uecarburlzaLlon of 1093 SLeel

Aprll 222011

MA1L 370 klneLlcs
ur Chen
ConcenLraLlon CradlenL
1homas Agasld Mlra ALlya Casey 8arbarlno uavld 8aruela
Crlffln 8eemlller 8achel 8eLhancourL klm SmlLh

1

Abstract
An experlmenL was conducLed by heaL LreaLlng 1093 sLeel for varlous Llmes aL a serles of dlfferenL
LemperaLures ln order Lo deLermlne Lhe cause of low surface hardness values uecarburlzaLlon ln Lhe
furnace was analyzed by measurlng Lhe decarburlzed layer and relaLlng lL Lo an experlmenLally
deLermlned acLlvaLlon energy 8esulLs were deLermlned Lo be lnconcluslve due Lo an lron oxlde on Lhe
surface of Lhe sLeel samples causlng varlablllLy ln Lhe slze of Lhe alpha ferrlLe decarburlzaLlon layer
Introduction
1he low surface hardness of heaL LreaLed 1093 sLeel spurred an lnvesLlgaLlon lnLo Lhe cause of Lhls
dlscrepancy 1he assumpLlon of reduced hardness due Lo decarburlzaLlon durlng heaL LreaLlng ls a hlghly
probable cause uecarburlzaLlon occurs aL elevaLed LemperaLures when enough energy ls provlded for
Lhe carbon near Lhe surface of Lhe sLeel Lo dlffuse ouL of Lhe alloy lnLo Lhe alr of Lhe furnace whlch
conLalns a lower carbon conLenL Lhan LhaL of Lhe sLeel WlLh Lhe carbon dlffuslng lnLo Lhe furnace alpha
ferrlLe ls lefL aL Lhe surface of Lhe sLeel 1o verlfy Lhe valldlLy of Lhls assumpLlon samples of 1093 sLeel
were heaLed aL Lhree LemperaLures for varylng amounLs of Llme and examlned Lo deLermlne Lhe slze of
Lhe decarburlzaLlon layer 1hls measured decarburlzaLlon layer (x) was compared Lo Lhe LheoreLlcal layer
and Lhe acLlvaLlon energy (C) was calculaLed uslng Lhe followlng equaLlons
(Lq 1)
(Lq 2)
uslng Lhese relaLlonshlps Lhe dlffuslvlLy (u) aL each of Lhe LesLed LemperaLures ls deLermlned and used
Lo flnd C lurLher Lhe resulLs are examlned for Lhelr conslsLency wlLh Lhese relaLlonshlps
Procedure
Pre-Heat Treatment
1hree bars of 1093 sLeel were cuL lnLo four equal secLlons creaLlng 12 samples noLches were cuL lnLo
Lhe samples ln order Lo dlfferenLlaLe beLween Lhe samples afLer Lhe heaL LreaLmenL 1he conLrol sample
dld noL have any noLches whlle Lhe samples Lo be heaL LreaLed had elLher one Lwo or Lhree noLches Lo
ldenLlfy Lhe lengLh of Llme ln Lhe furnace Lach sample's surface was sLerlllzed uslng an alcohol swab
Heat Treatment
A LoLal of nlne samples were placed ln Lhe furnace aL speclfled LemperaLures and Llmes (1able l)
2

TabIe I: Heat Treatment Temperature and Time

1emperature 1|mes
830C 1hr 3hr 3hr
86SC 1hr 2hr 3hr
900C 03hr 1hr 3hr

Samples were heaL LreaLed one aL a Llme ln Lhe furnace 8eLween heaL LreaLmenLs Lhe furnace was lefL
empLy and closed for flve mlnuLes Lo allow Lhe furnace Lo reach Lhe deslred LemperaLure before
beglnnlng Lhe nexL LreaLmenL All samples were normallzed aL room LemperaLure followlng Lhelr
speclflc heaL LreaLmenL AfLer normallzlng each sample Lhe bark" or surface oxlde layer was removed
Lo creaLe an even surface for vlewlng and analyzlng under an optical microscope.
etallograpby
Samples were cuL ln half exposlng Lhe crosssecLlon Lo vlew decarburlzaLlon layer and each sample was
mounLed ln cold seL epoxy A color codlng sysLem was used Lo ldenLlfy Lhe LemperaLure and Llme of Lhe
heaL LreaLmenL process for each sample Samples were allowed Lo cure ln Lhe epoxy for approxlmaLely
10 mlnuLes

All samples wenL Lhrough Lhe serles of 240 320 400 600 and 800 grlL sandpaper Lach sample was
sanded ln only one dlrecLlon on each grlL unLll a unlform surface was vlsually observed 8eLween each
lncrease ln grlL Lhe sample was roLaLed 90 degrees so LhaL Lhe new scraLches were perpendlcular Lo Lhe
scraLches from Lhe prlor grlL AfLer sandlng each sample was pollshed on boLh Lhe 6 m and 1 m
pollshlng pads ln order Lo provlde an approprlaLe surface for lmaglng Lach sample was Lhen examlned
under Lhe mlcroscope Lo ensure an adequaLe pollsh had been aLLalned

A 2 nlLal eLchanL was creaLed by mlxlng 30ml eLhanol and 1ml PnC
3
1he 2 nlLal eLchanL was
swabbed onLo each sample and lefL Lo reacL for 10 seconds Samples were Lhen rlnsed wlLh eLhanol
drled wlLh Lhe heaL gun and examlned under Lhe mlcroscope Lo ensure an adequaLe eLch lf a sample
appeared undereLched lL was llghLly pollshed on Lhe 1um pad and reeLched for anoLher 10 seconds lf
a sample appeared overeLched lL was llghLly pollshed on Lhe 1um pad and examlned under Lhe
mlcroscope Lo conflrm an accepLable condlLlon for mlcroscopy

Lach sample was examlned under Lhe opLlcal mlcroscope for valldaLlon as an approprlaLe sample as
sLaLed by Lhe Amerlcan SocleLy for 1esLlng and MaLerlals (AS1M) sLandard L1077 lmages of Lhe LoLal
and parLlal decarburlzaLlon layer of each sample were capLured CompuLer sofLware was used Lo
measure Lhe depLh of Lhe decarburlzaLlon layer aL mulLlple polnLs on each sample and Lhe average was
calculaLed for each sample
AddlLlonally mlcrohardness values were also Laken on each sample of 1093
3

esults
1he meLallographlc analysls of Lhe heaL LreaLed samples ylelded measuremenLs of Lhe average
decarubuzaLlon layer (x) values ln mlcromeLers 1hese x measuremenLs were found by summlng Lhe
average of Lhe flve measuremenLs of full decarburlzaLlon layer and Lhe flve measuremenLs of Lhe parLlal
decarburlzaLlon layer (llgure 1)

|gure 1 1he opt|ca| m|croscopy |mage of samp|e heated at 830C for 3 hours at
100x magn|f|cat|on shows measurements of fu|| and part|a| decarbur|zat|on |ayer
1hese x values are compared Lo Lhe LheoreLlcal x calculaLed uslng equaLlons 1 and 2 (1able ll)
1ab|e II Measured and ca|cu|ated decarbur|zat|on |ayer at each heat treatment

neat 1reatment k
ave
measured (m) Std Dev k
ave
ca|cu|ated (m)
830C 1 hr 1864 1464 9003
830C 3 hrs 20976 1443 13397
830C S hrs 3776 1209 20136
86SC 1 hr 12002 213 1113
86SC 2 hrs 14968 1063 13768
86SC 3 hrs 21366 1698 19312
900C 0S hr 13334 1468 10213
900C 1 hr 20842 202 14443
900C 3 hr 19762 1412 23013

lrom Lhls daLa Lhe dlffuslvlLy (u) aL each LemperaLure can be deLermlned by ploLLlng x
2
wlLh respecL Lo
Llme 1he slope of Lhls ploL ls equal Lo u ln m
2
/s lrom Lhese deLermlned u values C ls found by ploLLlng
ln(u) agalnsL 1/81 (llgure 2)
4

1he resulLs of Lhe mlcrohardness LesLs were deLermlned Lo be lnappllcable because AS1M L1077
volume 0301 sLandard sLaLes LhaL mlcrohardness LesLlng for measurlng effecLlve depLh of
decarburlzaLlon ls besL for carbon conLenLs below 033 wL C Slnce Lhe sLeels samples have a carbon
conLenL greaLer Lhan 033wL C mlcrohardness measuremenLs Laken could noL be used
iscussion
1he concenLraLlon of carbon ln Lhe 1093 sLeel samples 093C was slgnlflcanLly hlgher Lhan Lhe
concenLraLlon of carbon lnslde Lhe furnace whlch was assumed Lo be 0C 1he concenLraLlon gradlenL
creaLes a sLaLe of nonequlllbrlum ln Lhe sysLem 1he heaL LreaLmenLs provlded Lhe acLlvaLlon energy
needed Lo move Lhe reacLlon ln Lhe dlrecLlon of equlllbrlum Slnce Lhere ls a much hlgher concenLraLlon
of carbon ln Lhe sLeel samples Lhan ln Lhe alr of Lhe furnace carbon dlffused from Lhe sLeel Lo Lhe alr ln
Lhe furnace AL LemperaLures above 830

C Lhe carbon has Lhe drlvlng force needed Lo go Lo a lower

energy sLaLe ln Lhe alr wlLh a lower concenLraLlon of carbon 1hough Lhe slze of Lhe decarburlzaLlon
layer ls dependenL upon boLh Llme and LemperaLure LemperaLure has an exponenLlal relaLlonshlp wlLh
dlffuslvlLy 8ecause LemperaLure ls an exponenLlal facLor of Lhe Arrhenlus equaLlon lL has Lhe largesL
effecL on Lhe slze of Lhe decarburlzaLlon layer

lmages of Lhe decarburlzaLlon layer of Lhe 1093 sample heaL LreaLed aL 830

C for 3 hours provlde a vlsual

descrlpLlon of decarburlzaLlon afLer normallzlng (llgure 2) 1he measured value of Lhe decarburlzaLlon
layer was smaller Lhan Lhe calculaLed value due Lo Lhe bark" or lron oxlde layer LhaL formed on Lhe
ouLer surface of Lhe sLeel whlle ln Lhe furnace 1he bark formed from Lhe alpha ferrlLe presenL aL Lhe
decarburlzaLlon layer 1he alpha ferrlLe LhaL would have been measured ln Lhe decarburlzaLlon layer
wenL Lo form Lhe lron oxlde layer 1he loss of Lhe decarburlzaLlon layers was relaLed Lo Lhe Llme and
LemperaLure of Lhe heaL LreaLmenL Slnce LemperaLure has Lhe largesL effecL on Lhe formaLlon of Lhe
decarburlzaLlon layer lL was also Lhe mosL heavlly welghLed varlable on Lhe amounL of bark formaLlon
1he assumpLlon was made LhaL all of Lhe bark formed from Lhe alpha ferrlLe decarburlzaLlon layer
3


|gure 2 In the m|crostructure of the surface of th|s heat treated samp|e a
d|st|nct decarbur|zat|on |ayer of a|pha ferr|te can be seen as we|| as an area of
part|a| decarbur|zat|on

1he mlcrograph of Lhe sample LreaLed aL 830
o
C for 3 hours shows a relaLlvely large decarburlzaLlon
layer ln comparlson Lo Lhe mlcrograph of Lhe 830
o
C for 3 hours sample Lhe 900
o
C for 3 hour sample
has a smaller decarburlzaLlon layer (llgure 3) 1he assumpLlon ls LhaL as Lhe decarburlzaLlon layer
decreases Lhe bark formaLlon lncreases AlLhough one would assume LhaL as LemperaLure lncreases
decarburlzaLlon would lncrease lL ls noL a safe assumpLlon slnce Lhe Arrhenlus equaLlon along wlLh
LquaLlon 1 and LquaLlon 2 are noL solely relaLed Lo decarburlzaLlon dlsLance buL also Lo bark
formaLlon As sLaLed prevlously LemperaLure plays an exponenLlal role ln Lhe decarburlzaLlon process
and Lherefore causes compllcaLlons ln Lhe samples 1he sample aL 900
o
C for 3 hours proves LhaL as Lhe
sLeel ls heaLed aL hlgher LemperaLures whlle keeplng Llme consLanL some of Lhe alpha ferrlLe ls
oxldlzed Lherefore reduclng Lhe decarburlzaLlon layer and lncreaslng bark formaLlon ln a sense Lhe
900
o
C 3 hour sample was over heaL LreaLed" whlch ls why Lhe decarburlzaLlon layer decreased


6


|gure 3 1h|s m|crograph shows a re|at|ve|y sma|| decarbur|zat|on |ayer A
sma||er decarbur|zat|on |ayer can be attr|buted to the ox|dat|on of the outer |ayer
of a|pha ferr|te

1hrough Lhe average measured values of Lhe decarburlzaLlon layer Lhe dlffuslvlLy (u) aL each
LemperaLure found by ploLLlng x
2
wlLh respecL Lo Llme 1he slope of Lhls llnear relaLlonshlp glves u ln
m
2
/s loLLlng Lhe dlffuslvlLy and LemperaLure we were able Lo deLermlne Lhe acLlvaLlon energy of Lhe
reacLlon (llgure 4)


|gure 4 1he s|ope from the p|ot of the natura| |og of the d|ffus|v|ty (D) w|th respect to 1]k1 was
ca|cu|ated to determ|ne the act|vat|on energy
y 62971x 32679
8 04392
264
263
262
261
26
239
238
237
236
102L04 103L04 104L04 103L04 106L04 107L04 108L04 109L04 110L04
|
n
(
D
)
1]k1
Determ|nat|on of
7

1he bark was ground Lo a powder and evaluaLed uslng Lhe x8ay ulffracLomeLer (x8u) 1he resulLs
showed Lhe bark Lo conLaln Lwo lron oxldes MagneLlLe (le3C4) and PemaLlLe (le2C3) (Appendlx A)

1he experlmenLally deLermlned acLlvaLlon energy of 63 k!/mol had a 37 error from Lhe LheoreLlcal
acLlvaLlon energy of 148 k!/mol uesplLe Lhe calculaLed value of Lhe acLlvaLlon energy belng much lower
Lhan Lhe LheoreLlcal Lhls may explaln why some of Lhe decarburlzaLlon layers were larger Lhan Lhe
calculaLed values slnce less energy was requlred for Lhe carbon Lo dlffuse ouL of Lhe sLeel
onclusion
A decarburlzaLlon layer was examlned and measured for all heaL LreaLed samples yeL Lhe
experlmenLally deLermlned acLlvaLlon energy had a 37 error uue Lo Lhls hlgh value of error and Lhe
varlablllLy of Lhe resulLs lL was deLermlned LhaL Lhe resulLs are lnconcluslve lurLher research ls needed
on Lhe relaLlon of Lhe alpha ferrlLe Lo Lhe formaLlon of Lhe bark oxlde

Appendlx A

oxide powder
00-033-0664 (*) - Hematite, syn - Fe2O3 - Y: 50.00 % - d x by: 1. - WL: 1.5406 - Hexagonal (Rh) - a 5.03560 - b 5.03560 - c 13.74890 - alpha 90.000 - beta 90.000 - gamma 120.000 - Primitive - R-3c (167) -
01-082-1533 (C) - Magnetite, syn - Fe3O4 - Y: 50.00 % - d x by: 1. - WL: 1.5406 - Cubic - a 8.39700 - b 8.39700 - c 8.39700 - alpha 90.000 - beta 90.000 - gamma 90.000 - Face-centered - Fd-3m (227) - 8 -
Operations: Background 0.309,1.000 | mport
oxide powder - File: oxide2.raw - Type: 2Th/Th locked - Start: 20.000 - End: 80.000 - Step: 0.100 - Step time: 3. s - Temp.: 25 C (Room) - Time Started: 1303159680 s - 2-Theta: 20.000 - Theta: 10.00
L
i
n

(
C
o
u
n
t
s
)
0
100
200
300
400
500
600
700
800
900
1000
1100
1200
1300
1400
1500
2-Theta - Scale
20 30 40 50 60 70
1