cooled copper mold: 13-water ture and p r e s s u r e the pellets rapidly s i n t e r jacket: 14-ingot: 15-cooled and cohere. The sintering i s accompanied bottom of the mold. by partial degassing of the pellets (i. e . , elimination of dissolved hydrogen, magnesium, and other volatile i m purities). In a vacuum of about 0.1 to 0.01 m m Eg, an a r c may be formed because of the presence (in the narrow gap between the consumable electrode and the melt) of ionized vapors of titanium and gaseous impurities and metals evolved during the melting (sodium, magnesium). It is recommended that the melting be c a r r i e d at a residual p r e s s u r e (in the furnace chamber, i. e . , in the space above the melting crucible) of 0.01 m m Hg. Under these conditions the actual p r e s s u r e in the crucible ( a t the molten metal level) is always higher and attains about 0.1 mmHg. The selection of vacuum pumps for the a r c furnace will be determined by the necessity for evacuating the g a s e s evolved and f o r maintaining a certain vacuum. The magnitude of the gas flow created during the melting

209

depends on the concentration of volatile substances in the sponge and on the melting rate. Thus, at a melting r a t e of 5 kg/min the gas flow i s :
Concentration of gases in the sponge, I/kg Gas flow*. mm Hg l/sec..

... .... 0.25 ..................... 15.9

1 63.4

2 126.8

Vacuum arc furnaces may operate on alternating or direct current, but furnaces operating on d i r e c t c u r r e n t (in which the electrode s e r v e s a s the cathode and the melt as the anode) a r e p r e f e r r e d , In DC furnaces the melting consumes about 2 1 3 of the total amount of energy, and in A C furnaces about 112. As a consequence, a l a r g e r volume of the metal is in the molten state and favorable conditions a r e created f o r the production of a uniform ingot. The c u r r e n t i s rectified with the aid of motor generators or (which i s much m o r e convenient) with selenium or germanium rectifiers. A c u r r e n t of 8000 to 9000 amp at 25 to 30 V is used f o r ingots 350" in diameter. Under these conditions the melting r a t e v a r i e s between 3.7 and 4.5 kg/ minute. The consumption of electrical energy is about 4.5 to 4.8 kwh/kg titanium. A stable arc discharge must be maintained during a r c melting, and formation of s e c o n d a r y a r c s between the electrode and the crucible walls a s well as the g l o w d i s c h a r g e must be prevented. Glow discharge is observed between the a r c electrodes at a certain c r i t i c a 1 p r e s s u r e of the gases. During the glow discharge the a r c between the electrodes is extinguished and the discharge takes place through the bulk of the ionized gas, in the entire volume of the crucible. The glow discharge is accompanied by a strong luminescence of the gas. A magnetic coil (solenoid) fitted off-center with respect to the crucible is used to stabilize the a r c in the vacuum. The external magnetic field p r e s s e s the secondary a r c s against t h e crucible walls ( a s concentric rings) and they become so long that the applied voltage i s not sufficient to main tain them. Moreover, the presence of the magnetic coil r e s u l t s in the mixing of the metal, which improves i t s uniformity. Secondary a r c s between the electrode and the crucible walls a r e ob served during the gas ionization caused by the spattering of the molten titanium. The spattering may be reduced to a minimum by working at a s m a l l distance between the electrode and the melt ( 5 to 10") and using a narrow gap between the electrode and the crucible wall ( 2 0 to 30"). Moreover, the spattering impairs the quality of the ingot surface, since droplets of the metal solidify on the crucible walls without joining the bulk of the ingot, which must be machined to a considerable depth, thus reducing the useful yield of the metal. The l o s s e s a r e often a s high a s 10 to 15%. The melting of titanium alloys The refractoriness and mechanical properties of titanium a r e improved by the addition of alloying additives (manganese, aluminum, chromium, vanadium, molybdenum, iron, and other metals) in concentrations of a few percent.

* In vacuum technology
t the time.

___--__--__---_

the gas flow has the dimensions p.V/: where p is the pressure,

the volume and

210

Since only a s m a l l amount of the metal is in the molten s t a t e during a r c melting, i t is difficult to e n s u r e the uniformity of the alloy. It is advisable to mix the alloying additive with the sponge before p r e s s i n g the consumable electrode. A less satisfactory method consists in a uniform distribution over the melting zone of pellets made of the alloying additives. I n all these methods, the production of a uniform alloy r e q u i r e s a second melting, in which the ingot produced by the first melting is used a s the consumable electrode.

Production of solid titanium by powder metallurgy techniques f 2 f Solid ductile titanium may be produced by powder metallurgy techniques, provided the oxygen and nitrogen contents of the starting powder do not exceed the maximum permissible limit. The starting m a t e r i a l s used in the production of titanium by powder metallurgy techniques a r e titanium produced by t h e r m a l reduction with calcium hydride ( s e e Section 49) or by grinding a titanium sponge. Titanium sponge is ductile, which makes grinding difficult. Most often titanium sponge is converted to a powder by saturation of the sponge with hydrogen; titanium hydride i s then formed, which is brittle and i s readily ground. The saturation with hydrogen is c a r r i e d out in hermetical ly sealed s t e e l r e a c t o r s . The sponge is f i r s t heated in vacuum to 800", then cooled to 400 to 450" and purified (nitrogen- and oxygen-free) hydrogen is admitted. A t f i r s t the reaction is quite vigorous and is accompanied by the evolution of heat. Hence, the hydrogen must be admitted gradually, and must be diluted in the f i r s t stage with an i n e r t gas. Very brittle titanium hydride is obtained when the hydrogen content attains 3 to 4 % . The hydride i s ground in ball m i l l s to a particle s i z e of 0.05 to 0.12". This corresponds to a powder with an average specific surface of 740 cm2/g. F i n e r hydride powders actively absorb nitrogen and oxygen. Powdered titanium hydride, being m o r e brittle, i s m o r e difficult to p r e s s than the powders obtained by grinding the sponge. The b a r s obtained a r e l e s s strong but a r e sintered m o r e rapidly than b a r s made of non hydrogenated titanium powder. Titanium o r titanium hydride b a r s a r e p r e s s e d in s t e e l diesunder p r e s s u r e s from 3 . 5 to 8 tons/cm2. Large b a r s weighing 50 kg or m o r e a r e most conveniently pressed by the use of the hydrostatic technique discussed in the chapter "Molybdenum". The sintering is c a r r i e d out at 1000 to 1300" in vacuo (10-4mmHg). The satisfactory sintering obtained a t these relatively low temperatures may be attributed to a certain extent to the fact that an cy -+ transformation (i. e . , the conversion of the hexagonal form of titanium into the cubic form) takes place a t 880, which i n c r e a s e s the mobility of the atoms. Moreover, the sintering of titanium hydride is favored by the dissociation of the hydride (which takes place at these temperatures), which also i n c r e a s e s the mobility of the atoms. During the sintering of b a r s made of titanium hydride powders, the bulk of the hydrogen is evolved at 500 to 800, which means that the temperature r i s e in this range must be slow. The temperature is then increased to 1200 to 1300".
21 1

The final porosity of articles made of titanium hydride does not exceed 270, the linear contraction being 1 2 to 14%. Thus, the density of a r t i c l e s made of p r e s s e d titanium hydride (density when p r e s s e d 3.2 to 3.8 g/cm3) sintered f o r 8 hours a t 1300O i n c r e a s e s to 4.45g/cm3. Because of the l a r g e contraction during the sintering, hydride powders cannot be used f o r the production of standard s i z e a r t i c l e s , The l i n e a r contraction during sintering (for 15 hours a t 1000" and 4 hours a t 1200O) of c o a r s e titanium powders produced by the grinding of titanium sponge is only 4 to 570. The production of solid m e t a l involves intermediate forging (compression) of the m e t a l followed .by a second sintering. One of the advantages of powder metallurgy is that the resulting alloys a r e uniform. The pulverized alloying constituents must be throughly mixed before p r e s s i n g with titanium or titanium hydride powder. Large b a r s weighing 50 to 6 0 kg may be sintered in vacuum induction furnaces. In addition to the conventional powder metallurgy method, another method used in titanium metallurgy consists of combined pressing, sintering, and mechanical treatment under p r e s s u r e . A description of the method is given below. The titanium powder i s tightly packed in a steel tube, the tube ends a r e sealed by welding, and the tube with the powder is hot-rolled at 900". The s t e e l shell protects the metal against oxidation. After the rolling the s t e e l shell is cut and is readily separated from the titanium sheet a s a result of the formation of a thin intermediate l a y e r of a titaniumi r o n alloy. The mechanical properties of titanium produced by powder metallurgy techniques a r e virtually the s a m e a s those of arc-melted titanium. How ever, powder metallurgy cannot supersede the melting method because of the limited dimensions of the bars. The powder metallurgy method is m o r e suited to m a s s production of s m a l l sized titanium and titanium-alloy a r t i c l e s . Its advantages a r e : the economical utilization of the metal (finishing operations a r e few), production of a r t i c l e s of any desired density, and the production of uniform alloys.

212

Chapter V
ZIRCONIUM
53. GENERAL DATA ON ZIRCONIUM

Zirconium belongs to Group IV of the periodic system. It was discovered by Klaproth in 1789 in i t s most abundant mineral - zircon. This mineral has been known since antiquity and has been used a s a precious stone under various names (hyacinth, jacinth, jargon). Berzelius was the f i r s t to p r e p a r e metallic zirconium (in 1824) in the form of a very impure powder, by reduction of potassium fluoro zirconate with potassium. P u r e ductile zirconium was prepared only in 1923 ( i . e . , 136 y e a r s a f t e r the discovery of the element) by thermal dissociation of zirconium iodide. This complex p r o c e s s , which was developed by van Arkel and de Boer, was until recently the only available method for the production of ductile zirconium. Methods for l a r g e - s c a l e production of ductile zirconium have only been developed in the l a s t decade. Zirconium compounds and metallic zirconium produced from o r e s always contain hafnium (the chemical analog of zirconium, which was discovered in 1923) a s an impurity. The concentration of hafnium in the zirconium depends on the nature of the starting raw m a t e r i a l and ranges from a fraction of one percent to s e v e r a l percent.

Properties of zirconium

11, 2 1

The appearance of solid zirconium r e s e m b l e s that of s t e e l ; zirconium powder has a dark-gray color. Zirconium appears in two corystalline f o r m s : the hexagonal Q = form (lattice p a r a m e t e r s a = 3.223 A , c = 5.123 A ) , which i s stable up to 862", and the face-centered cubic form p (lattice p a r a m e t e r s a = 3.61 A), which is stable above 862". Hafnium exists in the same forms. However, the transition from the a - t o the 8-form of hafnium takes place at a higher temperature (1310 f 1 00). Some of the physical properties of zirconium and i t s analog, hafnium, a r e given in Table 37. P u r e zirconium is ductile and can be readily worked by forging, rolling into sheets, and drawing. The presence of oxygen, nitrogen, carbon and hydrogen strongly affects the properties of zirconium, and i t s p r o p e r t i e s therefore depend on the method of production of the metal ( s e e Table 42).

213

TABLE 37 Physical properties of zirconium and hafnium Property Atomic number Atomic weight Density, g/cmS Melting point, "C Boiling point. " C Transition temperature to superconducting state, 'K Specific heat, cal/g, ' C 25 -100' 1000-1500' Linear expansion coefficient Total radiation (watt/cm2) at various temperatures,
"C

Zirconium 40 91.22 6.52 (for Q zr) 1852 i 10" 3600 0 .7

Hafnium

72 178.6 13.3 (for a - H D 2130 i 15: 5400 0.35

0.066 0.113 8 9 - lo-' at 20-7OO'C)

0.035

5 9 . 10-1 (at 0-10OO'C)

927 1127 1227 1327 Electrical resistivity, o h m . cm, at 'C: 20 800 Thermal neutron capture cross section, barns

2.03 5.40 7.20 10.0


41 * 143. 0.18* 0.02
_ _ . -

Zirconium i s stable in air. When the solid metal is heated to 400 to 600", it i s coated with an oxide film, but above 800" i t becomes rapidly oxidized with the formation of the higher oxide Zr02. Zirconium powder i s oxidized in the a i r with spontaneous ignition a t 180 to 285" (depending on the particle size). Zirconium actively absorbs hydrogen at temperatures a s low a s 300 to 400, with the formation of a solid solution of hydrogen in zirconium and of zirconium hydrides (ZrzH, ZrH, and ZrHz). The hydrogen can be expelled from the metal by heating to 1200 to 1300" in a high vacuum. Above 900" zirconium rapidly absorbs nitrogen and r e a c t s vigorously with carbon monoxide. Zirconium forms very hard and r e f r a c t o r y compounds with nitrogen and carbon - the nitride ZrN (mp 2930") and the carbide Z r C (mp 3530"). In contrast to hydrogen, oxygen and nitrogen cannot be removed from zirconium by heating in vacuo, The corrosion resistance of zirconium is superior to that of titanium and approaches that of tantalum and niobium. Thus, below 100" zirconium r e s i s t s hydrochloric and n i t r i c acids of all concentrations and of sulfuric acid of concentrations up to 50%. Zirconium is r e s i s t a n t to aqua r e g i a a t room temperature, dissolves in hydrofluoric acid and in concentrated sulfuric acid a t loo", but is not dissolved by aqueous solutions of alkali hydroxides.

214

Chemical properties

/ 171

Compounds of tetravalent zirconium a r e the only ones of practical importance; lower valency compounds a r e unstable and have been in adequately studied. Zirconium dioxide. Zirconium dioxide Z r Q is produced by ignition of zirconium hydroxide o r other zirconium s a l t s - sulfates, nitrates. The pure dioxide is white. After the ignition, Z r Q is virtually insoluble in hydrochloric, nitric and dilute sulfuric acids. The dioxide dissolves in hydrofluoric and in hot concentrated sulfuric acids. Zirconium dioxide is very stable and has refractory properties. The heat of formation of Z r Q is 259.5 kcallmole, its melting point is 2700 to 2900". Z r Q exists in s e v e r a l crystalline modifications. The monoclinic form is stable up to 1000 to llOOo, the tetragonal at 1100 to 1900", and the trigonal above 1900. Zirconium dioxide is amphoteric. The fusion of ZrOz with alkalies yields zirconates ( s a l t s of zirconic acid), while the dissolution of the di oxide in acids r e s u l t s in the formation of the respective s a l t s - zirconium sulfates, chlorides and nitrates. In aqueous solution these s a l t s hydrolyze with the formation of zirconyl (ZrO") s a l t s , e. g. :
ZrC1,

+ H,O

= ZrOCI,

+ 2HC1

Zirconates. Salts of metazirconic ( H 2 Z r Q )and orthozirconic ( H 2 Z r 0 4 ) acids a r e known. Most zirconates a r e insoluble in water. S o d i u m a n d p o t a s s i u m z i r c o n a t e s a r e produced by the fusion of Z r Q with alkali hydroxides, sodium carbonate or potassium carbonate. They a r e sparingly soluble in water, but undergo hydrolytic decomposition by reactions of the type: Na, ZrOs $- 2H,O 2 ZrO (OH), 2NaOH.

C a l c i u m a n d m a g n e s t u m z i r c o n a t e s (CaZr-03 and MgZr03) a r e produced by heating a mixture of Z r Q with CaO or MgO powders to 1400 to 1600". The melting points of calcium and magnesium zirconates a r e 2 3 5 0 and 2150" respectively. Sulfates. The anhydrous zirconium sulfate Z r (Sod2 ( a white crystalline substance) is formed when ZrOz i s heated with concentrated sulfuric acid. Its dissolution in water is accompanied by hydrolysis and complex-forma tion: Zr (So,), HzO 2 (ZrO) SO4 H,S04, ZrO (SO,) HzS04ZH, [ZrO (SO,),] - zirconylsulfuric acid.

Upon evaporation of the sulfuric acid solution, zirconylsulfuric acid crystallizes out a s the trihydrate H,[Zro(SO~,]* 3&0. The formula of the normal zirconium sulfate Zr( S04)2 4Hzo is occasionally incorrectly ascribed to the trihydrate; from the standpoint of chemical composition both formulas a r e identical. When a weakly acid solution containing zirconyl sulfate is boiled, hydrolysis takes place with the formation of sparingly soluble basic zirconium sulfates whose composition can be expressed by the general formula xZrOp. ySOs zHnO. In the basic sulfates the m o l a r r a t i o Z r Q : SO, > 1.

Halides. Z i r c o n i u m c h l o r i d e a n d o x y c h l o r i d e . Zirconium tetrachloride ZrCI, is one of the starting m a t e r i a l s for the production of zirconium metal. The chloride is formed a s a r e s u l t of the high-tempera t u r e interaction of chlorine with a mixture of zirconium dioxide and carbon or zirconium carbide: 210, 2C 2C1, = ZrCI, 2CO; ZrC 2C1, = ZrC1, + C.

+ + +

The chloride is a white crystalline powder which sublimes a t low temperatures. The vapor p r e s s u r e over the solid s a l t is 1 atm at 330". The s a l t m e l t s a t 437" under a p r e s s u r e of 18.7 atm. The heat of formation of solid Z r C , is 2 3 1 . 9 kcal/mole, the heat of sublimation i s 28.5 kcal/mole. Zirconium chloride is very hygroscopic and is hydrolyzed in aqueous solutions (and in humid a i r ) with the formation of zirconium oxychloride ZrOCla (zirconyl chloride). Zirconium oxychloride is stable in aqueous solutions. It crystallizes out a s the crystalline hydrate ZrOClz' 8 H z 0 . The c r y s t a l s a r e in the shape of tetragonal p r i s m s with well-defined cleavage. The s a l t i s readily soluble in water and sparingly soluble in concentrated hydrochloric acid ( s e e Figure 1 1 3 ) . ZrOC12. 8 H 2 0 may be dried in a i r without decomposition. In dry a i r the hydrate l o s e s some water and is converted into the dihydrate ZrOClz 2 H z 0 . Quantitative dehydration takes place at 180 to ZOO". The heat of formation of anhydrous ZrOCl2 i s 246 kcal/mole. Z i r c o n i u m b r o m i d e a n d i o d i d e . F r o m the standpoint of their properties Z r Br, and ZrI, resemble zirconium chloride. Like the chloride they a r e readily sublimed. The vapor p r e s s u r e over the solid s a l t s attains one atmosphere a t 3 5 7 " for ZrBr, and a t 4 3 0 " for ZrI,. At elevated p r e s s u r e s these s a l t s melt a t 450 and 499" respectively. Zirconium iodide is used to p r e p a r e high-purity zirconium metal by the thermal dissociation method. Z i r c o n i u m f l u o r i d e s a n d c o m p l e x f l u o r i d e s . Zirconium fluoride ZrF, i s formed by the interaction between gaseous hydrogen fluoride and zirconium dioxide at 500 to 600, or when zirconium dioxide is heated with acid ammonium fluoride.

ZrO,

210, + 4HF ZrF, + 2H20; + 2NH,HF, = ZrF, + 2NH, + 2H,O.


=

Zirconium fluoride is a white substance, which boils at 908". The heat of formation of ZrFl is 456 kcal/mole. The crystalline hydrate Z r F 4 . H 2 0 is precipitated on the addition of hydrofluoric acid to a solution of zirconium chloride or sulfate. The fluoride dissolves in the presence of excess hydrofluoric acid; this is due to the formation of the complex acid:
ZrF,

+ 2HF

H, [ZrF,].

Similarly, the dissolution of zirconium dioxide (or hydroxide) in hydrofluoric acid is accompanied by the reaction

210, 6HF Z H ,[ZrF,]

+ 2H,0.

Complex s a l t s - fluozirconates K2ZrF, and Na2ZrF6- a r e formed in solutions containing potassium and sodium ions.

216

Potassium' fluozirconate is of importance in technology. This s a l t readily crystallizes out of solutions a s colorless rhombic prisms. The solubility of this s a l t is strongly affected by the temperature ( s e e Table40), so that i t can be purified by recrystallization. The fluozirconate is stable in the a i r and i s not hygroscopic. The s a l t is used a s starting m a t e r i a l f o r the production of metallic zirconium. Nitrates. Zirconium hydroxide dissolves in n i t r i c acid. Depending on the concentration of the nitric acid, the s a l t crystallizing out of solution i s either the normal nitrate Z r ( N Q ) 4 5H20 o r zirconyl nitrate ZrO(N03),. 2Hz0 and dizirconyl nitrate Zr,Q(NQ), 2H20. In manufacturing p r o c e s s e s zirconium is occasionally isolated a s zirconyl nitrate. Phosphates. On the addition of orthophosphoric acid or sodium phos phate t o an acid solution of a zirconium compound, a white precipitate of the phosphate ZrHz(P04), s e p a r a t e s out; it is converted upon ignition into the pyrophosphate ZrP20,. Zirconium phosphate is insoluble in water and in sulfuric and hydrochloric acids of concentrations up to 20%. This makes it possible to s e p a r a t e zirconium f r o m almost all other elements. Zirconium carbide and nitride. The carbide Z r C and nitride ZrN a r e very hard, refractory substances wit'n metallic properties (metallic l u s t e r , electrical conductivity). The melting point of the carbide i s 3530", that of the nitride i s 2930". The two compounds have the s a m e c r y s t a l lattice (f. c. c. ) and form a continuous s e r i e s of solid solutions. The heat of formation of ZrC i s 48.2 kcal/mole, that of ZrN is 82.2 kcal/ mole. The carbide is produced by heating a mixture of zirconium dioxide and carbon at 1900 to 2000", through the reaction:

ZrOz

+ 3C = ZrC + 2CO.

If the process is c a r r i e d out in the presence of nitrogen, the carbide produced always contains some nitrogen. Zirconium nitride is formed by the reaction of nitrogen with zirconium powder a t 800 to 1000" o r between gaseous ammonia and zirconium tetrachloride.

Uses of zirconium / I S / Zirconium, i t s alloys and chemical compounds a r e used in various branches of industry. Its main u s e s at present a r e : 1) nuclear energetics; 2) electronics; 3) pyrotechnics and manufacture of ammunition; 4) machine building; 5) production of s t e e l s and nonferrous metal alloys; 6 ) production of r e f r a c t o r i e s , ceramics, enamels, and glasses. Zirconium metal o r zirconium-based alloys a r e used in the f i r s t four industries listed above. Nuclear energetics. In view of the development of nuclear energetics, zirconium attracted attention in 1950 a s a possible s t r u c t u r a l m a t e r i a l for nuclear power r e a c t o r s , with the result that ductile zirconium and zirconium-based alloys began to be manufactured. The value of

217

zirconium a s a s t r u c t u r a l m a t e r i a l in nuclear technology is explained by its low t h e r m a l neutron capture c r o s s section (about 0.2 barn), i t s high r e s i s t a n c e to corrosion and favorable mechanical properties. The low t h e r m a l neutron capture c r o s s section of zirconium was not at f i r s t evident, since zirconium usually contains 0.5 t o 370 of hafnium which h a s a much higher capture c r o s s section (about 115 barns). Accordingly, before zirconium could be used in nuclear technology it was n e c e s s a r y to solve the difficult problem of the separation of zirconium f r o m i t s chemical analog - hafnium, Zirconium and i t s alloys a r e used in nuclear power r e a c t o r s which operate a t temperatures which a r e too high for the use of aluminum to be possible. Zirconium is used to make protective shells for uranium fuel elements, coolant circulation tubes, and other s t r u c t u r a l p a r t s . The heat resistance of zirconium and i t s resistance to the effect of water and s t e a m may be improved by the addition of tin (1.4 to 1.670) a s well a s of s m a l l amounts of iron (0.1 to 0.1570), chromium (0.08 to 0.1270)~and nickel (0.04 to 0.0670). The alloy containing these additives is known a s Zircalloy-2 /1, 2, 1 6 / . Like molybdenum, zirconium is used f o r alloying uranium used a s nuclear fuel; the addition of zirconium improves the mechanical strength and corrosion resistance of the uranium. Electronics. The use of zirconium in vacuum tubes i s due to i t s ability to absorb gases, and s o to maintain a high vacuum in electronic instruments. F o r this purpose zirconium powder of ductile zirconium is applied to the surfaces of anodes, grids, and other heated p a r t s of vacuum tubes. The application of zirconium to the surface of g r i d s in radio tubes s u p p r e s s e s grid emission. Zirconium foil is used in X-ray tubes with molybdenum anticathodes. The foil s e r v e s a s f i l t e r which r e n d e r s the radiation m o r e monochromatic. Pyrotechnics and manufacture of munitions. Zirconium powder, which h a s a low ignition temperature and a high burning rate, is used in this branch of technology. Zirconium powder is used a s the ignitor in mixtures f o r detonator caps, and in mixtures for photographic flash bulbs. Mixtures of zirconium powder with oxidants (barium nitrate or potassium chlorate) a r e used a s smokeless powder. Machine building. Until recently ductile zirconium and i t s alloys were used mainly in nuclear technology. However, with the further i n c r e a s e in i t s production volume and decrease in p r i c e zirconium may be effectively used as m a t e r i a l for chemical plant equipment; it is employed a s acidr e s i s t a n t m a t e r i a l f o r the production of centrifuge p a r t s , pumps, condensers, and evaporators; general purpose machine p a r t s (pistons, rods, shafts, etc.) - in turbine building (turbine blades and other p a r t s ) and in the production of medical instruments. Production of steels and nonferrous m e t a l alloys. Zirconium is extensively used as an additive to steel, serving to deoxidize the steel, to remove nitrogen, and to bind sulfur. In addition, zirconium is a valuable alloying element which is added to certain brands of a r m o r steel, s t e e l s f o r forged gun p a r t s , stainless steels, and high-temperature steels. Zirconium is added to s t e e l s in the form of ferrosilicozirconium (40 to 4570 Zr, 20 t o 2470 Si; remainder iron).

218

Zirconium is a component of a s e r i e s of nonferrous alloys (copper, magnesium, lead, nickel base alloys). Copper alloys containing 0.1 to 570 Z r can be hardened by thermal treatment (quenching and tempering). The tensile strength reaches 50 kg/mm2, which is 5070 higher than that of nonannealed copper. When a r t i c l e s made of pure copper ( w i r e s , sheets, tubes) are heated up to 200" their strength d e c r e a s e s considerably because of the r e l e a s e of cold-hardening. The addition of zirconium i n c r e a s e s the annealing temperature of copper to 500". The addition of s m a l l amounts of zirconium to copper i n c r e a s e s i t s strength with only an insignificant d e c r e a s e in the electrical conductivity of the copper. Zirconium is added to copper in the f o r m of a hardening copper-zirconium alloy containing 1 2 to 14% Zr. Copper-zirconium alloys a r e used f o r the production of electrodes for spot welding and f o r electrical bus b a r s when a high strength is required. Mangesium alloys containing zirconium a r e now increasingly used. The addition of s m a l l amounts of zirconium favors the formation of fine-grained magnesium castings and thus i n c r e a s e s the strength of the metal. Magnesium alloys with zinc and zirconium a r e very strong. Magnesium alloy with 4 to 570Zn and 0.6 to 0.770 Z r does not display c r e e p below 200" and i t s use a s a s t r u c t u r a l m a t e r i a l for jet engines has been recommended. Zirconium is added (in the form of a silicon-zirconium alloy) to lead bronzes. It promotes the dispersion of lead and completely prevents the segregation of lead in the alloy. Copper-cadmium alloys containing up to 0.3570 Z r a r e characterized by their high strength and electrical conductance. Zirconium is a component of a number of corrosion-resistant alloys. Thus, an alloy containing 5470Nb, 4070Ta and 6 to 770 Z r has been proposed a s a substitute for platinum. Manufacture of refractories, porcelain, enamels, glazes and glass. These branches of industry at present account for m o r e than 5070Of the total consumption of zirconium; zirconium is used in the form of i t s minerals (zircon and baddeleyite) and chemical compounds (zirconium dioxide, zirconates, zirconium diboride). Zirconium dioxide (mp 2700 to 2900") and the mineral zircon (ZrSi04) a r e used a s r e f r a c t o r i e s . P u r e zirconium dioxide is unsatisfactory a s a refractory m a t e r i a l because of i t s t h e r m a l instability which causes a r t i c l e s heated to high temperatures to crack upon cooling. This is associated with phase t r a n s formations in the dioxide; such transformations a r e accompanied by volume changes which cause the cracking. The cracking is prevented by the addition of s t a b i l i z e r s - magnesium or calcium oxides. The stabilizers dissolve in the zirconium dioixde forming a solid solution with a cubic c r y s t a l lattice, which is stable a t both high and low temperatures. Zirconium dioxide o r the minerals zircon and baddeleyite are used in the production of r e f r a c t o r y bricks for metallurgical furnaces, crucibles f o r melting m e t a l s and alloys, refractory tubes, and other articles. Zirconium m i n e r a l s o r zirconium dioxide a r e added to certain brands of porcelain used f o r the production of insulators for high-voltage lines, f o r high-frequency circuits, and f o r the spark plugs in internal-combustion engines. Zirconium-containing porcelain has a high dielectric constant and a low expansion coefficient.

219

111111I11111111111111111111111111 111 111111111111II1I1III11111II.IIIIllllllllllllllllllll1

I I II 111111.IIIIIIIIIIIIIII111111111111111111

l11111111 I

Zirconium dioxide and zircon (purified from iron) find extensive u s e They impart to the enamel a white color and resistance to acids and completely replace tin oxide which is in s h o r t supply. Zircon and zirconium dioxide a r e also components of certain brands of glass. The addition of Z r G improves the resistance of the glass to alkali solutions. A l a r g e proportion of high-quality zirconium concentrates is used in foundry practice f o r the production of casting molds and rods; powdered zircon is applied to the surface of molds to obtain c a s t s with satisfactory surface properties. Other U s e s . Mention should be made of the use of zirconium sulfates (double zirconium ammonium sulfate, etc. ) a s tanning agents in the leather industry. The total production of zirconium concentrates outside the USSR is about 170,000 tonslyear.

as enamel components.

Uses of hafnium 1161 The commercial production of hafnium and i t s compounds i s only 6 to 8 y e a r s old. Interest in hafnium a r o s e mainly in connection with nuclear technology in which it is a component of the control rods and reactor shields (hafnium has a high thermal neutron capture c r o s s section 115 barns). It is used a s metal, as hafnium dioxide and a s hafnium diboride (HfB2). Another promising use is in the production of refractory materials. Hafnium carbide (mp 4000") or a solid solution of hafnium carbide in tantalum carbide (75% TaC) with a mp of 4200" may be used. Other refractory compounds of hafnium a r e also being studied.

54.

MINERALS, ORES, AND ORE CONCENTRATES

The abundance o f zirconium in the E a r t h ' s crust is 0.025% by weight. In the o r d e r of abundance it precedes some common metals such a s copper, zinc, tin, nickel, and lead. There a r e about 20 known zirconium minerals. They a r e concentrated mainly in granite and basic (nepheline-syenite) pegmatites. Zirconium minerals may be divided into three groups: 1) zirconium dioxide (the mineral baddeleyite and its varieties) ; 2) zirconium orthosilicates (zircon and its varieties); 3) the zirconosilicates of sodium, calcium, iron and other elements (eudialyte, eucolite, catapleite, etc. ). The minerals baddeleyite and zircon a r e at present the main industrial s o u r c e s of zirconium. In addition, the minerals eudialyte and eucolite may s e r v e a s a source of zirconium but their zirconium content is much lower. B a d d e 1e y i t e consists of almost pure zirconium oxide. The common impurity is hafnium (up to a few percent) and m o r e r a r e l y uranium (up to 1%) and thorium (up to 0.2%) a s well as mechanical contaminants minerals containing SiOz, Ti&, Fez%, etc. The purest samples contain

220

,,

..

2fw 21UU 2wo

1 2300 1 j , ZrU,

2400

(ss)

\ \ \

-1

I
\.+

1900

-I

/sou
1700
1600

I1 -

'

IPSl4 + 1 17?5*lU 1:1 1 3

I ;

ZrQi+(ss)

t sto4

1675*5*"A zrsi04

of a baddeleyite- zircon mixture containing approximately 7 5 to 82% ZrOz, 2 to 5% Fe&, 6 to 1770 si%, 1.3 to 1.770 A l z q and 0.2 to 0.8% Ti&. The major impurities in the concentrates a r e silica, iron, aluminum, and titanium. High-quality concentrates a r e produced by gravitational beneficiation on concentration tables (resuiting in the removal of the lighter Al- containing minerals) and electromagnetic

22 1

Zircon-containing o r e s a r e enriched by gravitational methods in conjunction with magnetic and electrostatic separation. The separation of zircon f r o m quartz is c a r r i e d out by concentration on tables. nmenite and garnet (a silicate of aluminum, iron, calcium, and other elements) are separated by magnetic methods with the use of weak fields (zircon is non magnetic). Monazite is separated by magnetic separation in strong fields. The rutile and the zircon are separated by electrostatic methods or by flotation. A m o r e complete removal of the iron (which is of importance when the zircon is to be used as a component of enamels) is attained by treating the concentrate with hot sulfuric acid. The approximate composition of zircon concentrates is shown in Table 38.
TABLE 38 Approximate composition of zircon concentrates,% TiO,

66.42 61.31 66.3

31.20 33.0 34.8 31.0

5.5 Traces 1.7 2 . 3

The l a r g e s t zircon deposits outside the USSR a r e in Australia, India, and Brazil. In the USSR, zircon is found in the U r a l s , in the Azov alkaline massif, and in other regions. E u d i a 1y t e a n d e u c o 1i t e . The composition of eudialyte corresponds ClIz. Eucolite is a variety to the empirical formula (Na, Ca)6Zr[Si60~8][OH, of eudialyte, containing Fez+ions. A mineral of intermediate composition, mesodialyte, is also known. Eudialyte contains 11.6 to 17.3% NazO, 1 2 to 14.570 Z r Q , 3.1 to 7.1%FeO, 47.2 to 51.2% S i Q and 0.7 to 1.6% Ce. The mineral contains small amounts of r a r e - e a r t h metals (Ce, La, Y) and niobium. Eudialyte is colored various shades of pink or crimson. The mineral is readily decomposed by acids. Eudialyte and eucolite a r e usually found in alkaline magnatic rocks (nepheline syenites). Deposits a r e found in the USSR, Norway, Greenland, Transvaal, Brazil, and other countries.

55. PRODUCTS FORMED IN THE PROCESSING OF ZIRCON CONCENTRATES Zircon concentrates, which a r e the main zircon raw material, s e r v e a s the starting m a t e r i a l f o r the production of ferrosilicozirconium, f e r r o zirconium and zirconium compounds: zirconium dioxide, potassium fluozirconate and zirconium tetrachloride.

222

Ferrosilicozirconium is produced directly by smelting zircon concentrates. Technical zirconium dioxide s e r v e s as raw m a t e r i a l f o r the smelting of ferrozirconium and is used for the production of re fractories. The high-purity dioxide is used f o r the production of highquality refractory a r t i c l e s (crucibles, tubes, etc. ). Potassium fluozirconate and zirconium tetrachloride a r e used mainly for the production of metallic zirconium. The main methods f o r the production of zirconium dioxide, potassium fluozirconate and zirconium chloride f r o m zircon concentrates a r e reviewed below.

56. METHODS FOR THE DECOMPOSITION O F ZIRCON CONCENTRATES

Zircon is a v e r y stable mineral. It is decomposed by various pyro metallurgical methods by way of decomposition of zirconium silicate. The following methods a r e used for the decomposition of zircon: 1. fusion of zircon with sodium hydroxide, with the formation of sodium z i r conate ; 2. sintering with lime o r chalk, with the formation of calcium zirconate; 3 . fusion with potassium fluosilicate KzSiF6, with the formation of potassium fluozirconate; 4. chlorination of a mixture of zircon and carbon to produce zirconium tetrachloride; 5. reductive smelting with coal to zirconium carbide ZrC, which is chlorinated to produce zirconium chloride. The f i r s t two methods a r e commonly used for the production of technical and pure grade zirconium dioxide. If necessary, a processing stage involving the separation of hafnium from zirconium is included. The method of fusion with KzSiF6, developed in the USSR, can be readily combined with the separation of zirconium and hafnium by fractional crystallization of the complex fluorides ( s e e Section 6 3 ) . The fourth and fifth methods a r e used in the production of zirconium tetrachloride .
57. DECOMPOSITION OF ZIRCON BY FUSION WITH

SODIUM HYDROXIDE A flow sheet of the decomposition of zircon by fusion with sodium hydroxide is shown in Figure 111. Fusion. The fusion of z i r c w with sodium hydroxide yields sodium zirconates and silicates:
ZrSiO,

+ 4NaOH = Na,ZrOs + NaSiO, + 2H,O.

Other products which a r e most probably formed a r e the orthozirconate Na4Zr04, complex zirconosilicates (such as NazZrSiOJ, and other sodium silicates, e. g . , NazSi205, Na6Siz0,.

223

Titanium present a s an impurity (in the form of rutile and ilmenite) r e a c t s with sodium hydroxide to yield sodium titanates. Iron and aluminum a r e present in the melt as f e r r i t e (NazO* Fez%) and aluminate ( N a 2 0 .A l & ) . The fusion with sodium hydroxide is c a r r i e d out in iron crucibles o r c a s t i r o n v e s s e l s a t 600 to 650". The sodium hydroxide is added in a 50% e x c e s s (over the stoichiometric amount needed) in o r d e r to ensure quantitative decomposition of the zircon, The melt is granulated by pouring into water a thin jet.

I
Zircon concentrate Crushing Fusion Water *)Leaching NaOH

7
7
4
Residue NazZrOs, ZrO(OH), , FeZOs,Na, TiOs , HSiOs

Solution Na,SiO, , Na AlO,, NaOH and other

HC1

Solution of ZrOC1,

First variant

7 I
-

Second variant Leachin;

Leaching

Residue H2SiOs

Solurion of ZrOSO, HdZrO( SOJz1

1
H2S04

Residue H,SiO,

Crystallization of ZrOCl, ' 8 H 2 0

Hydrolytic separation of the basic sulfate

Isolation in the form of H 2 [ Z r q S 0 ~ 2 13Hz0 '

J.

FIGUW 111. Flow sheet of the processing of zircon by fusion with sodium hydroxide.

During the fusion or leaching m e a s u r e s must be taken to prevent spattering of the alkaline melt or of solutions. An exhaust hood must be fitted over the v e s s e l (or crucible) in which the fusion is c a r r i e d out. An "alkaline m i s t " (consisting of a stable sodium hydroxide aerosol) which affects the integument and the r e s p i r a t o r y system is formed in poorly ventilated rooms. This i s difficult if fusion p r o c e s s e s have to be c a r r i e d out on a l a r g e scale. The rapid corrosion of s t e e l o r c a s t iron v e s s e l s by the alkaline melt i s another disadvantage of the process. The fusion with sodium hydroxide has the advantages of requiring a relatively low temperature and of having a high output rate.

224

Leaching the melt with water. The leaching of the alkaline melt with water is c a r r i e d out in iron tanks fitted with s t i r r e r s or in drum-type leaching equipment of the ball type, in two stages with intermediate decantation. During the leaching with water most of the silicon p a s s e s into solution in the form of sodium silicate Na2Si03. At the s a m e time there i s a partial hydrolysis of sodium zirconate with the formation of zirconium hydroxide : NazZr03 3HzO = Zr (OH), 2NaOH.

In the second stage leaching of the residue with water, partial hydrolysis of sodium silicate also takes place a s the alkalinity of the solution decreases: Na,SiO,+ 2H,O = H,Si03 2NaOH.

Sodium titanate remains in the residue together with the zirconium while sodium aluminate p a s s e s into solution. The f e r r i t e is decomposed a s 'follows : Na,O . FezO, HBO= Fe,O, 2NaOH.

Thus, the residue from the leaching of the melt with water contains sodium zirconate, zirconium hydroxide, residual sodium silicate, silicic acid, sodium titanate, iron oxides, and a certain amount of sodium zirconosilicate. The approximate composition of the residue i s : 80 to 84% ZrOz, 8 to 1 2 % SO2, 4 to 670Na20. The residue is forwarded to acid leaching. Acid leaching. Either hydrochloric or sulfuric acid is used f o r the leaching. Zirconyl chloride solutions a r e formed in the f o r m e r c a s e , while solutions containing zirconyl sulfate and zirconylsulfuric acid a r e formed in the l a t t e r case:
Na,ZrO, + 4HCl = ZrOC1, 2NaCl+ 2H20, Na,ZrO, 2HzS04 = ZrOSO, Na,S04 2H,O, ZrOSOa HzSO, Hz [ Z f l (m4)zl.

+ +

Compounds of iron and other elements p a s s into solution. A p a r t of the silicic acid remains in the insoluble residue together with the undecomposed zircon, while another fraction is solubilized a s a colloid. Coagulants (e. g., carpenter' s glue) a r e added to the solution to coagulate the silicic acid. The voluminous silicic acid precipitate entrains a p a r t of the zirconium. The dissolution in hydrochloric acid is c a r r i e d out in heated earthen ware or rubber-lined iron tanks fitted with s t i r r e r s . Tanks made of a synthetic r e s i n - faolite - may also be used. The dissolution in sulfuric acid i s c a r r i e d out in lead-lined tanks o r in tanks lined with acid-resistant ceramic tiles. The selection of the acid to be used f o r leaching depends on the required purity of the final product - zirconium dioxide ( s e e below). The solution is separated from the residue (consisting mainly of silicic acid and the undecomposed mineral) and zirconium compounds a r e precipitated f r o m the solution ( s e e Section 59).

225

58.

DECOMPOSITION OF ZIRCON B Y FUSION WITH LIME

Heating zircon with l i m e leads to the formation of calcium zirconate and silicates in accordance with the equations:

ZrSiO, ZrSiO,

+ 2CaO = CaZrO, + CaSiO,; + 3Ca0 = CaZrO, + Ca,SiO,.

Dicalcium silicate Ca2Si04is preferentially formed under the conditions used in the process, Unlike zircon, calcium zirconate i s readily decomposed by acids and zirconium may thus be solubilized. The flow sheet of the p r o c e s s i s shown in Figure 112.

I Zircon concentrate I

Sintering LeachinE with acids

.1
First variant

1
Second variant

6 1 0 ' 7 0

HC1

I
First leaching in the cold

Leaching

1
Solution CaC1, ,colloidal H z S i Q . Fe. etc.

J
Residue
P

1 l -

c/

H2S04

Recidue CaSO,. HzSi03

HC1 25-3070

Second leaching with heating

7
Residue HZSiO,

Solution ZrOCIZ

1
Isolation of ZrOClZ. BHZO

Solution HdZrO( SOd21 ZrOS04

1
frnmralution Isolation of H2[Zr4S04)z 3 HzO Isolation of the basic sulfate by hydrolysis

F Z i r { ? G j

i- -

FIGURE 112.

Flow sheet of the processing of zircon by sintering with lime.

Sintering with lime. Zircon and lime interact at a noticeable r a t e only at 1400 to 1500". However, the addition of alkali-metal and alkalinee a r t h chlorides (in particular calcium chloride) reduces the p r o c e s s temperature to 1000 to 1100". The accelerated r a t e of the reaction in the presence of chlorides is attributed to the formation of an intermediate liquid phase (CaC12 melts at 774") which dissolves some of the calcium
226

oxide. In addition, i t is possible that the reaction is catalytically accelerated because of the chlorinating action of calcium chloride:
2CaC1, + ZrSiO, = ZrC1, ZrC1, + 3Ca0 = CaZrO,

+ Ca,SiO,; + 2CaC1,.

During the sintering the i r o n oxides r e a c t with l i m e to form calcium f e r r i t e C a ( F e q ) 2 ; ilmenite r e a c t s to form calcium tftanate. Chalk may be used instead of lime when the reaction r a t e is somewhat lower. The amount of the l i m e or chalk added is 100 to 150% of the stoichiometric (for reaction (.2)), while the amount of CaClz is about 1 / 5 (by weight) of the amount of l i m e added to the charge. The sintering is c a r r i e d out on the s o l e of a muffle furnace or in a r o t a r y furnace a t 1000 to 1100". The sintering time is 8 to 1 0 hours. Under these conditions the degree of decomposition reaches 90 to 9470. The decomposition of zircon concentrates by sintering with lime and calcium chloride has the following advantages over the fusion with sodium hydroxide: the reagents used a r e much cheaper and l a r g e - s c a l e industrial processing i s simpler since rotary furnaces may be used. Hydrochloric o r sulfuric acids a r e used to leach the sintered m a s s . Leaching with hydrochloric acid. The sintered m a s s i s f i r s t treated with 5 to 10% HCl in the cold. This r e s u l t s in the dissolution of the excess calcium oxide and calcium chloride but calcium zirconate i s not affected. Moreover, processing with cold HC1 results in the decomposition of most of the calcium orthosilicate. The colloidal silicic acid formed is separated together with the solution. The second leaching is c a r r i e d out at 85 to 90" with 2 5 to 3070 hydrochloric acid. A s a result zirconium is solubilized:

CaZrO, + 4HC1 = ZrOC1,

+ CaCI, + 2H20.

Silicic acid is formed a t the s a m e time a s a result of the decomposition of residual calcium silicate. A solution of c a r p e n t e r ' s glue i s used to coagulate the silicic acid. The solution i s separated from the residue and zirconium is precipitated by one of the methods described below. Leaching with sulfuric acid. When the sintered m a s s is leached with sulfuric acid, calcium remains in the residue a s a component of the sulfate. The sintered m a s s i s gradually introduced into concentrated sulfuric acid; the reaction is accompanied by strong heat-up of the mixture, which aids in the separation of the silicic acid (silicic acid i s dehydrated upon heating). At the end of the reaction the solution i s diluted, filtered, and zirconium isolated.
59. ISOLATION OF ZIRCONIUM FROM HYDROCHLORIC AND SULFURIC ACID SOLUTIONS

At present t h e r e a r e t h r e e industrial methods for the isolation of zirconium from hydrochloric and sulfuric acid solutions. 1. Precipitation of zirconium oxychloride ZrOClz * 8Hz0 f r o m hydro chloric acid solutions.

221

2. Hydrolytic precipitation of basic zirconium sulfates; this may be c a r r i e d out from both hydrochloric and sulfuric acid solutions. 3. Isolation of the crystalline zirconylsulfuric acid hydrate Hz[ZrO(SO,),] 3 H z 0 f r o m sulfuric acid or zirconium oxychloride solutions.

Isolation of zirconium oxychloride The method is based on the fact that ZrOClz * 8 H z 0 is sparingly soluble in concentrated hydrochloric acid while being readily soluble in water and dilute HC1 ( F i g u r e 113). The solubility of the oxychloride has a minimum (10.8 g/ 1) a t HC1 con centration of 318 g / l ; in dilute HC1 solutions the solubility is 40 to 50 times higher 71. In addition, the solubility of ZrOClz strongly depends on the temperature ( i t s solubility in concentrated HC1 at 70" is about five times that a t 20"). Hydrochloric acid is concentrated by evaporation of the solution. During the evaporation the HC1 concentration must remain below 20.270 (about 220g/l) or an azeotropic mixture w i l l be formed*. Even in acid of this concentration the solubility of ZrOClz 8H20 is not high ( - 25 g / l ) and, after the solution is cooled,
about 70 to 9 0 % of the zirconium
initially present in the solution may be
isolated a s crystals, MCl concentration, g/l The method yields zirconium compounds of a very high degree of purity since most FIGURE 113. Solubility of ZrOC1,of the impurities (Fe, Al,a p a r t of titanium, 8 H Z 0 in hydrochloric acid a t 20". and some others) remain in the hydrochloric mother liquor. Also the s a l t may be purified by recrystallization. Zirconium oxychloride can be readily converted to other pure zirconium compounds : zirconium hydroxide, dioxide, fluorides, nitrate o r sulfate. To do this zirconium oxychloride is dissolved in water, zirconium hydroxide precipitated by the addition of ammonia and the hydroxide is then dissolved in the respective acid. Zirconium dioxide prepared by this method contains 99.6 to 99.8% Z r Q .

Hydrolytic precipitation of basic sulfates This method is extensively used in industry, a s i t can b e applied to both sulfuric and hydrochloric acid solutions.

--------------* Azeotropic solutions are solutions which

distill without a change in composition (i. e., the liquid and vapor phases have the same composition).

228

I
Isolation from sulfuric acid solutions. In sulfuric acid solutions zirconium is present in the form of zirconyl ions Z r d ' and of complex anions [ZrO(S04)2]2-;the ratio of these species is determined by the equilibrium

(Zro)*+ + 2 (so,)2- 2 [Zro

The presence of these two ionic species i n solution h a s been confirmed by recent studies c a r r i e d out with the aid of ion-exchange resins. At sulfuric acid concentrations up to 0.5 to 1.0 N ( 2 4 to 50 g / l ) zirconium is sorbed on both cation and anion exchange resins. At h.igher sulfuric acid concentrations adsorption on cation exchange r e s i n s c e a s e s , which shows that all the zirconium is in the form of a complex anion. Accordingly, when carrying out the hydrolysis, it is not sufficient to adjust the pH of the solution to that corresponding to the hydrolysis of zirconium sulfate; the worker must also ensure that t h e r e i s no l a r g e excess of S@- ions since the extent of formation of complex ions depends on the sulfate ion concentration in the solution r a t h e r than on its acidity. Hydrolytic p r e c i pitation of zirconium in fact does not take place when a sulfuric acid solution containing a l a r g e excess of the acid is neutralized with soda ash o r ammonia. The hydrolysis takes place only if a p a r t of the So"ions is removed from the solution, e. g . , by the addition of BaClz o r CaClz When the sulfate-ion concentration in the solution is reduced to a certain minimum (about 0.55 to 0.6 molefmole Z r Q ) , it is possible to achieve nearly quantitative hydrolytic precipitation of basic zirconium sulfate from sulfuric acid solutions a s a result of the hydrolysis of zirconyl sulfate. The composition of the precipitated basic sulfates may be described by the The precipitated general formula xZrOz. ySOs. zHpO (the ratio x : y > 1). basic sulfates have a variable composition, depending on the initial acidity and zirconium content of the solution. The molar ratio Z r Q : SQ in the precipitates ranges from 1 : 0.75 to 1 : 0.2. The hydrolysis is c a r r i e d out in dilute solutions with a zirconium content of 40 to 60g/1, which a r e neutralized with soda ash to pH = 2 -3. In o r d e r to prevent contamination of the precipitates with iron, f e r r i c ions a r e reduced to f e r r o u s ions with iron turnings ( s e e Section 45). The solution is heated to 70 to 80" when up to 98% of the zirconium is precipitated a s the basic sulfate. The voluminous precipitate of the basic sulfate s o r b s iron, aluminum, and other s a l t s present a s impurities. The impurities a r e only partly removed when the precipitates a r e washed. Isolation of the basic sulfate f r o m hydrochloric acid solutions. It is much s i m p l e r to isolate the basic sulfate from hydrochloric acid than from sulfuric acid solutions since the difficulties associated with the presence of an excess of sulfate ions a r e absent. The basic sulfate is precipitated from hydrochloric acid solutions containing 40 to 6 0 g/1 zirconium by the addition of sulfuric acid (about 0.55 molefmole Z r Q ) followed by neutralization and dilution to a resultant acidity of about 1 t o 1.5gram HC1 p e r liter. The solution is heated to 70 to 80" and the basic sulfate ( Z r Q : SQ r a t i o of about 1 : 0.3) precipitates out. Soluble sulfates (Na2S04, MgS04) may be added instead of sulfuric acid. The precipitate contains 97 to 98% of the total zirconium. The basic sulfate precipitates a r e washed, separated by filtration, dried and ignited (to remove SQ) a t 850 to 900" in muffle furnaces lined with r e f r a c t o r i e s of a high alumina content.
229
/9/*

The method yields technical grade zirconium dioxide containing 97 t o 98% Z r q . The main impurities are: 0.5 to 1.5% Ti(&, 0.2 to 0.8% Si(&, 0.1 to 0.15% Fez@, 0.2 t o 0.5% CaO, 0.2 to 0.4% PzO, , and up to 0.3% S@. Additional purification of the basic sulfate is required to obtain a product of a Bigher degree of purity.

Isolation of zirconium as the zirconylsulfuric acid hydrate The addition of concentrated sulfuric acid to concentrated aqueous solutions of zirconium sulfate or chloride r e s u l t s in the precipitation of crystalline zirconylsulfuric acid hydrate Hz[ZrO(SO& J . 3Hz0. The dependence of the solubility of the crystalline hydrate on the sulfuric acid concentration is given in Table 39; it will be s e e n that the solubility in the H2S0, concentration range of 46.7 to 57.470 is v e r y low.

TABLE 39

____
Y

H2SO4

2 1 0 , g/ 100

g
concentration, %

concentration, '70
31.2 33.1 35.6 39.6 42.5 44.1 46.1

solution

Zr02 g/lOO g solution

16.20 9.60 5.30 3.51 1.03

70.5 72.9
. . . ~.

0.14 0.15 0.50 2.0

The highest degree of separation of the zirconium (94 to 95%) is obtained under the following conditions. One volume of concentrated sulfuric acid is added to two volumes of the sulfate solution containing 120 to 130 g / i zirconium (or to the oxychloride solution containing 200 to 220g Z r / l ) . This r e s u l t s in the separation of a white crystalline precipitate which s e t t l e s rapidly and is separated by filtration through porous earthenware filters. F u r t h e r purification is accomplished by dissolving the precipitate in water (one kg of precipitate in one liter of water) and reprecipitating the crystalline hydrate by the addition of concentrated sulfuric acid. This yields a product of a v e r y high degree of purity in which the concentrations of iron, copper, and s i l v e r a r e less than lo-' % each, and the concentra tions of calcium, sodium, magnesium, and silicon a r e below each. P u r e zirconium dioxide may be produced by igniting the crystalline hydrate of zirconylsulfuric acid at 850 to 900" 1 8 1 . It is advantageous to use the method based on the isolation of zirconyl sulfuric acid for the purification of the basic sulfate precipitates prepared by hydrolysis.

1455

230

6 0 . PROCESSING OF ZIRCON BY SINTERING WITH

POTASSIUM FLUOSILICATE This method w a s developed in the USSR by Sazhin and Pepelyaeva 131. The method is based on the reaction occurring when zircon is heated with potassium fluosilicate: ZrSiO,

+ KaSiF,,

-+

K,ZrF,

+ 2SiOI.

The potassium fluozirconate obtained is leached with water and the salt is then crystallized out of solution.

Zircon concentrate

.1
Crushing KzSiF6. KCl

Water with 1% HCl

J
,
Leaching

Sintering

Wash watet

SiO,, undecomposed mineral Washing

J. Residue

.L
KzZrF, solution

Crystzllization

I+

KzZrF6 ( HO crystals

Mother liquor

Solid residue

To waste

Removal of Hf by fractional crystallization

Precipitation of zirconium hydroxide

NHPH

7 Mother liqlmr
Precipitation of Zr(OH)r

To isolation of Hf

Sintering

1$
FIGURE 114. Flow sheet of the processing of zircon by fusion with potassium fluosilicate.

23 1

The potassium fluosilicate required for the p r o c e s s is readily prepared f r o m a relatively cheap substance - sodium fluosilicate, which is a byproduct of the manufacture of hydrofluoric acid and cryolite. The inter action of a saturated solution of Na2SiF6with a saturated solution of KC1 c a u s e s precipitation of K2SiF6:
Na,SiF,

+ 2KC1

--f

&SiF,

+ 2NaCI.

This can be explained by the fact that the solubility of K2SiF6 ( 1.13 g / l ) is much lower than the solubility of Na2SiF6 ( 6.7g/l). A flow sheet of the processing of zircon by sintering with potassium fluosilicate is shown in Figure 114. The zircon concentrate is crushed to a particle s i z e of l e s s than 0.074" and mixed with potassium fluosilicate and potassium chloride. The potassium chloride intensifies the degree of decomposition of the mineral. The K2SiF6is taken in 50% e x c e s s over the stoichiometric amount. The degree of decomposition of zircon is strongly affected by the sintering temperature of the charge. When the p r o c e s s is c a r r i e d out a t 650 to 700" the degree of decomposition reaches 97 to 9870 *e. The sintering is c a r r i e d out in r o t a r y furnaces. The sintered m a s s consists of burnished and slightly fused grains with a particle s i z e of 0.5 to 1 0 " . It is crushed to a particle s i z e of l e s s than 0.15 m m and leached with 1% HC1 (at a solid :liquid ratio of 1 : 7) at 85". The leaching time is 1.5 to 2 hours.

TABLE 40 Solubility of K2ZrF6 in wateI

1.

"C
1.22 1.55 1.92 2.31 2.94

Solubility,

g/lOOg. H2O
60 IO 80 90
100

10 20 30 40 50

3.81 5.06 6.90 11.11 23.53

The mixture is allowed to stand (the temperature of the s l u r r y must not exceed 80") and the c l e a r solution is forwarded to the crystallization stage. If the solution is filtered, the filtration must be c a r r i e d out at a high temperature in o r d e r to prevent premature crystallization. The solubility of K2ZrF6i n c r e a s e s sharply with increasing tempera t u r e ( s e e Table 40). As a result, 75 to 90% of the zirconium present in the solution may be crystallized by cooling the solution (depending on the initial zirconium concentration). Ammonium hydroxide is added to the mother liquor, and the precipitated zirconium hydroxide is returned to sintering. A s impurities accumulate in the mother liquor, they a r e periodically discarded. The K2ZrF6

--------------* Because of the difficulties involved in

maintaining a constant temperature in rotary furnaces, there a r e wide fluctuations i n the degree of decomposition of the concentrate and it is usually lower than this value.

232

c r y s t a l s prepared by this method f r o m zircon concentrates from the Azov region of the USSR have the following composition: 31.9 to 3270 Z r +Hf, 27.2 to 27.670 K, 39.9 t o 40.0570 F, 0.044 to 0.04570 F e , 0.041 to 0.04270 Ti, 0.06 to 0.0770 Si, 0.006 to 0.008% C1, 1.5 t o 2.570 Hf (on the zirconium). Hafnium is separated f r o m zirconium by fractional crystallization. All leaching and crystallization operations may be c a r r i e d out in s t a i n l e s s s t e e l equipment; the corrosion of such equipment is negligible because the inner walls a r e coated with a film of insoluble fluorides. In o r d e r to convert i t into zirconium dioxide, potassium fluozirconate is dissolved in water (25 to 3 0 g K,ZrF6 in one l i t e r of water a t 50 to 60") and the resulting solution is poured into a solution of ammonium hydroxide; the N H 4 0 H is taken in 15070 e x c e s s over the stoichiometric amount r e quired f o r the reaction: K2ZrF,

+ 4NKOI-I = Zr (OH), + 4NH$ + 2KF.

The fluoride s a l t s a r e removed f r o m the zirconium hydroxide by washing with water containing 0.570 NHS, the hydroxide is separated by filtration, dried and ignited a t 900" to zirconium dioxide. The precipitation with ammonia f r e e s the zirconium f r o m the iron and titanium which a r e bound in ammoniacal fluoride complexes and thus only a s m a l l fraction i s precipitated. Thus, zirconium dioxide containing 0.005% F e and 0.00570 T i was prepared f r o m fluozirconate containing 0.03270 F e and 0.012 Ti. This p r o c e s s is characterized by i t s s m a l l number of operations and i t s simplicity. A s w i l l be shown below, hafnium can be readily separated from zirconium by this p r o c e s s , and zirconium dioxide of a high degree of purity can be produced.

6 1 . D E C O M P O S I T I O N O F Z I R C O N BY R E D U C T I O N WITH CARBON TO C A R B I D E O R C A R B O N I T R I D E / 1 , 2 /
Zirconium carbide or carbonitride are produced by reducing the zircon with carbon; during the reduction most of the silicon is removed as S i 0 whose vapor pressure at 2000 to 2200" is close to one atmosphere. The main reactions occurring during the reduction of zircon with carbon are: ZrSiOl 6C = ZrC S i c ZrSiO, 5C = ZrC + S i ZrSi0, 4-4C = ZrC S i 0 ZrSiOl 3C = Zr Si0

+ +

+ 4CO;

+ 4CO; + 3CO; + 3CO.

In the presence of a suitable amount of carbon in the charge, the reduction occurs mainly in accordance with reaction (3). Silicon monoxide is also formed by a side reaction involving the reduction of zircon by silicon: ZrSiO, Si = ZrO, 2.50.

Moreover. the reduction involves a number of other side reactions, one of which is the formation of the easily melted zirconium silicide ZrSi,. When using a charge containing 18 to 20% carbon (in the form of powdered coke), 95 10 96%of the silicon is removed as S i 0 by carrying out the reduction in an electric-arc furnace. A diagram of one type of one-phase a r e furnace used for the reductive melting of zircon a t one plant in the USA is shown in Figure 115. T h e furnace bath consists of a steel cylinder 2.14111 in diameter. The pressed carbon a t the bottom setves as one of the electrodes. The second carbon electrode. which is about 6 0 0 m m in diameter. is placed in the center of t h e furnace. A part of the charge adjacent t o the walls does not take part in the reaction and serves as a protective thermal insulation layer. Low-voltage

233

In such a c a s e the losses of heat through marginal radiation on the walls ar e minimal. T h e radiation losses through t h e cover of the furnace a r e also very small since t h e electrode serves as a shield. T h e total furnace power is 750kw.
(- 50V) current is supplied to the electrodes and this necessitates the use of a short arc.

FIGURE 115. Diagram of an arc furnace for the carbidization of zircon.


cover; 3-electrode; 4-fused part layer serving as thermal in sulator; 6-pressed carbon (lower electrode); I-current lead to t h e lower electrode; 8-exhaust gases; 9-exhaust hood; 10 charge. 1-steel jacket; 2-asbestos

of the charge (carbide); 5-charge

The furnace is allowed to cool, the mass is discharged by rotating the furnace, and the carbide particles are separated from the fraction of the charge which has not reacted. T h e consumption of electrical energy is 8000 to 11,000 kwhlton zirconium carbide. One of the disadvantages of the process is t h e fact that the losses caused by dust formation during the charging and discharging of t h e furnace ar e high. To reduce dust formation the charge is fed into paper bags. However. even in this case the mechanical losses reach 8%. Since the pores of the charge contain nitrogen. t h e product formed.in the fusion is usually the carbonitride ( a solid solution of zirconium nitride in zirconium carbide). The carbonitride formed has the following approximate composition: 75 to 85% Zr, 3 to 5 % C, 2 to 4 % Si, up to 2% N , 1 to 2 % F e . up to 2 % T i , and 1 to 1 0 % 0.

62.

PRODUCTION OF ZIRCONIUM TETRACHLORIDE

Zirconium chloride may be prepared by chlorination of three products : zircon concentrates, zirconium dioxide, and zirconium carbide (or carbo nitride).

Chlorination of zircon concentrates

A mixture of zircon concentrate and carbon can be chlorinated at 900 to 1000" at a r a t e which is sufficiently high to be of practical value. The main

234

reaction involved is:


ZrSiO,

+ 4CI1+ 4C = ZrC1, + SiCI, + 4 0 .

A mixture of the pulverized concentrate, carbon and a binder (either coal t a r o r sulfite-pulping liquors a r e used a s the binder) is pelletized. The pellets a r e coked a t 700 to 800". The chlorination of the pellets is c a r r i e d out in shaft furnaces lined with Dinas [silica r e f r a c t o r y ] bricks. Since the chlorination reaction is endothermic, constant inflow of heat is required to maintain the furnace temperature at 800 to 1000". The p r o c e s s may be c a r r i e d out without external supply of heat if a certain amount of a i r is introduced into the furnace together with the chlorine; the heat evolved in the combustion of a p a r t of the coal in the charge is sufficient to maintain the temperature. The difference between the volatilization temperature of ZrCll (330O) and the boiling point of SiC14 (58")p e r m i t s the condensation of the zirconium chloride in the p r i m a r y condensers which a r e maintained a t 150 to 180, and the condensation of liquid silicon chloride in the secondary condensers in which the temperature is maintained a t -loo by cooling with water o r with a s a l t solution. The silicon chloride finds use in the production of various organosilicon compounds and may be used a s the starting m a t e r i a l for the production of semiconductor grade silicon. Zirconium chloride produced by the above p r o c e s s is contaminated with silica, which is formed a s a r e s u l t of the interaction of a p a r t of the SiC14with moisture and oxygen which may be present in the furnace. F o r the s a m e reason zirconium chloride contains the oxychloride ZrOC12. The d i r e c t chlorination of zircon in mixture with carbon h a s recently found increasing industrial use. The disadvantages of the p r o c e s s (the high temperature) a r e compensated by the elimination of the expenses involved in the pretreatment of the concentrate in o r d e r to separate the bulk of the silicon (e. g . , by carbidization) and by the production of a valuable by-product - silicon tetrachloride.

Chlorination of zirconium carbonitride 11, 4 f


T h e carbide and carbonitride a r e chlorinated a t a low temperature (350 to 450") and the heat evolved i n the reaction is sufficient to maintain the process even in small chlorinators. T h e chlorination reactions are: ZrC 2C1, + ZrCl4 C 202 kcal ZrN

+ + + +X I ,+ ZrCl4 + N, + 160 kcal

A diagram of a furnace used for the chlorination of zirconium carbonitride in a plant in the USA is shown in Figure 116. T h e chlorination is carried out in a shaft furnace about o n e meter in diameter and about 2.5 m high. T h e chlorine enters the furnace through a perforated graphite plate over which there is a layer of coke particles. T h e process is started by heating the coke layer with hot air. T h e carbo nitride is then passed over the heated layer and the chlorine is fed to the furnace. Intensive chlorination of the carbonitride starts a t 400" and is maintained by the reaction heat evolved. T h e condenser is a nickel cylinder 1250 mm in diameter and 2 4 5 0 m m high. T h e temperature in the condenser is maintained a t 150'. At the exit from the condenser zirconium chloride is screened on a sieve. T h e screened fraction is degassed in a low vacuum i n order to separate SiC14 and TiC1,. T h e zirconium tetrachloride produced is contaminated with oxygen and with dust particles entrained from the furnace. T h e product is bright-yellow.

235

FIGURE 116. Diagram of the apparatus for the chlorination of zirconium carbide.
1-chlorine entry tube; 2.-hatch; 3-coke; 4-carbide; 5Dinas bricks (299 mm); B-ZrCl, condenser; 7-expanding coupling; 8-bin (used only during the charging); 9-carbide; 10 -exhaust; 11-flexible coupling; 12- sieve; 1 3-electro magnetic vibrator; 14-fine fraction; 15-coarse fraction.

The cost of zirconium chloride produced by chlorination of zirconium carbonitride is higher than that of the chloride produced by direct chlorination of zircon. T h e process has the advantages o f a low chlorination temperature, simplicity of operation, and the production of technical chloride of a higher degree of purity.

Chlorination of zirconium dioxide

1, 4/

Zirconium dioxide is usually chlorinated if it is the end product of the separation of hafnium from zirconium and if in the production of zirconium metal the pure zirconium dioxide has to be converted into the chloride. The following main reactions a r e involved in the chlorination of zirconium dioxide in a mixture with carbon:

-ZzrO, 1
2

C + Cl, = -ZrCI, 1
2

+-COCO,. 1
2

(3)

236

At temperatures above 700" the chlorination preferentially takes place in accordance with equation (l), which involves the g r e a t e s t d e c r e a s e in free energy. The chlorination is c a r r i e d out in shaft furnaces charged with the pelletized raw materials. A diagram of one of the furnaces used and the chlorination apparatus is shown in Figure 117. The furnace consists of a s t e e l shell lined with Dinas bricks. The internal dimensions of the shaft a r e : diameter 650 to 700mm, height about 2000". The current is supplied to t h r e e graphite plates mounted in the furnace lining.

FIGURE 117. Apparatus for the chlorination of a pelletized mixture of zirconium dioxide and carbon.
1-charging bin: 2-feeder; 3-quartz bricks: 4-manholes (for cleaning); 5-steel shell; 6-insulation bricks; 7-cooled copper electrode for the supply of the current; 8-graphite electrode with a sleeve; 9-chlorine supply tubes; 10-gas exhaust (nickel tube); 11- first condenser; 12-air jacket; 13-partition; 14-discharge hole; 15-air heater; 16-fan; 17-cyclone-type condenser; 18-water scrubber; 19-scrubber irrigated with an alkali solution; 20-exhaust tube for the gas (to the cooler); 21-circulation pump; 22-tube for the entry of the wash water.

The charge contains 81% zirconium dioxide, 14.5% g a s black and 4.570 dextrin (which a c t s a s a binder). The mixture is moistened with water and pelletized with a r o l l e r p r e s s (the s i z e of the pellets is 3 5 X 25 X The pellets a r e dried a t 140". 20"). When starting the furnace, the pellets a r e charged t o a level somewhat above the middle of the electrodes. Carbon plates a r e laid o v e r the l a y e r of pellets (from the center to each electrode); the plates a r e in contact

237

with the electrodes and s e r v e f o r the initial heating of the pellets. An additional amount of the pellets is then fed to the furnace, until the pellets r e a c h the required level. The electrical conductivity of the hot pellets is high enough to p e r m i t heating the charge by a d i r e c t passage of e l e c t r i c current. The furnace temperature in the vicinity of the h e a t e r s is main tained a t about goo", and in the upper zone of the furnace a t 250 to 500". Under normal operating conditions the voltage applied to the electrodes is 15 V and the c u r r e n t is 250 amperes. The chlorine pipelines, the charging tubes, the g a s exhaust lines, and the condensers a r e made of nickel sheets. The condensers a r e fitted with jackets within which t h e r e is a circulation of heated air. The temperature of the f i r s t condenser is maintained a t 150 to 200" and of the second condenser a t 100 to 150". The excess chlorine is absorbed in s c r u b b e r s (irrigated with alkaline water) which a r e fitted after the condensers. The consumption of electrical energy is 3 kwh/kg of zirconium chloride. The degree of extraction of zirconium into the chloride is 92 to 93%. Zirconium chloride prepared by one of the above methods contains a number of impurities. The chloride is purified by sublimation ( s e e Section 65).

6 3 . METHODS FOR THE SEPARATION OF HAFNIUM AND ZIRCONIUM

The zirconium used in nuclear power plants must contain l e s s than 0.01 70hafnium, whereas zirconium m i n e r a l s always contain hafnium. The hafnium content of zircon is usually between 0.5 and 270, but some zircon varieties (e. g . , cyrtolite) contain much l a r g e r amounts of hafnium. Recently, the production of pure hafnium and hafnium compounds has also attracted interest. Many methods for the separation of these elements have been studied, the most important being: 1) fractional crystallization of the fluoride complexes; 2) fractional distillation (rectification) of the halides and other compounds; 3) extraction with organic solvents; 4) ion- exchange methods ; 5) selective reduction of the chlorides.

Fractional crystallization of complex fluorides The fractional crystallization of the complex fluorides KzZrFs and K2HfF, has found industrial application in the USSR 1 3 1 . The molar solubility of the hafnium s a l t is about 1.5 times that of the zirconium s a l t ( s e e Table 41). A s a result, fractional crystallization r e s u l t s in the concentration of hafnium (together with the niobium and iron present as contaminants) in the mother liquor. Studies c a r r i e d out in the USSR have shown that 16 to 18 successive recrystallizations reduce the hafnium content of KzZrF6from 2.5% (in % Z r ) to a few thousandths of one percent / 3 / (Figure 118).

238

Number of crystallizations FIGURE 118. Change in the hafnium content of K2ZrF, as a function of t h e number of successive crystallizations.

FIGURE 119. Flow sheet of the separation of hafnium from zirconium by fractional crystallization of K2ZrFs.

A flow sheet of the fractional crystallization p r o c e s s is shown in Figure 119. The K2ZrF, c r y s t a l s a r e dissolved in w a t e r at 90" in stainless steel vessels; each fraction of the crystals (except the l a s t two fractions)

239

is dissolved in the mother liquor of the preceding crystallization. The solid :liquid r a t i o during the dissolution is 1 : 7, which corresponds to a K2ZrF6concentration of 0.5 m o l e s / l (or about 140 g/l). The solutions a r e cooled to 1 7 t o 19". The solubility of KzZrF6 at that temperature is 16.3 g/l. The c r y s t a l s formed a r e allowed to settle, the mother liquor is decanted, the mother liquor from the preceding stage is poured on the c r y s t a l s and the next crystallization stage performed.

TABLE 41
Solubilities of zirconium- and hafnium-potassium hexafluorides Solubility ratio K2HfF6/ K2ZrF6
0.1942 0.1008

1.5 1.51

The f i r s t and second mother liquors which a r e most concentrated in hafnium a r e withdrawn from the crystallization cycle ( F i g u r e 119) and evaporated to 1 / 5 to 1 / 6 of the initial volume. The K2ZrF6 c r y s t a l s precipitated from the evaporated solution and the hafnium present in the c r y s t a l s as an impurity a r e returned to the f i r s t crystallization stage, while zirconium hydroxide containing about 6% Hf is precipitated from the mother liquor by the addition of ammonia; the hydroxide is used a s the s t a r t i n g compound for the production of p u r e hafnium. With the above method, theyield of pure K2ZrF6c r y s t a l s (with a 0.010/0) is 800/. hafnium content

Separation by extraction Zirconium and hafnium may be separated by selective extraction (from aqueous solutions) with various organic solvents: organophosphorus compounds, ketones, and amines. As an example, we shall consider the extractions with tributyl phosphate and with methyl isobutyl ketone. The extraction methods a r e distinguished by their high output; they can be c a r r i e d out a s a continuous process. Extraction with @%butyl phosphate (TBP)*,Tributyl phosphate extracts zirconium and hafnium from solutions containing the oxychlorides or n i t r a t e s of these elements. The extraction is c a r r i e d out most con veniently from nitrate solutions containing f r e e n i t r i c acid. In the organic phase zirconium and hafnium n i t r a t e s a r e present a s complexes with TBP. The following reaction takes place during the extraction

_____-_____--__

Zr02+ 2H+

+ 4 N 0 3 + PTBP:
240

Zr (NO,),

2TBP+ HpO.

T h e properties of tributyl phosphate have been described above.

The equilibrium constant of the reaction is

Y=
The r a t i o

[Zr (NO,), 2TBP ] [Zr02+] [H+]* [NOT] [TBP]

where a is the distribution coefficient. Hence


aZr

=K

[H+]*

[TBPj.

This equation shows that the degree of extraction i n c r e a s e s with increasing acidity, nitrate ion concentration and T B P concentration. A s is evident from Figure 120, the distribution coefficient of zirconium (az r) is higher than that of hafnium a t all nitric acid concentrations / 1O / . The higher extractability of zirconium is due to the lower degree of dissociation of i t s nitrate. The degree of dissociation of hafnium nitrate d e c r e a s e s sharply with increasing nitric acid concentration. Hence, the separation coefficient

%r aH
f

d e c r e a s e s from 12 to 4 a s the H N Q

concentration is increased from 6 N to 9N (Figure 120). The extraction should be c a r r i e d out from solutions containing 5 to 6 N HN03 (i. e. , a t high separation coefficients). Since the T B P has a high density and viscosity, i t is usually mixed with i n e r t diluents. The extraction is c a r r i e d out in columns or in e x t r a c t o r s of the m i x e r - s e t t l e r type ( F i g u r e s63 to 66). The extraction in a fourteen-stage extractor of the m i x e r - s e t t l e r type is schematically shown in Figure 121. The feed solution containing 125g / l Z r a , 5 M H N Q and 2.470 Hf (in70 Z r ) is introduced at the fifth stage. Here it i s combined with the washing solution (5.4 M HN03) which is moved from the f i r s t stage in countercurrent to the organic solvent, which is introduced at the 5 6 7 8 9 fourteenth stage. A 40% solution of TBP in Concentration HNO,, M n-heptane (C7H6)is used a s the extractant. The FIGURE 120. Dependence of the zirconium and the nitric acid a r e reextracted
distribution and separation cofrom the organic phase with water. To recover
efficients of zirconium and
the nitric acid, the aqueous solution is evaporated
hanfium on the nitric acid conto dryness and n i t r i c acid and nitrogen oxide centration. vapors a r e trapped. The extraction purifies zirconium not only f r o m hafnium but also from a number of other-impurities (Al, Ca, Fe, Mgl Si, Ti). Extraction with methyl isobutyl ketone (hexone) / 20/. Methyl isobutyl ketone is used in industrial practice a s an extractant f o r solutions of
24 1

zirconium and hafnium sulfates or oxychlorides containing ammonium thiocyanate NH&NS. The composition of zirconium compounds in such solutions has not been thoroughly studied. It can be expected that in addition to the simple thiocyanates Me(SCN)d the solutions also contain complex anions such as [Me(SCN)6]2-, (where Me=Zr o r Hf). The organic solvent is preliminarily saturated with thiocyanic acid HCNS. The hafnium is preferentially extracted into the organic phase, and the separation coefficient may he as high as 80.

Feed solution ( 1 2 5 g / l ZrO,


40%

+ 5MHNOs;

Hf/Zr = 2.5%)

TBP i n

FIGURE 121. Flow sheet of the separation of zirconium and hafnium by continuous extraction. Each square represents one mixer-settler stage.

The aqueous solution (reffinate) containing pure zirconium (the Hf content is 0.005%) is treated with the pure solvent which extracts HCNS. Zirconium is then separated f r o m the solution by one of the methods described above.
Solution of Zr(Hf)OSO, or Zr(Hf)OCl, + N b C N S

Extraction with methyl isobutyl ketone containing HCNS Aqueous phase Zr (up to 0.005% Hf) HCNS

11

J . = -1

Organic phase
(Hf

+a

part of Zr)

Washing with HC1 Aqueous phase (Zr)

3.

~-

Washing with methyl isobutyl ketone Methyl isobutyl ketone

Reextraction with sulfuric acid Aqueous solution of HfOSO,

-1

1 7Aqueous solution
. I
To the precipitation of the zirconium

Organic phase Recycled

To the tation of hafnium

FIGURE 122. Flow sheet of the separation of zirconium and hafnium by extraction with methyl isobutyl ketone from solutions containing thiocyanates.

242

The organic phase (extract) containing hafnium and some zirconium is washed with hydrochloric acid to extract the zirconium. The hafnium is then reextracted with a sulfuric acid solution and hafnium compounds a r e isolated from the solution. A flow sheet of the p r o c e s s is shown in Figure 122. Ion exchange methods*
Zirconium and hafnium may b e separated by ion exchange on cation exchange (when the zirconium and hafnium are present in the solution as cations) or anion exchange resins. T h e cation exchange separation from nitric acid solutions / l o / will b e described as an example. ZrOZf and HfOz+ are first sorbed on the cation exchange resin a t the top of t h e column or in a separate column until i t becomes saturated. They are then eluted with 0.5 M HzS04. Zirconium is eluted first and moves to the bottom of the c o l u m n a h e a d o f thehafnium. With a column of sufficient length this method may be used t o extract 95 -98% zirconium with a hafnium content of less than 0.01%

I
FIGURE 123. Flow sheet of the separation of zirconium and hafnium by ion exchange.
1-saturation column; 2-column for separation by elution; 3 pump; 4-storage tank for the starting solution containing Zr(Hf)(iSO& + 2 MHNOS; 5-storage tank containing 0.5M H,SO, used for elution of zirconium; 6-storage tank for 1 . 5 M H2S04 used for elution of hafnium; I-collector for solution of pure zirconium in H z S 0 4 ; 8-collector for the hafnium-containing sulfuric acid solution.

During the elution use is m a d e of the different tendencies of ZrO2+ and Hf02+ ions to form CZrO(S04)zJz~ and [HfO(S0,),12- complex ions. T h e flow sheet of the process is shown in Figure 123. T h e solution (17 g/l Zr and 2 M HN03) is circulated for a while through the saturation column. A 0.51 M solution of HzS04 is then passed first through the saturation column and then through the separatory column (which is packed with cation-exchange resin in the H-form) until hafnium appears in the solution a t the exit of the column. The filtrate, which is collected. contains zirconium.

___-----___--- * T h e principle of the ion exchange


has been discussed in Chapter VI.

methods for the separation of elements with similar properties

243

T h e elution is then continued with 1.5 M H,SO, t o extract the fraction containing the hafnium and the remaining fraction of the zirconium. This method may be used to recover 93% of the zirconium with a Hf content of 0.035%. The drawback of the ion exchange method is its low output (about 0 . 0 5 g Zr/hr per c m z of column cross section). It gives, however, a sharp separation and may b e used as a continuous method.

Rectification

'

This method of separation i s based on the differences between the boiling points of some zirconium and hafnium compounds - chlorides, complexes of Z r and Hf chlorides with phosphorus oxychloride, and alcoholates. The rectification of chlorides is the most important since the separation yields the pure chlorides ZrC14 and HfC14 which may be used directly for the production of the metals. The p r o c e s s is complicated by the fact that at atmospheric p r e s s u r e the zirconium and hafnium chlorides undergo sublimation at temperatures below their melting points. They can be melted only under the p r e s s u r e of their own vapor. Thus, ZrC1, m e l t s a t 437" under 18.7 atm. At that temperature the vapor p r e s s u r e of hafnium chloride is 31.8 atm. The ratio of these p r e s s u r e s (i. e . , 1.7) is the separation coefficient for one rectification stage. These data show that zirconium and hafnium can be separated by rectification, but the rectification must be c a r r i e d out under p r e s s u r e in o r d e r to have chloride vapors in equilibrium with the liquid phase. Research is being c a r r i e d out in this direction, and columns for rectification under p r e s s u r e a r e being developed. The rectification of "complex chlorides", i. e . , substances of the general formula 3 M e C14. 2POCl3 (where Me i s Z r o r Hf), has been studied m o r e thoroughly and has been used on a pilot-plant scale. Recent studies have shown that these "complex chlorides" a r e not s e p a r a t e species but azeotropic mixtures 1111. They a r e readily prepared by heating Z r Q with phosphorus pentachloride o r by the reaction of the chloride with phosphorus oxychloride: 3ZrC1, 2POC1, = 3ZrC1, . 2POC1,, 3210, 6PC1, = 3ZrC1,. 2POC1, 4POC1,.

The "complex chlorides" have low boiling points (the bp of Z r s a l t is


100") and t h e i r boiling points differ by 5". The boiling point of 3ZrC14. 2POCl3 i s 360 rt 1" and that of 3HfC14. 2P0Cl3 is 355 f 1".

According to different authors, the separation coefficient for the above two compounds is 1.14 to 1.16. Successful rectification may be c a r r i e d out in sieve-tray g l a s s columns and in nickel columns. A single rectification in a column with 50 t r a y s yielded a zirconium fraction containing < 0.00570 Hf (the initial Hf con centration was 1%). The yield of zirconium was 40%. Although the rectification of the complex chlorides is highly effective, it has the disadvantage of yielding a final product which cannnot be used directly for the production of the metal. A multistage processing method must be employed. The complex chloride is decomposed with an alkali solution to s e p a r a t e the phosphorus (which is dissolved a s the phosphate Na2HP04)and to p r e p a r e Z r a . The ZrOz is then chlorinated to yield ZrC1,.

244

. .......

..111.111111.

I..

, . I

....

I , .....I

I
T h e r e is a s i m p l e r method in which the complex chloride is fused with NaCl and is then subjected to thermal decomposition yielding sodium chlorozirconate 1211:
3ZrC1, 2POC1, 6NaCINa2ZrC16%

300-400'

2NaCI

+ ZrC1.t.

SNa,ZrCI,

+ 2POC1,t;

In this c a s e zirconium chloride is obtained directly. In another method, the complex chloride vapor is passed through a layer of coke heated to 800". The following reaction takes-place:
3zrc14 ~ P O C I ~ 2

~~ Z ~ C 5 I ,2 ~ ~ 1 2 ,~0t.

Phosphorus trichloride (bp 75") can be easily separated from ZrCll / 2 1 / .

Selective reduction of chlorides Zirconium chloride is reduced m o r e readily than hafnium chloride. Thus, at 400 to 450" ZrCIl is reduced by zirconium powder to the non volatile chlorides ZrCl3 and ZrC12. The reduction of ZrC1, (containing 1.5% Hf) by zirconium powder, at 420", yields a sublimate containing 30% Hf; the yield of Hf is up to 95%. The reduction product is a mixture of ZrCl,, ZrClz and the excess zirconium powder; i t s hafnium content is reduced to 0.05 to 0.1%. Hafnium-free zirconium may be isolated from the reduction product by heating in vacuo; this causes decomposition of the lower chloride by the reactions: 2ZrC1, -+ ZrC1, + ZrC1, and PZrCI, + Zr + ZrC1,. Hafnium-free zirconium chloride i s distilled off in the process. The residue containing the ZrClz and the zirconium contaminated with hafnium is returned to the reduction stage. The hafnium-rich sublimate ( u p to 3 0 % Hf) may be processed by the extraction and ion exchange methods described above. Aluminum may be used instead of zirconium powder a s the reducing agent. Quantitative reduction may then be c a r r i e d out at temperatures a s low a s 300". The hafnium-rich sublimate is contaminated with large amounts of aluminum chloride, but i t s separation i s not difficult. The selective reduction method is characterized by a high effectiveness and can be readily combined with the subsequent p r o c e s s e s for the conversion of chlorides to the m e t a l s ,

64.

MANUFACTURE OF ZIRCONIUM

The technological problems accompanying the production of ductile zirconium a r e s i m i l a r to those involved in the production of metallic titanium, which were discussed above.
245

Like titanium, zirconium actively absorbs oxygen, nitrogen, and hydrogen. The solubility of oxygen in zirconium reaches 40 at.% (10.6% by weight). Zirconium contairring m o r e than 0.2% oxygen cannot be mechanically worked. The solubility of nitrogen in zirconium is about 20 at.%. The mechanical properties and corrosion resistance of pure zirconium a r e strongly affected by i t s nitrogen content. The solubility of hydrogen in a-zirconium is up to 5 at.%, and in p-zirconium i t is higher. The presence of hydrogen in zirconium (even at concentrations of 0.003 70) markedly reduces the impact strength of the metal. Carbon, CO, and C 0 2 r e a c t with zirconium at high temperatures, yielding the refractory carbide Z r C (mp 3530"). The presence of carbon has little effect on the . mechanical properties of zirconium but reduces its corrosion resistance to water at high temperatures. In parallel with the similarity of their properties, there is a noticeable similarity in the methods of production of zirconium and titanium. These methods shall accordingly b e discussed m o r e briefly. They may be classified into three groups!: 1. Thermal reduction with metals: a) reduction of zirconium tetrachloride by magnesium; b) reduction of potassium fluozirconate KzZrFB by sodium and of zirconium fluoride by calcium; c) reduction of zirconium dioxide by calcium o r calcium hydride. 2. Electrolysis of molten salts. 3. Thermal dissociation of zirconium iodide. The main commercial method f o r the production of ductile zirconium is based on the reduction of ZrC14 with magnesium. Zirconium powder (which is used a s such) is produced by reduction of K2ZrF6by sodium and by reduction of Z r Q by calcium or calcium hydride. Electrolytic methods f o r the production of zirconium have been increasingly used. Thermal dissociation of zirconium iodide is used for the production of zirconium of the utmost purity. When necessary, the method is used f o r the refining of zirconium sponge o r powder. The strength of pure zirconium is not high enough and i t s corrosion resistance is markedly affected by the presence of very s m a l l amounts of certain contaminants (e. g . , nitrogen). The mechanical properties and the corrosion resistance a r e improved by the addition of alloying elements.

65. THERMA.L REDUCTION OF ZIRCONIUM CHLORIDE WITH MAGNESIUM /1,4,12, 20/ The reduction of zirconium chloride by magnesium is c a r r i e d out by the Kroll p r o c e s s ( s i m i l a r to the reduction of titanium chloride by magnesium) and is based on the interaction of ZrC14 vapor with liquid magnesium : '?rC14 (gas)

+ 2Mg (liquid) + Z r (solid) + 2MgC12(liquid) + 52.5 kcal.

The crude zirconium chloride, which is prepared by one of the above described methods (see Section 62), is purified by sublimation in a separate vessel (the separate process) o r in the apparatus used for the reduction (the combined process).

246

The s e p a r a t e p r o c e s s

Purification of zirconium tetrachloride by sublimation. Zirconium tetrachloride is usually contaminated with the oxychloride ZrOClz, with 1 t o 270i r o n (as FeC13), occasionally with chromium ( a s CrC13), and very s m a l l amounts of titanium and silicon. In addition i t may contain ( a s a mechanical contaminant) p a r t i c l e s entrained from the furnace during the chlorination. In o r d e r to remove i r o n and chromium, FeC13 and CrC13 are preliminarily reduced by hydrogen (at 200 to 300") t o FeClz and CrClZ. The lower chlorides have high boiling points (FeCl2 1030", CrClz 1300") and a r e not volatilized at the sublimation temperature of ZrC14 (450 t o ZrOClz is a l s o not volatilized under these conditions. 6 6 0 ' ) . A diagram of the apparatus used f o r the purification of zirconium chloride is shown in Figure 124. A crucible (made of a Ni-Cr alloy*) containing crude zirconium chloride is placed in a s t a i n l e s s s t e e l retort. In o r d e r to c r e a t e a l a r g e r volatilization surface the chloride is distributed on t r a y s placed one over the other. The r e t o r t is covered with a hermetically sealed lid made of lead-antimony eutectic (mp 247"). This alloy may be either cooled until solid (in o r d e r to maintain a g a s impermeable seal) or maintained in the molten s t a t e when the lid s e r v e s a s a safety valve. A coil-shaped, a i r - or water-cooled chloride condenser, which is introduced into the r e t o r t , is fastened to the lid. The r e t o r t is placedin a furnace with t h r e e temperature zones. The alloy is frozen by cooling, the v e s s e l is evacuated and then filled with hydrogen (these operations a r e repeated twice) and the reduction of i r o n chloride is c a r r i e d out in the lower p a r t of the r e t o r t a t 200 to 300". The HC1 vapor formed is then evacuated together with the hydrogen. The reduction is repeated, the alloy in the hydraulic s e a l is melted and the chloride is sublimed by gradually increasing the temperature in the lower p a r t of the r e t o r t f r o m 400 to 650". The temperature of the upper p a r t of the r e t o r t and the lid is maintained a t 400" (above the sublimation temperature of ZrC14) in o r d e r to ensure that the chloride condenses only on the coil, which is cooled f i r s t with a i r and then with water. The sublimation in the r e a c t o r 700" in diameter and 1750" high continues for about 30 hours. The sublimation yields about 200kg of purified chloride. The iron content in the purified product is reduced to 0.1 % , and the iron is present a s FeC12. Additional amounts of i r o n a r e subsequently removed in the reduction of the chloride vapor to the metal. The yield of the sublimed chloride depends on the purity of the starting chloride (mainly on the ZrOCl2 content) and ranges from 90 to 96%. The apparatus is cooled and the lid with the chloride condensate is t r a n s f e r r e d to the apparatus f o r thermal reduction by magnesium. The density of the condensate on the coil is about 2.2 to 2.3 g/cm3. Reduction. The reduction is c a r r i e d out in an apparatus resembling the sublimation vessel (Figure 125). A steel crucible containing the required amount of magnesium b a r s (a 20% excess of magnesium is taken)

* The alloys used in Inconel X .

with the composition: 72.7% Ni. 15%Cr, 7% Fe. 1%Nb. 2.5% Ti, 0.75 A l . 0.7% Mn. 0.4% Si, and 0.05% C.

247

is placed on the floor of the vessel. The converter with sublimed zirconium chloride is placed i n the upper p a r t of the vessel. The lid is sealed in the s a m e way a s in the sublimation apparatus. A drip pan and shields a r e fitted over the crucible to prevent radiation heat transfer and the penetration of solid zirconium chloride particles into the crucible.

FIGURE 124. Diagram of vessel used for the purification of ZrC14 by subli mation.
1-valves for the supply of hydrogen and the evacuation and removal of gases; 2-heater of the hydraulic seal; 3-hydraulic seal (Pb-Sb alloy): 4 coil (condenser); 5-retort; 6-crucible containing crude ZrC1,; I and 8-heaters.

FIGURE 125. Diagram of vessel used for the reduction of ZrC14 by magnesium.
1- retort; 2-hydraulic seal; 3 condenser with the sublimed chloride; 4-crucible; 5 and 6 shields; 7-lid; 8-valves; 9 heaters; 10-reaction mixture.

The apparatus h a s three heating zones: the lower zone containing the crucible, the intermediate zone containing the chloride, and the upper zone, which s e r v e s for the heating of the ring-shaped groove and the lid. Before the beginning of the reduction the apparatus is evacuated and filled with argon (the evacuation and filling is repeated in o r d e r to ensure complete removal of air). The lower zone is then heated to 825", i. e . ,

248

t o a temperature above the melting point of MgClz which is formed in the reduction process. At the s a m e time the intermediate zone of the r e t o r t is heated to 450 to 500, i. e . , to the temperature required f o r the volatilization of zirconium chloride. At the end of the reduction the temperature of the intermediate zone is increased to 650". The reaction of the gaseous ZrC14 with the liquid magnesium r e s u l t s in a continuous removal of chloride f r o m the gaseous phase and thus in the volatilization of f u r t h e r amounts of chloride from the coil. The presence of an i n e r t gas in the r e t o r t reduces the volatilization r a t e of the chloride and thus reduces the reduction rate. Rapid volatiliza tion takes place in ;he absence 'of'an i n e r t gas, and the reaction is SO f a s t that the reaction m a s s may be o w h & a t e dand the crucible melted. This is due to the fact that the reduction is exothermic. The p r e s s u r e , which may i n c r e a s e a s a result of the overheating of the reaction m a s s during the reduction is automatically decreased by maintaining the Pb-Sb alloy in a molten s t a t e ( a t 250 to 300"). Two hundred kg of ZrC1, a r e reduced in one operation in a v e s s e l 700" in diameter and 1750 m m high, yielding about 75 kg of zirconium. The duration of the p r o c e s s is 2 4 to 30 hours. The zirconium yield is 93%. The container i s cooled to room temperature and the crucible is taken out. The zirconium is at the bottom of the crucible in the form of a sponge. It is protected against oxidation by a dense layer of magnesium chloride.

The combined p r o c e s s The vessel used for the combined sublimation of the chloride and i t s subsequent reduction with magnesium is shown in Figure 126. Its use eliminates the t r a n s f e r of the sublimed ZrC1, from one vessel to another, which brings the chloride into contact with a i r and causes i t s partial hydrolysis. Moreover, carrying out both p r o c e s s e s in a single vessel reduces the total p r o c e s s time by about 3 5 70. In the combined vessel (Figure 126) a container with zirconium chloride is fitted over the crucible containing the magnesium b a r s . The vessel is covered with a lid fitted with a coil. It is sealed with the aid of a Pb-Sb alloy seal, a s described above. At f i r s t the vessel i s evacuated, filled with hydrogen and the iron and chromium chlorides a r e reduced at 300". It is evacuated to remove the HC1 vapors, then filled with argon and the reduction is c a r r i e d out. The lower zone of the apparatus is heated to 825", which also causes an i n c r e a s e in the temperature of the intermediate zone (to about 400 to 450'). The ZrC1, is volatilized and r e a c t s with the molten magnesium. When the apparatus is charged with 250 kg of ZrCl,, the reduction time is about 15 hours. The temperature of the intermediate zone is thenincreased to 650" in o r d e r to volatilize the remaining ZrC1, which condenses on the water-cooled coil. The total p r o c e s s cycle has a duration of 44 hours, a s compared with about 64 hours in the s e p a r a t e process.

249

1.'1

I1

I I

I I

II

I11

1111

11111

111

II.I11111111111111111111

I III~IIIIII11111111111111111111

FIGURE 126. Diagram of apparatus for the combined purification of ZrC1, by sublimation and reduction of vapor with magnesium. 1-crucible with crude ZrC1, ; 2-coil condenser: 3-hydraulic seal (Pb-Sb alloy): 4-heaters; 5 retort; 6-shields; I-crucible with magnesium.

The vacuum- thermal distillation of magnesium and magnesium chloride Magnesium and magnesium chloride a r e separated from zirconium by vacuum distillation at 900 to 920, as in the case of titanium. A diagram of the vacuum-thermal distillation apparatus is shown in Figure 127. The crucible used f o r the reduction, together with the reaction mixture, is placed upside downin the r e t o r t . The r e t o r t i s evacuated to l o q 5 m m H g and the crucible is heated to 825; a s a r e s u l t the to bulk of the magnesium chloride descends through a funnel into a ringshaped stainless-steel container placed below the crucible. The remaining magnesium and chloride a r e distilled and condense in the lower, cooled p a r t of the r e t o r t . The upper p a r t of the r e t o r t which contains the crucible is heated with the aid of a lifting electrical furnace. The space between the r e t o r t and the furnace is evacuated in o r d e r to prevent bending of r e t o r t walls under the atmospheric p r e s s u r e . The removal of magnesium and magnesium chloride is accompanied by the removal of the hydrogen absorbed by the zirconium. The duration of the distillation is 1 2 to 1 6 hours. The r e t o r t is cooled f o r a long time in vacuo. In o r d e r to accelerate the cooling of the r e t o r t

250

the vacuum furnace is raised when the temperature of the upper p a r t drops to 400'. Various types of sponge a r e formed in the crucible after the distillation. The bulk (70%) consists of a dense metal containing almost no magnesium o r magnesium chloride.

FIGURE 127. Diagram of installation for the vacuum-thermal purification of zirconium sponge.
1 -retort; 2-collector for molten MgC1,; 3-lifting electrical furnace; 4-thermocouple; 5-crucible containing the sponge; 6-funnel; I-rubber gaskets (water-cooled vacuum seals); 8-oil-cooling system for the lower part of the retort; 9-water circulation heat exchanger for the cooling of oil; 10-diffusion pump; preliminary vacuum pump; 12-lifting device.

The sponge is extracted from the crucible by means of a pneumatic rack p r e s s o r a chisel in an argon atmosphere and is ground in a conical c r u s h e r to a particle size of about 6 mm. The l a r g e particles a r e crushed with the aid of a 200 ton p r e s s . Uniform 100 kg batches of sponge a r e prepared by mixing various types of sponge in mixers. The approximate impurities content (in 70 by weight) of the zirconium sponge is:

25 1

. . 0.003 . . <Z.IO-' . . 0.0095 . . (0.002 co . . 4 0 0002 Cr . . -0:0015


A1 B
C cd

cu

Fe

H f

2 N

. . . 0.0085 . , 0.01 . . 0.002 . . . 0.004

. . .0.003 . . . 0.065

Ni 0

Pb S i

Ti

. . <0.0002 .0.064.1 .. . 0.007 . . 0.002 . . ~0.005 . . <0.001

66. REDUCTION OF POTASSIUM FLUOZIRCONATE BY SODIUM

The complex fluoride KzZrFs is reduced by sodium i f the metal powder produced need not be of a high degree of purity (e. g . , for use in pyro technics o r in electronics). As compared with the reduction of zirconium chloride, the reduction of potassium fluozirconate has the advantages of being a simple production technique and of being convenient in handling since the s a l t is not hygroscopic and is stable in the air. The reduction is c a r r i e d out with metallic sodium since the sodium fluoride formed in the reaction can be readily leached with water out of the bulk of the zirconium. The reduction i s based on the reaction: KIZrF,

+ 4Na = Zr + 4NaF + 2KF.

The heat of reaction is about 2 6 4 kcal/kg charge ( K 2 Z r F 6+ 4Na). The heat of reaction does not suffice to support a spontaneous reaction and external heating i s required. The reduction may be c a r r i e d out in hermetically sealed s t e e l r e a c t o r s of the "bomb" type, which a r e heated to 800 to 900". After cooling the solidified m a s s is taken out of the crucible by crushing with a pneumatic hammer. In o r d e r to s e p a r a t e the s a l t s from the powder, the m a s s is ground in a wet mill and leached with water in tanks fitted with s t i r r e r s . Since the reaction m a s s contains an excess of sodium which r e a c t s vigorously with water, the m a t e r i a l is gradually introduced into the water in small portions. In o r d e r to leach out the iron, the m a t e r i a l is treated with dilute hydrochloric acid, rinsed with water, separated by filtration and dried a t 60". The drying m u s t be c a r r i e d out with caution since the m a t e r i a l has a tendency to b u r s t into flame spontaneously.

67. REDUCTION OF ZIRCONIUM DIOXIDE B Y CALCIUM AND CALCIUM HYDRIDE

Zirconium dioxide i s reduced by calcium o r calcium hydride under conditions resembling those used f o r titanium ( s e e above). Even in the presence of a l a r g e excess of calcium (50 to 1000/) and even if the reaction is c a r r i e d out in hermetically sealed apparatus, the powder obtained a f t e r leaching the calcium oxide contains 0.3 to 1 % oxygen and 0.03 to 170nitrogen. When the reduction is c a r r i e d out in a purified argon atmosphere, the nitrogen content depends on the nitrogen content

252

of the calcium. The powders produced a r e suitable f o r use in vacuum technology, pyrotechnics, photographic flash powders and for military uses. The reduction may be c a r r i e d out in hermetically sealed r e a c t o r s ; when using s m a l l s i z e r e a c t o r s , the lid is welded before each reduction. The reaction is conducted at 950 to 1100". The calcium oxide is leached with dilute hydrochloric acid. The oxygen content is reduced by holding the zirconium powder in a molten calcium bath o r in calcium vapor; the use of calcium vapor is m o r e effective. The difficulties experienced in freeing the zirconium f r o m oxygen a r e due to the fact that the oxygen is in the form of a solid solution. The reduction by calcium hydride is c a r r i e d out at approximately the s a m e temperatures a s those used in the reduction with calcium (900 to 1100") ; the product is zirconium hydride, which is used f o r the production of zirconium a r t i c l e s by powder metallurgy techniques.

68.

PRODUCTION OF ZIRCONIUM BY ELECTROLYSIS

Many workers found that zirconium may be produced by electrolytic reduction of i t s halides in molten salts. The best r e s u l t s were obtained when the electrolysis was c a r r i e d out in baths consisting of a solution of K2ZrF, in NaCl o r KCI. Foreign workers recommend the use of a bath containing 2 0 % K2ZrF, + 80% NaCl / 11. Soviet workers pointed out the advantages of electrolytes containing 25 to 3 0 % KzZrFs and 70 to 75% KC1 (the decomposition potential of KC1 i s higher than that of NaC1; the solubility of K F in w a t e r i s higher than the solubility of NaF, which facilitates the separation of the electrolyte from zirconium powder) 1 2 2 1 . The cathodic reaction involved in the electrolysis i s : Z r F i - z Zr4+ 6F-; ~ r ' + 4e + ~ r .

Fluoride ions a r e discharged at the anode. However, the elemental fluorine reacts immediately with the chloride ions. Thus, chlorine i s evolved at the anode. The anodic reaction i s :
4F-

+ 4NaCl-

4e

4NaF

+ 2Cl,t.

Thus, potassium and sodium fluorides gradually accumulate in the bath. The optimum conditions of electrolysis in NaCl and KC1 baths a r e s i m i l a r . A high degree of separation of zirconium from hafnium is obtained by electrolysis of fluoride-chloride melts. The potential of hafnium is lower than that of zirconium and hafnium accumulates in the melt. Thus, the cathodic deposit from a melt containing K2ZrF6with 0.6% Hf ( i n % of the Z r + Hf) contains only 0.05 70Hf / 231. In o r d e r to produce high- quality zirconium powder, the fluozirconate used must be of a high degree of purity and the p r o c e s s must be c a r r i e d o u t in a purified argon atmoswhere. The design of one type of hermetically sealed electrolyzer f o r the production of zirconium i s shown in Figure 128. Thepuregraphite crucible, which is heated by means of a graphite heater, s e r v e s a s the

253

anode. A s t e e l o r molybdenum rod may be used as the cathode. A carbonblack packing is used a s t h e r m a l insulation. I t is very important t o d r y in advance the crucible, the whole electrol y z e r and the initial s a l t s used in the process. The electrolyte i s purified by preliminary electrolysis a t a reduced potential (1.5 to 2V) using graphite cathode. This causes the deposition of metallic contaminants m o r e noble than zirconium. The cathode is then replaced and the electrolysis is c a r r i e d out at 7 5 0 to 860' at a potential of 3.5 t o 4V and a c u r r e n t density of 2.5 to 4 amp/cm2. The metal is deposited on the cathode in the form of c o a r s e c r y s t a l s 0.3 to 0.5 mm in size. The rate of deposition is 0.5 g / a m p . h r , the c u r r e n t efficiency is 6 0 to 65%.

FIGURE 128. Diagram of a hermetically sealed electrolyzer for the production of zirconium.
1-graphite crucible serving as the anode; 2 graphite electrical heater; 3-heater electrodes: 4-hermetically sealed jacket; 5-chamher for cooling the cathodic deposir in argon; 6-shield.

When a sufficient amount of zirconium has collected, the cathode is lifted into the cooling chamber. The cathodic m a t e r i a l contains about 2570 zirconium and 75% s a l t s (NaC1, KCI, KF, NaF). It is crushed and leached with hot water. The zirconium powder is rinsed with alcohol or acetone and dried. It has the following composition: 99.8 to 99.9% Zr, 0.03 to 0.0570 C, 0.01 to 0.027'0 N, and 0.04 to 0.07% C.

254

When graphite crucibles a r e used for the electrolysis, the duration of the electrolysis is limited by the gradual saturation of the graphite crucible with the melt. This drawback is eliminated in an electrolyzer developed by Soviet scientists. The bath is made of s t a i n l e s s s t e e l and is cooled with water. The electrolyte in contact with the walls solidifies, creating a c r u s t which protects the walls against corrosion 1 2 2 1 . After remelting in vacuum the electrolytic zirconium h a s mechanical properties s i m i l a r to those of zirconium produced by t h e r m a l reduction of zirconium chloride with magnesium. If proper equipment f o r continuous o r semicontinuous electrolysis is developed, the electrolytic production of zirconium will b e able to compete with the method based on t h e r m a l reduction with magnesium.

69. THE THERMAL DISSOCIATION (IODIDE) METHOD 1 1 . 6 , 13, 141

Zirconium of the highest degree of purity is usually produced by refining the metal by thermal dissociation of zirconium iodide; the nature of the p r o c e s s was described above in the chapter on titanium ( s e e Section 51). The iodide purification method is based on a reversible reaction which is c a r r i e d out in a single vessel.

zr + 2 1 ~or crude 7

250-300' 470--550",

~ ~ 1 ~ 1 3 0 0 - 1 4 0 0 , + ' ~ 21 ~ ~.

'(vapor)

purified

J.

(vapor)

The r a t e of deposition of zirconium on the filament is determined by the r a t e s of transfer of iodide to the filament surface and of iodine to the zirconium to be purified. This is the cause of the relatively low r a t e of the iodide process. Moreover, the p r o c e s s r a t e depends on the tempera t u r e s of the filament and of the crude zirconium. The optimum filament temperature is 1300". The zirconium being purified is maintained at 250 to 300 o r 470 to 550". A s h a r p drop in the p r o c e s s r a t e is observed a t 300 to 450". This is attributed to the interaction a t these temperatures of Z r h with zirconium, with the formation of the nonvolatile Zr13: 3Zr14 + Z r
+

421-13

with the result that the concentration of Z r b in the gaseous phase decreases. The vapor p r e s s u r e of ZrI, i n c r e a s e s above 450" because of the disproportionation of the lower chlorides: ZZrI, e ZrIz

+ Zrb,

2ZrIz e Zr14. Industrial equipment f o r the iodide p r o c e s s is made of anickel-chromium alloy. Its design is the s a m e a s that described e a r l i e r ( s e e Figure 106).

255

In o r d e r to maintain the required temperature of the walls adjacent to the crude zirconium, the apparatus is i m m e r s e d in a molten-salt thermo stat. The filament temperature is controlled by the method described in the chapter on titanium. Zirconium sponge (prepared by t h e r m a l reduction of the chloride with magnesium) o r powder (prepared by reducing zirconium dioxide with calcium o r by electrolysis) may be used a s the starting material. About 5 0 g of iodine a r e introduced into the v e s s e l f o r each kg of crude zirconium taken. A number of V-shaped zirconium filaments is usually mounted inside the vessel; the filament diameter is about 2 mm. Zirconium rods, 25 to 3 0 mm in diameter and up to 2 m long, a r e obtained at the end of the process. Up to 50 kg of purified zirconium is produced in diameter in one vessel. Depending on the conditions, a rod 25" can be grown within 3 0 to 40 hours. The iodide purification p r o c e s s removes oxygen and nitrogen (since zirconium nitrides and oxides do not r e a c t with iodine) and metallic contaminants which do not form volatile iodides. The concentration of impurities in zirconium produced by the iodide p r o c e s s i s , 70:

. . . 3.10-3 . . . 2.10-' . . . 3.10-3 Cu . . .<5.10-5 Ti . . . Ca . . .<5.10-3


Si Fe Ai

Mn
Ni

.<1-10-3 3 .. . . Mo . . .<l.10-3
.<1.10-3

. . .~1.10-~ . ...
3.1OP3

C r

..

H C

o ...

Sn N

. . .~1.10-~ . . . 1.10-3
1.10-2
. . . l.10-3
. . . 1.10-2

70.

PRODUCTION OF SOLID ZIRCONIUM

A s in the c a s e of titanium, solid zirconium is produced by melting in a r c furnaces with consumable electrodes. The design of such furnaces has been described above ( s e e Chapters 1 1and IV). The electrodes a r e prepared in advance by p r e s s i n g zirconium sponge. If necessary, the alloying additives a r e introduced during the melting. The melting is c a r r i e d out in vacuo o r in an i n e r t g a s atmosphere (argon, helium). In most c a s e s uniform ingots a r e produced by remelting in which the ingot made in the f i r s t melting is used a s the consumable electrode. The melting r a t e is l l O k g / h r (for ingots 152" in diameter) and the consumption of electrical energy is 1.2 kwh/kg; in the c a s e of ingots 254" in diameter the melting r a t e is 270kg/hr and the consumption of electrical energy is about 0.8 kwhlkg. A s m a l l fraction of the zirconium produced is compacted (mainly to yield s m a l l a r t i c l e s of i r r e g u l a r shape) by powder metallurgy methods. Zirconium hydride powders a r e preferentially used in powder metallurgy. Electrolytic powders and powders prepared f r o m zirconium sponge ( s e e chapter on titanium) may also be used. The metal o r hydride powders a r e pelletized by pressing under 6 to 8 tons/cm2. The pellets a r e sintered in vacuo (lom4 to Hg) a t 1 2 0 0 to 1300". Pellets with a density close to the theoretical, which may be p r e s s u r e worked, a r e produced by using suitable sintering times. Zirconium alloys

256

may also be produced by powder metallurgy. The metal produced by powder metallurgy usually has higher oxygen and nitrogen contents than the metal produced by vacuum melting of the sponge. It has a lower corrosion resistance. The mechanical properties of solid zirconium prepared by various methods a r e compared in Table 42.
TABLE 42 Mechanical properties of zirconium prepared by the iodide process and by thermal reduction with magnesium Tensile strength Elongation of a 50.8 mm
sample, % Vickers hardness.

Zirconium type

kgl mmz

kg/mm2

Sponge m a d e by thermal reduction with magnesium. Molten in arc furnace. Forged and rolled a t l O O O " , cold-rolled with 30% reduction. Annealed for 1 hour a t 700". Prepared by the iodide process. Molten in arc furnace. Forged and rolled a t 790". Sub sequent processing as for sponge zirconium.

44.5

30

180

25.0

36

104

257

Part Two
THE RARE-EARTH METALS

Chapter VI
THE RARE -EARTH METALS (LANTHANIDES)
71.
GENERAL DATA O M RARE-EARTH METALS

The place of the lanthanides in the periodic system and their electron s t r u c t u r e 13, 101 The lanthanide group comprises 14 elements with atomic numbers from 58 (cerium) to 71 (Mecium) which a r e located in row 6 of the periodic table following lanthanum, and whose properties resemble those of lanthanum. For this reason 1a n t h a n u m is usually included in the group and the elements a r e known a s lanthanides - e. g., resembling lanthanum, Moreover, s c a n d i u m a n d y t t r i u m , which belong to Group 111, a r e chemical analogs of lanthanum, and which (especially yttrium) a r e almost always present together with lanthanides in minerals a r e also considered a s lanthanides. Lanthanides usually have a place apart in the periodic table, at the bottom of the table ( s e e Table 1). They all have very similar physicochemical properties. This is due to the peculiar structure of their electron shells. It is well known that the chemical and many physical properties of the elements depend mainly on the s t r u c t u r e of the external electron shells. In the lanthanides the s t r u c t u r e of the two outer shells (the 0 and P shells) remains the s a m e a s the charge on the nucleus (i. e . , the atomic number) is increased, since the transition from one element to another is due to the filling of the inner 4f-electrcn level ( s e e Table 43). The maximum number of electrons in the f-orbital is 14, and this is also the total number of the lanthanides. The 4f-electrons a r e relatively unaffected by external factors since they a r e shielded by the electrons in the outer orbitals. Hence, they have little effect on the chemical properties of the lanthanides. In the ground state the lanthanide atoms (except gadolinium and lutecium) have no electrons in the 5d-orbital ( s e e Table 43). However, the transfer of an electron from the 4f- to the 5d-orbital requires a small amount of energy. The valency of 3+ characteristic of the lanthanides is associated with the transfer of one electrom the the 4 f - to the 5d-orbital. The valency bonds then involve two electrons from the outer 6s-orbital and one electron from the 5d-orbital. In addition to the valency of 3+, some lanthanides may have valencies of 2+ or 4+ (Table 43). These "abnormal" valencies a r e attributed to the variation in the binding strengths of 4 f electrons with their number. The binding strength increases a s the number of 4 f-electrons v a r i e s from 1 to 7 and again from 8 to 14,i. e., the maximum binding strengths correspond to a half-filled and a completely

258

filled 4f-orbital. Hence, gadolinium and lutecium have the most stable f-orbital configurations. A valency of 4+ is exhibited by cerium and praseodymium (the f i r s t f-electrons a r e easily t r a n s f e r r e d to the W-orbital) and terbium and dysprosium, which follow gadolinium in the s e r i e s . A valency of 2+ is exhibited by samarium, europium, and ytterbium, i. e., in elements in which the number of electrons in the f-orbital is either equal o r close to 7 o r 14.
TABLE 4 3 Electron structure, valency, ionic radii and ionization potentials of the lanthanides ~ ~

.
m

F
U
#

N-shell (n=4)

o-shell ( n = 5 )
0

2 ol

2
~ ~

.d I

m >

k%
E . .O+ -* m 4

'g
?
57La 58ce 59Pr 60Nd

5s

5P

..
5

8
E
U
~

2 4s 2 2 2 2 2 2 2 2 2 2 2 2 2 2 2 4P 4d 10 10 10 LO
10 10 10 10 10 41

L
2 2 2 2 2 2 2 2 2 2 2 2 2 2 2

:. t
9&
1.22 1.18 1.16 1.15 36.2 37.2 37.5 37.8 38.2 38.2 38.8 38.6 39.4 39.5 40.0 40.2 40.3 40.8 41 .O

6 6 6
6

2
3 4

61Pm
62Sm 63Eu 64Gd 65Tb

6 6
6

6
6

66Dy
67Ho 68Er 69Tu 70Yb 71Lu

6 6 6 6 6

10 10 10 10
10

5 6 7 7 9 10 11 12 13 14

10

2 2 2 2 2 2 2 2 2 2 2 2 2 2 2

6 6
6 6

6
6 6

6
6 6

6
6 6

3+ 3.4+ . 4 (51H 3-t 3+ 2.3+ 2.3+ 3+ 3.4+ 3 (4)+

:+

6
6

3 2 k
3+

1.13 1.13 1.11 1.09 1.07 1.05 1.04 1.04 1 .oo 0.99

__

The lanthanides a r e subdivided into two groups: the cerium group [ ( L a ) ,Ce, Pr, Nd, Pm, Sm, and Eu] and the yttrium group [Gd, Tb, Dy, Ho, Er, T u , Yb, Lu, and (31. This subdivision was originally based on the differences in solubility between the double sulfates of the lanthanides and sodium o r potassium. However, subsequent studies revealed periodic variations of some properties within the lanthanide s e r i e s , which were in agreement with their separation into two subgroups. Thus, there i s a n analogy in the variations of the valency s t a t e s within the two subgroups ( s e e Figure 129) and in the color of the ions: the color of solutions of the trivalent ions of the f i r s t seven elements resembles, in the r e v e r s e o r d e r , the color of the next seven elements ( s e e Table 44). The variations in the magnetic properties of the trivalent ions also have a periodic nature ( s e e F i g u r e 130). On the other hand, t h e r e a r e some properties that vary gradually within the s e r i e s . Thus, the atomic and ionic radii d e c r e a s e continuously a s the atomic number is increased ( s e e Table 43). This phenomenon, which is known a s the "lanthanide contraction'' is attributed to the gradual de c r e a s e in the basicity of the elements in the transition from cerium to

259

lutecium and is the cause of the different solubilities of the Ianthanides and the stability of their complex compounds.
TABLE 44
~~

_-

Color of the trivalent lanthanide ions


~

__

Atomic number 57 58 59 60

Number of electrons in the f orbital of the trivalent iron

__

__
4f'andf

Y and 4
La colorles

-f'and 41:

4fZand4f:

4f3and 41'

41' and 41
.

U6and 41'
-.

41'

Ce colorles

Pr greenyellow
Nd redviolet Pm pink

61

62
63
64

Sm yellow

Eu )ale-pink
Gd olorless Tb pale-pink Dy palegreenyellow

65
66

67

Ho dark tellow

68
69
70

Er
pink Tu palegreen Yb olorless

71

Lu :olorless

__

. .

-~

FIGURE 129. Valencies of the lathanide elements.

260

Atomic number FIGURE 130. Magnetic moments of the trivalent lanthanide ions.

Brief historical note on the discovery of the lanthanides 11, 101


T h e discovery of the lanthanides has a complicated history. T h e mixture of lanthanide oxides ("earths") isolated from minerals was originally believed to be a single element. First discovered was the earth yttria. by the Finnish chemist Gadolin in 1794; i t was found in the vicinity of Ytterby, Sweden, in a mineral which was subsequently named gadolinite. A few years later (in 1803) Klaproth and Berzelius simultaneously isolated a new "cerite earth" from the "heavy stone bastnlsite". For a long t i m e yttria and ceria were believed to be the same earth and to belong t o the same element. It was only in 1839 that Mosander isolated cerium oxide, lanthanum oxide, and "didymium" ( a mixture of neodymium and praseodymium oxides) from ceria. Four years later h e also isolated terbium and erbium oxides from yttria. The subsequent separation of the oxides and the discovery of new elements was accelerated by the discovery of spectroscopic analysis. In 1878 Marignac isolated ytterbium from erbia, and in 1879 Cleve found erbium, thulium and holmium in erbia. In the same year Lecoq d e Boisbaudran prepared samarium from ceria. and in 1880 Marignac isolated gadolinium. In 1813 Mendeleev was the first to use the fractional crystallization of the double ammonium nitrates for the separation of lanthanum and "didymium". Using the same method, von Welsbach showed in 1885 that "didymium" is a mixture of neodymium and praseodymium. Dysprosium, europium, and lutecium were discovered later. Only the element with the atomic number Z = 6 1 remained undiscovered in 1901. As we know now. this c a n be attributed to the fact that it has no stable isotopes. Element 61 was prepared only in 1947 (by Marinsky and Glendenin) from the uranium fission products in a reactor, and was named T h e isolated promethium isotope with a mass number of 147 has a half-life by them p r o m e t h i u m of about 3.7 years. Although the discovery of the lanthanides was completed towards the beginning of the twentieth century. many of them were not isolated in pure form and they were not thoroughly studied. New, more effective methods for the separation of the lanthanides have been developed within the last 15 years. By now a l l the lathanides have been prepared both as pure compounds and pure metals.

Physical properties 1 7 , 10,121 The lanthanides a r e silvery-white metals. Some of them a r e paleyellow (e. g., praseodymium and neodymium). They have close-packed hexagonal o r face-centered cubic crystal lattices, with the exception of samarium (rhombohedral) and europium (body-centered) (Table 45). La, Ce, Pr, Nd, Sm, Gd, Tb, and Yb have m o r e than one allotropic modifica tion.

261

P
Cn cn

TABLE 45
Physical properties of the lanthanides. yttrium and scandium*

a-La y-Ce

'

57 58

, 6.162
6.768

920t5

3470

.
99.5 6.27 56.8
E

-3.710 1.871 1.825 1.828 1.821 1.802 2.042 1.802 1.182 3.33 2.84 2.7 3.3 3.2 -2.4 -2.335 -2.2 3915 11.2

8 0 4 t 5 13470
I

a
I

gonal

= 12.159 a = 5.1612
c

3058 110.3 3592 3860 3480 8.9 13.9

Q: N

I
a-Nd a-Sm

60
62
~

7.001

1024t5 1072+ 5 826t 10 1312.h 15 1 3 6 8 t 10 1380 i 20

3210 1670 1430 2830 2480

75.6
SO. 6

6.52

64.3

44

genal (Mg type) The same

7.536

Eu
a-Gd a-Tb

63 64 65

' 5.245
7.886

42.2 80.9 72.0 69.8 9.633 140.5 44,000 Close-packed hexagonal (Mg type) The same

n = 3.6725 c = 11.8354 a = 3.6579 c = 11.7992 a = 8.996 I 23"13' a = 4.5820


( I

-2.246 -2.2

1 8.253

7.136i

56
87

'
,

44 1100 64 166

HO

67 68

'
'

I 8.179 9.062 1 5 0 0 t 25 1 5 2 5 t 25

1 2330
2380 2390

Er

I
1

'

69.0

'

6.72 6.45

= a = c = a =
c

' 78.6

6.66 ' 107

= a = c = a =
c c

3.6360 5.7826 3.6010 5.6936 3.5903 5.6475 3.5773 5.6158 3.5588 5.5874

2.54 3.07
3.09 3.09 3.09 3.12

I
I

-2.2

,
1
I

-2.2
-2.2 -2.2 -2.1 -2.1

5730 5864

15.9

'
;

1.773

1
1.716 1.157

I
I

6433 6850

24.4

74-74 119.5

(0

0
3

1 3

S o m e of the physical properties of the lanthanides a r e shown in Table 45. The melting points of the elements belonging to the cerium subgroup are much lower than
those of the elements of the yttrium sub group. It is noteworthy that the boiling points of S m , Eu, and Yb, which may have a valency of 2+, a r e much lower than those of the other lanthanides. Mention must b e made of the large thermal neutron capture c r o s s sections of gadolinium, samarium, and europium. The p u r e lanthanides are ductile and a r e worked without difficulty (by forging, rolling). The mechanical properties a r e strongly affected by the concentration of impurities, and especially of oxygen, sulfur, nitrogen, and carbon. The tensile strengths and moduli of elasticity of the metals in the yttrium subgroup (except ytterbium) a r e higher than those of the metals of the cerium subgroup (Table 45). All the lanthanides a r e paramagnetic, except Gd, Dy, and Ho, which have f e r r o magnetic properties. cy -Lanthanum be comes superconducting at 4.9K and ,B-La at 585K. The remaining lanthanides do
not become superconducting even i f the temperature is reduced to a fraction of 1K.
Chemical properties

cN

/ 1, 2/

N m

m
c

cd

m
d 3

G
c

o
o

m 3

$ ZN : 8 N m
m
N N

2
m
0 ?
0 F

m - m m

The lanthanides a r e characterized by their high chemical activity. They form very stable oxides, halides, and sulfides, and r e a c t with hydrogen, carbon, carboncontaining gases, nitrogen, phosphorus, and a number of other elements. The metals a r e decomposedby water (slowly in the cold, and m o r e rapidly in hot water) and readily dissolve in hydrochloric, sulfuric, and nitric acids. The lanthanides a r e not attacked by hydrofluoric and phos phoric acids because of the formation of protective films of sparingly soluble salts.
Lanthanum, cerium, praseodymium, and neodymium a r e rapidly attacked by d r y and humid a i r at room temperature, while other metals a r e attacked slowly and retain their metallic luster for a long time. Cerium differs from the other

263

lanthanides in having an oxide C e 2 4 which is easily oxidized to the dioxide C e Q ; a s a r e s u l t cerium and cerium-rich alloys a r e pyrophoric. All lanthanides are rapidly oxidized in a i r a t temperatures above 180 to 200 with the formation of oxides of the type I n z Q (except cerium, praseodymi um, and terbium which form C e Q , Pr60il, and Tb4@).
TABLE 46

The melting and boiling points of lanthanide fluorides and chlorides of the LnxS type, in 'C /21 Fluorides Element melting 1430 1465 1373 1413 1410 1400 1 390 1380 1370 1360 1360 1350 1340 1330 1320 1390 boiling Chlorides melting 855 805 779 778 740 681 626 612 59 1 657 72 I 777 824 857 895 703 boiling 1750 1730 1710 1690 1670 Decomposes

La
Ce

2330 2330
2330 2330 2330

P r Sm
Eu
Tb
Gd

2330
2280 2280 2280 2230 2230 2230 2230 2230 2230 2230

DY E t Tu
Yb

Ho

Lu

1500 1490 Decomposes 1480 1510

The lanthanide o x i d e s a r e characterized by their chemical stability and high melting points. F o r instance, C e Q m e l t s a t about 2500" and Laz% above 2000". C e 2 4 , PrzQ, and Nd24 have hexagonal c r y s t a l lattices, while the oxides of the other lanthanides have cubic lattices. The Ln2Q oxides of La, Ce, Gd, Tb, Dy, Lu, and Y a r e colorless; Tu203 is white with a greenish hue; HozQ is pale-yellow, P r z 4 is greenishyellow, E u 2 4 is pale-pink, E r z 4 is pink, NdzQ is lilac-colored, Tb4011 is dark-brown, PrsOll is black-brown, and C e Q (when f r e e from other lanthanides) is yellow when hot and white when cold. Lanthanides absorb h y d r o g e n even at room temperature and r e a c t rapidly with hydrogen at 250 to 300" with the formation of h y d r i d e s of the type LnHzs8 (for Ce, La, and Pr) o r LnHz. The hydrides decompose when heated in vacuo (to above 1000") and a r e unstable in humid air. N i t r o g e n r e a c t s withlanthanides at 750 to 1000, yielding n i t r i d e s , mostly of the R N type. Their properties have not been thoroughly studied. The lanthanides react with hot carbon, hydrocarbons, Co and COz yielding c a r b i d e s of the LnCz type. The carbides a r e unstable in a i r and decompose water, yielding hydrocarbons (mostly acetylene, and some methane). The sulfides (ZnzS3, Ln3S4 and LnS) a r e formed when the lanthanide metals a r e heated in s u l f u r vapor. The s u I f i d e s have high melting points and a r e fire-resistant. The melting points of some sulfides a r e : La2S - 2100", Ce3S4- 2500, and NdzS3- 2200". At temperatures above 200" all halogens react rapidly with lanthanides. The h a 1 i d e s formed (of the LnX3 type) have relatively high melting and boiling points (Table 46). 264

The halides (except the fluorides) a r e hygroscopic and readily hydrolyze with the formation of oxyhalides LnOX. Lower halides (LnX2) a r e known to exist only in the c a s e s of samarium, europium, and ytterbium.

The properties of lanthanide compounds isolated from solutions a r e basic and are sparingly The l a n t h a n i d e h y d r o x i d e s LXI(OH)~ soluble in water and alkalies. Because of the decreasing basicity i n the transition from cerium to lutecium, t h e r e is a d e c r e a s e in the pH a t which the hydroxides s t a r t to precipitate and in their solubility products (Table 47).
TABLE 47 T h e pH of precipitation and the solubility product of hydroxides of the Ln(OH)s type
~~

Element

,H of precipita ion of nitrate solution

Solubility product of Ln(OH), a t 25"


~~

solution Gd
6.83 6.75
6.40

Solubility product of Ln(OH), a t 25'

La Ce Pr Nd Sm Eu

7.82 7.60 7.35 1.31 6.92 6.82

1.0. 1 0 - l ~

2.1 ' 10-22

1.3.
3.3. lo-"
2.9.
2.5.10-24

1 . 5 . lo-"
2.1.10-20 1.9.10-2' 6.8. lo-''

Er
TU Yb Lu Y

3.4.10-22

6.30 6.30 6.95

The hydroxide of Ce4+is precipitated from solution a t pH 0.9 to 1, and cerium may be separated from the other lanthanides in this way (after oxidation of Ce3+to Ce&).

Q4 r

Temperature, ' C FIGURE 131. Solubilities of some lanthanide sulfates a t various tempera
tures.

( a Ce Pr Nd Pm Sm Eu Ed Tb
FIGURE 132.

Py

fio Er Tm Yb Lu

Solubility of lanthanide oxalates in water a t 25'C.

265

FIGURE 133. Absorption spectra of solutions of chlorides of trivalent lanthanides,

266

The c h l o r i d e s , s u l f a t e s , a n d n i t r a t e s of the trivalentlanthanides a r e soluble i n water and in most c a s e s crystallize as crystalline hydrates of variabie composition. The solubility of the sulfates Lnz(SO43. 8Hz0 d e c r e a s e s sharply a s the temperature is increased from 0 to looo (Figure 131). The f l u o r i d e s and o x a l a t e s a r e sparingly soluble in water and dilute inorganic acids. The fluorides a r e precipitated either a s the crystalline hydrates LnF, -0.5HzO (La, Ce, etc. ) o r anhydrous s a l t s (e. g., of Pr and Nd). The most common composition of the oxalates is Lnz(CZO,), 1OHzO. The oxalates of the lanthanides of the yttrium subgroup a r e m o r e soluble in water than a r e the oxalates of the cerium subgroup (Figure 132). When heated to 500 to 600" the oxalates decompose yielding oxides of the LnzQ type. The lanthanide p h o s p h a t e s , c a r b o n a t e s , and f e r r i c y a n i d e s a r e also sparingly soluble i n water. Most of the simple lanthanide s a l t s have a tendency towards the forma tion of double o r complex s a l t s with ammonium and alkali metal s a l t s and with a number of s a l t s of bivalent elements. The most important of these a r e : the d o u b 1e n i t r a t e s of the lanthanides with ammonium nitrate Ln(N03),. 2NH4N03* 4Hzo and magnesium nitrate 3Mg(N03)z. 24H20, whose solubilities increase in the transition from lanthanum to gadolinium; the double lanthanide potassium ( o r sodium) sulfates (e. g., Lnz(S04)3.MeS04.nH20 w here n is 1 o r 2). F r o m the standpoint of their solubility in water, double lanthanide sulfates can be subdivided into three groups: sparingly soluble (La, Ce, P r , Nd, Sm), of medium solubility (Eu, Gd, Tb, Dy) and readily soluble (Ho, E r , Tu, Yb, Lu, u). This is extensively used f o r the advance separation of the lanthanides into the cerium and yttrium subgroups. The lanthanides f o r m complex compounds with many organic substances. The most important a r e the complexes with citric acid and with a number of aminopolyacetic acids: n i t r i l o t r i a c e t i c acid (NTA), e t h y l e n e d i a m i n e t e t r a a c e t i c acid (EDTA) and some other so-called "complexones" ( s e e p. 356). In most cases the stability of the organic acid complexes of the lanthanides increases from La to Lu, which is the principle of some methods of separation of the lanthanides. The lanthanide compounds a r e characterized by the fact that their absorption s p e c t r a a r e composed of s h a r p bands and lines. The rslative intensities of the bands in the wavelength range f5om 2000 to 7000 A a r e shown in Figure 133. In the visible (3900 to 7700 A ), the Ce3+, Gd", La3', Lu3+, and Yb3+ ions a r e colorless. The remaining lanthanide ions display characteristic colors.

Uses of r a r e - e a r t h metals f 4f The lanthanides are used a s metals, alloys, and chemical compounds in various branches of technology: f e r r o u s and nonferrous metallurgy, production of g l a s s and ceramics, the chemical industry, medicine, and agriculture. They s t i l l have other potential uses, the number of which i n c r e a s e s a s the properties of lanthanides, their alloys, and compounds become known.

267

F e r r o u s and nonferrous metallurgy. The lanthanides find extensive use a s additives in the production of steel, pig iron, and nonferrous metal alloys. They a r e used mainly as f e r r o c e r i u m o r mischmetall ( a lanthanide alloy in which the main components a r e cerium or c e r i u m and lanthanum). Lanthanide oxides a r e also used a s additiviesl51. The addition of lanthanides improves the quality of s t a i n l e s s and highspeed steels, high-silicon s t e e l s used in electrical technology, and re f r a c t o r y s t e e l s ; they i n c r e a s e the mechanical strength (and especially the impact strength), corrosion resistance, and r e f r a c t o r i n e s s , improve the workability of the steels, the surface of castings, and i n c r e a s e the recrystallization (grain growth) temperature of steel. F r o m 0.9 to 2.25 kg of mischmetall is usually added p e r ton of steel. The addition of cerium or mischmetall causes deoxidation, desulfurization, and probably denitridi zation of the steel. The lanthanide m e t a l s also improve the quality of pig iron (casting properties, hot ductility, resistance to oxidation, and strength). The addition of 0.3570 mischmetall to nichrome prolongs i t s s e r v i c e life a t 1000" by a factor of 10. F e r r o u s alloys with a high concentration of cerium group metals (70 to 7570lanthanides and 25 to 3070Fe) a r e pyrophoric and find extensive use in the production of flints ( f o r lighters) and a r e used in a r t i l l e r y for t r a c e r compositions. Mischmetalls or a mischmetall alloy with tin and magnesium may also be used for s i m i l a r purposes. The addition of lanthanide metals to various aluminum and magnesium alloys i n c r e a s e s their high-temperature strength. Aluminum-copper and aluminum-copper-silicon alloys containing 0.05 to 0.3570 cerium a r e used for the production of various a i r c r a f t engine parts. The introduction of mischmetall i n c r e a s e s the c r e e p strength of the high-strength fine-grained magnesium alloys (with zinc and zirconium) used in aviation. These magnesium alloys contain 0.5 to 4 % Zn, 0.6 to 0.770 Zr, and 1.25 to 2.7570 lanthanides. The production of glass and ceramics. The g l a s s industry is one of the m a j o r consumers of lanthanides. The introduction of lanthanides in the composition of glass produces g l a s s e s which absorb ultraviolet and infrared light. Glass containing 2 to 4% C e 2 Q is used for the production of protective goggles for g l a s s blowing and welding works. Praseodymium and neodymium h a v e a s i m i l a r effect. Cerium-containing g l a s s is r e s i s t a n t (does not fog) under the effect of radioactive radiation. It is' used in nuclear technology. Some lanthanide oxides a r e used in the production of optical glass. Thus, La203is one of the components of silicon-free g l a s s f o r photo graphic l e n s e s and periscopes. NdzO3together with V20, a r e added to optical g l a s s for photometers, Nicol p r i s m s and other instruments. Lanthanide oxides have been used for a long time a s decolorizers and colorants of glass. Thus, the addition of a small amount of C e Q de colorizes g l a s s while the addition of up to 1% of cerium oxide makes the g l a s s yellow while the addition of l a r g e r amounts makes it brown. Neodymium oxide colors the glass bright red, praseodymium oxide green, and a mixture of both produces a blue color.

268

Lanthanide oxides a r e extensively used in the optical industry a s abrasives in the polishing of glass. Cerium dioxide is used preferentially f o r this purpose (it is known a s "Polirit") in the f o r m of powders of various particle s i z e s which a r e obtained by varying ignition temperature. As polishing m a t e r i a l C e q is s u p e r i o r to the conventional polishing m a t e r i a l ("rouge") from the standpoints of polishing speed and quality. In c e r a m i c industry lanthanide oxides a r e used a s colorants and opacifiers f o r porcelain, glazes, and enamels, The u s e of lanthanide sulfides and oxysulfides a s r e f r a c t o r y and i n e r t substances f o r the production of metal-melting crucibles is contemplated. Nuclear technology. Lanthanides with a l a r g e thermal neutron capture c r o s s section - gadolinium, samarium, and europium (see Table 45) a r e of utmost importance f o r nuclear technology. The oxides of these m e t a l s a r e components of c e r a m i c shields used in atomic r e a c t o r s and nuclear engines. Lanthanum s a l t s are used a s coprecipitants ("carriers") f o r the p r e cipitation and separation of the transuranium elements formed in atomic r e a c t o r s (e. g . , f o r the separation of plutonium f r o m uranium, and isolation of neptunium). Electronics, X-ray and radio technology. In electronics lanthanides a r e used a s components of nonsputtering gas a b s o r b e r s (getters). Thus, one of the widely-used gas a b s o r b e r s of the T s E T O type consists of a finely-divided and sintered mixture of mischmetall with aluminum and thorium ( 8 0 % Th, 14.570mischmetall and 5.570 Al). Thorium actively absorbs hydrogen, but a s an absorber of nitrogen it is l e s s active than cerium and lanthanum. Neodymium oxide is used in electronic devices as a dielectric with a low linear expansion coefficient. Thulium has recently found a very important use; upon irradiation with r e a c t o r neutrons it becomes a source of y-radiation. The radioactive isotope (TmI7') formed a s a result of the irradiation is used for the production of portable s o u r c e s of soft X-rays for medical use and defecto scopy to replace the bulky X-ray equipment. Such instruments consume only 0.1 to 0.2 g of thulium oxide over a s e r v i c e life of one y e a r (the halflife of Tm17' i s 129days). The activity of thulium preparations may be regenerated by repeating the irradiation. The promethium isotope Pm14' (half-life 2.7years) i s used for the production of nuclear "microcells" in which the soft 0-radiation of the promethium is converted into electrical energy. Such microcells may be used in miniature receivers, hearing aids, and devices for rockets, artificial satellites, etc. E l e c t r i c illumination technology. In this industry lanthanide fluorides have been in use f o r a long time f o r the production of carbon electrodes f o r searchlights and movie p r o j e c t o r s ; the lanthanides i n c r e a s e the illumination intensity. The fluorides (mainly CeF3) a r e introduced a s components of the central p a r t of the electrode (the "core"). Chemical and light industry. Lanthanide compounds a r e used in the production of lacquers, paints, and phosphorescent compositions (phos phors) a s catalysts for the synthesis of ammonia and oxidation p r o c e s s e s in organic chemistry and a s chemical reagents in analytical chemistry and photochemistry.

269

Various lanthanide compounds a r e used f o r coloring and tanning of leather, and in the textile industry as mordants and water-proofing agents. Agriculture. Lanthanide compounds have been recently used in agriculture a s insectofungicides (pest control agents) and as microferti l i z e r s to accelerate plant growth.

72.

OCCURRENCE / 1 4 /

The natural r e s o u r c e s of lanthanides a r e quite large. Their total content in the E a r t h ' s c r u s t is O . O l % , i. e., the same a s the content of copper. The most common lanthanides a r e La, Ce, and Nd. Lanthanides with odd atomic numbers a r e l e s s abundant than their immediate neighbors with even atomic numbers ( s e e Figure 134).

FIGURE 134. Abundance of lanthanides in the Earth's crust, according to Fersman (thick line), Gol'dshmidt (thin line), and Vinogradov (dotted line).

More than 250 lanthanide-containing minerals a r e known. Of these, 60 to 65 a r e lanthanide minerals, i. e., minerals in which the total con centration of the lanthanide elements is higher than 5 to 8%. These minerals consist of phosphates, fluorides o r fluocarbonates, silicates and silicotitanates, niobotantalates and titanoniobates. The minerals usually contain some thorium and occasionally some uranium (Table 48).

270

TABLE 48 T h e composition of some lanthanide minerals


-

Mineral Monazite

Formula ( C e, L a . .

Approximate composition
~~

Density. g/ c m 3

.
4.9

. )PO4

5 0 -60% (Ce, La.. .),Os,22-31.5% P205, 4- 12% Tho,. 0.1 - 0.5% U, occasionally Z r O , (up to 7%) and SiO, (up to 6 % ) 5 2 - 6 2 . 6 9 YzOs contaminated with C e, Er, and other lanthanides, occasionally Tho,. UO, (up t o 5 % ). ZrO, (up to 3%). SnO,. SiO, (up t o 9%1 7 3 - 7 7 k ( C e. La, Pr),03, 19.8- 20.2% co, 6.2-8.5PF.

-5.5

Xenotime

YP04

4.45-4.59

Bastnasite Gadolinite

(Ce, La.. .)FC03 YZFeBe2SizOlo

4.8 - 5.2

10-13.770 FeO, 30.7-46.5% Y z 0 3 , 5--23%(Ce, La.. . ) , 0 3 2 , 3-24.5% SO,. 0 . 3 - 0. 4 ~0 Th 0 , . 9-10.2% BeO, impurities: C a , Mg
X

4.1-4.5

Orthite (allanite

Ca. Ce.. .),.(Al, Fe), Si3O,,(O, OH)

Up to 6 % C e2 0 s , up to 7 % (La, Ce),03, occasionally Be0 (up to 3.8%). up to 8% Y,03 (yttroorthite), impurity: Tho, 39.2-40%Ti02. 32-3470 ( C e, La.. ),03,8-1070 (Nb, Ta),O, 4.2-5.270 CaO, 7.8-9%Na20, impurities: Sr. K.Si, Th (0.5-0.6%)

1.1

Loparite

(Na, C a s C e. . . ), ( T i , Nb. Ta),Os

t.5-4.89

Eu xen ite- polycrase*

(Y. C e . C a. . . ) x ( T i . Nb. Ta),Os

18.2-27.7%(Y. Er.. .),Os, 0.2-4.370 ( C e , La.. . ) , 0 3 1 , 6 -3 0 % Ti 0 , , 4.3-41.470 Nb,Os, 1.3- 23% TaZO,, 1-570 Tho,. 0.4- 12% UOt
$6 -57.5%(Nb. Ta ), 0 5 , 31.42% Yz03. 0.9-6%(Ce. La.. .),03, up to 14%Erz03, 1-3.4% T h o , . 1.2 6% UO,, up to 670 TiO, , impurities: ZrO, , SnO,, W 0 3

k.78-5.37

Fergusonite

( Y. E r . C e, U) X (Nb. T a . Ti)O,

5.58-6.23

Samarskite

f , Er.. .),.(Nb, Ta)60,

6.4-14.570 Y2 0 s , 2.7-13.470 Er,O3, up t o 8 % ( C e , L a . . . ) , 0 3 , 21.1 46.8% Nb,O,, 1.8-27.0%Ta,0s, 0 . 4 - 2 % T h 0 2 , 4-167oUOz, i m purities: T i , Zr, Sn. and others
~~

5.6-5.8

T h e mineral is known as polycrase when the (Nb, Ta),05 : TiO, ratio is 1 : (4-6) euxenite when the ratio is higher than 1 : 3.

or lower, and as

There a r e wide fluctuations in the respective proportions of the individual lanthanides in minerals. Some minerals contain mostly the elements of the cerium group and only up to 570of the yttrium e a r t h s (e. g., monazite, bastnasite and loparite), while in others the elements of the yttrium group a r e predominant (e. g., xenotime, euxenite, and gadolinite).

271

The lanthanides accumulate in various types of magmatogenic, sedimen t a r y and metamorphogenic deposits*. Until the present the lanthanide industry has been based mainly on the exploitation of monazite p l a c e r s and deposits containing the mineral b a s t n a s i t e (bastnasite-calcite veins). One of the main s o u r c e s of elements of the cerium subgroup is m o n a z i t e , which is usually found in pegmatites, and occasionally in granites and gneisses. A s the bedrock is decomposed, the monazite p a s s e s into r i v e r and submarine p l a c e r s together with ilmenite, zircon, magnetite, and other minerals. The lowest monazite concentration in the worked p l a c e r s is about 1%. The l a r g e s t deposits have been found in India, Brazil, the USA, Australia, Madagascar, and Ceylon. Monazite concentrates containing 58 to 65% Ln203 a r e prepared by gravitational and magnetic beneficiation methods. F r o m these, the lanthanides a r e isolated a s by-products in the extraction of thorium. The approximate composition of monazite concentrates i s shown below (in 70) :
~(Ln)z03+Y,0s
c+03

La,03 PrzOs

Sm,03

. . . Ndz03 .

. . . . . . 10-12 . . . . . . . . . 1-1.5

. . . .

. . . .

......

. . . . 58-65 . . . . . .21-30 . . . . . . 1 6 18
3-4

poz . . . . . . . . . 3.5-6.5 " - . . . . . . . . .0.154.3 PZOS . . . . . . . . . 26-30 Fe,03 . . . . . . . . 0.5-0.8


U9Ua

Other

(Ln),03+Y,03

1-2

TiO, SiO,'

. . . . . . . . .

.........

0.2-1.7 1-2

L o p a r i t e is one of the minerals of commercial value; i t h a s a complex composition and occurs in nepheline syenties and in many pegmatite veins. Loparite o r e s a r e easily enriched yielding concentrates which contain 80 to 90% of the mineral. In the processing of these o r e s the lanthanides a r e extracted together with niobium, tantalum, and titanium. Fergusonite and xenotite a r e among the minerals used a s a rich raw material f o r the extraction of lanthanides of the yttrium group.

73.

PROCESSING OF MONAZITE CONCENTRATES

Monazite i s the main raw material f o r the production of lanthanides and thorium. The processing of monazite yields two types of products: thorium concentrates and a technical grade mixture of lanthanide compounds. Two methods for the decomposition of monazite concentrates a r e used in industrial practice: 1) decomposition with sulfuric acid; 2) deccmposition with sodium hydroxide solutions.

---------------

Meramorphogenic deposits are deposits which have undergone changes (outside the wearhering zone) after their formation; these changes cause, in particular. a partial or total recyrstallization of the rocks with the formation of new structures and minerals.

2 72

The sulfuric acid method 110, 151 This method of processing monazite is based on the decomposition of the concentrate by concentrated sulfuric acid followed by the leaching of lanthanide sulfates and thorium sulfate with water. The thorium and lanthanide compounds a r e then isolated separately from the solution. Decomposition. The crushed monazite concentrate (particle s i z e 0.15 to 0.1 mm) is decomposed with concentrated sulfuric acid at 180 to 200". The consumption of sulfuric acid depends on the composition of the concentrate and is of the o r d e r of 1.5 to 2 tons/ton of concentrate, which is 2.5 to 3 times m o r e than the stoichiometric amount. It must be remembered that p a r t of the sulfuric acid is lost through evaporation. Depending on the particle s i z e of the concentrate the decomposition time ranges from 2 to 4 hours. The decomposition is c a r r i e d out in batch r e a c t o r s (made of s t e e l o r c a s t iron) fitted with a s t i r r e r o r in r o t a r y drum p r o c e s s e r s with automatic feeding of the sulfuric acid and the concentrate and with continuous discharge of the product. Measures must be taken for the quantitative absorption of sulfuric acid vapors by means of s c r u b b e r s o r electrostatic f i l t e r s . In batch reactors, the concentrate i s poured gradually, with s t i r r i n g , into sulfuric acid preheated to 200". At the end of the decomposition, the product i s a paste-like m a s s with a dark-gray color. The m a t e r i a l r e maining after a m o r e thorough separation of sulfuric acid (e. g., when the decomposition is c a r r i e d out in continuous drum-type p r o c e s s e r s ) i s light g r a y o r almost white. A higher degree of dehydration of silicic acid i s obtained in such c a s e s , and this facilitates the filtration of solutions in subsequent stages. The main reactions involved in the decomposition a r e :

2 W-4 PO, 3H2S04 (Lnh (SO,), 2H3P0,, Th3 (PO,), 6HzSOa * 31% (so,), 4H3PO4.

+ +

The titanium minerals - ilmenite and rutile - a r e also decomposed by sulfuric acid, yielding titanium sulfates. After the completion of the decomposition the resulting m a s s i s leached in lead-lined steel r e a c t o r s with cold water at a ratio of about 1 0 1 of water p e r kg of concentrate to be decomposed. Such dilution i s necessary in o r d e r to ensure the dissolution of the thorium and lanthanide sulfates. Towards the end of the leaching the temperature of the solution must not exceed 2 0 to 25" since the solubility of lanthanide sulfates d e c r e a s e s sharply with increasing temperature ( s e e Figure 1 3 1). The insoluble residue contains silica, zircon, c a s s i t e r i t e and some unreacted monazite. In addition, i t usually contains a l a r g e fraction of the radioactive daughter products of thorium and uranium. The solid residue is separated from the solution by settling and filtration and is then rinsed with cold water. The filtrate obtained contains the lanthanides, thorium, phosphoric acid, excess sulfuric acid, and contaminants such a s titanium, iron, etc. The filtration of the solution is made difficult because of the presence of colloidal silicic acid. A s a result, in some plants the pulp i s filtered through a layer of sawdust. The filtrate usually contains one of the thorium decomposition products - mesothorium (a radium isotope). It i s isolated from the solution by the

273

addition of barium chloride. The barium sulfate formed s e r v e s as a c a r r i e r for the coprecipitation of the isomorphous radium sulfate. Isolation of thorium and lanthanides from the eulfate solutions. The various technological methods of processing monazite concentrates with sulfuric acid differ in the methods of isolation of the thorium concentrate and the lanthanide concentrate from the sulfate solutions.

Dilute NH40\

7
Sulfate solution Dilution and neutralization to pH 1.0

Settling, decantation, filtration, and rinsing

Precipitate: ThP,O,, a fraction of the lanthanides (thorium concentrate)

Filtrate Neutralization t o pH 2.3

To the production of pure thorium compounds

s Precipitate:

Setrling, decantarion, filtration


.

~-

Filtrate T / N H 4 0 H Neutralization to pH 6.0

. i

the major fraction lanthanides of the

T o the production of pure lanthanide compounds

Precipitate : uranium. the residual lanthanides (uranium concentrate 1%

Solution

To the extraction of uranium


FIGURE 135. Flow sheet of the processing of sulfate solutions by the stepwise neutralization method.

Discarded

The most widely used method is based on a stepwise neutralization of the solutions with the precipitation of thorium and lanthanide phosphates, taking advantage of the different pH of precipitation of thorium phosphate and lanthanide phosphates: thorium phosphate is precipitated from sulfate solutions at pH 1.0 while the lanthanide phosphates a r e precipitated at pH =: 2 . 3 . A flow sheet of the process is shown in Figure 135.

274

I
The neutralization is usually c a r r i e d out with dilute solutions of ammonium hydroxide which a r e gradually added into the s t i r r e d sulfate solution. The thorium is separated by neutralizing the solution to pH 1 and heating to a boil. About 99% of the thorium is precipitated a s the sparingly soluble thorium pyrophosphate ThP20, -2HzO. Towards the end of the neutralization the total concentration of lanthanides in the solution m u s t not exceed 2'3'0in o r d e r to keep the sparingly soluble double lanthanide ammonium sulfates in solution. Nevertheless, 5 to 870 of the lanthanides (in % of the total amount present in the solution) is copreci pitated with the thorium. Since the starting solution contains a much l a r g e r amount of lanthanides than of thorium, the precipitate usually contains about 5070of thorium phosphate and the r e s t is lanthanide sulfates (and possibly phosphates) which a r e sorbed by the volumhous precipitate. After the separation of the precipitate of thorium phosphate the solution is neutralized with ammonia t o pH 2.3. This causes precipitation of the major fraction of the lanthanides, probably a s acid phosphates of the Lnz(HPO,), type. After settling, the precipitate is separated by filtration and used f o r the production of pure lanthanide compounds ( s e e below). The filtrate which contains certain amounts of lanthanides and uranium is neutralized with ammonia to pH 6.0, which causes precipitation of uranyl hydroxide together with the lanthanides remaining in the solution. The precipitate contains up to 1%U and s e r v e s a s the concentrate used for the extraction of uranium. The composition of the precipitates and the distribution of thorium, lanthanides and uranium in the fractions from the stepwise neutralization of sulfate solutions is shown in Table 49 a s an illustrative example.
TABLE 49 Composition of precipitates and distribution of thorium. lanthanides, and uranium in the fractions from the stepwise neutralization o f sulfate solutions / 1 6 1 pH of xecipita tion :omposition of precipitates. converted into oxide mixture. 70
.. -

Product

Extraction into the given fraction,% of the initial amount

Tho, Washed thorium concentrate Washing solution Precipitate (lanthanide phosphates) Precipitate enriched in uranium
1.2 1.2 2.3 6.0
68.2

LnzO,
31.0

US08
0.8 0.5 3.3

99.1

0.2

99.3 96.7

The alkaline method

/ 10, 15/

A recent method based on the decomposition of monazite concentrates with alkali solutions is now being introduced in industrial practice. The main reaction involved in the p r o c e s s i s : & I $ ) PO,

+ 3NaOH

--f

Ln (OH),

+ Na,PO.,.

A sufficiently high degree of decomposition of the monazite is obtained by using a finely divided concentrate (96.570 with a particle sizeof 0.044mm),

275

a 4570 solution of NaOH at a ratio of 1.5 kg of NaOH p e r kg of monazite (i. e., 30070 e x c e s s over the stoichiometric amount) and by carrying out the decomposition f o r 3 hours a t 140". Increasing the processing tempera ture to 200" r e s u l t s in virtually quantitative decomposition but the hydroxide precipitates formed are sparingly soluble in acids, probably because of t h e i r partial dehydration. The main disadvantage of the u s e of sodium hydroxide f o r the decomposition of monazite is associated with the consumption of l a r g e amounts of the reagent. Soviet scientists were able to effect a considerable reduction in the consumption of sodium hydroxide (to 15070of the stoichiometric amount) and to accelerate the p r o c e s s by using a ball m i l l heated to 130" as the reactor. The abrasive effect of the balls a c c e l e r a t e s the reaction since i t destroys the hydroxide layer formed on the g r a i n s of the mineral. Since the p r o c e s s combines decomposition with grinding, i t does not require the u s e of a finely divided concentrate.

Grinding
1

Decomposition Wash waters

I
Dilution and mixing

.1
Filtration and washing Precipitate Lanthanide and Th hydroxides HCI Y Dissolution Filtrate Na3P04,NaOH

Evaporation and crvstallization Na3P0, 1 2 H 2 0 crystals By-product

7 Solution
NaOH

J-

J.
Solution NaOH (Na,PO, )

--r-

7
Dilution and neutralization, pH 5.8 Precipitate Th(OH),(U, lanthanides) (thorium concentrate)

1 .1
Solujion NaOH
7 J Lanthanide hy
Precipitation droxides

FIGURE 196.

Flow sheet of the alkaline method for the processing of monazite

276

The flow sheet of the alkaline method for the processing of monazite is shown in Figure 136. The concentrate is decomposed with alkali in steel reactors. The resulting s l u r r y is discharged into a tank in which it is diluted with the wash waters from subsequent stages (to a NaOH concentration of 300/0). In o r d e r to prevent the crystallization of sodium phosphate, the s l u r r y is heated to 100 to llOo, allowed to stand f o r one hour at that temperature (to obtain precipitates which are easy to separate by filtration), and is filtered while still hot. The hydrated cake is washed with water until the PzQ content is reduced to 0.470 o r less. The filtrate containing the Na3P04 and the e x c e s s NaOH is evaporated in an evaporator and the residue is taken f o r the crystallization of trisodium phosphate (Na3P0412H20)which is a valuable by-product*. Since the solubility of Na3P04 in NaOH solutions is v e r y low ( a t 20" the solubility of Na3P04 in 36% NaOH is only 1.3%), the m a j o r fraction of the sodium phosphate may be separated from the solution. The NaOH solution remaining from this stage is returned f o r the decomposition of the concentrate. The hydroxide precipitate is dissolved in concentrated hydrochloric acid, which is taken in about 125%excess with respect to the stoichiometric amount needed for the reactions:

+ 3HC1 + h C 1 3 + 3Hzo; Th(OH)4 + 4HCl -+ ThC14 + 4HzO.


Jh(OH), After heating f o r one hour to 80" the solution is diluted with water. The insoluble residue (quartz, zircon, etc. ) is separated by filtration. The resulting chloride solution is diluted and neutralized with sodium hydroxide to pH 5.8. This leads to the virtually quantitative precipitation of the thorium a s thorium hydroxide. Most (99.3% ) of the uranium and about 370 of the lanthanides in the solution a r e coprecipitated with the thorium. The thorium concentrate is used f o r the preparation of pure thorium compounds and f o r the extraction of the uranium in it. The approximate composition of the thorium concentrate is shown below inyo :
Th ............. 36.4 Lanthanides 1.45 U ............... 0.14 Fe ............... 2.21 T i ................ 6.13

.......

Si
P

.........

C1 Residue, insoluble in acids . . . . 23

.........

.........

4.41 0.44 0.36

After the precipitation of the thorium, a mixture of lanthanide hydroxides is precipitated from the filtrate by the addition of alkali. The dried lanthanide hydroxide precipitate h a s the following approximate composition:

............ 1 3 ................... 0.05 ................... 0.005 ................... 0.02 Si ................... 0.4 P .................... 0 . 1 c1 .................... 1.9
Lanthanides Th U Fe

--------------* Sodium phosphate is used in the production of fertilizers and in some other industries.
211

As compared with the decomposition of monazite by sulfuric acid, the alkaline method has the advantage of resulting in the separation of phosphorus from the thorium and the lanthanides in the f i r s t stage of the process, and yielding a useful by-product - sodium phosphate. The disadvantages of the alkaline method include a lower degree of decomposi tion (even if the concentrate is very finely divided) and the consumption of large amounts of sodium hydroxide.

74.

SEPARATION OF LANTHANIDES / l , 2,7,8,10-13/

Because of the similarity in their properties, the separation of the lanthanides is a difficult problem. E a r l i e r separation methods were based mainly on the differences in the solubilities of lanthanide compounds, The individual elements, in varying degrees of purity, were prepared a s a r e s u 1t of a l a r g e n u m b e r (sometimes several thousands of successive) f r a c t i o n a l c r y s t a l l i za t i o n s o r f r a c t i o n a1 p r e c i p i t a t i o n s The separation of some of the lanthanides was based on their oxidation to the tetravalent state (Ce, Pr, Tb) o r reduction to the divalent state (Sm, Eu, Yb). In such a case the separation was facilitated because of the great difference between the properties of tetravalent o r divalent lanthanides on one hand and the trivalent lanthanides on the other. Recently the separation of the lanthanides was greatly simplified and im provedbytheuseof i o n e x c h a n g e a n d e x t r a c t i o n ( w i t h o r g a n i c s o 1v e n t s ) m e t h o d s These new methods do not entirely supersede the fractional crystallization and precipitation methods. Fractionation methods a r e , however, now only used in the initial separation stages which produce fractions enriched in individual lanthanides. We shall briefly discuss the most common separation methods.

Fractional crystallization Two of the widely used crystallization methods a r e the fractional crystallization of the double lanthanide ammonium nitrates o r lanthanide magnesium nitrates (in the separation of the elements of the cerium group) and the crystallization of the bromates L n B r Q 9HzO (for the separation of the elements of the yttrium group). The relative solubilities of the double nitrates of the elements of the cerium group a r e shown in Table 50. The solubility of the double nitrates increases from lanthanum to samarium. When the cerium is separated in advance by the oxidation method (see below), a s e r i e s of fractional crystallizations of the double nitrates results in the concentration of lanthanum in the sparingly soluble fraction, of neodymium, samarium and europium in the soluble fraction, and of part of the lanthanum, praseodymium and p a r t of the neodymium in the intermediate fractions. The fractional crystallization is usually carried out in the sequence shown in Figure 137. The starting solution is evaporated and allowed to cool, yielding crystals and mother liquor which is separated by decantation.

278

The c r y s t a l s a r e dissolved in a f r e s h portion of the solvent, the solution is evaporated and the crystallization is repeated. The mother liquor from the f i r s t crystallization is subjected to a second evaporation and crystallization. The mother liquor from the second crystallization is then combined with the c r y s t a l s formed after the second evaporation of the starting solution. This yields three fractions, each of which is then subjected to crystallization in a similar manner and the mother liquors and crystals from successive crystallizations a r e combined.

Starting solution

so1

Intermediate fractions
FIGURE 137. Schematic presentation of the fractional crystallization process.

dissolution in water; I\ evaporation until forma tion of crystals; o mother liquor; a crystals; Y combined.

Each new operation increases the number of fractions in the crystalliza tion s e r i e s . When a given number of fractions is attained (usually 8 to 20), that number of fractions is subsequently maintained constant by combining the precipitates o r mother liquors of the end fractions from two subsequent crystallization s e r i e s o r withdrawing them from the process a s the end product.

TABLE 50 Relative solubilties of the double lanthanide ammonium and lanthanide magnesium nitrates (the solubility of the lanthanum salt is taken as unity)
-~

Salt Ln(NOs)~*2N~NO~.4HzO 2Ln(NC&. 3Mg(NOs),. 2 4 H z 0

1 . 2

1 . 2

1.5

3.8

_ _ ~ -

- .._____._ .-

___

279

The fractional crystallization of bromates of the elements of the yttrium group is c a r r i e d out in a s i m i l a r manner.

Fractional precTpitation The f r a c t i o n a l p r e c i p i t a t i o n o f t h e d o u b l e s u l f a t e s and hydroxides is m o s t frequently used. The precipitation of the double sulfates (Ln)2(S0J3 Na2S04-nHz0is extensively used f o r the separation of the lanthanides into cerium and yttrium groups and occasionally into t h r e e groups: 1) c e r i u m group (La, Ce, Pr, Nd, Sm) - sparingly soluble double sulfates; 2) terbium g.roup (Eu, Gd, Tb, Dy) - double sulfates of intermediate solubility; 3) yttrium group (Ho, E r , Tu, Yb, Lu, Y) - soluble double sulfates. In most c a s e s the u s e of the method is confined t o a rough separation into the c e r i u m and yttrium (including terbium) groups. The precipitation is c a r r i e d out from cold sulfuric o r nitric acid solutions containing sodium sulfate which is added either a s a solid or a s a saturated solution until the disappearance of the neodymium lines from the absorption spectrum of the sample. The separation of lanthanides by fractional precipitation of their hydrox ides is based on the different pH a t which the precipitation o c c u r s ( i t d e c r e a s e s f r o m lanthanum to lutecium, s e e Table 47). The solution containing the lanthanides is gradually neutralized by the addition of a dilute solution of ammonium hydroxide or by passing through the solution a s t r e a m of ammonia mixed with a i r (the Trombe method). The fractions precipitated at the different pH values a r e separated; a fraction enriched with samarium, europium, and the yttrium e a r t h s is obtained a t a low pH, followed by the intermediate fractions (Nc? + Pr contaminated with La), and finally a r i c h lanthanum concentrate (98 to 99% La) is produced at higher ( > 8.6) pH values. The fractional precipitations of oxalates (for the cerium group elements) and ferrocyanides (for the yttrium group) a r e also used for separation.

Selective oxidation The oxidation of cerium to the tetravalent s t a t e is the most commonly used method for its separation from the other lanthanides. Cerium i s readily oxidized by oxygen on drying a mixture of lanthanide hydroxides in a i r at 120 to 130 o r when a i r is passed through a hot suspension of the hydroxides. Other oxidizng agents such a s chlorine o r hydrogen peroxide a r e also used. After the oxidation of Ce to Ce4+the trivalent lanthanide hydroxides a r e dissolved in dilute (5 to 1 O q o ) nitric or hydro chloric acid while c e r i c hydroxide (CeOz nHzO) r e m a i n s in the insoluble residue. The residue contains 94 to 9 6 % C e q . In o r d e r to purify the cerium concentrate residue (i. e., to remove the other lanthanides), it is dissolved in concentrated nitric acid. P u r e cerium compounds a r e

280

isolated from the solution by crystallization a s c e r i u m ammonium hexa n i t r a t e (NHJ2Ce(NO& o r by precipitation as the basic c e r i u m nitrate which is formed when the solution is diluted and neutralized to pH z 0.7 to 1. Another method which is extensively used f o r the purification of cerium is the selective extraction of Ce(NO.Jr f r o m 6 t o 8 N n i t r i c acid by diethyl e t h e r o r tributyl phosphate (see below).

Selective reduction Reduction to the divalent s t a t e is used f o r the isolation of samarium, europium, and ytterbium from fractions enriched in these elements. The properties of the divalent samarium, europium, and ytterbium resemble the p r o p e r t i e s of the. alkaline-earth elements and in particular those of strontium and barium. Thus, in contrast to the sulfates of the trivalent lanthanides, t h e i r sulfates a r e sparingly soluble. The commonly used reducing agent is sodium amalgam which is mixed with an acetate solution of the lanthanide mixture to be separated:

Ln (CH,COO),

+ 2Na (Hg) =Ln(Hg) + 3CH,COONa.

The reduced elements a r e extracted from the amalgam by hydrochloric acid. Samarium and europium a r e separated by selective reduction of Eu3+ to Eu" with zinc amalgam o r zinc dust, which do not reduce Sm3+to Sm2+. This can be attributed to the different values of the oxidation-reduction potentials of s a m a r i u m and europium. The oxidation- reduction potentials of the lanthanides (with respect to hydrogen electrode) a r e (volts) :
La = La3f

~ m ~ f = ~ m ~ +. + .e .
Yb2f= Yb3++2 Eu2f = Eu3+ Pr3f = P I ' +

Y = Y3f Pr = Pr3f

+ 3; . . . . . . . . . . +3 9 . . . . . . . . . . . +3 e . . . . . . . . . . .
...... ..........

-2.37 -2. I -2.0


-1.72

+e . . . . . . . . . . -0.43 +e . . . . . . . . . . +1.60 ~ e 3 f= ~ e ' f +e . . . . . . . . . . + I .61

-I.

15

The hydrochloric acid solution containing Sm3+andEu3+is passed through a column packed withgranulated zinc amalgam. The solution is collected in dilute sulfuric acid when europium sulfate EuS04 is precipitated. Selective reduction of europium ( a s a means f o r i t s separation from s a m a r i u m and the other lanthanides) may also be accomplished by electrolysis using a m e r c u r y cathode. The electrolysis is c a r r i e d out in acetate solution containing potassium citrate. The potassium c i t r a t e and the lanthanides f o r m complexes of K3[Ln(C,QH7)2]type which a r e so stable that no precipitation of lanthanide hydroxides o c c u r s when the solutions are alkalized to pH 7 t o 9. (i. e., the pH a t which the electrolysis is c a r r i e d out). Adiagram of the electrolytic cell used for the reduction is shown in Figure 138. After the separation of europium the amalgam is washed with

28 1

water (to remove potassium, a p a r t of which is also separated) and europium is then extracted with hydrochloric acid. Another variant of the separation of samarium f r o m europium consists f the divalent sulfates o f these in the selective oxidation of a mixture o elements with n i t r i c acid: samarium is m o r e readily oxidized and is extracted into solution.

FIGURE 138. Diagram of an electrolytic c e l l for the reduction of europium on a mercury cathode. 1- glass vessel; 2- electrolyte; 3- pla tinum leaf anode; 4- mercury cathode; 5- stirrer; 6- terminals.

Separation with ion exchange r e s i n s Ion exchangers a r e sparingly soluble solids (organic o r inorganic) with active (ionogenic) groups which can be exchanged f o r ions present in solution. The most important ion-exchanging m a t e r i a l s a r e the synthetic organic r e s i n s which a r e extensively used, in particular, in extractive metallurgy of nonferrous and r a r e metals (for the extraction of ions from dilute solutions) and f o r the purification and separation of elements with similar properties. The modern synthetic ion exchangers (organic resins) a r e very large, interconnected molecular chains (cross-linked polymers) ; they form an insoluble three-dimensional lattice which swells in water. Active groups, which a r e capable of undergoing electrolytic dissociation, a r e "suspended" on the polymer chains. Ion exchangers a r e subdivided into two groups. C a t i o n exchange r e s i n s consist of an insoluble acid anion of high molecular weight and contain active groups (mostly S Q H o r COOH) which can exchange a hydrogen ion for a cation present in the solution. A n i o n exchange r e s i n s consist of an insoluble cation (base) of high molecular weight and labile, exchangeable anions. The most common active groups in anion exchange r e s i n s a r e the amino groups (-NHz, = NH, r N ) .

282

Cation and anion exchange r e s i n s a r e produced commercially in the f o r m of granulated m a t e r i a l s of various particle sizes. Like the conventional chemical reactions, the ion exchange process, obeys the Law of Mass Action and is characterized by the ion exchange constant. Thus, the reaction on cation exchange r e s i n s can be expressed by the equation:

RMi

+ Ma Z R 4 + M i ,

where MI and M Z a r e the cations to be exchanged, and R is the insoluble anion in the cation exchange resin. F o r a reaction involving exchange of cations of the s a m e valency, the ion exchange equilibrium may be expressed a s follows, in accordance with the Law of Mass Action:

where CI and C2 a r e the concentrations of M Iand M2 in the solution, N l and N2 a r e the respective concentrations on the resin, and K is the ion ex change constant. The equation f o r the exchange of ions of different valencies assumes the form:

where ZI and Z2 a r e the valencies of the ions. Cation and anion exchange r e s i n s may be "charged" with different cations and anions. It is thus accepted practice to speak about cation exchange r e s i n s in the H', Na+, NG+-forms, e t c . , and pf anion exchange r e s i n s in the S G , OH-, C1- and other forms. The ion exchange may be c a r r i e d out under static o r dynamic conditions. In the static method all the solution containing the ion to be exchanged is brought into contact (by stirring) with a certain amount of the resin, and the resin is then separated from the solution by filtration. The method is not very effective, especially in the case of low ion exchange constants when a l a r g e number of repeated operations is required. In the dynamic method (which is the one most frequently used) the solution is passed through a column packed with the resin. In such a case the resin is first saturated completely in the upper l a y e r s and subsequently in the lower layers. Ion exchange r e s i n s a r e characterized by their sorption capacity (for a given ion under certain defined conditions). The sorption capacity is expressed either in gram-equivalents p e r g r a m of r e s i n o r in percent of the weight of the resin, The sorption capacity of the resin depends on a number of factors: the value of the ion exchange constant, the filtration rate, the concentration and nature of the ions to b e exchanged, the grain size of the resin, the pH of the solution. When operating under dynamic conditions distinction is made between: a) t h ? working sorption capacity, i. e . , the amount of substance absorbed by the r e s i n until the "breakthrough" (the appearance of the absorbed ion in the filtrate); b) the total sorption capacity, i. e., the amount of substance absorbed before complete saturation of the resin

283

takes place. The r e s i n capacity is usually determined f r o m the efflux c u r v e s which show the dependence of the concentration of the exchanged ion in the filtrate a s a function of the volume of the solution passed ( F i g u r e 139). The volume of f i l t r a t e p r i o r to the "breakthrough" (section 'la") is a m e a s u r e of the working sorpti'on capacity of the resin. The a r e a SIcorresponds to the total amount of the ion absorbed in the r e s i n p r i o r t o the "break through", while the total area bounded by the discharge curve is a m e a s u r e of the total capacity of the resin. The ion exchange in cludes the following stages: 1) sorption of ion f r o m solution; the Volume of solution passed through filtration of the solution through the r e s i n the column bed is continued either to "breakthrough" FIGURE 139. Efflux curve for or to complete saturation of the r e s i n ; sorption on resin under dynamic 2) washing the r e s i n with water; conditions. 3) elution of the sorbed ion and regenera a-volume of solution passed prior tion of resin. to the "breakthrough"; s,-area In most c a s e s the elution is c a r r i e d out numerically equal t o the working with solutions containing ions with which the sorption capacity of the resin (prior r e s i n must be "charged" f o r the sorption to the "breakthrough"); C/Cp cycle. The elution curve has the shape ratio of the ion concentration in shown in Figure 140. During the elution the filtrate to its initial concentra tion. the concentration of the ion in the filtrate f i r s t i n c r e a s e s rapidly - " and then d e c r e a s e s just a s rapidly. When the sorption is c a r r i e d out f r o m solutions with a low concentration of the ion to be extracted, the sorption-elution cycle yields solutions in which the concentration of the ion is 1 0 to 100 times higher than in the starting solution.

FIGURE 140. Typical elution curve.

FIGURE 141. Elution curves for two ions ( M I and M 2 ) sorbed on a resin.

The different ions differ in their affinity f o r the resin, which depends mainly on the ionic charge and the s i z e of the hydrated ion. This is the principle of the separation of ions by the ion exchange method. The separation of ions of elements with s i m i l a r properties (e. g . , zirconium and hafnium, tantalum and niobium, lanthanide elements) is

284

best c a r r i e d out by the so-called ion exchange chromatography*. The mixture of ions to be separated is f i r s t sorbed in the upper p a r t of the r e s i n charge in a column. The ions a r e then eluted from the saturated layer. Depending on their affinity f o r the resin, the ions in the mixtulre a r e separated during the elution into individual zones which move along the column at definite r a t e s . The zone containing the ions with the lowest affinity for the r e s i n will have the highest r a t e of motion. The separation of ions MI and M2 (if the affinity of M Ifor the r e s i n is higher than that of M2)would yield an elution curve such a s that shown in Figure 141, in which each of the two curves corresponds to a sorption band containing one of the ions. At the exit of the column the filtrate fractions containing MI and M2ions a r e collected separately. Theoretically, each zone should contain the ions of only one of the elements to be separated. In practice, how ever, the zones a r e superimposed (the extremities of the elution curves often spread out) which makes quantitative separation difficult. The separation of ions with s i m i l a r proper ties may be made m o r e efficient by using eluent solutions containing some complex-forming agent which binds the ions to be separated into complex compounds of various degrees of stability. ron exchange p r o c e s s e s a r e c a r r i e d out in cylindrical columns charged with the granulated resin. A draining system (grid) which allows the solution which has passed through the r e s i n bed to be tapped without entrainment of resin particles is fitted to the lower p a r t of the FIGURE 142. Diagram of an filter (see Figure 142). installation for the separation Various types of cation exchange r e s i n s a r e used of elements by ion exchange. f o r the separation of lanthanides by ion exchange 1-column; 2-grid; 3-ion (the resin used in the USSR is KU-2, in the exchange resin; 4-upper USA Dowex 50 and others). These r e s i n s a r e layer of the resin, with the strongly acid cation exchange r e s i n s produced mixture of elements to be separated sorbed on it; 5 by copolymerization of styrene and divinyl pressurized vessel with the benzene and containing S03H as the active solution containing the elements group. to be separated; 6-vessel with The affinity of the lanthanides f o r the resin the eluent; I-collection vessel decreases from La3+ to Lu3+,i. e . , a s the size for the filtrate. of the hydrated ions is reduced. However, the affinities of the various lanthanides f o r the r e s i n differ only very slightly and the separation is not entirely satisfac tory. Better separation is obtained by eluting with solutions containing

* T h e chromatographic method was developed by the Russian botanist Tsvet in 1903, who intended t o use
i t for the separation of plant pigments. T h e term "chromatographic adsorption" was proposed by Tsvet
on account of the colored adsorption bands which were produced on the adsorbent during the separation.

285

organic compounds which form complexes of different degrees of stability with the lanthanides. The o r d e r of the elution corresponds to the relative strength of the anionic lanthanide complexes. As the eluent moves along the column (or a number of columns connected in s e r i e s ) the mixture of cations is separated into sorption zones (bands) which move at certain definite velocities. Various organic compounds which form lanthanide complexes a r e used f o r the elution: c i t r i c acid, nitrilotriacetic acid (NTA), and ethylenediaminetetraacetic acid (EDTA). EDTA is the eluent most widely used in the chromatographic separation of the lanthanides. It is an a-amino acid. EDTA is a tetrabasic acid with two nitrogen atoms:
HOOC -CH2 CH,

-COOH

\ N-CH*-CH~-N/ HOOC - CH,


/

CH2 - COOH.

Trivalent lanthanides form c h e 1a t e complexes with EDTA in which the nitrogen atoms a r e bound to the lanthanide ions by coordination bonds:
CH2COOH

CH2COO-

The stability constant of these complexes is given by the equation: (HEDTA)3- + Ln3+e H(LnEDTA) . [H(LnEDTA) ] K = [Ln3'] [HEDTA3'-] As is evident from Figure 143, the stability constants of the lanthanide complexes with EDTA increase from La to Lu, i. e . , with the atomic number of the element. The stability constants of the complexes of two adjacent lanthanides differ on the average by a factor of 2.4. This ex plains the high selectivity of EDTA as an eluent of the lanthanide cations from the resin. The lanthanides a r e separated by using a system of several columns charged with the resin and connected in series. The first few columns in the system s e r v e for the saturation of the r e s i n with the lanthanides During elution individual while the remaining columns a r e s e p a r a t o r y sorption zones a r e formed in these columns. In the saturation columns the resin is f i r s t charged with NHZ or Na+ ions (i. e. , the resin is used in the N&+ or Na' form). The r e s i n i n the s e paratory columns is in the Cu2+ form. The Cu2+ ions s e r v e a s r e t a r d e r s

286

which improve the separation. The copper complex with EDTA is m o r e stable than the complexes of most of the lanthanides with EDTA ( s e e Figure 143). Thus, when the eluent containing the complex lanthanideEDTA ions is passed through the separatory column, the copper ions p a s s into solution and d i s p I a c e the lanthanide ions from the soIution into the resin, thus retarding their motion. The lanthanide ions a r e displaced from the solution on to the r e s i n in an o r d e r corresponding to the strength of their complexes with EDTA, which r e s u l t s in the formation of better defined sorption zones. Usually the pH of the eluent is adjusted to 4 to 8.5 (depending on the operating conditions) by neutralization with ammonium hydroxide. Thus, the eluent contains the ammonium salt of EDTA. A dilute solution of EDTA (0.5 to 1%) is used a s the eluent since the ammonium s a l t s of the lanthanide-EDTA complex acid (NH4( LnEDTA) nHzO) and the copper complex Cuz(EDTA) 5Hz0 a r e not very soluble. The use of m o r e concentrated solutions causes the precipitation of s a l t s which clog the p o r e s between the r e s i n grains and thus interfer with the normal operation. The low EDTA concentration in the eluent limits the output of the separation p r o c e s s since the lanthanide concentration in the filtrate is low.

la&hMn&EuMn&2b&+fin4 Elements

T i m e , hours FIGURE 144. Elution curves of terbium. europium, and samarium sorbed on the resin by a 0.26 M so lution of EDTA, a t pH = 3.62.

FIGURE 143. Dependence of log K on the atomic number of t h e lantha nides ( K is the stability constant of the lanthanide-EDTA complex).

The filtrate fractions a r e collected separately at the issue from the column system. The first filtrate contains only copper ions. Fractions containing lanthanide ions appear only l a t e r a s the copper is eluted from the system; the o r d e r of elution of the lanthanides i s determined by the

287

stability of their EDTA complexes. Figure 144 gives the elution curves (with EDTA solutions) for the separation of three elements. The f i r s t wave corresponds to the elution of terbium, the second to europium, and the third to samarium, in accordance with the decreasing complex strength in sequence Tb-Eu-Sm. The EDTA solutions a r e regenerated by acidification to pH 0.5 to 1. This causes the decomposition of the lanthanide complexes and the p r e cipitation of the sparingly soluble EDTA. If the initial specific charge on the column* with the lanthanide mixture is small ( l e s s than 5% on the weight of the resin) and the total length of the separatory columns is l a r g e enough, each filtrate fraction at the i s s u e from the l a s t column contains only one of the elements to be separated. The chromatographic method may be used to prepare individual lanthanides of a high degree of purity, but the yields a r e low. Accordingly, in most cases the elements a r e preliminarily separated by some other method and the ion exchange method is used only for the separation and purification of like elements.

Extraction The separation of the lanthanides by extraction is based on the different distribution coefficients of the lanthanides between their aqueous solutions and organic solvents ( s e e Chapter 3, p. 122). The extractant most widely used for the separation of the lanthanides is TBP, which has been described above. Lanthanide nitrates and TBP 3TBP type. form complexes of the I J I ( N O ~ ) ~

Atomic number

FIGURE 145. Effect of the atomic number


on the distribution coefficients of lanthanide

nitrates between aqueous phase and TBP (HNO, concentration -15.6 M)

_________----- * T h e specific charge is the ratio of the weight of the


resin.

ions sorbed on the resin to the total weight of the

288

When lanthanide n i t r a t e s a r e extracted by T B P from solutions in which the HN03 concentration is above 5 M the distribution coefficients steadily i n c r e a s e with the atomic number of the element (Figure 145). The separation coefficients

=
=Z

f o r two adjacent elements depend on the

nitric acid concentration*. Thus, f o r Z-values between 57 (lanthanum) and 6 4 (gadolinium) the separation coefficients a r e 1.9 and 1.6 f o r solutions with nitric acid concentrations of 15.6 and 1 2 M respectively. The n i t r i c acid concentration may be reduced without decreasing the value of 0 by adding t o the solution certain n i t r a t e s (Al(NQ)s, LiNQ) which a c t a s salting-out agents, i. e., which displace the lanthanide n i t r a t e s into the organic phase. The separation coefficients for adjacent elements in the yttrium group ( Z = 64-71) a r e lower than in the cerium group. Hence, their separation by extraction is m o r e difficult than the separation of elements in the cerium group. The distribution coefficients i n c r e a s e with increasing concentration of the lanthanide in the solution; this makes i t advantageous to p r o c e s s l a r g e amounts of lanthanides. Two extraction methods a r e used for separation purposes: 1) in countercurrent in column extractors o r a s e r i e s of m i x e r - s e t t l e r extractors, with the aqueous and organic phases moving continuously in countercurrent ( s e e Figure 66) ; 2) in semi-countercurrent in which the aqueous phase is stationary while the organic phase is in motion and the mixture to be separated is added at the beginning to the f i r s t extractor (Figure 146).
Extractant

cP
FIGURE 146.

Diagram of semi-countercurrent extraction.


I, 11,111, IV, V -aqueous fractions.

1 -mixing chamber: 2-settler;

When using the first method the starting mixture of the lanthanides is separated into two fractions. Since the separation involves not two, but a l a r g e number of elements, the p r o c e s s must be repeated many times (i. e . , multistage extraction is required) and the number of separations must be at l e a s t n- 1, where n is the number of elements to be separated. However, the p r o c e s s has the advantage of being completely continuous, which is to be p r e f e r r e d for the separation of l a r g e quantities.

--------------

a Z + l and a Z are the distribution coefficients for elements with atomic numbers Z + l and Z respectively.

289

In semi-countercurrent extraction the starting mixture of elements to be separated is introduced into the f i r s t extractor and the remaining extractors a r e filled with a solution of nitric acid ( o r a solution containing nitric acid and the salting-out agent). The organic phase (TBP) is then fed into the system and p a s s e s successively through all the extractors. The contact between the aqueous and organic phases in each extractar r e s u l t s in the establishment of an equilibrium distribution of the elements between the two phases. Elements with lower atomic numbers concentrate in the f i r s t extractors while elements with higher atomic numbers con centrate in the l a s t extractors. Thus, a single passage of the extractant leads to the separation of the lanthanide mixture into a number of fractions. The technique is convenient when separation of small quantities is involved. Both undiluted TBP and solutions with a high concentrationof lanthanidenitrates (100-20Og/lof Ln203) areviscous. Thus, it is recommended that extractors of the mizer-settler type be used. The extractive separation method has a high output and i t s use is increasingly frequent.
Separation scheme Various separation schemes which combine the methods described above a r e used at present. As an illustrative example, a scheme showing the o r d e r of separation of elements of the cerium group i s given in Figure 147.

ce4+
concentrate-

0:idation Af C e arfd separation

&
To purification Fractional precipitation of th hydroxides or crystallization! of the double nitrates La concentrate

1-1

Fractional crystallization of the double nitrates (repeated a few :imes)

chromatography

Reduction with Zn

FIGURE 141. Sequence of separation of elements of the cerium group (Yt elements of the yttrium group).

290

The mixture taken for the separation usually contains hydroxides and is f r e e of contaminants. The cerium is separated f i r s t , by oxidation ( s e e p. 281). After the separation of the cerium, lanthanum concentrate is isolated from the nitrate (occasionally chloride) solution by fractional concentration of the hydroxides o r fractional crystallization of the double ammonium nitrates. The remaining elements a r e subjected to a rough separation into s e v e r a l fractions by fractional crystallization of the double nitrates. Neodymium and praseodymium a r e then separated by extraction and by the chromatographic method. Samarium and europium a r e isolated and separated from the inter mediate samarium-europium fraction by the reduction method (,see p. 281). The elements of the third fraction a r e separated by reduction, extraction and ion exchange. The above scheme still includes fractional crystallization as a means of rough separation. In m o r e modern schemes extractive methods a r e used for the separation into fractions.

75. CONTROL OF SEPARATION AND PURITY OF LANTHANIDE COMPOUNDS


Because of t h e simiiarity in the c h e m i c a l properties of the lanthanides, the conventional analytical methods (volumetric or gravimetric) cannot b e used to determine the composition of a lanthanide mixture. T h e only exceptions a r e cerium. which c a n b e determined by a number of methods based on its oxidation to the tetravalent state, and europium. which c a n be reduced with relative ease t o the divalent state. Thus, the analysis of the lanthanides is carried out with the aid of physical methods. These include: magnetic methods (measurements of the magnetic susceptibility). spectroscopic methods (absorption, arc, spark. and X-ray spectra), the chromatographic method, and the radiometric method. T h e measurement of t h e magnetic s u s c e p t i b i 1i t y is extensively used for the control of the separation of the elements during fractionation. It has the advantages of being a rapid method and of permitting the use of preparations in any form (solutions, solid substances). T h e lanthanides and their compounds are paramagnetic*. Their interaction with an external magnetic field draws them into the field. T h e magnetic permeability (and the molar magnetic susceptibility which is related to it) is measured with the aid of a magnetic balance. A curve showing the magnetic moments of the lanthanides as a function of their atomic numbers is given in Figure 130. T h e magnetic susceptibilities a r e strictly additive. Thus, the composition of a binary mixture of lanthanides may be determined with an accuracy of 2 - 3 P i f the magnetic susceptibility of each of the components is known. Spectroscopic methods are more accurate and offer more scope. T h e method most widely used for the determination of the degree of separation and purity of the lanthanide elements is a d s o r p t i 0 n s p e c t r o T h e absorption spectra of the lanthanides have been shown in Figure 133. B y using modern photome try photoelectric spectrophotometers. large amounts of most lanthanides can be determined with an accuracy o f f 1 . 5 % . However, the bands are not intense enough for the determination of small amounts of contaminants. S p e c t r o s c o p y i n t h e a r c and especially X - r a y s p e c t r o s c o p i c m e t h o d s are the most sensitive. At present. X-ray spectroscopic analysis is one of the most accurate methods for the determina tion of the composition of a mixture of lanthanides. T h e radiometric method (introduction of different radio-isotopes of lanthanides a t various stages of the process) is extensively employed to follow the separation of the lanthanides. This method is especially effective in the chromatographic separation of the elements. T h e determination of the m e a n a t o m i c w e i g h t may b e used as a simple method for the evaluation of the composition of binary lanthanide mixtures. T h e mean atomic weight is calculated from the oxide to-oxalate ratio. T h e oxide content is determined by igniting a weighed amount of the dry oxalate to constant weight, while the oxalate content is determined by dissolving the salt in sulfuric acid and titrating with permanganate solution. T h e determination of the m e a n a t o m i c weight is especially vaIuable in t h e control of the separation of yttrium-containing systems since its atomic weight differs sharply from t h e a t o m i c weights of the lanthanides.

_------------- * Lanthanum is diamagnetic.


291

76. MANUFACTURE OF RARE-EARTH METALS / 9 , 1 0 , 1 2 /


Because of the high chemical stability of the lanthanide compounds (oxides, halides), the pure metals and their alloys a r e produced by thermal reduction with m e t a l s or by the electrolysis of molten salts.

Starting compounds f o r the manufacture of m e t a l s Metallic lanthanides a r e produced mainly f r o m anhydrous lanthanide chlorides or fluorides. Lanthanide chlorides a r e prepared by the action of various chlorinating agents on lanthanide oxides; chlorine (in the presence of carbon), carbon tetrachloride, sulfur chlorides, ammonium chloride and hydrogen chloride a r e used as the chlorinating agents. The chlorination of oxides by chlorine in the p r e s e n c e of carbon r e s u l t s in the contamination of the chlorides produced with carbon since lanthanide chlorides do not distil under the p r o c e s s conditions used (600 to 800") and contain the excess carbon added to the starting mixture. Carbon tetrachloride* r e a c t s with lanthanide oxides at 700 to 800":
2(Ln),O,+3CC1,-+4(Ln)CI3+3C0,

(some CO and C0Cl2 a r e also formed). The chloride produced by the above method contains only a very small amount of carbon. This is explained by the dissociation of some CCll a t high temperatures : 2cc1, c1, + C,CI,, C,CI, Z 3C1, + 2C.

The chlorination with chlorine or Cell vapors is c a r r i e d out most conveniently in shaft furnaces charged with the pelletized m a t e r i a l (the pellets contain a mixture of the oxides with carbon when chlorine is used a s the chlorinating agent, or pelletized oxides when using CClJ. As is evident f r o m Table 46, the chlorides of most lanthanides melt within the range 700 to 850'. Thus, chlorination a t 800 to 850' yields a mixture of the molten chlorides which can be discharged periodically f r o m the furnace. Chlorides of a higher degree of purity a r e produced in the chlorination of lanthanide oxides by sulfur chloride, hydrogen chloride or ammonium chloride: the reactions involved a r e : 2(Ln),03 6S,CI,

(Ln), 0, + 6HC1 2(Ln)CI, + 3H,O;


( Ln),O, 6NH4CI 2 ( Ln)CI, + 6NH3 + 3H,O.

--f

4(Ln)CI,

+ 3 S 0 , + 9s;

--f

The simplest p r o c e s s is the one based on the use of ammonium chloride. In this p r o c e s s a mixture of the lanthanide oxides and NH&1 (in 1000Jexcess over the stoichiometric amount needed) is heated a t 200 to 300" until a sample withdrawn from the reaction mixture proves - - -- -- - -- --- - --

Carbon tetrachloride is a nonflammable liquid boiling a t 18".

292

to be fully soluble in water. The e x c e s s NH&l is then expelled by heating the chloride to 300 to 320 in vacuo (0.5 to 2 mmHg). The elimination of ammonium chloride must be quantitative to prevent contamination of the lanthanides by nitrogen. Since the anhydrous lanthanide chlorides a r e very hygroscopic, they m u s t be t r a n s f e r r e d rapidly f r o m one container to another when hot and out of contact with humid air. It is best to s t o r e the chlorides in an i n e r t gas. Anhydrous lanthanide fluorides may be prepared by dehydration of the hydrates precipitated from solution o r by the action of hydrogen fluoride o r ammonium bifluoride on lanthanide oxides. D e h y d r a t i o n o f f l u o r i d e h y d r a t e s . Lanthanide fluorides a r e sparingly soluble and a r e precipitated quantitatively from chloride, sulfate and nitrate solutions by the addition of hydrofluoric acid. They are precipitated either as the hydrated fluorides of the LnF3 0.5HzO type (La, Ce, and other fluorides) or a s the anhydrous fluorides (Pr and Nd). When a solution containing the fluoride precipitates is heated under an infrared lamp, the hydrated fluorides a r e converted into the anhydrous form. The precipitates a r e then separated by filtration, washed with alcohol, and dehydrated further by heating to 400" in a d r y i n e r t gas atmosphere (argon) under reduced p r e s s u r e (about 100" Hg). The fluorides prepared by the above method a r e still contaminated with the oxyfluorides LnOF. The formation of the oxyfluorides may be prevented by drying in a s t r e a m of hydrogen fluoride. F l u o r i n a t i o n o f o x i d e s . Fluorides of a higher degree of purity, which a r e not contaminated with oxyfluorides, a r e produced by the action of gaseous hydrogen fluoride on the oxides:

(Ln),O,

+ 6HF

2( Ln)F3

+ 3H,O.

A rapid reaction takes place at 550 to 575". The p r o c e s s may be c a r r i e d out in nickel or nickel-copper ( 7 0 % Ni, 30% Cu) tubes inwhichthe boats with the oxides move in a direction opposite to the direction of the gas flow. At 500 to 600" the r a t e of attack of d r y hydrogen fluoride on nickel is very low (about 0.9 "/year) because of the formation of a protective film of nickel fluoride. Graphite boats may be used. Hydrogen fluoride i s fed into the tube from s t e e l tanks. The unreacted gas a t the i s s u e from the tube may be absorbed either in a sodium carbonate solution or in con d e n s e r s cooled with dry i c e (solid C Q ) and returned to the fluorination process. Ammonium bifluoride NH,HFz may be used a s the fluorinating agent instead of hydrogen fluoride. Lanthanide oxides a r e fused with ammonium bifluoride a t 200":

( Ln),OS

+ 6NH,HF,

a( Ln)F, + 3H,O

+ 6NH4F.

The excess ammonium bifluoride and ammonium fluoride a r e then expelled by distillation at 450". The p r o c e s s yields a fluoride with a composition s i m i l a r to that of the fluoride produced by fluorination with gaseous HF. However, ammonium fluoride must be quantitatively r e moved in o r d e r to prevent contamination of the m e t a l s with nitrogen. Since this is r a t h e r difficult, it is best to p r e p a r e lanthanide fluorides by fluorination with hydrogen fluoride.

293

In contrast with lanthanide chlorides, lanthanide fluorides a r e not very hygroscopic. However, they can absorb gases from air. Hence, it is recommended that they be stored in an i n e r t gas atmosphere.

Materials f o r the smelting of the r a r e - e a r t h metals The purity of the metals thus produced depends on the impurity content of the starting m a t e r i a l s and of the crucible material used f o r the smelting of the metals and of the electrodes (when electrolytic methods a r e used). The data in Table 51 show the resistance of various refractory materials and metals to the action of molten lanthanides. O f the oxides, electrically fused magnesium oxide and beryllium oxide possess a satisfactory resistance up to 1200". Tantalum is the most resistant of the refractory metals, and may be used for melting lanthanides at temperatures up to 1700". Molybdenum is also quite resistant, and is frequently used a s the cathode in the electrolytic production of r a r e - e a r t h metals.
TABLE 51 Interaction of lanthanides with various electrode and crucible materials Material

Not attacked up to 1200'. Not attacked up to 1000". Not attacked up to 1250".

Behavior

React with molten lanthanides. Does not react with t h e metals in vacuo or in inert gases a t temperatures up to 1700". Not attacked by molten lanthanide halides. Does not react in vacuo or in inert gases a t up to 1400'. Not attacked by molten halides. Attacked slowly by the molten metals, not attacked by the molten halides. React with the molten metals a t various rates, depending on the temperature. Cast iron may be used to construct electrolytic baths for t h e production of technical grade mischmetall and cerium. Not attacked at low temperatures (800 to goo"), rapidly attacked a t high temperatures. Reacts slowly with molten metals, not attacked by molten halides Rapidly attacked.
~-

Molybdenum Tungsten
Copper, nickel, iron

CaF, (fluorite)
Graphite
Porcelain

Graphite is used both as an electrode and as crucible material; how ever, the metals which a r e produced in contact with graphite a r e always contaminated with carbon since graphite r e a c t s slowly with molten lanthanides. Electrolytic production of r a r e - e a r t h metals 1101 There a r e various methods for the production of lanthanides by the electrolysis of molten salts. Those most widely used a r e based on the

1455

294

electrolysis of anhydrous lanthanide chlorides dissolved in molten alkali or alkaline-earthchlorides. This method is used f o r the production of commercial quantities of mischmetall, cerium, lanthanum, neodymium, and other metals of the cerium group. During the electrolysis the molten metals a r e deposited on the cathode since their melting points a r e relatively low. The electrolytic production of metals of the yttrium group (with the exception of ytterbium) which have high melting points (1350 to 1700") is m o r e difficult. Electrolysis which would deposit the molten metals on the cathode is virtually impossible a t s u c h temperatures, because of the vaporization of the halides and the difficulties associated with the selection of bath and electrode materials. In o r d e r to obtain these metals in a liquid state without raising the bath temperature above 1100" the electrolysis is c a r r i e d out with liquid cadmium or zinc cathodes and the larithanides a r e obtained as an alloy with the cathodic metal. The r a r e - e a r t h metals a r e then separated from zinc or cadmium by vacuum distillation of the latter. This method has been used on a laboratory scale for the production of lanthanum, samarium, gadolinium, europium, and dysprosium.

FIGURE 148. Diagram of an electrolyzer for the production of tech nical cerium or mischmetall ( a n alloy of the lanthanides of the cerium group).
1 --steel body; 2.-graphite cathode; 3-cast-iron body; 4-graphite packing; 5-crushed chamotte packing; 6-current lead to the cathode; 7-graphite crucible; 8-graphite anode; 9-cast-iron ring; 10-molten electrolyte: 11-molten metal; 12-pivots for rotating the electrolyzer; 13-cast-iron crucible; 14-chamotte lining.

In most c a s e s the electrolytes used a r e based on an eutectic mixture containing 50% KCZ + 50% CaC12 (mp 660") o r a mixture containing 50% KCl + 5070 NaCl (mp 658'). Lanthanide chlorides a r e quite soluble in these

295

molten mixtures. The bath contains about 58 to 6 0 % LnCls and the r e s t consists of the alkali and alkaline-earth chlorides. The addition of small amounts of calcium fluoride is r e commended. A n electrolyzer used in Germany for the production of technical grade misch metal1 / l o / is shown in Figure 148. It consists of a graphite crucible (serving a s the cathode) on a grapliite bottom plate placed in a cast-iron body. The body is fitted with pivots which permit the bath to be rotated. The electrical current is supplied to the body, which is in electrical contact with the graphite plate, by means of a packed layer of graphite and pitch. The graphite crucible is fastened to the bottom plate with the aid of a cast-iron ring. The cast-iron body together with the crucible FIGURE 149. Diagram of an electrolyzer for a r e mounted within an iron jacket. The the production of pure rare-earth metals. space between the jacket and the crucible 1-furnace; 9-graphite crucible (anode); is packed with crushed chamotte and the 3-beryllium oxide crucible; 4-molten bottom is lined with chamotte bricks. cerium; 5-molybdenum cathode; 6 The bath volume is 30 l i t e r s . A lifting electrolyte; I -nickel crucible; 8-tubular graphite anode 100" in diameter is quartz shield; 9-asbestos gasket; 10-water cooled copper tube connecting t h e anode to the fitted in the center of the bath. At f i r s t the electrolyte is fused by current; 1 1 --graphite cover. passing current through the graphite rod placed between the electrodes. The rod is then removed and the bathismaintained in a molten state by the heat evolved in the passage of the current through the melt. Additional batches of the charge a r e periodically fed into the electrolyte and the process is continued until the crucible has been filled. The concentration of oxy chlorides LnOCl (which a r e formed by the interaction of lanthanide chlorides with moisture from the air) in the lanthanide chlorides taken for electrolysis must be a s low a s possible. The oxychlorides a r e not de composed by the current, but their presence in the melt causes partial precipitation of the metal on the cathode in a finely dispersed state ( a s a I' mist"). The fine metallic particles float to the surface of the bath and a r e oxidized if the electrolysis is c a r r i e d out in the presence of air. When the finely dispersed metal particles reach the anode they a r e chlorinated, yielding chlorides which dissolve in the melt. The appearance of these finely dispersed particles r e s u l t s in a sharp decrease in the current efficiency. The deposition of the metal in a finely dispersed state is also observed a t very high o r very low current densities. The electrolysis is c a r r i e d out at a potential of 12 to 15 V and a current of 2000 to 2200 amp (for a 30 l i t e r crucible) which corresponds to a cathodic current density of 3 amp/cm2. The current efficiency is about 6 0 to 70%. The bath temperature is 900 to 1100". During the electrolysis, chlorine is evolved at the anode. Hence, the bath must have a satisfactory exhaust system. After the crucible has been filled the bath is tilted and i t s

296

contents a r e poured into s t e e l pans preheated to 500 to 550. During the pouring the metal is protected against oxidation by the molten salt. The mischmetall or cerium produced by the electrolysis contains 94 to 99% lanthanide elements and a s e r i e s of impurities: carbon, calcium, aluminum, up to 1% silicon, 1 t o 2.5% iron, etc. A metal of a higher degree of purity may be produced by using electrodes made of metals which do not r e a c t with lanthanides (molybdenum and especially tantalum), by using pure mangesium and beryllium oxides f o r the lining of the crucible, and by conducting the electrolysis in an i n e r t gas atmosphere. Cerium of a higher degree of purity has been produced on a laboratory s c a l e by the electrolysis of its chloride using a molybdenum cathode, by A diagram of keeping the metal out of d i r e c t contact with the graphite. the electrolyzer used is shown in Figure 149. A graphite crucible serving a s the anode is protected against oxidation by placing it inside a nickel crucible, and a molybdenum rod is used a s the cathode. To prevent oxidation, the upper p a r t of the rod (i. e . , the p a r t outside the melt) is enclosed in a quartz tube. The molten metal is collected in a beryllium oxide crucible placed beneath the cathode on the bottom of the graphite crucible, The cerium produced under these 99.770 pure. The main contaminants a r e : F e 0.02 to conditions is 0.04/0, Si 0.06 to 0.670, Mo < O.Ol/o, Mg O.Ol%, Ni 0.4 to 0.670, A1 0.1 to 0 . 6 0 / , Be 0.2 to 0.970, Ca 0.05 to 0.2570. Electrolyzers of this type have been used f o r the production of relatively pure m e t a l s of the cerium group, a s well a s the higher-melting metals of the yttrium group. In the case of the yttrium group metals, molten zinc or cadmium (contained in a c e r a m i c crucible on the bottom of the graphite crucible) s e r v e s a s the cathode. The c u r r e n t is supplied to the liquid cathode by means of a molybdenum rod insulated with a porcelain tube (to prevent contact with the electrolyte). Zinc o r cadmium a r e easily removed from the cathodic alloy by vacuum distillation a t 900.

Metallothermic preparation of the lanthanides

/ 9, 10, 1 2 1

Metallothermic reduction of lanthanide halides (chlorides, fluorides) or oxides produces lanthanide metals of a higher degree of purity than the lanthanides produced by the electrolysis of molten s a l t s . The data in Table 52 (which contains values of the heats and f r e e energies of formation of lanthanide halides and the common metallic reducing agents) lead to the conclusion that the chlorides may be reduced by calcium o r sodium, and the fluorides by calcium. However, when sodium was used to reduce the chlorides, the lanthanide ingots produced could not be easily separated from the slags. The use of magnesium o r aluminum f o r the reduction of the halides yielded alloys of the lanthanide and the reducing agent, and the yield of the lanthanides (as alloys) was not high enough. Magnesium can be separated from the lanthanides by vacuum distillation above the melting point of the lanthanides; aluminum cannot, however, be quantitatively separated by this method. The best r e s u l t s as r e g a r d s yield, quality of the ingot, and purity of the m e t a l were obtained by using calcium to reduce the halides. This method may be used f o r the manufacture of all the lanthanides except

297

samarium, europium, and ytterbium, whose compounds a r e reduced only to the lower halides. A method was accordingly developed f o r the produc tion of these three lanthanides by reduction of their oxides with lanthanum and a simultaneous vacuum sublimation of the products.
TABLE 52
Heats and free energies of formation of lanthanide halides, k cal / g - at o m of chlorine and fluorine

Element

I-

Chlorides

Fluorides

A H2980 K

Pr ..

Nd Pm Sm

EU
Gd Tb

2 Er
TU Yb LU
Na ca

-86 24.7 -84 -82.7 -77.7 -81.7 -70.3 -78.7 -77.7 -77.3 -76.3 -71.3 -76 -98.3 -95.3

-82.3 -81 -80.3


-74

AH2989 H

AF2980

AFI 000'

- 69.7 - 68.7 - 68.3 - 87

-140.3 -138.7 -134.3 -136.7 -136 -13.5 -130.3 -134.7 -133.3 -132.7 -131.7 -130.7 -130.3 -125.3 -130.7 -136 -145.1 -131.5 -107.7

-134 -132.3 -131.3 -130.3 -129.7 -128.7 -124.3 -128.7 -127.3 -126.7 -125.7 -124.7 -124.3 -119.3 -124.7 -129 -139.1 -126 -102

-120 -118.3 -117.7 -116.3 -115.7 -114.7 -110.3 -114.7 -113.3 -112.7 -112 -111 -111 -106 -111.3 -112.6 -124.8 -112.8 89.7

Reduction of halides by calcium The reduction of the halides by calcium must be c a r r i e d out at tempera t u r e s above the melting point of the lanthanides in o r d e r to produce a metallic ingot. The slag must also be melted. It follows that the relatively low-melting metals such a s La, Ce, Pr, and Nd (mp in the range 800 to 1050") must be reduced under conditions differing f r o m those used f o r the reduction of metals of the yttrium group, whose melting points l i e between 1350 and 1650" (see Table 45). Reduction of chlorides. The low-melting lanthanides (La, Ce, Pr, Nd) may be prepared by the reduction of their chlorides or fluorides with calcium. Metals of a high degree of purity a r e produced by reducing anhydrous lanthanide chlorides in small hermetically sealed steel con tainers (bombs) lined with pure magnesium oxide o r with a dolomite mixture of calcium and magnesium oxides (Figure 150). The maximum reduction temperature is 1100". At that temperature MgO does not r e a c t with the lanthanides. However, partial reduction of the MgO (by the molten lanthanides) takes place at higher temperatures and the magnesium is dissolved in the melt. High-purity calcium (purified by vacuum distillation) a s grains with a particle s i z e of 0.6 to 1m m is mixed with the lanthanide chloride in a special chamber in a d r y argon at mosphere. The mixing in a d r y argon atmosphere is necessary toprevent absorption of moisture by the chloride and of nitrogen by the calcium.

298

__

..

-..

. ....

- ,

I . ,

..

The calcium is added to the mixture in a 15 to 2070 excess over the stoichiometric amount needed f o r the reaction: 2LnC13+3Ca = 2Ln + 3CaC12. When the process is c a r r i e d out on a s m a l l scale, the reaction heat does not suffice to m e l t the produced metal C a O packing and the slags. To increase the heat of reaction, iodine is added to the charge Fired refractory ( 0 . 3 to 0.7 moles iodine/mole chloride) lining together with an additional amount of calcium needed to form Car2*. Moreover, the introduction of iodine reduces the melting. point of the slag by the formation of a eutectic mixture of CaClz and CaI2. The hermetically sealed vessel containing the charge is heated to 700" in o r d e r to initiate the reaction. The metals a r e produced in C a O packing an average yield of 95'3'0, in the form of FIGURE 150. Reaction vessel (bomb) used dense ingots which a r e easily separated for the thermal reduction of lanthanide from the slag. The metals contain chlorides by calcium. about 270 calcium which is separated by remelting of the metals in vacuo, in magnesium oxide or beryllium oxide crucibles. Magnesium oxide crucibles cannot be used if the reduction is to be effected at temperatures above 1200" (e. g . , in the case of gadolinium). Tantalum is the most suitable reactor m a t e r i a l since i t does not r e a c t with most of the lanthanides below 1500 to 1600". The reduction is c a r r i e d out in a welded tantalum crucible covered with a perforated lid, in an argon atmosphere. High-frequency heating is employed (no heating additives a r e added to the charge). The lanthanides adhere strongly to the tantalum and the tantalum shell must be removed by mechanical means. In o r d e r to save tantalum, the crucibles a r e made of thin (0.02 to 0.06") foil. Reduction of the fluorides. Lanthanides with high melting points (Tb, Dy, Ho, Er, Tu, Lu, Y) cannot be produced by the reduction of their chlorides. The main difficulty i s the high vapor p r e s s u r e s of the lanthanide chlorides a t the temperatures (1500 to 1600') required for the production of ingots. The boiling points of lanthanide fluorides a r e higher than those of the chlorides (see Table 46). Moreover, a s compared with the chlorides, the fluorides have the advantage of being nonhygroscopic. Thus, their handling is m o r e convenient. According to American literature, the reduction of the fluorides by calcium is c a r r i e d out in tantalum crucibles in an argon atmosphere. When mixing calcium with lanthanide fluorides, c a r e is taken to reduce a s much a s possible the contact with the atmosphere. Calcium is added to the charge in a 10% excess over the stoichiometric amount needed f o r the reaction: 3Ca = Z n + 3CaF,.

The charge is packed in a tantalum crucible (which is degassed in advance by heating in vacuo), covered with a perforated lid and placed in

_______-_____-

T h e heat of formation of CaII is 128.5 kcal/mole.

299

the quartz tube of a vacuum induction furnace. In o r d e r t o degas the charge the crucible is heated slowly in vacuo to 600". At that temperature, pure argon is fed into the apparatus until the p r e s s u r e reaches 500mmHg and heating i s continued to the temperature a t which an active reaction begins between the fluoride and the calcium. Depending on the m e t a l to be produced, that temperature lies between 800 to 1000". The reaction is exothermic but the heat evolved is not sufficient t o r e a c h the required final temperature and the heating must be continued. In o r d e r to obtain a high yield of the m e t a l (in the ingot) the temperature a t the end of the process must be highar than the melting points of the m e t a l and the slag (the melting point of CaFz is 1418"). A temperature of 1450" is sufficient in the c a s e of the low-melting lanthanides and Gd,Tb, and Dy, while tem p e r a t u r e s 50 above the melting point are required f o r lanthanides with higher melting points. After the temperature of vigorous reaction has been reached, the reaction is completed within a few minutes; the maximum temperature is then maintained f o r another 15 minutes in o r d e r to obtain satisfactory separation of the metal from the slag. Under these conditions, the yield of the m e t a l (in the ingot) reaches 97 to 99%. The mixture is cooled and the b r i t t l e s l a g is easily separated from the metal. The main contaminant in the ingot is calcium (0.1 to 2%). To remove the calcium, the ingot is remelted in vacuo, in the s a m e crucible. Using this method, 300g of the metal may be produced in a crucible 50" in diameter and 200" high. The tantalum concentration in the light lanthanides (from La to Nd) is 0.02 to 0.0370, and in the heavy lanthanides 0.1 to 0 . 5 % . The concentra tions of other impurities a r e : Ca 0.0170, N 0 . 0 0 5 ~ 0 ,0 0.03 to O.l%, C 0.0075%, and F 0.005%.

Reduction of oxides with simultaneous distillation of the metal We have mentioned that samarium, europium, and ytterbium cannot be produced by the reduction of their chlorides o r fluorides with calcium. The compounds a r e reduced only to the divalent halides. A method f o r producing these three m e t a l s has been developed; the oxides a r e reduced with lanthanum and the metals formed (which have higher vapor p r e s s u r e s than that of lanthanum, s e e Table 45) a r e distilled a t the s a m e time:
S m , O ,

+ 2La .+ 2Sm t + L a , O , .

In one such p r o c e s s , the reduction is c a r r i e d out in a tall tantalum crucible with an air-cooled copper condenser fitted in i t s upper parts. A mixture of the oxide and lanthanum turnings (taken in a 20% excess) is placed in the tantalum crucible and i t s lower p a r t is heated to 1400" in a vacuum induction furnace. A vacuum of at l e a s t 1 0 - 4 m m H g is maintained during the heating. The beginning of the sublimation is accompanied by a s h a r p drop in the p r e s s u r e (to about lO-'mmHg) since the evaporating

* To prevent rontamination

with copper. the surface of the condenser is coated with a thin layer of the oxide of the m e t a l t o be produced. A suspension of the oxide i n alcohol is used for the coating.

300

... .

. ..

_..

..... ... -....--.

......-.-.--....-.. ....... ... ..

...I

metals actively absorb the residual gases. The metals a r e condensed at 300 to 400 when they a r e obtained in the form of a coarsely crystalline crust. Metallic powders a r e obtained at lower temperatures. Samarium, europium, and ytterbium, produced by the above process, a r e virtually f r e e of tantalum and lanthanum. The concentrations of C, N, 0, and H do not exceed 0.01 %.

301

Part Three
THE SCATTERED METALS

Chapter VII GERMANIUM


77.

GENERAL DATA ON GERMANIUM

On the strength of his periodic system, Mendeleev made in 1 8 7 0 a very accurate prediction of the properties of a hitherto unknown Group IV element "eka-silicon". In 1886, i. e . , 1 5 y e a r s later, Winkler discovered a new element, germanium, in the mineral argyrodite 4 A g $ GeS2. The proper t i e s of germanium coincided with the properties of "eka-silicon" as predicted by Mendeleev. Germaniumbecame of practical interest only during the Second World War as a result of the development of semiconductor electronics. The commercial production of germanium for that branch of technology began in 1945 - 1950.

Properties of germanium Germanium is a light-gray e l e y e n t , with a diamond-type cubic lattice and a lattice p a r a m e t e r a = 5.657 A (Figure 151). Each germanium atom is surrounded by four equidistant atoms positioned on the vertexes of a tetrahedron; the atoms a r e bonded through paired valency electrons. Some physical proDerties of germanium a r e listed below:
Atomic number Atomic weight.. Density, g / cm s : solid (25') liquid (1000') Decrease in volume on melting,% Melting point, " C Boiling point, " C Heat of fusion, kcal/g-atom Heat of vaporization, kcal/g-atom Heat capacity, c a l / g " C a t 0 -300C 9OO'C Heat conductivity a t 25"C, cal / cm . sec. ' C Linear expansion coefficient at: 0 -3OO'C 300-65O'C Mohs hardness Compressibility coefficient (up to 0-12,000 kg/ mm2). cm2/kg Surface tension, dynes/ c m Electrode potential (s. h. e.), V Magnetic susceptibility, CGS units T h er m al neutron capture cross section, barns
I,
I,

............................... 72.6 .............................. 5.323 ............................ 5.557 ................ 5.5 ............................. 958.5 ............................. 2690 .................... 8.3 ............. 84.0 ............ 0.0768 . ............... 0.085 ....... -0.14 .............................. 6.1 .lom6 ............................. 6.6. lo-' ................................. 6- 6.5

................................

32

..........................

...................... 600 ................... -0.15 ............... -0.12. ...... 2.8

1.4.10-'"

lou6

302

. . . . . . . . ..._....

.............

Even very pure germanium is brittle at room temperature, but it be comes ductile a t temperatures above 550O. Like silicon, germanium is a semiconducting material and thus finds increasing use in semiconductor electronics. In this connection, we shall briefly review the most important properties of semiconductors. The concept of semiconductors. The electrical properties of germanium /1 to 4, 12/. F r o m the standpoint of their electrical conductivity, sub stances a r e classified into three groups. Sub stances with a high electrical conductivity ( v ) , in the range io4 to 1 0 ' ohm-'. cm-' (i. e. , metals andalloys) a r e k n o w n a s c o n d u c t o r s . I n s u 1a t o r s a r e substances which conduct virtually no current (e. g. , quartz, mica, asbestos, etc.) ohm-'. and whose y is in the range lo-" to cm Substances with an intermediate conductivity (y = 1 0 ' to lo-'' 0hm-l.. cm-') a r e known a s s e m i c o n d u c t o r s ; this groups comprises a number of elements (silicon, germanium, selenium, tellurium, etc. ) and various compounds (some FIGURE 151. The crystal
oxides, sulfides, intermetallic compounds, etc. ).
line structure of germanium. The differences in the conductivity of these
Each atom is bound to four three groups of substances can be satisfactorily on explained by means of the so-called z o n e theory of a tetrahedron. which is summarized below. It is known that the electrons in atoms occupy certain discrete energy levels (quantum levels). The transition of an electron from one level to another is accompanied by a change in i t s potential energy. The electrons in the outer shell a r e known a s v a l e n c y electrons. Since a solid substance behaves as a single electronic system, the energy levels (zones)of the whole solid may be considered instead of the energy levels of the individual atoms, even though the magnitude of the energy levels of the solid differs from that of the individual atom. Each body has a zone which is completely or partially filled with electrons (known as the f i 11e d o r v a l e n c y z o n e ) a n d a c o n d u c t i v i t y (or e x c i t a t i o n ) zone. Thesetwo zones a r e separated by the f o r b i d d e n zone whose widthdetermines the conductivity of the substance**. As is evident from Figure 152, conductors, semiconductors, and insulators differ in the width of their forbidden zones. In conductors the forbidden zone is practically nonexistent, i. e., all the electrons in the valency zone readily p a s s (with a very small l o s s of energy) into the conductivity zone. In insulators the forbidden zone is wide. The transfer of electrons from the valency zone to the conductivity zone requires the expenditure of such l a r g e amounts of energy that these substances have virtually no electrical conductivity. In semiconductors, the forbidden zone is narrow. The transfer of electrons into the conductivity zone requires a relatively smaIl amount of energy (heating, illumination, o r the f a potential difference). application o

-'.

_______-_-_____
[or "band" theory.] ** It must be noted that the location of an electron at a certain energy level designates its potential energy
rather than its geometrical (physical) position.

303

Two types of conductivity may be distinguished in semiconductors e 1e c t r o n and h o 1e t y p e conductivity*. Both types of conductivity a r e the result of the disturbance (caused by heat, illumination, e t c . ) of the separate valency bonds between the paired electrons and the transfer of electrons into the conductivity zone (Figure 153). When an electrical field is acting on the body, the f r e e electrons move in a certain direction s known as n - t y p e creating an electrical current. Such conductivity i conductivity.

........ .. . . ...... ... .......... .......


a

C
%..*

1 . . . . . . .

'***.**-*'

h q :

m u h

FIGURE 152. Energy zones. a-conductors; b -semiconductors: c-insulators; I-conduc tivity zone; 11-forbidden zone; 111-valency (filled) zone.

However, another type of conductivity may also exist. An empty space (known a s a hole) remains when some external force causes the removal of an electron from the valency zone; this hole may be regarded as a positive charge of a magnitude equal to that of the electron charge. It is possible to have electron t r a n s f e r s (caused, f o r instance, by thermal fluctuations) from the filled lattice points to the adjacent unfilled points (holes). Such t r a n s f e r s create new holes. The electric field produces a directed motion of the holes which is equivalent to the motion of positive charges. Such conductivity is of the hole type ( p - t y p e ) .

FIGURE 153. T h e mechanism of n-type and p-type conductivity in germanium (each double line represents bonds between ger manium atoms consisting of paired electrons).
a-rupture of a n electron bond caused by the transition of an electron into the conductivity zone and the formation of a hole; b- donor atom in rhe l a t t i c e (e. g. , As? in which the fifth electron is weakly bound and passes into the conductivity zone: c-acceptor atom in the l a t t i c e (e. g., I n ' ? to which the missing electron is supplied ar the expense of the adjacent electron bond, in which a hole is left.

_______------- * [Or "n-type" and "p-type"

conductivity, respectively. ]

304

h the ideal c a s e (i. e., semiconductor c r y s t a l s not containing impurities) the number of f r e e electrons equals the number of holes, and the contribu tions of electrons and positive charges (holes) to the conductivity a r e equal. Such conductivity is known a s i n h e r e n t conductivity. The addition of s m a l l amounts of other elements may c r e a t e conditions such that the numbers of electrons and holes a r e not equal. In such a c a s e (depending on the nature of the additive) the conductivity is created preferentially through the motion of either electrons (n-type semiconductors) o r holes (p-type semiconductors) Some of the additives which c r e a t e an n-type conductivity in germanium (the so-called donor additives) are elements whose valency is higher than four, e. g. , arsenic, antimony, and phosphorus. This is because when a pentavalent element replaces a germanium atom in the lattice, the fifth electron is weakly bound to the atom and readily p a s s e s into the conductivity zone (see Figure 153). The additives producing p-type conductivity (acceptor additives) comprise elements with a valency below four - indium, gallium, aluminum, copper, zinc, etc. One unfilled lattice point remains when the atom of a trivalent element replaces an atom of tetravalent germanium. In such a case the number of holes is g r e a t e r than the number of electrons. The conductivity of semiconductors is strongly affected by external factors such a s illumination, heat, and electric field. The effect of a light beam on the conductivity is known a s the p h o t o e 1e c t r i c effect. One of the important properties of semiconductors is that a thin film which conducts the c u r r e n t only in one direction is created on the boundary when semiconductors of different types (n- and p-type) a r e brought into contact. This film is known a s the b a r r i e r layer. Accordingly, semi conductors may be used a s rectifiers. Some of the electrical properties of germanium, which a r e of value f o r i t s characterization a s a semiconductor, a r e listed below:

Width of forbidden zone, A E , eV Electron mobility*, p c m Z / V .sec Hole mobility*. p cmz/V.sec Specific resistance of high-purity germanium monocrystals ( a t 2 5 " C ) , o h m . c m .

................0.7 ................3900 ...................1900 ........... 55-60

* T h e mobility of the current carriers (p) is a measure of the


drift (motion) rate of electrons or holes in an electric field with an applied voltage of one volt.

--------------

The above value of the specific resistance is that of very pure germanium, and is close to the inherent resistance of germanium. The specific resistance of germanium is sharply reduced by the presence of impurities. As with all semiconductors, the specific resistance of germanium decreases with increasing temperature. There is a characteristic dependence of the electrical resistance of germanium on the pressure. Chemical properties of germanium. P u r e solid germanium is not attacked by a i r a t ambient temperature, but is rapidly oxidized at 600 to 700 with the formation of germanium dioxide.

305

The action of ammonia on germanium o r germanium dioxide a t 700 to 800" r e s u l t s in the formation of germanium nitride Ge3N2. The nitride is not decomposed by water, dilute alkalies, o r acids. It dissociates a t about 1000". A vigorous reaction between germanium and chlorine is observed a t room temperature, with the formation of GeC14 (bp 83"). Germanium r e a c t s with bromine and iodine on heating. Sulfur vapor r e a c t s with germanium, yielding GeS. Germanium is not attacked by hydrochloric o r dilute sulfuric acids. The metal is slowly dissolved by hot concentrated sulfuric acid, with the evolution of S Q . Nitric acid r e a c t s with germanium, yielding the hydrated dioxide GeOn nH20. The metal is readily dissolved by aqua regia. Alkali solutions have little effect on germanium, but i t is rapidly dissolved by molten alkalies in the presence of air. Germanium dissolves in dilute hydrogen peroxide solutions, yielding pe rge r m anic acid. Germanium f o r m s no carbides. It can be fused in a graphite crucible without significant contamination with carbon.

The properties of germanium compounds Compounds of tetravalent and divalent germanium a r e known; the former a r e m o r e stable. Some of the technologically important germanium compounds a r e described below. Germanium oxides and hydroxides. G e r m a n i u m d i o x i d e ( a white powder) is the main starting material for the production of metallic germanium. It is produced by the dehydration of the hydrate (Ge02-nHpO) precipitated during the neutralization of solutions containing germanium s a l t s (chloride, sulfate, nitrate). The melting point of GeQ i s 1115". Appreciable vaporization of the dioxide occurs above 1250". The heat of formation of G e Q is 128 kcall mole. The dioxide is reduced to the metal by hydrogen above 600". In contrast to silicon dioxide, germanium dioxide is appreciably soluble in water. The solubility at 20 and 100" is 0.4 and 0.9570 respectively. Germanium dioxide dissolves in alkalies with the formation of g e r m a n Alkali germanates a r e soluble in water; the germanates of the ates alkaline-earth and heavy metals a r e sparingly soluble in water but a r e readily decomposed by acids. G e r m a n i u m m o n o x i d e GeO (a dark-gray powder) is formed when germanium dioxide is reduced by hydrogen or carbon monoxide. Appreci able sublimation of GeO occurs above 700". The hydroxide Ge(0m2, corresponding to germanium monoxide, is precipitated a s an orangeyellow precipitate on the addition of alkali hydrdxides to solutions containing Ge2+ions. The hydroxide Ge(OW2is amphoteric. When dissolved in alkalies the hydroxide yields germanates (e. g . , NaHGeQ). Germanium sulfides. Germanium forms two sulfides: GeE& (white) and GeS (brown). G e r m a n i u m d i s u 1f i d e is precipitated by hydrogen sulfide from acid solutions. The acid must have a concentration of 4 to 5 N. GeS, undergoes hydrolysis in solutions of lower acidity:

306

GeS,

+ + 2) H,O zGeO,- n H,O + 2H&


(rt

P a r t i a l dissociation of GeS, (to GeS and sulfur) occurs in inert media a t temperatures above 400. The "apparent vapor p r e s s u r e " of Ge% at 550" is 2 . 5 10-3mmHg. Above 400" the disulfide is oxidized in a i r . The di sulfide dissolves in alkali sulfides o r alkali solutions with the formation of sulfogermanates. The dissolution of the disulfide in ammonium sulfide yields a sulfide with the formula (NH4),GezS7. G e r m a n i u m m o n o s u l f i d e GeS is p r o d u c e d a s a r e s u l t o f t h e reduc tion of GeSz by hydrogen or its thermal dissociation. It is characterized by its appreciable volatility at temperatures above 450". The vapor p r e s s u r e of GeS at various temperatures is given below:

Temperature. 'C Vapor pressure,"

.... ... . .. . 400


Hg

......

0.25

450 1.60

500 3.48

550 14.21

600 40.96

Germanium monosulfide is soluble in hydrochloric acid and alkali solutions. Germanium halides. Germanium forms volatile halides of the type GeX,, where X is F , C1, Br or I. Germanium tetrachloride GeC1, is of importance in technology. The chloride i s produced by dissolving germanium dioxide in 6 NHCl or by chlorination of germanium compounds (with chlorine). Germanium tetrachloride hydrolyzes in acids with a concentration below 6 N, with the precipitation of the hydrated germanium dioxide. GeC14 is a colorless liquid with a boiling point of 83". The chloride distils over when a hydrochloric acid solution containing germanium tetrachloride is boiled. The solubility of GeC14 in hydrochloric acid is strongly affected by the acid concentration. Thus, the solubility of GeC1, in 16 and 7.77NHC1 is 0.88 and 85.36g GeC14/1000g solution respectively. Germanium hydrides (germanes) Germanium f o r m s hydrogen compounds which resemble the hydrogen compounds of silicon (silanes) and of carbon (hydrocarbons). M o n o g e r m a n e GeH, i s a colorless gas at ambient temperatures. It condenses at -83". Monogermane may be prepared by the decomposition of magnesium germanide with hydrochloric acid:

GeMg,

+ 4HC1 = GeH, + 2MgC1,.

At temperatures above 280" GeH, i s dissociated into germanium and hydrogen. The homologues of GeH4 - Ge2H, and Ge3H, - a r e colorless liquids. In germanes hydrogen may be replaced by halogen atoms. Sucha substituted compound is germanochloroform GeHC13 - a liquid with a boiling point of 75".

Uses of germanium / 5 / As a semiconductor, germanium is of importance in electronics. It is used for the production of crystal rectifiers (diodes) and amplifiers (triodes). Crystal r e c t i f i e r s and amplifiers have s e v e r a l advantages over vacuum tubes: they require considerably less power and they have a longer service life; a s compared with vacuum tubes, semiconductor devices

307

a r e m o r e stable to mechanical vibration and impact, and a r e much s m a l l e r in size. Since no energy is used to heat the emitter ( a s in vacuum tubes), the efficiency of crystal amplifiers attains 40 to 50%. All these advantages favor the use of c r y s t a l rectifiers and amplifiers, especially so in complex computers, remote control, and r a d a r setups. Rectifying contacts (known as p-n o r n-p junctions) a r e produced in germanium monocrystals through the diffusion of impurities to the surface layer of a germanium plate. One commonly used method for the formation of a rectifying contact is the application of molten indium to the surface of a plate of n-type germanium. As a r e s u l t of the diffusion of indium atoms into the germanium, a thin layer with p-type conductivity is created at the contact s i t e . In addition to i t s use a s rectifiers in radio receivers, germanium has been recently used a s high-power rectifiers. of standard frequency AC, for currents of 6000 to 10,000 amp o r more. These r e c t i f i e r s a r e characterized by their high efficiency (the direct current produced is 95% of the alternating current used). Germanium rectifiers operate a t current densities which exceed several times the maximum permissible current density in selenium and other rectifiers. They a r e s m a l l in size and a r e especially convenient to move from place to place. Germanium triodes a r e extensively used for the amplification, generation, o r transformation of electrical oscillations. There a r e two types of triodes: n-p-nandp--n-p. In n - p - n triodes two l a y e r s of n-type germanium a r e separated by a thin layer with a p-type conductivity, while in p - n- p triodes the center layer has n-type conductivity while the external l a y e r s a r e of p-type ( s e e Figure 154).

FIGURE 154. Diagram of a germaniu..: triode


a
-I(

-p

-"triode:

b-p

- n -ptriode.

FIGURE 155. Diagram of a semiconductor triode with junction-type contacts.


1-germanium; 2-indium.

A diagram of a triode of the p - n -p type, in which p-type conductivity is created in the external l a y e r s by fusing drops of indium onto a plate of n-type germanium, is shown in Figure 155. Such triodes a r e small and weigh only a little over 1 gram. Like other semiconductors, germanium is used for the production of the r m ist o r s These devices make use of the fact that the electrical resistance of germanium is strongly affected by the temperature, which makes it possible to determine the temperature from the variation in the electrical resistance. Small germanium plates, serving as thermistors, may be used to measure the temperature anywhere in a room, in pipes,

308

vessels, and various mechanisms, s o that automatic signalization and control a r e readily effected, Thermistors a r e a l s o used in time relays and in instruments ensuring a gradual (at any desired rate) increase of the current in a circuit. Germanium is used f o r the production of p h o t o c e 11s with a b a r r i e r layer and of t h e r m o e l e m e n t s . Germanium f i 1m r e s i s t o r s a r e used in radio technology. A thin film deposited on glass by thermal dissociation of the gaseous monogermane Ge& o r germanium halide (GeC14, G I 4 ) has a resistance between 1000 o h m s and several megohms. Other uses of germanium a r e of secondary importance. Thus, for instance, uses have been proposed for some germanium alloys. An aluminum-germanium alloy (7470Al, 2170 Ge, 270 Fe, and 370Si) has been recommended for the production of vacuum tube cathodes. The use of the low-melting eutectic Au-Ge alloy (12% Ge, mp 356") has been proposed to obtain hard coatings on gold and for improving the quality of the gold solders. The presence of small amounts of germanium improves the fatigue and corrosion resistance of magnesium castings and increases their creep strength. The production of germanium in the Western countries in 1 9 6 0 to 1962 is estimated at 6 0 to 70 tons. The production volume at the present time is very probably higher. 78. OCCURRENCE

The concentration of germanium in the Earth's c r u s t is 7. l o m 4 7 0 by weight. Most of the germanium is scattered in silicates, sulfides, and complex sulfide minerals. The concentration of germanium in zinc, copper, lead, and iron sulfides ranges from a few thousandths to a few tenths of one percent. The highest concentrations of germanium (0.01 to 0.170 ) a r e encountered in low-temperature zinc blendes. There a r e several minerals (of the thio s a l t type) with high germanium contents. Some of these a r e listed below. A r g y r o d i t e Ag8GeS, contains 5 to 7% Ge (the element germanium w a s discovered in argyrodite). G e r m a n i t e Cu3(Fe, Ge, Ga, Zn)(As, S)4 (this formula is only approxi mate) contains 6 to 1 0 % Ge, 6 to 870 Fe, and 0.5 to 0.870Ga. This mineral was found for the f i r s t time in 1918 in the copper-lead-zinc o r e s of Tsumeb in South-West Africa. R e n i e r i t e (Cu, Fe)3(Fe,Ge, Zn, Sn)(S, A s ) 4 contains 6.37 to 7.870 Ge. The mineral was found in the copper-zinc o r e s of Katanga. In addition to sulfide ores, various c o a 1 s also s e r v e a s a source of germanium. The germanium content of coals of various types fluctuates between 0.001 and 0.0170, and the germanium is preferentially concen trated in coals of a low degree of metamorphization (anthracite coals contain almost no germanium). It has been found that the lower the ash content of the coal, the higher i t s germanium content 1131.

309

Behavior of germanium during processing of sulfide raw m a t e r i a l s During the beneficiation of sulfide o r e s containing s e v e r a l metals, the germanium is concentrated in the zinc, copper, o r mixed zinc-copper concentrates, depending on the form in which i t occurs in the raw m a t e r i a l (the germanium may be present either a s an isomorphous im purity in the sulfides o r as m i n e r a l s of the germanite o r renierite types). The production of zinc, In the pyrometallurgical (distillation) p r o c e s s for the production of zinc, zinc concentrates a r e subjected f i r s t to an oxidative and then to an agglomerative roasting in sintering equipment. Because of the high temperature doring the agglomerative roasting (1200 to 1300) a fraction of the arsenic, lead and cadmium in the m a t e r i a l to be roasted is vaporized. In the p r o c e s s a considerable fraction of the germanium ( a s chloride o r oxides) is expelled with the g a s e s and con centrates in the dust entrapped i n the electrostatic filters or some other dust trap. During the distillation of the zinc in r e t o r t furnaces germanium (which has a high boiling point) accumulates in the r e t o r t residues in which its concentration may occasionally reach a few hundredths or a few tenths of one percent. In the hydrometallurgical production of zinc most of the germanium is concentrated in the solid wastes from the leaching of zinc cinders. The concentration of germanium in the zinc sulfate solutions taken for electrolysis must not exceed 0.1 m g / l since the presence of germanium interferes with the electrolytic deposition of the zinc (prevents the formation of a continuous layer of zinc and causes dissolution of the zinc cathode). During the processing of zinc cakes (leaching residues) by the sublima tion method (the so-called rotary-kiln process), germanium is concen trated in the distillate (oxides). This is attributed to the volatility of germanium monoxide GeO. Thus, germanium may be produced a s a by-product of zinc manufacture f r o m the following m a t e r i a l s : dust from agglomerative roasting, r e t o r t residues, solid residues from the leaching of zinc cinders, and from oxides of the rotary-kiln process. The production of copper 1101. During the roasting of germaniumcontaining copper and copper-zinc concentrates, most of the germanium r e m a i n s in the cinders since the germanium dioxide formed is nonvolatile under the roasting conditions (800 to 850"). A fraction of the germanium may be l o s t through the interaction of GeOz with sulfides and the formation of GeS, which is volatilized at these temperatures: 3Ge0,

+ 4Fe (Zn) S

-+

3GeS

+ 4Fe (Zn) 0 + SOs.

In most c a s e s the roasted o r crude concentrates a r e taken for the smelting of matte in reverberatory furnaces. In addition, the matte is smelted directly from copper o r e s in shaft furnaces (the so-called "water jackets"). The distribution of germanium between the products of the reverberatory smelting of copper concentrates depends on the concentration of sulfur (in the form of sulfides) in the starting r a w material. In the smelting of

310

of the crude concentrates, most of the germanium (80 to 9 0 % ) p a s s e s into the matte, and the remaining germanium is distributed between the s l a g s and the dust. In the smelting of roasted concentrates most of the germanium ( 6 0 to 8070, depending on the degree of roasting) p a s s e s into the p r i m a r y slags, while the remaining 2 0 to 40% is distributed between the matte and the dust. The highest germanium content is usually observed in the dusts.
TABLE 53 Approximate composition of the dust formed in the smelting of copper concentrates in water-jacketed furnaces Copper concentrates from Copper concentrates from Katanga (Africa) Mansfeld (East Germany)

Element

%
Ge
Zn Pb

7 0
0.007-0.008
23.6-31.36
16.35-19.5
0.10-0.13
0.66-1.06
Not determined
0.6- 1.19
15.1 - 19.9
6.34-7.10

Cd cu As

61
Sulfur Bitumen

0.36 28 25 3 1.5 1.3 Not determined

i'

In the shaft smelting of copper concentrates, the degree of volatilization of the germanium i s higher than in reverberatory smelting. The d i s t r i bution of germanium after semipyritic smelting in one Soviet plant was a s follows: in the dust 4070, in the matte 1870, and in the s l a g s 4270. When using a charge with a germanium content of 3 g/ ton, the germanium concentration in the c o a r s e and fine dusts was 2 0 and 60 gfton respectively. The main