DUDLEY KNOX LIBRARY NAVAL POSTGRADUATE SCHOOL MONTEREY CA 93943-5101

LIBRARY
U.S.

NAVAL POSTGRADUATE SCHOOL

MONTEREY CALIFORNIA

MEASUREMENT OF
TRANSFERENCE NUMBERS
OF THE BISULFATE ION
IN AQUEOUS SULFURIC ACID

*****
Howard W. Gardner

MEASUREMENT OF
TRANSFERENCE NUMBERS
OF THE B I SULFATE ION
IN AQUEOUS SULFURIC ACID

by Howard W. Gardner
Lieutenant, United States Navy

Submitted in partial fulfillment of the requirements for the degree of

MASTER OF SCIENCE

United States Naval Postgraduate School Monterey, California

19

T-Jbrnry
V. s. Naval Postgraduate School Monterey, California

MEASUREMENT OP
TRANSFERENCE NUMBERS
OF THE B I SULFATE ION
IN AQUEOUS SULFURIC ACID

by

Howard W. Gardner

This work is accepted as fulfilling

the thesis requirements for the degree of

MASTER OF SCIENCE
from the
United States Naval Postgraduate School

ABSTRACT

The unusually high mobility of the hydrogen ion and the hydroxide ion
in water has been ascribed to a proton transfer chain mechanism. In re-

cent years it has been demonstrated that a similar mechanism results in

abnormal mobility of the bisulfate ion in absolute sulfuric acid.

In an

effort to determine the nature of the mechanism in the transition from chain

conductance by the hydronium ion to chain conductance by the bisulfate ion

in aqueous sulfuric acid solutions, Hittorf transference numbers were meas-

ured for the bisulfate ion in aqueous sulfuric acid above

707.

acid by weight.
The trans-

The transference numbers measured ranged from 0.07 to 0.14.

ference numbers obtained in these measurements are the result of physical

migration only and not a proton jump chain transfer; hence these values may
be considered as a confirmation of the existence of a chain mechanism in

this concentration range, but do not permit further determination of the

nature of the chain mechanism.

It is estimated that approximately two-

thirds of the current is carried by a proton transfer chain conductance

mechanism in the vicinity of the monohydrate composition.

The writer wishes to express his appreciation for the assistance and

encouragement given by Professor Richard A. Reinhardt of the U.

S.

Naval

Postgraduate School in this investigation.

ii

TABLE OP CONTENTS

Section
1.

Title

Page
1

Introduction
Experimental Procedure
Experimental Results

2.
3.

4 15

4.

Conclusions

22

Bibliography

23
25

Appendix

Composition of Sulfuric Acid Solutions

Appendix II

A Model of the Conductance Mechanism in Sulfuric Acid

30

iii

LIST OP ILLUSTRATIONS

Figure
1.

Page

Properties of Sulfuric Acid Solutions

Transference cell

3 5 8

2.
3.

Electrolysis circuit
Bisulfate ion transference numbers

4.

17

II-l
II-2

Conductivity according to model

Transference numbers according to model

39

44

iv

1.

Introduction

The abnormally high conductance of the hydrogen ion in water has been

known for many years and has been explained by a type of Grot thus chain

mechanism

p.,

2,

7J

Hammett and Lowenheim{9j measured the transference

numbers of barium and strontium ions in sulfuric acid and found that almost none of the current was being carried by the cation; they proposed a
similar mechanism of chain conductance in sulfuric acid.
More recent work,

especially that of Gillespie and Waeif [4,

theory.

6,

17}

has substantiated this

Similar abnormal mobility of the hydrogen ion has been found in

In general,

other hydrogen- bonded solvents.

there are apparently two prin-

cipal types of chain conductance mechanisms: conductance by an excess proton,

and conductance by a defect proton.

In the water system,

the hydrogen

ion conducts by the excess proton mechanism, and the hydroxide ion conducts

by the defect proton mechanism.

Similarly, in sulfuric acid the H3SO/

ion

conducts by an excess proton mechanism, and the HSOa" ion conducts by a defect proton mechanism.
The autoproto lysis of water is given by:

(1)

H2

H3 +

OH"

In sulfuric acid the autoproto lysis is represented by:

(2)

H S0 4 2

H S04 + 3

HS0 "'
4

In solutions of water in sulfuric acid, the ionization of the water at low

water concentration is essentially complete

H3 +

[^3}

(3)

H 2 S0

HSO4"

This ionization represses the autoproto lysis of the solvent; therefore, to

1

a good approximation,

in solutions with the mole fraction of sulfuric acid

greater than 0.5, the ionic species present will be the hydroniura ion and
the bisulfate ion.

At the concentration of the monohydrate of sulfuric acid, the numbers

of H2SO4 molecules and H2O molecules will be equal.
fore,

There should, there-

be approximately equal opportunity for proton defect conductance of

the HSO," ion through the sulfuric acid molecules and for the excess pro-

ton conductance of the H^0

ion through the water molecules.

Transference

number measurements were made on the sulfuric acid itself in the concentration range from 30 to 85 mole percent lUSO/ in the hope that information
could be obtained for comparison of the two conductance mechanisms.
The physical properties of aqueous sulfuric acid show remarkable var-

iation in the vicinity of equal molar concentration (see Pig. 1).

There

is a maximum of viscosity almost exactly at the monohydrate composition

[l2j.

Similarly there is a conductivity minimum at a mole fraction of sulThe viscosity- conductivity product

furic acid slightly above 0.50 [llj.

shows a sudden change between 40 and 50 mole percent sulfuric acid.

These
The

variations are undoubtedly related to molecular association phenomena.

hope that transference number measurements might provide additional insight

into the variations of the conductivity in this region provided additional

motivation for this study.

Unfortunately, the results obtained do not per-

mit much more than confirmation of the existence of the chain conductance

mechanism in this concentration range.

'

'

'
1
i

'

"^
1

I j

R M*

:#
.
.
.

<Ct >nd act ivi ty [l *l Uj

'-

.,

*
t

y
\ Jo
1 fi

\
1

ft U

1 lJ

Via

rrio'i

sit y-f- nru ur

<

v .product
/
/

..

.,

\ "V
v,

/
A >.T
.

"^-^
Sw

/
\
at) 3)

O 3 3

i
:

/
1

\
:

'a
Vij SCO sit y
,

i
2
d r.
.

>2j^

*
\

o e

H
y
0,

y
h
1

\
,

|:S
\

n
W
y
.

^
0. C

\
0. 8

\ \

\
0, [2

H >

0, 4

JU

'

Mo le FinACt ion
I

s0 '

I :

Fv
i

rc

Pr< >pej fti gs of SuJLfu ric! Ac id So Lut ior 6

H~

H
*

2.

Experimental procedure.

Transference numbers of the bisulfate ion were measured by the

Hittorf method in the system water-sulfuric acid in the concentration
range above
707.

sulfuric acid by weight.

The Hittorf method was chosen

for these measurements when consideration of other methods showed this

to be the only satisfactory method for direct transference number measure-

ments in this system.

There is insufficient knowledge of the system in

this range to permit selection of an appropriate indicator for use in the

moving boundary technique.

The cell potential method used by Hamer at low-

er concentrations requires a large number of consecutive cell potential

measurements; the number of measurements required becomes excessive at

higher concentrations because of heat of dilution effects [sj.

Furthermore,

for the purpose of this study, it is impossible to say with the cell potential measurements whether the conductance is by a chain mechanism or by

migration.

Initially an attempt was made to use a transference cell similar to

that of Hammett and Lowenheim [?J, modified to permit separation of the

cell compartments for weighing.

Difficulties were encountered in maintain-

ing the liquid junction, probably as the result of an air leak in the joint

or stopcock.

Furthermore the rather fragile legs of the liquid bridges

Consequently, this design was finally
2

were quite susceptible to breakage.

abandoned and the design shown in Fig.

was adopted.

In this cell the three electrolyte compartments were standard eight-

inch Pyrex test tubes.

tubing,

The liquid bridges were eight-millimeter Pyrex

the atmospheric openings of which were closed by rubber tubing

and pinchc lamps.

These liquid bridges could thus be filled by means of a

vacuum, and they provided a ready means of separating the electrolyte at

4

Vent

Vent

8-in. Pyrex Test Tube

Figure 2

Transference Cell

the end of each run.

Rubber stoppers were used as an expedient; however,

attack by sulfuric acid was rapid enough to make glass preferable.

Acid

attack on the rubber stoppers was reduced by wrapping the stoppers with
commercial saran film.
The cell was mounted on a removable "Flexofrarae"

support in the water bath during electrolysis.

The electrodes were of bright platinum mesh and were sealed in six-

millimeter soft glass tubing.

Electrical connections to the external cir-

cuit were made by a liquid mercury junction.

Considerable difficulty was

Hammett and Lowenhein

encountered in the problem of electrode selection.

[9 J reported the formation of ozone and other oxidizing agents at the anode

and the formation of sulfur dioxide and free sulfur at the cathode.

Exper-

imentation confirmed the liberation of free sulfur and sulfur dioxide at

the cathode.

Experiments were conducted with Hg.I^SO^ electrodes, but they

were unsuccessful; gas evolution and the liberation of free sulfur were re-

duced but not eliminated.

The unknown products of the cathode reaction

It was possible in the

made data from the cathode compartment useless.

analytical procedure to compensate for any peroxy sulfates formed at the

anode.

Production of ozone and hydrogen peroxide would not affect the reTherefore, it was

sults because of the requirement for electroneutrality.

possible to make all runs with bright platinum electrodes regardless of

the electrode reactions which might occur. The cathode compartment was ig-

nored, and only the anode and center compartments were analyzed.

Electrolysis was carried out using the circuit of Fig.

3.

The record-

ing millivoltmeter was calibrated for current measurement by direct com-

parison with the milliammeter.

The average current could thus be determined

from the record of the millivoltmeter.

The recorder was used in place of

the usual silver or copper coulometer since other errors in the procedure

would probably negate any increased accuracy gained by use of the more accurate coulometers.
In conducting the electrolysis,

the anode compartment test tube was

weighed dry to the nearest hundredth gram on a hanging-pan triple-beam

balance.
The three test tubes were filled to the desired level with stock
The electrode and liquid

acid of approximately the desired concentration.

bridge assembly was then put In place, and the liquid Junctions were filled.
The entire assembly was then placed in the water bath, where the cells were

immersed to the level indicated in Fig.

2.

The temperature of the water The solution was electrolyzed

bath was thermostatically controlled to 25C.

for 23-26 hours at about 30 ma.

During electrolysis the acid was protected

from moisture by silica gel drying tubes connected to the vents.

Upon com-

pletion of the electrolysis, the liquid seals were broken, the anode compartment was removed and weighed, and samples were taken from the anode
and center compartments and were titrated with sodium hydroxide.
The acid solutions were made by diluting
ic acid to the desired strength.
967.

C.P. reagent grade sulfur-

Sodium hydroxide solutions were prepared

The sodium hydroxide
C. S.

from Baker and Adamson

507.

sodium hydroxide solutions.

solutions were standardized against Baker and Adamson A.

reagent grade

potassium acid phthalate, minimum assay

99.97..

Three samples of about five grams of acid each were taken from each

compartment to be analyzed; the actual sample size was determined by the

acid concentration to suit the analytical procedure.

Samples were weighed

to the nearest tenth milligram in covered sample bottles and then quant-

itatively transferred to a flask for titration.

100 ml. of approximately

0.8 M sodium hydroxide was added by pipette; then titration to the end point

was completed with 0.2 M sodium hydroxide using phenolphthalein as an

7

0)
CO

a U

w
E
-C

o
X
H
CO

V
u 3
fcO

X O
y cr

o
<N

H
*-

>> -H

o
C

"3

CD

00

WW

a*

indicator.
The 0.8 M sodium hydroxide was standardized against potassium acid

phthalate by titration of 100 ml of the base from the pipettes used in the
acid titrations with an approximately 0.33 M solution of potassium acid

phthalate made by dissolving a weighed portion of the standard in water

and diluting to one liter.
The 0.2 M sodium hydroxide was standardized by

titration of weighed portions of potassium acid phthalate with the base.

The procedures reduced the error in volume measurement caused by wall ret-

ention of the solution.

The anode reaction may produce ozone, peroxymonosulfate, and peroxy-

disulfate as well as free oxygen.

Ozone production would not interfere

with the above analytical procedure since the hydrogen ion produced per
faraday is the same as that produced in liberation of free oxygen.
The

peroxysulfates, however, would introduce an error due to the different

acid equivalence per faraday for the production of peroxymono sulfuric acid
and peroxydi sulfuric acid.
To compensate for this possibility,
a modifi-

cation of the procedure of Treadwell and Hall QoJ which utilized hydrolysis
of the peroxysulfates was used.

After the endpoint was reached in the

If a detectable amount

titration,

the solution was boiled about 30 minutes.

of persulfate were present,

the solution would now be acidic and additional

titration to a second end point would permit treating the data as if no

peroxysulfates had formed.

The technique was tried with known acid and

potassium persulfate mixtures and found to be quite satisfactory, the results checking to about one part in 10,000 for total acid after hydrolysis

of the persulfate.

It was noticed in these practice titrations that the

presence of the persulfate rendered the first end point indefinite but that
the second was sharp.
In the actual runs there never was a detectable

change in the end point after boiling; this indicates that peroxysulfate

formation was too slight to be significant.

Since subsequent reduction of data vas based on the assumption that
the anode reaction is entirely represented by:

(4)

3 H
2

"

2 e"

it is necessary that any side reactions either must not affect the analy-

tical data or must be compensated for by the analytical procedure.

That

the above procedure satisfies these requirements is shown by consideration

of the reactions involved.

"

The probable side electrode reactions are:

(5) (6) (7) (8)

HS0 4

=
a.

H S
2 2

2 e"

HSO4"
9
II

2 H2

H 2 S0 5
6 H CT 3

+
+ +

H3CV 6 e"
2

2 e'

=
=

O3

+
2

4 H2

H2

H 3 CT

From equations (7) and (8) it can be seen that the production of ozone
or hydrogen peroxide results in the formation of one mole of hydrogen ion
per faraday, which is the same result as that obtained from reaction (4).

These reactions therefore will not interfere with the analysis of data on
the assumption of equation (4).

The peroxydi sulfate hydrolyzes in steps as follows [l3j:

(9)

H 2 S 2 <>8

H2

s =

H 2 so 4

H 2 S0 5 H2 2

(10) (11)

H 2 S0 5

+
=

H2
H2

H 2 S04
h o2

H2

The sum of these equations with the appropriate electrode reaction leads
to an equation equivalent to equation (4).
10

The hydrolysis of persulfate

according to these reactions is the basis for the quantitative analytical

procedure given by Treadwell and Hall
|_16j.

Since the hydrolysis proceeds

the hydrolysis,

through all the probable side reaction products except ozone,

as adapted to this procedure should provide satisfactory compensation for

the side reactions.

The net effect of this hydrolysis is that the total

base used will be the same as if all the acid had existed as H2SO4 and the

electrolysis had proceeded entirely according to equation (4).

The reduction of data was accomplished essentially according to the

method of Weissberger

[l8J

In each compartment of the transference cell

(12) v '

(n ).. v
11

'final
)

.. . (iO. x Tr initial

(AnJ '~ R'el.rn.

v i
,

(An v n) R'migr.
.

where:

(n

number of moles of component R in the compartment

after electrolysis.

(il ).

=
1

number of moles of component R in the compartment before electrolysis.

(Ani X

ex. rn.

change in the number of moles of component R in

the compartment by electrode reaction.

(An

).

change in the number of moles of component R in

the compartment by ion migration.

In the analytical scheme used for this work,

acid, which is equivalent to solely H^SO,.

the component detected is total

Since the sulfuric acid does not

enter into the effective electrode reaction, and because of the requirement
for electroneutrality,
'

(An
H 2 S04
"

el.rn.

is equal to zero, H

and:

(13)

(n

H S0 4 ) final
2

(n

H 2 S0 4 )initlal

(AnH

S0 4 )mi r -

The effects of chain conductance in the anode compartment are best

represented by the reactions:

11

(14)

H S0 2

HS0

"

p'

(15)

H
3

0"'

II

p"'

The protons represented in these equations are removed from the anode com-

partment by chain transfer; the net effect in the compartment is the for-

mation of bisulfate ions and water molecules from sulfuric acid molecules
and hydronium ions.
In order to find the relation between actual physical

migration of bisulfate ion and the analytical data, material balances were

written for the anode compartment.

ing quantities were used:
a =

For these material balances the follow-

the number of moles of HSO/

which entered the anode compartment

by physical migration, per faraday of total charge transfer.

b =

the number of moles of HSO/

formed in the anode compartment by

chain conductance, in accordance with equation (14), per faraday

of total charge transfer.
j.

c =

the number of moles of

II

which were transferred out of the

3

anode compartment by physical migration, per faraday of total

charge transfer.
d =

the number of moles of

HO

which were decomposed in the anode

compartment according to equation (15), per faraday of total

charge transfer, as a result of chain conductance.
Since the conductance is considered to result from the combination of these
two mechanisms alone and these ions are estimated to be the only ions pre-

sent in significant numbers, the following relation results:

(16)

The electrode reaction is assumed to be entirely represented by:

12

(4)

3 H

0''

e"

The following changes in number of moles present in the anode compartment

thus occurred per faraday:

Change in Number of Moles

K->0 3

HSO, A

H SO
2

HO
4
2

electrode reaction
ion migration

+1
-

3/2

+ a

chain transfer
total
1

- d
-

+b
a
-

-b
-

+ d -3/2 + d

c -d

Furthermore, electroneutrality requires:

1-c-d

which is in accordance with equation (16).

The equilibrium represented by:

(3)

HO'

'

HSO

"

HO

H S0
2

must exist at all times.

Since the analytical procedure determines the totthe total

al sulfuric present in both the ionized and the unionized forms,

change in moles of sulfuric acid in the anode compartment during electrolysis

is:

(17)

An
q

cA H J>U
2

= 4

q j~(a + b)
>-

b]

qa

where:

the total charge transfer in faradays the number of

Since the electronic charge on the bisulfate ion is one,

equivalents transferred per faraday is equal to

the transference number of the bisulfate ion

a;

this is by definition

[l8J

Therefore the transfer-

ence number of the bisulfate ion is given by:

13

(18)

hso 4 -

r^
\
J
.

According to the analytical procedure used:

/no.

equivalents NaOH

'

"l^SO^

[Vno 8 *c id sample

final

no . equiv. NaOH a
no
.

g.

sample

initial

...

sample wt.

(g) v&/

The analysis produced

^n

by application of equation (19) which,

in combination with the charge transferred,

permitted calculation of the

This trans-

bisulfate ion transference number according to equation (18).

ference number, as shown above, results only from the physical migration
of the bisulfate ion and not from the effective migration resulting from

conductance by the proton transfer chain mechanism.

14

3.

Experimental results.
i

A total of fourteen experimental runs were made.

The first one was

not used because proper precautions had not been taken against moisture,

and it was made with a Hg,Hg 2 SO, cathode which did not function properly

throughout the run.

One was discarded because the samples taken were too

small and the end point was over-run on addition of the 100 ml. of 0.3 M
NaOH.

One run was spoiled by the addition of the 0.8 M NaOH solution be-

fore the base had been standardized; the base was then found to be approx-

imately half strength.

Another run was spoiled by failure to record the

The results of the remaining ten

final weight of the anode compartment.

experimental runs are shown in Table

and are plotted in Fig. 4.

The calculation of the transference numbers involves small differences

between two numbers of nearly the same size.

The major sources of error

would thus be errors affecting the values of these two concentrations.

The

numbers were such that an error in analysis of one part in 15,000 would result in an error of about one percent in the transference number.
eral,

In gen-

the precision of the analyses was such that the accuracy of the analy-

ses was estimated as only about one part in 1,500.

From this it is esti-

mated that the error in the result from analytical errors may be as high as
about 20%, assuming that the combination of errors from the two analyses is

such that the probable error is equal to \J2

times the error resulting

from a single analytical error.

Absorption of atmospheric moisture would be an appreciable source of

error if the analysis did not depend on the difference between the com-

positions of two solutions which had been exposed to essentially the same

atmospheric conditions.

Since the center compartment and the anode com-

partment were in a similar environment at all times and since the acid
15

1 1

6-S

J
-*

go
r^-t

St
cr>

co

CO

CO 00 <t

CO
00 <t

o
co 00

00
1-4

CM
r-4

cm
t-t

r-4

0C+

O vO

00 <t

00
tf

r^

00

co

co

a:

o on o

r^
co
f-i

v>

vO rH

CM
r-t

r-l

o.

o CM
r-l

co -*
r-4

CO <r
r-4

ON
I
f-l

> c r-l 3 O
O*
(0

m O O O o

co

-ci-

o o o o

in

o oo o o

00

CO
r-l

vO ON

i-4

o o o

o o o o

m 00 o o o o

r*

CM

&0

o o o

i-t

o o o o

<Ji-

CO <*
r-l

r-l 1-1

o o o

o o o

r-l

u
c
4J
s-\

-iJ

c
1
i

00 CO CM 00

vO CO

o r^

CM <t

CO

CM

r>

on
00

vO
MT

l>

o r^

r-*

CO

vS
ov

r-.

m ON

o m

r^ o m *t

o o on
r

3
CO
f)
i-i

o
(0

Ci
r-4
Rj

60
*%,

H
r-l

>
H

XI *J

C h s
1-4

w
li (0 i-i

rl <u

3 V 41 o 5 "< a

VO

on CM 00

r-l

o ON o r^ o
T-l

CM CO
1-1

t-l

m
i-i r i-4

00 CM
r-l

CM

vO
r-4

CO vO
<t
r-4

CO
r*

r
o> ft

m m vD
<r

r-4

r*

I-*

ON
t-l

r-i

r-l

1-4

o 00 rH o

r-4

a
X
tl

> vO

CM

vO
^5

r*.

m o r^

CM UO

si-

<f
(-4

*
i-*

CO vO ON

in <t

m
<fr

ON ON 00
t*

-h

CO <J

tj-4

>-*
r-.

O V o a O r-t

O.

CN
r-4

r-l

O CM
CO

CO 00

o
CO

ON

CO vO

m m

m CO m

o o m

1-4

00

CO
00

V ^S M a r-l M c 3 c3 CO o

r>-

ON ON
r-4

vO

O
CO
r-l

t^

o ft

-xt

O 00
CM

m *
CM

r-l

ft

CO

o CO o CO

o m
r* CM

00 CO CM

rH
CO CM

CN
r-l

vO
CM

3
04

CO

<-i

CM

<!

vo

co

on

16

" "- 1 !
,

r-

^.^_ iiijs

OIKby

tir^

te
.

'

Q 1
z;

f - i,

?
'
:

Ll

<D M 2 c
a
1

o
|

T
I

.'

T*

r,

C .1

>

1
';.
>

'

1
n
()

0, 2

0. 4

0. 6

0.8

1.

Mo le
_
',

Frcjtct

ton

ff

^
i

".

F] gu< r^c 4
.
.

ti i
-rj

nil Mil rut. e

on Tr ins fer (*nr :fi

ilum hers

'

'
1

|
.

1
-

3
....
-

H
.

-H

1 ?
i
. i

f-j

compositions were approximately equal, it may be assumed that the absorption

of atmospheric moisture was approximately the same in each compartment.

The

error from this source should then be small in comparison with the other
errors.

A small mass change will occur as

rolysis.

a result of gas evolution during elect-

For oxygen evolution this loss would amount to about 0.2 gram or

about one part in 300.

A small loss of acid, probably of the order of one

This

half gram will occur by physical removal with the electrode assembly.
loss will introduce another error of the order of one percent.

Thermal diffusion errors may have existed but were probably negligible
in comparison with the other errors.

The design of the transference cell

did not permit cooling of the liquid bridges in the thermostat.

Free sulfur

formed in the catholyte and its diffusion through the liquid bridge could be

observed.

In no case was the free sulfur observed to move over one third of

the way up the leg of the bridge;

this indicates that the effect of diffusion

on the results should be small.

Errors from measurement of the average current arise from calibration
of the millivoltmeter and variation of current. The calibration is only

accurate in this case to about one-fourth milliampere; this represents an

error of approximately one percent.
The current varied in distinct steps

with the operation of the thermostat on the water bath; the difference in
current caused by this effect was approximately one milliampere.
This var-

iation shoved up on the recorder trace as short pips.

The current was aver-

aged based on the high value then adjusted downward to correct for the lower

average value.

The errors from this effect are estimated to be negligible

in comparison with the other errors.

In aggregate,
it is estimated that the accuracy of the transference
18

numbers obtained is of the order of

257..

This estimated accuracy is in-

dicated in Fig.

4.

As a first check on the results, Maiden's rule

ing to this rule:

[10J was applied.

Accord-

(17)

^j
X]

constant
ionic conductance of species
J

where:

=
=

7?

viscosity of the solution

This relation was extended to estimate:

(18)

?o

X so H

fA* T^-~
"

where:

jc^=

limiting ionic conductance in water for HSO

50 at 25C.

X H j =

[15]
(?

y
j\

viscosity of pure water = 0.0094 poise

25

TjjgQ,- =

migration transference number of HSO4"

equivalent conductance of solution based on the mean
ionic concentration.

Ar

c^_

1000

= =

specific conductivity of the solution

mean ionic concentration

Concentrations of the ionic species present were determined from the Raman
spectra studies of Young and others [19, 20J.
The results of the appliIt

cation of this principle are shown in Table

2.

is apparent that in

most cases the approximation from the experimental measurements is about half that predicted from the limiting ionic conductance.
However, the fact

that the values are of the same approximate magnitude indicates that the

measured values are probably a reasonable representation of the migration

19

of bisulfate ions in these solutions.

20

Table

Comparison of

70J(

products

HSO
2

1000
4
c

f
[12]

T
4

At ^
.14 .14
.27

0.31
0.31
0.48

11.5

235 235
123 123 123 123
128

.096

11.5
12.3

.096
.195

.27

.14 .13
.11

.27

0.48
0.48

12.3 12.3 12.3

10.3

.195 .195
.195

.25 .21 .20 .28

0.48
0.57

.10 .12

.190
.183

0.61
0.82

9.6 4.1
3.8

131 110

.09
.12
.07

.30 .60
.37

.186
.191

0.84

104

^ AHS04
c+

0.47

calculated from data of Young, Maranvllle and Smith \\&\

on the assumption of HSO," and H-0'" only

21

4.

Conclusions.

The measurements made in this study, by the low value of migration

transference number,

tend to confirm the existence of a chain conductance

mechanism in the middle concentration range for sulfuric acid solutions.

If the hydronium ion is assumed to have approximately twice the mobility

by migration,

the chain mechanism of conductance must account for approxIt is

imately two-thirds of the current carried through the solution.

impossible to distinguish by this method of measurement which chain mechanism may in fact be carrying the current.

Since chain conductance reduces the individual ionic lifetimes,

it

reduces the ability of the ions to conduct by physical migration.

The

mechanisms of chain conductance and physical migration may thus be considered competing mechanisms for ion which may conduct by either a proton

transfer chain or by migration.

A model was developed utilizing the idea

of competing mechanisms to predict the conductivity of sulfuric acid sol-

utions; this model, which is empirical in nature,

II.

is presented in Appendix

In its present state of development,

the model is not satisfactory,

but its qualitative agreement with published conductivities is sufficient

to permit hope that it may provide a basis for development of a better

model for conductance in these solutions.

22

'

BIBLIOGRAPHY
1.

D. Bernal and R. H. Fowler, "A Theory of Water and Ionic Solution with Particular Reference to Hydrogen and Hydroxyl Ions", J_^ Chem.

J.

Phys.
2.

1,

515 (1933).

M.

Eigen and L. de Maeyer, "Hydrogen Bond Structure, Proton Hydration and Proton Transfer in Aqueous Solution", The Structure of Electrolytic Solutions (V. J. Hamer, Ed.), pp. 65-88, John Wiley and Sons, Inc., New York (1959).
R.

3.

J. Gillespie and S. Wasif, "Solutions in Sulfuric Acid. Part IX. The Electrical Conductivity of the Water- Sulphur Trioxide System in the Region of the Composition of Sulphuric Acid, 1953 , J. Chem. Soc. 209-215, (1953).
;

5.

R. J. Gillespie and S. Wasif, Part XI. 'Solutions in Sulphuric Acid. Densities and Viscosities of Some Sulphuric Acid Solutions", J^ Chem. Soc 215-221 (1953).
,

6.

R. J. Gillespie and S. Wasif, "Solutions in Sulphuric Acid. Part XII. Electrical Conductivity Measurements", J^ Chem. Soc, 1953 , 221-231

(1953).
7.
S. Glasstone, K. J. Laidler, and H. Eyring, The Theory of Rate Processes Chapter X, pp. 552-575, McGraw-Hill Book Company, New York
,

(1941).
8.

J. Hamer, "Temperature Variation in Transference Numbers of Concentrated Sulfuric Acid as Determined by the Galvanic Cell Method' J. Am. Chem. Soc 57, 662-667 (1935). W.
,

9.

L.

P. Hammett and F. A. Lowenheim, "Electrolytic Conductance by Proton Jumps; The transference Number of Barium Bisulfate in the Solvent Sulfuric Acid", J. Am. Chem. Soc , 56, 2620-2625 (1934).
S. Karned and B. B. Owen, The Physical Chemistry of Electrolytic Solutions 3rd Ed., Reinhold Publishing Corporation, New York (1958).
,

10.

H.

11.

E. Kunzler and W. F. Giauque, "The Change in Electrical Conductivity of Aqueous Sulfuric Acid Near Absolute H 2 S0^ and H2SO^.H 2 0", J. Am. Chem. Soc , 74, 804-6 (1952).

J.

12. Landolt-Bornstein,
,

Zahlenwerte und Funktionen aus Physik Chemie Astronomie Geophysik Technik Vol. IV, Part 1, Springer-Verlag, Berlin (1955).
, ,
,

13.

W. M.

Latimer and J. H. Hildebrand, Reference Book of Inorganic Chemistry (revised Ed.), The Macmillan Company, New York (1940).
W. Mellor, A Comprehensive Treatise on Inorganic and Theoretical Chemistry (revised Ed.), The Macmillan Company, New York (1940).

14.

J.

23

15.

B.

B. Owen and R. W. Gurry, "The Electrolytic Conductivity of Zinc Sulfate and Copper Sulfate in water at 25", J^ Am^ Chem. Soc. 60 3074 (1938).
,

16.

F.

P.

Treadwell and W.
Vol.
II,

Ed.,
17.

pp.

T. Hall, Analytical Chemistry Ninth English 522-3, John Wiley and Sons, Inc., New York (1945).
,

S. Wasif, "Properties of Sulfuric Acid Solutions. Part I. Transport Number Measurements in Sulfuric Acid and Oleum Solutions", J_^ Chem.

Soc
18.

1955

372,

(1955).
, ,

A. Weissberger

(editor), Techniques of Organic Chemistry Vol I. Physical Methods of Organic Chemistry Part IV, Chapter 46, pp. 3049-3111, Interscience Publishers, Inc. New York (1960).
,

19.

F. Young, L. F. Maranville, and H. M. Smith, "Raman Spectral Investigations of Ionic Equilibria in Solutions of Strong Electrolytes", The Structure of Electrolytic Solutions (W. J. Hamer, Ed.), PP. 48-63, John Wiley and Sons, Inc., New York (1959). T.
,

20.

T.

F. Young and G. E. Walrafen, "Raman Spectra of Concentrated Aqueous Solutions of Sulphuric Acid", Trans Faraday Soc. 57, 34-39 (1961).
. ,

24

APPENDIX

COMPOSITION OF SULFURIC ACID SOLUTIONS

To estimate the concentrations of various ionic species present in

sulfuric acid solutions,

the values of Young, Maranville and Smith

[18J

were plotted in the dilute concentration range.

The values for the high

concentration range were scaled from the graphical presentation of Young

and Walrafen
fl9]

and were plotted and smoothed with the values for the

dilute solutions.

Density data for stoichiometric concentrations were

[l2]
.

taken from Landolt-Bornstein

Values from this plot were then tab-

ulated for intervals of 0.05 mole fraction H SO, in thi6 appendix as

Table 1-1.
In the development of a model for the conductance mechanism in sul-

furic acid solutions a "relative abundance" was used (Appendix II); this
is defined by:
c

(I-D

Z
k
= =

where:

Y.

the relative abundance of species i the molar concentration of species

c.
i

and subscripts

and k indicate any of the molecular or ionic species present in the sol-

ution.

These relative abundances are tabulated in Table 1-2.

25

cj 0)

oCM

Os

OS

en

m
so"

CM

m
en

sO
CM

CO
r-t

CO
iI

00

vl-

en

o CO

CO CM

H 8
i-i

<r
to CM

O
0)

m
* t-t

sO CO

CO

o o

CO
r-\ cj
f CO

in

<t CM

CO
cn

en

OS

<*
r

o
so

CO

sO

CO

cn

cn

CO

DC

m CM
tS)

u
CO

co

^
co

h iH

o CO
X

m m

CM
1 1 1 1

sO

1
1

1 1
1

1
1

** .o
c4

O o H CJ M o

4)

I I

Os

00

r-

<t

in OS

CM

6*
CO

6*
CO

h
r-t

*
co

Os
sO*

00

m o

OS
-

rH

o
CO

r^

<t

00 OS

sO
sO*

en

r4

A U
H

St

4J <X

O CM

CO

OS r*
OS

<r
CM

i-< r-

>ct

m
St CO

m m

<t

CO

OS CO

3 o co

<*
r>

*
St

St <*

Os

<t
CM

in f*
r-l

00 Os CM
r-l

CO OS

sO Os

Os oo

CO r^

sO
CM

CO CM

o CM

m
r-l

O
r-l

JC
H

4J

o O

O CO

<t

CM

CO EC

o o o o o

00
r-l

m
r-*

i-i

vO
CM

Os st

o Os
sO

CM

CO SO
00

St
f-4

so
st
r-

l-l

r-4

sO
CM

m
CO

5
<*

s*
CM

sO
00

co CO

CO CO sO

CM CM

m
o m

m
in lO

sO

r^

J3

*-S
OJ -

O O

r-l

<t O CO
CM

o
r-4

m
r-t

o CM
26

m CM

co

m CO

aj

in <t

sO

"V

o
r

"4

9 u V
r-4

O CM
:.

H
id

9 O V

O CO
CM

<r

vO

r^

rH

CM

CO

o m

r^

vO

vO

CO

3
I

m sm

o>
4-

~\

CJ

o en w

<o <r

CO
co

CO rH

r>.

CM

o CO
CO

m CM
(S;

uX

m m

CO
cj

M
t>
-J-

m m o CO m re
o

2 o M H

CO

4J
i-l

co

*^

rH

CO

CO

<~i

CO

'

rH

rH

V X w

O
I

rC

co

opa
I

1
I

o o
2 O

CJ S3

c
iH
i

K
i

CM

to CM

01

CJ

M 4J d V

<*

cj CM

w
rfl

s
x o
n

o
H

O 4J a

u H

(A
c_>

cj

1
cj*

CO

r-t

I I

<fr

CO

CM

CO cj

oCM oCO o

8CM
S3

O
4)

3
r
iH

c o

oCM

vO CO

o
o

CM rH CM

CM CO

CO

o o

cj*

cj

CJ

09 EC

<t

CO

CM

U3
co 4J CM CO EC

O o

<*

U">

\0 00

CO

o
00

o CO

ii

NO

r^

r^

CO

CO

CO

.fl

-?
4) "-<

H
4J

O O

to

2 K

O to

st

m
r

o CO

m
co

o as

m C\

CM

O
t-t

27

O CO

UO
1

CO
1
1

r-

vO
-l

no

1
1

1 1

1 1

1
l

m EC

CM

0>

in CO

m
CM

<T>

CM iH

o CO

CO

co

iH

CM

as vO \o ft
r-*

m
CM
CM

so CO CO
00 CO

NO

m
CO

<-H

SO
CO CO

vO

1/0 r*.

CO

w u

H ^8
EC

co

CO

NO CO

o CO

CM

CM

M S

3^ 3 H M
H 3

r <
EC

rxo

o
en
oo

CO

m
CM

CT>

o
1-4

CO

00 CM CO

o
M0
CM
CM

en

CO
CM

00 CM

CO

\0 CO

ft
-tf

CO CO

CO

>
r-4

< m

CM

o o o

<f
CO oo

r-t

i-l

vO

A CO

t-<

CM

t-t

W0 ft ft

o
a>

r>*

r-t

CO
CO

\0 v

vO

00 CM

OS
t-t

CO
l-t

M3

O
s
CM
EC

oo
00

cm

r-

vO CM

CO CO

U0
t-I

CO CM
EC?

o CM

in CM

o CO

uo
CO

m vo

o
r-

28

O OT
td

in

m
co

m
vO

00 vO

CM
1

ITi

<

vj-

CO

CM

m m
m

co
in

^8
4-1

CO CM

O CM

o o
c-'-

CM
I

oCO
EC

m
o CM

o
as co

CM CM

CM

m o

m
o

i
i
i

i i

o 10

-*
CM

CO

t-H

o m

<f

as

co vO

as f*

as CO

o o
t-l

EC

^
V
i-l

o CO

<f

m p

o CO

m
CO

o as

m Os

r-i

iSsT

29

APPENDIX II A MODEL OF THE CONDUCTANCE MECHANISM IN SULFURIC ACID

In attempting to explain the mechanisms of conductance in the middle

concentration range of aqueous solutions of sulfuric acid, a model was

developed which permitted reproduction of the conductivity as a function
of concentration to a fair degree.
The model does not,
in its present state

of development, permit accurate reproduction of the conductivity; it also

depends on a number of arbitrary assumptions regarding values of parameters

which may be sufficiently in error to make significant changes in the details of the development.
The basic idea behind this model, however,
is

believed to be sound, and it is therefore presented in the hope that further

study may permit its development in a more accurate form. Since chain conductance depends on an ion or molecule of the proper
sort to participate in the chain mechanism being in the immediate vicinity

of the ion which is transferring or receiving the proton,

the ability for

chain conductance to take place should depend on the relative numbers of

the two species which are present in the solution.

This is the key assump-

tion in the development of this model.

For comparing relative numbers of

components present in the solutions, the "relative abundance" was defiped;

this is a type of mole fraction given by the following equation:
c

(II-l)

Yi

k
where:
Y. 1

relative abundance of species

i i.

molar concentration of species

i

and subscript
ic species,

indicates an individual species from all molecular and ionk,

represented by the subscript

30

present in the solution.

For

the purposes of this model,

acid, molecular water,

the species considered were molecular sulfuric

bisulfate ions,

sulfate ions, hydronium ions, and

=

solvated hydronium ions in sulfuric acid, ie. H3O.H2SO4"

H,SO

'"

[_20j
I.

The values of the relative abundances used are tabulated in Appendix

A further quantity, the

defined to be the
sura

'complementary relative abundance", Y*, is

of the relative abundances of all species which can

In

participate in the proton transfer mechanism with ionic species "j".

this development these complementary relative abundances are given by:

Y* "

+
3

4-

v* "

V
H o S0,

HSO.

Y* H S0
5
5

+ -

Y 1^0

Y*
SO.

--

Y
HSO, 4

Y
H SO,
2

In the development of this model,

the conductance is assumed to be

the result of two competing mechanisms,

ionic migration by bodily motion

The ability of an ion to part-

and chain conductance by proton transfer.

icipate in the migration is reduced by the shorter lifetime of an ion

which results from its participation in a proton transfer.

Eigen and

deMaeyer

f_2 J

cite the lack of spectroscopic evidence for the existence

of the hydronium ion as evidence for a short lifetime of the hydronium ion
in dilute water solutions.

The ability of an ion to participate in a pro-

ton transfer is proportional to the fraction of the ions or molecules in

the solution v/ith which it is capable of exchanging a proton;

that is,

this ability is proportional to Y*

It then follows that the ability of

an ion to participate in the migration should be proportional to 1-Y*.

31

Maiden's rule is assumed to apply to the migration of ions during the

ionic lifetime, hence the following relation is established:

(ii-*)

Xj.
Aj
A

'

*j
=

f'O-Y^
ionic conductance by migration only.

where:

i^f

The value for

is equal to the normally reported values for

ex-

cept in the case of the hydroxide ion and the hydronium ion where the chain

conductance in aqueous solutions at infinite dilution is primarily by the

chain conductance.
The conductivity due to proton transfer by a chain mechanism is assumed to be proportional to the fraction which that species represents of the

total number of particles in the solution; that is, conductivity by the

chain mechanism is proportional to Y

Further, the conductivity result-

ing from chain conductance is proportional to the rate of proton transfer,

which is proportional to Y*.

Gillespie has found evidence that the chain

;

mechanism is viscosity dependent (_6j

although the nature of the viscosthe chain conductance was assumed

It may be

ity dependence is not definitely known,

to be inversely proportional to the viscosity of the solution.

assumed that the conductivity due to the chain mechanism is some function of the total concentration; it was found empirically that the best approximation with the other assumptions was that the conductivity by the chain

mechanism is inversely proportional to the cube root of the total concentration.

The conductivity resulting from the chain mechanism alone by an

individual ionic species is thus given by:

32

(II-3)

Yj

cK "

1?

(% c"7V/3

where:

^' &ha,n

conductivity of the ionic species by the proton

transfer chain mechanism

proportionality constant.

For want of better information, k is assumed to be constant and independent

of the ionic species involved.
The total specific conductivity of the sol-

ution is then given by:

(H-4)

-'

Vj cKoln

\j

/i<?

r.

or

In applying this model, all values for

were obtained from pub-

lished values or estimated.

The published values for the hydronium ion

obviously could not be used since they include both effects of chain conductance and migration.
No information was available for the H_ SOion.

The values used for the computations with this model were:

ion

A
"

HSO

4
4

50

p2,
rioj

15)

h so

"

80
80 40

H
3
H,

(assumed)
(assumed)

.so/
33

The values for the migration terms of the viscosity-conductivity pro-

duct were calculated at intervals of 0.05 mole fraction H SO

2

These val-

ues were then subtracted from the known experimental values of the viscos-

ity-conductivity product.
compared to the values for

The difference between these two values was then

^_ Y.Y*.

Comparison of these curves resulted

in the final selection of the reciprocal cube root concentration dependence

as the best approximation to the desired form for the curve.

The value for

k was determined by calculating, at intervals of 0,05 mole fraction sul-

furic acid,

the value of k required to fit that point.

The values of k

were then averaged and the average value of k was used to calculate an esti-

mated curve.

The results of these calculations are tabulated in Table II-l,

The values of the conductivity and the viscosity-conductivity product esti-

mated by the model are plotted together with the experimental values for

comparison in Fig. II-l.

From the calculated values in the model, it is possible to estimate
the transference numbers by these various mechanisms.

For migration conductance

'

V-

'j

XjO-Y/")
=? Ye*r

(II-6)

T<j
fj

rr\>qr.

where

f"<<^r

transference number of the j-th ionic species

by migration only.

Test
For chain conductance:

conductivity estimated by the proposed model.

(II-7)

chair\
=

where

Ij

chain

transference number of the J-th ion by the

proton transfer chain mechanism.
34

A plot of the estimated transference numbers

values are tabulated in Table II-2.

is shown in Fig.

II-2;

these

This model is at best highly speculative in its detailed form; however,

the general concept of competing mechanisms is believed to be sound The most serious apparent

and possibly subject to further development.

discrepancy in the fit of the model to the observed values is in the concentration range more dilute than the monohydrate.
The difference between

the form of the experimental curve and the curve derived from the model

probably indicates a significant variation between the model and a good

description of the actual mechanism.

The reciprocal cube root dependence of chain conductance on the total

concentration is difficult to justify.

It would seem natural that the con-

ductivity should be proportional to the actual number of charge carriers in

a unit volume;

this is contrary to the form of the concentration dependence

used.

The difference may lie in the assumption that the proportionality

It may also be that

constant, k, is independent of the species in question.

over the concentration range the variation of the total number of charge carriers in a given volume is not significant.
Then the proportionality to

the reciprocal cube root of concentration is equivalent to direct proportion-

ality to the mean distance between particle centers.

It is thus possible

to explain the proportionality to the reciprocal cube root of concentration

as being equivalent to direct proportionality to the mean distance of charge

motion per proton transfer.

The large number of interrelated and unknown variables makes it im-

possible to derive a model which could be justified on the basis of the

conductivity data for sulfuric acid alone; by a trial and error adjustment
of a number of variables it should be possible to achieve a good fit to the
35

known conductivity data without any assurance that the model represents
the actual process.

Further work with other hydrogen bonded solvents

might permit development of a similar model to that worked out in this

case.

Correlation between solvents might then permit

more accurate dev-

elopment for a model of proton transfer conductance.

36

v; v>

vO CM

CM
r-l

CM 00

00

iH CM

\C

00

o>
csi

o CM

CO CM

ON CM

on en

en

on

r*.

00

CM

cn

CM

On NO CM CM

CO

CM
cn CM

ON

cn CM

m CM

vO CM

m CM

cn CM

cn CM

en CM

cn CM

cn ON

1
>v JO

8 m

r* CM
-*

<*
t

r*
CM

si"

cn

o en

r-t

00

ON CM

o on

o s*

On

CM

-* ON cn

o ON
cn

oo 00 cn

&
>
1-1

4J

>
>-

<?
1

Nj

--

1 1

r M
M H
l I-t

-a

s o
i

00

CM

s cn o

00 ON cn

CM 00

cn

cn cn cn

ON

m CM o

CM

cn

CM

CM 00 CM

VO CM cn

o p^ cn

m o o
ct

cn

3 o

CJ

f\f

u.' s"-'

o
1-1

0. iw * etf

O V
(X w
U-l

5?

o
o
c
K|->
\

m
ON

o m

<t n

ON <*

cn

NO
On

00

m
On

VO
h

<-H

vO

co

r*.

on

cn

vo

00

o o

I
i
i

*
r*. r-

vO
r-.

CM 00 n
ON

oo CM

cn

ON
r*.

o r

n
vO

ON
c?>

p*
sa-

in 00

ON en
vO

cn

cn

CM

ON

.2

8 CM a

o CM
37

m CM

o cn

m cn

m <r

o m

m m

vO

r^.

1*
VI

V
k.

\j

vO CM
r-4

-*
i-H

r-

o o

<J\

en vO

CM en

-< CM
Csl

vO

u" r^i-l

cm

O CM

en
r-t

r^

o o o

CM Ov n

-J

O
<4-

r*

CM
-3-

rH ft <t

00
CTi r~<

I I

l-H

>-

<f *

m en <t o

l-l

en en

rW rH <n

r r-t

>-

JO

5
->

en
SSL

00 vO
CM

en

o o
o

en
vO

o en

CM CM

en

IaI-

38

m
r-.

o CO

m
00

o o\

u->
<T>

o o

38

/
I

1
i

( \/

* o ft
\

h*
*

'exper iflSJJLfcl il [14 ]

"
'

1
.

X ft D
4
+

P,

^
^^H

lit

f|.

imat< Hi

>y

model

4->

(I
l

/~~V*
^-,
r-|
r

6 StiiU&l ;ed
"~~~~
!

by model
>

H
>

==

4-i

8 3

* c' 1

4 *t

*f

ex ^exAme. iltii

A ^'" V Vfa
1

U
o
_.

W\
|{

11/
u

A\

V
1

^
\
r
1

V
\\
1

11

2
a;

Vj

i ) r. tr

PEQ

^<

H
1

rj

.2

0. 4
Hoj |.e
1

6
TI

0. 8

1.

h'ra ctf on

IS' 2 S0 4

....

.
'

Fi (P*

't

.
.
.

c OIK ktc Livity Ac COJ "dii *g to Mot lei

r
.

3*
i

^_

id

CO

cm

NO

nO ON

m CO

vO

^3"

00

m NO

nO

co
CO ro

O CM
CO

o in
CM

JC

vO
1 1
1 1 1 1 1

1 1
1

1
1

u
i

O O

NO

o o

m o lO m
EC

u
d 00
l
i

CM

M M
H

08 I-

V
.O

1| z 6
V o+
c j M v

o o

CO

o o

vO

o o

o o

ON

O CO

m m

03

e
at

1 1 1

r^ vO
I*.

o CO
r^

<t

CO vO

ON CO NO

CO

r-t

CO

CO

1^ <f

CM CO

m CO
CM

CO

CO CO CM

ON CO

o CO

u
00

a
+

m CO
o

NO

CM

NO <f rH

CO
CO
r-i

CM CM CM

IT>

r*.

cs

ON CO CO

ON

NO CO

CO CM CO

r* ON
CM

tf

NO
CM

CM CM

co CO
l-l

o CO X

t) r-<

C
00
as

o CM

m CM

o CO

m CO

o
<t-

n <f

o m

m m

m NO

40

<

<

00 c*
<~\
4-t

u~l

U"

<r

O
i-H

f. r

-H

1 1
1

vO

01

o to
sa

m
n

o o

u
G O

u ^-^
^N
1-t

a)

A
2s

P
oO
Nl

jM ^
v O C j w V
M
00

CO

4J

P
t)

\s
CM

f^.

o CO un k

00 CM

f. f>

i-H

1 1
1

<*

M
4>

C
J

ft
^-\

XI

H XJ

CM

o
en
32

o o

1
en

en
*r\

o CM
r-i

r-. r

o <f o

o CM o

1 1 1

t) r-l

O CO

m
r

o CO

m 00

o
CTv

m o>

o o ft

41

'

00 ON

as

CM

IA
< CM

o n CM

o
-d-

CM

St lA CM

CM

St CO

CO

r*. -

o
op o

md

a r^

K
X.
(0

>
I

CM
1

a o

CO

f
4-

CO CO CM

2 CO

in

CM
r*. \r\

op 3

r>.

o CO

CM

CM

CM NO

<7\

r*.

sO
St

NO

CO CM

St CM

CM

O CM

NO NO
i-4

CM CO
F-l

CO

eH

CO IV

CO

CM CM

tT\

CM

NO CO

f-4

r<

CO

CM

* CM

tA
1 1
1
1

t 1

1
1

u
00

1
o
St

st
CO

rv

NO

NO <t

CM CO

o CM o

CO

CO

CO

i I
i

i
i

i
i

I
i

i
i

A
-H tO

o CM

A CM

O CO

uo CO

<f

m
<t

A g

m lA

A NO

O
r*.

*m
SB

42

c o
H

<
o

O 00

CM

m m
00

ON CO

c> CM on

ON CO ON

1 1

JZ

4 V
St 00

ON vO oo

CO

o ON

cs

d ON

1 1

o CO
PC

<t

00
f

3
CM
4J

^-g

m o

o st O

v>

eg

CM
I

ft?

s-'
I I

1 1 1

1 1 1

1
1

1 1

1
1

O CO

St

u
bO
1
1

1 1

1 1 1

1
I

1 1

1
1

1
1

O CO

6-!
4) t-*

O CO

st
ITI

r-

CO

r>

oo

o ON

on

o o
T-4

CM

43

~
1 |

| |
|

~i
1

'

]l:
<

'

.1

..0

_
.:..

J
f

Xo.tal cajtic ii traj laHerd ili acco rolling to Ha tier e

>

k
L

r ' s
i

f~ ^
N
s.

TotaJ
*

aiti an,

by mode

,fc

><

HSC'4"' ;ha in

~\
i

S.
Jt 4>
vj fi

1
.

\
\
r

n
0)

+ ii rt Cllc tin H U ~u
-

J
Lio m-V \

'

1
0)

.4

/
\y
N.

""

<3

H o'
3

^
\l

r.iij ;r.i

H
ri U

//
/

^N

\\

" .

NN
\^

y.
/

\
^s_
6
0. 8
i

y
C

^W^
.;
|

HF>U

mi|gra tiy n

J
f

Jj

M lol< >

F rac tic n

s(

q
r

Ti i

T'l d IS urc - iri

rr-

f
(.

k .,w.i 1 *-! *
1

M
11

^ M r. vv R +

<

ir

1*4

-'
*
"

h j

' 1

S^entoiuansference^nu^oH

6 \ 9768 002 01050 LIBRARY^

DUDLEY KNOX

HitloMlwIIll;