Citation
A fundamental electrochemical investigation of bromoaluminate and mixed chlorobromoaluminate room temperature molten salt systems

Material Information

Title:
A fundamental electrochemical investigation of bromoaluminate and mixed chlorobromoaluminate room temperature molten salt systems
Creator:
Boon, Jeffrey Allen
Publication Date:
Language:
English
Physical Description:
134 leaves : illustrations, charts ; 28 cm

Thesis/Dissertation Information

Degree Divisions:
Department of Chemistry, CU Denver
Degree Disciplines:
Chemistry

Subjects

Subjects / Keywords:
Electrochemistry ( lcsh )
Salts ( lcsh )
Electrochemistry ( fast )
Salts ( fast )
Genre:
bibliography ( marcgt )
theses ( marcgt )
non-fiction ( marcgt )

Notes

Bibliography:
Includes bibliographical references (leaves 130-132).
General Note:
Submitted in partial fulfillment of the requirements for the degree, Master of Science, Department of Chemistry
Statement of Responsibility:
by Jeffrey Allen Boon.

Record Information

Source Institution:
University of Colorado Denver
Holding Location:
Auraria Library
Rights Management:
All applicable rights reserved by the source institution and holding location.
Resource Identifier:
19967711 ( OCLC )
ocm19967711
Classification:
LD1190.L46 1988m .B66 ( lcc )

Full Text
A FUNDAMENTAL ELECTROCHEMICAL INVESTIGATION
OF BROMOALUMINATE AND MIXED CHLORO-
BROMOALUMINATE ROOM TEMPERATURE MOLTEN
SALT SYSTEMS
by
Jeffrey Allen Boon
B.S., University of Colorado, 1986
A thesis submitted to the
Faculty of the Graduate School of the
University of Colorado in partial fulfillment
of the requirements for the degree of
Master of Science
Department of Chemistry
1938


This thesis for the Master of Science degree fay
Jeffrey Allen Boon
has been approved for the
Department of
Chemistry
by
Sandra S. Eaton
Date Z-£>
>


Boon, Jeffrey Allen (M.S., Chemistry)
A Fundamental Electrochemical Investigation of
Bromoaluminate and Mixed Chloro-Bromoaluminate
Room Temperature Molten Salt Systems.
Thesis directed by Associate Professor John A. Lanning
In this project, the chemical and electrochemical
properties of the bromoaluminate molten salt system were
determined. With knowledge of the bromide system, the chemical
and electrochemical properties of the mixed chloro-bromoaluminate
molten salt system were determined, allowing further work on the
applicability of these systems as electrolytes in high density
electrical storage devices.
The qualitative electrochemical behavior of the bromide
and chloride systems were compared. The systems were found to
behave similarly, but the basic bromide system had two oxidative
processes, compared to the single process of the chloride
system. The bromide system was also shown to have a smaller
electrochemical voltage window.
The fundamental electrochemical parameters of the bromide
species were determined for the basic bromide melt system. All
of the parameters, except the diffusion coefficients, compared
well with the literature values of the bromide species solvated
in acetonitrile. The ionic properties of the molten salt system
were shown to effect the diffusion coefficients of the bromide
species, and are approximately two orders of magnitude smaller in
the melt than in acetonitrile.


iv
The bromine formed by the oxidation of bromide was shown to react
with bromide to form tribromide. The rate and equilibrium
constant of the reaction were investigated and discussed.
Although neither parameter was explicitly determined/ both should
be considered when performing electrochemical experiments in
melts containing bromide.
Finally, a preliminary investigation of the mixed melt
system was performed. When MEIB or MEIC was added to a chloride
or bromide melt respectively, the expected results were
obtained. It was observed, however, that a minor addition of
MEIB to a chloride melt dramatically increased the peak current
for the oxidation of the chloride. 27A1 NMR showed the halides
of the tetrahaloaluminate species were in exchange.


DEDICATION
To my parents, Clyde and Eva, who taught me to observe
and question the world around us, and to my wife, Cindy, who is
always so patient with me while I do.


ACKNOWLEDGEMENTS
I wish to acknowledge the assistance of everyone at Frank
J. Seiler Research Labs, but especially the following. Dr. John
S. Wilkes who put up with my endless questions without ever
seeming to become impatient. Mr. Fred Kibler who never
questioned the necessity of any of the stupid things I asked him
to make for me. Mr. John Lloyd Pflug who was always happy to run
samples for me or help me with the instrumentation.
I would also like to thank my research advisor, Dr. John
Lanning for not only being a mentor, but also a friend.


CONTENTS
CHAPTER PAGE
1. INTRODUCTION......................................1
History of Molten Salts.........................1
History of Non-Aqueous Bromide Research.........4
Analytical Problem Definition...................5
2 . CHEMICAL BACKGROUND............................. 8
Synthesis and Purification of Reagents.........8
Molten Salt Definitions........................14
Synthesis of the Molten Salts..................18
Physical Properties.......................... 22
Summary........................................28
3. ELECTROCHEMICAL BACKGROUND......................30'
Description of Apparatus.......................30
Conventions....................................33
Reference Electrodes...........................33
Working Electrodes.............................37
4. DETERMINATION OF FUNDAMENTAL
ELECTROCHEMICAL PARAMETERS.......................40
CV of Selected Melt Compositions...............40
Formal Electrode Potentials....................53
Number of Electrons Transferred................55
Chemical Additions.............................82
Diffusion Coefficients
85


\
.. .\
Vlll
CHAPTER PAGE
Heterogeneous Rate Constant................... 96
5. DISCUSSION OF THE CHEMICAL PARAMETERS............109
Reaction Rate.................................109
Equilibrium Constant..........................113
Summary.......................................116
6. MIXED CHLORO-BROMOALUMINATE
MOLTEN SALT SYSTEMS.............................118
Addition of MEIB to a Chloride Melt...........118
Addition of Two Dissimilar Melts..............122
Summary................. ....................128
BIBLIOGRAPHY.........................................133
APPENDIX 1
130


TABLES
Table PAGE
1. Coefficients for the Calculation
of the Density....................................26
2. Coefficients for the Calculation
of the Viscosity..................................26
3. Number of Electrons Transferred by RDLSV..........60
4. n for the Transfer of Several Different Moles
of Electroactive Species and Electrons............66
5. Summary of the Transfer Coefficient and
the Number of Electrons Transferred...............82
6. Diffusion Coefficient as a Function
of the Melt Composition...........................89
7. Diffusion Coefficient as a Function
of the Melt Composition...........................92
8. Heterogeneous Rate Constant as a Function
of the Sweep Rate
101


FIGURES
Figure PAGE
1. XH NMR of a Chloride Melt Before
and After Chlorination........................... 6
2. 1H NMR and CV Showing Purity of MEIB..............10
3. UV-VIS of Tetramethylammonium Tribromide..........15
4. Anionic composition diagram.......................17
5. The CVs Showing Slow Acquisition of Equilibrium
When AlBr3 is Mixed With MEIB.................... 19
6. Phase Diagrams of Bromide and Chloride Melts......23
7. Density Diagrams of Bromide and Chloride Melts...24
8. Viscosity Diagrams of Bromide and Chloride Melts.27
9. Density Coefficients and the Natural Log of the
Viscosity as a Function of the melt Composition..29
10. Electrochemical Conventions.......................34
11. Reference Electrodes..............................36
12. Working Electrodes................................38
13. The CVs of Neutral Melts..........................41
14. The CVs of a Bromide Melt at Slow and
Fast Scan Rates...................................43
15. The CVs of Slightly Acidic Melts..................45
16. The CVs of Fully Acidic Melts.....................48
17. The CVs of Slightly Basic Melts...................49
18. The Peak Potentials and Limiting Currents
as a Function of the Bromide Concentration........50
19. The CVs of Fully Basic Melts......................52


XI
Figure PAGE
20. Formal Potential as a Function of Sweep Rate....54
21. CV as a Function of Sweep Rate...................57
22. Potential as a Function of ln[(id-i)/i]............61
23. Ep-Ep/2 as a Function of Sweep Rate...............64
24. Tafel plots of an N = 0.4883 Bromide Melt........69
25. Tafel Plots of an N = 0.4688 Bromide Melt........70
26. UV-VIS Calibration Curves..........................74
27. Coulometric Cell...................................75
28. UV-VIS of an Aliquot Removed from the
Coulometric Cell after Passing 35 Coulombs........77
29. Transference Numbers of the Anionic Constituents
as a Function of the Melt Composition............80
30. Addition of Tetramethylammonium Tribromide
to a Basic Bromide Melt...........................84
31. The Limiting Current as a Function of the
Square Root of the Rotation Rate...................88
32. The Current as a Function of the Square Root
of the Inverse of Time.............................91
33. Proposed Structure of a Chloride Melt..............95
34. Peak Separation as a Function of Bromide
Concentration......................................98
35. Peak Separation as a Function of ..................99
36. The Current and the Semiintegral as a Function
of Time for an N = 0.4600 Bromide Melt............103
37. The Potential as a Function of ln[ (nico-m)/m]....106
38. The ln(Q) as a Function of the Potential..........107
39. The CV of an N = 0.4900 Bromide Melt at
Various Scan Rates................................112
40. Potentiometric Titration Curves...................115
41. Addition of MEIB to a Chloride Melt...............119


Xll
Figure PAGE
42. Addition of Two Dissimilar Melts.................124
43. 27A1 NMR of Several Compositions of
Mixed Chloride-Bromide Melts.....................126
44. 27A1 NMR of a Mixed Melt at Various
Temperatures
127


CHAPTER 1
INTRODUCTION
History of Molten Salts
In August 1948, Frank H. Hurley was awarded a patent for
depositing aluminum from a molten salt bath made from ethyl
pyridinium bromide and aluminum chloride.1 This bath was unique
because the molten salt was a liquid at room temperature. In
1951, Hurley and Wier improved the utility of the system by
showing that not only aluminum, but a range of other metals
including silver, copper, nickel, lead, and iron could
2
efficiently be plated from the bath. For approximately the next
fifteen years, the electrochemical and physical properties of the
bath were studied, and the ability to plate out metals, which had
been added, in the form of a salt, to the bath, was improved.
In 1968, the Air Force, through the Air Force Office of
Aerospace Research (AFOAR), began to investigate the feasibility
of using this type of molten salt system as an electrolyte for
3
high density energy storage applications. In 1978, John C.
Nardi, Charles L. Hussey, and Lowell A. King developed a series
of organic substrates including methyl-, ethyl-, propyl-, and
butylpyridinium chlorides and patented the use of molten salts as
4
battery electrolytes. There are several theoretical advantages


2
of using a molten salt instead of a conventional aqueous system
as an electrolyte.
Current molten salt systems are stable liqui'ds over a
much greater temperature range than aqueous systems. Conditions
demanded by the military and encountered in space vary
dramatically. Currently, military specifications require
equipment to operate from -40 to +74 C. Lovell Lawrence Jr., a
space systems designer for the Mercury project, indicates, the
equipment on earth orbiting satellites must be able to withstand
5
temperatures ranging from -65 to +110 degrees centigrade. It is
desirable to find a battery system which is operable in either
extreme. Depending on the composition of the molten salt, liquid
ranges from -90 to over +200 C are possible. Compared to
aqueous systems, which are liquid from approximately -20 to +120
C, it is apparent the molten salt systems would be more
appropriate than aqueous systems for space and military
applications.
Molten salt systems are comprised totally of ionic
species. Theoretically, all of the ions in a melt are capable of
carrying a portion of the current in a cell. In aqueous systems,
the water molecules, which are incapable of carrying current, are
present only to solvate the ions which do carry charge. The
weight of the water is, therefore, non-productive, and decreases
the energy density of the cell. Since the molten salt is
completely ionic, it is theoretically capable of obtaining a
higher energy density than aqueous systems, making smaller and
lighter cells possible. This will, in turn, allow for the


3
production of smaller and lighter vehicles, which are inherently
more energy efficient.
Because water passivates aluminum electrodes, thereby
decreasing the efficiency of the electrode, aluminum electrodes
are impractical for use in aqueous battery systems. Lead is the
principal battery electrode material used in aqueous battery
systems today. Since the molten salt systems do not show a
tendency to passivate aluminum electrodes, the molten salt
systems allow for the use of an aluminum couple at an aluminum
electrode. Changing to an aluminum electrode system will offer
dramatic weight savings, with all of the inherent benefits
explained above.
The use of aluminum as an electrode material not only
increases the energy density of a battery by decreasing the
weight of the cell, but also by increasing the theoretical power
stored in the cell. The standard electrode potential of aluminum
is higher than that of lead. This means the power inherent in an
aluminum couple is greater than the power of a lead couple, so
the energy density of a cell using an aluminum electrode will be
greater than one using a lead electrode. By the same argument,
the power available from an aluminum couple is higher than that
available from a zinc couple, which is typically used in "dry
cell" energy storage systems.
In 1981 the organic constituent of the molten salt was
changed from the substituted pyridinium bromide to a
disubstituted imidazolium chloride.* Many of these systems were
characterized, and it was found that melts made from


4
1-methyl-3-ethylimidazolium chloride (MEIC) and aluminum chloride
had the characteristics most suitable for use as an electrolyte.7
History of Non-Aqueous Bromide Research
In the late 1950s, Alexander I. Popov and David H. Geske,
published a series of papers dealing with the electrochemistry of
halogen and mixed halogen systems. The sixteenth in this series
investigated the electrochemistry of bromide and mixed iodo-, and
g
chloro-bromide systems. Since that time, the use of
electrochemistry, as a means of investigating non-aqueous bromide
systems, has received considerable attention.
It is possible to monitor the progress of bromination
reactions in organic chemistry by following the appearance or
disappearance of the bromide oxidation wave. Desbene-Monvernay
9
et al. used electrochemistry to observe the progress of the
bromination of styrene with tribromide. The concentration of
bromide was followed by rotating disk voltammetry. The limiting
current of a species is proportional to the concentration of the
species. As the concentration of the bromide decreased, the
limiting currents also decreased until all of the bromide was
consumed.
The fundamental work of Popov and Geske has also helped
to illuminate the electrochemistry occurring in the Zinc-Bromine
battery systems. In 1987 Adanuvor et al. studied the effects of
the reaction of bromine with bromide to form tribromide on the
10
electrochemical behavior of a bromine electrode. Mastragostino 12
12
and Daniel J. Eustace have studied the effect of
et al.


5
complexing the bromine on the electrochemistry of a Zinc-Bromine
battery.
Analytical Problem Definition
The most promising battery system to date, employing a
molten salt electrolyte, has been the use of a reversible
aluminum electrode as the cathode and a reversible chlorine
13
electrode as the anode. Recent work at Frank J. Seiler
Research Labs shows that molten salts made from MEIC and aluminum
chloride tend to chlorinate the imidazolium cation in the
presence of chlorine14, according to the reaction:
Figure 1 shows the proton NMR of an MEIC-aluminum chloride melt
before and after bubbling chlorine through it at elevated
temperatures for approximately 72 hours. The pair of singlets at
approximately 8 ppm, which are due to the protons attached to the
number 4 and 5 carbons, disappears as a result of their being
exchanged for chlorines. A singlet appears at approximately 13
ppm which is due to the HC1 formed during the chlorination
reaction. The singlet at 10 ppm can be observed to break up into
a series of three singlets, corresponding to zero, one, and two
chlorine substitutions. This series of three singlets coalesces
back into a singlet after the 4 and 5 positions are completely
chlorinated.
The absorbances due to the two methyls and the ethyl
protons are split into higher multiplets upon chlorination.
H H
Cl
Cl


6
A)
H
4.H
^
CH- >
\
ch2ch,
fl
vl

10
$/ppm
B)

A

/JL

rX.A
UujJ.
10
5
S/ppm
0
Figure 1. 1H NMS of an N = 0.333 chloride melt A) before, and
B) after chlorination.


7
Through integration, it is apparent the peak due to the proton in
the 2 position (at 10 ppm) is less than half of its original
intensity. This indicates the imidazolium ring is also
chlorinating at the 2 position, albeit at a much slower rate.
The tendency of the imidazolium ring to chlorinate makes
the use of the chlorine-chloride couple, in an MEIC-A1C13 melt
electrolyte, impractical. Previous work at Seiler indicates the
14
melts show no tendency towards bromination. The molten salt
formed by combining l-methyl-3-ethylimidazolium bromide (MEIB)
with aluminum bromide (AlBr3), though, tends to decay, through a
process involving the autooxidation of bromide to tribromide,
after several days. For these reasons it was suggested the most
promising system might be to solvate bromine (or bromide) in an
MEIC-A1C13 melt. To understand this system, the electrochemistry
of the pure bromide system must be investigated. This
information will help to understand the electrochemistry
occurring in the mixed chloro-bromoaluminate system and will pave
the way for future studies on the applicability of these systems
as energy storage devices.


CHAPTER 2
CHEMICAL BACKGROUND
Synthesis and Purification of Starting Materials
Acetonitrile. All of the acetonitrile (Fisher, Certified
Reagent) used in these experiments was refluxed over anhydrous
phosphorous pentoxide for at least two days before use.
Nitrogen, which had been dried by passing through a "Dreirite"
anhydrous calcium sulfate column, was continually flushed over
the top of the still. The acetonitrile was removed at the time
of use.
1-methylimidazole. The 1-methylimidazole (Fluka, Puriss)
was vacuum refluxed over, anhydrous barium oxide for at least two
days before collecting by distillation.
l-methyl-3-ethylimidazolium bromide (MEIB). The
l-methyl-3-ethylimidazolium bromide was formed by the
substitution reaction:
H

CHj
H
f".
H
CHjCHjBr
w
CH, V
Br*
CH,CH,
220 mL of 1-methylimidazole were combined with 250 mL of
acetonitrile and 250 mL of ethyl bromide (J.T. Baker "Baker
Analyzed") in a 2 liter round bottom flask. A Teflon stir bar


g
was added to the reaction vessel and a condenser with a drying
tube at the top was put on the reaction flask. The mixture was
cooled with an ice bath for approximately 8 hours, to slow the
reaction down, and then refluxed for four days while stirring.
500 mL of ethyl acetate (J.T. Baker, "Baker Analyzed"
HPLC grade) were added to the reaction mixture and the flask was
shaken continuously until all of the MEIB was recrystallized.
The mother liquor was forced off through a glass tube by applying
a pressure of nitrogen which had been dried with a "Dreirite"
column. The MEIB was dissolved in 250 mL of acetonitrile and
recrystallized with ethyl acetate, as above, for a total of five
times. The resulting MEIB was washed with ethyl acetate a final
time and dried under high vacuum for approximately 12 hours. The
reaction flask was taken into the dry box, and the MEIB was
transferred to a brown glass bottle which had been dried for at
least 12 hours and put into the dry box while still hot.
Since the recrystallization scheme does not remove any
unreacted 1-methylimidazole, care was taken to allow the reaction
to go to completion. The purity of the MEIB was determined by
two methods. Figure 2 shows the 1H NMR of a solution containing
approximately equal parts of MEIB and 1-methylimidazole. The
singlet at 3.7 ppm, the doublet centered at 7.0 ppm, and the
singlet at 7.5 ppm are due to the 1-methylimidazole. The
presence of protonic impurity (typically due to insufficient care
being taken during the preparation of the MEIB, or through the
absorption of water from the atmosphere of the dry box), in a
melt made with the MEIB, was determined electrochemically. The


10
A)
S/ppm
Figure 2.
melt.
A) NMR of MEI3, and B) CV of an N = 0.500 bromide


11
cyclic voltammogram (CV) of a neutral or basic melt, which
contained any protonic impurity, would show an irreversible
reduction wave at approximately +0.100 V on a platinum electrode,
when referenced against a chloride reference cell. The CV of a
neutral bromide melt, containing protonic impurity, is shown in
Figure 2. If there was any question whether the wave was due to
the reduction of the proton, a CV of the same melt was run on a
glassy carbon working electrode. Since a very high over
potential is required to reduce protons at a glassy carbon
electrode, the wave would disappear. Contaminated MEIC was
removed from the dry box and recrystallized as described above.
l-methyl-3-ethylimidazolium chloride (MEIC). The MEIC
was prepared by a modification of the method described by Levisky
is
and Wilkes. 220 mL of 1-methylimidazole were combined with 200
mL of acetonitrile in a 750 mL pressure vessel. 250 mL of ethyl
chloride were added to the solution by passing the ethyl chloride
gas through a dry ice-isopropyl alcohol gas condenser. A stir
bar was added to the reaction mixture and the vessel was sealed
with a green neoprene stopper which was clamped in with a
stopcock puller. The reaction mixture was warmed very gently on
a hot plate while stirring. A blast shield was immediately
placed around the reaction vessel and caution was exercised at
all times due to the relatively high pressures involved. The
reaction was allowed to proceed for two weeks. The reaction
flask was removed from the hot plate, placed in an ice bath, and
allowed to cool for approximately half an hour. The stopper was
carefully removed from the reaction vessel, and the reaction


12
mixture was transferred to a 2 liter round bottom flask, while
streaming dry nitrogen over the mixture. The MEIC was purified
and stored in the same method as described above for MEIB.
Aluminum bromide. The aluminum bromide was purified by
distillation. A thirteen inch long, thick wall, glass tube, with
a one inch inside diameter, was sealed on one end and a male
19/20 ground glass joint was put onto the other. Ten, three
inch, aluminum wires, 1 mm in diameter, (Alfa Products, 99.999%)
were put in the tube. 100 grams of aluminum bromide (Alpha, 98%)
and 0.5 gram of sodium bromide wdre added to the tube. A gas
inlet valve connected to a female 19/20 ground glass joint was
used to seal the tube. The tube assembly was removed from the
dry box and a vacuum was applied for fifteen minutes. The tube
was flame sealed while still under a vacuum and annealed in an
annealing furnace.
The cooled tube was hung vertically in a tube furnace
which had a temperature gradient with the bottom of the furnace
at 220 C and the top at room temperature. All of the aluminum
bromide was allowed to become and remain molten for six hours.
The tube was withdrawn from the tube furnace at a rate of
approximately half an inch a week. The purified aluminum bromide
condensed at the top of the tube, while the impurities remained
at the bottom, as the tube was withdrawn.
The cooled tube was wrapped with aluminum foil and
strapping tape in order to minimize the dissociation caused by
ultraviolet light. When needed, the tubes were brought into the
dry box, all of the tape and aluminum foil was removed, the tubes


13
were scored on each end, wrapped in a cloth towel, and broken
open. The aluminum bromide was collected and stored in a brown
glass bottle which had been dried in the drying oven for 12 hours
and transferred to the dry box while still hot.
During the later stages of the research, the aluminum
bromide was purified by vacuum sublimation. The sublimation was
done outside of the dry box, and after completion, the sublimator
was brought into the dry box, where the purified aluminum bromide,
was collected. The aluminum bromide was sublimed three times
before use. No discernible difference was observed between the
aluminum bromide purified by distillation and that purified by
sublimation.
Aluminum chloride. Aluminum chloride (Fluka, Puriss) was
purified and stored as the aluminum bromide was above, except
sodium chloride was substituted for sodium bromide, and the
sealed tube was withdrawn from the tube furnace at a rate of
approximately an inch an hour. The tube was not wrapped with
aluminum foil and strapping tape, since the aluminum chloride
showed no tendency to photodissociate.
Tetramethylammonium bromide. The tetramethylammonium
bromide (J.T. Baker, Reagent grade) was recrystallized from 95%
ethyl alcohol three times.
Tetramethylammonium tribromide. The tetramethylammonium
tribromide was prepared by the method described by Chattaway and
16
Hoyle. Nine grams of tetramethylammonium bromide were
dissolved in 60 mL of 95% ethyl alcohol. Three mL (9.2 grams) of


14
bromine were added to the solution, and the mixture was boiled
down to a volume of approximately 25 mL. The mixture was allowed
to cool, and the product was collected by vacuum filtration. The
orange needle-like crystals were recrystallized from 25 mL of
ethyl alcohol which had been saturated with bromine. This
recrystallization procedure was repeated a total of three times.
The melting point of the purified product was 119-120 C, which
compares well with the literature value of 118-119 C.17 The
resulting product was analyzed by spectrophotometry.
The UV-VIS spectra of the reaction product, bromine, and
the tetramethylammonium bromide, as well as a mixture of bromine
and the reaction product, were taken with methylene chloride as a
solvent. All of the solutions were prepared immediately before
the spectra were taken to minimize any dissociation which might
occur. These spectra are shown in Figure 3. The spectra of the
product compared well with the spectra attributed by Buckles et
1 s
al. to the tribromide anion.
Molten Salt Definitions
When the MEIB and the AlBr3 are mixed, a set of
dissolution reactions occur. As aluminum bromide is added to the
MEIB, the MEIB is solvated and the reaction:
AlBr 3 + Br + MEI+ ----> AlBr4 + MEI +
occurs. As more aluminum bromide is added, all of the bromide is
consumed, and the aluminum bromide begins to react with the
tetrabromoaluminate by the reaction:


c) i.r
M
1.1 -
t
3
5
3 i.< -
t.i -
lot 5oo 4oi mo ioo ?o# ioo
WRvnliiiylli/nm
Figure 3. UV-VIS spectra of A) reaction product, B) bromine, C) tetramethyammonium bromide, and
D) bromine plus the reaction product.


16
AlBr4 + AlBr3 ---- Al2Br7~
No unreacted aluminum bromide is present in any melt composition,
so a term, called the apparent mole fraction of aluminum bromide
(N), is introduced. The acid-base properties of the system are
usually described by the reaction:
Al2Br7 + Br -----> 2 AlBr4
In a method completely analogous to the chloride
19
system, where the anionic make up of the melt was calculated
assuming the association reactions listed above proceed
completely to the right. Figure 4 shows the anionic composition
of the melt as a function of the apparent mole fraction of
aluminum bromide. For the chloride system, the equilibrium
constant of the acid-base reaction was found to be approximately
. , 13
10i0 , which suggests the previous assumption is justified.
Although the formation constant of the AlBr4~ was not determined
in this work, bromoaluminate molten salts will be assumed to
behave in a similar manner.
There are three compositions of particular interest in
the working range (0.33 <. N 0.66) of a melt made from MEIB and
aluminum bromide. At an apparent mole fraction equal to 0.33,
the bromide and the tetrabromoaluminate are present in equal
concentrations. At an apparent mole fraction equal to 0.50 the
only anionic species present is the tetrabromoaluminate. At N =
0.66 the only anionic species present is the heptabromoalumin-
ate. This allows for the selection of the species present and


MOLE FRACTION ANION
0.0 "-----1------1------'-----L------1-----L------1-----1-------1-----1------1-----L-
0.0 0.1 0.2 0.3 0.4 0.5 0.6
MOLE FRACTION AIBra
Figure 4. Anionic composition diagram of the bromide melt system.
i


18
their concentration in any melt, by changing the amount of the
inorganic salt added to the melt.
Two distinct regions, in the working range of the melts,
are apparent from the composition diagram of the MEIB-aluminum
bromide molten salt system, shown in Figure 4. Any melt with an
apparent mole fraction of aluminum bromide less than 0.50 is
called a "basic" melt. In these melts, the bromide ion acts as a
Lewis base. Any melt with an apparent mole fraction of aluminum
bromide greater than 0.50 is said to be an "acidic" melt, because
the heptabromoaluminate acts as a Lewis acid. A melt with an
apparent mole fraction equal to 0.50 is called a "neutral" melt
because it lacks the Lewis base, bromide, and the Lewis acid,
heptabromoaluminate.
In the pure melt systems, a melt made from MEIB and
aluminum bromide is called a bromide melt. A melt made from
aluminum chloride and MEIC is called a chloride melt. In the
mixed melt systems, all of the fractions will be referenced to
the amount of the chloride melt present.
Synthesis of the Molten Salts
A melt is made by slowly adding the appropriate amount of
the inorganic salt to a known amount of the organic salt. Since
the resulting dissolution reactions are highly exothermic,
approximately 0.20 gram increments of the inorganic salt were
added, and the mixture was allowed to cool between additions.
It was found that, unlike the chloride system, the
dissolution reactions in the bromide melt are appreciably slow.
Figure 5A shows the cyclic voltammogram which resulted from a


A)
B)
D)
E)
Figure 5. CV of an A) N = 0.500 bromide melt at T = 10 min., B) at T = 15 min., C) N = 0,480
bromide melt at T = 25 min., D) at T = 24 hrs., and G) N = 0.480 bromide melt after 24 hrs. h
to


20
bromide melt which was "titrated" to neutrality in a method
20
analogous to that described by Lipsztajn and Osteryoung. The
CV was taken approximately thirty minutes after the initial
addition of aluminum bromide to the MEIB. Figure 5B shows the CV
of the same melt 20 minutes later. The cyclic voltammogram of
the melt which resulted from adding enough MEIB to this
supposedly neutral bromide melt to theoretically result in a melt
with an apparent aluminum bromide mole fraction of 0.490 is shown
in Figure 5C. This CV was taken approximately 25 minutes after
the initial CV was taken. The melt was allowed to stand for
twenty hours and another CV was taken. This CV is shown in
Figure 5D. The melt was "retitrated" to neutrality and enough
MEIB was again added to the melt to obtain an N = 0.490 melt.
The CV of this resulting melt is shown in Figure 5E.
The reduction wave, which appears in Figure 5B, continued
to grow until it became the cathodic limit of the melt. The
decrease in the electrochemical voltage window, and the decrease
in the magnitude of the oxidation waves as time progressed
indicate a slow acquisition of equilibrium conditions. This
indicates either the reaction:
MEIB
MEI+ + Br'
or the reaction:
AlBr3 + Br"
AlBr,-
must be appreciably slow. No attempt was made to quantify the
time required to reach equilibrium, other than to insure


21
equilibrium was reached by allowing the melt to stand for twenty
hours.
Bromide melts were also observed to degrade over time.
Several hours after a bromide melt was made, it started to turn a
distinctive brown color, commonly associated with bromine or
tribromide species. The color became darker as time progressed,
and after several days, the melt appeared black. No
electrochemical consequence of the color was observed, but to
insure knowledge of the system, the melts were not used after
severe discoloration occurred.
Because of the slow kinetics and the problems with
degradation of the melts, a routine was developed for working
with the bromide melts. Melts were made as close to N = 0.50 as
possible without "titration", and allowed to remain at 60 C
overnight. The melts were "titrated" to neutrality, as
determined by cyclic voltammetry, and the amount of MEIB or AlBr3
needed to make the desired melt was added to the neutral melt.
The relatively small additions of MEIB and AlBr3, needed to make
the desired melt composition, did not make a noticeable
difference in the CVs over time.
The melts were used for no more than three days before
being discarded. The discarded melts were destroyed by
hydrolysis. The melts were removed from the dry box and slowly
added to an excess amount of ice. Once the melt was completely
hydrolyzed, it.was washed down the drain.


22
Physical Properties
The electrochemical properties of a species are seriously
affected by the physical properties of the material in which it
is solvated. The rate at which ions are transported to the
electrode double layer affects the observed cell currents and
potentials. The rate a species diffuses through a solvent is
affected by the viscosity of the solvent. Also, the densities
were required to calculate the concentrations of the species.
The physical properties of the chloride molten salt
21
system were determined by Fannin et al. and Hussey et al.
22 2 3
determined the physical properties of the bromide system.
The phase diagrams, densities, and viscosities of the two systems
are presented and compared below.
Phase diagrams. The phase diagrams of the chloride and
bromide melt systems are shown in Figure 6. Although the bromide
system consistently melts 20 30 degrees higher than the
chloride system, it is still properly called a room temperature
molten salt. Both, the bromide and chloride molten salt systems
show a distinct dystectic point at N = 0.500. Both systems also
have several compositions which became glasses instead of
freezing. These glassy transitions were reproducible and were
never made to freeze.
Densities. The densities of several chloride and bromide
melt compositions are shown in Figure 7 as a function of
temperature. The densities of the bromide melts were fitted to
the equation:


Figure 6. Phase diagrams of A) the chloride melt system, and B) the bromide melt system.



pens


25
p = a + bT
(1)
where T is the temperature in degrees Celsius. The coefficients,
a and b, were expanded to the forms
a
+ 3iN + 3aN + dgN
(2)
and:
b = b0 + b]N + b2N2 + b3N3 (3)
The values for the coefficients are given in Table 1.
Viscosities. The viscosities of several melt
compositions as a function of inverse temperature are shown in
Figure 8. Neither the chloride nor the bromide systems exhibited
Arrhenius behavior, but the kinematic viscosity of both systems
followed the Vogel-Tammann-Fulcher (VTF) relationship. The
kinematic viscosities of the bromide system were fitted to the
equation:
In r| = K/(T-T0) + 1/2 In T + In A (4)
where T is the absolute temperature. The parameters K, T0, and A
were fitted by linear regression to the equation:
f = c0 + Cj^N + c2N2 + c3N3 (5)
for melts with compositions greater than 0.30 and less than or
equal to 0.50, and:
f = c0 + Cj/N + c2/N2
(6)


26
TABLE 1 Density Coefficients for Equations 1-3.
0.30 < N < 0.50
X ax bx
0 1.0588 1.2907 x 10-3
l 3.8594 -1.0448 x 10~2
2 -2.8945 1.1519 x 10-2
3 0.0000 0.0000
0.50 < N < 0.75
X ax bx
0 -5.8410 1.7946 x 10-2
l 3.7009 x 101 -9.0089 x 10-2
2 -5.6829 x 101 1.4121 x 10-1
3 3.0474 x 101 -7.4099 x 10-2
Table 2- Viscosity Coefficients for Equations 4-6,
0.35 < N < 0.50
cO C1 c2 c3
To -4.0191 x 101 1.9867 x 103 -4.5874 x 103 2.4562 x 103
K 3.3582 x 103 -2.0888 x 104 4.8630 x 104 -3.3051 x 104
In A -2.8631 x 10,1 1.8526 x 102 -4.4919 x 102 3.4788 x 102
0.50 < N < 0.75
c0 C1 c3
T0 -8.8020 x 102 1.3485 x 103 -4.2335 x 102
K 9.3588 x 103 -1.1072 x 104 3.4117 x 103
In A -3.2068 x 101 3.5633 x 101 -1.0983 x 101


2.5 2.7 2.9 3.1 3.3
io/t/k
Figure 8. Viscosity diagrams of A) the chloride melt system, and B) the bromide melt system.


28
for melts with compositions between 0.50 and 0.75. The values of
the coefficients are given in Table 2.
Summary. In the chloroaluminate molten salt system, the
typical working composition range is dictated by where the system
forms a homogeneous liquid at room temperature (0.333 <. N
0.667). Since the bromoaluminate system requires only a slightly
elevated temperature to become liquid throughout the same range,
the same range will be studied.
As mentioned in the discussion of the anionic composition
diagram, the working range of the molten salts can be divided
into three unique systems. These regions have been termed
"basic", "neutral", and "acidic" depending on the composition of
the melt.
The non-Arrhenius behavior exhibited by the viscosities
is an indication of the nonideality of the molten salt systems.
The nonideality affects, and must be taken into account when
determining, many of the electrochemical properties of the molten
salt system.
Figure 9A shows a plot of the density coefficients as a
function of the composition of the melt. Figure 9B shows a plot
of the natural log of the absolute viscosity as a function of the
melt composition. The phase diagram, as well as the plots shown
in Figure 9, shows a distinct break at N = 0.500 giving support
to the concept of unique physical, and thus electrochemical,
behavior in each region. Subsequent analysis of the bromide
melts will take advantage of this behavior.


ln(vicolly/cp) -b10*/g/cm,C
29
t
Figure 9. The A) Density coefficients, and E) Viscosity at
several temperatures, as a function of the bromide melt
composition.


CHAPTER 3
ELECTROCHEMICAL BACKGROUND
Description of Apparatus
All of the electrochemical experiments were performed in
a Vacuum Atmospheres dry box filled with an atmosphere of ultra
high purity (UHP) grade helium. The atmosphere in the dry box
had a combined water, oxygen concentration of no more than 10
ppm. Oxygen and water were continually scavenged by passing the
atmosphere through a commercial drying train.
The drying train was filled with molecular sieves to
absorb the water, and the oxygen was reduced onto a bed of copper
impregnated ceramic beads called "Ridox". The drying train was
regenerated by heating and passing a mixture of 90% nitrogen and
10% hydrogen through it. The absorbed water is driven off by the
heat, and the reduced oxygen combines with the hydrogen to form
water, which is also driven off. The dry box had a set of two
dry trains one of which was either being regenerated or was ready
for use while the other was being used.
The drying trains are supposed to last indefinitely, but
during the course of the experiments it was found the quality of
the atmosphere inside the boxes degraded faster than expected.
The use of bromine (or bromide, which will auto-oxidize to form
bromine) poisons the molecular sieves and the Ridox in the drying


31
train. When this point is reached, the only course of action is
to change out the molecular sieves and the Ridox. Once this was
discovered, care was taken to minimize the use of free bromine in
the boxes. The Vacuum Atmospheres manual suggests placing a
condenser in the circulation line in front of the drying trains.
This has not been done yet so no comment on the effects of this
idea can be made.
A light bulb which had a hole punched in the glass
envelope was used to indicate the quality of the atmosphere
inside the dry box. Any time the light bulb filament lasted for
less than a week, the dry box was switched over to the unused dry
train and the old dry train was regenerated. The identity of the
contaminant in the dry box could be roughly determined by the
color of the residue left on the glass of the bulb when it burned
out. A bluish film results from a high water concentration, due
to the formation of WO3. A white to yellowish film is formed
when there is a high concentration of oxygen from the formation
of W2O5.
Access was provided to the dry box through a port which
was held under a vacuum for fifteen minutes and then refilled
with the atmosphere from inside of the box. This procedure was
repeated three times. This process continually removes the
atmosphere from inside the box, and any chance impurities, and
replaces it with the UHP helium from the tank.
All of the electrochemical experiments were performed
with a Princeton Applied Research (PAR) model 173
potentiostat/galvanostat equipped with a PAR model 179 coulometer


32
plug-in module. Any potential or current programming was
provided by a PAR model 175 programmer. All of the potentials
were monitored on a DANA model 5330 digital voltmeter. The
working electrodes were mounted in a Pine MSR electrode rotator.
Any scans taken at less than a volt a second were recorded on a
Houston Instruments Model 2000 Omnigraphic X-Y recorder. Any
scans at a volt a second or greater were fed into a Nicolet model
204A storage oscilloscope and then output to the Omnigraphic
recorder. All of the CVs presented in this work were acquired by
initially scanning in the anodic direction.
Since not all compositions of the bromide melts are
liquid at room temperature, a block heater was used to heat the
melts. Temperatures were monitored with a Doric model 412A
Trendicator, using a type K (Chrome1-Alume1) thermocouple.
All of the N.M.R. spectra were run on a JEOL 90Q FTNMR.
All of the molten salts were run neat. All of the proton and
carbon spectra were referenced against a sealed glass capillary
containing hexamethyldisiloxane (HMDS). The capillary caused a
minor degree of line broadening and some small spinning
sidebands, but the spectra were all usable. No reference was
used in the aluminum spectra.
All of the UV-VIS spectra were run on a Hewlett Packard
model 8450A Diode Array Spectrophotometer. Unless otherwise
stated all of the samples were solvated in dry acetonitrile. All
of the spectra were run in 10 mm quartz cells.


33
Conventions
Except where otherwise noted, Electrochemical Society
conventions have been used throughout these experiments. Figure
10 represents these conventions. Potentials become more negative
to the right, and reductive processes become favored. These
potentials are termed "cathodic". Currents resulting from a
reduction process are shown as positive and above the horizontal
axis (typically, quadrant 1). Positive potentials, which are
called "anodic" potentials, are to the left, and favor oxidative
processes. Oxidative currents are negative and towards the
bottom of the graph (typically, quadrant 3).
Reference Electrodes
The purpose of a reference electrode is to provide a
stable potential to which the potential of the working electrode
is compared. Any reference electrode must have a migration of
ions, and thus material, across the interface between the
reference electrode material and the analyte solution.
To date, in the chloride molten salt systems, the
standard reference electrode used has been an aluminum wire in an
N = 0.600 chloride melt. The reference couple is thus the
reduction of the heptachloroaluminate by the half reaction:
4 A12C17- + 3 e -----> A1(s) + 7 A1C14
A fine asbestos fiber separated the reference compartment from
the sample compartment. Any chloride melt which crossed the
interface merely altered the apparent mole fraction of aluminum


34
Figure 10
Electrochemical conventions and nomenclature


35
chloride in the chloride melt being studied. Any chloride melt
which would cross the interface to a bromide melt, though, would
contaminate the bromide melt, with chloride containing species.
Although a bromide reference would not cause
contamination of the bromide melts, the bromide melts suffer from
thermal- and photo-decomposition. The change in the composition
of the reference electrode caused by the decomposition could
result in a drift in the potential of the reference electrode
over time.
In an attempt to detect any leakage from the reference
electrodes, a background cyclic voltammogram, referenced against
a chloride reference electrode, of an acetonitrile solution
containing tetraethylammonium perchlorate (TEAP) as a supporting
electrolyte was taken. The CV of this electrolyte solution is
shown in Figure 11A. Allowing the reference electrodes to remain
in this solution for six hours did not noticeably alter the CV.
Since the reference electrode was removed from a melt when an
experiment was not in progress, during the lifetime of a melt,
the reference electrode would not remain in a melt for more than
a total of six hours.
To facilitate relating potentials referenced against the
chloride and bromide electrodes back to more common reference
systems, the potentials for the reduction of ferrocene, in an
acetonitrile solution containing TEAP as a supporting
electrolyte, against the chloride and bromide reference cells was
determined. Figure 11B shows the CV of the ferrocene system
against the bromide electrode, and Figure 11C shows the CV of the


36
B)
C)
Figure 11. CV of A) acetonitrile with a TEAP as a supporting
electrolyte, B) ferrocene in acetonitrile using a chloride
reference electrode, and C) ferrocene in actonitrile using a
bromide reference electrode.


37
ferrocene against the chloride system. The bromide reference
cell is 0.018 volts cathodic, and the chloride cell is 0.069
volts anodic of the normal hydrogen electrode (NHE).
Both reference electrodes seem to be comparable. Neither
noticeably contaminated the acetonitrile/TEAP solutions and the
voltages read against these electrodes were comparable. The
chloride reference cell was used throughout the rest of the
experiments, because it would allow a direct comparison to a
wider range of previous work, and would avoid problems with an
unsteady reference potential due to the degradation of the
bromide melt. In order to minimize any possible contamination
problem, care was taken to minimize the flow across the interface
by using the smallest asbestos fiber possible and minimizing the
pressure head of the melt in the reference cell.
Working Electrodes
Figure 12 shows the CVs of an N = 0.490 bromide melt on
glassy carbon, tungsten, and platinum working electrodes
respectively. The glassy carbon and tungsten electrodes show
very broad oxidation waves. This indicates the oxidative couples
have a higher degree of electrochemical irreversibility on these
electrode materials. On the platinum electrode, however, the
couples are much more distinct, indicating a relatively minor
degree of irreversibility. Platinum also tends to show the
reduction of any protonic impurities which may be present.
Because protonic impurities are the major source of contamination
of the molten salt systems, the use of platinum will allow for


38
Figure 12. CV of an N = 0.480 bromide melt on a A) glassy
carbon, B) tungsten, and C) platinum electrode.


39
the continual monitoring of the purity of the melts. For these
reasons, it was decided to use platinum through out the rest of
the experiments.


CHAPTER 4
DETERMINATION OF FUNDAMENTAL
ELECTROCHEMICAL PARAMETERS
CV of Selected Melt Compositions
An electrochemist's primary method of gaining a
preliminary understanding of an electrochemical system is by a
technique called cyclic voltammetry (CV). With the information
obtained with cyclic voltammetry, an electrochemist can develop
other experiments to further investigate the system. As noted in
the discussion of the composition diagram, there are several
compositions of the molten salt systems, that have relatively
simple compositions. These compositions were analyzed initially
in an attempt to understand the more complex compositions. The
CVs of the chloride and bromide molten salts are shown for
comparison, but the discussion is limited to the bromide system.
Neutral melts. The only species present in a neutral
bromide melt are the tetrabromoaluminate and the imidazolium
cation. The CVs of a neutral bromide and a neutral chloride melt
are shown in Figure 13. As can be seen there are no oxidative or
reductive processes, of import, occurring other than at the
electrochemical limits of the melts. As the concentration of the
tetrabromoaluminate is decreased, by adding more aluminum
bromide, the anodic limit shifts cathodic, and the cathodic limit


41
Figure 13. CV of an H = 0.500 A) chloride, and B) bromide melt.


42
remains stable. The anodic limit of the bromide melt occurs at
+1.725 V and the cathodic limit occurs at -0.885 V when
referenced against the chloride reference electrode.
The dependence of the anodic limit on the concentration
of the tetrabromoaluminate indicates the anodic limit is due to
the oxidation of some form of the tetrabromoaluminate. The half
reaction ascribed to this process is:
4 AlBr4 ----> Al2Br7 + Br2 + 2 e~
This proposal is supported by high scan rate CV
experiments. Figure 14 shows the CV of a neutral melt with the
potential scanned at both high and moderate rates. At the
moderate scan rate, no waves are observed other than the melt
limits. As the scan rate is increased, a reduction wave, and the
associated reoxidation waves, are observed. The longer the
potential is held at a value greater than the anodic limit, the
larger the reduction wave occurring at +0.175 V, and the
reoxidation waves occurring at +0.425 and +1.075 volts, versus
the chloride reference electrode, become. The reduction wave
shows the same characteristics attributed to the reduction of the
heptabromoaluminate discussed in the section on slightly acidic
melts below.
As the potential enters the anodic melt limit, the
oxidation dictated by the above half reaction occurs. At
moderate scan rates, the heptabromoaluminate completely diffuses
away from the electrode surface before a potential sufficiently
cathodic enough to reduce the heptabromoaluminate is reached. At


43
A)
B)
Figure 14. The CV of a neutral bromide melt at A) 50 mv/sec, and
B) 500 mv/sec.


44
higher scan rates, the heptabromoaluminate does not completely
diffuse away from the electrode surface. Since the
heptabromoaluminate is still in the vicinity of the electrode, it
will be reduced when a sufficiently cathodic potential is
reached. The aluminum is plated out onto the electrode and
subsequently reoxidized, as discussed below.
The cathodic limit is independent of the concentration of
the tetrabromoaluminate. Since the only other species present is
the imidazolium cation, the cathodic limit must be due to the
reduction of the organic cation. No reoxidation of the products
formed in the reduction of the imidazolium cation is seen. This
indicates an electrochemically irreversible reaction as is
usually seen in the destruction of an organic substrate.
Slightly acidic melts. As aluminum bromide is added to a
melt, the melt becomes more acidic. In a neutral or an acidic
melt, such an addition will cause the concentration of
tetrabromoaluminate to decrease and the concentration of
heptabromoaluminate to increase. The species present in an N =
0.510 bromide melt are the same as those present in the neutral
melt plus a relatively minor concentration of the
heptabromoaluminate. The CVs of an N = 0.510 bromide and N =
0.510 chloride melt are shown in Figure 15. The CV appears the
same as that for a neutral melt, but with the addition of a
reductive process occurring at +0.235 volts, and the reoxidations
occurring at +0.350 and +0.670 volts vs. the chloride reference
electrode. The limits occur at the same potentials and are
ascribed to the same processes as for the neutral melt.


45
Figure 15. CV of an N = 0.510 A) chloride, and B) bromide melt.


46
The peak current of the redox couple increases as the
concentration of the heptabromoaluminate increases, indicating
the couple is due to some form of the heptabromoaluminate. The
reduction wave has a characteristic sharp break at the base line
which is commonly associated with a nucleation process occurring
prior to the deposition of a metal. The reoxidation wave does
not follow a typical square root of time decay relationship. The
sharp decay indicates a process which is either passivating the
electrode surface, or is not diffusion controlled. The
combination of these two observations argues that as in the
chloride system, this redox couple is the result of the half
reaction:
4 Al2Br7 + 3 e -------> Al(s) + 7 AlBr4
Two oxidation waves are observed. Both are due to the
reoxidation of the deposited aluminum off of the electrode
surface. The initial wave is due to the oxidation of aluminum
off of the electrode which is deposited on top of a aluminum
monolayer. The second oxidation wave results from the oxidation
of the aluminum which is plated (or bound) to the platinum
surface of the electrode. It is easier to oxidize aluminum off
of aluminum than if is to oxidize aluminum off of platinum.
Fully acidic melts. As the concentration of the
heptabromoaluminate is increased, the peak current, due to the
reduction of the heptabromoaluminate increases until it becomes
the melt limit. This cathodic limit occurs at +0.250 V vs. the
chloride reference electrode. An N = 0.667 chloride melt and an


47
N = 0.667 bromide melt are shown in Figure 16. Since there is no
longer any tetrabromoaluminate present in the melt, the anodic
limit becomes the oxidation of the heptabromoaluminate. The
anodic limit occurs at approximately +1.930 V vs. the chloride
reference electrode. The half reaction of this oxidation process
is:
2 Al2Br7" ----> 2 Al23r5 + Br2 + 2 e
Slightly basic melts. As the concentration of bromide is
increased, by adding MEIB to a neutral melt, two redox couples
begin to appear. Figure 17 shows the CVs of an N = 0.490
chloride melt and an N = 0.490 bromide melt. Figure 18 shows
peak potentials and currents as a function of the bromide
concentration in the melt. The peak potentials shift anodic as
the concentration of bromide increases.
The peak currents increase linearly with respect to the
bromide concentration indicating the oxidative processes are due
to some form of the bromide ion. The fact the currents do not
pass through the origin, in Figure 18, is probably due to the
presence of a small amount of protonic impurity which will react
with the tetrabromoaluminate and the bromide, thereby decreasing
the effective aluminum bromide concentration.
The only species present in a bromide melt are the
heptabromoaluminate, tetrabromoaluminate, and bromide anions, and
the imidazolium cation. If the oxidation waves were due to some
form of the heptabromoaluminate, these waves would also be
present in the CV of an acidic melt. If the waves were due to


48
B)
Figure 15. CV of an N = 0.657 A) chloride, and B) bromide melt.


49
Figure 17. CV of an N = 0.490 A) chloride, and 3) bromide melt.


Paak Curranl/mA Paak Potantla^V
Figure 18. A) peak potential, and B) peak current of the
oxidation waves as a function of the bromide concentration.


51
oxidation of the tetrabromoaluminate, they would also be seen in
the neutral and acidic melts. If the waves were due to the
oxidation of the imidazolium cation, they would be present at all
times, because the imidazolium cation is present in all melt
compositions. For these reasons it was decided the two oxidation
waves must be due to some form of the bromide ion.
Fully basic melts. As the concentration of the bromide
is increased further, the point is finally reached where the
first oxidation wave becomes the anodic limit. Figure 19 shows
the CVs of an N = 0.333 chloride and an N = 0.333 bromide melt.
The cathodic limit, as in a neutral melt, is the irreversible
reduction of the imidazolium cation. The anodic limit of the
fully basic bromide melt occurs at +0.41 V and the cathodic limit
occurs at -1.70 V vs. the chloride reference electrode.
Melt summary. The neutral and acidic bromide melts have
been shown to behave electrochemically in a manner completely
analogous to the chloride system. The basic bromide system,
though, has two oxidative processes, other than the anodic limit,
occurring. All of the evidence, both positive and negative,
indicates the oxidation waves present in the basic bromide melt
system are due to some form of the bromide anion.
While Popov and Geske7 have attributed the first of these
waves to the oxidation of the bromide to tribromide and the
second to the oxidation of the resulting tribromide to bromine,
molten salt systems are different enough from typical solvent
systems to necessitate the verification of these assignments.


52
A)
B)
Figure 19. CV of an N = 0.333 A) chloride, and 3) bromide melt.


53
The rest of this portion of this project is aimed at, first,
verifying these assignments in the molten salt system, and
second, determining the electrochemical properties of the bromide
species in the basic melts.
Formal Electrode Potentials
The potential, halfway between the cathodic and anodic
peak potentials of a process, is called the conditional or formal
electrode potential (E1) of the process. The formal electrode
potential is defined in a manner unique from the standard
electrode potential in order to account for any possible solvent
interactions. The standard electrode potentials which are found
in most tables are usually derived using thermodynamic
considerations. Experimentally, it is usually not possible to
obtain conditions such as activity coefficients of unity. This
non ideality of the experimental conditions can drastically
affect the electrochemical properties of a species.
The potentials, halfway between the anodic and cathodic
peak potentials, for both of the processes are shown as a
function of the square root of the sweep rate in Figure 20. The
potential for the first wave is not dependent on the sweep rate.
The formal electrode potential for the first oxidative process is
0.684 volts versus the chloride reference cell.
For the second wave, the potential halfway between the
cathodic and anodic peak potential is dependent on the sweep
rate. This occurs because of an unequal dependence of the anodic
and cathodic peak potentials, of the second wave, on the sweep
rate. In order to account for this dependence on the sweep rate.




55
the potential was extrapolated back to a sweep rate of 0. This
resulted in a value of +1.01 volts for the formal electrode
potential for the second wave.
Number of Electrons Transferred
Knowledge of the species involved in an electrochemical
process combined with a knowledge of the number of electrons
transferred, per mol of the species, often enables the
identification of the process. Cyclic voltammetry indicated the
two oxidation waves in a basic bromide melt are due to some form
of the bromide ion. The processes occurring in the basic bromide
melts can be identified if the number of electrons transferred in
each process can be determined.
When the heterogeneous rate constant (the rate at which
electrons are transferred from the electrode surface to the
analyte solution) is not facile, the system is said to be
electrochemically irreversible. The degree of electrochemical
reversibility can range from completely reversible to completely
irreversible, with an intermediate degree being called an
electrochemically guasi-reversible system.
The waves of an electrochemically irreversible process
appear very broad and spread out, and are apparent by visual
inspection. While harder to detect than the irreversible case,
an electrochemically guasi-reversible system can be distinguished
from an electrochemically reversible one by observing the
oxidation and reduction potentials of a couple as a function of
the rate at which the potential is varied. Unlike a reversible


56
system, the peak potentials of a wave, .in a quasi-reversible
system, will shift as a function of the sweep rate.
Figure 21 shows the cyclic voltammogram of an N = 0.490
bromide melt at various sweep rates. The potentials of the peaks
are also shown as a function of sweep rate in Figure 21. Since
both of the oxidation waves shift anodic as a function of the
sweep rate, it is apparent the oxidative processes occurring in a
basic bromide melt system are electrochemically
quasi-reversible. This indicates the rate of electron transfer,
from the electrode surface to the electroactive species, must be
taken into account.
To account for the slow rate of electron transfer, an
indication of the degree of irreversibility must be introduced to
the equations describing electrochemical behavior. This is
accomplished by adding a term, called the transfer coefficient
(a), to the equations. The transfer coefficient is a unitless
parameter which modifies the number of electrons transferred (n),
per mol of electroactive species consumed, in the rate
determining step.
Experimentally, there are several methods of determining
the number of electrons transferred in, and of accounting for the
transfer coefficient of, an electrochemical process. Several
techniques, such as rotating disk linear sweep voltammetry,
linear sweep voltammetry, and constant current amperommetry, can
be used to determine the product of the transfer coefficient and
the number of electrons transferred. Other techniques, typically
various forms of coulometry, have been devised which can be used


Figure 21. A) The CV of an N = 0,4800 bromide melt at different scan rates, and B) the peak
potentials as a function of the scan rate.
tn


58
to determine the number of electrons transferred, independent of
the transfer coefficient.
There are two methods of separating each term of the
product of the transfer coefficient and the number of electrons
passed from the other. If the number of electrons transferred
has been determined independently, the transfer coefficient is
easily obtained from the product. If it is not possible to
determine the number of electrons transferred independent of the
transfer coefficient, both terms must be estimated. Since severe
restrictions are placed on the value each term may have, it is
usually a fairly simple matter to assign a value to each.
Rotating disk linear sweep voltammetry (RDLSV). As a
potential is applied to an electrode surface, any species in the
region of the electrode, which is electroactive at that
potential, will be converted by an electrochemical process. As
the species is consumed, diffusion to the electrode surface will
begin to occur. In a voltammetric experiment with no convective
means of mass transport, such as in cyclic voltammetry and linear
sweep voltammetry, the current passing through the cell will
reach a maximum and then decay back towards zero.current as the
electroactive species is depleted from the region surrounding the
working electrode.
If the electrode is rotated, forced convection to the
electrode surface will occur. The induced convection moves the
electroactive species from the bulk of the solution to the
vicinity of the electrode surface, where it will be reduced or
oxidized. At some minimum rotation rate, the rate the


59
electroactive species is brought to the electrode surface,
through a combination of convection and diffusion, will be
greater than the rate it is removed from the solution by the
electrochemical process.
In rotating disk linear sweep voltammetry, the cell
currents reach a limiting value, called the limiting current, as
distinguished from the peak currents observed in experiments with
no convective means of mass transport. When the species is
brought to the electrode surface, by both the convective and
diffusive processes, the measured current is proportional to the
rate at which the electroactive species is transported by each
process. Since, in rotating disk linear sweep voltammetry, the
rate of the convective process is altered in a known manner, the
observed current becomes proportional to the rate at which the
material is brought to the surface by the diffusive process.
If the potential is swept from a non-Faradaic region (a
potential where none of the species present in the analyte
solution are electroactive) to a Faradaic region (a potential
where a current begins to flow through the cell due to the
oxidation or reduction of at least one electroactive species) a
current will begin to flow. For an oxidative process, the
relationship of the limiting current (i^) to the current (i)
observed at any point on the decreasing portion of the I-E curve,
24
as a function of the potential is given by the equation :
E
RT
(l-oc)nF
In
(7)


60
A plot of the potential as a function of In (i^-i)/i
should be linear with a slope of RT/(l-a)nF. Figure 22 shows
this relationship for both of the oxidative processes in an N =
0.480 bromide melt for several different rotation rates. Table 3
gives the values of (l-a)n for both of the processes at the
various electrode rotation rates.
The observed current-potential curve is actually the sum
of two curves resulting from separate anodic processes. The two
waves are not resolved enough to determine the product of the
number of electrons and the transfer coefficient for the first
wave independent of the product for the second wave. Since
current observed for the oxidative process is the sum of the
currents of the two oxidative processes, the current due to the
first process is the difference between the currents observed for
the first and second waves. Experimentally, this is accounted
for by taking the difference between the slopes of the lines
TABLE 3 The Product of the Transfer Coefficient and the
Number of Electrons Transferred as a Function of
the Rotation Rate (u).
(l-a)n
(Cd/S ) Wave 1 Wave 2
300 1.020 0.623
500 0.960 0.573
700 0.918 0.493
900 0.890 0.480

0.947 0.542


61
i
Figure 22. The potential as a function of In [I^-I/I] for A) the
first, and B) the second oxidation waves for several different
melt compositions.


62
obtained when the potential is plotted as a function of
ln[(ij-i)/i], for each wave.
It is usually difficult to distinguish a transfer
coefficient less than 0.3 from a completely irreversible case and
greater than 0.7 from an electrochemically reversible system. If
the degree of irreversibility is observable, 0.5 is a reasonable
initial estimate of the transfer coefficient, as experimental
values of the transfer coefficient typically fall in the 0.3 to
0.7 range. Since a discrete number of electrons must be
transferred in any elementary electrochemical process, the number
of electrons transferred, per mol of reactive species, is
restricted to having one of only a few values.
Even though a discrete number of electrons must be
transferred in an elementary electrochemical process, the number
of electrons transferred, per mol of electroactive species, need
not be an integer (i.e. if two electrons are passed in a process
which consumes three molecules of the electroactive species, n
equals 2/3). After rounding n to a value corresponding to an
integer number of electrons transferred, the transfer coefficient
can be redetermined. Using this process, it was determined one
electron was transferred in each of the electrochemical
processes. With the number of electrons transferred, the
transfer coefficient was found, from the product of the transfer
coefficient and the number of electrons transferred, to be 0.59
for the first oxidation wave and 0.46 for the second.
Linear sweep voltammetry (LSV). Chryssoulakis et al.
have demonstrated how the term (l-a)n can be estimated from the


63
25
slope of the leading edge of a voltanunetric wave. The greater
the electrochemical irreversibility, the broader the wave and
thus the smaller the slope. Figure 23 shows the difference
between the potential of the peak current (Ep) and the potential
corresponding to the current at half the peak height (Ep/2) for
the first oxidation wave as a function of the square root of the
sweep rate. By extrapolating the values for Ep-Ep/2 back to a
sweep rate of 0, the working values were obtained. The average
of the working values for Ep-Ep/2 was +0.082 V.
If the process were electrochemically reversible, the
26
peak separation would be given by the equation s
EpEp/2
2.2 HI
nF
(8)
If the process were electrochemically irreversible, the peak
27
separation would be given by the equation :
1.857 RT
Ep Ep/2 = (i_a)nF
(9)
Equation 8 indicates 0.77 electrons would be transferred per mol
of electroactive species, if the system were electrochemically
reversible. Equation 9 shows the product of the transfer
coefficient and the number of electrons transferred would equal
0.65, if the system were electrochemically irreversible. If it
is assumed a has a value of 0.5, n would have a value of 1.3 for
the electrochemically irreversible case.
There are two methods of evaluating the quasi-reversible
system. First, assuming the oxidative process is actually half


0.0 0.2 0.4 0.6 y2 0.8 1.0 1.2
(Sweep Rale/V/s)
Figure 23. Ep-Ep/2 as a function of the sweep rate for the first oxidation wave.


65
way between the reversible and irreversible cases
(quasi-reversible with a transfer coefficient of 0.5), the number
of electrons transferred can be estimated by the average of the
values for the two cases. This method results in a value of 1.04
electrons transferred for every mol of electroactive species
consumed.
In the second method, a more rigorous approach to the
28
quasi-reversible system is given by the equation:
Ep~Ep/2
A RT
nF
(10)
The term A is a complex term describing the kinetics of the
electron transfer and the concentration of the electroactive
species in the region of the electrode. A has a value of 2.2 in
the reversible system and 3.7 in the irreversible system.
Assuming a mid-range value of 3.0 for A gives a value of 1.05 for
the number of electrons transferred.
In both of these techniques, the actual number of
electrons transferred must be between 0.77 and 1.30 corresponding
to the electrochemically reversible and irreversible cases
respectively. Table 4 lists a series of possible number of
electrons transferred (X) and moles of electroactive species
involved (R), along with the resulting value for n. To obtain a
value of n within the specified range, it can be seen that at
least 4 electrons must be transferred for every 5 molecules of
electroactive species converted, but no more than five electrons
can be transferred for every 4 molecules of electroactive species
converted.


66
TABLE 4 n for Several Different Moles of Electroactive
Species and Number of Electrons Transferred.
X R n
1- 1 1.00
4 5 0.80
5 6 0.83
5 4 1.25
6 5 1.20
As mentioned previously, the degree of electrochemical
reversibility appears to be approximately halfway between the
irreversible and reversible cases. The number of electrons and
molecules of the electroactive species required to give a
reasonable value of n would be quite large, requiring a complex
mechanism to explain the transfer. The simplest explanation,
fitting all of the observations, is the transfer of one electron
for each molecule of the electroactive species consumed,
resulting in a value of 1 for n.
Amperometry. If a current is forced to flow, under
steady state conditions, through a cell, a potential great enough
to oxidize and reduce enough electroactive species to carry the
current will be induced. The difference between the induced
potential and the "rest" potential observed when no current is
flowing is called the overpotential (n)/ and is measured in volts.
The relationship between the induced current and the
overpotential required to carry the current is' described by the
29
Tafel equation :


67
r| = a + b In i
(ID
where a and b are terms which depend on the conditions under
which the experiment is performed. If the experiment involves a
completely irreversible process and high overpotentials, the
. 30
Tafel equation takes on the form:
n
anF
RT
In iQ -
anF
RT
In i
(12)
If the electrochemical process under study is
irreversible, equation 12 indicates, a plot of the natural log of
the applied current as a function of the observed overpotential
will have a linear portion with a slope of anF/RT for the
cathodic, and (l-a)nF/RT for the anodic branch. The
extrapolation of the linear portions to a 0 overpotential gives
an intercept equal to the natural log of the exchange current
(iQ) of the system.
31
Bard and Faulkner describe a method, originally put
forth by Allen and Hickling, for analyzing systems which are
electrochemically quasi-reversible. If the Butler-Volmer
equation is rewritten to the form:
where:
In
l_enfrl
In iQ anfri
(13)
a plot of ln[i/(l-enfrl) ] as a function of the overpotential, a


68
straight line should result. The slope of the line will equal
- overpotentials, and the intercept is equal to the natural log of
the exchange current.
An N = 0.4883 and an N = 0.4688 bromide melt were made.
Each melt was put in an electrochemical cell in which the
platinum working electrode was rotated at a rate of 750 rpm, and
the potential was observed for a series of imposed oxidative
currents.
Figures 24A and 25A show the graphs of the natural log of
the current as a function of the overpotential (called a "Tafel
Plot"), for the first oxidation wave of each melt composition.
As can be seen there are not obvious linear portions of the
curves. This non-linear dependence reinforces the belief that
the oxidation waves in the basic bromide melt system are due to
electrochemically quasi-reversible processes.
Figures 24B and 25B show the plot of ln[i/(l-eI1^rl) ] as a
function of the overpotential (called an "Allen-Hickling Plot"),
for the first oxidative process. In both of these graphs there
is a portion which is linear. The slopes result in a value for
(l-a)n of 0.43 and 0.47, for the N = 0.4883 and the N = 0.4688
bromide melts respectively. Using the method of determining the
transfer coefficient and the number of electrons transferred from
the product, as discussed in the section on linear sweep
voltammetry, it was found that one electron was transferred for
the first oxidative wave. The transfer coefficient was found to
be 0.57 for the N = 0.4883 melt and 0.53 for the N = 0.4688 melt.


69
B)
Figure 24. The A) Tafel plot and B) Allen-Hickling plot for the
first oxidation wave of an N = 0.4883 bromide melt.


70
A)
n/v
B)
Figure 25. The A) Tafel plot and B) Allen-Hickling plot for the
first oxidation wave of an N = 0.4588 bromide melt.


71
Even ideal Tafel and Allen-Hickling plots have a limited
range (usually from 100 to 200 mV of overpotential) of
linearity. Below approximately n = 50 mV, the reverse
electrochemical reaction becomes noticeable, and shifts the
current more negative. Above approximately n = 200 mV, the
observed current is limited by mass transfer effects, causing the
observed current to be less than would be expected by the Tafel
relationships. The oxidation of the bromide to bromine shows a
Allen-Hickling curve whose linear region is limited, not only by
the usual processes described above, but also by the overlap of
the second wave.
Because of the factors limiting the linear portion of the
curve, the position of the line drawn to the Allen-Hickling plot
is suspect. Irrespective of the position of the line, the
linearity of the Allen-Hickling plots confirms the
electrochemical quasi-reversibility of the bromide oxidation.
Although the placement of the line, to the curve, should affect
the calculation of the number of electrons transferred, per mol
of electroactive species, in practice, the constraints put on the
values n may have, allows for a reasonable deviation in the
positioning of the line without affecting n. These
considerations indicate that all of the error, in placing the
line to the Allen-Hickling plot will manifest itself in the
transfer coefficient. Because of this, the values of a,
determined by this method, must be considered suspect.
When there is no net current flowing through the cell,
there is still an equilibrium forward and reverse current flowing


72
at each electrode. At the working electrode, this current is
called the exchange current. When the exchange current is
normalized for the area of the working electrode, it is called
the exchange current density (jQ). The exchange current was
found to be 0.20 mA for the N = 0.4883 melt and 0.19 mA for the N
= 0.4688 melt. This corresponds to an exchange current density
of 0.71 mA/cm2, and 0.66 mA/cm2 for the N = 0.4883 and N = 0.4688
melts respectively.
Coulometery. A method of determining the number of
electrons transferred in an electrochemical step, independent of
the degree of irreversibility, is constant potential coulometry.
In this technique, a potential great enough to induce the
electrochemical process under investigation, is maintained across
a cell containing a known amount of analyte, and the amount of
charge required to convert the analyte is observed.
The amount of charge required to consume a given amount
of the analyte is determined by one of two means. In the first
method, called exhaustive coulometry, a potential is applied long
enough to consume all of the analyte. Once all of the analyte is
consumed, the current will cease to flow. In the second method,
either the formation of the product, or the consumption of the
analyte is followed by an independent analytical technique.
In either technique, the charge passed (Q) is related to
the amount of analyte (N) present in the sample by the equation:
Q = nFN (14)
For the first method, once the amount of charge required to


73
convert a known amount of analyte is found/ it is a simple matter
to determine the number of electrons transferred per mol of
analyte. In the second method, if the amount of charge passed is
plotted as a function of the amount of analyte converted, a
straight line with a slope of nF should result.
Since the products, of the oxidative processes, have been
tentatively identified as tribromide and bromine, UV-VIS
spectroscopy can be used to monitor the progress of the
coulometry. Calibration curves of bromine and
tetramethylammonium tribromide were made, and are shown in Figure
26. Dry acetonitrile was used as the solvent, and 10 mm matched
quartz cells were employed. The bromine was found to have a
maximum absorbance at 398 nm, and a molar absorptivity of 179.
The tribromide ion was found to have a maximum absorbance at 269
nm, and a molar absorptivity of 5110. The linear range of both
was found to acceptable and the degree of linearity was
excellent. The maximum concentration of bromine in the linear
portion of the calibration curve was found to be 15 mM, and the
maximum linear concentration of the tribromide was found to be
0.30 mM.
An N = 0.4979 bromide melt was made and a cell was
constructed as shown in Figure 27. The working electrode was a
platinum mesh and the counter electrode was an aluminum wire in
an N = 0.6061 bromide melt. Care was taken to have the melt
level in the counter cell equal to that in the working cell in
order to minimize the hydrostatic pressure and the resultant
convective mixing. A potential of +1.250 V, referenced against


Abiorbinca Abiorbmci
74
[BrJ/mH

Figure 27. The UV-VIS calibration curves for A) tribromide, and
B) bromine.


75
Figure 27. Coulometric Cell.


76
the usual chloride reference electrode, was applied between the
working and the counter electrodes. This potential is
sufficiently anodic enough to cause both of the oxidative
processes.
Aliquots of the melt were taken at different times during
the course of the experiment. The aliquots were dissolved in
acetonitrile and the UV-VIS spectra of the resulting solutions
were taken. Figure 28 shows the UV-VIS spectra of an aliquot of
melt after approximately 30 coulombs were passed through the
cell. The absorbance at 269 nm is due to the tribromide anion,
indicating the product of one of the electrochemical steps must
be tribromide.
Figure 28 also shows the spectra of the same solution
approximately 15 minutes after the dilution. Obviously, the
concentration of the tribromide observed in the diluted aliquot
decreased over time. Since the standard solutions
(tetramethylammonium tribromide in acetonitrile) were shown to be
stable over time, some component of the melt must allow for the
reaction of the tribromide in solution. For this reason it was
not possible to determine the concentration of the tribromide in
the aliquot and thus a Q vs. N plot could not be constructed.
When the potential is applied to the cell, an oxidative
process is forced to occur in the working compartment of the
cell, and a reductive process takes place in the counter cell.
The bromide in the working compartment of the cell will be
oxidized by a process which will be determined at the conclusion
of the experiment. The most likely cathodic process, occurring


Abtorbanc* Abtorbanc*
77
A)
Wavtltngth/nm
Wavingth/nm
Figure 28. UV-VIS of an aliquot A) immediatly, and 3) 10 minutes
fter being removed from tile coulometric cell.


78
in the counter cell, is the reduction of the heptabromoaluminate
by the reaction:
4 Al2Br7~ + 2 e ---- Al(s) + 7 AlBr4
As the potential is maintained across the cell, a
positive charge will build up in the working compartment of the
cell, due to the consumption of bromide, and a negative charge
will build in the counter cell, due to the formation of
tetrahromoaluminate. In order to maintain charge balance in each
of the compartments, ions must migrate between the two
compartments. Since a positive charge is built up in the working
compartment, and a negative charge is built up in the counter
cell, the charge will be balanced by cations migrating from the
working compartment to the counter cell, and anions migrating
from the counter to the working compartment of the cell.
The only cations present, in the molten salt system, are
the imidazolium cations, thus all of the positive charge will be
transferred by the imidazolium ions. A couple of different types
of anions, though, are present in the counter cell compartment.
The negative charge will be transferred by both the hepta- and
the tetrabromoaluminate. Although the tetrabromoaluminate is
stable in the working compartment, any heptabromoaluminate
reaching the working compartment will react with the bromide ion,
by the reaction:
A12 B rj
+ Br
2 AlBr4~
effectively decreasing the bromide ion concentration.


79
The fraction of the charge carried by the different ionic
species present in the bromide melt, called the transference
number (t) of the species, was determined by Hussey et al.23 It
was found that 76% of the charged is carried by the imidazolium
cation, irrespective of the composition of the melt. The
transference numbers of the anionic species, though, are
dependent on the composition of the melt. This dependence is
shown in Figure 29. Since the melt in the counter cell had a
composition of approximately N = 0.61, graphically it was found
that 10% of the total charge transferred was carried by
heptabromoaluminate anions, and 14% of the charge was carried by
the tetrabromoaluminate.
Correcting for the volumes of the aliquots removed, 46.49
coulombs of charge were passed to exhaustively oxidize the
bromide. Since, at the beginning of the experiment, there were
29.401 gms of an N = 0.4979 bromide melt in the working
compartment of the cell, there were initially 5.3987 x 10-4 mols
of bromide present. Adding 10% to the charge passed, to account
for the migration and subsequent reaction of the heptabromoalum-
inate anions with the bromide anions, 51.14 coulombs of charge
were effectively consumed by the oxidative process. Substituting
these values into equation 14 results in a value of 0.98 for n.
This can be rounded to the transfer of one electron for each
molecule of bromide oxidized.
Summary. All of the techniques gave results consistent
8
with the processes first proposed by Popov and Geske for bromide
dissolved in acetonitrile. The rotating disk linear sweep


Transfer Number
figure 29. The transference numbers of the anionic species as a function of the melt composition.
00
o


81
voltammetry showed one electron was transferred in each of the
oxidative processes. While amperometry and linear sweep
voltammetry were not able to indicate the number of electrons
transferred in the second process, the techniques did support the
idea that one electron was passed in the first oxidative process.
The coulometry indicates not only that one electron is
transferred for the combination of the first and second anodic
processes in the basic bromide melt, but also that the product of
one of the oxidative processes is tribromide ion. Since the
other techniques indicate one electron is transferred for each of
the individual processes, the two processes must be linked
through a chemical reaction. The coulometry indicates the
overall (sum of the two individual processes) process is:
Br ----> Br + e~
which can also be written as:
2 Br- ----> Br, + 2 e
This evidence indicates the couple responsible for the
first wave is:
2 Br
Br2 + 2 e
The chemical reaction:
Br2 + Br ----> Br3
occurs to remove the bromine produced in the first oxidative step
from the melt. The second oxidative process is described by the


82
reaction:
2 Br3 -----> 3 Br2 + 2 e
The number of electrons transferred and the transfer
coefficients are summarized in Table 5 for the different
3 2
techniques. Iwasita and Giordano studied the electrochemistry
of lithium bromide and bromine in an acetonitrile solvent
system. Using a platinum working electrode and a silver-silver
bromide reference electrode, they determined one electron was
transferred in each of the oxidative processes. They also found
the transfer coefficient ranged from 0.44 to 0.52 for the first
wave and 0.49 to 0.63 for the second. The values for the number
of electrons transferred and the transfer coefficients found in
this work agree well with those determined by Iwasita and
Giordano.
Chemical Additions
Since the current observed in a voltammetric process is
proportional to the concentration of the electroactive species.
TABLE 5 Summary of the Transfer Coefficient and the Number of
Electrons Transferred for the Various Techniques
Technique Wave 1 Wave 2
a n a n
RDLSV 0.407 1.0 0.542 1.0
LSV N.A. 1.0 N.A. N.A.
Amperometry Coulometry 0.549 1.0 n = N.A. 1.0* N.A.
^number of electrons transferred for the first and second
processes combined.


83
increasing the concentration of the species will increase the
observed current. To take advantage of this, tetramethylammonium
tribromide was prepared and added to the melt. If, as proposed
above, the second oxidation wave of the basic bromide melt system
is due to the oxidation of the tribromide anion to bromine, the
current of the second wave should increase, upon the addition of
tribromide.
The CV of an N = 0.490 bromide melt is shown in Figure
30. Upon addition of the tribromide anion to the melt, the rest
potential (the potential observed when no current is flowing,
defined by the types and concentrations of the species present in
the system) shifted from -0.053 V to +0.762 V vs. the chloride
reference electrode. The CV of the resulting solution is also
shown in Figure 30. Since the current of the second oxidation
wave increased markedly, this wave is probably due to the
oxidation of the tribromide anion.
Electrochemistry is a notoriously bad technique for the
resolution of electrochemical processes with similar oxidative or
reductive potentials. Since the typical reversible voltammetric
wave is on the order of 50 mV wide at half of the peak height,
any other electroactive species with a formal electrode potential
within approximately 50 mV of the actual species will be
indistinguishable from the analyte species. For this reason,
chemical additions can not be used as absolute proof of the
species involved in an electrochemical process, but it is another
piece of evidence to support the assignments made in the previous
section.


84
Figure 30. The CVs of an N = 0.4900 bromide melt A)before, and
B)after adding tetramethylammonium tribromide.


85
Diffusion Coefficients
An important property in the electrochemistry of a system
is the diffusion coefficient of the electroactive species in the
solvent being used. The rate a species is transported to a point
in space, by the diffusive process, is called the flux of the
species. The flux (J) of a species, at any point in space, is
described by Fick's first laws
3C
J = D ---- (15)
9x
where D is the diffusion coefficient of the species and 8C/8x is
the concentration gradient at that point in space.
Since the diffusion coefficient of a species is a
constant, for any defined solvent system, the flux of the
species, and thus the current observed for the process, is
proportional to the concentration gradient of the species at the
electrode surface. By altering the concentration gradient in a
known manner, the diffusion coefficient is readily obtained.
The area around the electrode, where the only means of
mass transport is through the diffusive process, is called the
diffusion layer. Since, at the outer edge of the diffusion
layer, the concentration of the electroactive species is equal to
the concentration in the bulk of the solution, and the
concentration of the species at the electrode surface can be
controlled to equal zero, it is apparent that altering the
thickness of the diffusion layer will alter the concentration
gradient.


86
Two techniques are typically used to alter the thickness
of the diffusion layer in a well defined manner. In the first
technique, called rotating disk linear sweep voltammetry, the
thickness of the diffusion layer is altered by forcing
hydrodynamic transport of the species towards the electrode
surface. In the second technique, called chronoamperometry, the
thickness of the diffusion layer varies as the region around the
electrode is depleted of the electroactive species.
Rotating disk linear sweep voltammetry. As an electrode
is rotated in the analyte solution, hydrodynamic convection
towards the region of the electrode occurs. Although the
electrode is rotated, the region immediately next to the
electrode surface remains stationary relative to the electrode
surface. The angular velocity of the analyte solution, with
respect to the electrode surface, increases as the distance from
the electrode increases. The faster the electrode is rotated,
the thinner the diffusion layer becomes.
At slow rotation rates, the diffusion layer is relatively
thick, and thus the concentration gradient is small. As the
rotation rate is increased, the thickness of the diffusion layer
decreases, and the concentration gradient increases. As the
rotation rate is increased, the current observed for the
electrochemical process will also increase, due to the increased
concentration gradient.
The current-rotation rate relationship is described by
33:
the Levich equation


87
__ 2/ 3 1/2 -1/6 * , , .
^lim = 0.620nFAD co v Cr (16)
where i]_j_m is the limiting current at any given rotation rate, to,
in rad/s. D is the diffusion coefficient, and Cr* is the
concentration, of the analyte in the bulk of the solution. The
kinematic viscosity of the analyte solution is given by v. If
the limiting current is plotted as a function of the square root
of the rotation rate, a straight line, with a slope of:
m = 0.620nFAD2/3v_1/6Cr* (17)
should result. Any deviations of the plot of the current as a
function of the rotation rate from linearity indicates the
electroactive species is being brought to, or away from, the
electrode surface by a process other than the hydrodynamic
convection induced by the rotating electrode.
Figure 31 shows the limiting current of several bromide
melt compositions as a function of the square root of the
rotation rate. Since the waves are a combination of two
processes, the limiting current of each process is determined by
extrapolating the linear portions of both, the rising portion of
the voltammogram and the observed plateau. The intersection of
the two extrapolations corresponds to the limiting current of the
process. The diffusion coefficients for each species are listed
in Table 6. The plots for each rotation rate show good
linearity, indicating the system is under diffusion control.
Chronoamperometry (ChAmp). When a potential is placed
across a solution containing a species which is electroactive at


Figure 31. The limiting current as a function of the square root of the rotation rate for several
melt compositions.