Professional Documents
Culture Documents
Lead Anodes Electrorefining
Lead Anodes Electrorefining
in
THE FACULTY OF GRADUATE STUDIES
Department of Metals and Materials Engineering
We accept this thesis as conforming to the required standard
In
presenting
degree
at the
this
thesis
in partial fulfilment of
University of
department
or
by
his
or
requirements
for
an
her
purposes
may be granted
representatives.
It
is
permission.
Department
The University of British Columbia
Vancouver, Canada
DE-6 (2/88)
extensive
by the head of
understood
that
publication of this thesis for financial gain shall not be allowed without
Date
advanced
the
copying
my
or
my written
Abstract
In the B e t t s process for lead electrorefining the noble i m p u r i t i e s originally
p r e s e n t i n the b u l l i o n f o r m a s t r o n g a n d adherent layer of s l i m e s . W i t h i n this
layer the e s t a b l i s h e d i o n i c c o n c e n t r a t i o n gradients c a n lead to
reactions. The
from a
secondary
thermodynamic
s h o w e d that: (A) c o n c e n t r a t i o n
[ii]
Table of C o n t e n t s
Abstract
ii
List of Tables
iz
List of Figures
xi
Nomenclature
xz
Acknowledgments
zzv
I Introduction
II M e t a l l u r g y of L e a d
III P l a n t Practice i n L e a d E l e c t r o r e f i n i n g
IV T h e A n o d i c Process
I. I n t r o d u c t i o n
II. T h e p h y s i c a l m e t a l l u r g y of the lead anodes
III. I n d u s t r i a l practice
IV. S l i m e s e l e c t r o c h e m i c a l b e h a v i o r
12
12
12
14
15
V T h e C a t h o d i c Process
I. A d d i t i v e s c o n t r o l a n d electrochemistry
II. S t a r t i n g sheet technology
III. C e l l electrolysis p a r a m e t e r o p t i m i z a t i o n
IV. B i p o l a r r e f i n i n g of lead
22
22
23
23
25
Chapter 2 F u n d a m e n t a l s of the E l e c t r o c h e m i c a l M e a s u r e m e n t
Procedure
I. C o m p o n e n t s of t h e A n o d i c Overpotential
II. T r a n s i e n t E l e c t r o c h e m i c a l T e c h n i q u e s
A. Current interruption techniques
B. A C i m p e d a n c e t e c h n i q u e s
1. A i m s a n d l i m i t a t i o n s of the A C i m p e d a n c e s t u d i e s
26
26
28
28
31
32
Chapter 3: E x p e r i m e n t a l Procedure
34
I. E l e c t r o c h e m i c a l E x p e r i m e n t s
A . E l e c t r o c h e m i c a l cells
1. B e a k e r C e l l
2. R e c t a n g u l a r cell
34
34
34
35
[iv]
B. E l e c t r o d e s
37
1. W o r k i n g electrodes
37
' (a) M a t e r i a l s
37
(b) B e a k e r c e l l
38
(I) P u r e l e a d w o r k i n g electrodes
38
(II) L e a d b u l l i o n w o r k i n g electrodes
38
(c) R e c t a n g u l a r cell
38
(I) P u r e lead w o r k i n g electrodes
39
(II) L e a d b u l l i o n w o r k i n g electrodes
39
2. Reference electrodes
39
3. C o u n t e r electrode
40
C . Electrolyte
41
D. T e m p e r a t u r e c o n t r o l
43
E. I n s t r u m e n t a t i o n
43
1. W e n k i n g potentiostat
43
2. S o l a r t r o n e l e c t r o c h e m i c a l interface a n d frequency r e s p o n s e
analyzer
47
(a) Generalities:
47
(b) D e s c r i p t i o n of the e x p e r i m e n t a l set-up
48
II. Electrolyte P h y s i c o - C h e m i c a l Properties
50
A. p H measurements
50
B. E l e c t r i c a l c o n d u c t i v i t y
50
C. K i n e m a t i c v i s c o s i t y
50
D. D e n s i t y
51
C h a p t e r 4 E l e c t r o r e f i n i n g of L e a d In A Small-Scale E l e c l x o r e f i n i n g C e l l :
A Case Study
I. I n t r o d u c t i o n
IL P r e s e n t a t i o n of R e s u l t s
A . A n o d i c overpotential m e a s u r e m e n t s
1. Stage I
2. Stage H
3. Stage m
B. A n a l y t i c a l c h e m i s t r y
1. B u l k electrolyte c o n c e n t r a t i o n s
2. Inner s l i m e s electrolyte c o n c e n t r a t i o n s
C . C h a r a c t e r i z a t i o n of the s l i m e s layer
1. M e t a l l o g r a p h y of the s t a r t i n g lead anode
2. A n a l y s i s of the s l i m e s layer p h a s e s a n d c o m p o u n d s
(a) S E M a n a l y s i s
(b) X-ray diffraction
C h a p t e r 5 A n o d i c a n d R e s t Potential B e h a v i o r of P u r e L e a d i n
H S i F - P b S i F Electrolytes
I. Overview of P u r e L e a d D i s s o l u t i o n i n H S i F - P b S i F Electrolytes
Under Galvanostatic Conditions
A . A n o d i c overpotential i n the absence of large c o n c e n t r a t i o n
g r a d i e n t s i n the anode b o u n d a r y layer
2
52
52
54
56
56
65
67
69
69
71
74
75
80
80
85
86
86
86
C h a p t e r 6 E l e c t r o c h e m i c a l B e h a v i o r of L e a d b u l l i o n Electrodes i n the
Presence of S l i m e s
I. I n t r o d u c t i o n
II. A C Impedance C h a r a c t e r i z a t i o n of the S t o r t i n g W o r k i n g
Electrodes
A . A C b e h a v i o u r i n the absence of a d d i t i o n agents i n the b u l k
electrolyte
B. A C b e h a v i o u r i n the presence of a d d i t i o n agents i n the b u l k
electrolyte
III. DC a n d AC S t u d i e s i n C o r r o d e d Electrodes
A . S t u d i e s u n d e r galvanostatic, potentiostatic, a n d c u r r e n t
interruption conditions
1. E x p e r i m e n t a l r e s u l t s
(a) V a r i a t i o n of the a n o d i c overpotential as a f u n c t i o n of the
electrolysis time a n d the c u r r e n t i n t e r r u p t i o n time
(b) C h a n g e s i n the i m p e d a n c e as a f u n c t i o n of the s l i m e s
layer t h i c k n e s s a n d of the c u r r e n t i n t e r r u p t i o n time
2. A n a l y s i s of the e x p e r i m e n t a l d a t a
(a) R e l a t i o n s h i p between the DC a n o d i c overpotential a n d
the DC c u r r e n t d e n s i t y
(I) D a t a a n a l y s i s
(b) P r o p o s e d analogue representation of a lead b u l l i o n
electrode covered w i t h a layer of s l i m e s
(I) D a t a a n a l y s i s
(i) C a s e I: i m p e d a n c e s p e c t r a o b t a i n e d i n the
wesence of a net F a r a d a i c c u r r e n t
ii) C a s e II: i m p e d a n c e s p e c t r a o b t a i n e d i n the
absence of a net F a r a d a i c c u r r e n t
115
115
117
118
120
124
124
124
124
133
151
151
153
161
169
170
183
[vi]
C h a p t e r7
P h y s i c o - C h e m l c a l Properties of H S i F - P b S I F Electrolytes
a n d t h e i r R e l a t i o n s h i p to the T r a n s p o r t Processes A c r o s s the S l i m e s
Layer
191
I. I n t r o d u c t i o n
191
II. Average S l i m e s Electrolyte C o m p o s i t i o n
191
III. E h - p H D i a g r a m s
194
A . (F)-Si-H 0 s y s t e m
195
B. (Pb-F)-Si-H O s y s t e m
196
C . (Sb-F)-Si-H O (As-F)-Si-H 0, a n d (Bi-F)-Si-H 0 s y s t e m s
199
IV. P h y s i c o - C h e m i c a l Properties of H S i F - P b S i F Electrolytes
201
A . p H , density, viscosity, a n d activity of H S i F s o l u t i o n s
201
B. D e n s i t y , viscosity, a n d electrical c o n d u c t i v i t y of H S i F - P b S i F
electrolytes
203
V . M a t h e m a t i c a l M o d e l : N u m e r i c a l S o l u t i o n of the N e m s t - P l a n c k
Flux Equations
214
A . C a s e A : c o n s t a n t y, v a l u e s
215
B. C a s e B: c o n s t a n t y v a l u e s i n the presence of eddy d i f f u s i o n .... 2 1 7
C . A n o d i c overpotential v a l u e s derived f r o m the m a t h e m a t i c a l
model
222
2
S
C
R
B
u m m
o n c
e c o m
i b l
a
l
m
i
r y
u s i o n s
e n d a t i o n sf o rF u r t h e rW o r k
o g r a p h y
226
234
238
Appendices
A p p e n d i x 1 M a t h e m a t i c a l M o d e l : N u m e r i c a l S o l u t i o n of the
N e r n s t - P l a n c k F l u x E q u a t i o n s a n d Its A p p l i c a t i o n to the B e t t s
Process
265
A p p e n d i x 2 A n a l y t i c a l S o l u t i o n of the N e r n s t - P l a n c k F l u x
Equations
276
291
A p p e n d i x 4 A n a l y t i c a l C h e m i s t r y of Electrolyte S o l u t i o n s
Containing PbSiF -H SiF
6
312
322
A p p e n d i x 6 C u r r e n t I n t e r r u p t i o n a n d AC Impedance M e a s u r e m e n t s
u s i n g the S o l a r t r o n Devices
330
A p p e n d i x 7 S o l u b i l i t y of P b S i F 4 H 0
334
A p p e n d i x 8 S o l u t i o n of F i c k ' s S e c o n d L a w E q u a t i o n U n d e r C u r r e n t
Interruption Conditions
337
A p p e n d i x 9 E x t e n d e d V e r s i o n of T a b l e s Presented i n C h a p t e r 6
348
A p p e n d i x 10 K r a m e r s - K r o n i g T r a n s f o r m a t i o n s of NUD E l e m e n t s
360
A p p e n d i x 11 P r o g r a m s U s e d to Generate the E h - p H D i a g r a m s
Presented i n C h a p t e r 7
363
[viii]
L i s t of Tables
Chapter 1
T a b l e 1.1 B e t t s lead eleclxorefining i n the w o r l d
10
19
Chapter 4
T a b l e 4.1 C h a r a c t e r i s t i c s of experiment 1 X 2
55
85
Chapter 5
T a b l e 5.1 R e s u l t s of the a n a l y s i s of the c o n c e n t r a t i o n overpotential
i n c r e a s e s d u r i n g the first s e c o n d after a p p l i c a t i o n of the c u r r e n t steps
T a b l e 5.2 V a r i a t i o n of the B
and
96
p a r a m e t e r s w i t h c u r r e n t d e n s i t y .. 107
Chapter 6
Table 6.1 C h a r a c t e r i s t i c s of the experiments presented i n c h a p t e r 6
T a b l e 6.2 Average times r e q u i r e d to m e a s u r e the AC i m p e d a n c e
spectrum
116
117
158
159
159
175
[ix]
179
181
181
184
188
Chapter 7
T a b l e 7.1 C o m p o s i t i o n of the electrolyte s a m p l e s extracted f r o m anode
s l i m e s o b t a i n e d u n d e r i n d u s t r i a l operation o f the B E P (Corninco Ltd.) .. 192
T a b l e 7.2 V a l u e s of the coefficients A a n d Q i n E q u a t i o n 7.17
T a b l e 7.3 C h a n g e s i n the osmotic a n d activity coefficients as a
f u n c t i o n of the Ionic s t r e n g t h
202
203
204
206
212
Summary
Table S. 1 S u m m a r y of i n f o r m a t i o n t h a t c a n be derived f r o m u s i n g
transient electrochemical techniques
233
[x]
L i s t of Figures
Chapter 1
F i g . 1.1
lead
F i g . 1.5 L e a d anode m i c r o s t r u c t u r e
13
F i g . 1.6
15
17
18
+ 3
F i g . 1.11
C o n t i n u o u s p u r i f i c a t i o n of electrolyte v i a p u r i f i c a t i o n c o l u m n
21
22
Chapter 2
F i g . 2.1 E l e c t r o c h e m i c a l cell arrangement
27
Chapter 3
F i g . 3.1 A s s e m b l y u s e d i n the experiments performed w i t h the b e a k e r
cell
35
36
37
40
44
45
47
49
Chapter 4
F i g . 4.1 L e a d anode top view
F i g . 4 . 2 A n o d i c overpotential (uncorrected for
of the s l i m e s layer t h i c k n e s s
54
changes as a f u n c t i o n
57
F i g . 4 . 3 O u t e r a n d i n n e r A, B, a n d C reference electrodes a n o d i c
overpotential response to c u r r e n t i n t e r r u p t i o n s (during a n otherwise
galvanostatic experiment)
58
59
59
F i g . 4 . 6 D e t a i l of the r\ response to c u r r e n t i n t e r r u p t i o n s m e a s u r e d b y
the outer reference electrode (at different s l i m e s layer thickness)
60
61
F i g . 4 . 8 D e t a i l of the T\ response to c u r r e n t i n t e r r u p t i o n s m e a s u r e d b y
the i n n e r A reference electrode (at different s l i m e s layer thickness)
62
F i g . 4 . 9 D e t a i l of the i i response to c u r r e n t i n t e r r u p t i o n s m e a s u r e d b y
the i n n e r B reference electrode (at different s l i m e s layer thickness)
63
F i g . 4 . 1 0 D e t a i l of the T\ response to c u r r e n t i n t e r r u p t i o n s m e a s u r e d
a
65
[xii]
66
68
70
the c u r r e n t i n t e r r u p t i o n time
70
F i g . 4 . 1 7 C h a n g e s i n the l o c a l c o m p o s i t i o n of the s l i m e s electrolyte as a
f u n c t i o n of the m o v e m e n t (from the s a m p l i n g point) of the
a n o d e / s l i m e s interfac
71
F i g . 4.18 C h a n g e s i n the l o c a l c o n c e n t r a t i o n of the total S i a n d F
p r e s e n t i n the s l i m e s electrolyte as a f u n c t i o n of the m o v e m e n t (from
the s a m p l i n g point) of the a n o d e / s l i m e s interfac
72
74
75
76
F i g . 4 . 2 3 L e a d anode m i c r o s t u c t u r e s . A n o d e "A" . A l l m i c r o g r a p h s
c o r r e s p o n d to the s a m e observation point( p o i n t #2 F i g . 4.21)
78
F i g . 4 . 2 4 L e a d anode m i c r o s t u c t u r e s . M i c r o g r a p h s c o r r e s p o n d to
different o b s e r v a t i o n p o i n t s
79
81
82
83
84
[xiii]
Chapter 5
F i g . 5.1 P o t e n t i a l difference between a fixed reference electrode a n d a
c o r r o d i n g anode i n the absence of a d d i t i o n agents
87
89
90
91
92
93
94
96
97
98
F i g . 5.11
of a
99
100
101
F i g . 5.14
c u r r e n t i n t e r r u p t i o n time
F i g . 5.15
102
104
s u b t r a c t i n g the Rg value)
105
[xiv]
107
108
109
110
F i g . 18
112
113
114
Chapter 6
F i g . 6.1 Impedance d i a g r a m (Argand plot) of a t y p i c a l l e a d b u l l i o n
electrode (Exp. CC1-6) u n d e r rest potential c o n d i t i o n s a n d i n the
absence of a d d i t i o n agents
119
120
121
123
125
127
128
[xv]
129
130
132
(From E x p . CA6)
133
134
135
136
137
138
140
141
142
143
145
[xvi]
147
148
149
150
154
155
157
157
164
171
174
177
178
179
[xvii]
182
F i g . 6.43 M o d i f i e d R a n d i e s analogue c i r c u i t
183
184
186
187
189
Chapter 7
F i g . 7.1 S y s t e m (F)-Si-H 0 at 2 5 C
196
F i g . 7.2 S y s t e m (Pb-F)-Si-H 0 at 2 5 C
197
a n d loga
198
pb+2
F i g . 7.4A S y s t e m (Sb-F)-Si-H O at 2 5 C
199
F i g . 7.4B
S y s t e m (Bi-F)-Si-H 0 at 2 5 ' C
200
F i g . 7.4C S y s t e m (As-F)-Si-H O at 2 5 C
200
201
F i g . 7.6 E l e c t r i c a l c o n d u c t i v i t y , density, a n d v i s c o s i t y of H S i F - P b S i F
electrolytes
2
207
213
216
218
219
[xviii]
a n d p ^ x l O " c m sec" )
F i g . 7.12 V a r i a t i o n i n the s l i m e s electrolyte c o m p o s i t i o n a s a f u n c t i o n
of the d i s t a n c e from the s l i m e s / b u l k electrolyte interface, w h e n
c h a n g e s i n activity coefficients a n d i n eddy d i f f u s i o n are a c c o u n t e d for
(CaseB)
4
220
221
222
223
225
[xix]
Nomenclature
a,
a dXi)
pb
+2
+2
interface).
slope, mV (Eq. 5, Chapter 6).
Bj
b,b b
2
Vzc
C
C
+2
t2
rb
C,o
C,
+2
C
CPE!
loo
CPE
D ^2
<Di
e(t)
Econtroi
F
iron.
i\,
+2
[xx]
I
I (t)
&0
Total molar ionic strength = /, = 0.5{Af (v,Z +v^ )+M {y& + VjZ )} = 4x[PbSiF(5]
+ 3x[H SiFe]
2
IiHzSifj
hbsiFj
L
Characteristic length of the electrode, [cm]
m
Slope defined in Eq. 4, Chapter 5, [sec" ^
m
Molality, [mol of solute/Kg of solvent]
M
Molarity, [mol/1]
N,
Water mol fraction with respect to species i.
n
Number of electrons involved in electrode reaction.
IrI, Iy 1
Statistical parameters defined in Appendix 9.
R,
Analogue parameter that represents charge transfer resistances associated
with the lead dissolution processes, [Qcm ].
Analogue parameter related to the DC conductivity of the slimes electrolyte,
[Qcm ]
r
Analogue parameter related to the DC resistance of the slimes electrolyte,
[Qcm ]
r
Analogue parameter related to the charge transfer resistance associated with
the lead dissolution process, [Qcm ]
Rb
Resistance of the electrolyte present between the reference electrode and the
anode boundary layer, [Clem ]
R,
Charge transfer resistance, [Qcm ]
RD
Diffusional (DC) resistance, [Qcm .]
Rata
Resistance of the film created by the addition agents, [Qcm ]
Rm
"Apparent" average resistivity of the electrolyte present across the slimes
layer, [Qcm ].
Rp
Polarization resistance, [Qcm ]
Rs
"Uncompensated" ohmic resistance, [Qcm ].
R
Universal gas constant, [8.3114 J mol" deg"]
s,
Stoichiometric coefficient in electrode reaction
T
Absolute temperature, [K]
0
v and v_ Number of cations and anions into which a mole of electrolyte dissociates.
+
w
x,
[xxi]
x
Z,
Z(jco)
Zj
Z*
IZI
Z
Z
Z
Xjotai
a / s e
a/sl
CPE
2/lBC
ZnJt)
Z
Z
sl/be
sI/se
Zw
[M]
[M ]
[M ]
{M }
fn
+n
+n
|= -N/ZI+Z )
2
I
Total molar ionic strength = 4x[PbSiF6l + 3x[H SiF6]
[HaSiFglu H SiF concentration at the total molar ionic strength of H SiF -PbSiF
mixtures, [M],
[PbSiFg] + [HaSiFJ PbSiF and H SiF concentrations in the electrolyte rnixtures, [M],
[PbSiFelK PbSiF concentration at the total molar ionic strength of H SiF -PbSiF
mixture, [M].
t
[xxii]
y,
AO
O
<J> 0
^zc
A ^
(1
A H ^ and ApK^ Equivalent conductivity of the pure H SiF and PbSiF solutions at
the total ionic strength of H SiF -PbSiF mixtures, [cm eq" CI' ]
X,
Individual equivalent conductivity of ions i, [cm eq' Ci' ]
2
T)
r)
n
r|
rj^,
Total ohmic drop across the slimes layer, [mV] (Eq. 32, Chapter 7)
r\
ru
K^fc
KiPbsiFj,,
ac
TI
K
K [x^
+2
ov
x
t
x
K_
H,
v
p
05
[xxiii]
0)
min
max
Abbreviations
AAS
AC
BEP
cd
C.P.V.
CPE
DC
E.M.F.
EPMA
FFT
FRA
PCR
RM.S.
SACV
SEI
SSM
SEM
SSR
[xxiv]
Acknowledgments
I want to express my sincere appreciation to my research advisor, Dr. Ernest
Peters who guided me constantly throughout my stay at U.B.C. I want to thank
him for all his teachings, and most of all for having the persistence of bearing
with me.
My stay at U.B.C. was made possible due to scholarships from; Universidad
National Autonoma de Mexico (U.NJLM.) and The University of British Columbia,
and through supplementary funding from the Natural Sciences and Engineering
Research Council of Canada. I want to thank these institutions for their support.
I want to express my gratitude to Dr. R.C. Kerby (Corninco Ltd.) and Dr.
C.C.H. Ma (Dept. of Electrical Engineering, U.B.C.) for providing me with a variety
of new ideas. My sincere thanks to Dr. Charles Cooper for taking the task of
proofreading the thesis and contributing with constructive comments.
I want to thank my beloved wife, Diana, whose positive and encouraging
attitude gave me the strength to accomplish this work.
I also would like to thank my parents and brothers for giving me moral
support throughout this work.
[xxv]
Introduction
T h e B e t t s process for lead electrorefining treats lead b u l l i o n c o n t a i r i i n g 1-4%
i m p u r i t i e s . A n t i m o n y , arsenic, b i s m u t h , a n d a variety of m i n o r metals i n c l u d i n g
silver a n d gold are a m o n g s t these i m p u r i t i e s . T h e b u l l i o n i s cast into anodes of
a b o u t 1 m , w e i g h i n g between 2 0 0 a n d 3 0 0 k g . These anodes are electrolytically
2
electrolyte
1 If bismuth does not corrode with lead it would enrich at the anode surface until lead corrosion is
stopped.
[1]
The single, most important parameter to study was found to be the ionic
concentration gradients in the electrolyte entrapped within the slimes layer. One
of the direct effects of the presence of these concentration gradients is the
precipitation of secondary products within the slimes layer. The characterization
of these secondary products and their effect on the dissolution of noble compounds
was studied by several methods. Among these, transient DC (direct current) and
AC (alternating current) electrochemical techniques were used to find the extent
to which precipitation of secondary products and entrapped electrolyte
concentration gradients affect the refining cycle. The electrolyte composition
within the slimes and the accompanying potential gradient were obtained by
mcorporating tap holes in typical lead anodes. The study and characterization of
this entrapped electrolyte required the use of several analytical procedures,
including titTimetic analysis and atomic absorption spectroscopy. The
characterization of the physico-chemical properties of this entrapped electrolyte
was achieved by measuring the electrical conductivity, viscosity, and density of
synthetically prepared solutions whose compositions approximated those of inner
electrolytes as found from anode tap holes or calculated from theory. In addition
to this, sampling of the slimes layer and characterization of the contained
secondary products was an important part of this work. Scanning electron
microscopy and X-ray diffraction techniques were employed to detect the phases
and elements in the slimes layer.
A mathematical model based on the Nernst-Planck flux equations was
developed to describe the establishment of concentration gradients within the
slimes layer. This model predicts these gradients when combined with
relationships between concentrations and fundamental solution properties (i.e.
activities, mobilities, diffusion coefficients). The full application of this model will
occur only when more experimental data on these fundamental properties become
available.
To explain the presence of secondary products and their stability range from
a thermodynamic perspective, computer generated Eh-pH diagrams were drawn.
[2]
C h a p t e r 1 L i t e r a t u r e Review
I Introduction
This chapter deals mainly with the Betts electrorefining process for lead, as
described in the literature. In this process lead bullion is purified by transferring
most of the lead from a soluble anode to a cathode through a lead-containing
electrolyte while leaving behind impurities in an adherent anode slime. The
electrochemistry of this process involves the properties of anode slimes and of
entrained electrolyte. Thus, the processes of anodic corrosion of lead and transport
of lead ions through the entrained electrolyte are essential to the understanding
of the Betts process. In this chapter, the physical metallurgy of lead anodes, which
affects the slimes adherence is also discussed.
n M e t a l l u r g y of L e a d
Lead is an ancient metal. It was used by the Romans for components of their
water distribution systems, and for that purpose it had to be malleable and ductile
IH. The ancients made lead of acceptable purity probably by smelting lead ores
under conditions that prevented arsenic, antimony, and other hardening elements
from reducing to the metallic phase. This could be accomplished in most cases
by producing high lead slags such as those still produced in fire assaying. In those
days it was necessary to avoid excessive copper in the ore, but arsenic, antimony,
bismuth, nickel, iron, etc. were reliably held in the slag by maintaining the high
oxidizing conditions of lead silicate based slags. Lead recoveries were low - not
better than 85% from the highest grade hand picked galena ore. When high lead
recoveries were found to be obtainable by coke-based blast furnace reduction,
lead so produced was too hard, usually because of its copper, antimony, and
arsenic content. The function of lead refining became both a softening process
and a method of recovering silver and gold [2].
Nowadays the extraction process can be conveniently portrayed in the
two-step flowsheet shown in Fig. 1. The first step involves bullion production from
the sulphide concentrate and the second step the refining of bullion to the final
product.
The conventional route for bullion production requires sintering of the
concentrate to produce a lead oxide containing product, which is then reduced
in a blast furnace with metallurgical coke to produce lead bullion. The KTVCET
[3]
Metallurgy of Lead
and QSL processes represent two relatively recent commercial developments 131
that replace both the sintering - blast furnace combination with a single furnace
that treats concentrates, and reduce both the costs of lead smelting and the
environmental impact.
The KIVCET [4] process replaces the sintering/blast furnace operations with
a flash smelting step. In this process, lead sulphide is oxidized to lead bullion and
sulphur dioxide in a stream of oxygen. In a second step, the bullion and slag flow
under a weir to an electrically heated settling hearth where coke breeze or coal is
added to reduce the residual lead oxide in the slag and produce a final bullion
(for refining) as well as a low-lead slag.
Lead Sulphide
Concentrate
Bullion
Lead B u l l i o n
Production
Sulphur
Dioxide
Fig. 1 Flowsheet for Lead
ExtractionfromSulphidic
Concentrates.
Refining of
Bullion
Residues
99.99+ %Pb
Byproduct
Metals
[4]
Metallurgy of Lead
T
DROSSING
COPPER DROSS
OR MATTE
1
T
SOFTENING
ANTIMONIAL SLAG
Fig. 2
[Sb]
ZINC-SILVER
CRUST
T
DEZINCING
METALLIC ZINC
<C.Mg)
IZnJb)
T
FINAL REFINING
component
[] brackets indicate a minor impurity
(Bi)
DEBISMUTHISING
L e a d [6].
0 brackets indicate a major impurity
(ZnS)
IS>I
T
DESILVERISING
component
BISMUTH DROSS
CAUSTIC DROSS
T
MARKET LEAD
> 99.99% Pb
1 Fig. 2 was taken as is from the literature and does not contain inputs required for material balances.
[5]
Metallurgy of Lead
Decopperized lead
Casting kettle
73_J
Starling
g sheel
^ m machine
Starling sheet
I Electrolytic cell
Electrolyte
Scrap anode
Deposit cathodes
Mechanical scraper I
Scrap
Filtrate-
Wa
Cenlrituge
Slime
1
1
Pig lead
1 Nitric acid media are not suitable because, in the presence of free acid, nitrate is reduced (to nitric oxide
gas) at the cathode, preferentially to plating of lead. At higher pH, where the nitrate ion is much more inert
to reduction, the electrical conductivity of the electrolyte is much too low for an economic practice.
[6]
HI P l a n t P r a c t i c e i n Lead E l e c t r o r e f i n i n g
A detailed description of the Betts Electrorefining Process (BEP) can be found
in Betts' book [271 and numerous patents [28-32]. The fundamentals of the process
described there remain applicable to all the plants which electrorefine lead
nowadays.
The BEP process is used in Canada [33-36], China [22,37,38], East Germany
[39], Italy [40-42], Japan [43-51], Peru [52-541, Rumania [55.56], Russia [57-591, U.S.A.
[60] and West Germany [61-63]. The average annual production of lead by BEP is
approximately 1,000,000 tons. Since the production of refined lead in the
non-socialist countries (for which good figures are available) is close to 4,700,000
tons per year [64], it can be assumed that up to about 20% of the world lead
production is refined by the BEP. Two variations of the BEP, the sulphamic acid
and the fluoboric acid processes are operated in Italy 141] and in West Germany
[61] respectively.
The Betts Electrorefining Process normally utilizes a fluosilicic acid (HaSiFe)
electrolyte containing lead fluosilicate to electrorefine impure lead anodes into
pure lead cathodes. The fluoborate and the sulphamate processes are identical
to the fluosilicate process except for the substitution of the electrolyte. Fluoboric
acid has not been used extensively due to its relatively high cost. HBF is a stable
acid with a good electrolytic conductivity and a high solubility of the lead salt. It
is also used in lead plating baths [65-67], lead fluoborate-fluoboric acid
rechargeable batteries and in the recovery of lead from spent batteries [68-72].
4
The sulphamic acid process has two main limitations when compared to the
fluosilicic acid process [40,58,73]: firstly, sulphamic acid decomposes rapidly at
current densities larger than 100 Amp/m resulting in high reagent replacement
cost; secondly, the free acid is a crystalline solid of limited solubility, and limited
ionization in aqueous solution, leading to a low conductivity of sulphamic acid
solutions (at most, half of the conductivity of equivalent fluosilicic acid solutions).
This results in larger power costs in refining. The significant advantage of the
2
1 Commercialization of any new lead processes faces a lack of need for plant expansion in this industry
and so must be justified on the basis of conversion or replacement of existing capacity.
[7]
sulphamic acid process is that it is the most efficient process for the removal of
tin from impure lead (Sn remains in the slimes) [74-77]. Plants that use the
fluosilicic process remove Sn prior to electrolysis by using the Harris process 1781.
When anodes containing significant Sn concentrations are refined by the Betts
process, tin dissolves with lead at the anode and co-deposits at the cathode. The
lead - tin alloy may be sold, or some post treatment of the cathodes is necessary
to remove Sn [79].
The wide use of H SiF in lead electrorefining is due to its low cost. Fluosilicic
acid is produced as a by-product of the treatment of phosphate rock in fertilizer
manufacture [801. Fluorides and silica contained in phosphate rock form
fluosilicate and are separated from the fertilizer product. Fluosilicic acid is also
produced as a by-product during the dissolution of apatite with sulphuric acid
[54,80]. During the electrorefining operation, fluosilicic acid is consumed by
entrapment in the anode slimes and by volatilization from the surface of the
electrolyte. The acid develops significant vapour pressures through volatile
decomposition products according to the reaction,
2
H SiF ^2HF(g)
2
+ SiF (g)
4
SiF and HF are both corrosive and toxic and are removed from the tankhouse
atmosphere by adequate ventilation.
Table 1 shows some of the operating parameters of various lead
electrorefining plants. The wide variation in electrolysis conditions seen in this
table does not seem to have a strong influence on the final quality of the refined
lead. For example, lead concentrations in the electrolyte can be varied between
30 and 270 g/1 without affecting seriously the refined lead quality. The electrolyte
recirculation rates are also varied widely, with no apparent correlation to other
operating parameters. Extremely high electrolyte velocities might improve mass
transfer across electrode boundary layers, but can also nullify the additive effects,
worsening the deposit quality and causing short circuits [81].
4
Table 1 also shows that the current densities employed in the Betts process
fall in the range of 120 to 230 Amp/m . Higher current densities have been
achieved through the use of galvanodynamic techniques such as current
modulation and periodic current reversal (PCR). The current modulation
technique consists of decreasing the current density (e.g. from 220 to 160 Amp/m )
in small steps [33,87]. Each constant current density step is determined on the
2
[8]
(a)
>
(b)
ft
200
a)
l , extrapolated curve
c
120
<u
o
ft
IH
>
Q
80
0)
J-l
30
60
90
o
ft
in
IH
>
IH
IH
Tl
O
9)
Va
4!
<
120
ISO
40 JH
180
60
<
90
120
150
180
Fig. 4 Changes in the anodic overpotential value, ri ,during lead electrorefining [87]
A
(a) Conventional galvanostatic process (b) Galvanodynamic process in which the cell current is continuously
decreased during the refining cycle to fix the value of T | .
A
basis of an anode overpotential value which increases as the slimes layer thickens
and decreases when the current density is reduced. The upper limit for the anode
overpotential is usually deterrnined by Bi dissolution from the anode. Fig. 4 shows
the current density program that can be applied to the elechrorefiriing circuit
without reaching the critical overpotential value for Bi dissolution. The higher
average current density possible with current modulation reflects in a shorter
electrorefining cycle and higher refinery production.
Periodic current reversal (PCR) in lead electrorefining is widely employed in
China {381. PCR involves frequent short reversals of the electrolysis current
direction 188]. This reduces the concentration polarization in the slimes layer and
levels the cathodic deposit by selectively dissolving projections. High current
efficiencies, good cathode quality, low electrolyte losses, low energy consumption
and a decrease in the number of short circuits have been reported through the
use of PCR [22]. A 16% increase in free acid was found in the slimes layer due to
PCR. [22]
[9]
TABLE
1 . B e t t s l e a d e l e c t r o r e f i n i n gi n t h e w o r l d
Takehara. Japan U3]
236-270
75
135
Harima,
Japan 1*6]
Shenyang Smelter,
China [sal
90-100
110-115
74
104-124
38-58
400 Sb, 800 Fe,
110 As, 28 Zn.
290 Sn. 0.2 Ag
1000
300-450
200-300
32-45
Pressure Tank with
centrifugal copper
pump
18-25
1.5-1.7
Electrolyte:
Pb. g/1
Total HjSiF,. g/1
Free HjSlFj, g/l
Others: rag/1
52-61
2-5 Bi. 0.5-2 Cu, 1-2 As,
130-200 Sb
Additives Consumption:
Aloes, g/ton
Ligniri Sulphonate, g/ton
Glue, g/ton
Others, g/ton
Temperature, C
Circulation Apparatus
Recirculation Rate, 1/min
Add loss. Kg/ton
600-1100
28-43
35-38
40-43
1.8 m'/min x 20 m Head
30
1.6-2.7
40
2.3
30
4
120-140
0.5-0.6
15000 Amp S70V
147
0.47
10000 Amp 9200V
185
0.55
5000 amp 50v 13000
amp 8 50v
93
175-180
Current:
Amp/m , Cathode
Cdf Voltage, V
Current, Kw per Generator
1
Current Efficiency. %
Energy Consumption Kwh/ton Pb:
Electrolysis
Mechanical
95.6-98.7
93
154-157
17.8-20.7
143
30
154-172
0.46
2800-3500
96.30
120-130
30-40
Anodes:
Casting Wheel, 18 Moulds
Casting Technique
Composition
Length,Width,Thickness,
Mode of Suspension
Life, Days
Scrap. %
Anode Spacing, mm
Weight, Kg
mm
Casting Wheel
.68% Sb.
0.073% Cu.
0.044% Sn
970x740x35
cast lugs
8
26.5
110
280
920x620x23
Suspended Lugs
380-440
1200x800x24
cast lugs
7
30
100
250
0.6-1.0
10-20
4-5
.8
10
7
1
10
4
.8-1.
8-11
2.5
20-25
95
1403
Cathodes:
Starting Sheet
Production Technique
Thickness, mm
Weight, Kg
Life, days
Anode Slimes:
Composition
Removed After ? days
Percentage of Anodes
Scrubbing Technique
12-15% Pb,
8-12% Bl.
0.2-0.4% Te
7
2.9
1.3
2.5
1.1-1.4
500x130x155
300x100x150
28. 29
vinyl chloride with steel
frame
1150x920x1450
(inner size)
320x75x120
42, 43
prefabricated concrete,
PVC lining
32, 33
Reinforced Concrete
with asphalt or PVC
lining
45000
99.999
68000
99.999
27000
99.999
54450
99.99+
2.4-3.6
Rotating Brush
Tanks:
Length, Width. Depth, cm'
Number of Anodes, cathodes
Construction Materials
Pb Annual Production, ton
Pb Average Content %
[10]
(continuation)
Cominco,
Canadaiss]
[881
Kamloka.
Japan [so]
75 (60-80)
141
90 (90-100)
8 Bi. 65 Sb, 4000 SiO,.
1.6 Cu, 12 Sn. 5 As,
17 In, 70 Tl
30-50
120
100
75
120
60
600 HF. 5. Sb. 0.5 BL
0 SlOj. solids
75
130
75
1 Bi. 150 Sb
550 (Calcium)
550
180
600
40
Centrifugal Pumps
12
3
35-45
Volte Pump 18.5Kw x 2
156
0.5-0.6
Mercury Rectifier 1520
Motor generator 360
135
0.55
20000 Amp O60V
Electrolvte:
Pb.g/l
Total H,S1F g/1
FreeH,SiF,.g7l
Others: mg/1
Additives Consumption:
Aloes, g/ton
Lignin Sulphonate, g/ton
Glue, g/ton
Others, g/ton
Temperature, T !
Circulation Apparatus
170 (Aloin)
250 (Calcium)
40 (38-43)
Centrifugal Pumps
27 (27-45)
2
600
35
Storage tank with epoxy
lined^gump
15 (32% pure)
40
2
Current:
Amp/m , Cathode
CellVoltaBe. V
Current, Kw per Generator
1
Current Efficiency. %
Energy Consumption Kwh/ton Pb:
Electrolysis
Mechanical
230 (max)
.3-.5
6300 Amp(5 day), 5400
Amp(7 day)
1000 kw
90-95
185
0.45
150
92
90
96
168
50
195
90
143
35
165
130
Casting Wheel, 12
Mould
0.5% Sb. 0.3% Ag.
0.3% Bi, 0.1% Cu,
0.002% Sn
Casting Wheel
864x660x30
730x710x25 (immersed
surface)
Suspended Lugs
Anodes:
Casting Technique
Composition
Length.Width.Thickness, mm
Mode of Suspension
Life. Days
Scrap, %
Anode Spacing, mm
Weight, Kg
1140x990x20
Suspended Lugs
Shoulder Type
6
40
130
200
4
49
100
150
6
45
110
Mechanized production
with on line casting of
ribbon
1
60-70 (Final Wdght)
Continuous Drum
Casting
.6
80 (Final Wdght)
.7
13
Cathodes:
Continuous Drum
Casting
Starting, Sheet
Production Technique
1
6.3
5 or 7
Thickness, mm
Weight, Kg
Life, days
Anode Slimes:
5 or 7
3
Conveyed By Monorail
Between Rubber
Scrapers
6
1% (solids)
pneumatic stripping
268x82x112
24. 25
Asphalt Lined Concrete
(old) Polymer Concrete
(new)
455x95x130
40, 41
Glue Lined Concrete
500x130x160
43, 44
Vinyl Chloride Resin
and Concrete
144000
99.99
15000
99.99
72000
99.99
30000
99.99
Composition
6
1.4
Rotating Brush
Tanks:
Length, Width, Depth, cm
Number of Anodes, cathodes
Construction Materials
3
[11]
Introduction
TV T h e A n o d i c Process
A. Introduction
The Betts process, as practiced by all refineries, depends on the formation
of an adherent, porous anode slimes layer during electrolysis. The slimes layer
consists of undissolved impurities which are removed mechanically from anode
scrap after the anodes are withdrawn \ If the slimes do not adhere to the anode
during electrolysis they will settle through the solution and to some extent be
mechanically entrained in the cathode deposit. There is no indication in the
literature of a lead refining process in which (as in copper refining) slimes fall is
encouraged. Further, the recovery of slimes from the bottom of the cell (as in
copper refining) is perceived to be more costly. To form an adherent slimes layer,
certain elements (mainly As, Sb, and Bi) must be controlled within a narrow
composition range and/or ratio in the anode, and the anode casting process must
be designed to control the rate of solidification to optimize the microstructure of
the cast bullion.
1 During normal lead electrorefining practice some slimes do fall, and are cleaned out of the cells at very
infrequent intervals (months).
[12]
Fig. 5 Lead anode microstructure. From an anode currently being used by Cominco Ltd. Chemical composition
as described in Table 1. Top view, air cooled side.
1) Impurity phases of the solid solution type (SST): the Pb-Bi system.
2) Impurity phases of the precipitation type (PT): the Pb-Sb system.
3) Impurity phases of the eutectic type (ET): the Pb-As and the Pb-Ag
systems.
The strength of slimes adhesion to the anode was quantified by Tanaka [94]
through observations of slimes fall and slimes morphology. Tanaka summarized
the results of these studies with the following relationships:
1) The greatest slimes adherence is obtained when impurity phases of the
SST type are present in the anode in concentrations greater than 0.23% wt.
A SST concentration lower than this critical value produced a slime that
easily slides off the anode.
2) The addition of a third element to the eutectic systems increases
significantly the slimes adherence.
3) Water quenching of the anodes increase slimes adhesion particularly in
[13]
Industrial practice
the FT and E T systems. However, the very small slimes particles formed
during quenching may also promote slimes detachment and mechanical
entrainment in the cathodes.
The increase of slimes adhesion by water quenching relates directly to the
lead anode solidification rate, which influences the growth and distribution of
impurity-containing phases. Especially important are the eutectic forming
systems where it is known that the solidification parameters (growth velocity,
temperature gradient in the liquid and growth mechanism) and system parameters
(volume fraction and impurities content) can lead to anomalous eutectic
structures 1951. Such structures have been reported to occur in the following
systems (95] :
1
C. I n d u s t r i a l practice
As Table 1 shows, the range of anode impurities used in lead refining varies
from plant to plant. In every refinery the anode composition is kept vrittiin narrow
limits to obtain an adherent slimes layer. The need for production of anodes with
homogeneous properties has led to control methods for cooling rates and casting
techniques. In addition to the use of a casting wheel, straight horizontal 144-46]
and vertical [47] anode casting systems are employed.
Fig. 6 shows the straight type horizontal lead anode casting system currently
used in Japan. Although the vertical and horizontal casting processes were
originally developed to save space (over that occupied by a casting wheel), they
were carefully designed to achieve uniform cooling rates during the casting of the
1 A broken lamellar structure consists of a near regular array of "broken" plates and occurs in systems
that contain less than 10% of the faceting phase [95]. The complex structure consists of an array of plates
that are regular over small areas around a well defined spine [95].
[14]
ID
01
Anod*
Casting Mould
Cooling
Electrode
Assembling
Machine
Lifter
Lifter
anodes. Such new techniques for anode casting as well as for optimizing heat
treatment have been patented by Japanese companies 1961, but no information
was found that would indicate the nature of changes in microstructure resulting
from these newer techniques.
D . S l i m e s e l e c t r o c h e m i c a l behavior
The slimes layer formed in the BEP undergoes structural and chemical
changes during its growth. As the slimes layer thickens, the different phases and
compounds present react with the entrained electrolyte creating secondary
products. The Ej, - pH diagram for the PbSiF -H SiF system (ion indicates that
both lead fluoride and silica can precipitate at higher pH values. This has been
related to the effect of electrolysis parameters on secondary processes (such as
these precipitations) that take place witxiin the slimes layer during electrolysis
[97-1001. The transport processes within the slimes layer in the context of these
secondary reactions have also been the subject of several studies [102-108].
6
A simple physical model to study the role of the slimes layer during anodic
dissolution rate in electrorefining systems was developed by Reznichenko et al.
[1091. In this model, the slimes layer was represented by a silver gauze diaphragm
(representing the noble impurity) electrically connected to the anode (a highly
pure base metal) and placed at a (variable) distance from the anode corresponding
to the slimes - electrolyte interface. The distance between the anode and the gauze
was varied to simulate the effect of an increasing ohmic drop between the anode
and the slimes layer. During electrolysis, concentration gradients normally
[15]
Slimes electrochemical
behavior
Equation 1 shows that the larger the ohmic drop the larger the equilibrium
concentration of ions of the noble metal and the larger its dissolution rate.
Although this model oversimplifies the different phenomena taking place within
a slimes layer, it provides an insight as to the effect of ohmic drop buildup in the
slimes layer and the dissolution rate of noble impurities normally left in the slimes.
Reznichenko et al. 11091 applied their model to the study of the Cu-Ag system in
which they found a logarithmic relationship between the dissolution rate of noble
impurities and the ohmic drop potential.
Early research on the behavior of BEP slimes has focused on relating the
anodic overpotential to the cathode purity [lio.ni]. Among the impurities most
closely followed in the cathodic deposit is Bi [112.1131. Even though the BEP has
a large selectivity for the removal of Bi, it has been found that towards the end of
the electrorefining cycle such selectivity can be lost. It seems that the buildup of
concentration gradients and the precipitation of secondary products within the
slimes layer can initiate steep increases or discontinuities in the anodic
overpotential at a certain slimes thickness. This would produce a large increase
in the rate of dissolution of such impurities, and subsequently in the deposition
of impurities in the cathodic deposit. Fig. 7 shows how Bi contamination in the
cathode increases with anodic overpotentials above a critical value (about 200
mV). This has been recognized as a general behaviour for bismuth in lead refining
by the BEP [33,87,110, ill]. On the other hand, the use of periodic current reversal
(PCR) [39,98] has been found to increase the minimum anodic overpotential at
which noble impurities start to dissolve.
[16]
500
[17]
Slimes electrochemical
behavior
CO
62
Bath level
Fig. 8 Dimensions of the anodes
used in Wenzel's experiments
showing the position and size of the
electrolyte sampling wells [97].
CM
98
::i::e-
T
"
tj * -e-
-Q-
* 8
due to the pH increase several reactions may occur: SiF " decomposition to SiF
and F" (pH > 3); precipitation of S b 0 (pH > 4.9) and precipitation of PbO (pH >
7).
Table 2 shows the results of chemical and diffraction analysis, which seems
to indicate the presence of these secondary compounds within the slimes layer .
From an analysis of the concentration of noble impurities in the entrapped
electrolyte, Wenzel et al.were able to conclude that when the SSM exceeded 5%,
as determined from the Pb-Bi-Sb ternary diagram, permissible impurities in the
anode were too high to obtain an acceptable impurity level in the refined lead.
2
1 Note that since slimes oxidize rapidly in air, the analysis may indicate oxide phases where metallic
phases were present in the in-situ slimes.
[18]
Fig. 9 Changes with electrolysis time of the relative concentrations of Pb* and H* within the slimes layer with
respect to their bulk values [97,115].
2
Bulk electrolyte composition: [PbSiFJ = 0.4 M, [HjSiFJ^ = 0.6 M, T= 35 *C, I = 200 Amp/m , Stationary
Electrodes. Vertical Axis: Pb and H* concentration ratio between the electrolyte sampled within the slimes layer
and the bulk electrolyte.
(a) Anode with 0.46% Sb and 0.24% Bi (= 1.73% SSM) (b) Anode with 0.92% Sb and 0.24% Bi (= 3.80% SSM)
2
+2
Note: No extrapolation of the inner electrolyte concentrations at zero slimes thickness done here.
1
2
Anode
Composition
%wt,(rest Pb)
X-ray
Diffraction
Analysis
Pb
Sb
Bi
Si
0.46% Sb,
0.24% Bi
56.98
7.62
3.95
11.97
6.66
0.46% Sb,
0.24% Bi
58.02
0.92% Sb,
56.57
9.40
4.28
14.09
3.33
0.92% Sb,
0.24% Bi
6.81
5.55
11.27
6.65
PbO, Sb 0 ,
PbF , Pb Si0 .
BtjOg
2
0.24% Bi
4
PbO, SbaOg,
PbF , Bi 0
3
PbO, Sb 0 , PbF ,
2
BiaOg
53.77
9.37
9.20
12.22
4.19
PbO. Sb 0 . PbF ,
BisOa
2
[19]
Wenzel's X-ray diffraction analysis on the slimes products agree with the
findings of Isawa et al. 1116,117], who studied the slimes composition of synthetic
and industrial lead anodes and found the presence of the following compounds
1
One compound that has been difficult to characterize is AgaSb 194], whose
presence has been detected in both synthetic and industrial lead anodes. The
morphology of this compound has been found to be a function of the anode cooling
rate and of the electrolysis conditions.
The electrolyte used in lead electrorefining contains impurities whose
concentration ranges from a few ppm and several g/1, and have their origin in the
anode from which they are dissolved (i.e. C u , Sn ) or from the acid
manufacturing process (i.e. phosphorus species). Their steady state
concentrations in the electrolyte depend on the electrolysis parameters and slimes
thickness. They can affect both the anodic and cathodic processes through
changes in fundamental electrochemical parameters such as the exchange current
density [Q, the symmetry factor (a), and the electrical double layer capacity (CdJ \
Measurement of these parameters usually involve transient electrochemical
techniques, such as polarization scans, current interruptions, and AC impedance
studies.
+2
+2
1 The term electrical double layer is used to describe the arrangment of charges and oriented dipoles
constituting the interphase region at the boundary of an electrolyte [144, p.630]
The exchange current density, i , is a measure of the rate of equilibrium potential and sensitivity to
interference [145, P.IO]
The symmetry factor, a, determines what fraction of the electrical energy resulting from the displacement
of the potential from the equilibrium value affects the rate of electrochemical transformation [144, P.923]
0
[20]
Slimes electrochemical
+3
+2
behavior
single current step transient method [118]. He found that under an oxygen
atmosphere the presence of the above ionic species increased the anodic
overpotential value. Thus, they may inhibit the dissolution of some noble species
including lead.
50
100
150
200
[21]
Fig.
11
Continuous purification o f
electrolyte v i a purification c o l u m n
[125,126].
V T h e C a t h o d i c Process
A . A d d i t i v e s c o n t r o l and e l e c t r o c h e m i s t r y
In the absence of additives, lead deposits with a very small overpotential,
and tends to form rough, porous deposits or dendrites that result in short circuits.
To obtain deposits that are flat, smooth, and free from projections, special reagents
are added to the electrolyte. These "addition agents" increase the cathodic
overpotential (actually called "inhibition"), and change the kinetic parameters (to,
a, and e.d.l.) under comparable electrolysis conditions [120-123].
[22]
B. S t a r t i n g sheet technology
The cathode starting sheets used in the BEP are usually cast on a rotary
drum casting machine. The stiffness of these sheets is increased by adding
approximately 15 ppm Sb to the melt and by impressing wrinkles on the cathodes
right after they are produced [791. Procedures to avoid dross incorporation in the
lead cathodes have also been developed [137]. In this section of the electrorefining
process high levels of automation have been achieved [138].
C. C e l l electrolysis parameter o p t i m i z a t i o n
The overall electrorefining process has been studied using statistical
correlations of some of the variables measurable in an operating plant [39.54.98].
To optimize the electrolysis parameters when the BEP is run at high current
densities (> 200 Amp/m ), factorial design of experiments at three levels for four
variables has been used by Lange et al.[98]. They varied PbSiF and H SiF
concentrations, temperature, and current density, at constant values of addition
agent (glue), anode composition, cell geometry, and electrolyte recirculation. The
influence of these parameters on the average anodic and cathodic polarization,
on the cell voltage, and on the specific energy consumption was obtained by
employing regression analysis correlations. In addition, changes in these
parameters were correlated with the cathode quality and appearance as well as
2
[23]
with the consumption of glue. The optimum electrolysis parameters found for a
current density increase (from 200 to 300 Amp/m ) were a Pb concentration of
60 g/1 (down from 80 g/1), a glue addition maximum of 1500 g/ton Pb (up from
1250 g/1), a free H SiF content of 120 g/1 (no change) and an electrolyte
temperature of 35 to 40C. Under these electrolysis conditions, Lange et al. found
that Sb and Bi slimes layers could be subjected to anodic polarizations as high
as 280 mV without serious impairment of the cathode quality.
2
+2
140
120
cO
100
CO
ti
80
(*)
W^f
cu
= 10 sec
>
O
o
4 0
< * ) ^ = r ^ = 20.ec
rH
T3
O
20
ti
20
40
backward
. . . i . . . i . . . i . . . i .. .
60
80
100
120
140
[24]
hours during the 6-day long anode cycle. In addition to this, cathodes are
exchanged daily to avoid short circuits and to obtain a deposit of acceptable purity
[57].
Lange et al. also studied the use of PCR in the BEP. Fig. 12 shows the anodic
overpotential dependance on the PCR parameters. Thiet [39] found that the use
of PCR decreased the anodic overpotential values. The reduction in the anodic
overpotential values did not reflect in lower PCR energy consumption probably
because it was offset by unproductive energy consumption during the reverse
current phase of the cycle. Thiet also showed that the quality of the PCR cathode
deposits was very high [39].
D. B i p o l a r r e f i n i n g of lead
Among other alternatives to optimize the lead electrorefining process, the
use of a bipolar configuration looks most promising. In this configuration only
the terminal electrodes are connected to the source of current. Between these
electrodes a large number of bipolar electrodes can be incorporated. One side of
these electrodes will corrode anodically, while pure lead will deposit on the other
side. The Impurities will be left behind as a slimes layer on the anodic side. The
advantages of operating the BEP in a bipolar mode include PCR, C.P.V., optimum
lead and free acid contents, and the use of jumbo electrodes (~4m , possible
because there are no bus-bar connections)\ The process has been operated in a
pilot plant where it proved to be superior to the parallel process [1421. The rationale
for the current use of the parallel system is that a high economic investment is
required for the substitution and implementation of the bipolar process. If new
plants to electrorefine Pb are to be constructed in the future, very likely they will
be assembled on the bipolar configuration.
2
1 In the bipolar mode by-pass currents are reduced by increasing the area of the electrodes so that the
ratio of bypass to electrode area in the cell decreases.
[25]
Chapter 2
F u n d a m e n t a l s of the E l e c t r o c h e m i c a l Measurement
Procedure
Early in the development of this project it was found that the form of ionic
concentration gradients within the lead anode slimes layer needed to be studied.
To study these concentration gradients without causing major disruptions to the
system, "in-situ" experimental techniques were considered. Among these,
transient electrochemical techniques appeared to be particularly advantageous
and consequently were used extensively in this work. The study of the
electrochemical response of typical lead anodes was complemented by
measurements on the physico-chemical properties of H SiF -PbSiF electrolytes.
What follows in this chapter is a description of the electrochemical parameters
germane to this work, and of the transient techniques that were used.
2
I. C o m p o n e n t s of the A n o d i c Overpotential
The anodic overpotential was measured by following the difference in
potential between the lead anode ("working electrode") and a suitable reference
electrode. A high purity lead wire located within a Luggin capillary and contacted
with cell electrolyte was chosen as the reference electrode. Its use was possible
because of the high reversibility of lead in the H SiF -PbSiF electrolyte system.
Furthermore, the amount of current that passes through the reference electrode
is limited to the current drain of the meter (a few nanoamperes). The "anodic
overpotential" \ T\ , measured by such an electrode would also be free of any
junction potential between its tip and the slimes/electrolyte interface provided
the bulk electrolyte is well mixed. In the present work (unless specified) all
potentials are given with reference to such an electrode. The third electrode (a
lead foil counter electrode) is required to complete the electrical circuit. The
electrochemistry of this counter electrode is of no importance to the anodic
processes. Fig. 1 shows a simplified view of this three electrode arrangement.
2
1 This "anodic overpotential", r\ or potential difference between the anode and the reference electrode,
includes an ohmic component, x\ , and long range potential gradients as well as interfacial overpotential
components.
Al
[26]
changes the value of T| mainly through changes in its migration and concentration
overpotential components. Additionally, an ohmic contribution is included in this
TJ measurement.
A
[27]
ac
AC
AC
AC
n. Transient E l e c t r o c h e m i c a l Techniques
The amount of information that non-transient electrochemical techniques
are able to provide is rather limited [ 11-14]. Significant enhancements in electronics
in the last 25 years have supplied the electrochemist with a wide range of transient
techniques. Reviews on the use, applications, and limitations of these techniques
are available [11.15-21]. Among these transient techniques, current interruption
and AC impedance were used extensively in this work. These techniques were
chosen due to their potential for resolving the T\ components.
A
1 The migration potential, <t>, can be calculated by solving the Nernst-Planck flux equations. Appendixes 1
and 2 describe how these equations can be solved. A4> can be neglected in the presence of a substantial
excess of supporting electrolyte.
[28]
halted instantaneously. This will be the case if contained electrical double layers
are not uniformly charged \ The relaxation time required for the double layer to
equilibrate can be estimated from the following relationship :
2
TD =
...
The discharge of the double layer can also take place through Faradaic
reaction. The time constant for this reaction can be estimated using the following
equation:
Ft.
T/
The third major process that can take place upon current interruption is the
relaxation of concentration gradients. The time constant for this unsteady state
process can be estimated from the following relationship:
" =D
Typical values for x and x are close to 0.5 msec [7]. Thus, the electrical double
f
layer will be equilibrated and discharged in approximately 1 msec. After this time,
concentration gradients will relax. The larger the time constant the longer it will
take for the system to reach its open-circuit or equilibrium potential. Another
process that can take place upon current interruption is the dissolution and/or
re-precipitation of secondary products within the slimes layer. This process may
originate one or several potential arrests [351 *.
1 Double layers may not be uniformly charged if the difference in potential at the electrode/solution
interface is not uniform.
2 L is the characteristic length that controls the current distribution [20,33].
3 In electrolytes of uniform composition, potential drops upon current interruption have time constants of
the order of 7x10* msec .
4 These potential arrests are sometimes referred as Flade potentials.
[29]
1 The time domain data refers to the potential and current transients dependance with time. When these
data are Fourier transformed, the frequency spectrum of the system is obtained (time-domain o
frequency-domain). For a more complete description of this transformation see appendix 3.
[30]
Transform (FFT) algorithm [40-43]. The software packages used in this work
incorporate this algorithm . Knowledge of the FFT algorithm, its application range
and limitations, are required to understand the data generated. Appendix 3
reviews some of the relevant steps required for implementing the Fourier
transformation.
1
B. AC impedance techniques
The use of AC impedance to study electrochemical systems spans over a
period of more than 100 years. Kohlrausch, as early as 1854 [44] proposed its use
for the determination of the electrical conductivity of electrolytes. Later, with an
AC bridge, he obtained the conductivities of a large number of electrolytes [45],
Contributions by Warburg [46] and Randies [47] set the theoretical and
experimental foundations from which AC impedance measurements have evolved.
Nowadays, the technique is routinely used for the analysis of a wide range of
electrochemical systems [48-54]. Electrical engineering theory has been heavily
used by electrochemists to interpret the results, and reviews on its characteristics,
advantages, limitations, and implementation can be found in the literature
[14.19,48.55-57].
In a typical AC impedance experiment, a small sinusoidal signal (either voltage
or current) is applied to an otherwise DC system. The AC output response is
followed as the frequency of the input signal is changed. The ratio of the input
and output signals is known as the transfer function [58,59]. Thus, the impedance,
Z(jco), is a transfer function. The system response can be assumed to be linear by
limiting the amplitude of the input signal to a few mV (or a few mA). The possibility
of obtaining an electrical analogue circuit by using AC impedance is a strong
driving force for its application. Knowledge of this analogue allows the modelling
and prediction of the output signal when the input is known.
The linkage between current interruption and AC impedance techniques
arises through the use of the transfer function. Theoretically, upon Fourier
transformation of the current interruption data, the transfer function obtained
should match the one obtained by AC impedance. That this connection exists is
shown in appendix 3. On the other hand, experimental artifacts make the analysis
of data generated in the time-domain valid only for a very limited frequency range
[31]
The search of electrical analogues that match the response of the system
and have physical meaning is strongly pursued in this thesis.
[33]
Chapter 3: E x p e r i m e n t a l Procedure
I. E l e c t r o c h e m i c a l E x p e r i m e n t s
A . Electrochemical cells
1. Beaker Cell
A I L wide mouth (0=11 cm) polyethylene beaker was adapted to function
as electrochemical cell (Fig. 1). This cell was used to study the anodic behavior
of small working electrodes (i.e. working electrodes whose geometric area was
between 1.4 and 3 cm ). A Lucite tight cover was used to hold the electrodes in
a fixed position and to avoid loss of water due to electrolyte evaporation. The
cover had holes through which the electrical connections to the electrodes and
the tubes used for electrolyte sampling were taken out of the cell. Previous to
each experiment the cell was thoroughly washed several times using deionized
water. After the cell was dried, the electrodes were firmly positioned In the cell.
Then the electrolyte was introduced to the cell through a plastic tube. During
the experiment, this tube was used to obtain bulk electrolyte samples. Electrolyte
was added only at the beginning of the experiment in volumes that ranged
between 320 and 400 ml. Mixing of the bulk electrolyte was by either magnetic
stirring or recirculation , . Electrolyte recirculation provided the best
experimental reproducibility and was preferred to magnetic stirring. After the
electrodes were positioned and the electrolyte introduced, the cell was covered
and sealed using generous amounts of silicone rubber . Only after the silicone
rubber had "dried to touch" (i.e. after approximately 2 Hrs.) was the cell immersed
in the constant temperature water bath.
1
1 H SiF containing solutions can etch glass [1]. Thus, glass laboratory ware was avoided as much as
possible.
2 Geometric areas do not consider surface rugosities.
3 Electrodes were immersed in a 10% Vol HN0 solution and washed with deionized water prior to their
introduction to the cell.
2
4 Magnetic stirrer bar coated with Teflon* (1=2.54 cm, 0=0.95 cm) spinning at low speed by using a
Tek-stir magnetic stirrer model S8250-1.
5 As illustrated in Fig. 1 electrolyte was recirculated through plastic tubes located at opposite sides of the
cell and at different electrolyte depths.
6 Electrolyte recirculation rates were between 5 and 6 ml/min. A Masterflex pump catalogue No. 7553-20
and Masterflex Tygon tubing catalogue No. 6409-14 were used.
7 100% silicone rubber, RTV Silastic 732.
[34]
Rectangular cell
F i g . 1 A s s e m b l y used in the
experiments performed with
the beaker cell
(A)
W o r k i n g electrode
(B)
Counter electrode
(C)
Reference electrode
(D)
C u wire
(E)
(F)
Electrolyte
recirculation inlet
(G)
Electrolyte
recirculation outlet.
2. Rectangular cell
Rectangular Lucite cells were used to study the electrochemical behavior
of working electrodes whose exposed geometrical area ranged between 30 and
40 cm . Fig. 2 describes the dimensions and design features of this kind of cell \
A Lucite tight lid was used to cover the cell. In this lid, holes of appropriate size
were drilled. Through these holes, the electrical connections to the electrodes
and the tubes used for electrolyte sampling were taken out of the cell. Bulk
electrolyte was recirculated using lateral cell inlet and outlet facilities.
2
1 Cell dimensions were modified according to the characteristics of the experiment. Fig. 2 shows the
actual dimensions of the electrochemical cell employed in experiment LC2 to be discussed extensively in
Chapter 4.
2 Electrolyte recirculation rates were between 5 and 6 ml/min.
[35]
Rectangular cell
Electrolyte from the top of the cell was brought to the bottom continuously.
During the electrolysis of lead electrodes, electrolyte samples taken at different
cell locations showed no significant concentration differences.
Working electrode
Counter electrode
Reference electrode
Cu wire
Plastic tube used to
withdraw bulk
electrolyte samples
(F) Electrolyte
recirculation inlet
(G) Electrolyte
recirculation outlet
(H) Lucite side walls
(I)
Lucite strips.
Drawings are not scaled.
All measurements in mm.
(A) Cell tridimensional view
(B) Lateral view
(C) Frontal view.
Fig. 2 shows the position of the two small Lucite strips used to support the
anode. The presence of these strips allowed the electrolyte entrapped within the
slimes layer to flow downwards. Additionally, wide Lucite strips were used to
center the anode in the cell and to improve current distribution. Electrolyte
volumes were fixed at the begirining of the experiment and ranged between 300
and 400 ml. When bulk electrolyte samples were withdrawn from the cell,
electrolyte of the initial composition was added to the cell to maintain constant
the electrolyte volume. As in the beaker cell case, after the electrodes were placed
and electrolyte added, the lid was fitted to the cell and sealed afterwards using
silicone rubber. Again, only after the silicone rubber had dried to touch, was the
cell immersed in the constant temperature bath.
[36]
Materials
B. Electrodes
1. Working electrodes
(a) Materials
Working electrodes were produced from pure lead and from typical lead
bullion anodes provided by Cominco Ltd. Typical anodes used in the Betts
electrorefining process were taken from the anode refining wheel and some of
its sections were sent to UBC. No particular details regarding the anodes cooling
rates or about the pyrometallurgical steps previous to their casting were
provided. According to Cominco, the anodes were cast under normal operating
conditions and that's where the term "typical" comes from. The electrochemical
studies here pursued required the anodes to form strong and adherent slimes.
As was later found, this set of anodes did produce slimes which did not fall from
the electrodes and remained attached to them during the dissolution stages.
1
B.2
Fig. 3 Sections of the lead
bullion anodes used to prepare
working electrodes.
A.1
Fig. 3 shows the anode sections that were cut from the anodes and sent
to UBC. The anode A strip was sent in July 1986. This strip was cut in smaller
pieces and from the area indicated in Fig. 3A, electrodes were produced. Sections
of anode B were sent in March 1987. From anode B, electrodes were produced
mainly from its center part (section B - l shown in Fig. 3B). The working
electrodes were prepared so as to study the electrochemical behavior of the
1 Pure lead working electrodes were made out of Tadanac ingots (Pb >99.99%)
[37]
"mould" cooled anode face and the "air" cooled anode face. Thus, electrodes
were fabricated by cutting perpendicular sections of the anodes. Depending on
the type of electrochemical cell employed, different electrode sizes were used.
(b) Beaker cell
Small working electrodes whose exposed geometric area ranged between
1.4 and 3 cm were prepared as follows:
2
2 The deformed layer was measured metallographically and in some cases was as thick as 0.1 mm.
[38]
Relgrence electrodes
2. Reference electrodes
A pure lead wire was used to measure the difference in potential between
the bulk electrolyte and the working electrode. This wire was mounted in a plastic
tube (0=2 mm). The tube was bent at one end, and a plastic tip was inserted
there. Fig. 4 illustrates this Luggin-Haber reference electrode arrangement. The
reference electrode tip was placed between 2 and 5 mm away from the original
position of the working electrode. This distance remained constant during the
4
1 Size and location of the holes is provided when specific experiments in which electrolyte samples were
taken and inner potentials obtained are analyzed.
2 From the inner slimes layer 100 u l of electrolyte were slowly extracted (over a period of 3-6 hours). A
100 u.L Unimetrics removable needle syringe was used to withdraw the inner electrolyte samples.
3 The difference in potential between this wire and the anode was used as an indication of the slimes
layer conductivity.
4 Johnson Matthey, 99.95% Pb, 0=1.0 mm.
5 Eppendorf pipette tips 5-100 |xL
[39]
Counter electrode
experiment. The tip of the reference electrode was located facing the geometrical
center of the working electrode. After the cell was assembled and the bulk
electrolyte had penetrated the reference electrode compartment, this was sealed
by using silicone rubber. Thus, the electrolyte surrounding the lead wire had the
same composition, temperature, and atmosphere as the bulk electrolyte.
A
/
/
1 ..
*H J
,
3. Counter electrode
Pure lead foils were used as counter electrodes. In the beaker cell case the
lead foil surrounded the working electrode whereas in the experiments which
3
1 This sort of experiments were only carried out using the rectangular electrochemical cell in which the
size of the working electrode was large enough to incorporate these potential measuring probes.
2 The electrolyte entrapped within the slimes layer will be called "slimes electrolyte".
[40]
Electrolyte
used the rectangular cell, the lead foil was located facing the working electrode
and @42 mm away from it. In both cases, the area of the counter electrode was
larger than the area of the working electrode. This was done to improve the
current distribution and also to avoid formation of dendrites. When the beaker
cell set-up was used, the geometric area of the counter electrode was between
10 and 15 times larger than the exposed geometric area of the working electrode.
In the experiments performed using the rectangular cell, counter electrode areas
were between 1.2 and 1.4 times larger than their working electrodes counterpart.
In all the experiments, electrical contact was incorporated by soldering a Cu wire
to the counter electrode. The place where the electrical contact was made was
isolated from the electrolyte by a silicone rubber coating.
C. Electrolyte
Technical H S i F obtained as a by-product of the treatment of phosphate
rock [2] was neutralized with either PbC0 or PbO to prepare mother electrolyte
solutions. The reactions that take place upon neutralization of the acid are:
1
H SiF +PbC0
2
...1
=> PbSiF + H 0
...2
H SiF + PbO
2
1 Acid composition: 2.03 M H SiF and 0.40 M Si0 . This acid was provided by Cominco Ltd.
2 Phosphate rock is treated with H S0 to produce HF which is later contacted with Si0 to produce
H SiF . The sequence of reactions that take place is [2]:
2
...a
...b
]0
HiF
...c
...d
3 Depending on the extent of the neutralization of the acid, a mixture of lead oxides and fluorides in
addition to silica compounds can precipitate.
4 The presence of colloidal Si0 nH 0 in this system has been reported in the literature [3-7].
2
[41]
Electrolyte
A titrimetic analysis routine was set up to analyze the electrolyte for H SiF ,
Si0 , HF, and PbSiF . In this routine, Pb is analyzed via complexometric titration
with EDTA [8,9]. H SiF , Si0 , and HF (if present) are determined by titration with
LiOH. Details of this procedure are provided in Appendix 4.
To obtain electrolytes of the set compositions, the mother electrolyte solutions
were diluted with H SiF and deionized water \
2
The electrolyte samples withdrawn from the slimes layer were analyzed by
using Atomic Absorption Spectroscopy (AAS) and specific ion electroanalytical
techniques. In these samples, Pb was determined via AAS using the absorption
line at 283.3 nm.
4
In the Sb determination via AAS, matrix effects can affect the analysis [ioi.
Thus, Sb standard solutions were prepared by adding known amounts of PbSiF
and H SiF so as to match the lead and acid content of the samples. The absorption
line at 231.1 nm was used to determine Sb via AAS. The total Si content of the
entrapped electrolyte was also determined via AAS. Si was determined using the
absorption line at 251.6 nm and a nitrous oxide - acetylene flame. Three different
standard solutions were prepared as follows:
6
[42]
Wenking potentiostat
The total amount of fluorine ions present In the entrapped electrolyte was
estimated by using an ion sensitive electrode \ Standards were prepared by using
NaF solutions. Additionally, calibration curves were obtained using H SiF and
PbSiF -H SiF standards. Gran's plot and standard addition techniques were
incorporated in these measurements [11,12].
2
D. Temperature control
The electrochemical cells previously described were immersed in a constant
temperature water bath. This bath had a volume of @9 L and was covered with
styroform. The bath was stirred using magnetic bars and/or air sparging. The
bath temperature was controlled by using a YSI model 71 temperature controller.
A thermistor probe was used to monitor the bath temperature and immersion
heaters were used to maintain it. This experimental set-up allowed temperature
control within 1.5 C of the set point.
After the bath had reached the set temperature, the assembled
electrochemical cell was introduced and at least 3 hours were allowed for the cell
to reach thermal equilibrium before the electrochemical experiment began.
2
E. Instrumentation
The electrochemical instrumentation involved the use of a variety of electronic
equipment. A Wenking potentiostat was used in the first half of this work and a
Solartron Electrochemical Interface together with a Solartron Frequency response
analyzer were used in the second half. The improvements in the electronics of the
Solartron devices enable complex experiments to be performed. What follows is
a description of the different electrochemical arrangements used. Also, computer
control of the electrochemical experiments and data acquisition will be explained.
/. Wenking potentiostat
1 An Orion Fluoride electrode model 94-09 together with a double junction reference electrode Orion
model 90-02 filled with 1M NaN0 in the outer chamber was used in these measurements.
3
[43]
Wenking potentiostat
terminal panel was used to address the DT2805 board. This panel acts as an
extension of the DT2805 card and simplifies the process of computer interfacing.
All the electrical connections are done in the DT707 terminal panel which sends
and receives information from and to the DT2805 board. Besides from this
terminal panel, a control panel was also used to link the potentiostat, computer,
and electrochemical cell connections.
OSCILLOSCOPE
ELECTROCHEMICAL
CELL
CONTROL
DT
707
PANEL
Fig. 5 Experimental set-up
using the Wenking
potentiostat
SCREW TERMINAL
PANEL
WENKING
IBM XT COMPUTER
ASYST/ASYSTANT
POTENTIOSTAT
SOFTWARE
DT 2805 BOARD
1 Asystant menu-driven software version 1.02, and Asyst command-driven software version 2.10.
[44]
Wenking potentiostat
Fig. 6 Detail of the connections required to Interrupt the current and to follow the cell response
(A) Three terminal cell
(B) Four terminal cell
[45]
Wenking potentiostat
Calibration of the routines used for interrupting the current was done by
using "dummy" cell electrical circuits. Appendix 5 provides a description of these
measurements. The algorithm used in the computer programs to perform the
current interruptions is also described in Appendix 5. By doing these calibration
runs, it was found that the Wenking potentiostat halts the flow of current to the
cell almost immediately (within 10 (isec ). On the other hand, after the short
circuit was opened it was found that the current did not immediately recover its
previous value. The rise time of this process was of the order of milliseconds and
was dependent of the current going through the cell (e.g. see Fig. 7). In addition,
due to hardware and software limitations, the data acquisition system was able
to follow the decay in potential only within 1 msec after current interruption .
Subsequent points were sampled at various acquisition rates. This resulted in
the current interruption routine being able to resolve decays whose time
3
1 As shown in Fig. 6, the reference terminal of the Wenking potentiostat is no longer used for connecting
the reference electrode of the electrochemical cell.
2 Mercury wetted relay Elect-trol model 31511051.
[46]
Generalities:
constants were larger than 10 msec. As the time constant of the diffusion
processes under study were several orders of magnitude larger than 10 msec,
this instrumental constraint was not critical.
158
Fig. 7
Current step
resulting from halting the
flow of current to the
electrochemical cell using
the Wenking potentiostat
Current was halted within
-10 usee.
.88
68.
128
188
T i n e , usee
248
380
[47]
[48]
WE
RE1
RE2
Working
CE
Solartron
Electrochemical
Interface
The dissolution of pure lead and of typical lead anodes was studied using
the beaker electrochemical cell previously described. In these studies, all leads
were shielded to avoid pick-up noise. During the galvanostatic and potentiostatic
experiments current was interrupted using the "sampled" IRg compensation
routine built into the SEI. In these measurements, only the value of the
uncompensated resistance was obtained .
The AC impedance of pure lead and of typical lead anodes was measured
under a variety of experimental conditions. Impedance measurements were
made under rest potential conditions, in the absence and in the presence of a
slimes layer, and in the absence and presence of DC current . All the AC
impedance spectra were obtained using sampling rates and integration times
set to 200 cycles per frequency followed by a 5 sec break . The SEI bandwidth
1
1 The difference in potential between the working and reference electrodes and the current flowing to the
cell are continuously recorded by the SEI. During the AC experiments, the SEI removes the DC component
(if present) from the potential and current waveforms before sending them to the FRA. The fundamental
frequency of these waveforms was used to obtain the impedance of the system.
2 Integration times were reduced by using the auto-integration routine built-in the FRA. A long integration
cycle was chosen (1% error with 90% confidence)
[49]
pH measurements
type chosen for experiments under potentiostatic control was the type E
(maximum bandwidth 24kHz) and for experiments under galvanostatic control
the type B (bandwidth >100KHz).
n. E l e c t r o l y t e P h y s i c o - C h e m i c a l Properties
A . pH measurements
The pH of H SiF containing solutions was measured using a liquid
membrane pH electrode and a double junction reference electrode .
Measurements were made at room temperature. In these measurements, no
attempts were made to correct for liquid junction potentials. pH measurements
were also performed using pH sensitive paper.
2
B. Electrical
conductivity
C. Kinematic viscosity
Kinematic viscosity of H SiF -PbSiF solutions was measured using
Cannon-Fenske routine viscometers for transparent liquids. The viscometers were
2
3 Cell constant = 1.05. Cell constant was obtained by measuring the electrical conductivity of a NaCl
saturated solution and of a 0.100 M KCI solution at different temperatures.
4 Sample volumes vary between 25 and 30 ml.
5 Thermocouples type E Chromel(+)-Constantan(-). A 0 "C reference junction was used during the
temperature measurements.
6 Omega DC controlled solid state relay (SSR) model SSR 240 D10. Control voltage 2-32 Voc.
[50]
Density
D. Density
The density of H SiF -PbSiF solutions was measured using 25 mL
Binghmam-Type pycnometers. Pycnometers were calibrated by measuring the
density of deionized water. Measurements were made at room temperature. From
3 to 5 density measurements were made until the obtained mean density values
were within 0.2%.
2
[51]
Introduction
C h a p t e r 4 E l e c t r o r e f i n i n g of Lead i n a Small-Scale E l e c t r o r e f i n i n g
C e l l : A Case S t u d y
I. I n t r o d u c t i o n
During the electrorefining of lead by the Betts electrorefining process, ionic
concentration gradients become established within the slimes layer due to the
restriction that the slimes layer presents to the movement of ions. Additionally,
the presence of the electric field created by the passage of current contributes to
the formation of concentration gradients within this layer.
The Betts electrorefining process is carried out under nearly galvanostatic
conditions. Under these conditions, a constant flux of lead ions is generated at
the anode/slimes interface. These ions are driven towards the cathode by
diffusion, migration, and convection. In the absence of a slimes layer, mixing of
electrolytes removes the migration and diffusion restrictions. Thus, if pure lead
were to be dissolved, the concentration gradients generated would be present only
within 100 to 1000 |im of the anode/bulk electrolyte interface where mixing
disappears. On the other hand, during the refining of lead bullion, the slimes
layer generates an environment in which ionic concentration gradients can be
present over distances larger than 10 mm. These concentration gradients
contribute to gradients in electrode potential applied to slimes filaments, and
thereby affect the dissolution of noble impurities. Primarily, this gradient arises
due to the larger concentration of lead ions in the vicinity of the anode interface
with respect to their concentration in the bulk electrolyte. Lead ions move out of
the slimes layer mainly by diffusion and migration. As lead ions are released, their
positive charge must be neutralized to achieve local electroneutrality. This leads
to reverse movement of SiF and repulsion of H from the anode/slimes interface.
Electrical neutrality has to be observed throughout the slimes layer and this
dictates a strong relationship between the SiF " , H \ and Pb concentration
gradients.
2
+2
Given the previous relationships, the slimes layer will have a large average
concentration of Pb and SiF " ions and a small average concentration of H ions
with respect to their bulk electrolyte values. Moreover, there will be gradients of
their concentrations throughout the slimes depending on the slimes thickness
and the current density. The interrelationship between the diffusion coefficients,
ionic mobilities, and activity coefficients of the above ions plays a significant role
+2
[52]
Introduction
[53]
Presentation ol Results
n. P r e s e n t a t i o n of R e s u l t s
The study of the dissolution of a typical lead anode under galvanostatic
conditions was done by using the rectangular electrochemical cell described In
Fig. 3.2.
42
^
//
maaauremontm
In mm
Fig. 1 is a top view of the lead anode showing the locations where the potentia
sensing probes were incorporated as well as the point from which electrolyte
samples from the inner slimes electrolyte were withdrawn. Particular
characteristics of this experiment are provided in Table 1. Current interruption
at preset slimes thickness was implemented by using the circuit described in
Figs. 3.5 and 3.6.
Experiment LC2 was carried out in three stages which are described in the
following paragraphs:
Stage I: Galvanostatic dissolution during 300 Hrs. During this stage, current
flow to the cell was halted for 138 msec every 3 Hrs.
Stage II : Long current interruption for 190 Hrs
Stage HI: Further galvanostatic dissolution for 30 min., during which current
was halted every 10 min for 1.8 sec.
[54]
Presentation of Results
During these stages, electrolyte samples were taken from the bulk electrolyte
and from the inner slimes solution at preset times. After the completion of stage
m , current was halted and the corroded anode was left standing in the cell for
48 Hrs. After this time, the electrolyte was slowly drained and a dilute H SiF
solution (pH =1.5) was used to displace the concentrated solution in the slimes
layer. Afterwards, the corroded anode was removed and dried in a vacuum oven
at low temperature (t < 40C). Slimes samples for observation in the SEM were
prepared by using a low viscosity resin and a vacuum imbibition technique 11,21
. Unsupported samples of the same slimes were analyzed by X-ray diffraction.
2
0.01% Sn, 0.02% Cu, 0.14% Bi, 0.25% As, 1.12% Sb, 81 oz/ton Ag.
(Air cooled face exposed to the electrolyte)
Anode Dimensions
Cathode
Reference Electrodes:
Additives Concentration
=2 g/1 aloes, =4 g/1 of lignin sulphonate (added only at the beginning of the
experiment)
Electrolyte Volume
320 ml
Electrolyte Temperature
401.5' C
6 m 1/min
139 Amp/m
Instrumentation
[55]
Stage I
1 Measurements on the slimes electrical conductivity were made by following the difference in potential
between a bare Pt wire inserted in the slimes layer and the lead anode. In these measurements, it was
found that the difference in potential between the Pt wire and the lead anode was negligible. The Pt wire
and the anode appeared to be short-circuited indicating the high electrical conductivity of the slimes
filaments.
2 Oscilloscope readings were consistent with computer measurements.
[56]
Stage/
180
160
*
3
X
Outer
140
Inner A
Inner B
Inner C
* 120
n
+
+
o
ft
100
5 80.
0
0
60.
40.
I *
. tt
tt
tttt
tt
20.
^
00
.0
I
2.
I
4.
I
6.
1
I
+ +
8.
Slimes T h i c k n e s s ,
10
12
14
16
MM
Fig. 2 Anodic overpotential (uncorrected for TJQ) changes as a function of the slimes layer
thickness.
Stage I in Table 1.
Position of the reference electrodes:
Outer: Located in the bulk electrolyte, =5mm away from the slimes/electrolyte Interface
Inner A, B, and C: Located within the lead anode at =3, 6, and 8.5 mm away from the
slimes/electrolyte Interface
[57]
Stage I
200
>
150
a
o
&
100
>
O
o
40
60
100
Fig. 3 Outer and inner A, B, and C reference electrodes anodic overpotential response to current interruptions (during a n
otherwise galvanostatic experiment).
Abscissa values reflect current interruption number. Current interruption measurements where made every 3 Hrs. (i.e. every
0.138 m m slimes). The first current interruption was made i n the absence of slimes and the 100th current interruption was made
at a 13.8 m m slimes thickness.
[58]
Stags/
11
.
(A)
|
100
"0 I
20
(B)
80
80
100
60
?0
(B)
SO
80
100
S3
"
80
90
100
Fig. 5 Detail of the rfo response of the inner B (Fig. A) and inner C (Fig. B) reference
electrodes to current Interruptions (during the whole electrorefining cycle).
[59]
Stage I
6 1
1 2 8
1 8 8
2 4 0
1 2 8
1 8 8
2 4
3 8 8
. 8 8
8 8
6 8 ,
1 2 8
1 8 8
2 4 8
3 8 8
6 8
1 2 8
1 8 8
2 4 8
3 8 8
[60]
Stage/
30.
i-
23.
20.
IS.
10.
S.0
.00
.00
2.0
4.0
i i i6.0
8.0
10.
Slines Thickness, nn
12.
14.
16.
+2
[61]
Stage/
1 2 0
1 8 0
2 4 0
3 0 O
0 0
6 0 .
1 2 0
1 8 0
2 4 0
Fig. 8 Detail of the response to current Interruptions measured by the inner A reference
electrode (at different slimes layer thickness).
X axis: Time, msec
[62]
+2
42 .
38 .
** *
34 . -
38.
26 .
22 .
.80
:
300
74 . -
, .
60 .
1 .
10.2
. 1
128
188
248
MM
388
70 .
66 .
62 ,
58 ,
300
54 .
.00
13.7
MM
1
60 .
120
180
240
300
1 The reference electrode potential is related to the lead ion concentration. Knowledge of the activity
coefficients of the various species that are in the vicinity of these electrodes is required to estimate these
ionic concentrations.
[63]
Stage /
reference electrodes (see Figs. 3 and 5B). The presence of a highly concentrated
Pb region between this electrode and the anode/slimes layer interface can
account for this behavior.
+2
9.8
<B>
26 .
13
22 .
5.0
18 .
1.0
14 .
10 .
-3.0
-7.8
.88
60 ,
180
120
240
300
6.0
.00
40 .
49 .
36 .
45 .
32 .
41 .
28 .
37 .
24 .
33 .
20 .
.00
10.2 nn
8.1 nn
12.3 nn
60 ,
120
180
i l i i I i iI
240
300
29 .
.00
60 ,
120
180
240
300
13.7 nn
60 .
120
180
240
300
[64]
Stage II
180.
2.08
4.00
6.00
8.00
10.0
12.0
16.0
SI i lies Thickness, nn
Fig. 11 Anodic overpotential (corrected for riJ changes as a function of the slimes layer
thickness.
Stage I in Table 1.
Position of the reference electrodes:
Outer: Located in the bulk electrolyte, =5mm away from the slimes/electrolyte interface
Inner A, B, and C: Located within the lead anode at =3, 6, and 8.5 mm away from the
slimes/electrolyte interface
2. Stage II
After forming a 13.8 mm thick slimes layer, current was interrupted for 190
Hrs and the lead anode polarization was followed as a function of time. By
[65]
outer
inner
inner
inner
Outer
Inner B
-40
i-
20
i.i.i.
crrri vrw, m
40
60
80
100
120
Time
Since
Current
140
160
180
200
Interruption, H r s
Fig. 12 Difference in potential (uncorrected for r\^) between the reference electrodes and the lead anode as a
function of the current interruption time.
Stage II in Table 1. After anodic dissolution up to a 13.8 mm thick layer of slimes (Fig. 2) current was halted
for 190 Hrs and the polarization was followed as a function of time.
The arrows indicate the polarization values prior to current interruption.
Details of the polarization decay in thefirstmilliseconds after current interruption can be seen in Figs. 6H, 8H,
9D, and 10D.
Stage III
100
>
90
80
<
.03
'+->
C
CD
-t->
O
Q.
i_
0)
>
O
o
T3
O
C
<
70
'-
'-
'-
'-
60
50
**'
.^.......\~*-r
Stage m in Table 1.
After interrupting the current for
190 Hrs (Fig. 13), current was
applied back to the cell and the
T | values shown in this plot were
obtained.
40
30
20
10
*
t
0
-10
- f - Outer;
!
i
!' Inner fA
:
F Inner iC
,
10
15
20
25
30
Time, min
1 Among these mixed electrochemical processes are: (a) local concentration cells (b) cementation
reactions (c) Re-dissolution of PbF : PbF +2e~ = Pb* +2F~
2
[67]
Stage III
[68]
Fig. 13 shows how the anodic overpotential values increased during the
application of current back to the cell. Concentration gradients within the slimes
layer become established very rapidly: In 30 min only =0.02 mm of slimes were
formed yet the T) values increased by more than 40 m V . This shows that the
slimes layer does hinder appreciably the flow of ions. Current interruptions
during 1.8 sec were applied during this stage to study the characteristics of these
concentration gradients (Fig. 14). At time zero no significant decay in potential
is observed upon current interruption in any of the reference electrodes. As
concentration gradients become established, larger overpotential drops can be
observed upon current interruption.
1
B. Analytical chemistry
1. Bulk electrolyte concentrations
The changes in composition in the bulk electrolyte during Stage I are shown
in Fig. 15. A continuous depletion of Pb in the bulk electrolyte is seen to take
place during the formation of the slimes layer. As the current efficiency for the
refining process was close to 100%, this depletion can only be associated with
Pb concentration enrichment in the electrolyte within the slimes layer.
Concurrently to the Pb depletion, there is a continuous, yet small increase in
acid concentration which is related to H depletion within the slimes layer . No
major changes in the bulk electrolyte concentration of hydrolyzed Si0 were
detected during this stage .
+2
+2
+2
The changes in the bulk electrolyte composition during Stage n are shown
in Fig. 16. Upon current interruption, there is an increase in the Pb bulk
electrolyte composition, and a decrease in the acid concentration. Once current
is interrupted, Pb and SiF " diffuse out of the slimes layer while H diffuses in.
+2
+2
1 In the same period of time, the outer anodic overpotential rose by less than 2 mV in the absence of the
slimes layer (Fig. 2), as compared to a 65 mV rise in the presence of a 13.8 mm thick layer of slimes
(Fig. 11).
2 Material balances indicated that the slimes electrolyte can have average [PbSiF ] higher than 1 M while
[H SiFe] can be lower than 0.5 M.
6
3 Si0 as determined by titration via the LiF-LiOH technique described in Appendix 4 is merely a
composite of all dissolved species containing at least 1 oxygen atom. The general formula Si (OH) F "
with 1<y<4, and'- < 6 accounts for the existence of these species.
2
q
z
[69]
0.9
- .a...
0.8
o
E
\"7
0.7
[H2SiF6]
0.6
CD
O
c
0.5
o
O
03
[Pb+2]
0.4
TS
LU
[Si02]
A/S/02/
_1
"~~"~--0
I
I
8
10
12
1 I
14
L.
1<
Slimes Thickness, mm
0.9
_
o
E
o
08
GL.
0.8
0.7
0.6
CD
O
e
o
O
0.5
9
a
0.4
[H2SiF6]
Fig. 16
Changes in the bulk
electrolyte composition as a function
of the current interruption time
CD
TJ
Stage II in Table 1.
[PbSiF6]
LU
-XCO.
[Si02]
i
20
11
i40i i 60
i i i 80
i i i100
ii
i i ii i i ii i i ii i
120 140 160 180 200
1 1 11
11
[70]
concentrations.
10
Fig. 17 shows the changes in the composition of the inner slimes electrolyte
during stage I as the anode/slimes interface moves away from the sampling
point. As can be seen, Pb concentrations at this point are not as high as the
average values predicted from mass balance computations. Furthermore, the
acid balance is also somewhat larger than expected . However, Pb
concentrations were found to be between 3 and 6 times larger than the
corresponding bulk electrolyte values. Differences between mass balances and
local compositions, may be due to presence of precipitates and also to the fact
that sampling was made only at one point within the slimes layer. Fig. 17 also
shows that negative changes in the acid concentration accompany positive PbSiF
variations. This indicates that SiF " exerts an influence in the transport processes
within the slimes layer. The Si0 concentrations shown in Fig. 17 indicate that
large amounts of these species are contained in the inner slimes electrolyte.
+2
+2
[71]
Analysis of ionic species of noble impurities in this electrolyte showed that AsO
was present in concentrations of -0.17 m M whereas [BiO ] was lower than 0.01
m M . Additionally, [SbO ] in the entrapped electrolyte was of the order of 0.2 m M .
Furthermore, no major changes in the concentration of these noble species were
detected during the refining cycle. These small amounts of noble impurities in
the inner electrolyte indicate that they do not react significantly at the
corresponding overpotential levels shown in Fig. 12 \
+
(A)
1.9
(B)
<
1.8
1.6
1 5
5f
1 4
\ \\\ \v
1.2
Chemical Analysis
10
.. .
0
,
2
10
Fig. 1 Changes in the local concentration of the total Si and F present in the slimes
electrolyte as a function of the movement (from the sampling point) of the anode/slimes
interface
Stage I in Table 1.
Electrolyte samples taken from a fixed point located -3 mm away from the slimes/bulk
electrolyte interface were analyzed for total Si and F using three different analytical
techniques.
(A) Changes in the total concentration of Si-bearing species
(B) Changes in the total concentration of F-bearing species
From chemical analysis data
1 Analysis of the cathode at the end of stage III confirmed that impurities dissolution was not significant.
Cathode impurities concentrations were: 0.0003% Cu, 0.0010% Sb, 0.0016% Bi, <0.0003% Sn,
<0.0001% Ag, and <0.0001% Tl.
[72]
1.2
[H2SiF6J
-0
1 fZT"
1 3 -
Fig. 19
Changes in the local
composition of the slimes electrolyte
as a function of the current
interruption time.
0.8
o
E
c
o
3 0.6
Stage II in Table 1.
From a fixed point located ~3 mm
away from the slimes/bulk electrolyte
interface electrolyte samples were
withdrawn at preset times and
analyzed for the species shown in the
plot.
/S/027
[PbSiF6]
40
80
120
160
200
[73]
(A)
(B)
7.5 | -
7 /
13
O 6.5
1.2 O
5.5
X
20
40
60
80
Chemical Analysis
20
Fig. 20 Changes in the local concentration of the total Si and F present in the slimes
electrolyte as a function of the current interruption time.
Stage n in Table 1.
Electrolyte samples taken from a fixed point located -3 mm away from the slimes/bulk
electrolyte interface were analyzed for total Si and F using three different analytical
techniques.
Total Silica and total fluorine were calculated by adding the concentrations of all the Si and
F-bearing species.
(A) Changes in the total concentration of Si-bearing species
(B) Changes in the total concentration of F-bearing species
From chemical analysis data
C. Characterization
[74]
Corroded electrode
Uncorroded electrode
de/slimes
Interface
Air cooled f a c e
Fig. 22 shows the microstructure of the lead anode ("air" face, location #1
in Fig. 21) \ A variation of the so-called "honeycomb" structure can be observed
in Fig. 22. As found from EPMA , the inner grains in this microstructure have
2
1 This sample was prepared by polishing up to 600 grit followed by 5 urn alumina. Afterwards, the sample
was chemically etched with a polishing-etching solution of the following composition: 20 ml CH COOH
(concentrated), 42 ml H Oj, (30%), 40 ml HN0 (concentrated), and 70 ml of glycerine [ 6 - 8 ] .
3
2 EPMA in etched samples is not recommended as irregular absorption of x-rays resulting from
topography affects the analysis. Thus, the electron probe microanalysis shown in Fig. 22C only indicates
qualitative changes in concentrations.
[75]
(C)
Sb
2
31
23
2
3
2
3
Pb
87
51
62
80
82
81
83
Bi
3
0
0
14
0
8
2
[76]
The changes in the anode microstructure at different parts of the lead anode
can be seen in Figs. 23 and 24 (locations #2 and #3 respectively of Fig. 21) \
E P M A performed on this sample shows that there is less Pb in the grain
boundaries than inside the grains. Furthermore, Bi seems to be present with
lead throughout the whole microstructure. This analysis also shows a eutectic
phase that is rich in As and Sb (points 6 and 7) whereas Sb and Ag-rich phases
can be observed In the proximities of the grain boundaries (points 8 and 9) and
in some cases within the grains. The precipitates within the grains have a random
composition. Most of these precipitates are Pb-rich compounds. Additionally,
Cu-As compounds (not identified in Fig. 23) can also be seen inside the grains.
The microstructures compared in Fig. 24, show the continuity of the honeycomb
structure throughout the sample.
3
1 These samples were not etched. They were prepared by polishing up to 600 grit followed by using 0.5
urn alumina. Backscattered electrons were used to reveal the anode microstructure.
2 As these samples were not etched, relative changes in the probe microanalysis are significant and
represent semi-quantitative changes in the elemental composition of the samples.
[77]
Cu As Ag Sb
1
8
0
0
7
1
0
0
7
1
0
0
1
0
7
0
7
0
0
0
6
0 12 2
0 11 1 13
0
4 28 10
0
6 23 7
1
8
0
0
Pb
87
87
88
89
89
77
73
54
62
87
Bi
5
5
4
4
4
4
3
3
2
5
[78]
SEM analysis
[80]
SEM ana/ysis
Fig. 25 Microstructure of the slimes layer @2mm away from the slimes/electrolyte interface
(position #2 Fig. 21)
Backscattered
Electron
Image
[81]
016929 30KV
X 4 0 0 7 5 u m
Electron
Images
As
Ag
Sb
Pb
Si
1
2
3
12
6
3
8
56
56
79
40
25
14
0
20
35
36
[82]
SEM analysis
Fig. 27 Microstructure of the slimes layer @12mm away from the slimes/electrolyte
interface (position #3 Fig. 21)
Backscattered
Electron
Image
[83]
SEM analysis
(B)
016924
30KV
X400"''
'75um
Electron
Images
15
63
Pb
19
[84]
X-ray diffraction
Inner Slimes
(=12 mm from the slimes-electrolyte
interface)
XXX
XX
XXX
SbA
XX
Bi
XX
Sb
XX
PbF
Si0
Sb 0
2
PbO (yellow)
Ag Sb
3
XX
Cu Sb
XX
AsA
XX
PbSiOj
Bi 0
2
The presence of PbF in these samples and its relative larger concentration
in the inner slimes is a supplementary indication to the presence of Si detected
by EPMA that hydrolysis of SiF " takes place.
The presence of metallic Bi, and Sb, together with some of their oxides was
as expected from electron probe microanalysis . The presence of mtermetallic
compounds (i.e. A^Sb and Cu Sb) was also expected from these analysis.
2
1 With multiple phases present there were overlapping peaks that caused problems with positive,
unambiguous identification.
2 The presence of oxides can be the result of oxidation of the slimes after they were dried.
[85]
1 As explained in Chapters 2 and 4, R, [Qcm ] is the specific resistance of the electrolyte and is related to
T\ by the following relationship: Ti =IR .
2
[86]
Anodic overpotential in the absence of large concentration gradients in the anode boundary
occurs nearly reversibly (n, =0 mV) \ The small magnitude of the generated
ac
(A)
(B)
80
>
i I I I | I i I | I I i
| i
i i | i i i | i i i | i i i | i i i | i I I | i i i
Fig. 1 Potential difference between a fixed reference electrode and a corroding anode In the
absence of addition agents.
Experimental conditions: Galvanostatic Experiment, Current Density= 200 Amp m" ,
Electrode Area 1.50 cm , [PbSiF l=1.31 M , [H SiF ]=0.30 M . T=4Q1.5"C, bulk electrolyte
electrical conductivity K=220 mmhos cm" . Beaker electrochemical cell, electrolyte volume
=300ml, stationary electrolyte. Wenking potentiostat-Data Translation Board-IBM XT
computer.
2
1 iiae determination was done by using AC impedance techniques described in section III.
2 In the case shown in Fig. 1 concentration gradients in the Nernst boundary layer span over a region
between 100 and 1000 urn thick. By comparison, during the refining of impure lead, concentration
gradients are present throughout the whole slimes layer (i.e. the Nernst boundary layer spans over
several mm).
[87]
Correlation between the anodic overpotential and the presence ol addition agen
During the refining of lead, addition agents (i.e. aloes and lignin sulphonate)
are normally used to modify cathodic reactions. The presence of addition agents
can also affect anodic reactions through complex adsorption mechanisms [lj. In
general, the r| values observed during the galvanostatic dissolution of pure lead
were found to increase in the presence of addition agents in the bulk electrolyte .
Moreover, when an excess of these additives is added to the electrolyte, to the
extent that suspended material is visible, they collect on the anode and promote
very large overpotential values. This behavior is shown in Fig. 2A where it can be
seen that in a few hours T| rises from 0 to =1000 mV. In this case, current
interruption measurements indicate thatTj accounts for most of the overpotential
values (Fig. 2B). In the presence of purely resistive films in the anode surface, T)
can be described by the following equation:
A
...1
T] = IR =I(R +R J
a
fi
Rt was found to be equal to 1.4 Qcm from the ohmic drop value obtained a
few seconds after the electrolysis began. Any extra increases in TI are due to the
presence of Rfli . The T j produced by this film along with its resistance are plotted
in Fig. 2B. It seems as if the addition agents form or generate a highly resistive
film that may be counter productive to the refining process . Furthermore, the
presence of this film favours the development of concentration gradients in the
anode/electrolyte interface. Thus, the rj values obtained upon current
interruption increase continuously with time (see inset Fig. 2A) indicating that
the stagnant zone at the anode surface is thickening.
2
[88]
Correlation between the anodic overpotential and the presence of secondary products that precipitate on the anod
(A)
(B)
Fig. 2 Overpotential changes during the galvanostatic dissolution of pure lead (in the
presence of excess quantities of addition agents)
Experimental conditions: Current Density = 180 Amp m" . Electrode Area 31.0 cm ,
[PbSiF ]=1.31 M , [H SiF ]=0.30 M , =2 g 1 aloes and =4 g 1 lignin sulphonate (suspended
material was visible), T=401.5C. Rectangular electrochemical cell, electrolyte volume
=320ml, bulk electrolyte recirculation rate =6 ml min" . Wenking potentiostat-Data
Translation Board-IBM XT computer.
2
fflm
C. Correlation
189]
Establishment of Ionic Concentration Gradients in (he Anode Boundary Layer and their Relationship to the Anodic Ov
soluble PbSiF .4H 0 salt can precipitate \ Fig. 3 shows how the formation of
secondary products can be promoted by dissolution of the anode using high
current densities. rj is the main component of the anodic overpotential values
shown in Fig. 3. Moreover, TJ is related to the porosity and tortuosity factors
resulting from the presence of precipitated products in the anode vicinity.
Concentration overpotential also contributes to the rj increases shown in Fig. 3.
6
6000
0)
E
\
CL
E
<
o
>
CM
5000
<
5^
Anodic Overpotential
4000
Current Density
OJ
3000
c
Q
-1'
CL
i_
2000
OJ
>
O
OJ
1_
13
1000
T5
O
C
O
i._i
1400
2800
Time,
c
'(/)
<
4200
sec
[90]
Establishment ol Ionic Concentration Gradients in the Anode Boundary Layer and their Relationship to the Anodic O
obtained \
The division of the anodic overpotential in its ohmic, activation, and
concentration components was attempted by studying the anodic response of
pure lead to current steps. Prior to the experiment -0.5 mm of the exposed surface
of the working electrode was removed by anodic dissolution at low current density.
This aided in obtaining reproducible results. Variation in rj as a result of the
presence of addition agents was studied using the same cell and electrodes . The
thickness of the hydrodynamic boundary layer was not controlled, yet, by fixing
the recirculation rate of the bulk electrolyte, reproducible results were obtained .
A
CD
c
CD
v_
i_
The transient period for both current rise and fall was
smaller than 10 usee.
(J
Time
Fig. 4 shows the characteristics of the current steps used to study the anodic
behavior of pure lead. After the application of each current step, current was
2 The experiment started by analyzing the T] changes using an additives-free electrolyte. Subsequently,
this electrolyte was slowly substituted with an electrolyte of matching composition but containing addition
agents.
a
3 The hydrodynamic conditions in the vicinity of the anode/electrolyte interface are more nearly a product
of convection than of electrolyte recirculation. However, electrolyte recirculation assured uniform
electrolyte composition between the reference electrode and the anode/electrolyte interface.
[91]
interrupted and the decay potential was followed until it reached its former rest
potential value (=0 mV). Then, the current density was increased and the next
current step was applied.
A. In the absence of addition agents in the bulk electrolyte
Fig. 5 shows the changes in rj due to the application of the current steps.
Right after current is applied there is an abrupt change in n, \ This steep
overpotential change arises due to the presence of the n and can be described
by the following relationship (see also Fig. 6) :
A
r\
(immediately
>
250
<
200 -
from tower
to higher
current
densities
OJ
-*->
150
Q.
i_
CD
>
Increasing
Current
Density
100
x>
o
c:
<
...2
300
50
10
I i i i I
15
Time,
20
25
sec
30
1 The first T I reading was taken 100 msec after the application of the current step. Oscilloscope
measurements at shorter times indicated that the T | value at time zero could be obtained by extrapolation
of the measured data.
a
2 tin prior to and during the application of the current step was determined by using AC impedance
techniques (see Section III). In addition the ohmic drop compensation routine built-in the electrochemical
interface was used to obtain r\ (see Appendix 6).
a
[92]
The Rg value obtained from the slope of the plot shown in Fig. 6, was nearly
equal to the Rg value obtained prior to the application of the current steps. The
activation overpotential, if present, would be observable by a non-zero intercept
or as a curvature near the origin. Evidently, activation overpotential is not
controlling the dissolution of the lead anode.
The T| values compensated for the initial r| (Fig. 6) are presented in Fig. 7.
The overpotential obtained immediately after application of current is very close
to zero (which implies that rj =0 mV) . Subsequent increases in the anodic
overpotential are due to changes in the concentration of ionic species within the
anode boundary layer . Thus, increases in r| are nearly equal to changes in its
rj component.
A
ac
300
IL
240
>
E
CL
o
o
E
O
; I/"
/7
iii i i iiiiiI i i
i i
400
800
ii1
i i i:i ii1
1200
iiiiiii1 ii
1600
200
1 As described by Eq. 2, at time t=0, T| = 0. Thus, the overvoltage at t=0 becomes equal to the initial value
of the activation overpotential [3,p. 356].
C
2 As the anode dissolves, an extra increment to t\ can also be present in these measurements. n,
changes can be neglected during the first seconds after the application of current.
a
[93]
40 r
U
^
<r
S
frt
.
~
OJ
i
1
<
~~~~
:
:
U
j-
o c
j_
35 jf
"
"~~
i
T
'
/
f /
increasing
Current
I
Density
10
15
Time,
20
25
30
sec
The anodic overpotential response during the first seconds after the
application of the current step was modelled by using the analytical solution of
Fick's diffusion equation under unsteady state conditions. To solve this equation
the following assumptions were used:
1) Absence of Migration
2) Absence of Convection
3) Unit Activity coefficients
4) Dissolution of lead is not controlled by kinetics (r| =0).
5) Linear semi-infinite conditions
Description of the boundary conditions required to solve this equation are
provided in Appendix 8 along with its analytical solution. From this solution, the
following relationship between the concentration overpotential and the square
root of time should be observed if the dissolution process is dorninated by diffusion:
ac
eXP
RT
nFC . y
b
p +2
By defining:
[94]
O, =exd
*|_ RT
21
<b =m -jr +b
x
...4
should be obtained.
By using the data shown in Fig. 7, and assuming that rjJO.tJ^A, the plots
shown in Fig. 8 were obtained. In this figure, a linear relationship between <J>j and
the square root of time is limited to times smaller than 1 sec \ From curve fitting
these data to Eq. 4, the data shown in Table 1 were obtained. Correlation
coefficients close to 1 were obtained in almost the whole range of current densities
studied. The average diffusion coefficient, D , agrees with the values reported
in the literature . This diffusion coefficient, while dominated by the movement
of Pb , also includes the effects of other ions (SiF and H among others). Changes
in the transference number of the lead ions will take place when concentration
gradients are fully developed and this will affect the absolute value of this
coefficient . The presence of these concentration gradients is observable in the
analyzed data, as departures from linearity and as changes in the value of the
intercept
from unity .
pb+2
+2
The agreement between the solution of Fick's equation and the results of the
current step experiments confirms that in the absence of addition agents, the
lead dissolution process is controlled by diffusion.
1 As seen in Fig. 8, the larger the applied current density the smaller the linear region between the square
root of time and O,.
2 Diffusion coefficients for lead in HCI0 -Pb(CI0 ) electrolytes at 25 "C reported in the literature are:
4
ro+1
p4+1
4 2
3 The solution of Fick's second law presented in Appendix 8 implies that the transference number of lead
is zero. Any departures from this transference number will affect the obtained diffusion coefficient.
4 Notice also that b decreases as the current density increases.
D
[95]
Table 1 Results of the analysis of the concentration overpotential increases during the first second
after application of the current steps. Fick's Second Law approximation.
Current
Density,
TUry
Amp-m"
50
100
150
250
Coefficient, r
1.00
0.99
^ ^ ^ ^ ^ ^
Regression
0.0371
0.0720
+2
cm -sec
2
0.927
0.979
5.1E-06
5.4E-06
0.983
0.1614
"""""""""""
5.0E-06
6.7E-06
300
400
0.97
0.96
0.1946
0.2775
ii/iiiiiiiiiiii
0.957
6.6E-06
5.8E-06
500
800
0.95
0.91
0.3624
0.5672
0.963
0.951
5.3E-06
5.6E-06
1002
1398
0.89
0.83
0.8205
1.1494
0.964
0.959
4.2E-06
4.1E-06
1598
0.80
1.1921
0.940
5.0E-06
5.30.8xl0
[96]
>
Fig. 9
Decay in the anodic
overpotential (corrected for TIQ) as a
function of the interruption time, t .
<
c
cu
c>
CL
i_
<D
>
O
o
TJ
O
C
<
i iiiiiiii iiiiiiiiiiiiiiiiii i
4
Time, s e c
10
0> = m [V^-VS + ^
D
...5
where:
0 = exp
2
RT
Eq. 5 is applicable only when ti is very small (less than 1 sec) because
convection will stop the thickening of the anode boundary layer. Fick's second
law could not be used to analyze the T| response upon current interruption
A
1 T|n was measured prior to the interruption of current by measuring the impedance at high frequencies
and by using the current interruption routine built into the SEI. Upon current interruption, the observed
ohmic drop coincided with that obtained in these measurements.
[97]
because the boundary conditions are not known precisely. Nevertheless, what the
data in Fig. 9 show is that, upon current interruption, concentration gradients
relax and this relaxation can be followed by monitoring the anodic overpotential
dependance with time.
B. In the presence of addition agents in the bulk electrolyte
If addition agents are added to the bulk electrolyte, the rj response changes
(Fig. 10). Upon subtracting the initial T| value from the r\ readings, the remaining
overpotential was positive (Fig. 11). This rj value decreased during the first
milliseconds after the application of current. After this initial decrease, changes
in TJ were a function of the applied current density. In any case, TJ increased only
up to the point at which convection stops the thickening of the boundary layer.
A
Fig. 10
Anodic overpotential
response (uncorrected for T|Q) of a
pure lead electrode to the current
steps described in Fig. 4.
200
>
E
<
175
150
to the 77
response
current density
from
.K5
to a different
-4'
CL
OJ
>
O
u
step
densities
125
Increasing
Current
Density
100
75
50
<
TJ
-4'
C
OJ
Experimental
conditions:
2.34 cm ,
electrode
area
[PbSiFs]=0.37 M,
[H SiF ]=0.82 M, [SiOJ=0.13 M,
T=401.5"C, =2gl" aloes and
=A g l' lignin sulphonate, beaker
electrochemical cell, electrolyte
volume =310ml, bulk electrolyte
recirculation rate =6 ml min".
Solartron
Electrochemical
Interface-IEEE Card-IBM XT
computer.
25
i 111 1 1 1 1 1 1 1 1 1 1 1
11 i i 11 1 1 i 1 1 11 1 1 i i 1 1 1 11 i 11 1 1 11 i i
10
15
20
25
30
Time, sec
[98]
(A)
(B)
188
. r-
18.8688.
8
8
t
"
' ' ' I' I ' ' ' I ' ' I'
r11 i l i 1111 11 l 1111 i i 11 i 1111 i i 111 i IB
1
1.818
8
.8
81111 1
6
128
.
8
.88
28.6
48.8
(1.8
Time,
sec
88.8
'I ' ' I ' I ' ' I ' ' ' I ' 'I
188
.
24.8
38.8
Time, sec
188.
Fig. 11 Anodic overpotential response (corrected for initial T|J of a pure lead electrode to the
current steps described in Fig. 4
From the
data shown in Fig. 10 the T| values shown in Fig. 12 were subtracted. Each
curve corresponds to the values obtained at different current densities (in Amp m") as
shown to the left of every curve. For the different current densities applied, TJA values at t=0
were as follows:
Q
50
150
250
406
600
10
8.2
9.6
11.5
9.1
(A) Long range variation of the TIA response (corrected for initial TIJ. T| increases up to the
point at which convection stops the thickening of the boundary layer.
(B) Detail of the rj increases.
a
the region where concentration gradients can be found. Upon passage of current,
concentration gradients become established, changing the resistance of the film
and the kinetics for lead dissolution. The non-zero intercept of the anodic
overpotential curves indicates that inhibition is present as a result of this adsorbed
film. On account of the synergistic effect between addition agents, activation
overpotential, and concentration overpotential, no simple analysis of the T]
transients can be performed. Yet, the ri response indicates clearly that the
addition agents affect the anodic reaction for lead in such a way that r| can no
longer be considered to be zero.
A
ac
[99]
288.
- *
Anodic O v e r p o t e n t i a l
(right
after
applying current
step)
168
128 ,
88.8
48.8
.888
.888
' i
128.
I I
_I_I_L
248 .
488
368.
688,
Fig. 12 Comparison between the anodic overpotential value obtained right after application
of current (*) and the uncompensated ohmic drop obtained from the high frequency intercept
of the impedance spectrum (O)
TJA values derived from the data presented in Fig. 10. T| values obtained from the high
frequency AC measurements under rest potential conditions (i\ = IRJ.
FT
[100]
188.
MM M M H M M M M - * M M
4 8 8 . AMP/M
cd =
88 . 8
12 . 8
O v e r p o t e n t i a l decay c a u s e d by
the p r e s e n c e of a d d i t i o n a g e n t s
9.68
T3
C
7 .28
-M
o
muii*i
&
4.88
48 .8
c5
2.48
,0004^-L.
-a
o
<2
.888
' '
.688
' ''
1 . 2 8
1 . 8 8
'
' '
1
2 . 4 8
3 . 8 8
2 8.8
BOO-j-
.888
. 6 8 8
'
1 . 2 8
1 . 8 8
2 . 4 8
3 . 8 8
Time, sec
Fig. 13 Changes in the anodic overpotential upon application of a current step.
From the data presented in Fig. 10.
Upon application of current TJA increases from 0 to 96 mV and decreases slowly afterwards.
From A C measurements under rest potential conditions an Rg value of 2.09 2cm was found.
Thus, Tl =2.092cm * 40 mAmp cm =83.6 mV
The inset plot shows the changes in r\ (mV) as a function of time (sec) after correcting the
anodic overpotential for r\ .
2
Finally, the ohmic drop measured by either fast current interruption or high
frequency AC impedance measurements predicted T) values which were slightly
larger than measured (Fig. 14). The difference between these values is partially
due to the presence of a film resistance,
(so-called "film inhibition').
a
[101]
Introduction
20 .0 -
i
i 6 . 0
Unaccounted
o h n i c d r o p c a u s e d 1>J
t h e p r e s e n c e at a d d i t i o n a g e n t s
12 .0
88. e
s
8 . 8 8
6 6 .0
4 .aa -
II
.000+'
.eee
a.
>.
a 44 . 0
o
cd=
1 1 1 1 1 1 1 1 1 1 1 1 1 1 1i
.800
1.60
li
i * 1 i
2.40
i i
1 i i
1 i
i 1 i
i t 1
4 . 8 8
3.20
400.
2 2 .8
Fig. 14 Changes In the anodic overpotential as a function of the current Interruption time.
Experimental conditions as described in Fig. 10.
After applying current (1=400 Amp m") for 320 sec (t t =320 sec) current was halted and the
overpotential decay was followed as a function of time.
Upon current interruption % decreases abruptly (from 105 to 12 mV). From AC
measurements prior to the interruption of current an R, value of 2.17 Qcm was found.
Thus. Ti =2.17flcm * 40 mAmp cm =86.8 mV.
The inset plot shows the changes in T | (mV) as a function of time (sec) after subtracting T |
from the first r^ reading.
2
HI. A C Impedance
A . Introduction
In the previous section, D C (direct current) transient techniques were used
to study the dissolution of pure lead and its relationship to the components of
the anodic overpotential. In this section, a complementary study using AC
techniques is presented .
1
[102]
1 The bulk electrolyte has a uniform composition and its properties are a function of its geometrical
capacitance C and bulk resistance R . From these values, the dielectric relaxation time of the bulk
electrolyte, T , can be obtained (x =R C ) [7].
2 A sinusoidal current waveform with an amplitude of 21.3 Amp/m R.M.S. was swept from lower to higher
frequencies while the impedance was measured at every frequency. Application of DC currents of the
same order of magnitude of the amplitude of the AC waveform (see Fig. 7) resulted in overpotentials lower
than 5 mV.
g
[103]
obtained indicates that diffusion in the electrode boundary layer is the only
mechanism controlling the dissolution/deposition of Pb . At high frequencies,
+2
the interception of the impedance curve with the Z% axis is equal to the value of
the uncompensated ohmic resistance, R g . When this value was subtracted from
1
the impedance curve, a 45 straight line was observed in the impedance diagram
(Fig. 16) .
2
1xE
--
ra<l/sco
1 .00
2 . 50
6 . 28
I S . 78
D
3
9 . 63
-4
99 . 58
6
298 .
O
628 .
1378 .
1771 .
z
120
z_
. B9B
--
N
I
. 060
. 030
. 000
. O0O
. 320
. 640
Zra*l
.968
ft-CM*
1 .60
1 .28
Fig. 15 Impedance diagram of pure lead under rest potential conditions (in the absence of
addition agents)
Experimental conditions: electrode area 2.34 cm , [PbSiFel=0.35 M, [H SiF ]=0.84 M,
[SiO ]=0.13 M, T=401.5"C, beaker electrochemical cell, electrolyte volume =300ml, bulk
electrolyte recirculation rate =6 ml min" . Solartron Electrochemical Interface-IEEE Card-IBM
XT computer.
Impedance curve obtained under galvanostatic control, A C waveform amplitude 21.3 Amp m'
RM.S.
2
A total of 65 experimental points are plotted. Some of the frequencies (In rad sec") at which
these points were sampled are indicated in the diagram
From the high frequency intercept of this plot with the Zg, axis, the Rg value can be obtained
(Rg= 1.40 Qcm ).
1
1 For a fixed distance between the reference and the working electrodes, this value was very reproducible
(variations were less than 0.5%). These R values were used to obtain r\ prior to the application of the
current steps (see section ll.b).
s
2 Impedance diagrams in which the real part of the impedance is the abscissa, Z, and the negative of
the imaginary part of the impedance is the ordinate, -Z,, are also known as Argand plots. In an Argand
plot the AC frequencies at which the impedance was measured are also shown.
[104]
iaer-xE a
.896 -
Experimental
Analogue node I
.872
rad/sec
N
i.ee
6.28
15.78
39.65
99.58
258.
628.
1578.
1771.
* 2.58
.848
6
0
.824
.eee
.eee
.824
.848
.872
Zreal fi-CM"
i i i
.896
xE 8
I
.128
i i
To model the diffusion processes that take place in the electrode boundarylayer, distributed elements have been used [81 . Among these distributed
elements, the Constant Phase Angle Element (CPE) has been used extensively [71.
The impedance of the CPE element can be described by the foUowing equation:
112
1 All real electrical analogue elements are distributed in space, i.e. their absolute value changes with
position due to their finite size. Diffusion processes are distributed over the electrode boundary layer, and
constitute a classical example of a distributed element [8].
2 A "distributed element" is a component in an analogue model that represents properties of the system
distributed over macro distances, such as ionic concentration gradients across the Nernst diffusion layer.
[105]
[7.8.12]:
...7
Zw,~ = ^(/"w)^
When * =0.5 the Warburg coefficient can be obtained from the Bj value
(Bj = ov). Thus, ^ is a subset of the generalized CPE response. For the Pb/Pb
equilibrium reaction ov is given by the following equation:
7jC
+2
V,- =
When
(n ^-
RT
(nF)
RT
..8
C D
pb
pb
The impedance curve shown in Fig. 16 was curve fitted to Eq. 6 from which
the following values were obtained: B = 0.139 Q cm sec' and ^20=0.5. When these
values were incorporated in Eq. 8, D was found to be equal to 2. lxlO" cm /sec.
For different experiments this number varied by as much as one order of
magnitude . Nevertheless, it was always observed that diffusion was the only
controlling mechanism for lead dissolution/precipitation. Activation polarization
if present would have been observed in the Argand diagram as an arcfromwhich
2
Vzc
pb+2
1 Warburg studied the establishment of concentration gradients in the electrode boundary layer upon
application of an AC voltage. By solving Fick's second law under AC conditions, Warburg found that a
square root frequency dependance of the impedance should be observed if the process was controlled by
diffusion. This square root dependance is equivalent to observing a 45" relationship between the real and
the imaginary components of the impedance (or a
value of 0.5). The assumptions under which
Warburg solved Fick's second law are similar to those described in the previous section to solve the DC
transient case: presence of a supporting electrolyte, unit activity coefficients, absence of convection. Any
departures from these conditions will be reflected in departures from the predicted theory.
2 Notice that Warburg impedance is strictly valid only when 4*^=0.5. Unfortunately, in the study of
diffusion processes by AC techniques, ^ values of 0.5 are the exception rather than the rule [13]. By
introducing a generalized form of the Warburg impedance the physical meaning of each of the involved
parameters may change. For example, variations in the *F value have been related to the presence of
irregularities in the electrode surface at the micrometer level. Also, D
no longer represents an
ZC
M+2
absolute diffusional coefficient but rather an integral value related to all the ionic species present across
the diffusion layer.
3 Lack of well-defined hydrodynamic conditions could be attributed to this large variation in the observed
diffusion coefficients. As can be seen in Eq. 9 when 4 ^ departs from 0.5 it is not possible to obtain the
diffusion coefficient unless the thickness of the Nernst boundary layer is known.
[106]
6.eS|-xE-2
i.ee
2.SB
6.28
13.78
D 39.65
4 99.58
a 258.
0 628.
888.
4> 1487.
y2C
xE-2
4.88
6.88
the exchange current density, i^, and the double layer capacitance, C^, could have
been obtained (to obtain the so-called Randies circuit [161). Evidently, for lead
dissolution in the absence of addition agents rj ->0 (i.e. *-*).
ac
Table 2 Variation of the B, and ^Vjc fitting parameters with current density
Current
Density,
Amp/m
100
150
150
200
Bi,
0.139
0.139
0.074
0.066
0.070
0.056
0.50
0.50
0.43
0.43
0.43
0.43
Qcm sec
2
yzc
AC impedance curves obtained ~2000 sec after the application of the set current density.
The amplitude of the applied AC waveform was set to 21.3 Amp/m RM.S.
The experimental impedance curves were fitted to Eq. 6 using between 60 and 66 points
Regression coefficients and residuals for the regression curves were ( l<co<2000 rad/sec ]:
r >0.96 with 1 y 1 < IO" Q cm
r >0.95 with 1 y 1 < 10 Q cm
2
9t
[107]
observed In the absence of a net DC current. Straight lines with slopes close to
39 were obtained (i.e. *Fzc~0.43) while the B values were between 0.07 and
0.05 Q cm sec'** and appear to change with current density (see Table 2).
x
rad/sec
.863
2 . 40
4
O
V
I
X . 80
. 48
2 . se
I S .78
99 . 3 8
628 .
396S .
25815.
137834.
198786.
81
<*
C
...
N
I
1 . 20
.688
.000
.888
Fig. 18 Impedance diagram of pure lead under rest potential conditions (in the presence of
addition agents)
Experimental conditions: electrode area 2.34 cm , [PbSiF ]=0.37 M, [H SiFe]=0.82 M,
[SiO ]=0.13 M, T=401.5C, =2 g l" aloes and =4 g l" lignin sulphonate, beaker
electrochemical cell, electrolyte volume =310ml, bulk electrolyte recirculation rate =6
ml min' . Solartron Electrochemical Interface-IEEE Card-IBM XT computer. Impedance curve
obtained under galvanostatic control, AC waveform amplitude 21.3 Amp m" RM.S.
2
A total of 130 experimental points were obtained. Some of the frequencies (in rad sec" ) at
which these points were sampled are indicated in the diagram (^=1.74 Qcm ).
1
1 "Inhibition intensity" in Winand's terms [19,20] includes activation overvoltage as well as other
polarizations caused by addition agents.
[108]
The impedance spectrum obtained when addition agents are present in the
bulk electrolyte is shown in Fig. 18. There are big differences between this
spectrum and that obtained in the absence of addition agents (compare Figs. 15
and 18). The presence of two weU-defined arcs can be seen in Fig. 18. The arc
observed at large frequencies ( a 4 0 0 0 rad/sec) is related to the charge transfer
for the Pb/Pb reaction [i^, C^) whereas the arc which spans over frequencies
smaller than 4000 rad/sec is associated with the presence of an adsorbed layer
of addition agents that affect the movement of ions from/towards the electrode
surface (and the region where concentration gradients can be present) \ The Rg
value was not affected by the presence of the addition agents.
+2
8 .08
6 . 4 0
4 . 8 8
Ci
r a d / s e c
. 8 6 3
. 4 0
2 . 5 0
15
. 7 8
21.3
9 9 . 3 8
-4
6 2 8
3 9 6 5
2 5 0 1 5 .
Anp/n
A M P / M
4 . 3
1 5 7 8 3 4 .
2 2 2 9 5 3 .
91
t
...
N
I
. 2 8
. 6 8
. 0 0 0
. 8 8 8
L
4 . 8 8
Zreal
7 . 2 0
.60
12
.0
ft-CM*
Fig. 19 Impedance diagrams of pure lead under rest potential conditions obtained at two
different amplitudes of the applied AC waveform (in the presence of addition agents)
Experimental conditions: as described in Fig. 18.
The AC amplitude values shown in the plot refer to their RM.S. value
A total of 130 experimental points were obtained. Some of the frequencies (in rad sec ) at
which these points were sampled are indicated in the diagram.
Ra. was subtracted from both impedance curves.
]
1 This low frequency arc was only observed after a small Faradaic current was applied (less than 50
Amp/m for 400 sec), after which, it was always present in the impedance curves obtained under rest
potential conditions.
2
[109]
(B)
Fig. 2 0 Analogue circuits used to model the high frequency response of the Impedance
curve shown In Fig. 15.
The analogue shown in (B) is obtained when
small values of the Bj parameter.
ZCPE-^.
(A) Analogue used to represent the high frequency region of the impedance spectrum:
Z(j(6) = R,+
(B) Analogue used to represent processes taking place in the electrical double layer:
[110]
The impedance of the analogue circuit shown in Fig. 20A can be described
as follows :
2
Z(/co)=/?,+
lVJ
...11
The CPE included in the analogue shown in Fig. 20A is used only to subtract
the high frequency part of the arc observed at low frequencies from the charge
transfer arc. This CPE is not intended to represent any particular process but
only to subtract the higher frequency data.
The values of the analogue elements obtained from the curve fitting process
are as follows (Fig. 21):
R = 1.665 Qcm (5%), Cd,=20.6 fiF/cm (3%),
B!=149.25 Qcm sec ' (5%), and ^=0.74(12%). From the Revalue ^ was found
to be equal to 82 Amp/m (5%). For the same electrode these values were
reproduced within 15%. On the other hand, small changes in the bulk electrolyte
composition, electrolyte temperature, bulk electrolyte recirculation rate, and
electrode roughness appreciably affected the and C^ values. In any case, for a
fixed electrolyte composition, i values were not smaller than 70 Amp/m
(70<(o<500 Amp/m ) whereas C^ values were between 18 and 30 |F/cm . Since
in the absence of addition agents a charge transfer arc was not found (io-x), it
is concluded that the presence of the addition agents affects significantly the
kinetics for the Pb/Pb equilibrium reaction. The decrease in the *<, values as a
result of the presence of addition agents can be attributed to: (A) changes in the
2
ct
>l zc
+2
1 The Stern-Geary equation has also been used to relate the steady-state corrosion current density, 4
and the polarization resistance, R [25]:
p
\ 2.303(8,
'< + BI)
PP
^corr
' 1_
2 Analytical representation of the impedance was obtained by using the Laplace plane techniques
described in Appendix 3.
[Ill]
Fig. 21 High frequency section of the impedance diagram shown In Fig. 18.
Impedance curve was fitted to the electrical analogue shown In Fig. 20A (R. was computed
from the high frequency Intercept of the Impedance curve with the Z, axis).
From the curve fitting process, the following values were obtained: B,= 149.250 Q cm sec" ,
^=0.741, F^l.SSSQcm , 0^=20.6 uF cm .
2
zc
Some frequencies (in rad sec" ) are shown for both the experimental (solid line) and the
regressed data (dotted line).
1
Quality of Fit Parameters (51 experimental points were fitted to obtain the values of 4
parameters):
rVO.973, l y a l ^ ^ x l O Q cm
rVO.941 I y 1 =6.16xl0" Q cm
1
^=0.985 ly^l^l.iexlO^Q'cm
electrochemically active surface area and (B) variations in the current distribution
in the anode vicinity. An effective decrease in the value (considering that the
electxochemlcally active surface area is the same in the cases presented in Figs.
16 and 21) would mean that the kinetics for lead dissolution and deposition have
become "less" reversible.
[112]
1 .68
1 .28
rad/sec
.863
.48
2 .58
15 .78
9 9 .38
4
628 .
3965 .
25815.
0
157834.
198786.
Rest
In
Potential
the presence
Conditions cd = 8
o f a c d = 188
flny/h'
a
91
.888
, 488
.A
.888
.888
J L_l_
. 788
- i
L_
.48
Zreal
2.18
ft-CM*
2 .88
3 .58
Fig. 2 2 Impedance diagrams of pure lead obtained in the presence and in the absence of a
net Faradaic current (in the presence of addition agents)
Experimental conditions: As described in Fig. 18.
The AC amplitude was the same in both cases (21.3 Amp m
RM.S)
[113]
increases *' . The decrease in the size of the low frequency arc may be the result
of a defined boundary layer whose size has compacted due to the presence of a
fixed electric field created by the passage of a net Faradaic current .
a
.see
A
_
. 168
4
a
O
. 128
- *
rid/tic
.863
.31
1 .58
7.91
3 9 .63
199 .
996 .
4991 .
9959 .
15783.
In
the presence
of a c d
= 58
In
the presence
of
= IBB
a cd
Anp/n>
fthp/h'
91
p.
. 888
.848
.888
.888
. 868
TJ
128
Zreal
.188
ft-CM*
I I l
, 248
. 388
Fig. 23 Detail of the Impedance diagrams obtained In the presence of a net Faradaic current
(in the presence of addition agents)
Experimental conditions: As described in Fig. 18.
The AC amplitude was the same in both cases (21.3 Amp m" R.M.S)
2
1 The exchange current density is also a function of the concentration of the electroactive ion in the
electrode surface [22-24].
2 Notice how the radius of the high frequency arc obtained in the presence of a net Faradaic current
(^-.08 Qcm , Fig. 23) is at least 20 times smaller than under rest potential conditions (Ret~1 -6 Qcm ,
Fig. 21). The "apparent" charge transfer resistance decreases until at high current densities R ~0. The
term "apparent" charge transfer resistance refers to the R* values that may be observable upon passage
of a net Faradaic current when studying irreversible systems using AC techniques.
2
ct
3 Notice that addition agents are also dispersed by the anodic process because: (A) the interface is
retreating, and (B) there is a flux of Pb in the opposite direction. Thus, decreases in the size of the low
frequency arc as the current density increases can be due to a depletion of addition agents in the anode
boundary layer.
+2
[114]
+2
All experiments shown in this Chapter were performed atT=40l .5C . using
the beaker electrochemical cell (Fig. 3.1) and the Solartron equipment (Fig. 3.8).
Data acquisition and control of the experiment were performed using an IBM XT
personal computer via the IEEE interface (Fig. 3.7). Electrodes were prepared from
the same anode (anode A, Fig. 3.3) as described in section 3.1 .B. Electrolyte was
recirculated continuously at 6 rnl/min (cell volumes varied between 300 and 320
ml). The geometrical area of the electrodes was used to compute the current
density and the impedance per unit area. In experiments in which addition agents
were added to the electrolyte, any insoluble precipitates were removed prior to
the introduction of the electrolyte to the cell. The assumed concentration of
addition agents prior to the filtering operation was 2 g/1 aloes and 4 g/1 lignin
sulphonate. The compositions of the electrolytes used in the various experiments
presented in this Chapter are shown in Table 1.
1
1 Anode composition: 0.01% Sn, 0.02% Cu, 0.14% Bi, 0.25% As, 1.12% Sb, 81 oz/ton Ag.
[115]
Introduction
Table
Experiment
Number
Electrode
Face
Addition
Agents
CA2
mould
yes
CA4
mould
[PbSiF ] [HaSiFe]
mol/1
mol/1
yes
0.27
0.45
0.69
0.77
Bulk
electrolyte
mol/1
[SiOJ
Conductivity,
K,atT=40 C,
mmhos/cm
Characteristics of
the experiment
0.14
300
Galvanostatic
dissolution at
1=200 Amp/m
Potentiostatic
dissolution at
0.14
315
Econtn>l=220 mV
mould
CA5
CA6
mould
air
CC1
air
CC2
yes
yes
no
no
0.45
0.45
0.35
0.36
0.76
0.73
0.84
0.83
0.14
Galvanostatic
dissolution at
1=800 Amp/m
Galvanostatic
dissolution at
1=200 Amp/m
Galvanostatic
dissolution at
1=200 Amp/m
Galvanostatic
experiment.
315
0.14
320
345
0.09
345
0.09
In the presence of a net Faradaic current the impedance spectra were obtained under galvanostatic control
except in Exp. CA4 in which the curves were obtained under potentiostatic control. Under current
interruption conditions all the impedance spectra were obtained under Galvanostatic control.
1 Under galvanostatic conditions application of AC frequencies in excess of 3x10 rad/sec often resulted in
phase shifts produced by the Solartron Electrochemical Interface due to bandwidth limitations. On the
other hand, under potentiostatic control, frequencies as high as 4x10 rad/sec could be applied without
observing phase shifts. A phase shift is a displacement of the capacitative component of the impedance
curve towards negative values [1-3].
4
2 For example if 0.063<ox6300 there are l o g ^ j = 6 decades of frequency swept and the impedance is
measured at 6x20=120 discrete points.
[116]
Table2
Frequency Range
O W rad/sec
o w . rad/sec
Time, Hrs
0.063
408407
l.l
0.63
408407
.30
n. A C Impedance C h a r a c t e r i z a t i o n of t h e S t a r t i n g W o r k i n g
Electrodes.
Prior to the anodic dissolution of lead bullion electrodes their impedance was
obtained under rest potential conditions. Table 3 summarizes the characteristics
of the obtained spectra . The differences between the various spectra are
explained in the next paragraphs.
1
Table 3 Summary of the values of the electrical analogue parameters obtained under rest potential conditions
Frequency Range
Experiment Addition
and Sweep Agents
Number
Pot/Gal
Control
rad/sec
rad/sec
R.,
Qcm
Ra,
Qcm
On,
uFcm
CA2-7
yes
Gal
560
99588
1.42
CA4-7
CA5-7
yes
yes
llltllll
Pot
315
560
250159
111740
1.02
1.19
CA5-4
CA6-2
yes
yes
Gal
Gal
560
177
111740
157834
1.20
1.916
CC1-5
CC1-6
no
no
Gal
Gal
628
628
15783
44482
0.90
0.89
CC2-7
CC2-2
no
no
Gal
Gal
560
560
17709
28067
0.85
0.85
0.501
64
iilslii!
llliill
65.30
0.725
48.9
32.6
Qcm sec'"
^zc
Amp/m
23.71
0.64
269
1X45
23.90
0.39
0.64
72
186
liiiilll
35.5
llliill
2.88
0.56
0.55
157
70
0.57
0.45
1.88
1.88
0.59
0.59
1 The statistical parameters related to the quality of the curve fitting procedure are presented in
Appendix 9.
[117]
The impedance curve shown in Fig. 1 was curve-fitted to Eq. 1 to obtain the
Bi and
values. B and *zc were found to be equal to 2.88 Q. cm sec* and 0.45
respectively. By comparing these values with those obtained when pure lead was
studied (Bi-0.14 Qcm sec" ' and ^^-0.5) it is seen that
remains practically
unchanged, while Bj shows a marked increase in its absolute value . The increase
Vzc
M ZC7
1 Due to software limitations in the graphics program the axes of the Argand plots are not marked
-Zj as in the main text but as Zreal and -Zimag respectively.
and
aB
= B V0i)*
l
= B <Q
l
thus,
Za^w^cos^Vzc
Z = -B (o" ysin^
zc
2c
m = tan-Yzc
from which:
V
z c
= -tan
Thus, it can be seen that *P does not depend on B, which is only a multiplying factor (i.e. the impedance
curve shrinks or contracts according to its value).
ZC
[118]
. 138 p i t
8
A.
ExperiMental
flnaloout
nodel
. 128
a
a
.888
.888
'
628.
1117.
1987.
3533.
283.
11173
19869
33333
28866
44482
Zred, Qcm
.288
Fig. 1 Impedance diagram of a typical lead bullion electrode (Exp. CC1-6) under rest
potential conditions and In the absence of addition agents.
R, was subtracted from the Z* component of the impedance (Fv=0.89 Qcm ).
Some frequencies (in rad/sec) are shown for both the experimental (solid line) and the curve
fitting data (dotted line).
The impedance curve was fitted to a CPE element from which the following values were obtained
B!=2.88 Qcm sec"* ^=0.45.
2
ZC
1 As a result of surface irregularities at the micrometer level, the thickness of the electrode boundary layer
may not be uniform and that results in increases in the B, value.
1119]
the diffusion coefficients cannot be obtained the shape of the impedance curves
unequivocally indicates that there are no charge transfer limitations for the
dissolution/precipitation of Pb from/to the lead bullion electrodes.
1
+2
. D O
*
*
1 .68
4
a
O
rad/sec
* 063
.40
2 .50
15.78
99.58
628.
3965.
23813.
62832.
99588.
. 488
ana
1.28
I I
1.88
I I
I I l_l I
2.48
3.88
' I
i i
' I
4.28
Fig. 2 Impedance diagram (Argand plot) of a typical lead bullion electrode (Exp. C A2-7) under rest potential
conditions and in the presence of addition agents (R, was not subtracted from Z^)
1 The lack of knowledge of the electrochemically active surface area poses a serious hindrance for the
computation of the diffusion coefficient: the penetration depth of the AC wave and the thickness of the
boundary layer have to be uniform across the electrode in the whole range of frequencies for meaningful
diffusion coefficient values to be obtained.
[120]
^ , + fl (/Q>)
e
Z(/co) =/?,+-
The Rg values were obtained directly from the high frequency intercept of the
impedance curves with the
axis. The four remaining parameters in Eq. 2 (R^,
C^, B and ^zc). were obtained by curve fitting the experimental data to Eq. 2.
l t
a
*
a
o
>
368 .
11.17 .
2229 .
4448 .
8B76 .
177B9.
33333.
71364.
62B32.
99339.
a
So
csf
aa
_L_i_ J
I I
I I
L_
. 42
'
. . .
'
. . .
Z^,, Qcm,2
2
Fig. 3 Detail of the high frequency region of the Impedance diagram shown in Fig. 2
Rg was subtracted from the Z* component of the impedance (R =1.42 Qcm ).
Some frequencies (in rad/sec) are shown for both the experimental (solid line) and the
curve fitting data (dotted line).
The impedance curve was fitted to the analogue circuit shown in Fig. 5.20 from which
the following values were obtained B,=23.71 Q cm sec'*^. ^=0.64, Fc^O.501 Qcm , and 0^=64
uF cm' .
2
1 For a description of the characteristics of this circuit see Chapter 5 section III.c.
2 Again, it is worth repeating that the presence of a CPE in the circuit shown in Fig. 5.20 does not
represent a purely diffusional process and is included only to aid in the curve fitting procedure to obtain
the R and C , values.
d
[121]
GB IxE
r A d / f i c
*
D
a
a
*
Ol
-J1Illl1lll
.00
.20
sea.
1117 .
3339 .
4448 .
8876
177B
3333
7 8 9 8
7 0 3 8
1117
.
9
5
4
4
4
.
.
.
.
0.
xE
I l I 1 I I I I I l l I I l l l I l l l I l l l I l l i_J
.48
.68
.80
l . O
Fig. 4 Detail of the high frequency regions of the impedance diagrams obtained under
rest potential conditions
(A) Exp. CA4-I fB) Exp. CA5-J
The values of the derived analogue values are shown In Table 3.
[122]
In experiments CA4, CA5, and CA6 the electrolyte composition was kept
constant, yet, the derived kinetic parameters varied widely \ These variations
can be attributed to differences in the electrode roughness among the different
electrodes . In any case, the exchange current densities were not lower than 70
Amp/m (70<io<270 Amp/m ) while the double layer capacitances varied between
25 and 66 |F/cm . These large values of ^ indicate that dissolution of the lead
present in the lead rich phases takes place under nearly reversible conditions
and that the noble impurities present in the lead bullion are not significantly
affecting the kinetics for lead dissolution (or deposition).
a
aa
ea
rad/se o
. B63
.4a
2 . 30
IS . 78
99 . s e
28 .
3965 .
238 13.
63833.
99388.
4
O
-*-
Potvirit
i os-t.-axt.ic:
GAIvAnosiatlo
Control
Control
. 4 o e
Zrea,, Qcm
Fig. 5 Impedance spectra obtained under potentiostatic (solid line) and galvanostatic
control (dashed line).
From Exp. CAS (Sweeps CA5-1 and CA5-4 in Table 3).
1 The same electrolyte was used in Exps. CA4, CA5, and CA6 but since the amount of solids filtered prior
to the introduction of electrolyte to the cell varied in each case, the final addition agent content may vary.
2 Small changes in the bulk electrolyte recirculation rate, electrode microstructure, and cell temperature
may also produce the observed changes in the parameter values.
3 Notice that large capacitance values are associated with high exchange current densities indicating that
electrochemically active surface areas are different in every case.
[123]
Variation of trie anodic overpotential as a function of the electrolysis time and the current interru
1 Theoretically both curves should have been identical if the processes under study behaves linearly.
Instrumental artifacts (i.e. Potentiostat bandwidth) may also contribute to the observed discrepancies.
5 mV R.M.S does not exactly produce a sinusoidal current waveform of 35 Amp/m of amplitude and vice
versa.
2
2 The fraction of electronic current going through the slimes filaments was assumed to be negligible
compared to the amount of current crossing the anode/slimes interface. The validity of this assumption
was confirmed by measuring the distance between the anode/slimes electrolyte interface and the
slimes/bulk electrolyte interface at the end of the experiment which was in agreement with the computed
value. Furthermore, the slimes composition does not seem to change much with slimes thickness.
[124]
Variation of the anodic overpotential as a function of the electrolysis time and the current interrupti
(A)
(B)
1
acu
-M
o
&
1>
o
o
1.88
3.SB
3.4B
mm Slimes
7.28
9.88
(C)
-.BBS
1.88
3.68
3,48
7.28
9.88
mm Slimes
70B .
I
73
C
-M
o
&
360
420
Si
o
200 .
140 .
. 00
Fig. 6 Overpotential response of a typical lead bullion anode as a function of the slimes
thickness (Exp. CA2).
The AC impedance was measured at preset slimes thicknesses indicated by the spikes
shown in the curves.
CA) Anodic overpotential as a function of the slimes thickness (uncorrected for initial
(B) Cell potential as a function of the slimes thickness (uncorrected for initial TIJ.
(C) Anodic overpotential measured by the counter and reference electrodes as a function
of the slimes thickness (corrected for Initial TIJ.
[125]
Variation of the anodic overpotential as a function of the electrolysis time and the current interru
1 Abrupt dissolution of phases containing noble impurities was not observed in the case study presented
in Chapter 4 because a low current density was applied, the electrode was larger, only a 14 mm thick
slimes layer was formed, and the maximum value of T| was <200 mV.
2 The potentiostat follows continuously (among other parameters) the difference in potential between the
reference electrode and the current. At finite sampling times (i.e once every 3 min.), these data are
digitized and saved as "DC" data. If during the digitization process the AC waveform was being applied, a
net DC overpotential and DC current may be observed as "spikes".
3 A 400% increase in the current density (from 200 to 800 Amp/m) resulted in a corresponding reduction
in the amount of lead that could be removed before impurities dissolve at an excessive rate (i.e. at similar
T1A values, the equivalent amount of lead dissolved is ~4 times smaller).
a
[126]
Variation of tie anodic overpotential as a function ol the electrolysis time and the current interruptio
.88
.aa
.68
1.2
1.8
rnrn Slimes
2 .4
3.8
Fig. 7 Anodic overpotential (corrected for initial TIJ measured by the counter and
reference electrodes as a function of the slimes thickness (Exp. CAS).
Galvanostatic conditions 1=800 Amp-m
Detween the slimes and the electrolyte promoting their earlier dissolution. In
addition, passage of larger currents increases the "ohmic" drop component of
the anodic overpotential \
Fig. 8A shows the changes in current density as a function of the
electrolysis time during a potentiostatic experiment (Exp. CA4 in Table 1). In
this experiment the dissolution of noble impurities present in the slimes layer
was restricted by liiinting the potential difference between the reference electrode
and the lead anode to 220 mV. From the numerical integration of the data
presented in Fig. 8A, the amount of lead dissolved as a function of the electrolysis
time was found (Fig. 8B). At short electrolysis times, large current densities can
flow because ionic transport proceeds relatively unhindered. As the slimes layer
thickens, its presence restrains the flow of current up to the point at which only
very small currents can flow through the cell. These decreases in current are
[127]
Variation ol the anodic overpotential as a function of the electrolysis time and the current interrupti
(A)
2.S <r'3<
. 00
. Ea
1.2
JL .
mm Dissolved
2.4
3.a
Fig. 8 Current density changes as a function of the electrolysis time and of the amount
of lead dissolved (Exp. CA4. potentiostatic conditions E
=220 mV)
control
IA) Changes in the anodic current density as a function of the electrolysis time
(B) Changes in the amount of lead dissolved as a function of the electrolysis time. This
curve was obtained from numerical integration of the data presented in CA)
(C) Current density changes as a function of the amount of lead dissolved
[128]
Variation of the anodic overpotential as a function of the electrolysis time and the current inter
also the result of the presence of secondary products which block the movement
of ions even further until the lead dissolution process is nearly halted (Fig. 8C).
At long electrolysis times, lead ions are still being generated whereas
concentration overpotential no longer changes and the precipitation of
secondary products must take place for rj to remain constant.
During experiment CA4 the difference in potential between the reference
and working electrodes was kept constant and equal to 220 mV ( ^ ^ , = 2 2 0
mV). As a result of the presence of r| between such electrodes, the potential
difference applied between the working electrode and the slimes/bulk electrolyte
interface is not constant. The anodic overpotential changes as a function of Rg
and of the current density can be described by the following equation:
c
mm Dissolved
1129]
Variation of the anodic overpotential as a function of the electrolysis time and the current interruptio
(B)
786
788
568
548
281
f-H
148
S
<u
o
&
228
68.1
I
48.8
-188
.88
72.8
Time, Hrs.
.88
.32
.64
.u.!...,.M;
o
24.8
428
18.
(C)
28.
.96
1.3
f
38.
1.6
M
46.
38.
Time, Hrs.
I
73
fi
<u
-M
O
&
I
T3
O
Time, Hrs.
Fig. 10 Anodic overpotential changes upon current interruption (Exps. CA2, CA5, and CA4).
T | prior to current interruption was not subtracted from these measurements
n
The AC impedance was measured at preset times indicated by the spikes shown in the curves.
(A) From Exp. CA2
[130]
Variation of the anodic overpotential as a function ol tie electrolysis time and the current interrup
1 The AC wave may have changed the conditions at the anode/slimes electrolyte interface by promoting
precipitation and/or hydrolysis reactions at that interface. Under these conditions, Faradaic currents can
divert to the slimes filaments and cause the excursions in potential shown in Fig. 11.
[131]
e r
Variation of the anodic overpotential as a function ol the electrolysis time and the current intenvptio
(A)
7 8 0
c-
.0
(B)
7 B B
. pr
5 6 8
4 2 0
e
*
o
a.
01
2 8 0
. 0 0 0
.
0 0 0
t
_J
i. .
80
. 60
S l l n e s
T h i c k n e s s ,
40
1_
nn
Fig. 11 Anodic overpotential (corrected for initial r\^j as a function of the slimes
thickness (Exps. CA6 and CC1).
(A) From Exp. CA6
CC1
[132]
Changes in the impedance as a function ol the slimes layer thickness and of the current interruption
788
668 .
548'
560< r
428
388
428
189;
68.
.88
T| prior to current
interruption was not
subtracted from these
measurements
The AC impedance was
measured at preset times
indicated by the spikes shown
in the curves.
288
.82
.84
.86
.88
.18
1481
128
Tine, Hrs
1133]
Changes in the impedance as a function olthe slimes layer thickness and ol the current interrup
4. 0
r
* ad/s e c
M M
A
B
C
m-
1 . 56
7.91
39 . 65
3. 2
2. 4
nE
F
G
I
199 .
996 .
4991 .
J
K
<
-
.31
s
SI i ne 0. 8
1. 6
2. 2
3. 1
3. 7
4. 7
5. 3
6. 0
6. 6
7. 2
7. 8
15783.
25015.
ft
c
N
1.6
000
5 . 40
00
ft-CM*
F i g . 1 3 Impedance spectra obtained during Exp. CA2 at slimes layer thicknesses between 0.8 and 7.8 mm.
Each impedance curve was obtained at a different slimes thickness as indicated in this Argand plot
[134]
Changes in the impedance as a function ol the slimes layer thickness and ol the current interruption
(A)
2 . a a
r a d / s e a
. 0 6 3
. 3 5
*
1 . 3 9
1 1 . 1 7
62 . 8 3
-
3 5 3 .
1987 .
O
1 1 1 7 4 .
3 9 6 4 7 .
6 2 8 3 2 .
x .6 8
1 . 2 B
Dl
t
N
. 8 0 0
. 4 0 0
8 0 0
'III11III
1 . 2 8
1
1_
(B)
- . 8 3 0
0
1
u
91
158
Fig. 14 Impedance spectra obtained during Exp. CA2 at a slimes layer thickness of 8.4 mm
i
1 In a Bode plot, the high frequency arcs are better resolved by analyzing the variations in the phase
angle (dotted line, right vertical axis) as a function of the logarithm of the frequency.
[135]
Changes in the impedance as a function ol the slimes layer thickness and of the current interruption
(A)
2
0 0
A
X .
6 0
r A d / l c
. 0 6 3
. 4 4
3 . 13
2 2 . 2 9
1 5 8 .
1 1 1 7 .
7 9 1 1
.
3 6 8 8 2 .
2 2 2 9 3 3 .
3 5 3 3 5 4 .
ti
O
1 . 2 8
bi
4
E
N
1
1 1 1 1 1 1 1 1 1 11
0 0 0
4 0 0
0 0 0
2 . 4 0
1 i
1 i
4 . 0 8
1 i
1 i
5 . 6 8
1 i
~"~~^
Z r e a l
1 i i
7 . 2 8
1 i
i
18 .
8 . 8 8
ft-CM*
(B)
.
"i"
0 3 0
030/
T 2
-
^ ^
'
"~
"
'
"
*
N
a . 6
1
-
- ^_
- - o^fo
X .
**
X .
1 - ' '
X.
- . , 1 5 0
>^
\
>
4 .
'
\ .
*- ^
2 . 4
- 1 .4 8
- . 8 8 8
1 . 4 8
L O G
<
F r e < i ,
t t
2 . 8 8
r a d / s e c
'
ii
'
'
4 . 2 0
1 1 II 1
51
rad
Phase A n g l e ,
1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1 1
10
i
5 .6 0
Fig. 15 Impedance spectrum obtained during Exp. CA2 at a slimes layer thickness of
8.65 mm .
l
[136]
Changes in the impedance as a function of the slimes layer thickness and ol the current interruption
3.90
*
O
Fran
Fron
OC D * t *
Current Interruption
3 .38
2 .70
2 . 10
0"
41
*
<
s
tl
&
s
e
o
e
1 . SO
= ft 6
i
.90S
- .000
1.80
3.60
5.40
SIines Thickness, n n
7 .20
9 . 00
Fig. 16 Changes In the value of R,, as a function of the slimes thickness (Exp. CA2).
From the AC and DC data obtained i n Exp. CA2.
1 Current interruption was done prior to obtaining the AC impedance spectrum. Appendix 6 describes how
the current interruption measurement was implemented.
2 The invariability of the R, values at slimes thicknesses smaller than 8 mm confirms the results obtained
in Chapter 4 in which upon current interruption t\ remained constant (Fig. 3.8).
a
[137]
Changes in the impedance as a function ol the slimes layer thickness and ol the current interwptio
1.2
6 .3
7
49 . 9
141 .
396 .
1117 .
n n
&
3149 .
8876 .
14067 .
ft
B
C
D
. 96
A.
72
ft c
48
>
*
J*
22295.
0.63
0.
1.
1.
1.
MG
64
20
76
89
rad/sec
000
400
800
1.20
Zreal , ft CM
mm.
1 . 60
. 00
F i g . 17 Impedance spectra obtained during Exp. CA5 at slimes layer thicknesses between 0.64 and 1.89
Each impedance curve was obtained at a different slimes thicknesses as indicated in this Argand plot. R,
was subtracted from the Z component of the impedance.
[138]
Changes in the impedance as a function of the slimes layer thickness and of the current interruption
(A)
(B)
3.8 r-
.Mr-
; riJAtc
- A 63
.64
>
4
&
0
.48
Qf
.888
mm
I I
J.M
9.96
39,63
158.
628.
2381.
9959,
39646.
S68B2.
I I I
II
I I
II
I I
.248
II
.488
I 1
Z^j,
II
II
I t
.728
II
I I
IIIIII
.968
,2
I I
1.8
-.588
L
1.28
flcm
.588
1.58
2.58
3.58
4.58
Fig. 18 Impedance spectra obtained during Exp. CA5 at a slimes layer thickness of 2.2
(R, was subtracted from the Z* component of the Impedance).
(A) Argand plot fB) Bode plot
i.ee
.eee
.6ee
l.ae
Slimes Thickness,
l.se
mm
2.48
3.98
[139]
Changes in the impedance as a function ol the slimes layer thickness and of the current intanvption
(A)
I
r a d l / s e c
1.
a4
68
O
1.
28
D i s o I v e
nn
.63
A .99
6 .28
19 . 87
62 . 83
199
.
628 .
1987 .
6284 .
7911
.
ft
B
C
dl
8 . 87
1 .39
1 .77
2 . 86
2 . IS
2 . 52
n
E
F
. 888
. 488
.888
. 888
1.
2 . 48
28
3 .
real.
68
4 . 88
ftcm*
6.
88
(B)
688 ,
Dittol
nn
G
H
I
2
2
2
2
2
.
.
.
.
.
va1
76
834
837
841
865
368 .
A.
248 .
tk
O
4
>
. 863
. 44
3 . 15
22 . 29
158 .
1117 .
7911
.
56882.
396469.
488487.
I
208 .
488.
608.
880 ,
_L
Fig. 20 Impedance spectra obtained during Exp. CA4 at slimes layer thickness between
0.87 and 2.87 mm.
Each impedance curve was obtained at a different slimes thickness as indicated in this
Argand plot. Ft, was subtracted from the Z* component of the impedance.
[140]
Changes in tie impedance as a function of the slimes layer thickness and of the current mtemiptio
While the impedance curves obtained during Exps. CA2 and CA5 increased
uniforrnly with the slimes thickness, those obtained during Exp. CA4 showed
marked increases and variations in their magnitude and shape (Fig. 20). This
different behaviour is believed to be the result of the precipitation of secondary
products across the slimes layer. Thus, impedance arcs whose time constant
is large are present throughout the whole electrolysis cycle . The absence of
high frequency arcs is a clear indication that Faradaic dissolution of noble
impurities did not occur in this experiment.
1
1.30
r-
"a
U
1.24
1.18
cj
O
J..J.2
T 3
1.06
o
C
1.00
.000
j i_
.600
_]
1.20
1.80
l_
2.40
3.00
mm Dissolved
Changes in Rg as a function of the amount of lead dissolved are shown in
Fig. 21 . Again, minor variations in the Rg values result from changes in the
concentration of the electrolyte between the reference electrode and the
slimes/bulk electrolyte interface. As lower currents go through the cell Rg
returns to its original value because such variations in concentration disappear.
2
1 As indicated by the frequency values at which the capacitative pan" (i.e. the imaginary part) of the
impedance reaches a maximum value, the relaxation processes nave very large time constants (of the
order of sec).
2 These changes in R were used to compute the anodic overpotential as a function of the amount of lead
dissolved (Fig. 9).
9
[141]
Changes in tie impedance as a function of the slimes layer thickness and of the current interruption
2 .00
1 .0
1 .28
a
a
rad/sec
17 .7
62 .8
223 .
791 .
2807 .
99S9 .
3S33S.
125375.
222953.
353354.
After
8 . 17
Hrs
After
86
Hrs
.9
.880
.488
6-
.000
2 .88
4----'-Q-
2 .68
3.28
Zreal
3.88
ft-CH*
l_
I
4.48
5 .88
Fig. 22 Argand plot showing the changes in the impedance spectra obtained in the presence of a layer
of slimes and in the absence of a net Faradaic current (Exp.
CA2).
The anode was corroded up to the formation of -8.7 mm of slimes (Fig. 6). Subsequently, current was
halted for -115 hrs (Fig. 10A). During this period, impedance spectra were obtained at preset times as
indicated in this Argand plot.
Curve A was obtained -0.17 Hrs after current interruption
Curve B was obtained -86.9 Hrs after current interruption
Changes in the impedance as a function of tie slimes layer thickness and of the current interrupt
o ,. .
s
3 .28
2.98
2.68i
OJ
4-1
CO
s.
6
o
o
2.38
2.88
.888
24.8
48.8
72.8
Time, Hrs.
. sa
r a d / s e
. 6 3
. 4 0
_
2 . 5 0
9 . 9 6
3 9 . 6 5
1
6
2
9
3
4
o
. 3 0
3
2
5
9
9
4
8
8
0
5
6
4
.
.
1
9
4
8
128 .
*
4
96.8
.
.
6 .
5 .
A f t a r
. 6 3
H r s
A f t e r
7 . 8 8
H r s
A f t e r
3 8
H r s
. 4
. 2 8
. 1 0
. 0 0
1 . 4
1 . 7
2 . 0
2 . 2
Z^u, Qcm
2 . 5
2 . 8
Fig. 24 Argand plot showing the changes in the impedance spectra obtained in the presence of a layer of slimes
and in the absence of a net Faradaic current (Exp. C A5).
The anode was corroded up to the formation of -2.2 mm of slimes (Fig. 7). Subsequently, current was
halted for -46 hrs (Fig. 10B). During this interval, impedance spectra were obtained at preset times as indicated
in this Argand plot.
Curve A was obtained -0.63 Hrs after current interruption
Curve B was obtained -7.88 Hrs after current interruption
Curve C was obtained -38.4 Hrs after current interruption
[143]
Changes in the impedance as a function of the stones layer thickness and of the current interrup
18.a
2a.a
3e.e
48.a
se.e
Time, Hrs.
Qcm ).
2
with plates of large area separated by a small gap, the whole being in a vacuum whereas C is the
capacitance of a parallel plate condenser when an isotropic material is present between the plates [5].
0
[144]
Changes in the impedance as a function of the slimes layer thickness and ol the current interruption
[145]
Changes in the impedance as a function of the slimes layer thickness and of the current mtenvption
(A)
a.
a
4
a
*1 . 2
8 0
4 0
r a d / s e c
. 6 3
2 . s a
9 . 9 6
3 9 . 6 5
1 5 8 .
6 2 8 .
2 5 8 1 .
9 9 5 9 .
1 V 7 8 9 .
2 8 8 6 8 .
rtn
A
B
C
D
E
F
G
H
I
J
X
J i i i
(B)
Z ^ , Ocm
LJ
S I t
.
.
1
2
3
S
5
6
6
7
7
i i
n e s
4 3
9 8
. 4 7
. 3 9
. B 6
. 3 7
. 7 9
. 3 2
. 6 4
. 8 6
. 4 9
L
5 . 0
(C)
Qcm
Fig. 27 Impedance spectra obtained during Exp. CA6 at slimes layer thicknesses
between 0.43 and 717 mm.
Zreai,
Each impedance curve was obtained at a different slimes thickness as indicated in this
Argand plot. R, was subtracted from the
component of the impedance.
(A) Impedance spectra acquired in the region where potential excursions where not observed
(Fig. 11A).
OB) Impedance spectrum obtained at ~3.7 mm of slimes
(C) Impedance spectrum obtained at ~7.7 mm of slimes
[146]
Changes in the impedance as a function of the slimes layer thickness and of the current interwp
In Exps. CA6 and CC1, changes in Rg were observed only in the region
where the anodic overpotential excursions occur (Fig. 29). Upon current
interruption, the impedance spectra show the presence of high frequency arcs
whose size decreases as the current interruption time increases (Fig. 30). Also,
Rg decreases as a function of the current interruption time up to a limiting value
which in the case of Exp. CA6 is nearly equal to that observed at the beginning
of the experiment (compare Figs. 29A and 31).
1 Exps. CA6 and CC1 were performed using working electrodes from different sides of the lead anode
and in the presence and absence of addition agents (see Table 1). In both experiments the potential
excursions appeared at about the same slimes thickness (-3.2 mm). Consequently, the outset of the
excursions in potential must be related to changes in the slimes electrolyte rather than in the slimes layer
or in the anode.
2 In Figs. 27B, 27C, 28B and 28C, the maximum of the imaginary part occurs at 2x10 <co<3x10 rad/sec,
thus 5x10" <T<3.3X10^ sec.
4
[147]
Changes in the impedance as a function olthe slimes layer thickness and of the current interrupt
(A)
1 . 0 8 r
.880
4
O
.608
rd/sec
6.3
1 7 .7
4 9 .9
141 .
396 .
1117 .
3149 .
8876 .
11174 .
17789.
i 8 . 6 3
nn S i t
. 14
.51
1 .37
2 .22
3 .50
B
C
D
H I S
rad/sec
.480
.280
i
.8884
. 000
I i i i I i i i I i i i I
.330
.640
. 960
Z ^ . Qcm
(B)
1 .60
(C)
8.68
rid/sec
6.3
AC tueei dene it
t 25.8
99.6
396.
0 1378.
6.48
4 6283.
6 25813.
0 99582.
396463,
* 488487
4.88
1 . 28
8,88
4.5 m Slinei
6.48
4.88
rid/sec
6.3
AC tvtep dene it
25.8
99.6
396.
0 1378.
4 6283.
0
2581S.
0 99582.
396463,
* 488487
5.3
nn Slinet
CJ
C!
3.28 -
3.28
1.68 -
1.68
,800
.888
1 1 1 1 1
12.8
Z ^ . Qcm
mm.
' ' ' '' ' ' ' ' ' '' '
2.48
4.88
1
1 1
9.68
I I I I I
12.8
Fig. 28 Impedance spectra obtained during Exp. CC1 at slimes layer thicknesses between 0.14 and 5.3
Each impedance curve was obtained at a different slimes thickness as indicated in this Argand plot. R,
was subtracted from the
component of the impedance.
(A) Impedance spectra acquired in the region where potential excursions where not observed (Fig. 11B) (B)
Impedance spectrum obtained at -4.5 mm of slimes (C) Impedance spectrum obtained at -5.3 mm of slimes
[148]
Changes in the impedance as a function of the slimes layer thickness and of the current interruption
and
Fig. 29 Changes in the value of the R, as a function of the slimes thickness (Exps. CA6
CCD.
(A) From the AC and DC data obtained in Exp. CA6
(B) From the AC data obtained in Exp. CC1
[149]
Changes in the impedance as a function oltte slimes layer thickness and of the current interruption t
4
o
1 .
CJ
. 6 3
3 . I S
13 . 7 8
7 9 . 11
3 9 6 .
1 9 8 7
.
9 9 3 9 .
4 9 9 1 8 .
1 3 7 8 3 4 .
3 3 8 1 3 2 .
A4> t
r-
8 . 6 3
H r .
A f
t e r
2 . 4 4
H r s
A f
t a r
1 1 2 . 8
H r
8 8
Fig. 30 Argand plot showing the changes In the Impedance spectra obtained in the
presence of a layer of slimes and in the absence of a net Faradaic current [Exp. CA6).
The anode was corroded up to the formation of -8 mm of slimes (Fig. 11A).
Subsequently, current was halted for -113 hrs (Fig. 12). During this interval, impedance
spectra were obtained at preset times as indicated in this Argand plot.
Cury.e A was obtained -0.63 Hrs after current interruption
Curve B was obtained -2.44 Hrs after current interruption
Curve C was obtained -112.8 Hrs after current interruption
6 9
u
a
ee
08 !
ea i
ea
Ohrj
ns
P .
oC J '
a
P
.08
24 .
48 .
72.
Time. Hrs.
96.
120
[150]
x >2
+b
...5
With
1 Despite the different definitions of concentration overpotential available in the literature there is no
conclusive evidence that any of them describes accurately the physical phenomena involved, yet, there is
an agreement that because of the presence of concentration gradients, an ohmic drop is included in the
concentration overpotential measurement (Compare refs. [23-24] and [8-9]).
2 The ohmic part in Eq. 5 of TI has more physical meaning when it is equal to /
0
t h a n when it is
equal to i^Jj^-^dx
[151]
Under steady state conditions , Eq. 4 indicates that small changes in the
applied current density ought to result in changes in TJ due exclusively to the
ohmic component of the slimes electrolyte . If linearity between applied current
density and the observed concentration overpotential is observed, the average
resistivity of the slimes electrolyte, F^, could be obtained from Eq. 5. In addition,
the parameter b should provide complementary information about the extent
to which concentration gradients vary across the slimes layer .
a
1 As explained in Chapter 4, even though the external current may had been halted, processes that
support the passage of internal currents may still be present after current interruption.
2 Steady state conditions are such that the concentration gradients across the slimes layer remain
constant during the measurement of the T|-I relationship.
3 As Eq. 4 indicates, Ohm's law is not useful in a region where concentration gradients are present.
Nevertheless, in such a region, an integral value of the changes in conductivity can be obtained.
4 Notice that R and b are a function of the local electrolyte conductivity and concentration gradients
across the slimes layer.
5 AC and DC currents and overpotentials are terms that can easily be confused: An AC wave varies
sinusoidally as a function of time and can be described by the following equation:
C
where: M is the amplitude of the waveform, co is its frequency, * is the phase angle, and t is the time. By
knowing M , , and <B, the DC "instantaneous" component of the AC waveform can be obtained.
e
[152]
Data analysis
The steps that were followed to analyze the rj spikes produced by the AC
current waveform are as follows:
A) From the TJ readings obtained during the application of the AC current
waveform, r) , was subtracted *.
B) The resulting anodic overpotential and the current density were
curve-fitted to a straight line according to Eq. 5.
A
1 The exact values for the amplitude and frequency of the waveform may change depending on whether
a oc current is applied or not.
2 This is equivalent to implying that all the current flow is at the anode/slimes interface without any
significant Faradaic current crossing the slimes/electrolyte interface.
3 R, is known prior to and after the application of the current steps. An average Ft, value can be used to
correct the r\ readings.
4 Notice that the anodic overpotentials shown in Figs. 6C, 7, and 11 are all corrected only for the initialr\
Onn=IR)- The approach in this section is to analyze the anodic overpotential corrected forr| present at
the local time the current sweeps were applied.
A
[153]
Data analysis
(A)
(B)
M.lr
268.
236.
212,
cu
Q
ti
188,
U
38.8
164.
i i i t
12,1
i i i
i i i
12.3
i i i
i t i
12.6
i i i
Ii
i i
Ii
12.9
i i
i i i
33,0'
13.5
13.2
i i i i i i i i i
12.8
12.3
i i i i i
i
12.6
12.9
13.2
13.3
Time, Hrs.
Time, Hrs.
(C)
38.8
* *
>-> 48.8
a
C
38.8
<U
&
I...
^ "
11
M I r 111111111111
.888
52.8
''''''
184.
111111111111'
156.
288.
268.
[154]
Data analysis
The current density changes observed during the application of one of the
AC sweeps (Exp. CA2) are shown in Fig. 32A. The anodic overpotential spikes
(corrected for TJQ) that result from variations in the cell current are shown in
Fig. 32B. The linear relationship between these two quantities can be seen in
Fig. 32C. This relationship was found to be frequency independent, indicating
nearly steady state conditions. Further analysis of the r\ spikes produced
during the application of the AC waveform at different slimes thickness showed
that linearity was observed only up to anodic overpotentials smaller than 250
mV (Fig. 33).
A
240
200
160
. h
120
8 B .
40
. 0
140
360
320
280
240
200
160
164.
1 8 8 .
236
261
2 6 0 .
236
140
1 6 4 ,
188.
2 1 2 .
2 3 6 .
3 6 0 .
[155]
Data analysis
Table 4 Analysis of the spikes produced during the application of the A C waveform,
in the presence of a net DC current (Exp. CA2, Figs. 32 to 35)
Parameters Derived from Regression
Analysis Computations
Slimes
Slope
Intercept,
Qcm2 b, (mV)
Thickness
mm
|y | ,mv
2
From Eq. 4
3
b_
p,
m
QcmIR..
IR,
Experimental
mV
mV
mV
0.62
30.8
2.54
12.1
0.80
0.850
6.8
42.9
7.78
41.9
45.8
2.23
6.4
28.1
1.45
0 986
1.63
73.9
73.3
6.48
54.4
3.10
41.8
12.6
96.1
95.5
0.988
2.15
1.30
6.93
5 33
75.3
59.8
76 5
150.4
151.8
0.981
7.38
3 93
098
5.95
82.4
163.7
88.8
171.2
170.0
0.961
7.68
4.57
0.93
0.920
7.98
93.1
1017
194.8
193 3
6.56
5 23
375.1
092
0.905
7.95
113.9
110.9
224 8
7.18
633.5
221.8
5.71
1.03
0888
9.27
124.9
265.2
2614
7 79
140 3
1192.5
7 22
089
0.051
5.90
4.96
3.74
361.0
96.5
8.41
4386.0
457.5
436.0
0.558
10 86
9 40
360.6
1.97
543 3
520.0
5688.7
1827
8 65
Between 18 and 20 experimental points were used to obtain the regression coefficient. These points
were collected during -55 min and correspond to digitized samples taken when the frequency of the applied
AC waves was between 0.063 and 6.3 rad/sec.
Abstracted from Table 4, Appendix 9
[156]
M e
hi
.eaa
i I i I i i I i i I i i I i I i i I i I i I
l.sa
s.se
s.4a
7.2a
9.ea
i i i
.aaa
i i
l.sa
i i I i
3.6a
i i
i i
3.4a
i i
7.2s
i i
9.aa
mm Slimes
mm Slimes
Fig. 34 Changes In the value of the resistance of the slimes electrolyte as a
function of the slimes thickness (Exp. CA2, Table 4)
(A) R,,, (Qcm ) changes (B) p (Qcm) variations
2
.eee
I . S B
3.6a
s.4e
7.2a
9.00
mm Slimes
Fig. 35 Changes in the parameters b and IR with relationship to the experimental
variations of the anodic overpotential (Exp. CA2, Table 4).
[157]
Data analysis
Table 5 . A Analysis of the spikes produced during the application of the AC waveform.
in the presence of a net DC current (Exp. CA6)
From Eq. 4
Slope
Rn..
Qcm
Intercept,
b, (mV)
|y | ,
mV
2
p , Qcm
m
, mV
mV
m m m
Experimental
mV
5.14
32.0
0.17
3.56
0.14
1.96
4.75
57.8
3.06
1.52
0.99
1.59
4.98
0.40
76.6
78.3
5.37
15.64
3.52
1341
189
0.85
2.55
130.5
6.64
2.70
113.6
0 88
10 16
216
4.06
52.5
168.2,
166.1
7.49
3.13
140.9
0.97
4.09
2.31
4.18
60.9
201.9
205.2
4 experimental points were used to obtain the regression coefficient These points were collected during
-12 min and correspond to digitized samples taken when the frequency of the applied AC waves was between
0.63 and 6.3 rad/sec.
Abstracted from Table 5.a, Appendix 9
il?
3 Use of linear polarization techniques such as SACV (small amplitude cyclic voltammetry) should provide
the same results [10-13].
[158]
Data analysis
Table 5.B Analysis of the spikes produced during the application of the
AC waveform,
Slope R ,
Qcm
Intercept, b,
(mV)
2.23
3.62
3.83
1.64
0.99
60.8
0.89
0.91
0.95
0.63
2.44
3.83
43.07
112.8
iiii^iiii;:
44.5
lllliiliillli
|y r\mv
p , Qcm
21.95
41.19
62.43
22.09
21.63
2.82
4.57
4.84
88
Experimental
TR^+b, mV
mV
60.8
48.4
44.5
5.3
0.2
82.1
WmmMi
44.0
9 experimental points were used to obtain the regression coefficient These points were collected
during -27 min and correspond to
digitized samples taken when the frequency of the applied AC waves was
0.2
between 0.63 and 6.3 rad/sec.
Abstracted from Table 5.b, Appendix 9
The TJ spikes observed in Exps. CA6 and CC1 were also analyzed according
a
to Eq. 5. The results obtained from such an analysis are shown in Tables 5A
and 6. A linear relationship between overpotential and current was only found
in the region where the potential excursions were absent. Furthermore, a
quasi-linear relationship between overpotential and current was also found
under current interruption conditions (Table 5B, Exp. CA6) *.
Table 6 Analysis of the spikes produced during the application of the
AC waveform,
1.37
1.79
2.22
2.65
3.07
3.50
0.52
0.63
0.74
0.81
0.92
0.96
32.4
38.0
42.8
48.0
52.4
58.4
|y l ,
mV
2
b
lR
p , Qcm IRm.mV
IR.,+6,
mV
Experimental
mV
0.98
0.97
0.98
0.95
0.93
0.93
0.1
0.3
0.4
1.1
1.7
2.3
3.21
3.11
2.97
3.06
2.94
3.12
3.80
3.51
3.34
3.05
2.99
2.74
10.1
12.2
14.4
15.7
17.8
18.7
42.6
50.2
57.2
Mmm
70.3
il7!lll:
42.8
50.6
57.8
64.4
70.9
78.1
5 experimental points were used to obtain the regression coefficient These points were collected during
-12 min and correspond to digitized samples taken when the frequency of the applied AC waves was between
0.63 and 6.3 rad/sec.
1 In Table 5B, r and |y| vary widely yet the predicted (IR +o) and experimental (T| ) overpotentials are
nearly equal.
2
[159]
Data analysis
iii) A relationship that relates the slope b to the average slimes electrolyte
resistance. Rn,, and to the local electrolyte concentrations cannot be inferred
yet \
The observed differences in the values of the r| components (IRn, and b)
at a fixed slimes thickness are related to the bulk electrolyte composition, to
the presence of addition agents, and to the changes in the slimes
physico-chemical properties (i.e. porosity, tortuosity).
A
this ratio, the smaller the restriction for the movement of electrolyte across the
slimes layer and the lower the observed overpotential. In Exp. CC1 the ^- ratio
is ~3 and remains nearly constant. On the other hand, at similar slimes
thickness, for the other experiments this ratio shows marked decreases (from
2.5 to 1.3 for Exp. CA2 and from 3.4 to 1.6 for Exp. CA6).
1 i.e. b values cannot be derived from R values and vice versa : Average electrical conductivities may be
equal yet local concentrations may be different.
2 The amount and nature of addition agents incorporated during the refining of Pb using a sulphamic acid
electrolyte has also been shown to have a strong impact on the permeability of the anode slimes [25].
3 Notice that whereas Exps. CA2 and CA6 were carried out using the lead anode "mould" cooled face,
Exp. CC1 was performed using the "air" cooled face. The microstructures of these electrodes were found
to be similar. Thus, the anodic overpotential variations can hardly be related to microstructural differences
between the different electrodes.
m
[160]
interruption conditions 1=0 and rj does not have an ohmic component, yet, by
A
the electrolyte present within the slimes was found to have a different
conductivity than the bulk electrolyte.
(b) Proposed analogue representation of a lead bullion electrode covered
with a layer of slimes.
So far in this thesis it has been shown that the response of electrochemical
systems to A C waveforms can be used to obtain kinetic and diffusional
parameters. Resistors, capacitors, and CPE's have been used to link changes
in the AC spectra to associated physico-chemical parameters. A general model
is to be proposed to analyze the observed AC impedance spectra. This model is
based on a set of assumptions. From this model and from the experimental
data, several electrical analogues are developed to examine the AC impedance
data and to find the link between these analogues and the physical phenomena
they may represent.
Fig. 36 shows a general analogue model of a lead bullion electrode covered
with a layer of slimes. Six interfaces can be identified in this figure. Each one
of these interfaces has an associated impedance :
2
1 i.e. the pseudo-steady state observed during the relaxation of the concentration gradients in the
entrained electrolyte was slowly displaced by applying a sinusoidal current waveform.
2 The impedance of the reference electrode is neglected in this analogue.
[161]
a / 8
(O=0
where:
TJ (t) = overpotential (compensated for rid observed upon passage of current
as a function of the electrolysis time. By using a reference electrode reversible
to Pb no corrections for liquid junction at the reference electrode/electrolyte
interface have to be incorporated in the T\(t) values.
+2
[162]
By analyzing the path of the DC current on its way from the anode to the
bulk electrolyte a deeper insight in the Z^. components can be obtained. Thus,
as shown in Fig. 36 there are two main paths for the DC current: One through
the anode/slimes electrolyte interface and the second through the anode/slimes
interface. If current enters the slimes filaments it can either leave them through
the bulk electrolyte (path A) or through the slimes electrolyte (path C). On the
other hand if any current crosses the anode/slimes interface it can either go
across the slimes electrolyte (path B) or return to the anode via a ground loop
(path D).
For current to go through the slimes filaments it will first have to overcome
a resistance associated with Z . As the slimes filaments were found to be
grounded to the anode
such resistance must be negligible. Any current
entering the slimes filaments can only produce Faradaic work at the
slimes/slimes electrolyte interface if and only if the activation energy barrier of
such process is overcome. This requires large overpotentials and ionic gradients
across the slimes layer. If any current diverts towards the slimes filaments it
can either go across the slimes/slimes electrolyte interface (path A) or across
the slimes/bulk electrolyte interface (path C). The experimental evidence is that
Faradaic reactions of the slimes compounds are insignificant at overpotentials
less than 200 mV. For example, the amount of bismuth corroded is less than
30 ppm, too small to account for significant Bi corrosion currents crossing the
slimes filaments. Thus, during most of the electrorefining cycle, any current
going through the anode/slimes interface can only result in charging of the
a/Sl
1 Tja is equal to the current density, I, times the uncompensated ohmic resistance, R, (n =IR ).
n
[163]
Proposed analogue representation ola lead bullion electrode covered with a layer of sl
Anode/Slimes
i
i
Slimes/Electrolyte
Interface
Interface
i
i
Zsl/be
S l i m e s : f i l a m e n t
sl/se
W
5 *
L
llllllllllillllllllllllllll
a/se
S
Bulk
Electrolyte
Slimes
Layer
[164]
Proposed analogue representation of a had bullion electrode covered with a layer ol sli
electrical double layer of the slimes filaments *and are insignificant. Thus, from
this analysis it can be assumed that for the D C case, up to r\(t) values equal to
200 mV: Z i->0, Z ^ - ^ , and Z
a/a
al/be
-*>.
From the previous description it can be assumed that almost all of the
applied D C current flows across the anode/slimes electrolyte interface and
continues in its way towards the bulk electrolyte by overcoming the diffusional
impedance Z^r (path B). All this current is transferred to the electrolyte mainly
as a result of the Faradaic transfer of Pb to the slimes electrolyte. For such a
transfer to take place, an energy barrier associated with Z
has to be overcome.
Such impedance is small (lead dissolves almost reversibly), yet it can increase
if conditions at the interface change (i.e. as a result of the presence of secondary
products blocking the interface). Once Pb ions are transferred to the slimes
electrolyte, an ionic current is established. The resistance that the ions find in
their movement across the slimes is a function of Zw which is an intensive and
position dependent quantity. The larger the effective distance the ions have to
travel before reaching the bulk electrolyte the larger this impedance. The
complex migraUonal/diffusional processes taking place across the slimes
electrolyte can be described by Z ^ . Furthermore, if as a result of the movement
of ions across the slimes electrolyte/bulk electrolyte interface a diffusion layer
is established, a semi-infinite Warburg impedance has to be included as well,
although for well mixed bulk electrolytes its presence can be neglected. Once
the Pb ions reach the bulk electrolyte they are transferred to the cathode by
convection, migration and diffusion.
+2
a / s e
+2
+2
s ] / b e
1 Notice that under DC conditions a capacitor will act as an open circuit and will have an infinite
impedance. The impedance of a capacitor, Zc, is given by: z =^, thus at co=0, Z,.-*.
c
[165]
Proposed analogue representation ola lead bullion electrode covered with a layer of sl
current can only be related to changes in the value of Z^. A relationship between
the observed b values and Zy, may also exist, yet, impedance values obtained
at frequencies other than zero have to be provided (see Eq, 9).
The previously introduced DC model also must be consistent with the
experimental evidence observed under current interruption conditions.
Depending on the extent to which concentration gradients had been established
prior to the current interruption, an overpotential decay will always be observed.
Eventually, the electrolyte compositions inside and outside the slimes electrolyte
equilibrate, ionic movement virtually stops, and the impedance disappears as
the overpotential vanishes (i.e. as n . ^ - 0). On the other hand, as a result of
the ionic concentration differences between the slimes electrolyte/bulk
electrolyte and the anode/slimes electrolyte interfaces, a series of concentration
cells may be established because electronic current can flow across the slimes
filaments (path D). This process may consist of (as an example) continued
corrosion of the lead anode, accompanied by deposition of noble elements in
the electrolyte, as Bi , SbO , or Ag* on slimes filaments. These internal currents
may be too small or difficult to measure, however, they can affect the
characteristics of the overpotential decay and consequently of the impedance.
As concentration gradients disappear, the rate of these internal processes
declines, up to the point at which they become negligible.
+3
[166]
analogue model will be proposed. The validity of the proposed model will be
tested by analyzing the changes in impedance in the presence and in the absence
of a net Faradaic D C current.
In the previous analysis it was found that all the applied D C current can
be assumed to flow across the anode/slimes electrolyte interface. On the other
hand, a superimposed AC wave can either cross the same interface and/or divert
through the slimes filaments. If it diverts to the slimes filaments, it can cross
the slimes/slimes electrolyte interface without actually perfonriing any Faradaic
work. This will happen if the slimes layer can be considered to have only a
capacitative component. For these processes to take place, the AC waveform will
have to produce alternating potential fields at the slimes/slimes electrolyte
double layer interface. This will result in large capacitative effects being observed
in the impedance spectrum. Thus, if any AC current crosses the slimes/slimes
electrolyte interface, equivalent impedances related to the capacitative part of
the slimes impedance should be observed even in the absence of a net Faradaic
current. As the experimental impedance spectra obtained under current
interruption conditions did not indicate the presence of large interfacial areas,
AC current transfer across the slimes layer in the presence of a net Faradaic
current can be neglected.
The experimental data indicate that while a net D C current is being applied,
formation of Pb is the preferred reaction. Such a reaction takes place without
a significant energy expenditure and is the path of least resistance for the flow
of current. If this is the path for least resistance for D C current, it can also be
assumed to be the path of least resistance for the AC current. Thus, assurning
that all the AC current crosses this interface, the overall impedance will have
only three components: Z , 7^,, Zw (notice the similarity between the D C and
the AC cases) . Accordingly, in the absence of a net Faradaic current, the polarity
of the double layer can be switched and/or its potential difference changed
simultaneously in the slimes/electrolyte and the anode/slimes interface. Only
then, the capacitative and resistive process associated with these impedances
can be observed as changes in the impedance of the system. Consequently,
+a
a/ae
1 Notice that in the worst scenario in which in the presence of a DC current the AC current actually crosses
the slimes/slimes electrolyte interface and causes Faradaic reactions, a wide non-uniform current
distribution can result. If this had taken place the system would had been changed to the extent that
steady state concentration gradients during the AC measurements would not have been observed, the
system would have oscillated to the extent that the impedance measurement could not have been taken.
[167]
1 These abrupt impedance changes upon Faradaic reaction of the slimes filaments can be seen by
comparing Figs. 13 and 14.
2 Isothermal temperature of the slimes electrolyte can be assumed on the basis of sufficient thermal
conductivity of H SiF -PbSiF solutions and of the large porosity of the slimes layer, contributing to
convection.
2
[168]
Data analysis
5) In the presence of a net Faradaic current the lead anode and the slimes
layer are equipotential.
6) Changes in the microstructure of the slimes layer as a result of blockage
of ionic flow (i.e. by the precipitation of secondary products) can be incorporated
within the Warburg diffusional impedance or by additional impedances
connected in series with it \
7) Blockage of the anode/slimes electrolyte interface (i.e. by the
precipitation of secondary products) inhibits charge transfer processes and
increases the Faradaic impedance, Z .
a / s e
9l/se
8 l / b e
1 Non-uniform porosity across the slimes layer can be expected when secondary products precipitate or
re-dissolve.
2 Capacitance effects associated with the noble compounds present in the slimes layer are more
significant as the TI approaches values larger than 200 mV (i.e. in the region where their Faradaic
reaction can occur).
a
[169]
(i) Case I: impedance spectra obtained in the presence of a net Faradaic current
The impedance spectra obtained in Exps. C A 2 . C A 5 . C A 6 , and C C 1
were found to be described accurately by two different yet related electrical
circuits (see Fig. 37). The values of the parameters associated with the
proposed electrical circuits were obtained by curve fitting the experimental
data to the theoretical impedance functions.
1
CPEL
= b,m^'
ZCPE, = Wmf*
CPEO
Kijnf
-+B U^'
-ir+
z=
Rj
2
...io
with
R
/?2
and
D =
1
D =
2
x,
and
\ = D!^
1 As presented in the various Tables shown in Appendix 9 "described accurately" means that from a
statistical regression analysis perspective, good correlation existed between the experimental and the
curve-fitted data. The number of parameters involved in the curve fitting routine was never in excess to
that required to obtain significant fits according to statistical rules (i.e. as derived from ANOVA statistical
tables).
2 This process required the use of complex non-linear square fitting routines . The Zj component of the
total impedance was used to obtain the values of individual parameters in the different electrical circuits.
Once an initial set of values was obtained, they were used to compute the Z component of the
impedance and improve the accuracy of the fitting process.
3 For an in-depth review of the characteristics of the ZARC circuits see ref. [is].
[170]
Data analysis
Circuit A . l
"ZARCl
W
Ffgwrey tewree I
CPE,,
ZCPE
"ZARC2
I^
F r
W 7 "PP |
Circuit A . 2
ZARCl
ZARC2
| Bthaon
CPE,-,
-CPE.
Wv
Wr
a / s e
sl/se
1 R was chosen to represent charge transfer resistances associated with the lead dissolution processes,
while R was related to the DC conductivity of the slimes electrolyte. CPE, represents the distributed
nature of the anode/slimes and the slimes/slimes electrolyte interface while CPEa represents the
presence of a distributed capacitance generated by the concentration gradients present in the slimes
electrolyte.
1
2 The high frequency term refers to the part of the impedance that was observed at co> 100 rad/sec.
[171]
Data analysis
z > = QO'co)
Z. =r.
=r
a / s e
sl/se
A(oo
z*,
((,
1 Notice that by neglecting Z , in circuits A and B the only elements that disappear are CPE and CPE
respectively.
2 Once the parameter values were obtained the impedance at co=0 was found to be nearly equal for the
two electrical circuits shown in Fig. 37.
w
[172]
Data analysis
1173]
Data analysis
(A)
,7M
.3*8
2.88
1.8
6.3
22.1
79.1
A
-
(B)
28i.
4 996.
6 3333.
0 12537.
17789.
28868.
rti/itc
.863
.31
1.3S
7.91
39.63
199.
996.
4991.
23813.
28868.
. 8 0 nn
I U M I
1.68
1.28
2.23 nn f l i n t t
SP
EwrlmtUI
AnAlogut no4ll
' ' ' '' ' ' ' ' ' ' ' '' ' '^ ''' '
1
.288
. 428
(C)
3.88
4.88
.863
.31
1.3B
7.91
a4 39.63
199.
996.
0 4991.
23813.
* 2B86B.
9.88
4i 1
S.9S
nn
7,28
s
a
ft
3.88
(D)
rtJ/ste
.S68
a
4
ft
a
3.48
.863
.31
1.38
7.91
39.63
199.
996.
4991.
13783.
23813.
7.79 nn l t n
- Cxytritieittl
I i ii I ii I i iI it I i > Ii ii Ii i i Ii i i I i ii I
3.88
4.88
3.88
-.888
1.88
3.68
3.48
7.28
QCTTl2
Zreaj, Qcm
Z^,
Fig. 38 Correlation between the experimental and theoretical Impedance spectra
(Exp. CA2, circuit A.2, Table 7).
2
1174]
Data analysis
Table 7
RA^I R, + R
Slimes
^zci
Qcm
Thickness
i i c m sec
1
,sec
Qcm
,sec
b,
^ZC2-
i i cm sec
RA,IOUI
Qcm
nun
0.80
2.23
3.10
5.33
0.21
0.42
0.64
1.06
0.73
0.64
0.59
0.60
33.50
14.09
9.39
10.19
0.0010
0.0042
0.0102
0.0225
5.95
1.54
0.51
0.0373
6.56
7.18
7.79
4.70
5.13
4.30
0.56
0.55
0.59
IllBili
iiiillliiii
ii.il
0.3323
0.3593
0.2030
0.44
1.17
1.62
3.17
0.77
0.69
0.69
0.71
iilill! llllll
iiMil
2.08
4.65
0.98
0.96
0.90
5.55
3.84
4.11
4.60
0.037
0.176
0.260
0.592
0.65
1.58
2.26
4.23
4.06
0.795
4.99
0.82
1.583
1.992
2.484
5.99
7.21
8.95
2.05
b) Xg values are small (of the order of msec) while x values are large
i
Ri
(of the order of sec). The separation in the magnitudes of these time constants
decreases as the slimes layer thickens .
c) The largest changes in the values of the derived analogue parameters
appears at slimes thicknesses larger than 6.5 mm. These changes can be
1 R^totai is expected to be larger than R because in addition to include ionic processes it also includes
resistances associated with charge transfer processes.
m
[175]
Data analysis
Ri
[176]
Dataanafysa
(A)
(B)
J.M
IB.B
3.U
CO
4J
8.81
XRI
R A . total
TR2
Ri
Ra
1.31
l.N
t
.318
2.88
3.68
3.48
7.28
9,88
3.68
mm Slimes
3.48
7.28
9.68
mm Slimes
(C)
12.8
P A , total
Pi
P2
o Lit
CJ
R1
4-4
Ato
Rj. [toon*]
4.81
CO
0,
K
'
[Qcm] =
Slimes
Compare R
2.48
Ail0U
' I
.881
,888
II
1.88
I
3.61
3.48
/?,
[Qcm ]
=;
Thickness
; r
[cm]
i7.28
i I i i 9.88
Ii i
mm Slimes
[177]
Data analysis
.7(0
.560
.063
.31
1.S8
7.91
a 39.63
4 199.
6 996.
0 4991.
250IS.
* 28068.
420
*t
,89 nn Stines
280
Experimental
Anilogue noJel
.140
.080'
700
.888
1 This tail was found to be described by the CPE element shown in the analogue circuit B.1.
2 Notice that B is nearly equal to the B, value obtained under rest potential conditions and in the
absence of addition agents [B ~1.9 Q cm sec while B,~2-3 Q. cm sec' (see Table 3, Exps. CC1-5 to
CC2-2)]
a
y 2 C
yzc
[178]
Data analysis
Slimes
Thickness
r
Qcm
a
r
Qcm
^ZCa
Ratotal
Qcm
Q cm sec'^
F cm"
mm
0.80
2.23
3.10
, . , , ^ , , , , , , .
5.95
6.56
7.18
7.79
0.43
1.15
1.64
, , , , , , , , , , . , . . .
i i i i i i i i i !
4.36
4.80
5.29
0.23
0.49
0.66
0.90
0.026
0.102
0.146
0.398
1.13
1.60
2.40
3.82
0.570
0.749
0.975
1.35
11.48
7.46
8.59
- j f
9.88
10.48
11.25
12.61
ZCb
0.58
0.53
0.56
0.60
0.61
0.61
0.62
0.64
0.66
1.64
2.30
4.25
5.00
5.96
7.20
9.12
(A)
(B)
mm Slimes
mm Slimes
Fig. 41 Variation of the derived electrical analogue parameters as a function of
the slimes thickness (Exp. CA2. Circuit B.2, Table 8)
CA) Changes
Ra.totai
p ,totai
B
[179]
Data analysis
a / s e
The two electrical circuits shown in Fig. 37 were also used to analyze
the impedance spectra from exp. CA5. The parameters obtained from the
curve fitting process are shown in Tables 9 and 10. From the analysis of
the data presented in these tables, precipitation of secondary products
appeared to take place between 1.2 and 1.76 mm of slimes. Furthermore,
the large R value at 1.89 mm of slimes and the small time constant
associated with it, appear to indicate that charge transfer processes were
strongly hindered as a result of the continuous precipitation of secondary
products (see Table 10).
x
[180]
Data analysis
Ri.
Thickness
Qcm
Q.
\ ,sec
bi,
^ZCl
cm sec
2
b,
Ri
Qcm
,sec
2 cm sec
RA^DHI
R, +
RA^OUI
Qcm
mm
0.64
1.20
1.76
1.89
0.101
0.017
0.069
0.817
0.91
0.68
0.83
0.58
75.04
1.47
0.35
13.64
0.0007
0.0015
0.1423
0.0079
0.32
0.83
1.61
0.99
8.19
11.10
9.26
6.77
0.68
0.59
0.53
0.81
0.0088
0.0121
0.0373
0.0934
0.42
0.85
1.68
1.81
Table 10 Electrical analogue parameters derived from the fitting of the impedance
data obtained in Exp. CA5 (Circuit B.2)
Slimes
Thickness
mm
Qcm
0.64
1.20
1.76
1.89
0.38
0.72
1.50
Qcm
0.04
0.08
0.16
0.32
cm'
^ZCA
Q cm sec'
0.017
0.036
0.063
0.062
16.10
16.10
11.39
Qcm
Vzc
ifiliilllllf
Re.total
0.65
0.64
0.57
0.62
0.42
0.80
1.60
1.82
The impedance spectra obtained in Exp. CA4 was not analyzed directly
by any of the two proposed electrical analogues because of the complex
characteristics of the observed spectra. On the other hand, enough
information on the precipitation of secondary products was obtained by
solely plotting the maximum
and -Zg values as a function of the slimes
thickness, as obtained from the following relationships:
-Z
=[Z ]
hnax
at a> = a>
<)
[181]
Data analysis
1888.
er
Zraal (nax)
o -Zinag (nax)
' I ' ' I ' ' I ' i I ' ' I ' ' I ' ' I
1.28
1.88
2.48
3.88
mm Dissolved
As can be seen in Fig. 42, both components increase at the same rate
up to @2.1 mm, after which sudden increases in their values are observed.
This point appears to mark the onset of precipitation of secondary products
This precipitation is continuous and produces exponential increases in the
impedance towards the end of the electrolysis cycle. These impedance
increases result from both charge transfer and diffusional restrictions
created by the presence of these products. Secondary products act as a
barrier for ionic movement across the whole layer of slimes and create a
nearly motionless region in which only very small currents can flow. Thus,
this represents the lirrutlng case in which Z^-***. As there is no other path
for current flow (see Fig. 37) other than through Path A \ the dissolution
process has to stop (so called "anode passivation").
1 Exp. CA4 was performed under potentiostatic conditions and, as found by analyzing the cathodic
deposit, impurity dissolution did not take place.
[182]
Data analysis
(ii) Case II: impedance spectra obtained in the absence of a net Faradaic cur
Upon current interruption the impedance spectra were found to be
capable of being described by the modified Randies analogue circuit shown
in Fig. 43. Such a circuit has the following impedance :
1
Z(/'G>)=-
/? +5 (/'u))"
c(
.14
cDL
-0
Glwv-
CPE
1 Notice that the fact that the impedance could be fitted to Randies circuits implies that: Z ^ - * - , and
while
and Z have finite values.
Zjiybe-^oo
[183]
Data analysis
(B)
(A)
6M
177.
443.
1117.
28(7.
Ttn.
17781.
44483.
111741
28MU
* 333334
g
4
t
a
- A m l t f i i no4tl
.48*
.3tB
r*4/stc
177.
396.
4 888.
1987.
a 4448.
4 9939.
a 22293.
a 49914.
78384.
111748,
* EiyrtMnt*1
AIM 10utftfl4tl
.248
.128
'' >
1
I
1.28
lii.ii.
l.M
,2
.888'
R,. i l c m
0.17
0.64
12.03
86.90
0.044
0.112
0.070
0.048
23.75
167.42
206.32
274.37
B L Qcm sec"
2
Vzc
3.83
!!!!!i!i6l?: aiii;:i
:
3.i3
2.23
^zc
0.19
0.33
0.31
0.30
[184]
Data analyse
d)
is nearly constant .
4
a / s e
+2
1 Remember that both the t\ and C , measurements are given with respect to the geometrical area of the
electrodes.
2 Ro, and C | are also a function of ionic concentrations at the anode/slimes electrolyte interface and the
roughness of the anode.
3 B decreases also as a result of the re-dissolution of secondary products and ensuing tortousity and
porosity changes.
4 For ionic diffusion across a porous electrode O.S^^O.25 [13,16,19-21]
d
[185]
Zmu. "cm
Z^j. flcm
Table
Time, Hrs
0.63
11.67
34.59
Ret, Qcm
m
0.023
Cd,, uF cm'
804.67
1588.19
1308.61
B Qcm sec'
2
J'.71
0.70
Vzc
^zc
0.21
0.25
0.25
[186]
Data analysis
c) Bj values are smaller than in Exp. CA2 because the region over which
concentration gradients span is ~3 times smaller.
d)
values are smaller than in Exp. CA2. Thus, due the larger
presence of secondary products there is a larger blockage for ionic flow and
a reduction in the available space for ionic diffusion.
7.Br-xE-2
A
Expertnental
Analogue node I
14.1
31.5
78.5
D 158.
4 353.
5.6
rad/sec
6.3
4.2
791.
1771.
2229.
3533.
ci
2.8
2.8
Z^y, Qcm
3.8
2
Fig. 46 Argand plot of a typical lead bullion electrode (Exp. CC2) in the presence of a 2.2 mm layer of
slimes
AC impedance spectra obtained under rest potential conditions 20 Hrs after current interruption
R, was subtracted from the Zj, component of the impedance
The impedance curve was fitted to a CPE elementfromwhich the following values were obtained Bj=
0.14 cm sec'**, ^ ^ . 4 2 .
Compare with Figs. 5.16 and Fig. 6.1
2
[187]
Data analysis
+2
+2
1 In the absence of slimes, the A C impedance spectra were also fitted to a Randies circuit from which a B,
value of 1.88 Ci cm sec'** was obtained (see Table 3). Thus, as a result of the uniform corrosion of the
electrode, surface irregularities observed at the slimes/bulk electrolyte interface diminish at the
anode/slimes electrolyte interface and this results in equivalent B, values for lead bullion and pure lead
working electrodes (see section II.A).
2
2 Upon current interruption concentration gradients relax and local ionic concentrations decrease. This
results in re-dissolution of the precipitated products. Upon re-dissolution ionic species are released
creating an environment in which bounded diffusional processes take place (i.e. the precipitated products
during their re-dissolution maintain a fixed Pb* concentration at the slimes/slimes electrolyte interface).
2
[188]
Data analysis
(see Fig. 47). These phenomena can be followed by analyzing the variations
in the analogue parameters as a function of the current interruption time
(see Table 13).
(A)
(B)
ExpurtiMitil
' f h u l o m c na<lel
l.M
tl
2.1 Hri
r*4/t*c
.63
a.St
3.96
D 39.63
0 138.
4 628.
6 2381.
0 9939.
25813.
39646.
r4/ic
& .63
2.38
9.96
39.(3
138.
628.
2381.
9939.
, r i.
11111111
9.3 Hri
1
68
1 1
22293.
33333.
1 1 1
1 1
11
I.I
(D)
1.88 r
* ExperineiWI
- Amloiut nodtl
r*4/ttc
.863
.33
1.99
11.17
a
a 63.83
4 333.
ft 1987.
a 11174.
23813.
39647.
ii
I
.(88
r*4/tic
.863
.33
1.99
11.17
0 62.83
4 333.
6 1987.
0 11174.
23813.
39647.
DM*' i i I i i i I
.888
. 488
ii I i iiIiiiI ii iI i i
. 888
1.28
i I i
ii Iii
1.(8
i I
iii I
2.1
Z ^ , iicnr
Z ^ . Qcm,2
Fig. 47 Correlation between the experimental and theoretical impedance spectra
(Exp. CA4, current Interruption conditions. Table 13).
From the AC data presented in Fig. 26. Circuits A, 1 and A.2 (Fig. 37) were used to
fit the experimental data. The derived analogue parameters are shown in Table 17
Each spectrum was obtained at the following current interruption times:
(A) 2.1 Hrs CB) 9.3 Hrs (C) 11.2 Hrs (D) 19.6 Hrs
[189]
Data analysis
Time, Hrs
\ ,sec
bi,
^zci
il cm sec
Qcm
V R
*ZC2-
2.08
9.33
0.87
0.32
0.73
0.79
liPiiii
11.18
19.55
034
0.32
0.73
0.72
0.00057
lllllillll
75.11
0.00052
b,
2
1.77
0.00167
0.43
0.00062 """'""2
l.6
"""""" 0.47
:
0.99
0.99
x ,sec
Ri
O cm sec
Qcm
108.92
R,
0.55
0.59
111111111
1.70
Qcm
0.77
0.56
0.34
0.40
2.64
1 58
1.33
1.31
The A C sweeps obtained at current interruption times longer than 11 hrs were fitted to the
analogue (circuit A.1). Thus, 2 additional parameters were obtained:
ZZARCI-ZZARC2-CPEO
Time, Hrs
b,
^ZCo
Q. cm sec
2
11.18
iiiiiiiiiiiiiii
0.050
0.74
0.046
0.79
[190]
Introduction
C h a p t e r 7 P h y s i c o - C h e m i c a l Properties of HaSiFe-PbSiFe
E l e c t r o l y t e s a n d t h e i r R e l a t i o n s h i p t o t h e Transport Processes
A c r o s s t h e S l i m e s Layer
I. I n t r o d u c t i o n
In the previous chapters, the electrochemistry of lead bullion electrodes and
of the phases and compounds present in the slimes was shown to be linked to
the ionic concentration gradients across the slimes layer. The extent to which
these gradients change was found to be a function of the current density and the
bulk electrolyte composition. From an analysis of the ohmic and diffusional
components of the anodic overpotential, increases in ri were linked to changes
in ionic composition across the slimes layer. ri changes were also related to
precipitation of secondary products and to hydrolysis of the acid. Precipitated
products hinder the ionic movement across the slimes layer, increase TJ , and
thereby enhance the dissolution rate of noble impurities.
A
In this chapter lead and acid concentration gradients within the slimes layer
are related to hydrolysis of the acid and precipitation of secondary products. The
mean ionic compositions present within the slimes layer during the industrial
operation of the BEP are used to indicate the magnitude of such gradients.
Additionally, Eh-pH diagrams are used to describe the conditions under which
secondary reactions across the slimes layer can take place. The shape of the
concentration gradients and their effect on the anodic processes are assessed by
solving numerically the Nemst-Planck flux equations in their fundamental form.
Some of the experimental data required to solve these equations were derived
from an analysis of the physico-chemical properties of H SiF -PbSiF electrolytes
(pH, viscosity, density, and electrical conductivity) and from data available in the
literature.
2
n. Average S l i m e s E l e c t r o l y t e C o m p o s i t i o n
Table 1 shows the range of electrolyte compositions found within the slimes
layer at the end of the electrorefining cycle during normal industrial operation of
the B E P . Pb concentrations within this electrolyte are between 5 and 10 times
x
+2
1 In-situ electrolyte concentrations of the slimes electrolyte are likely to be smaller than the reported
compositions shown in Table 1. During the extraction of electrolyte samples from the separated slimes,
some re-dissolution of precipitates (mainly of PbF ) may have taken place.
2
[191]
Table 1 Composition of the electrolyte samples extracted from anode slimes obtained under
industrial operation of the BEP (Cominco Ltd.)
Lead electrorefining test conditions: Refinery cells, cd = 220 ATTI , 5 days, T=35-40*C,
2
Test
Anode Composition
Anode
Number
wt%
Face
Assay, wt%
M (moll )
1
m M
(mmol l )
Efficiency
i l l i l l i l l l Illl
Bi
Ag
111 I l l l
Pb
Pb
1.86
1.98
HjSiF,
Sb
Bi
As
0.17
0.14
0.28
0.17
#2 Anode 6
HI
042
o.n
III
031 0.14
ill HI
0.36 0.16
1.28
1.83
%
84
1.52
1.52
0.28
0.28
77
79
103-1
0.14
0.10
96
103-2
76
104-1
78
104-2
0.17
0.17
94
In the BEP the only cause for lost current efficiency Is cathode-to-anode short circuits and current lost to ground.
Without shorts, a lead bullion electrorefining cell gives 100% current efficiency 152).
1 Thiet [1] found the following mean ionic concentrations within the slimes layer:
at 200 A/m
at 250 A/m
at 300 A/m
2
2
2
[192]
As shown in Table 1, the concentration of noble impurities (As, Sb, and Bi)
in the entrapped electrolyte is very small. The indicated Bi and As concentrations
are of the same order of magnitude as found from insitu measurements (see
Section 4.2.B Chapter 4) while Sb concentrations are larger \ Within the slimes
layer the concentration of noble impurities is small because the noble phases do
not react to a significant extent and cementation reactions maintain their
concentration at small levels (see Eq. 1.1). Furthermore, the presence of
equilibrium quantities of these impurities in the recirculating electrolyte (see
Table 1.1, Electrolyte composition row) indicates that these impurities can be
tolerated up to a certain level without serious impairment of the cathode quality.
This was also indicated by Thiet Ul and others 16-91, who showed the relationship
between the concentration of noble species in the bulk electrolyte and the cathode
composition.
The mean Pb concentration across the slimes layer for all the data presented
in Table 1 is ~2.15 M while the mean acid concentration is ~0.20 M . Assuiriing
linear concentration gradients across the slimes layer , the maximum amount
of Pb at the anode/slimes electrolyte interface can be considered to be close to
4 M while the H SiF concentrations at such an interface can be considered to be
smaller than 0.10 M . Even at this large Pb concentration, PbSiF .4H 0 is not
expected to crystallize because of the low acid concentrations . Furthermore, the
presence of small amounts of acid at such an interface aids in the stabilization
of PbSiF by suppressing its hydrolysis.
+2
+2
+2
...I
...2
'
1 Larger Sb concentrations shown in Table 1 as compared to those shown in section 4.2.B can be
attributed to partial re-solution during the filtering process of antimony oxides present in the slimes layer.
[193]
Eh-pH Diagrams
...3
Both PbF and Si0 have been identified and quantified in the slimes layer
(e.g. see Table 1.2) [5.11,12]. As early as 1908 Betts [ i l l already indicated the
presence of Si0 in the slimes layer and related it to H S1F decomposition. He
found that -2% of Si0 in the slimes could correspond to ~1 kg of H S1F loss per
ton of lead produced. In 1966 Cominco Ltd. reported Si0 assays in well washed
slimes that were of the order of 1.6% [131. Thus, according to Eq. 3 and assuming
a production of slimes of ~25 kg per ton of lead, an estimated loss of H S1F of
-0.96 kg/ton of Pb can be obtained. This can account for as much as 50% of the
acid losses. Furthermore, PbF in the slimes can be as high as 19.6% and can
account for most of the lead found in the slimes layer. In Table 1.1 only one refinery
(Cerro del Pasco) reports SiO assays in the slimes layer. From their reported value
(0.4% SiOJ an estimated loss of H SiF of ~0.13 kg/ton of Pb was obtained .
2
HI. E h - p H Diagrams
The Eh-pH diagrams shown in this section were generated using the c.s.i.RO.
THERMODATA computer program [14,151 . A list of the INPUT (*.inp)filesused In
the generation of these diagrams is provided in Appendix 11. Also, a list of the
8
1 Hydrolysis of the SiF " ion takes place in 3 steps which involve intermediate splitting of F" and formation
of SiFsIHaO]" (see section lll.a).
2 At 25 "C the maximum solubility of monosilicic acid H Si0 is 10" M while the solubility product for PbF
is 7.8x10" [10].
6
3 Notice that in Table 1.2 f 7-).,,% is between 4.1 and 5 while in PbF (7-L* is 5.45. Some fluorine may
be present in compounds other than PbF but most of it can be accounted as PbF .
4 As shown in section lll.b washing of the slimes can lead to further hydrolysis of the acid and removal or
formation of precipitates.
2
5 This number was obtained assuming that 1.4 kg of slimes are produced per ton of lead.
6 c.s.i.R.o. means Commonwealth Scientific and Industrial Research Organization (Mineral Engineering
Division, Port Melbourne, Australia).
[194].
(F)^hH,0 system
a >3
5tf " + 3H 0
6
.. .4
SiF~ +2H 0
2
SiF [H OT
5
+H 0
SiF
+ F~ +2H 0
...6
H^iOs
<=>
...5
+ F~
SiF
<=>
+ 3H 0
SiF [H OT
<=>
+ AHF
...7
In the acidic range (pH<l) formation of SiF [H 0]' and SiF [H 0] takes place
5
[24]:
Si F; + H 0 + H <=> SiF [H 0 ]~+HF
2
... 8
.. .9
1 The diagrams were drawn assuming unit activities for all the solid species species and T=25 C.
2 Either H Si0 or Si(OH) can be formed during the hydrolysis of the acid [16,17]. The polymerization of
orthosilicic acid, Si(OH) , has been shown to be very slow [18].
3 The dashed vertical lines in Fig. 1 extend the region in which HF, HF \ and F may be present.
2
[195]
(Pb-Fj-Si-HjO system
F i g . 1 S y s t e m (FJ-Si-HaO at 2 5 C
A
SiF (H 0) + HF
B
C
SiF (H 0)- + HF
HF + SilL, (g)
D
E
F
G
H
HF - + SiH,(g)
F j O ^ + HjSiO,
SiF "
H Si0 + F
F + SiH*(g)
-U
-3 " -2
PH
25/F + Si0 + 4H 0 3SiF [H 0\
6
-2
... 10
Si
<
[H-SiFJ + iSiO-]
B. (Pb-F)-Si-H 0 system
2
+2
.. .3
A 3
~0
...IZ
siF?
Pb
+2
[196]
(Pb-F)-Si-Hfi system
Assuming that
0^0 = 3 / . ^ = % ^
precipitation can take place are given by the following equation (see also Fig. 3):
pH = 1.78 - 0.25 l o g [ ^
and a
pb+2
... 13
F i g . 2 S y s t e m ( P b - F ) - S i - H 0 at 2 5 C
2
+2
+2
+2
Once the activities of the individual species are known, Eq. 13 can be usee
to obtain an upper limit for the ionic concentrations across the slimes layer to
avoid obstruction of the ionic flow due to Si0 and PbF precipitation
2
1 The extra weight of these precipitates can also lead to falling slimes.
[197]
(Pb-Fj-Si-HjO system
1.2
1.4
1.6
1.8
2.2
PH
Further analysis of Eq. 3 indicates the difficulties encountered during the
washing of the slimes to obtain accurate PbF and Si0 assays. Depending on the
pH and ionic strength of the washing solution, PbF and Si0 may either precipitate
and/or re-dissolve indicating erroneous assays. Only by slowly displacing the
concentrated electrolyte with solutions of varying PbSiF -H SiF composition and
matching pH but different strength, can the extent to which these secondary
reactions take place be diminished . This process involves a trial and error
procedure using "slices" of slimes taken at different slimes thickness. In a first
try the composition of the slimes electrolyte is measured and solutions with similar
pH but diminishing ionic strength are prepared. The slimes are put in contact
with these solutions until all the SiF " and Pb are extracted. Once these ions
are removed, the slimes can be washed with dilute HN0 solutions at the same
pH as before.
2
+2
1 Notice that for the washing process to be successful, activities of Pb* , H*, and SiF " are required to
remain in a fixed ratio given by the following equation (see Eq. 12):
2
[198]
As shown In Fig. 4, Sb, As, and Bi are noble with respect to lead and their
dissolution will not take place unless the potential and/or the pH rises permit
the formation of oxides and/or soluble species \ Sb is the least noble of these
three elements and is expected to react earlier than the others. Upon dissolution,
Sb forms soluble oxides some of which have been detected in lead slimes (e.g. see
Tables 1.2 and 4.2) [301. Arsenic also forms highly soluble oxides such as the
amphoretic arsenious anhydride, A s 0 1311. When As is exposed to moist air it
becomes covered with oxidation products [31]. This poses a problem in its
identification in the slimes layer. Bi does not form oxides to the same extent as
Sb and As, but rather, forms polynuclear complexes such as ByOH)^* [301. These
ions have been found to be the predoiriinant species at pH=1.5 and total Bi
concentrations of 0.01 M [301.
3
Sb + SiFV
Sb + SiF5(H 0)- + HF
SbO + SiF SbtT + S i F ^ j O Y + HF
Sb 0 + SiF "
Sb 0 + SiF - + HF
S b A + S i F ^ O V + HF
Sb O + SiF, + HF
Sb + HF
2
Activity of Sb soluble
species = 10"*
1 The fat dashed lines in Figs. 4A, 4B, and 4C mark the stability region of Pb* according to Fig. 2.
2
[199]
F i g . 4 B S y s t e m (Bi-FJ-Si-HaO at 2 5 C
SuV
Bi +
B i + SiF5(H 0)+HF
BiO + SiF - + HF
E
F
Bi + SiF (H 0y + HF
5
+3
Bi + HF
Activity of Bi soluble
species = IO
-4
4.0
F i g . 4 C S y s t e m ( A s - F ) - S i - H 0 at 2 5 C
2
1.1
Sh
A
B
D
E
F
As + SiF "
As + SiF (H 0)- + HF
s
AS40 + S i F
6
AsO + S i F
+
l
4
2
6
+ HF
AsO + SiF (H 0y + HF
+
Activity of As soluble
-0.25
species = 10"
3.5
U.0
[200]
IV. P h y s i c o - C h e m i c a l Properties of H S i F - P b S i F E l e c t r o l y t e s
2
To model the concentration changes across the slimes layer, data on the
physico-chemical properties of H SiF -PbSiF electrolytes are required. In the next
section, the experimental data generated in this work are analyzed alongside data
available in the literature.
2
solutions
3
Fig. 5 Changes in pH as a function
of the electrolyte composition
CL 2
1 1
T=25"C
Initial Electrolyte Composition:
[H SiF ]=2.040 M, [SiOJ=0.40 M
1
-
0
1C"
:
5
10"
10"
10~
10~
10
[H SiF ], mol/l
2
H SiF
2
HSIF;
<=> HSiF;
<=> SiF;
+
+
...14
H
H
pK =\.19
2
...15
[201]
Arkhlpova et al. [27,28] have indicated that the pH of H SiF solutions (among
other properties) is a function of j. and varies according to the way the acid was
produced.
Leonte et al. [351 measured the density, viscosity, specific heat, and electrical
conductivity of solutions containing between 0.3 and 2 M H SiF . They showed
that due to the technique under which their acid was produced, minor impurities
such as P 0 and S0 ~ were present in concentrations that increased with the
acid composition. On the other hand, no j values were reported in their samples.
Leonte et al. derived empirical equations for the density and viscosity changes
as a function of the acid composition :
2
p = 0.9982 + (8.53 10" -1.96 10"r)w - (2.99 IO" + 8.06 lO^w) (t - 20)
3
... 16
x\=Ae
...17
A and Q = empirical coefficients which depend on the solution composition (see Table 2).
T = temperature K
t = temperature "C
Table 2 Values of the coefficients A and Q in Equation 17 [35]
[H SiFJ, M
0.00
1.219X10"
1965.93
0.334
0.663
1.908x10*
2.027xlO'
1857.05
1860.40
0.972
1.298
2,434xlO
2.937X10'
2.485X10-
1834.36
1803.40
1.708
1884.70
Folov et al. [361 measured the osmotic and activity coefficients of H SiF
solutions at 25C using isopiestic measurements and H S 0 as standard. As can
be seen in Table 3, y do not change significantly with ionic strength.
2
1 Sohnel et al. [37] also present a p-[H SiF ] correlation from which the apparent partial molar volume of
HgSiFs at infinite dilution, , was obtained: <J>v =23.5 cm mol" at 17.5"C.
2
[202]
Table 3 Changes in the osmotic and activity coefficients as a function of the molal ionic strength, [36]
molality, m
Mean Activity
Coefficient, y
0.10
0.20
0.30
0.60
0.931
0.611
0.577
0.30
0.40
0.90
0.927
0.934
0.562
0.555
0.50
0.60
1.50
1.80
0.944
0.954
0.553
0.553
0.70
0.80
2.10
0.964
0.975
0.559
0.559
0.90
1.00
2.70
3.00
0.987
0.999
0.564
0.571
1.20
1.40
3.60
4.20
1.026
1.060
0.588
0.612
conductivity
of H SiF -PbSiF
B. Density, viscosity,
electrolytes
and electrical
Two different values of j. are indicated in Table 4, and each one of them can
be related to the presence or absence of Si0 or HF in H SlF -PbSlF electrolytes:
2
1 H SiF -PbSiF electrolytes were prepared by neutralizing H SiF solutions with PbO.
2 The dynamic viscosity, TJ, is the ratio between the applied shear stress and rate of shear.The cgs unit of
dynamic viscosity is one gram per centimeter per second and its called one poise (symbol P). One
centipoise, 1 cP = 10' P.
3 The kinematic viscosity, v, i s a measure of the resistive flow of a fluid under gravity, the pressure head
being proportional to the density of the fluid [44]. The cgs unit of kinematic viscosity is one centimeter
squared per second and it is called one Stoke (symbol St). 1 centistoke, cSt = 10" St.
2
[203]
Electrolyte Composition, M
ratio
[SiOJ
[HF]
0.4
s1
2040
(a (51
5.0
5.0
T=40'C
T-40"C
To25.5 "C
525.1
T-40'C T 23 C
-
443.7
1.229
From mother solutions prepared using acid sample S1 (Technical HaSiFg supplied by Cominco Ltd.):
v17
v18
v19
v20
V21
k6
v13
vii
v12
zl4
V14
K2
kt
v4
0
0
0
0
0.27
032
0.32
0.34
0.35
0.36
0.36
iioiii
0.837
0934
0.795
lloMil
IlKll
0.36
v10
v9
05
.9
0.64
v16
v15
l l l0.81
iiil
v3
v8
v7
v2
0.215
0.332
ilblMOlI
0.847
1.017
0.692
096
0.98
Itilili
1.21
0.830
0792
0.067
1.010
0.849
0 363
0.165
0.083
0.716
0.331
0.097
0.156
0.446
v6
ALIO
AL1
k4
v5
vi
sol 42
pbaOl
vO
sol 6
1.61
1.62
1.75
1.75
1.8
0.306
0.092
0.069
0.119
0.124
0.151
w71
3.064
1.21
1.45
1.46
1.53
0.04
0.05
0.06
0.13
0.20
0.09
0.09
0.13
011
0.07
0.04
0.09
l i i b l l i lliioli
5.0
5.2
5.2
o
l l l b l l l : 111211l l l s i l l l
"""""" 'S'"""""""" 5.2 ,,,,, ,^,,,,,.
108.7
160.5
183.1
303.9
g
0
5.0
341.9
5.5
5.3""
245.3
b""
1990
l
l
l
b
l
l
l
l
i
&
4
l
l
l
297.9
5.4 """
3312
i i l b l l l l l i M i i llliilli l $ p l i
5.7
5.5
282.4
331.8
o
lllblll
iiislii
i
l
l
e
l
l
l
l 1391
5.6
5.4
345.6
290.9
o
lilolli I i l b l l l l l i i6.0i i ;
llililf 295.4
6.0
63.9
80.2
b""'
0.15
363.5
iiisiiii
11P
1
1
5.5
""""'5:2 " "
2844
345.1
0.14
005
1971
lilolil iislli l
ilppl
I
l
S
l
l
l
0.02
5.9
5.4
173.8
143.8
o
0
123.4
lilolli
lllbli
loll
1
1
1
5
0
1
1
1
0.12
5.6
5.2
300.7
251.2
o
231.1
0.06
l
l
l
b
l
l
l
lll&ill
6.0
6.0
177.2
144.1
o
o
0.00
157.2
o
libllll
0.09
5.7
5.0
249.3
205.8
lliiii
llliill
128.1
190.2
215.9
359.6
400.0
293.7'
237.4
354.4
llliill
llf&Ill
lliwsill
lltllil l i i f l l
iiiiii lllSiii
I
l
b
5
i
4
i
l 010
0.317
o
0
liiblli
o
0
ll5ltl:l
6.0
6.0
6.0
0.40
6.1
0677
0.699
0.714
0.753
0.796
1029
1.036
1056
1.091
0.843
0.962
1.016
0.961
0773
0.726
0.791
0.827
1.161
1.216
1.228
1.217
1.085
1.215
0.778
1.137
1.084
0.989
0.935
0977
1.170
1.109
1.090
1.115
0.681
0.79
0 712
08
.61
0.831
0757
0.73
0 737
0.831
0807
0.771
0 785
p l l s p l l 11^:711i i p p l
6.0
l l l6.0
iill
llliiii
l i i i l l ! l l l i0i l l ll&ii
6.0
6.0
l l i b l i f l l l i i l l ll&ii l i i i l l
6.0
6.0
ij,,,,,.
0
0.697
0.724
0754
0.822
0.893
256.2
:|l236l5ll
201.3
liwiil
209.6
213.5
196.1
163.9
170.6
171.6
llliill
lliilililllliil
211.5
173.0
1.300
1.387
2.74
2.88
ll^ll
3.25
0
0
0
0
0.63
0.66
0
llolsfllli
.................g, ..............
0.15
iiifci!
6.2
ll&all
6.0
1142
1320
1.304
1.306
1.281
1.325
1.408
1.374
1.414
1421
1.209
1 546
0.887
0.846
1570
1.537
llliill
1582
1.612
6.1
1.397
From mother solutions prepared using acid sample VV71 (Technical Reactive BDH Chemicals):
w65
0.22
iilbllll
132.0
108.5
0.923
0.714
l
l
l
l
b
l
l
i
l
l
l
M
i
i
w62
0.07
6.0
209.9
174.9
1.153
0.785
1.46
o
b
w61
0
1.500
0.895
I i l b l l l l lioloalll WMtim
l
i
l
b
l
i
e
l
i
i
i
M
i
i
6.0
0.12
w60
2.58
0
247.7
203.0
1.958
1.066
o
w64
w50
w63
w51
1122
llaSSIIl
241.8
170.0
196.7
1:4:207.91:::!!!l67t6lll
203.5
161.3
1.979
2319
2.520
1.064
1.199
II'PSSII
1.26
1 293
1.469
1.676
1.837
1860
1.934
1.942
2.000
v p
[204]
electrolytes
(-
...18
[HJiFJ + lSiO-]
As can be seen in Table 4, for pure H S1F solutions, the amount of dissolved
SiO increases with acid strength without significant changes in
(samples
VI7 to V21). This relationship holds even in the presence of PbSlF . The amount
of dissolved Si0 appears to decrease with the PbSlF concentration but by
comparing the j. values it can be seen that (ji)b remains close to 5 even in the
presence of large amounts of PbSiF (e.g. compare
and
in samples V13,
V7, and AL1). This indicates that soluble silica is present as H S1F reaction
products such as those shown in Eq. 10. Similar Si0 solubilities for H SiF -PbSiF
electrolytes are reported in the literature [io]. In the absence of H S1F , minor
amounts of HF are required to avoid PbSiF hydrolysis (Samples W 5 0 to W65).
As seen in Table 4 only when HF is present can j. reach values larger than 6.
2
2 l3
1 The presence of Si0 appears to increase p and v. On the other hand, HF appears to have a larger
effect on K than on p or v.
2
2 From these correlations K, p, and v, values that are within 6% of the experimental measurements can
be obtained.
3 A simple relationship between K, [PbSiF ] and [HjSiFJ, for the range of compositions shown in Fig. 6A
could not be found because of the changes in K at [H SiF ]>0.4 M. Eq. 19 although cumbersome
describes accurately the electrical conductivity changes as a function of [PbSiFj and [H SiF J.
6
[205]
-T-AKJX^+^11^3+A x
12
+ A x + A XJXJ+A x .x
13
14
15
...19
p^A^A^+Aj^
...20
v^A^AjV.+Ajvf+A^
...21
Where:
K = electrical conductivity at 40 "C, [mmhos cm' ]
v - Kinematic viscosity ( coefficient of) at 40 C. |cSt)
Pb = density at 23 'C. [g cm" ]
x X2, x , y y , y : variables related to the H SiF and PbSlF concentrations as follows:
x, = [PbSiFg]
Xj = [PbSiF ]+ [HaSiFd
x = 2 x [H SiF ]
y, = [PbSiF ]
y = [H SiF6l
y = [PbSiF ]x[H SiF ]
Aj to A are constants whose value is given in Table 5.
1
lf
lf
15
Table 5 Coefficients in the empirical electrical conductivity, viscosity and density correlations
Coefficient
Electrical
conductivity, K, (Eq. 19)
-1.212E+08
5.666E+01
2.244E+07
1.473E+07
2.981 E+07
1.212E+08
-3.090E+02
-3.717E+07
-2.713E+01
1.273E+01
-6.062E+07
1.677E+02
9.293E+06
3.563E+00
1.129E+01
A,
liifiitiiiliiiiiiiii
A,
A*
As
Ae
A
Ae
Ag
A
A,,
7
10
A12
A
Au
A
13
15
Kinematic Viscosity,
v., (Eq.21)
Density, p.
(Eq. 20)
9.951 E-01
3.223E-01
1.668E-01
6.423E-01
4.340E-02
4.320E-02
1.606E-01
[206]
T= 40C
0.2
0.4
0.6
0.8
1.2
1.4
1.6
1.B
[PbSiF ], mol/l
6
[207]
= xA
+ (1 -x)A ^
PbSiFt
.. .22
iFt
where:
Kg*
^mixt -2{[PMiF
n r r D i , c , T]i +, [/W
ri/c.Tii
]}
6
J W .
Aw
W
2[PbSiF]
A _. .
lt
WP
_ KeW/,
2-[/7 S/F ] ,
2
6 /
x--
Thus, for example, for an electrolyte mixture containing 0.36 M PbSiF and
0.252 M H SiF at 40 C:
6
I = 2.196
t
j = 1.44
{PbSiF
x = 0.66
[PbSlF ] = 0.549
W ^
[H SlF ] = 0.732
% =334.9 A
It
It
A^BO^xSO
A =80.0
W /
miF
WjOT
=228.8
1 ^ = 2x0.612x131.2=160.6
nX, = constant
...23
1 This iwvalue compares well with the experimental K value shown in Table 4 (K^= 160.6 vs K=167.1
mmho crrr for solution V14).
2 Walden's rule is based on a model for ionic mobility that obeys Stokes law in the laminar flow region
[51].
[208]
H.=-
...24
The validity of Walden*s rule has been confirmed in a variety of systems [46].
Recently, Peters et al. 1471 have shown that in strong chloride mixtiires, Walden's
rule is not followed by H because its movement in electrolytic solutions is not a
function of the viscosity of the medium but of the activity of water. Similarly, in
H SiF -PbSiF electrolytes Walden's rule is expected to be followed by Pb and
SiF ~ while T)X + values should decrease with increasing ionic strength as the water
activity decreases.
+
+2
1) H SiF and PbSiF are completely dissociated and the formation of ion-pairs
can be neglected.
2) Only three ions contribute to the total electrical conductivity: Pb , SiF " ,
and FT.
3) The electrical conductivity in concentrated H SiF -PbSiF electrolytes can
be represented by the following relationship:
2
+2
K=iic .z x -l
1
25
equivalent conductivity of Pb
+2
(i.e. X - = X + ) \
Sif
pb
5) Xi values in pure H SiF and PbSiF solutions are the same as those of
H SiF -PbSiF iriixtures of equivalent ionic strength:
2
= 2C
%lSP 6l/
,
thus,
= C -X
H+
pb+2
H+
pb+2
+ 2-C
+ 2 C - X _ =--
SiF;2
SiF
-X SiF
SiF
..26
4[PbSiF ] .X
= 2 [HiSiF^-X^
lt
pb+2
l-iHJiF^-X^ ..27
1 Individual equivalent conductivities at infinite dilution and at 25 C reported in the literature are:
X _ = 59[53] and X , =65 [46]
OT
pt 2
[209]
pb ' Kb
+1
=2
=2
'
pb+2
'
+ 2 [HJiFJ
H '
+
...28.a
V+
...28.6
{ t * W J + [PbSiF ]}
V+ (2 [Z/^/FJ + 4 [ / W J }
...28.c
pb+2
Assumptions 3, 4, and 5 are linked through Eqs. 26 to 28 . They are valid only
to the extent to which X, values from binary solutions (i.e. from Eqs. 26 and 27)
can be used to obtain the conductivity of ternary mixtures (i.e. from Eq. 28). Such
correspondence has been found to occur in chloride electrolytes 148] for which
Kr
Thus, to obtain X, values and their dependance with concentration, the
2) Obtain the composition of pure H SiF and PbSiF solutions at the total
2
ionic strength and the specific conductivities of these solutions (From Eq. 19):
[PbSiFJn = \
with K =
[H SiF ] = \
2
3) Obtain
with K = % s -
It
and
pb+2
X+
H
tf
6]/i
KiPbSiFfo
r
+1
4[PbSiF ]
6
4) Introduce
"
It
pb+2
'
-H-$ 6-u
iF
Kb*
2[H SiF ]
2
M H SiF at 40 C ( r ^ = 0.769):
2
[210]
I = 2.196
IPbSWJ
= 1.44
[PbSIF ) = 0.549
=87.8
6 It
[H SiF6] = 0.732
2
lt
= 40.0
K, =334.9
= 188.8
W/
+2
H+
1 This twvalue compares well with the experimental K value shown in Table 4 (K = 172.9 vs K=167.1
mmho crrr for solution V14). The difference between these values is partially due to innaccuracies in the
computation of K values using Eq. 19.
nM
2 Water activities for this system are not available. Yet, by analysis of similar phenomena taking place in
other systems [47] this analogy can be drawn.
[211]
Table 6
Sir
Solution
Solution Properties
Composition, u From Eqs. 20,21, and
Strang*
22
[PbSiFJ [H,SiFJ
200
0.00
1.88
1.52
1.17
0.85
0.06
^rrrtxt Pmtxt
Iii!
208
0.43
249
mm
0.35
0.61
0.87
0.10
0.10
25
0.49
0 76
0.88
0.98
0.25
000
0.00
000
0.00
0.00
0.00
0.00
0.00
0.10
0.40
0.90
1.00
1.25
1.50
0.49
350
iisi!
248
liiiii lisnll
0.86
0.98
337
1.640 i i b i i i
1.610 7.72
1.517 i i i i i
1.429 6.00
1345
1.218 4.00
0.663
0.695
0.753
0.828
0.855
1.040
1106
1.215
1.346
1.390
1.434
0.653
0.720
0.786
0.820
0.664
0.908
0.941
Iflll Illl!
000
i i66i 0.647
0.00
0.667
143
169
189
It
0.902
0.887
0.849
0.823
0.808
0.802
liill
Binary
Properties Properties
Individual
Equivalent
(PbSlF]=J [HS1F1=J
Conductivities
[PbS.F.1
liool?
ilfii
1.50
[HjSiF,]
mi llell
208
1.93
2.57
624
Walden's
Product
WMm V
Eq.
izlil
I269I1
27.0
954
94.1
Ural
143
174
110.4
0.70
0.18
34
0.23
1.72 1614311
liii
0.86 ' 126 ' 1.14
427
3.43
5.47 WMm
ll&f
6.16
192 l 0 5 ' 532
1.54
48.3
41.7
36.8
32.6
31.1
238.4
205.1
150.2
106.4
98.6
l i i i
1.056 1.45
0.36
62
0.48
249
1.109 116911 11616711 l i 6 i l 16:9611
1.135 3.31
0.83
123
1.10
421
1 170 11111 l-iiil
1.38 i i i i i
1.205 4.97
1.24
167
1.66
491
1.231 5.59
1.40
181
1.86
511
1.249
43.0
38.6
37.1
35.3
33.7
32.4
214.6
1.00 '
1.33
IIMI
456
liii WMW
iilii
0.40
1.60
3.60
ImM
0.716
0.729 1.317
0.764 1.398
0.805 1.479
38.7 .
38.6
IIMI :l37.9lf
262
37.0
141.1
134.5
1213
117.6
120.0
131.9
WmW
346
34.8
,,,,,,,,
Ileal:
255
33.3
321
33.7
36.4
37.0
0.50
0.47
0.00
llioii
0.20
0.33
0.33
0.27
0.16
0.47
0.68
164.4
0.19
137.4
118.7
117.1
0.21
0.62
0.62
0.58
0.52
0.51
0.52
163.1
0.17
0.18
0.19
0.05
Islil Iiiii
0.24
www
lire!
338
0.24
0.24
IIPIlltll IH!iiiii iiiii IS! Illl! wWliiPii 1
11111
0.21
005 Ilil l i i i WwW
245 8
56.0
367 161.2 019
Wmw
1.027
1.124
135.0
Transport
numbers
29.C
218
I9I2II
IliSl l;564:l WWii
WMW
189
2.00
526
31.5 100.0
332
35.6
liii Iflff.
007 l l a l l
0.48
248
0.90
1.10
421
1.38 l: 46i!l
1.66
491
1.86
511
0.00
0.00
0.00
0.00
lllfifl
4.00
5.00
6.00
249
32.7
1378:1 33.6
153.6
421
34.5
131 8
35.7
114.6
491
36.9
511
104.9
37.5
Isolli
142.8
133.2
125.4
121.6
Illlll www
Mil Illlll lUti
11!
?
!
lito-l Illlll l i i f l wWii
iiiii Wmw
0.40
illol&il
1.00
1.25
1.50
68
131
142
168
189
0.53
120
1.33
1.67
2.00
268
llall
455
492
526
120.4
52.1 :|24|il
68
42.3 209.3
36.4
1455 l l a l l
35.6 135.1
142
168
33.6 114.0
189
31.5 100.0
162.8
156.8
31.7
335
34.2
35.9
37.4
laMil
129.8
121.7
119.0
081
0.83
0.82
0.81
0.79
iiiii:
0.23
0.24
0.77
Illl!
Iiiii
0.50
0.50
0.50
0.50
0.50
0.76
0.76
000
0.00
0.00
0.00
0.00
0.00
H+
interface most of the current should be carried by Pb and SiF while at the
slimes/bulk electrolyte interface most of the current should be carried by H . This
should create significant concentration gradients as Pb and SiF " are ~5 times
less mobile than H .
+2
+2
[212]
O
01
(A)
300
01
H'
A
250 .
(B)
o
CD
p" 200 -
y ' 200
150
150
cj
-
p' 250
Zf
V
01
300
CM
50 "
3
0.5
77^2
1
1.5
2.5
50
<
r.
(3
0.5
1.5
2.5
It"
Ii*
Fig. 7 Changes in Walden's product as a function of the square root of the ionic strength
(A) Changes in Walden's product for H SiF -PbSiF mixtures of different compositions
(B) Changes in Walden's product for binary H SiF and PbSiF solutions
2
D =
&RT _\RT
.30
Thus, for a solution conteining 0.35 M PbSiF and 0.84 M H SiF , at 40 "C,
^ = 35.8 and ^ = 137.1, from which: D = 3.8 x 10" and ^ = 5.0x10^. This
diffusion coefficient for Pb
is nearly equal to that obtained from
chronopotentiometric measurements (seeTable 1 Chapter 5). The correspondence
between these values and the validity of Walden's rule for species other than H
support the assumptions and the model used to derive Individual equivalent
conductivities from electrical conductivity measurements.
6
H+
+2
1 As shown in Appendices 1 and 2 this equation can be modified to include the effects of the effective
charge of species i exposed to the electric field, Z*. When such modification is included: D, = D/".
[213]
V. M a t h e m a t i c a l M o d e l : N u m e r i c a l S o l u t i o n of the N e r n s t - P l a n c k
Flux Equations.
To obtain local ionic concentrations and potential gradients across the slimes
layer as a function of the electrolysis conditions the Nernst-Planck flux equations
were solved using a finite interval numerical technique (see Appendix 1). The
differential equations that were solved simultaneously are as follows [50]:
1
d<D
dlna
CD c
a
/(r)
-+U.c =
dx
"dx
d\na
dlna
...rn
ZF
rfq>
dQ>
...[/v]
Z C +Z C +Z C =0
a
Eq. i states that the only species being generated at the anode is Pb whereas
Eqs. if and iii indicate that H and SiF " do not react and their equilibrium
concentrations are only a function of the migration and diffusion gradients - .
Eq. iv is the electroneutrality condition.
+2
1 =D?+D : D is the eddy diffusion constant which can be considered to be equal for all the ions.
Furthermore, it can be modified so as to incorporate changes in porosity and tortuosity of the ionic flow
across the slimes layer for different anode compositions and/or electrolysis conditions.
2 Notice that when the molecular diffusion coefficient is smaller than the eddy diffusion constant, the
overall diffusion coefficient is nearly the same for all the components. In this case, eddy diffusion
practically equalizes the effective diffusivities for all the components [so,p. 238]
E
where Z* is the effective charge of species i exposed to the electric field, z* <Z,.
[214]
obtained from \ values (see Eqs. 25 to 27) \ From the computed ^ values D"
values were obtained using the Nernst-Einstein relationship (Eq. 30). To these
values, D were added to obtain corresponding 2} values . Individual ionic
coefficients were assumed to remain constant
in the whole range of
concentrations. Changes in y,as a function of the ionic strength require knowledge
of activity changes (and water activities) in H SiF -PbSiF iriixtures unavailable
in the literature.
The solution of the Nernst-Planck flux equations was obtained by increasing
the number of variables analyzed. This was done by first neglecting eddy diffusion,
then incorporating it in a re-run. Potential and ionic concentration terms are
represented by the following symbols:
3
+n
A . C a s e A ; constant y v a l u e s
(
1 By using the X, values obtained at 40 C, concentration and potential profiles are also obtained at that
temperature.
2D changes also must be known in advance. D is not only a function of the electrolyte composition
but of the slimes thickness as well. In the initial computation of the concentration profiles this number can
be assumed to be zero. This value can be changed once local concentrations are known (i.e. as a result
of changes in v and p as a function of the slimes thickness).
E
3 As shown in Table 3 y does not change significantly with ionic strength and therefore this assumption
is not unreasonable.
[215]
(A)
s. 6a
EO
a
E
33
o
cu
c
o
13
(B)
43 . 0
33 . 0
5
fi
o
o
cu
c
o
23.07 . 3 6
a
o
<u
o
13.0-
3 . 0 0 i
0 0 0
i
1.2
+2
+2
[216]
Fig. 8 shows the concentration and migration potential changes across the
slimes layer for two different current densities. Both [Pb ] and [SiF ] increase
linearly as the slimes layer thickens. These increases are accompanied by a
quasi-exponential decay in [H ]. As a direct consequence of the presence of
concentration differences across the slimes layer, a migration potential becomes
established. Variations in d> are given with respect to the bulk electrolyte (<I> = 0
at the slimes/bulk electrolyte interface). The larger the slope of d> with respect to
the distance from the bulk electrolyte, the larger the electric field.
1
+2
By comparing Figs. 8A and 8B, it can be seen that: (a) increases in current
density result in corresponding changes in {Pb } , {SiF " } , {H }, and <X>; (b) the
concentration values found at normal slimes thickness are impossibly high.
+2
diffusion
1 The electric field, X, at any point within the slimes layer is equal to the negative gradient of the
electrostatic potential at that point:
v-
~~dx
Thus, the electrical field increases in the opposite direction to the electrostatic potential. Likewise, positive
charges move in the direction of the electric field [51, p.348].
[217]
Cass B: constant
153. ,_
8.8
1 ... I
2
4
i .
D i s t a n c e from t h e
18
Anode/slimes
'
I
12
J!
Interface,
14
J_J
16
mn
Fig. 9 Changes in the potential of the slimes electrolyte as a function of the distance from the anode/slimes
interface.
From the experimental data presented in Chapter 4, Fig. 4.12.
As the slimes layer thickens the potential for a fixed point away from the anode/slimes interface decreases.
concentrations are continuously taking place. These changes are not only
associated with natural convection but also with precipitation of secondary
products. This changes the physico-chemical properties of the slimes layer and
significantly hinders the ionic movement. Thus, from the data presented in Fig. 10,
it can be inferred that precipitation of secondary products takes place after -10
mm of slimes have been formed: While the difference in potential in the middle
of the slimes remains nearly constant, a steep increase in the potential difference
between the inner B and inner C reference electrodes can be observed after a
slimes thickness of ~ 10 mm. As a result of the precipitation of secondary products,
the inner potential rises because ions can no longer move at the same rate.
The predicted ionic concentration ratios presented in Fig. 8 are several fold
larger than those experimentally measured (e.g. see Fig. 4.21, and 13,41).
Furthermore, their magnitude indicates that large differences in viscosity and
density across the slimes electrolyte can be present. These differences lead to the
development of natural convection. The extent to which natural convection can
[218]
25
8
I
\
3
28
0
0
15
18
4
0
X
a,
fiiiter-
1 liner
B
I * n e r A - I n n e r
i n n e r B - i n n e r C
J
I
l_
_1 1_
J L
J 1_ _L J I l_
4
6
8
18
12
14
D i s t a n c e from the Slimes/Bulk E l e c t r o l y t e i n t e r f a c e ) mm
j i i_2
16
Fig. 10 Changes in the potential difference between the outer and inner reference electrodes as a function of the
distance from the slimes/bulk electrolyte interface.
From the experimental data presented in Chapter 4, Fig. 4.12.
The difference in potential between the inner reference electrodes was divided by their separation distance and
plotted in this figure.
x +a
=|pln1
;
...31
1 Eddy diffusion as expressed by Eq. 31, incorporates mixing of slimes electrolytes as a result of both
natural and forced convection. It becomes extremely small at the anode/slimes interface and reaches a
maximum value (in this model) at the slimes/bulk electrolyte interface. At a fixed distance from the latter,
the contribution of eddy diffusion increases with time, because this point becomes more remote from the
anode/slimes interface. Therefore, eddy diffusion leads to an unsteady-state electrolyte composition and
thus to a time-dependent polarization at points within the slimes layer that are at a fixed distance from the
slimes/electrolyte interface.
[219]
where:
x = Total slimes thickness, [mm)
x, = distance from the slimes/bulk electrolyte interface at which point D is to be computed,
totai
[ x, < XfetJ,
[mm]
a and P = arbitrary positive constants whose value depends on the electrolysis conditions:
a [mm] and p [cm sec" ]
2
4 . ee ixE-3
, e e. 0a8 0
4.00
8.00
1 2 . 0
16.
Fig. 11 Variation of the Eddy diffusion constant, D , as a function of the distance from the slimes/bulk
electrolyte interface, (a=lxl0" mm and P=5xl0"* cm sec")
E
D values were computed for three x,^ values (4,10, and 14 mm) using the following equation:
E
~
*, + a
D = IP-ln^_-1
F
Eq. 31 can be used to obtain D values across the slimes layer for a
E
fixed
+2
[220]
I
z i . e
6
o
C
o
33
p i**t i i I
eee
i i i i i i i i i i i i i i
2.36
i i i i I i
i i i i i
l i i i
7 . 0 8
4 . 7 2
i I
9 . 4 4
i i i i i
I I
ee
72.0-
36
1. 4 4
. O
o
B
08 =_
1 .
CS1F6-2]
b*aiB. 0 -
[SiF - ]
.720
o
U
24.0
.
g
2
360
8 .
eee^''
. 000
1
00
1 1
1 2 . 0
+2
+2
[221]
60
[SiF - ]
2
1 . 28
ii
. 960
- c
g*
33 .
640
320
8 .
000-k
. 0 0 0
4.00
00
8.00
12.0
16
Fig. 13 Variation in the slimes electrolyte composition as a function of the distancefromthe slimes/bulk
electrolyte interface, when changes in eddy diffusion are accounted for (Case C).
[PbSiFg]^0.28M [HzSiFJfc = 0.74M a = l x l 0 " mm
P = 5xlfX cm sec"
Right scale: Changes in <t> Left scale: Changes in [Pb ], [SiF ' ], and [H ]
2
+2
cd = 139 A m .
2
Derived ratios: D = 72.0 mV, {Pb }=3.48, (SiFy } = 1.38, and {H+} =0.57
+2
1 The predicted mean ionic concentrations within the slimes layer can be made to match those
experimentally measured by changing the eddy diffusion parameters a and p.
[222]
(A)
400
360
320
280
O
3
"S
240
o
O
1
B
200
_L
.
.00O
4.00
8.00
_L
12.0
16.0
CJ
(B)
3
1 . 40
^1.20
i>E
0-
- Density g / c m
i .
0 0
Kinematic Viscosity, c m / s e c
2
CJ
CO
800
4J
K E
600
.000
4.00
_i
i_
8.00
12.0
:0
16.0
[223]
+2
The results of the numerical integration of the data presented in Fig. 14A,
according to Eq. 32, are shown in Fig. 15. The concentration overpotential shown
in this figure was computed according to the following equation:
RT
a
n
Z F
PB+1
Tl = Trin
...33
a^ (bulk)
+2
Using Eqs. 32 and 33, and the computed d> values, the total anodic
overpotential can be obtained as follows:
'H>ri>ril
fert
+ q>
...34
As shown in Fig. 15, the computed anodic overpotential, r^, is smaller than
that experimentally measured, ri . The difference in potential between ru and rj
decreases as the slimes layer thickens until it practically disappears towards the
end of the electrorefining cycle (at -14 mm of slimes). The smaller
values with
respect to TJ at smaller slimes thicknesses, are due to the fact that TJ was measured
as a function of time, (i.e. with respect to a moving interface) whereas ru was
obtained from concentration gradients computed at a fixed time (i.e. with respect
to a fixed interface). That is, the model predicts anodic overpotentials only after
a pseudo-steady state has been established. Thus, the lowest experimental rj
values shown in Fig. 9 (those obtained at the end of the electrorefining cycle) are
nearly equal to the computed
shown in Fig. 15. The small differences between
A
1 Changes in K, p, and v were computed from the local ionic concentration changes using Eqs. 19'to 21.
[224]
.eee
4.ee
8 . e e
1 2 . e
l e . a
Slimes Thickness, mm
Fig. 15 Comparison between the experimental (unsteady-state) and predicted (steady-state) anodic
overpotentials.
Experimental anodic overpotential from Exp. LC2 (outer reference electrode measurements corrected for
uncompensated ohmic drop).
Predicted anodic overpotential from the solution of the Nemst-Planck flux equation and from the data presented
in Fig. 13.
The computed overpotential contains contributions from ohmic resistance, concentration overpotential (of Pb* ),
and migration potential (caused by unequal diffusivities of anions and cations). The ohmic and concentration
overpotentials are shown.
1
[225]
Summary
The Betts electrorefining process (BEP) for lead is successful because it
retains most noble impurities in adherent anode slimes while depositing rather
pure (>99.99%) lead cathodes. The anode slimes account for only 1 to 4% of the
weight of dissolved lead, and so represent an enrichment of a factor of 25 to
100 in the noble impurities (particularly precious metals) present in lead
bullion.
2) It is well known that anode polarization must be limited to less than 200
mV to avoid bismuth dissolution and transfer to the cathodic deposit. This,
together with the reported X-ray diffraction data provides evidence that in-situ
slimes contain metallic bismuth.
3) Electrolyte extracted from within slimes layers shows that it is markedly
enriched in lead fluosilicate and somewhat depleted in fluosilicic acid. The change
in the inner slimes electrolyte concentration is more abrupt towards the
slimes/anode interface and accounts for the hydrolysis of SiF " ions:
2
Summary
[227]
Summary
Chapter
[228]
Summary
ac
Summary
1 A "distributed element" in an analogue model represents properties of the system distributed over
macro distances, such as ionic gradients in solution across the slimes layer, or such as slimes
filament/electrolyte interfaces distributed from the anode/slimes to the bulk electrolyte/slimes interface.
[230]
Summary
assessed experimentally only if the slimes are properly washed and then kept in
a dry atmosphere.
From changes in K as a function of composition for pure H SiF and PbSiF
solutions, individual equivalent conductivities (A,) in H SiF -PbSiF mixtures were
obtained by assuming that A, = X + . From the ^ values and ionic
2
pb
pb+2
ph+2
pb
1 a), is the overall diffusion coefficient for species i, and includes the contributions of molecular and eddy
diffusion.
[231]
Summary
The inclusion of mixing due to eddy diffusion (equal for all ions) accounts
for the unsteady-state observations if the eddy diffusion contribution increases
at a fixed distance from the slimes/electrolyte interface as the anode/slimes
interface retreats. The form of the proposed eddy-diffusion component
D = | p In |
equation. If a and P in this equation are adjusted so that the model fits the
polarization across the whole slimes layer, it also predicts quite accurately the
polarization of intermediate sections, even though these polarizations decline (i.e.
are in unsteady state) with respect to the outer reference electrode measurements.
The mathematical model can be used to predict (a) ionic concentration
gradients (b) total polarization values across anode slimes. The thermodynamics
of the system can be incorporated to predict the onset of hydrolysis and the
dissolution of noble elements.
[232]
Summary
DC
Faradaic
Slimes Secondary
Current
Comments
DC/AC
Products
Analysis
No
No
No
No
AC
No
No
Yes
No
AC
Yes
No
No
No
AC
Yes/
No
Yes
Yes Yes/
No
DC
+2
d!
Yes/
No
No
Yes Yes/
No
DC
Yes/
No
Yes
Yes Yes/
No
AC
Yes/
no
No
Yes Yes/
No
AC
di
[233]
Conclusions
1)
2)
3)
The filaments contained witiin the slimes layer are highly conductive and
are grounded to the anode.
4)
5)
Upon current interruption the anodic polarization decays first abruptly (as
the uncompensated ohmic drop disappears) and then slowly ( due to the
presence of a back E.M.F. created by ionic concentration gradients that decay
slowly).
6)
7)
[234]
Conclusions
[235]
R e c o m m e n d a t i o n s for F u r t h e r W o r k
1) The Nernst-Planck model should be tested and modified to make it predict
the concentration and potential gradients observed in the industrial operation of
the Betts electrorefining process.
2) The eddy diffusion coefficient witiiin the slimes layer is a function of the
electrode height, as well as slimes permeability and slimes thickness. Its value
should be measured and related to electrode polarization for various electrode
heights, especially practical heights found in the industry.
3) The experimental procedures generated in Chapters 4 to 6 can be used to
expand the applicability range of the mathematical model here proposed. An
experimental separation of the parameters accounted for in the model calls for
the use of a horizontal anode (below the cathode) configuration. The immediate
effect of this change will be the presence of large ionic concentrations
approximating those computed from the model in the absence of natural
convection. Furthermore, current interruption decays will have time constants
related strictly to molecular diffusion and migration without the confusion
introduced by natural convection. Vertical electrolysis using electrodes of different
height should also provide information on the presence of natural convection and
its relationship to the establishment of concentration gradients. Studies using
rotating disk electrodes at very low rotation speeds (to avoid detachment of slimes)
can also be used to improve the predictions of the model. All these studies can
be complemented by measuring, at preset times, the resistance of the slimes
electrolyte and the liquid junction and concentration overpotentials (e.g. by using
a superimposed small amplitude cyclic waveform as described in Chapter 6,
section III). These values can be compared directly with those predicted by the
mathematical model.
4) The transport processes witxiin the slimes layer could be related to the
Navier-Stokes equation as an ultimate goal in modelling.
5) The extent to which concentration gradients become established across
the slimes layer can be further studied by adjusting the flow of current to avoid
precipitation of secondary products and impurities dissolution. Periodic
interruption of current together with periodic current reversal can be used to
accomplish those objectives.
[236]
[237]
B I B L I O G R A P H Y Chapter 1
[I]
[2]
[3]
[4]
[5]
[6]
[7]
[8]
[9]
T. Nakamura, F. Noguchi. Y. Ueda, and H. Ito, "A Basic Study on the Oxidation
Smelting of Lead Sulphide", The AusIMM Adelaide Branch, Research and
Development in Extractive Metallurgy [Proc. Conf.], Australia, p. 85-94 (1987) .
[10]
G.C. Quigley and J.V. Happ. "Kinetics of Arsenic Removal from Lead Bullion:', The
AusIMM Adelaide Branch, Research and Development in Extractive Metallurgy
[Proc. Conf.], Australia, p. 187-193(1987).
[II]
J.N. Greenwood, "The Refining and Physical Properties of Lead', Metall. Rev., 6
(23), 279-351 (1961).
[12] L. Gmelln, "Handbuch der Anorganlsche Chemie", Verlag Chemle-GMBH,
Weinheim/Bergstrasee, Germany, System No. 47 Pb [B2], (1972) p. 364-369.
[13]
[14]
M. WadaandT. KobayasM,"Treatrnento/SUmesfromI^adElectrorefining;',
Kogyo Kaishi, 90 (1035), 370-373 (1974).
Nippon
[238]
Bibliography Chapter 1
[ 15]
[16]
[17]
[18]
[19]
European Patent Application, Applic. No. 81301045.1, Filed March 12, 1981, Int.
CL C22 B 3/00, C22 B B / 0 4 .
Recovery",
[20]
U.K. Patent
[21]
[22]
[23]
D.S. Flett
(1985).
[24]
[25]
[26]
Lead
[27]
[28]
A G . Betts, "Electrolytic
January 9, 1902.
[29]
A G . Betts,
[30]
[31]
A G . Betts, "Treating
Dec 8 , 1906.
[32]
AG.
Bibliography Chapter 1
[33]
[34]
[35]
Anonymous, "The Diamond in the Lead Refinery", Cominco Magazine, 23 (8), 17-21
(1962).
[36]
J . B . Hull. "Froducing Lead at Trait', Eng. Min. J . . 139 (6), 34-38 (1938)
[37]
[38]
[39]
[40]
[41]
G. Tremolada and L. Abduini, "Lead Refining with Sulphamate Bath at The A. Tono
e C", Symposium on Sulphamic Acid and its Electrometallurgical Applications,
Milan, p 353-366 (1967).
[42]
Ph
Erzmetall, 22 (7), (1969) p 312; also i n supplement volume same year, p. 128.
[43]
[44]
[45]
[46]
A Ohta, A Ichinose and T. Kohno, "Lead Electrolysis With High Current Densi
atHarimaWorks", T M S paper selection A81-21. TMS-AIME, Warrendale, PA (1981).
[47]
[48]
[49]
[50]
[240]
Bibliography Chapter 1
[51]
[52]
T.E. Harper."The Betts Process at OroyaPeru S.A.", Trans. Amer. Inst. Min. Metall.
Engrs., 121, 283-288 (1936)
[53]
T.E. Harper And G. Reinberg,"Modified Betts Process Yields Refined Lead Bullion",
Eng. Min. J., 136 (3), 119-120 (1935).
[54]
[55]
[56]
[57]
[58]
[59]
[60]
C L . Mantell,
p. 314-322.
[61]
[62]
[63]
[64]
[65]
[66]
[67]
H. Silman, "The Uses of'Hydrofluoric, Fluoboric and HuosiUcic Acids and their Salts
in Electroplating Practice",Galvanotechnisches Kolloqium., 6th, p. 79-88 (1980).
[68]
[69]
" I n d u s t r i a l E l e c t r o c h
[241]
Bibliography Chapter 1
[70]
[71]
[72]
[73]
[74]
[75]
[76]
[77]
[78]
[79]
C.J. Krauss. "Cathode Deposit Control in Lead Electrorefining:', J. Met., 28 (11), 4-9
(1976).
[82]
[83]
[84]
[85]
[86]
Private Communication, H. Kubo, Kamioka Mining Co. Ltd.. January 10, 1988.
[87]
C.J. Krauss "Electrolytic Lead Refining", Canadian Patent 1020491, Nov. 8, 1977.
[88]
I.D. Entchev, N.T. Kuntchev, G A Haralampiev, A.D. Milkov, B.Y. Sotirov, and
D.A Petrov,"Electrolytic Lead Refining", Canadian Patent 928246, June 12, 1973.
[89]
Bibliography Chapter 1
[90]
[91]
[92]
[93]
M. Hansen, "Constitution of Binary Alloys", 2nd ed..McGraw Hill Book Co., New
York, 1958, p. 41, 173, 325 and 1101.
[94]
T. Tanaka, "Adhesion of Anode Slime on the Anode Surface in the Electrolytic Refi
ofLead', Metall. Trans. B, 8B, 651-660 (1977).
[95]
[96]
Y. Nishimura, "Method for Casting a Large Lead Anode Plate", Canadian Paten
1019132, Oct. 18, 1977.
[97]
F. Wenzel and K. Hein, "Jn/luence of the Anode Sludge on the Purity of the Cathode
Metal in the Electrolytic Refining ofLead', Neue Hiitte, 19 (5), 263-267 (1974).
(Translation National Translations Center, NTC 11F-81-11594).
[98]
H.J. Lange, K. Hein and N.K. Thiet."ElectroIytfc Lead Refining at Increased Current
Densities", Neue Hiitte, 28(4), 136-141 (1983).
[99]
[ 101] H. Bombach, K. Hein, J. Korb and H.J. Lange, "Investigation on the Chemical Stability
ofLeadElectrolytes ContainingHexafltuyrosilicicAcid', Neue Hiitte, 31 (9). 347-351
(1986). (Translation MTNTEK TR-1265).
[102] B.G. Ateya and H.W. Pickering, "Effects of Ion Migration on the Concentrations and
Mass Transfer Rate in the Diffusion Layer of Dissolving Metals", J . Appl.
Electrochem.. 11, 453-461 (1981).
[103] J.M. Hornut. G. Valentin and A Storck,"The Film Model for Determining the Effect
ofIonic Migration in Electrochemical Systems", J. Appl. Electrochem., 15, 237-24
(1985).
[104] O. Wein, "The Effect of Migration on Ixuninar Diffusion Layers", Collect. Czech
Chem. Commun., 53, 697-720 (1988).
[243]
Bibliography Chapter 1
[109] L.I. Reznichenko and AD. Pogorelyi,"Relationships Governing the Dissolution of the
Metals of the Anode Slime During Electrolytic Refining", Sov. Non-Ferrous Met. R
(Engl. Transl.), 4 (6), 226-229 (1976) .
[111] W. Eisert,"Electrolytic Refining of Lead by the Betts Process", Neue Hiitte, 4 (1)
29-35 (1959).
[112] I. Pajak. M. Kwarcinskl. A. Bojanowska and B. Kwarcinska,"Electrolytic Refining of
Lead Having a High Bi Content', Pr. Inst. Met. Nieze. 5 (3), 121-130 (1976).
[113]P.Mechenov,"ElectroreJ^nlngo/Lead. Mlnno DeloMet. [Sofia], 17 (6). 30-35 (1962).
chemical Abstracts: 57 [1962]- 14776.
,,
[117] M. Isawa and T. Tanaka,"Research on the Anode Slime From Electrolytic Lead
Refining", Nippon Kogyo Kaishi, 78, 519-526 (1962).
[118] F. Miyashita and G. Miyatani,"On the Anode Phenomena on the Lead Electrolyt
Refining", Technol. Rep. Kansai Univ.. 13. 81-92 (1972).
[119] F. Miyashita and G. Miyatani, "On the Influence of the Impurities in the Electr
on the Anode surface in the Lead Electrolytic Refining", Technol. Rep. Kansai U
14, 61-71 (1973).
[ 121] F. Miyashita and G. Miyatani, "The Influence of Surface Active Agent on the Pr
of the Electrolytic Deposition ofLead', Nippon Kogyo Kaishi, 9 0 (1039), 617-620
(1974).
[122] F. Miyashita and G. Miyatani, "The Influence ofDissolved Oxygen and Surface Acti
Agent on the Electrode Reaction ofLead Electrolytic Refining", Nippon Kogyo K
9 2 (1060), 431-437 (1976).
[123] F. Miyashita and G. Miyatani, "Jn/Iuence of Electrolytes and Surface Active Agent
on the Surface Reaction of Lead Electrode", Nippon Kogyo Kaishi, 9 4 (1085),
485-489 (1978).
Bibliography Chapter 1
p. 317-323 (1967) .
[126] G. Baralis and M. Marone,"Production of 99.9999% Ag, Pb, Sb, and Sn", Metall.
Ital., 7, 494-501 (1967) .
[127] R C . Kerby and H.E. Jackson,"Organic Levelling Agents for Electrolytic Lead
Refining", Can. Metal. Quart., 17, 125-131 (1978).
[139] R C . Kerby and C.J. Krauss, "Bipolar Refining ofLead', U.S. Patent 4416746, Nov.
22, 1983.
[140] R C . Kerby."Electrolyte/or Bipolar Refining ofLead', Canadian Patent 1174198,
September 11, 1984.
[141] RC. Kerby, "Bipolar Refining OfLead', Canadian Patent 1126684, June 29, 1982.
[142]
[144] J.O'M. Bockris and AK.N. Reddy, "Modern Electrochemistry", V. 1 and 2, Plenum
Press, New York, 1970.
[245]
Bibliography Chapter 2
B I B L I O G R A P H Y Chapter 2
[I]
[2]
[3]
J.D.E. Mcintyre and W.F. Peck Jr, "An Interrupter Technique For Measuring The
Uncompensated Resistance of Electrode Reactions Under Potentiostaic Cont
Electrochem. Soc. 117 (6), 747-751 (1970).
[4]
[5]
[6]
[7]
[8]
[9]
[10]
M. Hayes and AT. Kuhn, "Techniques For The Determination of Ohmic Drop
Half-Cells and Full Cells: A Review", J. Power Sources, 2, 121-136 (1977/78).
Z. Nagy and J.T. Arden, "Error Analysis of The Coulostatic Technique and
Comparison To Other DC Relaxation Techniques For The Measurement of Kineti
Electrode Reactions", J. Electrochem. Soc, 130 (4), 815-822 (1983).
K.J. Vetter, "Electrochemical
Kinetics,
Theoretical
and Experimental
[12]
[13]
[14]
[15]
Bibliography Chapter 2
[16]
J.R Macdonald ed., "Impedance Spectroscopy". John Wiley and Sons, New York,
1987.
[17]
[18]
[19]
[20]
[21]
[22]
[23]
[24]
B.V. Tilak and B.E. Conway, "Overpotential Decay Behavior - J. Complex Electro
Reactions Involving Adsorption:', Electrochim. Acta, 21, 747-752 (1976).
[25]
B.V. Tilak. C G . Rader, and B.E. Conway, "Overpotential Decay Behavior - II.
Pathways Involving Discharge, Recombination and Electrochemical Desorpti
Adsorbed Intermediates", Electrochim. Acta, 22, 1167-1178 (1977).
[26]
[27]
[28]
[29]
[30]
C.K. Dyer and J.S.L. Leach, "Electrode Impedance Measurements At Short Tim
By Square-Wave Technique", Electrochim. Acta, 19, 695-700 (1974).
[31]
[32]
RP. Buck, "Diffuse Layer Charge Relaxation At The Ideally Polarized Electrode
Electroanal. Chem., 23, 219-240 (1969).
[247]
Bibliography Chapter 2
[33]
[34]
[35]
[36]
[37]
[38]
A A Pilla, "A Transient Impedance Technique For The Study of Electrode Kinet
J. Electrochem. Soc, 117 (4), 467-467 (1970).
[39]
[40]
[41]
K.G. Beauchamp and C.K. Yuen, "Digital Methodsfor Signal Analysis", George Allen
& Unwin, Great Britain, 1979.
[42]
[43]
CD. McGillem and G.R Cooper, "Continuous and Discrete Signal and System
Analysis", 2nd ed., Holt Rinehart and Winston, USA 1984.
[44]
[45]
[46]
[47]
J.E.B. Randies, "Kinetics ofRapid Electrode Reactions", Disc. Farad. Soc, 1, 11-19
(1947).
[48]
[248]
Bibliography Chapter 2
[50]
[51]
[52]
G.W. Walter," A Review ofImpedance Plot Methods Used For Corrosion Perform
Analysis of Painted Metals", Electrochim. Acta, 26 (9). 681-703 (1986).
[53]
[54]
[55]
S. Venkatesh and D.-T. Chin, "The AUernating Current Electrode Process", Isr.
Chem., 18, 56-64 (1979).
[56]
[57]
[58]
[59]
[60]
[61]
[62]
W.H. Smyrl and L.L. Stephenson, "Digital Impedance For Faradaic Analysis in
Corrosion in Oxygenated .1 N HCl', J . Electrochem. Soc, 132 (7), 1563-1567
(1985).
[63]
Handbook", AT. Kuhn ed., John Wiley & Sons, Great Britain 1987.
[64]
[65]
[66]
[67]
[249]
Bibliography Chapter 3
[68]
[69]
[70]
[71]
R Francis," The use of Small amplitude Cyclic Voltammetry to Study the Corr
of Copper and Copper Alloy s", Paper 19 in ref [21].
[72]
BIBLIOGRAPHY Chapter 3
[I]
"Fluosilicic Acid', Dechema Data Sheet, Serial No. 49.235.050.1, October 1959.
[2]
[3]
[4]
[5]
[6]
[7]
V.M. Masalovich, G JV. Moshkareva and P.K. Agasyan, "Study ofComplex Formation
inSokitions ofHydrofluoric and Silicic Acid', Russ. J. Inorg. Chem., 2 4 (2), 196-198
(1979).
[8]
[9]
[10]
[II]
[250]
Bibliography Chapter 4
[12]
B I B L I O G R A P H Y Chapter 4
[1]
[2]
[3]
AS. Gioda. M.C. Giordano, and V.Z. Macagno, "The Pb/Pb Exchange Reaction In
Perchlorate Acidic Solutions", J. Electrochem. Soc, 124 (9), 1324-1329 (1977).
[4]
S.H. Glarum and J.H. Marshall, "An Admittance Study of The Lead Electrode", J
Electrochem. Soc. 131 (4), 691-701 (1984).
[5]
C.J. Bushrod and N.A Hampson, "Anodic Behavior ofLead in Perchloric Acid I
Continuous and Interrupted Polarization ', Br. Corros. J., 6, 87-90 (1971).
+2
[6]
[7]
[8]
B I B L I O G R A P H Y Chapter 5
[1]
S.H. Glarum and J.H. Marshall. "An Admittance Study of the Lead Electrode",
Electrochem. Soc. 131(4), 691-701 (1984).
[2]
[3]
K.J. Vetter,
'Electrochemical
Kinetics,
Theoretical
and Experimental
AS. Gioda. M.C. Giordano, and V.Z. Macagno, "The Pb/Pb* Exchange Reaction in
Perchlorate Acidic Solutions", J. Electrochem. Soc. 124 (9), 1324-1329 (1977).
[5]
[6]
J.R Macdonald ed., "Impedance Spectroscopy", John Wiley and Sons, New York,
1987.
[8]
[9]
S. Venkatesh and D.-T. Chin, "The Alternating Current Electrode Process", Isr.
Chem., 18. 56-64 (1979).
[251]
Bibliography Chapter 5
[10]
[11]
[12]
G.W. Walter, "AReview ofImpedance Plot Methods Used For Corrosion Perform
Analysis of Painted Metals", Electrochim. Acta, 26 (9), 681-703 (1986).
[13]
[14]
[15]
[16]
J.E.B. Randies, "Kinetics ofRapid Electrode Reactions", Disc. Farad. Soc, 1, 11-19
(1947).
[17]
[18]
[19]
[20]
[21]
[22]
[23]
[24]
J. S. Newman,
[25]
" E l e c t r o c h e m i c a lS y s t e
[252]
Bibliography Chapter 6
B I B L I O G R A P H Y Chapter 6
[I]
[2]
[3]
[4]
[5]
N.E. Hill, W.E. Vaughan, A H . Price, and M. Davies, "Dielectric Properties and
Molecular Behaviour", Van Nostrand Reinhold Company, London 1969.
[6]
[7]
[8]
[9]
J.N. Agar and F.P. Bowden, "The Kinetics of Electrode Reactions I and II', Proc. R
Soc. 169A 206-234, (1939).
[10]
[II]
[12]
R Francis," The Use of Small amplitude Cyclic VoUammetry to Study the Corr
of Copper and Copper Alloys", Paper 19 in ref [in.
[13]
[14]
[15]
[253]
Bibliography Chapter 7
[16]
[17]
[18]
[ 19]
[20]
[21]
[22]
[23]
J. S. Newman,
[24]
[25]
G. Tremolada and L. Abduini, "Lead Refining with Sulphamate Bath at The A. Tono
eC", Symposium on Sulphamic Acid and its Electrometallurgical Applications,
p. 353-366, Milan (1967).
" E l e c t r o c h e m i c a lS y s t
B I B L I O G R A P H Y Chapter 7
[1]
[2]
H. J. Lange. K. Hein and N.K. Thiet, "Electrolytic Lead Refining at Increased Curren
Densities", Neue Hiitte, 28 (4), 136-41 (1983).
[3]
F. Wenzel and K. Hein, "Influence of the Anode Sludge on the Purity of the Cath
Metal in the Electrolytic Refining ofLead', Neue Hiitte. 19 (5), 263-67 (1974).
(Translation National Translations Center, NTC 11F-81-11594).
[4]
[5]
[6]
[7]
S. Hirakawa and R. Oniwa, "Method for Producing High Purity Lead by Removi
Bismuth', Canadian Patent 1023691, Jan. 3, 1978.
[254]
Bibliography Chapter 7
[8]
G. Baralis, "The Production of High Purity Lead', Lead 65 [Proc. Conf.]. Arnhem, p
317-323. 1967.
[9]
G. Baralis and M. Marone. "Production of 99.9999% Ag, Pb, Sb, and Sn", Me
Ital.. 7, 494-501 (1967).
[10]
[11]
A G . Betts, "Lead Refining By Electrolysis", J. Wiley and Sons, New York, 1908.
[12]
[ 13]
[14]
[15]
A G . Turnbull, "A General Computer Program For The Calcultion And Plottin
Thermodynamic-Stability Diagrams", Joint Australian Institute Of Mining And
Metallurgy/University Of South Wales, Extractive Metallurgy Symposium, Sidney
1977, 10 pp.
[16]
[17]
C E . Roberson and T.B. Barnes, "Stability of Fluoride Complex With Silica and I
Distribution in Natural Water Systems", Chemical Geology, 21, 239-256 (1978).
[18]
[19]
N.N.Golovnev, "The Influence ofpH on the Formation ofSiF ' in Aqueous Solu
Russ. J. Inorg. Chem., 31 (3), 367-368 (1986).
[20]
V.N. Krylov and E.V. Komarov, "Some Properties of Ftuorosilicic Acid', Russ.
Inorg. Chem., 16 (6), 827-829 (1971).
[21]
[22]
[23]
P.M. Borodin and N.K. Zao. "Equilibria in the Li^iF -HCl0 -H 0 by the F Nuclear
Magnetic Resonance Method', Russ. J. Inorg. Chem., 16 (12). 1720-1722 (1971).
[24]
19
[255]
Bibliography Chapter 7
[25]
[26]
[27]
L.N. Arkhipova and S.Ya. Shpunt, "Some Properties ofFluorosilicicAcid', Tr. NIUIFa,
208, 88-103, (1965).
(Translation National Translations Center. NIC 99D-85-22456).
[28]
[29]
[30]
[31]
[32]
[33]
[34]
[35]
[36]
[37]
[38]
[39]
H.P. Meissner and C.L. Kusik, "Double Salt Solubilities", Ind. Eng. Chem. Process
Des. Develop.. 18 (3), 391-394 (1979).
[40]
[41]
[256]
Bibliography Appendix 1
[42]
[43]
[44]
[45]
[46]
[47]
[48]
[49]
[50]
[51]
[52]
[53]
BIBLIOGRAPHY Appendix 1
[1]
[2]
[3]
[4]
[5]
B.G. Ateya and H.W. Pickering, "Effects of Mass Transfer In The Aqueous Phase
On Repassivation of Activated Surfaces and The Stability of Protective Films
Passivity of Metals [Proc. Conf.], RP. Frankental and J . Kruger, Eds., The
Electrochemical Soc, Corrosion Monograph Series, U.SA, P.350-368. 1978.
[257]
Bibliography Appendix 2
[6]
B.G. Ateya and H.W. Pickering, "Effects of Ion Migration on the Concentrations and
Mass Transfer Rate tn the Diffusion Layer of Dissolving Metals", J . Appl.
Electrochem., 11, 453-61 (1981).
[7]
[8]
[9]
[10]
[11]
C.Y. Chan and K.H. Khoo, "Re-determination of mean Ionic activity coefficients
the system HCl+KCl+H 0 at 298.15 K and correlations between Horned and Pitz
Equations", 1371-1379.
2
[12]
BIBLIOGRAPHY Appendix 2
[1]
[2]
[3]
B.G. Ateya and H.W. Pickering, "Effects of Ion Migration on the Concentrations
and Mass Transfer Rate in the Diffusion Layer of Dissolving Metals", J. Appl.
Electrochem., 11. 453-61 (1981).
BIBLIOGRAPHY Appendix 3
[1]
[4]
T.F. Bogart Jr., "Laplace Transforms: Theory and Experiments", John Wiley &
Germany 1988.
Inc.USA 1985.
[6]
1971.
[258]
Bibliography Appendix 3
[7]
AK. Walton, "Network Analysis and Practice", Cambridge University Press, Great
Britain, 1987.
[8]
[9]
[10]
[ 11]
[12]
[13]
[ 14]
[15]
A A Pilla, "A Transient Impedance Technique for the Study of Electrode Kineti
J. Electrochem. Soc, 117 (4), 467-467 (1970).
[16]
[17]
[18]
W.H. Smyrl, "Digital Impedance for Faradaic Analysis II. Electrcdissolution ofC
HCV, J. Electrochem. Soc, 132 (7), 1555-1562 (1985).
[19]
[20]
W.H. Smyrl and L.L. Stephenson. "Digital Impedance for Faradaic Analysis III
Corrosion in Oxygenated .1 N HCV, J . Electrochem. Soc, 132 (7). 1563-1567
(1985).
H. Gerisher, Z. Phys. Chem., 198, 286 (1951).
[21]
[22]
[259]
Bibliography Appendix 3
[23]
[24]
[25]
[26]
[27]
[28]
J.R Macdonald ed., "Impedance Spectroscopy", John Wiley and Sons, New York,
1987.
[29]
[30]
[32]
[33]
[34]
[35]
[36]
M. Hayes and AT. Kuhn, "Techniques for the Determination of Ohmic Drop
Half-Cells and Full Cells: A Review", J. Power Sources, 2, 121-136 (1977/78).
[37]
[38]
[39]
Handbook", AT. Kuhn ed., John Wiley & Sons. New York, 1987.
E.J. Muth, "Transform Methods", Prentice-Hall Inc., USA 1977.
[260]
Bibliography Appendix 4
[40]
[41]
[42]
[43]
[44]
[45]
M.W. Kendig, E.M. Meyer, G. Lindberg, and F. Mansfeld, "A Computer Analysis of
Electrochemical Impedance Data", Corros. Sci., 23 (9), 1007-1015 (1983).
[46]
S.D. Stearns, "Digital Signal Analysis". Hayden Book Company. Inc., USA 1975.
[47]
[48]
C D . McGillem and G.R Cooper, "Continuous and Discrete Signal and System
Analysis", 2nd ed.. Holt Rinehart and Winston, USA 1984.
[49]
[50]
[51 ]
BIBLIOGRAPHY Appendix 4
[1]
V.N. Krylov and E.V. Komarov, "Some Properties of Fluorosilicic Acid', Russ. J
Inorg. Chem., 16 (6), 827-829 (1971).
[2]
P.M. Brinton, L.A Sarver, and A E . Stoppel, "The Titration of Hydrofluoric and
Hydrofluosilicic Acids in Mixtures Containing Small Amounts of Hydrofluos
Acid', Ind. Eng. Chem.. 15 (10). 180-1081.
[3]
[261]
Bibliography Appendix 4
[4]
[5J
P.M. Borodin and N.K. Zao, "Equilibria in the Li2SiF -HCl0 -H 0 by the 19F Nuclea
Magnetic Resonance Method', Russ. J. Inorg. Chem.. 16 (12). 1720-1722 (1971).
[6]
[7]
[8]
[9]
[10]
W. Lange. "The Chemistry ofFluoro Acids", in Fluorine Chemistry, J.H. Simons ed.,
Vol. 1, p. 125-188, Academic Press (1950).
[11]
[12]
[13] Gmelins Handbuch der Anorganische Chemie. E.H.E. Pretsch ed.. V. 15 Si [B],
p. 614-653, Verlag Chemie, Germany (1959)
[14]
[15]
[16]
[17]
[262]
Bibliography Appendix 7
[18]
[ 19]
[20]
[21]
[22]
[23]
J.R Cooke and M.J. Minski, "Kinetics and Equilibria ofFluorosilicate Solutions wi
Special Reference to the Fluoridation of Water Supplies", J. Appl. Chem., 123-12
March 1962.
[24]
[25]
[26]
O.P. Subbotina, L.N. Arkhipova, and M.N. Tsybina, "Composition and Properties
of Fluosilicic Acid', Tr. NH Po Udobr. i Insektofungitsidam, 228, 56-60 (1976).
(Translation to Spanish ICYT: 5933).
[27]
[28]
A A Ennan, V.E. Blinder, and T.S. Borisenko, "The Properties of Fluosilicic Acid (A
Review)", Russ. J. Physical Chemistry. 51 (8), 1255, (1977).
[29]
C E . Roberson and T.B. Barnes, "Stability of Fluoride Complex with Silica and It
Distribution in Natural Water Systems", Chemical Geology, 21, 239-256 (1978).
[30]
[31]
[32]
[33]
BIBLIOGRAPHY Appendix 7
[1]
Biography Appendix 8
[2]
[3]
BIBLIOGRAPHY Appendix 8
[1]
[2]
C.J. Bushrod and N.A Hampson. "Anodic Behavior ofLead in Perchloric Acid I.
Continuous and Interrupted Polarization", Br. Corros. J., 6, 87-90 (1971).
BIBLIOGRAPHY Appendix 1 0
[1]
[2]
[3]
[264]
A p p e n d i x 1 N u m e r i c a l S o l u t i o n of t h e N e r n s t - P l a n c k F l u x
E q u a t i o n s a n d Its A p p l i c a t i o n t o T h e B e t t s Process
I. Objective:
To introduce an algorithm to solve numerically the Nemst-Planck flux
equations. From the solution of this equation the concentration of ionic species
and the potential profile throughout the slimes layer will be obtained as a function
of the electrolysis conditions.
n. A s s u m p t i o n s :
1) The Nernst-Planck flux equations can be applied in concentrated solutions
when they are solved in their fundamental form \
2) Dissolution of lead is the only Faradaic reaction. This reaction takes place
exclusively at the anode/slimes interface and proceeds without kinetic limitations
(i.e. i o - * ~ for Pb->Pb +2e) .
+2
3) Noble impurities originally present in the lead anode report to the slimes
and remain unreacted during the whole electrorefining cycle (i.e. i-0 for the
dissolution of Sb, Bi, As, and other noble impurities) .
4) Hydrolysis and secondary products precipitation can be neglected .
5) Mixing of electrolytes within the slimes layer can be accounted for by
mcorporating the eddy diffusion component, D in the overall diffusion term, 2}
3
1 This assumption implies that the cross coefficients in the Onsager phenomenological equations can be
neglected. Interrelationships between the involved ions can be incorporated within the absolute value of
activities and mobilities.
2 Assumptions 2 and 3 are based on the fact that upon dissolution of lead bullion electrodes, lead
dissolution at the anode/slimes interface is the main reaction. Furthermore, under normal operation of the
BEP the slimes remain unreacted and polarized.
3 Dissolution of noble impurities is a function the potential gradient across the slimes layer. Once the
potential gradient is known, the point at which noble compounds start to react can be predicted.
4 Assumption 4 implies that the concentration gradients across the slimes layer do not depend on the
presence of secondary products or hydrolysis. This assumption can be relaxed once changes in activities
as a function of position are known. This will allow the hydrolysis point to be predicted.
5 Eddy diffusion will be used to incorporate mixing due to natural convection within the slimes electrolyte.
Natural convection arises due to concentration differences that are created by the electrolysis itself.
Forced convection takes place in regions in which the velocity field is not influenced by the concentration
field around the electrode, and the density of the solution is constant [1,2]. Within the slimes electrolyte,
large concentration gradients exist and natural convection is perceived to be much larger than forced
convection.
[265]
Fundamental Equations
8) Only three ionic species are present (i.e. the PbSiF and H S i F molecules
6
Species a:
Pb
Species b:
SiF "
Species c:
+2
m. F u n d a m e n t a l E q u a t i o n s
Given the previous assumptions, the Nemst-Planck flux equations that need
to be solved simultaneously are as follows 13-61:
dlna
1
o c
N
d<& I(t)
-+u c =
dx
d\na
d\na
"dx
ZF
dQ>
Z.C.+Z C,+Z,C,
t
where:
Species a = Ft/
Species b = S1F
Species c = H
Q = concentration of species 1, (mol cm"].
Z, = charge number of species 1, [eq mol" !.
2
...rn
=0
...[iv]
1 The electrolyte composition for a fixed position across the slimes layer may change as a result of
hydrolysis, precipitation of secondary products and natural convection. Depending on the height of the
electrode (among other electrolysis parameters) local ionic concentrations may show gradients in both the
vertical and horizontal coordinates.
2 Electrical neutrality rules out the presence of charge imbalance at any point throughout the slimes layer.
3 H SiF -PbSiF are strong electrolytes and their dissociation is expected to be very high.
2
4 D can be considered to be equal for all the ions. Furthermore, it can be modified so as to incorporate
changes in porosity and tortuosity of the ionic flow across the slimes layer for different anode
compositions and/or electrolysis conditions.
E
5 Notice that when the molecular diffusion coefficient is smaller than the eddy diffusion constant, the
overall diffusion coefficient is nearly the same for all the components. In this case, eddy diffusion
practically equalizes the effective diffusivities for all the components [7].
[266]
Fundamental
Equations
and
= 10,
^ total
Eq. i states that the only species being generated at the anode is Pb whereas
Eqs. ii and iii indicate that H and SiF " do not react and their equihbrium
concentration is only a function of the migration and diffusion gradients. Eq. iv
is the electroneutrality condition.
+2
Anode/Slimes
Slimes/Electrolyte
Interface
nterface
Slimes
Layer
Lead
Anode
Bulk
Electrolyte
Fig. 1 Coordinate system used to
obtain the numerical solution of the
Nemst-Planck flux equations.
=-x
x
X =0
0
Z;F
where Z* is the effective charge of species i exposed to the electric field, Z* <z,-.
2 The Betts electrorefining process for lead is normally carried out under nearly galvanostatic conditions
(i.e. / (t) = constant).
[267]
Derived Equations
TV. D e r i v e d E q u a t i o n s
Equation iv can be differentiated using the following relationship:
4
then,
d_
d_(aAj^-a'i^lda,
dx
dx^)
{Ci)
a dy
t
jf
y dx y}dx
t
rearranging terms:
d
dx
dx
dx _
Thus, the electrical neutrality equation (Eq. iv) can be equivalently expressed
as follows:
dlna d lny
a
dx
dx
+z c
dlna d\ny
b
dx
dx
+Z.C.
dlna d\ny
c
dx
dx
=0
...[v]
d lny
dlny _a
dx ~ "
y
d\na
dx
>
d lny
d\na
_
d\na
da>
=n
= q>
D a + uO = -n
a
...1
...2
[268]
Da +u *=0
e
...3
2C (a - y.) - 2C(Z -y ) + C (3 -y ) = 0
a
..A
c,=c;
Y.=7f
0 =0
a,
b
a,
y, y,
y,
<S>)
dlna,
dlnC, dlny,
dx
dx
dx
These equations can be used to obtain either ^ or ctj once the concentrations
of all the species are known.
V . F o r m u l a t i o n of the E q u a t i o n s t o be Solved N u m e r i c a l l y
Eqs. 1 to 4 constitute a system of differential simultaneous equations that
have to be solved in order to yield the values for the four unknowns as a function
of the distance from the slimes/electrolyte interface : a (x),a (x),a (x),(x).
a
By
D u . a - D u a = -riu,
a
a-Yl
. . . 5
1 For a fixed electrolyte concentration, the individual activity coefficients, y, and the individual activities
can be derived from theoretical and experimental information [8-12].
[269]
...6
(D \i a -n\i )
b
(D a a -Tl\i )
b
and
AL
a.
...7
li
+ C (a - ) = 0
a -n
-Ya
c Ye
a ( r
2 C ^ - 2 C
^ = 2C y -2QY + C Y + 2 ^ n
a
(:
Thus if initial values for the Y and other variables are provided, a finite interval
numerical approach could be used to solve for a \ Once a is known, a and a
c
can be obtained from Eqs. 7 and 5 respectively. The fourth unknown, cp, can be
obtained from either Eq. 2 or from Eq. 3.
VI. F i n i t e Difference E q u a t i o n s
The
differential
equations can be
[270]
'
Ax
a,dx ~ a{
from which:
aj =aj + Ax-af-a
+l
where the superscripts j and j+1 refer to different nodal positions (see Fig. 1
) and Axis the finite difference distance .
A similar equation can be written for the change in the activity coefficients:
l
-_
I<*Y,_
ytdx'Y,
71
Ax
from which the value of the activity coefficient at the next nodal point can be
obtained:
7j =^ + Ax-7j-Y .
+1
Local ionic concentrations can also be put in a finite difference form by using
the following equations:
dlna,
dlnCi dlny,
dx
a >
dlnCi
dx
dx
_ _
dlnji
'
dx~
dQ
dx
From the previous equation, the local values of the ionic concentrations can
be obtained as follows:
C/
+ 1
= Cj +
Cf-Ax-la,-^
+1
qV + Ax <D
[271]
Input data
VII. A l g o r i t h m for t h e N u m e r i c a l S o l u t i o n
Given the equations presented in the previous section the numerical problem
consists in finding the values of the activities and potential at the next nodal point
given their initial values (a/=>a/
, d>{=>Of ). To obtain these values, the
following variables have to be known at the following position:
+1
DJ
N{
+1
a{
YT
The concentrations at point j are known and the values of the above
parameters can be easily computed at that position. On the other hand the
concentrations at the next nodal point are not known in advance and an estimate
has to be used to compute the Y< values. The values for the i* parameters can
+1
be improved by repeating the iteration procedure until the assumed and computed
concentrations in the next nodal point are nearly equal. In the following
paragraphs the algorithm to solve such iteration problem will be outlined.
A. Input data
To solve the Nernst-Planck flux equations in their finite difference form, the
following information has to be known in advance:
at x=0 (i.e. in the bulk electrolyte):
r =C
c
= c
c =c:
c
and
q> = q> = 0
For a fixed value of the concentration of the ionic species, the following
variables have to be computed according to pre-established relationships :
1
1 i y values have to be provided also as a function of the slimes thickness, x , because of the presence of
eddy diffusion.
J
[272]
Algorithm
t=f(ci,ci,ci)
i=Aci,ci,ci)
i=f(c ,ci,ci)
j
ri =f{ci,ci,ci,x )
d^ficixuci^)
ri=f(ci,ci,c> )
\i=fZ,cl.ci)
Ni=f<ei,ci,ci)
K=Aci,cUC )
j
Ni=f(ci,ci,ci)
3>
From this value, the amount of lead dissolved can be obtained ,Ax, :
5
Ax [cm] = c-to
Q
i+1
i+l
Cb
cm
103.8grt>
\cm'
eq
x ^- S X \\2gPb
C
Thus, at any point, j , the distance from the anode/slimes interface, Xj, can
be computed from:
x.
[cm] =
Z Ar'
7=7
;=o
B. Algorithm
1) From the function I (t) obtain the time interval, At. From this value obtain
the changes in slimes thickness, Ax, for the different nodes. Set j =0
2) Set the node number to j=j+l
1 Even though the time intervals, At are equal and uniformly spaced, depending on the shape of the
current waveform, / (t), the distance intervals, Ax, may not be equal or uniformly spaced.
2 For any time interval Ar = t -t', At>0.
i+1
3 This value is an exact quantity as the shape of the / (t) waveform is known precisely.
4 The value for the current at the first nodal point (x=0) is given by: 7 =/(0)
1
5 Ax and x are negative quantities according to the frame of reference shown in Fig. 1. Thus both values
have to be multiplied by -1 before using them in their respective equations.
[273]
Algorithm
4) Given the initial electrolyte composition obtain the values of the following
parameters:
ci=c
c;=c;
c/=c;
vi
vi vi
yi
yl
yi
< at <
K AY Nj
D;
Di
c'-ti
2C y -2C y + C y + 2
a
a, =
DVi_
a -:a
h
" mi
a =-^-(Di^ -n at)
j
1 Notice that this assumption implies that for the first interval the activity coefficients do not change:
[274]
Algorithm
Cl +
Cl-Ax'-fc-yJ
C +
C' -to'-& -y ]
+1
values obtain:
yi
Ax'
J
h
Ax'
Yc
_iyi -yi
yi Ax'
+1
'
C{ \new)
+
<0.01
1 2 ) Once the convergence criteria have been fulfilled, go back to step 2 and
compute the concentrations and potentials for the next step.
1 The convergence criteria have to be fulfilled by two of the three computed concentrations (the third is
obtained from electrical neutrality). A successful convergence criterion was found by averaging the
assumed and computed concentrations and using the new value to improve the solution:
Xnew) + C \old)
i+
[275]
Appendix 2
Equations
A n a l y t i c a l S o l u t i o n of t h e N e r n s t - P l a n c k F l u x
I. Objective:
To solve analytically the Nernst-Planck Flux equations. This solution is to
be obtained for the case in which activities are equal to concentrations and eddy
diffusion is not present. From this solution the rmgration/diffusion ratio is derived.
Also an equation that describes local changes in resistivity of the entrained
electrolyte is obtained.
n. A s s u m p t i o n s :
1
1) Dissolution of lead is the only Faradaic reaction. This reaction takes place
exclusively at the anode/slimes interface and proceeds without kinetic limitations
(i.e.
for Pb->Pb +2e").
+2
2) Noble impurities originally present in the lead anode report to the slimes
and remain unreacted during the whole electrorefining cycle (i.e. Lj-^O for the
dissolution of Sb, Bi, As, and other noble impurities).
3) Hydrolysis and secondary products precipitation can be neglected.
4) Mixing of electrolytes within the slimes layer is neglected (i.e. D = 0 and
v=0).
5) Activities are equal to concentrations .
6) The water mol fraction is the same for all the ions and is equal to 1.
a
Species a:
Species b:
Species c:
Pb
SiF "
H
+2
1 These assumptions are similar to those used to obtain the numerical solution (see Appendix 1) except
for assumptions 4 to 6.
2 It can be shown that the solution here presented is also valid when individual activity coefficients are
constant (see Appendix 1).
[276]
Fundamental Equations
m. F u n d a m e n t a l E q u a t i o n s
Given the previous assumptions, the Nernst-Planckfluxequations that need
to be solved simultaneously are as follows [l-si:
" dx
-D
RT dx
dC
Z Fd&
_ i c -J-=r-
dx
dC
tb
RT dx
C
dx
...2
=0
Z F d<f>
...1
ZF
n
...3
RT dx
...4
Z C +Z C +Z C =0
a
where:
Species b = SiF
Species c = H
Species a = Pb
C, = concentration of species i, [mol cm' ].
Z, = charge number of species i, [eq mol' ].
>, = molecular diffusion coefficient [cm sec' ]
2
+2
1 l
Eq. i state that the only species being generated at the anode is Pb whereas
Eqs. ii and iii indicate that H and SiF " do not react and their equilibrium
concentration is only a function of the migration and diffusion gradients. Eq. iv
is the electroneutrality condition.
+2
IV. B o u n d a r y C o n d i t i o n s
Eqs. 1 to 4 constitute a set of simultaneous differential equations that have
to be solved subject to the following boundary conditions (see Fig. 1):
Z'F
Z* is the effective charge of species i exposed to the electric field, z' was assumed to be equal to Z
[277]
Derived Equations
Anode/Slimes
Slimes/Electrolyte
Interface
Interface
Bulk
Slimes
Lead
Anode
Electrolyte
Layer
x=x
x=0
Fig. 1 Coordinate system used to obtain the analytical solution of the Nernst-Planck flux equations,
at x =x
at x=0
C =C
C =C
C =C
C = Cf
C C
c
... 5
...6
C C
<D=q>
O=0
... 7
...8
V. Derived Equations
Eqs. 1 to 4 are to be rearranged so that the concentration and potential
profiles can be obtained as a function of the distance from the slimes/bulk
electrolyte interface, x. This is done as follows:
Dividing Eqs. 1, 2, and 3 by D , D , and D respectively and substituting the
corresponding Z, values in these three equations and in Eq. 4, the following
relationships are obtained:
A
dx
"RT dx
2FD
...
[278]
Derived Equations
dC
dx
dq>
=0
RT dx
F
-2C
...10
bh
dC
F d<b
+C
=0
dx RTdx
...11
2C + C = 2C
...12
Cc
dC
dC
** * *
a
F d
+(2C +C
2C
...13
dC
dx
dx
dC _
dx
...14
2FD
5, 6, and 7:
...15
= c ; + c + c ? + - ^ -
+c
0
c
...16
2FD C
a
'
...17
and
C = total bulk electrolyte concentration , C = C + C , [mol cm' ]
3
...18
...19
...20
[279]
Analytical Solution
VI. A n a l y t i c a l S o l u t i o n
All the terms to the right of the equal sign in Eq. 20 are known or can be
assumed. Thus, the problem consists in finding algebraic relationships between
the C , C , and C terms. These relationships were obtained as follows:
rearranging Eq. 10 and separating variables:
a
dx
= 2 C
dC
.21
RTdx
2Fd
.22
C ~RT dx
b
C =Cexp
( -2F_
RT
...24
[RT
...25
between C and <& can be obtained from Eq. 11 using b.c. 7 and 8:
c
...26
A relationship between C and C can be found by multiplying Eqs. 25 and
b
26:
c (c ) =cT(cTj
2
fr
the C (Cf
b
...27
K = C {c J
2
...28
[280]
Analytical Solution
c = ^ -
(C f
.29
...30
{C f
c
*2
~(C f
...31
FD C
...32
+ 4K = 0
2
Eq. 32 is a cubic equation that has to be solved for C given K K^, and
c
lt
lx
FD.C
. From the obtained C values, C , C , and <I> values can be obtained from Eqs.
29, 31, and 25 respectively.
Eq. 32 was used to obtain changes in the dimensionless concentrations and
in the migration potential as a function of the bulk electrolyte composition and
of the dimensionless parameter 0 (see Figs. 2 to 4):
c
0 =
Ix
FD C
...33
[281]
[282]
[283]
[284]
VII. Changes i n t h e S p e c i f i c R e s i s t i v i t y o f t h e E n t r a i n e d
Electrolyte
The specific resistivity of dilute electrolytes can be expressed as follows:
P,=
1
RT 1
FXZ.n.C,
F'lZfD.Q
where:
Pi= specific resistance, [flcm].
The dimensionless ratio between the resistivity of the bulk electrolyte and
that of the entrained electrolyte, p(/), is given by the following equation:
-
. _ _ lZ?DiC,(Bulk Electrolyte)
p(0-0)_;
p(0 = O)
lZ?D,C,(Slimes Layer)
/ c
( / ) =
onsta
1 Electrical conductivities within the slimes electrolyte are likely to decrease rather than increase (see
Appendix 1 and Chapter 7).
[285]
0,2 .
C, .2 mol/1 C , = .2
Pi f. A .'/L
C, " .8 mot/I C , .2
I
2
mol/1
9i.:?
mol/1
I
8
I '
10
'
12
'
I
14
16
'
20
18
- i
0
i '
| i i i
| i
4
i i
| i i i |
6
| i i
10
12
|
14
|
16
18
i i
i
20
Fig. 5 Effect of the variable on the dimensionless resistivity, p(I), for different bulk electrolyte
compositions.
[286]
V U I . C o m p u t a t i o n of t h e Diffusion F l u x
As shown by Eq. 1, the ionic flux has two contributions:
dC
-D
dx
a
Z F rfq>
a
Diffusion Flux
-DC
, =
" RT dx
Migration Flux
dC
The
dx
dx
2 + = 2
...34
dx
1 dC
F d<Z>
C dx
RT dx
d
l dC
2C dx
_ C dC
dx
..35
C dx
c
dC
C dC
dx
dx
C dx
dC
as a function of
dC
= -2-
dx
dx
dC
and
2C
dx
dC
dC
...36
C + 4C dx
c
4C
dC
...37
dx ~C + 4C dx
c
dC.
dx
1+
4C
2C,
C + 4C
c
C + 4C
c
dC
b
dx
C + 4C +
c
4C -2C
C. + 4C
2FD
[287]
dC _
dx
\C + 4Cl
2FD
.8C,-C
...38
dx
.38.a
2FD
The bracketed term in Eq. 38.a can be further simplified by using Eq. 29:
4iC,
C + 4C ^
Cc
SC -C
b
(c fJC f
c
jfi_
(cf
+ 4K
SK -(C f
U c
dC^
dx
2FD 8tf - ( C J
2
3
" dx
IF
8/T -(CJ
.39
IX. C o m p u t a t i o n of the M i g r a t i o n F l u x
The migration flux can easily be derived once the diffusion flux is known.
This can be inferred by rearranging Eq. 1:
_ F d<b I
dC
-ID
C
= +D " RT dx IF
dx
n
tt
The left hand terms in this equation can be simplified by using Eq. 39:
UD^
/ " ( C ) + 4AV
3
_ I '
~2F
(C f + 4K _ I '4K -2{C T
SK -(C )\ ~2F _ %K -(C f
c
IF IF SK -(C f
IF
' dx
From which the migration flux can be obtained as follows:
2
F d<3> '4K
I -2{C f
-2D C
=
RT dx 2F _ %K -{C f
2
...40
[288]
X . C o m p u t a t i o n of the M i g r a t i o n / D i f f u s i o n R a t i o
Both the migration and diffusion fluxes can be obtained once the C values
c
are known (see Eq. 32). The rnigration/diffusion ratio, R(I), can be obtained by
dividing Eq. 40 by Eq. 39:
Migration Flux , ^ 4/sT -2(CJ
^.~T . = /?(/)=
Diffusion Flux
(C f + 4K
...41
As shown in Fig. 6, the diffusional flux is always larger than the migrational
flux. Migration becomes increasingly more important at large values of the
dimensionless parameter 0 (e.g. at large slimes thicknesses and/or current
densities).
[289]
1-r
(A)
legend
0.2- ,
C .2
mol/1
Cjtt_ ._2_
moj^l C
C, =
0 +
.8 mol/1
C , .2
a
C , g .2
10
12
0=
lx
mol/1
.8 mot^l
mol/1
T-
16
18
20
FD X
n
0.8
(B)
6
o
3
0.6
C
O
in
5
o
cn
0.4-
0.2-
legend
C, .8 mol/l C , .8 mol/l
C, = 1.5 mol/l C , = .1 mol/l
10
0=
lx
-t
12
16
18
20
F i g . 6 Effect of the variable 0 on the migration /diffusion ratio, R(I), for different bulk electrolyte
compositions.
[290]
A p p e n d i x 3 T i m e D o m a i n To F r e q u e n c y D o m a i n
T r a n s f o r m a t i o n : T h e F o u r i e r T r a n s f o r m In Current Step
E l e c t r o c h e m i c a l Techniques
I. I n t r o d u c t i o n
The theoretical foundations of the Laplace transformation are described
extensively in the literature [1-4]. Laplace plane domain techniques are widely
used in the analysis of electrical circuits (5-71. Particularly, in the area of system
analysis Laplace analysis is extremely valuable 181. The availability of personal
computers with high processing power has boosted the use of this technique in
almost every scientific field. Among other advantages, the power of the technique
resides in its ability to decompose any signal into its fundamental components.
Furthermore, the possibility of obteuxiing the transfer function of the system is a
strong driving force for the implementation of this technique. Electrochemists
have applied Laplace techniques to the study of a variety of electrochemical
systems [9-191.
In this appendix, the one-sided Laplace transformation (or Fourier transform)
will be used to analyze a simple electrical circuit. The circuit here chosen follows
the electrical analogue of the electrode/solution interface proposed by Gerischer
[201 and Grahame [21]. The response of this circuit to a pulse of current will be
analyzed both in the time and in the frequency domains. The transfer function of
the system will be obtained by Fourier transformation of the input and the output
signals. The differences between the values obtained by the numerical (using the
FFT algorithm ) and the analytical Fourier transformations will be stressed.
x
[291]
Analytical solution
For this simple analogue, the study of the decay in potential upon current
interruption provides accurate information on these parameters. However, as the
system becomes more complex, analysis of the time-domain transients is
extremely difficult and the amount of information that can be obtained is limited.
Very often information on rj is the only value sought [33-381 but sometimes kinetic
and mass transport data are also required. Usually a "trial and error" procedure
has to be conducted in order to find an electrical analogue that matches the
response of the experimental system. That is, the configuration of the electrical
circuit is not usually known and experimental data have to be provided to create
an analogue circuit. The use of the Fourier transformation is a strong aid in such
an analysis. More complex circuits can be analyzed following the same technique.
n
(a)
(b)
e(t)
E(S)
cS
i(t)
-<s>-
R,
i(S)
Ri
A. Analytical
solution
1 In Fig. 2 L, represents the time that has elapsed since the current step was first applied.
[292]
Analytical solution
W(r)]= rf(t)e-*'dt=F{S)
t>0
...1
Jo
where:
T->lfW] = F(S) = Laplace transform of the time-dependent function, JW.
S = new algebraic variable which is independent of time.
where:
I(S) = Laplace transform of the current. i(t)
E(S) = Laplace transform of the potential, e(t).
...3
Thus, given i(t) and the configuration of the electrical analogue, the response
of the system, e(t), is obtained by following the 4 steps previously described. In
the following section these steps will be performed one by one for each current
step waveform (cases A and B).
Case A :
[293]
Analytical solution
Fig. 2 Current step functions used to analyze the circuit described in Fig. 1
(a) Current step with zero decay and rise times
(b) Current step with zero decay but with a finite rise time.
Step 1:
The input function, i(t), can be described by the following relationship:
i(0 = /: [l(r)-l(r-U + ia-O]
1
-4
{0
for
for
t<0\
,>o}
c components is described by the
lt
following equations:
L[/?J
L[/?J
L[c]
...6
=R
'cS
...7
...8
[294]
Analytical solution
while the Laplace transform of the time dependent function, i(t) can be
expressed as follows:
/(S)=|[l-^ +^]
-.9
Step 2:
The Laplace domain representation of the electrical circuit shown in Fig. 1 .a
is shown in Fig. l.b. Applying nodal analysis the impedance of the system is
obtained:
...10
where:
...11
Z (S) =
t
...12
Z (S) =
2
l+R cS
2
thus,
Z
...13
l ; c5
rearranging:
R.+R.
Z(5)=
...14
p-
Step 3:
E(S) is obtained by multiplying Eq. 14 times Eq. 9:
E(S) = I(S)Z(S)=j[i- -'
e
.15
+ e" ]
bS
rearranging:
k -^
R
(S) = -
^[l-^\r'1+ ^[l-^
T
f'']
S
S
+
-IS
[295]
Analytical solution
Step 4;
The inverse Laplace transformation of Eq. 16 gives
>
[ _ -^.-, ],
i
...17
Eq. 17 is the pursued analytical solution for the input function shown in
Fig. 2.a.
Case B:
Step 1:
The function i(t) can be described by
i (0
- t ) - r (t - t )]}
b
...
18
for
[t
r <
0
n
t > 0
...19
and m is the slope of the rise time curve which is defined as:
k
m =
...20
tc-h
...21
Step 2:
The impedance of the system is independent of the input and output
functions and is given by Eq. 14.
Step 3:
E(S) can be described as follows:
[296]
R +R
x
'Si'-" -;
V*2
R~c
.22
Step 4:
Taking the inverse Laplace transform of Eq. 22, e(t) is obtained as follows:
e(f) = k R [e '-e
''m-/.)] +
...23
.A
K'-O-
+ m{R +RJR c
so
ft
ed
The response of the circuit shown in Fig. l.a to the input signals expressed
by Eqs. 4 and 18 is described by Eqs. 17 and 23 respectively. The time dependance
of the input and output signals is presented in Fig. 3 assuming the indicated
parameters. Analysis of the response of the system in specific windows of time
can provide the unknown values of some of the parameters present in the circuit.
For instance, considering the time interval between t (i.e. after the capacitor is
fully charged but before current is interrupted) and t (i.e. just before current is
applied back), Eqs. 24 to 26 are derived from Eqs. 17 and 23:
[297]
(a1)
0.006
0.005
0.15
CL
>
0.004
t = 0.000 sec
<
t= 0.583 sec
0.003 -,t= 1.000 sec
0
+->0.1
c
t - 1.600 sec
3 0.002 J?f 25.0 n
O
/V 10.00 n
c
(a2)
0.2
c = 3000. / i F
0.001 -
Q_
0.05 -
f'
\
0.2
0.4
0.6
0.8
1.2
1.4
1.6
Time, sec
(bi)
0.006
0.2
0.4
0.6
0.8
Time, sec
1.2
1.4
1.6
(b2)
0.005
(= 0.000 sec
CL
E 0.004
<
t= 0.108 sec
f= 0.115 sec
e
0.003
tf
Rr
Z3 0.002
0.130 sec
25.0 0
Rz= 10.0
0.001
c = 300. fjF
0.026
0.052
0.078
0.104
0.13
Time, sec
Time, sec
Fig. 3 Response of the circuit described in Fig. 1 to the input functions described in Fig. 2 assuming the indicated
system parameters Figs, (al) and (a.2) correspond to Eqs. 4 and 17 respectively. Figs, (b.l) and (b.2) comply with
Eqs. 18 and 23 respectively
A) At t = t':
B) at t =
e(t) = k (R +R )
...24
k^
...25
U:
steep potential decay =
C) between
and t
e(t) = k R e
l
.26
[298]
Knowing R2, c is computed using Eq. 26. Analysis of other windows of time should
provide with equivalent values for the resistors and the capacitor.
For this particular example, the computation of the value of the electrical
parameters was straightforward. On the other hand, when the number of analogue
elements increases and/or time-dependant components are involved, this
analysis is overwhelmingly difficult. Thus, the use of Fourier analysis is
encouraged.
The Fourier transform is defined as [39-41]:
Fr/(f)] = j
F(j<o)
.. .27
(V^-T). The
functions used in electrochemistry are only defined for times greater than zero,
therefore Eq. 27 can be expressed as
Fr/(0] = (~f(t)e- 'dt=F(j(0)
J
Jo
...28
l+R cj(Si
2
The variation of Z(jco) with frequency is sketched in Fig. 4.a. In this plot, the
real part of the impedance is the abscissa and the negative of the imaginary part
of the impedance is plotted as the ordinate. Analysis of this plot (so-called Argand
diagram) can provide the value of the elements of the analogue circuit. Plotting
of log (co) vs. I Z(jco) I (so-called Bode plot) complements the overall information
[299]
required to analyze the system. The Bode plot for this circuit is shown in Fig. 4.b.
Information on the values and characteristics of the electrical parameters likely
to be present, can be derived from these plots (28.42.431.
The value of the real part of the total impedance can be obtained as the low
frequency interception of the impedance curve in the Argand diagram. In addition,
at high frequencies the value of Rj is attained:
when
when
(0 = 0
Z(j(n)=R +R
l
CD > oo
Z(ja)=R
...30
...31
...32
+ CJ(H
Thus, at
co = co =^
0
Thus, to obtain
the value of c, I Z(Jco) I is computed at this point and the frequency at which it
takes place is obtained from the Bode plot (Fig. 4.b). Complex non-linear
regression analysis [44,45] and deconvolution techniques [42,43] can also be applied
in the analysis of impedance data.
At this point, it is worth noting the fact already stated in the literature [28,29]
that the impedance obtained by Fourier transformation of data generated in the
time domain is also in character an AC impedance. That is, the results obtained
from the study of a time-invariant system should be equivalent in both cases.
[300]
Numerical Solution
-z
(a)
mag
Rt
CHWv
*2
y W=0
FVR
IZ!
eal
(b)
log wQ
log w
ni. N u m e r i c a l S o l u t i o n
The relationship between the Laplace and Fourier transformations expressed
by Eq. 28, produce the following representation of the impedance in the frequency
domain:
ZC/co) = Em
/(/co)
...33
[301]
Numerical Solution
where:
fn(t) = value of the digitized time dependent signalfl.t)taken at the nth time interval.
N = total number of sampled points taken
At = sampling interval, sec
Assuming that fit) is only defined for t>0 and that it has a finite number of
points, the following relationship is derived from Eq. 34:
F = F(j<x>)= lf (t)e
n
=o
m=0,l,...,A/-l
...35
,
m =0,1,...N-l
17
...36
The period, T, of the function f(t) is given by the product NAt. For a single
period, Eq. 36 shows that coAt spans in values comprehended between 0 and 2n
rad. It also shows that the frequency representation of the data generated in the
time-domain is given in multiples of the fundamental frequency ay
[302]
Numerical Solution
Thus, even though coAt values as high as 2n can be obtained (Eq. 36) the
highest meaningful coAt value is n rad. In practice, this value is reduced even
further to avoid as much as possible the presence of aliasing.
By applying the definition of the DFT given in Eq. 35 to Eq. 33, the discrete
Fourier transformation of the impedance can be expressed as follows:
where the DFT of the different functions involved has been incorporated.
Thus, the problem reduces to find the Fourier coefficients F for the e(t) and
i(t) functions and taking their ratio. This transformation has to account for the
restrictions imposed by leakage and aliasing and the F
coefficients must be
equally spaced data points. A large amount of easy-to-use programs are available
in the literature to implement this algorithm [1,46-481. In this particular work, use
of a built-in FFT routine present in a commercial software package was employed.
When the numerical time-domain tofrequency-domaintransformation is to
be performed, important consideration should be given to the values that the
sampling interval, At, and the function period, T, can adopt. To assure that the
maximum spectral content is to be obtained, At has to be chosen as small as
possible (see Eq. 37). In addition to this, the extent to which information in the
lower frequency range can be attained is regulated by the total number of points
sampled, N (see Eq. 36). Thus, the use of very small At values and a large number
[303]
Numerical Solution
'rod'
...40
2Ai
At
0.00781
0.00391
0.00195
0.00098
0.00049
64
128
256
512
1024
WW
32
64
128
256
512
16
32
64
128
256
1 Use of filters can help in obtaining frequencies as high as the ones implied by Eq. 36.
[304]
Numerical Solution
Figs. 5 and 6 show the frequency spectrum obtained when different input
functions and system parameters are employed. By comparing the Argand
diagrams in these figures it can be seen that they are similar. This is to be expected
as the only electrical parameter that was changed was the capacitor value. The
decrease in this value is observed in the Bode plots as a displacement of the
frequency towards higher values. As can be observed, an order of magnitude
decrease in the capacitor value reflected in a tenfold frequency increase required
to define the frequency spectrum.
[305]
Numerical Solution
(a)
(b)
0.006
0.2
0.005
CL
0.004
<
0.15 -
t= 0.550 sec
t= 0.583 sec
t = 1.000 sec
t= 1.600 sec
Rf 25.0 n
/?^= 10.0 Q
c= 3000. yuF
0.003
>
'+->
0.1
C
CD
i_
0.002
0.001
0.55
0.76
0.97
1.18
1.39
CD
o
CL
0.05
0.55
1.6
0.76
1.39
1.6
Time, sec
Time, sec
*
0.97 1.18
(c)
35
Numerical Solution
34
Analytical Solution
(d)
33
32
cn
cd
31
30
N
l
29
28
27
X Analytical Solution
26
Numerical Solution
25
20
40
Zreal, Q
10'
Freq, rad/sec
Fig. 5 Frequency response of the circuit depicted in Fig. 1 obtained assuming the parameters shown in (a).
The input (a) and output (b) functions were calculated according to Eqs. 4 and 17 respectively. These functions
were digitized using N=2048, At= 5.127x10"* sec.
Plots (c) and (d) describe the analytical (Eq. 33) and numerical (Eq. 37) frequency spectra. For this set of data cop
5.98 rad/sec, 0)^=975 rad/sec (Eqs. 37 and 40 respectively)
[306]
Numerical Solution
(a)
0.006
0.005
tcf 0.100 sec
0.004
>
t= 0.101 sec
t= 0.110 sec
- 0.003
C
CD
k_
k_
0.002
13
-t-<
C
CD
t = 0.200 sec
Rf
25.0 fl
Rf
10.0 n
o
CL
c = 300. yuF
0.001
0.1
0.12
0.14
0.16 0.18
0.2
0.1
0.12
Time, sec
(c)
G
CO
20
A1 1
Analytical Solution
Numerical Solution
25
30
0.14
0.16 0.18
Time, sec
(d)
35
34
33
32
31
30
N 29
28
27
X Analytical Solution
26
c
35
40
Numerical Solution
25
10'
Freq, rad/sec
Zreal. Q
Fig. 6 Frequency response of the circuit depicted in Fig. 1 obtained assuming the parameters shown in (a).
The input (a) and output (b) functions were calculated according to Eqs. 4 and 17 respectively. These functions
were digitized using N=2048, At= 4.883xl0' sec.
Plots (c) and (d) describe the analytical (Eq. 33) and numerical (Eq. 37)frequencyspectra. For this set of data
C0(=62.83. rad/sec, co^10240 rad/sec (Eqs. 37 and 40 respectively)
3
[307]
Numerical Solution
Figs. 7 and 8 show the effect that a different waveform period and sampling
interval have on the definition of the frequency spectrum at lower frequencies.
Increasing the waveform period from 0.03 to 0.10 sec decreased the value of co^
but increased % due to the fact that At was increased. Better resolution of the
frequency spectrum at lower frequencies is reflected in a well defined Bode plot
shown in Fig. 8.d.
Figs. 5,7, and 8 were obtained using the same electrical analogue parameters
and only changes in the characteristics of the input function were studied. Plots
c and d in these figures show that regardless of the shape and magnitude of the
input and output functions, the transfer function remains identical provided the
same system is being analyzed.
The advantages of the Fourier transformation in the analysis of time-domain
transients can be visualized by inspecting plots c and d in Figs. 5 to 8. The
numerical transfer function obtained matches the one derived analytically. If the
electrical circuit analogue had not been not known in advance, it could be inferred
from analysis of this frequency spectrum. Different circuits could be proposed
and their frequency spectrum calculated so as to match the numerical frequency
response. The numerical frequency response could then be curve fitted to the
theoretical circuit leading to the derivation of the values of the electrical
parameters.
Plots c and d of Figs. 6 to 9 also show that the information that AC impedance
and current interruption can provide is equivalent. This equivalence is due to the
fact that "lumped" time-invariant electrical parameters were considered. The
presence of distributed and/or time dependant parameters may give different
information from each technique as different systems maybe analyzed. If systems
where these parameters are present are to be studied, careful experimental
planning is required. Also, the information provided by each technique may be
used jointly to better characterize the system.
[308]
Numerical
Solution
(a)
0.006
CL
t=
0.004 -
0.100
sec
0.101
sec
0.15
>
t = 0.108 sec
t= 0.115 sec
If 0.130 sec
Rf 25.0 n
Rf 10.0 fl
c = 300. fjf
<
CD
v_
0.002
0.1
_i_
0.105
0.11
0.115
0.125
0.12
C
CD
-t'
CL
0.13
0.11
0.115
0.12
0.13
Time, sec
Time, sec
(c)
X
0.05 -
(d)
Analytical Solution
Numerical Solution
G
G
d>
CO
E
N
40
20
z
Freq, rad/sec
real, Q
Fig. 7 Frequency response of the circuit depicted in Fig. 1 obtained assuming the parameters shown in (a).
The input (a) and output (b) functions were calculated according to Eqs. 18 and 23 respectively. These functions
were digitized using N=2048, At= 4.883xl0' sec.
Plots (c) and (d) describe the analytical (Eq. 33) and numerical (Eq. 37) frequency spectra. For this set of data cop
209.44 rad/sec, 0^=34133 rad/sec (Eqs. 37 and 40 respectively)
5
[309]
Numerical
Solution
(a)
0.006
0.005
t = 0.100 sec
\ = 0.101 sec
0.004
<
f=
t = 0.115 sec
-M" 0.003
CD
i_
0.108 sec
R,= 25.0 n
0.002
/?= 70.0 0
2
c = 300. /xF
0.001
0.1
0.12
0.14
0.16
0.18
0.1
0.2
0.11
0.13 0.14
0.15
0.16
0.17
0.18
Time, sec
Time, sec
0.12
(c)
(d)
35
Analytical Solution
34
Numerical Solution
33
32
4 -
G
CC
31
30
N
N
29
28
27
26
40
20
Analytical Solution
Numerical Solution
25
10"'
Freq, rad/sec
Zreal, Q
Fig. 8 Frequency response of the circuit depicted in Fig. 1 obtained assuming the parameters shown in (a).
The input (a) and output (b) functions were calculated according to Eqs. 18 and 23 respectively. These functions
were digitized using N=2048, At= 1.465xl0' sec.
Plots (c) and (d) describe the analytical (Eq. 33) and numerical (Eq. 37) frequency spectra. For this set of data cop
62.80 rad/sec, 0^=10240 rad/sec (Eqs. 37 and 40 respectively)
5
[310]
Summary
IV. S u m m a r y
Equations were derived to describe the time domain response to a current
step for a simple RC electrical circuit. These equations were used to generate
theoretical time-dependant discrete functions that were subject to DFT by using
the FFT algorithm The complex process of analyzing electrochemical data in the
time domain was simplified by analyzing it in the frequency domain. The practical
implementation of the FFT algorithm was discussed. The limitations and
application range of the time domain to frequency domain transformation were
also explored. It was stressed that provided the same system is analyzed, current
interruption and AC impedance can provide the same information.
[311]
A p p e n d i x 4 A n a l y t i c a l C h e m i s t r y of E l e c t r o l y t e
Containing PbSiF -H SiF
6
Solutions
I. Objectives:
A) To describe the different techniques that have been reported in the
literature to analyze solutions containing H SiF .
B) To explain how a modified technique for analyzing H SiF solutions was
developed.
2
solutions
HiF
<=> HSiFl + H
HSiFl
<=> SiFf
...1
+ H
pK =\.19
...2
1 Solutions with more than 13.3% H SiF become enriched with HF due to preferential evaporation of
SiF . Reciprocally, SiF solution enrichment takes place upon evaporation of solutions with less than
13.3% H SiF [7,10].
2
[312]
SiF;
+2H 0
2
SiF
SiF [H 0]~ +H 0
5
SiF [H OY " + F~
+ 3H 0
SiF
<=> H^iOs
...3
F~ +2H 0
...4
+ HF
...5
In very dilute solutions, SiF " is completely dissociated [20.22] .This has
lead to the use of fluorosilicates (i.e. NajSiFe) for the fluoridation of water supplies
[23,24]. Data on the physico-chemical properties of H SiF solutions are reported
extensively in the literature [18,22,24-30].
2
Si ~ [HJIFJ + lSiOJ
2 SiF
is an octahedral ion.
[313]
...6
.. .7
Reactions 6 and 7 indicate that the amount of alkali required in the second
equivalence point should be equal to twice the amount used for the neutralization
of the acid. Disagreement in these values indicates the presence of other species
in solution. The presence of HF and soluble SiO compounds in H SiF solutions
a
have accounted for these discrepancies. In addition to this, it has been found,
that the hydrolysis of S1F " is kinetically slow [201. To circumvent these
2
a suitable salt to precipitate the hexafluorosilicate ions. When this salt is chosen
to be a fluoride salt, SiO can be determined in the same titration. Thus, if NaF
a
.. .8
.. .9
hexafluorosilicate ions are present as Na SiF . If any Si0 were present, it would
2
HF +NaOH ->NaF +H 0
2
...10
1 This titration has to be carried out at low temperatures (near 0 *C) to avoid dissolution of sodium
hexafluorosilicate. The solubility of Na SiF is: 6.52 g/l at 17C and 24.54 g/l at 100*C .
2 Titration of the solution near its boiling point aids in the kinetics of the hydrolysis reaction.
2
[314]
Background
the second titration was larger than twice as much the amount spent in the first
titration, HF was present in the sample. A reciprocal relation will indicate that
Si0 was present in the solution.
Experimentally this two-step titration technique is difficult to perform. The
main difficulties in the implementation of the technique are the cooling and
heating operations which have to be performed carefully. Dilution of the samples
can be a source of error as some unaccounted fluorosilicate hydrolysis can occur.
Additionally, any alkali added in excess during the first titration will react with
the hexafluorosilicate ions. Jacobson [3] points out that in addition to the
temperature of the solution during the first titration being very low, the solution
has to be concentrated so that the reaction products of the first stage do not
2
pass into the second stage before the first stage reactions are completed.
Notwithstanding these difficulties, this technique provides accurate acid analysis
and its use is frequently reported in the literature 15,181.
3. Analysis of H SiF -PbSiF electrolytes
2
Analysis of the electrolyte used in lead refiriing has been carried out by
modifying the two-stage titration previously described [7,311. In the analysis of
H SiF -PbSiF electrolytes, Pb has to be removed before the acid is titrated. Use
of Na S0 to separate Pb from the electrolyte was discarded due to potential
co-precipitation of NaaSiFg with PbS0 [311. To avoid this loss of SiF ~ , known
amounts of H S 0 are added to the sample to remove Pb as PbS0 (see reaction
11). The precipitation of PbS0 releases an equivalent amount of H SiF that
can be subsequently titrated using the two-stage titration previously described.
2
+2
PbSiF^+H^O^ H^iF^PbSO^ i
...11
[315]
the acid against LiOH adding LiF and/or HF. H SiF solutions were titrated at
2
Case
Case
Case
Case
When H SiF solutions are titrated against LiOH, the following reactions
2
take place:
... 12
... 13
+H0
Li SiF is the most soluble of all the alkali fluosilicates HO] . Thus, upon
3
neutralization of the acid (reaction 12), all the SiF " remains in solution. As
2
these ions are already in solution, heating and dilution preceding hydrolysis
via reaction 13, are not required. On the other hand, the hydrolysis reaction
has to be carried out very slowly to circumvent its kinetic limitations. Fig. 1 .A
shows the titration curve obtained under these conditions . Two equivalence
4
points are present during the titration of the acid. The first equivalence point
takes place at a pH between 3.5 and 4 and corresponds to reaction 12. The
second equivalence point takes place at pH values between 7.7 and 8.4 and
corresponds to reaction 13. While [H SiF ] can be calculated from the total
2
amount of LiOH spent in the titration, [Si0 ] cannot be quantified using this
2
technique
6 >7
2 Acid composition 2.03 M H SiF and 0.40 M Si0 . All samples were titrated against 0.8657 M LiOH.
2
5 0.5 ml of the acid sample S1 was diluted to 20 ml with deionized water and titrated using 0.8657 M
LiOH.
6 The presence of Si0 in the sample will be indicated by the difference in alkali volumes spent during the
first and second equivalence points.
7 [H SiFe], [PbSiF ], [HF], [SiOJ = concentrations of the indicated species in mol/l, [M].
2
[316]
HF + LiOH -LiF + H 0
... 14
H SiF + ILiF
2
Li SiF + 2HF
2
... 15
1 2 ml of 1.127 M HF were added to 0.5 ml of the acid sample S1. Afterwards the solution was diluted to
20 ml using deionized water.
2 2 ml of 1.127 M HF were added to 0.5 ml of the acid sample S1. Afterwards the solution was diluted to
20 ml using deionized water. To this solution = 0.5 g of CP LiF was added.
[317]
(B)
(A)
,...
i... i...
3
i..
6
10
mM LiOH
11
0
i.
11.. 11 ,
7
ii.i
8
10
mM LiOH
(A)
Technique).
[318]
(D) Titration against LiOH when LiF is added previous to the titration.
To the original acid sample, LiF was added and the solution was titrated
against L i O H . The two equivalence points described by reactions 13 and 14
1
can be observed in Fig. I.D. As in the previous case, accurate analysis of the
solution can be obtained by analyzing the titration curve.
Comparison of the different analytical routines:
The results obtained by analyzing the same acid sample using the routines
previously described are shown in Table 1. Similar concentrations can be found
by either of these techniques. The technique described in
Case
D (LiF-LiOH
technique) proved to give accurate acid analysis and was used extensively in
this work. This technique was used to analyze H SiF within 1.5 g/1, SiO
2
solutions were
analyzed by this technique, Pb was removed prior to the acid analysis by using
Li S0 :
2
PbSiF +Li0
6
... 16
[HaSiFd. M
[Si0 l. M
Two-step
titration:
Cold-Hot titration
2.03
0.40
[HF],
Comments
Quantification of HF or SiO
present in the sample is
possible.
2.04
Case B:
HF added.
2.05
0.44
Quantification of HF or Si0
present in the sample is
possible.
Case C:
HF and LIF
added.
Case D:
LiF added.
2.02
0.40
Quantification of HF or Si0
present in the sample is
possible.
2.02
0.40
Quantification of HF or Si0
present in the sample is
possible.
1 0.5 ml of the acid sample S1 was diluted to 20 ml using deionized water. To this solution = 0.5 g of CP
LiF was added.
[319]
Analytical routine
electrolytes (IAF-LiOH
1. Analytical routine
Pipette a 2-ml aliquot of the sample into a plastic centrifuge tube \ add
approximately 10-ml of 0.2 M L i S 0 . Mix well, then centrifuge for 5 minutes.
Decant the supernatant liquor into a 50-ml beaker. Wash the solid residue in
the centrifuge twice with 2-3 ml of distilled water, and add the washings to the
solution In the beaker. Add ~ 0.5 g of powdered CP LIF. Stir well and titrate the
mixture with LiOH up to a pH of 3-4 [first equivalence point). Record This volume
as V\. Continue the titration adding LiOH at a very slow rate up to a pH of
6.8-7.3. Record this volume as V [second equivalence point).
2
PbSiF +LJ' S0
2
...16
+ 2H 0
2
...15
...9
->LiF+H 0
2
...14
2 Add enough Li S0 so as to neutralize all the lead salt present in the electrolyte (reaction 15).
2
[320]
Calculation ol results
Record the volume of LiOH spent as Vi and obtain the moles of LiOH required
in this titration: m = V * [LiOH].
x
1i
... 13
Record the volume of LiOH spent as V and obtain the moles of LiOH required
2
3. Calculation
of results
[H SiF^=-/^-[Pb]
2
"sample ^
[HF]=-^--2*[H SiF \
^sample
[SiO j =
The
[321]
A p p e n d i x 5 C o m p u t e r Interfacing of the W e n k i n g P o t e n t i o s t a t :
C a l i b r a t i o n of t h e R o u t i n e s used t o Interrupt t h e Current
I. Objectives:
A) To explain some of the details involved in the computer interfacing of the
Wenking potentiostat.
B) To describe the algorithm used to interrupt the current.
C) To demonstrate how by using "dummy" cell circuits the current
interruption routines were checked.
D) To analyze the experimental results obtained in the time domain in the
frequency domain by using the FFT algorithm.
A. Current interruption using the Wenking potentiostat
In Chapter 3 the various connections required to interface the computer to
the potentiostat were described (see Fig. 3.6). Current was halted by using a
mercury wetted relay activated by the computer. This relay short circuits the
battery that controls the amount of current that goes through the electrochemical
cell. To activate this relay, a DC voltage of at least 3.5 V is required. Voltages
smaller than 0.5 V deactivate the relay allowing current to flow back to the cell.
The easiest and fastest way of controlling the relay operations was found to be
through the use of the digital output of the DT2805 board . By setting the output
value of the digital bits to 1, the relay was activated and the current halted .
Accordingly, resetting this value to zero opened the short circuit imposed by the
relay, allowing the re-establishment of the current flow.
112
1 When a digital value of 1 ("high" bit status), is output a voltage of =2.5 volts is generated, and when its
value is zero ("low" bit statusltne voltage output value drops to =0.2 V. This signal goes back and forth in
the shape of a pulse.
2 To generate the required voltages to activate the relay three digital output bits were connected in series.
[322]
grounded with respect to the working electrode. This aided in preventing common
mode voltages. Additionally, common mode voltages were avoided by not making
direct connections between the counter electrode and the DT2805 board.
B. Algorithm used to interrupt the current.
Concurrent interruption of current at preset times under galvanostatic
conditions was implemented by the hardware connections shown in Fig. 3.6. In
addition, appropriate software programs were developed. The characteristics of
these programs varied according to the number of channels sampled, the length
of the experiment, and the extent and frequency of the current interruptions.
The computer programs were written so as to handle up to a maximum of 8
analogue input channels. Additionally, control of the digital bits used to generate
the voltages required to control the relay's functions was incorporated in these
programs. Thus, data acquisition and setting and resetting of digital bits were the
tasks considered in the program's algorithm. In this algorithm, the speed at which
these operations took place was the most important parameter accounted for.
The different operations performed by the computer were programmed by
using the "foreground/background" operation mode. In this mode, all the tasks
performed by the DT2805 board are executed in the "background" while the
computer is free to do other operations in the "foreground". The rate at which the
background operations run is set by fixing the task period. Additionally, the
frequency at which each task runs is controlled by fixing the task modulo.
Moreover, tasks may start synchronously but they may not be instantaneously
activated. The task phase controls how many task periods have to take place
before the task is first executed. Furthermore, completion of a specific task may
activate/deactivate other tasks. The state of each task may also be altered from
the "foreground" operating mode. Also, tasks become idle, once the number of
times the task has been executed equals a preset value .This preset value is defined
as the number of iterations.
Table 1 shows the task set-up used to control the DT2805 operations. Typical
values for the task modulo, task phase, and number of times the task is executed
are provided in this table. The order in which each task is defined is important
as tasks are executed sequentially rather than simultaneously. Every task period,
the status of each task is checked. If the task status is idle, the task is skipped
and the next task is executed.
[323]
Function
Description
Task
Task
Number of
Modulo
Phase
Iterations
Task Status
Set bits
Reset bits
H a l t s flow o f c u r r e n t to the
1111^1111llllill^^^^
R e - e s t a b l i s h e s flow o f c u r r e n t
msec
50
50
to t h e c e l l
3
Acquire data
T r a n s i e n t Recorder Mode".
Tasks simultaneously
activated by T a s k 4
100
These tasks
c o n t r o l the c u r r e n t
before, after, a n d d u r i n g
Interruption cycle
current interruption.
4
No operation
OO
"Foreground" activated
o n c e T a s k 5 buffer i s
b y s t a r t i n g T a s k s 1, 2, a n d 3 .
Acquire data
25000
25000
3900
A l w a y s active
d u r i n g n o r m a l galvanostatic
conditions. This task controls
the current i n t e r r u p t i o n
f r e q u e n c y a n d the e x p e r i m e n t
length. Every certain n u m b e r
of acquired points T a s k 4 is
activated.
As was previously described, all tasks run at the same rate. This rate is fixed
by the value of the task period which for most of the programs was set to 4 msec.
The frequency at which tasks are executed is fixed by the task modulo. Thus, for
example, Task 5 was used to acquire a set of data points every 100 sec (i.e. 25000
times 4 msec). These data were sequentially put in a buffer array. The size of this
array was selected to control the current interruption frequency. Thus, every time
the buffer array was filled, data were saved in the computer's hard drive, and
Task 4 was activated. Once Task 4 is activated, Tasks 1, 2, and 3 start
simultaneously. These three tasks are executed sequentially. However, Tasks 1
and 2 start only after a certain number of tasks periods (or clock-ticks) have gone
by, whereas Task 3 is executed immediately. Task 3 acquires data at the highest
acquisition rate set by the task period (in this case at 250 Hz). As Table 1 shows,
100 data points are collected at this acquisition rate (i.e. during 400 msec). After
9 data points have been collected, Task 1 sends a high bit and halts the flow of
[324]
current. As this task is executed first, there is a time lag between the current
interruption and the acquisition of the tenth set of data points. This time lag is
due to the fact that the computer has to check Task 2 status before executing
Task 3 (i.e. there is a software overload). This time lag was measured with an
oscilloscope and was of the order of 1 msec. The current interruption length is
controlled by Task 2. Task 2 phase is 50, thus, current was interrupted during
41 clock-ticks (i.e. during 164 msec). After this time, Task 2 sends a low bit and
re-establishes theflowof current. One niillisecond after the short circuit is opened,
data are recorded. Subsequently, samples are taken every 4 msec during a period
of 196 msec. Once Task 3 is completed, Task 4 is deactivated and the data are
saved in the computer's hard drive. As Task 5 is active all the time, the previous
process repeats as often as Task 5 buffer is filled. The data acquisition program
ends when task 5 becomes idle (i.e. after 3900 data points or 108.33 Hrs.).
During the data acquisition run, an interactive procedure for changing the
A / D gain was incorporated in the computer program \ Thus, the DT2805 board
measured potentials lower than 100 mV with an accuracy of 0.3 mV. Potentials
between 100 and 1000 mV were measured within 1 mV.
C . C a l i b r a t i o n of the
1 The DT2805 board performs analogue to digital (A/D) conversions at a 12-bit resolution. Four input
voltage ranges are available in the board: -20 to 20 mV, -100 to 100 mV, -1000 to 1000 mV, and -10000
to 10000 mV. These ranges can be interactively changed by software control improving the resolution of
the acquired data.
2 Resistors were measured independently by using a Tech-300 digital voltmeter. Capacitor values were
obtained by using AC impedance techniques (see Appendix 6).
[325]
(A)
(B)
Time, msec
Time, msec
[326]
component
mAmp
4.88
4.89
0.97
4.93
0.95
4.90
0.49
0.99
2.46
2.45
4.91
24.61
0.97
4.91
0.96
4.90
0.50
0.99
2.46
2.45
4.93
24.65
0.50
0.99
2.46
2.45
4.91
24.65
4.89
,0.20
4.83
0.50
0.99
2.47
4.93
24.63
0.49
0.98
2.45
24.61
Theoretical values
C , \lF
Ra, Q
10.3
15.1
10.3
0
0
2.3
3147
10.3
4.9
3147
15.1
10.5
3147
10.3
2.3
1117
10.3
4.9
1117
15.1
10.5
1117
15.1
10.5
225
10.3
10.3
2.3
4.9
11.7
11.7
15.1
10.5
11.7
15.1
10.7
7.8
Expenmental Values
C, uF
R, Q
R* n
x
10.2
14.9
10.0
10.2
10.1
10.2
149
15.0
15.1
15.1
15.1
15.1
10.0
10.1
9.8
1
1
1
1
1
6
^
1
1
1
iiii;4 i9iii
15.0
15.0
15.2
15.0
15.1
14.4
13.7
13.7
12.2
13.2
13.6
12.2
14.8
14.9
24.7
24.9
25.1
25.0
25.2
24.7
25.0
25.3
25.2
:
0.0
0.0
2.3
2.1
4.9
4.8
10.5
10.3
10.2
10.2
10.2
10.1
2.3
2.2
5.2
4.9
10.4
10.3
10.2
10.1
10.3
10.2
11.6
11.6
12.9
12.2
11.6
0.0
0.2
0.1
0.7
0.4
0.2
0.3
0.1
0.7
0.3
0.0
0.0
Regression
No. points
Coeff., r
Considered
2952
2863
3062
2879
3289
3207
3178
3140
3146
3140
1339
1022
938
913
1115
1053
1026
1017
992
1005
II?233!Ii!
203
177
133
151
159
***
**
***
***
***
***
***
***
***
***
***
***
0.984
0.991
0.995
0.993
0.998
0.998
0.999
0.999
0.999
0.999
0.969
0.993
0.989
0.993
0.992
0.994
0.994
0.997
0.996
0.995
0.985
0.990
0.996
1.000
0.998
0.999
0.842
0.647
0.647
0.001
0.800
0.707
0.997
0.998
0.006
-0.004
-0.004
-0.500
1
1
3
5
4
4
10
10
10
10
10
10
3
4
3
4
6
6
6
6
6
6
3
4
4
3
4
5
3
3
3
3
3
3
4
4
3
3
3
3
[327]
Table 2 shows the values of resistors and capacitors obtained for different
current levels. The theoretical and experimental values of the components of the
electrical circuit are shown in this table. The first two rows of data show the circuit
parameters obtained when a single resistor is used as a dummy cell. In all the
other cases, the circuit shown in Fig. A3.1.A was used. The number of points
upon current interruption used to fit the data to Eq. A3.26 is also presented in
Table 2. The parameters used in the current Interruption routine to obtain these
data sets were the same in all the cases fTable 1).
Table 2 shows that as long as the time constant of the electrical circuit is
1
larger than IQ msec, accurate values of ths. parameters of the circuit can be
obtained. These large time constants are present in the sets where the 3147 and
1117 |iF capacitors were used. Circuits in which smaller time constants were used
did not provide accurate parameter values. In these circuits, only the total
resistance value (R + Ra) can be obtained. In both cases, the amount of steady
x
state current going through the circuit does not seem to have an effect on the
calculation of the parameters of the circuit.
D. Time domain tofrequency domain transformation: experimental results
In Appendix 3 the fundamentals behind the time domain to frequency domain
transformation were presented. Here, FFT techniques will be used to Fourier
transform the time domain transients presented in Fig. 1.
The FFT algorithm requires the data to be given as sets of 2 (q being a positive
q
integer number) equally spaced data points. Thus, to satisfy these requirements,
the transient functions shown in Fig. 1 were modified. This was done by curve
fitting these functions using high order polynomials over different time intervals.
Once the transients were curve fitted, they were reconstructed using a uniform
sample interval equal to 1 msec. By doing this, the functions shown in Fig. 1 were
represented by 400 points rather than by the original 100 points. To increase the
range of the frequency spectrum and to have the required 2 data points, the
q
transient functions were extended. Thus, a total of 4096 points were Fourier
transformed. The impedance of the system was obtained by Fourier transforming
1 The time constant for the RC circuit shown in Fig. A3.1 .A can be expressed by: x(sec) = RTC
[328]
the output (potential transient) and the input (current transient) functions and
taking their ratio. The results of this transformation are shown in Fig. 2. The
theoretical impedance spectrum is also shown in this figure.
(B)
(A)
Z, , Q
r
Freq, rad/sec
al
Fig. 2 Frequency domain representation of the time domain data presented in Fig 1.
The analytical solution was computed using the known parameter values: R - 15.1 f2,R = 10.2 i>,c = 3147 uF.
The numerical solution was obtained by applying the FFT algorithm to the experimental data presented in Fig. 1,
from which the following values were obtained: R, = 15.3 2, R =9.9 2, c = 2864 uF.
t
Fig. 2 shows that the circuit parameters derived from the experimental data
are within 10% of the theoretical values. These values are not as good as the ones
obtained by time domain analysis. However, if the electrical analogue had not
been known in advance, Fourier transformation would have shown some of the
characteristics of the circuit. Thus, time domain and frequency domain analysis
can be used concurrently to study and verify experimental data. As the information
provided by each technique is given by only changing the frame of reference (time
<=> frequency), the data can be interpreted from different perspectives.
[329]
Appendix 6
Current Interruption a n d
M e a s u r e m e n t s u s i n g t h e S o l a r t r o n Devices
A C Impedance
I. Objectives:
A) To describe how the Solartron Electrochemical Interface (SEI) performs
the "sampled" current interruption procedure.
B) To describe the implementation of the AC impedance measurements using
the S E I and the Frequency Response Analyzer (FRA).
A . Current interruption using the SEI
The SEI offers three ways of compensating for the value of IRg *. Two of them
(Feedback and Real Part Correction) require the value of the uncompensated ohmic
resistance, Rg, to be known in advance. The tiiird of these procedures, the
"sampled" IRg drop compensation, determines Rg by using a current interruption
routine. Use of this routine involved the quantification of the potential and current
values during current interruption conditions. These values are different from the
ones observed under steady state conditions as the S E I interrupts the current
continuously at frequencies as high as 18.5 KHz. Thus, by executing the "sampled"
interruption routine, the following parameters are obtained:
a) Factual This is the value of the potential actually being applied to the cell
under current interruption conditions.
b) C
: This is the value of the current actually being applied to the cell
under current interruption conditions.
actual
C) PsamiPe+hoid> This is the value of the potential 5 usee before current is applied
back to the cell.
From these parameters the value of the uncompensated resistance is
obtained using the following equation:
_
(Factual
Psample + hold)
The current interruption procedure was tested by using the dummy cell
provided by Solartron . The routine was applied under potentiostatic and under
galvanostatic conditions. A 18.5 KHz current interruption frequency was found
2
[330]
to give the best results. The Rg values obtained from this technique were witiiin
5%. of the theoretical values. Table 1 shows some of the Rg values obtained under
potentiostatic conditions using the Solartron dummy cell.
Table
steady-state conditions
mV
Current
uAmp
200
500
Potential
p
A
actual
Experimental
value
sample+hold
^actual
mV
mV
uAmp
20.86
234.0
192.3
24.32
1714
52.04
593
483
61.67
1784
1 If the system is controlled under galvanostatic conditions, the voltage waveform input by the FRA is
converted to a current waveform by the SEI. This current signal is superimposed on the DC current being
applied to the cell. The amplitude of this superimposed signal is obtained by dividing the sinusoidal
voltage amplitude by the value of the resistor used in the SEI to measure current.
[331]
(A)
eooo
(B)
c,
Hh
Hh
OWi
7000
6000
- 5000
<g
M
I
4000
M
4000
3000
2000
1000
KMX)
2500
4000
5500
'
7000
8500
2500
10000
Z |, ft
rea
10
10'
Freq,
10*
10
10
10
rad/sec
[332]
(A)
(B)
F
r
e
q
, r
a
d
/
s
e
c
Fig. 2 shows the frequency spectrum for a dummy cell with only one RC
component. Only one hump is obtained in the Argand diagram. The R and Rj
values obtained were within 1.0% of the expected values. Capacitor values
measured by this technique were within 5% of the nominal value (i.e. the value
provided by the manufacturer) \
x
1 The capacitor measurements done by A C impedance were used as standards for calibrating the current
interruption routines presented in appendix 5.
[333]
A p p e n d i x 7 S o l u b i l i t y of P b S i F 4 H 0
6
I.Objective:
To obtain the solubility of PbSiF -4H 0 in water and in H S1F as a
function of temperature and acid composition. Solubility changes are computed
from data available in the literature.
6
EL S o l u b i l i t y of pure P b S i F 4 H 0 as a f u n c t i o n of temperature
6
2464 4418
... 1
where:
jt, = anhydrous mole fraction (number of moles of the i-th component divided by the
number of moles of the mixture).
T= Temperature. K
xM
2
ank
+ (l-x )M
2
2
l
where:
W = weight of the anhydrous component contained In 1 Kg of solution [Kg of
component/ Kg solution!.
Manh = molecular weight of the anhydrous component (For PbSiF Mmh is equal to 349.34
g/mol)
M, = molecular weight of the solvent ( for water M, = 18 g/mol)
mh
From the
follows:
where:
W)^ = weight of the hydrated component contained in 1 Kg of solution [Kg of component/
Kg solutionl.
Afhyj = molecular weight of the hydrated component (For PbSlF .4H 0
is equal to
421.34 g/mol)
Af, = molecular weight of the solvent
6
[334]
Once W and W are obtained, the solubility of the salt can be obtained
from any of the following relationships:
and
anh
IOOOWU
...4
AWl-W^)
...5
lOOOW^ p
...6
c, =
2
where:
rrx, = molality of the solution [mol/Kg of solvent!
c = molarity of the solution [mol/1]
P, = weight percent: Kg of anhydrous i-th component contained in 100 Kg of solution,
p = solution density, [g/cm ]
2
Table 1 Solubility changes as a function of the temperature for pure PbSiF 4H 0 solutions.
6
Temp *C
mj, mol/kg
Pi,%
c *, mol/1
2
-1.0472
0.0897
0.6567
3.81
5.47
65.67
4.57
20
-1.0015
0.0997
0.6824
4.65
6.15
68.24
4.75
25
-0.9842
0.1037
0.6919
5.04
6.43
69.19
4.81
30
-0.9649
0.1084
0.7024
5.54
6.76
70.24
4.89
40
-0.9208
0.1200
0.7258
7.02
7.58
72.58
5.05
50
-0.8702
0.1348
0.7515
9.68
8.66
75.15
5.23
57
-0.8314
0.1474
0.7704
13.13
9.61
77,04
5.36
[335]
P, [PbSiFJ
68.97
0.98
67.96
7.34
56.5
13.93
43.1
25.82
23.95
39.65
10.38
,%
[336]
n. A s s u m p t i o n s :
1) Dissolution of lead is the only Faradaic reaction and proceeds without
kinetic limitations (i.e. io-**> for Pb->Pb +2e").
+2
m. F u n d a m e n t a l E q u a t i o n s
Unsteady state diffusion can be described by Fick's second law
[1,2] :
1
where:
C = concentration of species i [in this case Pb ], [mol cm" ].
D = Pb* molecular diffusion coefficient [cm sec" ]
t = time, [sec]
x = distance from the electrode/solution interface, [cm]
+2
1 In this appendix the equations presented by Bushrod et al. [2] are re-derived and explained in greater
detail. Furthermore, a relationship between concentration overpotential and local concentrations (not
included in Bushrod's derivation) is obtained.
2 i.e. C=C* for t=0 at any point x.
[337]
Fundamental Equations
Bulk
Boundary Electrolyte
Nernst
Pure
Lead
Layer
Anode
A
x=0
I
'0
x=x
C* = C(x,0)
...[ii]
Eqs. [i] and [ii] can be expressed in a general form by substituting the variable
C for the variable ix.
...2
u* = u(x,0)
The general solution of Eq. 1 subject to the initial condition 2 can be obtained
from their Laplace transform.
The Laplace transform, L, of the left and right terms of Eq. 1, is as follows:
= Ddx
Li\=Su-u
I Bt
[338]
Fundamental Equations
dx
D ~D
. . . 5
y(x) = k exp[mx]
...6
+ k
ex
y(x) = *,
. . . 8
[339]
from which:
u(x,S) = M(S) e x r j - x - ^ +N(S) e x p [ x ^ | j + .
...11
IV. B o u n d a r y C o n d i t i o n s and A n a l y t i c a l S o l u t i o n
For semi-infinite linear diffusion C(,t) = C* and
M(,S)
= y . Consequently,
<HII
+ j
...12
Eq. 12 will be solved for three different cases (see Fig. 2):
A) Application of a galvanostatic step (t[ < t)
B) Current Interruption in the presence of concentration gradients in the
Nemst-Boundary layer (tl < t < t" )
2
[340]
t
c
Fig. 2 Current step function for
which the concentration and
overpotential changes were solved/
c
1_
L.
CJ
Time
1 , .
1
+
ti
...13
where:
U = current density, [Amp-m"].
Z= Charge number of Pb [eqmol M
2
+2
...14
[341]
.15
hVI] J
exp
u(x,S)=ZF^DS'
2
...17
L- jAp[^VS]| ^I^p[.
!
=2
At
-k erfc
k '
2>/r
with
k=
Vo
Thus,
"
( J :
'''
) =
...18.a
ZFW| V? {-4DJ
X
...18.6
From Eqs. 18.a and 18.b, the concentration overpotential can be obtained
from the following relationship:
T] (x,t)=
Lb
c
ln-^ u
J
...19
from which:
[342]
Current interruption
Jt
-A
exp
...20.a
liDh
...20.6
B . C u r r e n t interruption
After applying current for t[ sec, a concentration profile that follows Eq. 18.a
is present. Thus, Eq. 12 is to be solved for the current interruption case using
Eq. 18.a as initial condition and Eq. 21 as boundary condition:
du
dx (0,0 = 0
...21
|(0,5) = 0
u\x,t )
x
.22
exp
[4oJ .
..23
' "V?
(0 5)=
M(S)-
1 1
ZF^D^DS
=0
...24
..25
M(S) = -
ZF^DS
Finally, incorporating Eqs. 25 and 18.a in Eq. 22, obteuxiing the inverse
Laplace transform and rearranging:
[343]
x (
u(x,t ) =
r
ZFyfD
erfc
-erfc
+u
..26.a
...26.6
ZFynD
The the concentration overpotential can be obtained from Eq. 26.b:
...27
electrode
After time tl current is applied back (see Fig. 1). The initial concentration
profile follows Eq. 26.a, and the boundary condition is given by Eq. 28:
dU .
du
...28
Hfi
-
...29
'l
.30
5
M(S)=
ZF^IDS'
...32
Finally, incorporating Eqs. 32 and 26.a in Eq. 29, obtaining the inverse
Laplace transform and rearranging:
[345]
(A)
.
87a
ess
N
o
C
C
O
.843
u
S
.838
e
o
N
+
.8
14
t, , ,
. % ae
ie.a
aa.a
Tine*
3a.a
sees
4a.a
se . e
(B)
aa
a.
i.
ea
i.
aa
808
**
0.
.
O
C
e
S
. 480
01
o
o
<J
A n n
IBB
ia.a
ae. a
T i
e ,
3a. a
sec
4 e . a
1_J
se. a
Fig. 3 Changes In u(0,t) and ii (0,t) upon application of the current step described in
Fig. 2.
c
C'=C(x,0)=[Pb l =0.8 moll" , i =i =50 Ampm , >=5xlO cm sec" . /,'= 16 sec, t\ = 32 sec.
+2
Bulk
[346]
488.
688.
Time, sec
1111111111111111111111111111111111111
888
488.
688.
Time, sec
Fig. 4 Changes In r\ {0,t) upon application of the current step described in Fig. 2.
{C*=C(x,0)=[Pb l =0.8 moll },
D=5xlO c m s e c
c
+a
Bulk
Fig.
A
C
D
ij, AmpnV
50
200
200
200
i , AmpnV
100
100
200
200
t[ =, sec
16
16
100
200
t' =, sec
2
32
32
800
600
[347]
Foreword:
The abstracted tables presented in chapter 6 are shown in this appendix
in their complete form. The names of the tables in this appendix were chosen to
correspond to those presented in chapter 6. Furthermore, the regression
parameters associated with the data shown in these tables are also presented
in this appendix alongside the main tables.
The regression parameters shown in this appendix were obtained from
analysis \ The square of the multiple regression coefficient, r , and the
ANOVA
square of the residual errors, l y l , were obtained from such analysis. These
r
\y NY-Y'\
in which:
y = Experimental data
y = Regressed data
y = parameter related to the variations vector , y with y = y + y .
2
the AC impedance data, are given with respect to the real, Z*, the imaginary,-Zj,
and the absolute impedances, IZI, (| Z |= ^Z^ + Z^
[348]
Table 3 Summary of the values of the electrical analogue parameters obtained under rest potential conditions
Frequency Range
Experiment Addition
and Sweep Agents
Number
Pot/Gal
Control
rad/sec
rad/sec
Qcm
Qcm
pFcm'
Ci c m sec
2
Amp/m
CA2-/
yes
Gal
560
99588
1.42
0.501
64
23.71
0.64
CA4-7
CA5-i
yes
yes
Pot
Pot
315
560
250159
111740
1.02
1.19
1.855
'"'0725'''''
lllPilf
12.45
23.90
IIKl!
CA5-4
CA6-2
yes
yes
Gal
Gal
560
177
111740
157834
1.21
1.20
0.854
1.916
CC1-5
CC1-6
no
no
Gal
Gal
628
628
15783
44482
0.90
0.89
CC2-J
0,01-2
no
no
Gal
Gal
560
560
17709
28067
iiliii
65.30
liii!!! llliill!!!
"liS""""
Illlll iiiiioiii
-
269
0.64
72
186
0.56
0.55
157
70
0.57
0.45
2.88
i i i i i ! lllslli
0.85
1.88
0.59
iy i .
r
Q cm
2
iy i .
iy ,
Q W
CA2-7
4.99X10
0.972
2.43X10
0.977
4.75X10"
0.971
CA4-7
CA5-7
2.92xia'
1.05x10'
0.995
0.976
4.74X10
6.53xl0'
0.986
0.967
2.45X10"
9.81xia
0.996
0.976
CA5-4
CA6-2
2.09X10'
3.10x10'
0.969
0.995
1.23X10
5.06X10
!lll!lii!!ll
0.992
1.96x1a
5.59X10
0.970
0.996
CC1-5
CC1-6
5.09X10"
1.02xia
0.876
0.970
2.19x10*
3.11x10*
0.966
0.987
3.51x10^
1.03xl0-
0.967
0.982
CC2-Z
CC2-2
8.00X1O'
1.48X10-
0.944
0.913
6.39X10"
0.975
0.984
1.16X10"
1.60X10
0.971
0.966
4.75x10
-4
[349]
Table 4
Slope R ,
Qcm
m
Intercept,
i,(mV)
|y | ,mv
2
p ,Qcm K,
m
mV
Experimental
TIA.
mV
mV
2 54
12.1
42.9
0.62
41.9
0.80
30.8
0.850
6.8
7.78
23.1
61.2
1.65
60.4
1.19
38.1
0.983
5.1
7.37
1.61
28.1
733
45.8
0.986
1.63
6.48
L45
6.4
73.9
223
41.8
95.5
54.4
0.988
6.93
12.6
1.30
96.1
2.15
3.10
48.6
109.3
0.992
6.72
1.26
2.50
109.7
3.72
8.8
61.5
133.5
0.952
6.70
1.20
135.2
3.16
73.7
86.0
4.72
76.5
150.4
0.981
7.38
151.8
59.8
0 98
3.93
5.33
88.8
170.0
0.961
7.68
171.2
163.7
0.93
4.57
82.4
5.95
194.8
193.3
0.92
7.98
101.7
0.920
375.1
5.23
6.56
221.8
224.8
7.95
110.9
1.03
5.71
0.905
633.5
7.18
113.9
262.4
265.2
9.27
140.3
7.22
0.89
0.888
1192.5
7.79
124.9
3.74
96.5
436.0
4386.0
5.90
8.41
4.96
361.0
0.051
520.0
182.7
5688.7
1.97
10.86
9 40
360.6
0.558
8.65
543 3
Between 18 and 20 experimental points were used to obtain the regression coefficient. These points were
collected during -55 min and correspond to digitized samples taken when the frequency of the applied AC waves
was between 0.063 and 6.3 rad/sec.
iiiiiii
llill&iiif
llliiili
llliiili:
[350]
Table 5 . A Analysis of the spikes produced during the application of the AC waveform.
in the presence of a net DC current (Exp. CA6)
Parameters Derived from Regression
Analysis Computations
Slimes
Thickness
mm
Slope
Qcm
0.70
0.98
1.25
1.47
2.11
2.39
2.66
3.06
5.37
5.58
5.79
6.00
6.22
6.43
6.64
6.85
7.06
7.28
7.49
0.36
0.48
0.67
0.56
1.00
1.21
1.15
1.52
1.89
2.87
2.16
2.08
1.95
2.53
2.70
2.76
3.36
3.98
3.13
Intercept,
b, (mV)
24.9
28.4
29.6
35.7
0.95
0.99
0.98
0.98
0.99
0.92
0.97
0.99
0.85
0.88
0.97
0.96
0.92
0.82
0.88
0.92
From Eq. 4
ly l ,
mV
p , Qcm
m
TD
nc mV
'Km
mV
Experimental
mV
0.17
5.14
3 56
7.0
32.0
32.6
0.03
37.7
3.05
4.90
9.3
39.0
0.11
2.28
5.36
13.0
42.6
44.2
0.04
3.28
3.83
10.9
46.7
47.9
0.14
1.96
19.5
4.75
57.8
59.5
!!tl3lii!l
40.0
1.80
1.69
5.08
23.6
63.6
65.7
iii^iiiii
0.71
209
4.31
22.3
68.8
70.5
47.0
0.40
1.59
4.98
29.6
76.6
78.3
15.64
134.1
2 55
3.52
130.5
36.8
81.4
5.14
15.37
137.2
1.46
55.8
138.3
98.6
3.74
143.7
4.33
2.35
42.0
140.6
105.7
2.61
3.47
146.2
149.7
6.85
40.5
113.0
10.92
2.98
37.9
150.8
154.7
3.13
3.94
49.2
106.1
42.73
2.16
155.3
160.0
illiliili
166.1
168.2
10.16
2.16
4.06
52.5
120.3
2.24
7.49
4.02
53.6
173.9
176.3
24.54
182
118.6
65.3
183.8
184.5
4.75
112.3
83.22
77.4
0.66
5.47
1.45
189.6
193.7
140.9
0.97
60.9
4.09
2.31
4.18
201.9
205.2
4 experimental points were used to obtain the regression coefficient. These points were collected during -12
min and correspond to digitized samples taken when the frequency of the applied AC waves was between 0.63 and
6.3 rad/sec.
llllPi!!!!
Table 5.B Analysis of the spikes produced during the application of the AC waveform.
under current interruption conditions (Exp. CA6)
From Eq. 4
Sloped,
Qcm
2
Intercept,
b, (mV)
|y | ,mv
2
p . Qcm
m
IRm, mV
IR^+ft, mV
Experimental
*u,
mV
0.63
,.,,,,,2.23,,,,,,.
60.8
0.89
21.95
2.82
0.0
60.8
82.1
48.4
3 62
2.44
0.91
48.4
49.4
41.19
QQ
3.83
3.83
44.5
0.95
62.43
484
0.0
44.5
44.0
154.01
9.08
2.93
31.5
0.79
3.70
0.0
30.0
31.5
14.33
2.59
24.0
0.64
208.29
3.27
0.0
24.0
21.7
30.14
41.96
1.92
6.7
0.88
2.42
0.0
6.7
6.7
43.07
1.64
5.3
0 87
22.09
2.07
0.0
5.3
5.1
22.74
1.24
54.97
0.98
3.3
0.37
0.0
3.3
2.9
65.17
1.00
2.3
066
21.53
1.26
0.0
2.3
1.4
91.44
-6.2
-6.2
1.15
0.95
7.60
1.45
0.0
-6.1
102.00
0 90
-0.4
0.70
17.22
1.14
0.0
-0.4
0.3
0.2
iiiiiiiitt
0.99
0.69
112.8
21.63
U5
0.0
0.3
9 experimental points were used to obtain the regression coefficient These points were collected during -27
min and correspond to digitized samples taken when the frequency of the applied AC waves was between 0.63 and
6.3 rad/sec.
[351]
Table 6 Analysis of the spikes produced during the application of the AC waveform,
in the presence of a net DC current (Exp. C C 1)
Parameters Derived from Regression
Analysis Computations
Slimes
Thickness
mm
Slope R Intercept,
Qcm
b, (mV)
From Eq. 4
ly | ,
mV
p , Qcm
IRm.
mV
Experimental
IRm+fc,
mV
mV
1.37
0.52
32.4
0.98
0.1
3.21
3.80
10.1
42.6
42.8
1.79
0.63
38.0
0.97
0.3
3.11
3.51
12.2
50.2
50.6
2.22
0.74
42.8
0.98
0.4
2.97
3.34
14.4
57.2
57.8
2.65
0.81
48.0
0.95
1.1
3.06
3.05
15.7
63.7
64.4
3.07
0.92
52.4
0.93
1.7
2.94
2.99
17.8
70.3
70.9
3.50
0.96
58.4
0.93
2.3
3.12
2.74
18.7
77.1
78.1
5 experimental points were used to obtain the regression coefficient These points were collected during - 1 2
min and correspond to digitized samples taken when the frequency of the applied AC waves was between 0.63 and
6.3 rad/sec.
[352]
Table 7 Parameters derived from the fitting of the impedance data obtained
in Exp. CA2 to the Zz^-Z^c
analogue (Circuit A.2, Figs. 6.38.39)
High Frequency Parameters
R-A,tottl
R +R
t
Slimes
Thickness
bi,
^zci
Qcm
sec
Ci cm
\ ,sec
D,
^zcz-
Qcm
sec
sec
1R
b,
2
,sec
il cm s e c
Qcm
mm
0.80
1.61
2.23
3.10
3.72
4.72
5.33
0.21
0.31
0.42
0.64
0.52
1.10
1.06
0.73
0.67
0.64
0.59
0.59
0.55
0.60
0.0062
0.0165
0.0296
0.0683
0.0503
0.1511
0.1045
5.95
1.54
0.51
0.1866
6.56
7.18
7.79
4.70
5.13
4.30
0.56
0.55
0.59
33.50
18.76
14.09
9.39
10.34
7.25
10.19
0.0010
0.0023
0.0042
0.0102
0.0064
0.0321
0.0225
0.44
0.0373
llliill
0.5415
0.3323
0.5671 iiiiieiiii 0.3593
3.45
0.88
1.17
1.62
2.22
2.50
3.17
1.29
2.08
4.65
0.77
0.69
0.69
0.69
0.68
0.70
0.71
0.079
0.21
0.30
0.39
0.42
0.59
0.69
!ll
llliill
0.98 llliiili
5.55
4.17
3.84
4.11
5.30
4.26
4.60
0.037
0.104
0.176
0.260
0.276
0.464
0.592
4.06
Illlll
lislli!
0.82
1.07
2.05
0.65
1.19
1.58
2.26
2.74
3.59
4.23
4.99
5.99
7.21
1.992
9.04
2.27
8.95
0.2030
2.484
11.11
All measurements refer to the geometrical area of the electrode.
Low and high frequency terms refer to the ranges of frequencies used during the deconvolution process.
'* Low and high frequency arcs were fitted to the whole frequency range (0.063<co<30000 rad/sec)
0.3873
1.94
0.96
0.90
|y|,QW
|y|,QW
0.80
0.0067
0.999
0.0661
0.777
-0.0594
1.009
1.61
2.23
0.0147
0.0242
0.999
0.999
0.0185
0.0136
0.978
0.991
-0.0038
0.0106
1.000
1.000
3.10
3.72
0.0174
0.0574
1.000
0.999
0.0107
0.0143
0.997
0.997
0.0067
0.0431
1.000
1.000
4.72
5.33
0.0328
0.1136
llllifiiollill!
1.000
0.0095
0.0449
0.999
0.997
0.0233
0.0688
1.000
1.000
5.95
6.56
0.4281
0.0563
0.999
1.000
0.1344
0.0083
0.993
1.000
0.2936
0.0480
0.999
1.000
7.18
0.1286
0.0054
1.000
0.3369
lillllHIII
0.1232
7.79
i!i!iH!i!!l
0.3324
1.000
1.000
0.0045
1.000
|y|,oW
2
A total of 6 parameters were curve fitted using between 110 and 113 experimental points.
For the impedance curves obtained at 0.8 and 1.6 mm slimes, r >0.99 when 1<CK30000 rad/sec
2
[353]
r
Qcm
r
Qcm
cm"
^ZCa
Qcm sec"
2
^B.total
Qcm
zc
mm
0.80
1.61
3.10
3.72
4.72
5.33
5.95
6.56
7.18
7.79
0.43
0.84
1.15
1.64
2.03
2.81
, , , , , , , ,
^ , , , , , , ,
4.36
5.29
0.23
0.026
11.48
0.38
0.068
8.06
0.49
0.102
"""7^"
0.66
0.146
0.73
0.178
8.84
0.80
IIIIIIITIIIII
0.90
0.398
lllilllllilllll
9.88
0.749
10.48
2.40
0.975
3.82
12.61
For the AC sweep done at 0.8 mm slimes: B = 1.92 Clem sec' , and 4
2
zc
0.58
0.66
1.22
0.53
0.56
0.57
0.59
0.60
0.61
0.61
0.62
0.64
2.30
2.76
3.61
4.25
5.00
5.96
720
9.12
ZCB
fitted to circuit B . l , all the other sweeps were fitted to circuit B.2)
(ZI
llliiili
|y | .
Q W
0.998
0.999
0.998
0.80
0.063
28068
0.012
1.61
2.23
3.10
3.72
4.72
5.33
0.063
0.063
0.063
0.063
0.031
0.057
0.063
0.063
28068
28068
28068
28068
28068
28068
5.95
6.56
7.18
7.79
0.063
0.063
28068
28068
0.063
0.063
28068
28068
0.181
0.226
0.241
0.253
0.015
0.028
0.077
0.124
|y | ,nW
0.059
|y | ,QW
0.791
-0.047
1.007
0.013
0.008
IllilPiil.
0.018
0.049
0.999
0.999
1.000
1.000
0.007
0.009
0.008
0.019
1.000
1.000
1.000
0.999
0.019
0.028
0.998
0.998
0.998
0.998
0.058
0.096
1.000
1.000
illiRil!
0.999
0.038
0.054
0.998
0.998
0.143
0.171
0.999
1.000
1.000
1.000
0.072
0.086
0.999
0.999
0.169
0.167
1.000
LOOO
0.995
[354]
Slimes
Ri,
Thickness
Qcm
^ZCl
sec
bi,
,sec
D ,
R,
Qcm sec
Qcm
sec
b,
2
,sec RA.total
O. cm sec
1
Qcm
mm
0.64
1.20
1.76
1.89
0.101
0.017
0.069
0.817
0.91
0.68
0.83
0.58
0.0013
0.0115
0.1972
0.0599
75.04
1.47
0.35
13.64
0.0007
0.0015
0.1423
0.0079
0.32
0.83
1.61
0.99
0.68
0.59
0.53
0.81
0.039
0.075
0.174
0.146
8.19
11.10
9.26
6.77
0.0088
0.0121
0.0373
0.0934
0.42
0.85
1.68
1.81
|y|,QW
ly,l .QW
Slimes
Thickness,
mm
|y i ,Q cm
0.64
0.005
0.996
0.001
0.990
0.005
0.996
1.20
1.76
0.013
0.045
0.997
0.995
0.002
0.002
0.994
0.998
0.012
0.043
0.997
0.995
1.89
0.062
0 998
0005
0 997
0 057
0 998
[355]
Table 10 Electrical analogue parameters derived from the fitting of the impedance data
obtained in Exp. CA5 (Circuit B.2)
Slimes
Thickness
mm
Qcm
0.64
1.20
1.76
1.89
0.38
0.72
1.44
1.50
Qcm
F cm"
0.04
0.08
0.16
0.32
0.017
0.036
0.063
0.062
^ Z C A
Qcm sec'
2
Qcm
Vzc
16.10
16.10
0.42
0.80
1.60
1.82
0.65
0.64
0.57
0.62
iiiiiiiiiiiiiiii
tzi
|y | ,
Q cm
0.002
0.999
6.28
6.28
6.28
0.63
22295
fiiiii
"22295'""
28068
|y l .nW
0.001
0.978
0.997
0.999
0.991
|y | ,QW
0.001
1.000
1.000
0.998
0.999
0.002
0.019
0.056
1.000
0.998
0.998
0.001
0.001
0.019
0.001
0.018
0.037
[356]
Rc. Qcm
0.17
0.64
12.03
21.32
58.58
0.044
0.112
0.070
0.063
0.063
0.066
0.049
0.044
0.046
0.048
Cdj.
72.08
lillililllll
76.82
78.06
86.90
nF cm'
23.75
167.42
206.32
247.86
267.78
284.79
illllll9l)ililil;si
272.15
276.70
274.37
B L
il cm sec
^fzc
3.83
3.63
3.13
2.69
2.62
0.19
0.33
0.31
0.32
0.32
0.32
0.30
0.30
0.30
0.30
iiiiiitiiHiiiiii
2.25
2.23
2.23
Time, Hrs
|y | ,Q cm
0.17
0.64
12.03
21.32
58.58
67.57
72.08
76.82
78.06
86.90
0.28
0.11
0.06
0.04
0.07
0.07
0.07
0.05
0.05
0.05
0.98
0.95
0.97
0.97
0.95
0.94
0.93
0.95
0.95
0.95
|y | ,QW
0.01
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
0.00
Il!;ii:i9iiiiill
0.99
0.99
Illli9l!f;!IIII
0.99
0.99
0.99
0.99
0.99
0.99
|y | ,"W
r
0.24
0.08
0.04
0.03
0.05
0.05
0.05
0.04
0.04
0.04
0.98
0.97
0.98
0.98
0.96
0.96
0.95
0.97
0.97
[357]
Time, Hrs
Ret, Qcm
063
2.44
0010
0.018
804.67
1350.32
1.11
0.85
0.21
0.24
4.26
6.07
0.030
0.041
1502.48
1625.51
0 83
0.82
0 26
0.27
7.88
9.70
0.046
0.046
1707.71
1719.34
082
0.83
0.28
0.28
11.67
17.58
0.021
0.024
1588.19
1579.10
0.71
0.71
0 25
0.26
23.45
30.27
0.009
0.023
1431.30
1452.57
0.66
0.69
025
0.26
34.59
0023
1308.61
0.70
0.25
Cd,, nF cm'
Bi,
ZC
llllllllllllllllll
ly l .
IIIBII tfcm
Time, Hrs
IZI
|y | ,QW
|y | ,nW
0.63
177.1
44485
0.026
0.958
0.002
0.956
0.012
0.979
2.44
4.26
177.1
177.1
35335
31492
0.023
0.014
0.913
0.936
0.000
0.000
0.983
0.987
0.016
0.009
0.937
0.955
6.07
7.88
177.1
177.1
31492
31492
0.014
0.017
IIollllll:
0.913
0.000
0.000
ilfJilll
0.985
0.010
0.011
0.951
0.939
9.70
11.67
177.1
17.7
31492
31493
0.015
0.013
0.922
0.979
0.000
0.001
0.986
0.990
0.010
0.010
0.946
0.985
17.58
23.45
17.7
11.2
31493
31493
0.013
0.019
0.980
0.971
0.000
0.001
0.992
0.989
0.009
0.015
0.985
0.978
30.27
34.59
17.7
17.7
31493
39647
0.012
0.015
0.978
0.976
0.001
0.001
0.990
0.989
0.010
0.012
0.983
0.981
[358]
Table 13 Electrical analogue parameters derived from the fitting of the AC impedance data
obtained in Exp. CA4 to the Z a ^ - Z a ^ c analogue (Circuit A . 2 , Fig. 6.47)
High Frequency Parameters
R, + R
\ ,sec
D,
TImt.
Ri.
Hra
Qcm
208
3.89
0 87
0.43
0 73
0.83
00092
0.0031
94.77
139.84
0.00167
0.00092
1.77
1.96
043
0.35
0 89
0.94
198
2.08
0 77
0.85
2.64
2.38
5.70
7.52
0 35
0.33
0 82
0.80
0.0026
0.0026
137.18
126.81
0.00071
0.00059
165
1.44
0 39
0.43
0.86
0.82
1.91
1.75
0 69
0.63
2.00
1.77
9.33
11.18
0.32
0.34
079
0.73
0.0029
0.0042
108.92
79.97
000062
0.00057
1.26
0.99
047
0.55
0.77
0.55
165
1.79
0.56
0.34
1.58
1.33
15.37
19.55
0.33
0.32
071
0.72
00047
0.0043
71.53
75.11
0.00054
0.00052
0.93
0.99
0.59
0.59
0.53
0.58
1.76
1.70
0.34
0.40
1.27
1.31
zci
Di,
bi,
V2C,
sec
CI cm
sec
Ra,
TCI-
Qcm
sec
b,
i i cm
b,
0
,sec
sec
The A C sweeps obtained at current interruption times longer than 11 hrs were fitted to the Z
analogue (circuit A.l). Thus, 2 additional parameters were obtained:
Time, Hrs
ZARC
RAJOISI,
Qcm
rZ ARC2-CPE
Z
^ZCo
Q. cm sec *'
2
11.18
15.37
0.050
0.054
0.74
0.73
19.55
0.046
0.79
[359]
A p p e n d i x 1 0 K r a m e r s - K r o n i g Transformations of NUD E l e m e n t s
I. Objective:
To obtain the Kramers-Kronig transformation of the non-uniform diffusion
element (NUD) used to fit the experimental data presented in Chapter 6.
n. F u n d a m e n t a l E q u a t i o n s
The Kramers-Kronig transformations (KKT) have been proposed as a way of
validating experimental data obtained from AC impedance measurements [l]. They
imply that causality, linearity, and stability were observed during the experimental
measurement. Basically, KKT indicate that there is a unique relationship between
the real (Z*) and the imaginary (-2y parts of the impedance. Thus, for a sufficiently
wide impedance spectrum, the real part of the impedance can be obtained from
the imaginary part and viceversa.
One of the mathematical forms of the KKT is as follows [2,3]:
(2(0^ f-rZaCO-Z^G))"
Z (G)) = 3
7C J J o
dx
...1
X -(>
2
where the total impedance as a function of the frequency, Z(co), is given by the
following relationship:
Z(to)=Z ((o)-yZ (co)
9l
...2
Eq. 1 indicates that if Zj, is known in the whole frequency range, -Zg can be
obtained. Thus, if the derived -Zg values match those experimentally measured,
the data follows the KKT.
To evaluate Eq. 1 at least two problems (one experimental and another
mathematical) have to be overcome:
1) Impedance data are not usually available in the whole frequency range
(0<uKo).
2) Eq. 1 has singular points at x=co.
Several experimental procedures and mathematical arrangements have been
used to evaluate the KKT [2,3]. The approach followed here is to evaluate Eq. 1 by
assuming that the NUD element fits the experimental data in the whole frequency
range:
[360]
Fundamental Equations
NUD
((0)
= B.O'co)
tanhfl (yco)
.. .3
y2C
Vac
2C
The steps followed to obtain the KKT of the NUD element were as follows:
1) From the e x p e r i m e n t a l d a t a obtained in Exp. CC
^zc values (e.g. for the sweep taken at 1.79 mm, B = 1.03 2c/w sec ,
B =0.91 sec"^, and ^ = 0 . 3 3 , Table 14, Chapter 6).
r
Vzc
1 i.e.
Z^-Z^co)
-co'
0
0
= ll,
[361]
Fundamental Equations
(A)
xE
.238
23a
eee
xE
288
isa
<s
t
N
I
108
,038
. 888
. eee
40
88.8
F r a q ,
128.
288
168
r j i d / s a c
Fig. 1 Kramers-Kronig Transforms of the impedance data fitted to the NUD element.
From the experimental data presented in Table 6.14 (sweep taken at 1.79 rnm.Bp 1.03 Q cm sec' , B=0.91 sec' ,
and zc=0.33).
(A) Comparison of the theoretical imaginary impedance component, -Z (co), with -Z^co) obtained by KKT of the
real component. (B) Detail of Fig. A at low frequencies.
2
Vzc
yzc
[362]
Appendix
(D-Si-H^O System
Input File:
TITLE
Eh vs pH diagram for the (F)-Si-H20 system
Activities of all components = 1
DIAGRAM
AXES
<-> (AQ)
H<+>(AQ)
H2 0
LIGANDS
H2SI03
HSI03<1->(AQ)
SI03<2-> (AQ)
SI H4 (G)
AREAS
SIF6<2->(AQ)
SI F4 [H20]2 (AQ)
SIF5[H20] <1->(AQ)
F<->(AQ)
HF (G)
HF (AQ)
HF2 <1-> (AQ)
F20(G)
LINES
02(G)
H2(G)
UNITS
JOU
FILE
cpdaiber
gibalber
cpdjandat
cpdnpldat
cpdnbsdal
cpdsgtdat
ACTIVITY
1.
Aa
COMPONENTS
1,1
AXES
ABSCISSA PH
H <+> (AQ)
ORDINATE EH
<-> (AQ)
LIMITS
abscissa 4.9.
ORDINATE-1.0 2.5
TEMPERATURE
298.15
PRESSURE
1.
AREAS
gibbs
check
all
OVERLAY
H2(G) = 2<->(AQ) + 2H<+>(AQ)
OVERLAY
2 H20 = 02 (G) + 4 <-> (AQ) + 4 H <+> (AQ)
1.,1.,1.,1.
PLOT
VISUAL
mark lower abscissa 12
-3.,-2.,-1.,0.,1.,2.,3.,4.,5.,6.,7.,8.
MARK LOWER ORDINATE 6
-.5,0.,.5,1., 1.5,2.0
label lower abscissa 14
4.,-4
-3.,-3
-2.,-2
0.0
1.0 1
2.0 2
3.0 3
4.04
5.0 5
6.0 6
7.0 7
8.0 8
9.0 9
LABEL LOWER ORDINATE 8
-1.-1
-.5-0.5
0. 0
0. 5 0.5
1. 1
1.51.5
2. 2
2.5 2.5
TITLE LOWER ABSCISSA
PH
TITLE LOWER ORDINATE
Eh
TITLE UPPER ABSCISSA
SYSTEM (F)-Si-H20 at 25 C
LEGEND
linetype dashed 0.5 0.5
OVERLAY
linetype full
IDENT
linetype dashed .3.7.3.7
LIGAND
linetype full
ORDINATE EH
MULTIPLIER
1.0
SPECIES <-> (AQ)
ABSCISSA PH
MULTIPLIER
1.0
SPECIES H<+>(AQ)
CURRENT PRESSURES IN ATM
1.0000
TEMPERATURE OF SYSTEM 298.15 K
G VS T COEFFICIENTS FOR SPECIES JOU
SPECIES
GA
GB
<->(AQ)
0.901164E-12 0.000000
H<+>(AQ)
-16.9655 0.000000
H2 0
-237190. 0.000000
H2 SI 03
-.102300E+07 0.000000
HSI03<1->(AQ) -955460. 0.000000
SI03<2-> (AQ) -887000. 0.000000
SI H4 (G)
-39310.0 0.000000
SI F6 <2-> (AQ)
-.213800E+07 0.000000
SI F4 [H20]2 (AQ) -.200200E+07 0.000000
SIF5[H20] <1->(A -.207300E+07 0.000000
F<->(AQ)
-278820. 0.000000
H F (G)
-273220. 0.000000
HF (AQ)
-296850. 0.000000
HF2<1->(AQ)
-578150. 0.000000
F2 O (G)
-4600.00 0.000000
02(G)
0.183206E-11 0.000000
H2(G)
0.180604E-11 0.000000
TOTAL LIGAND AREAS 2
TOTAL MAIN AREAS 7
A SI F6 <2-> (AQ)
B SI F4 [H20]2 (AQ)
C SIF5[H20] <1->(AQ)
D F<->(AQ)
E F2 0(G)
F H F (AQ)
G HF2<1-> (AQ)
Output File
INPUT WILL BE TAKEN FROM F:\PB07_0.INP
OUTPUT WILL BE WRITTEN TO F:\PB07_0.OUT
SPECIES 17 ELEMENTS 5 E H O Si F
UNITS CHANGED TO JOU
MAIN SPECIES PER AREA 1
LIGAND SPECIES PER AREA 1
[363]
(Pb-F)-Si-H Q System
2
(Pb-FJ-Si-HaO System
Input File
TITLE
Eh vs pH diagram for the (PW>SI-H20 system
Activities of all components = 1
DIAGRAM
AXES
<-> (AQ)
H <+> (AQ)
H2 0
LIGANDS
H2SI03
HSI03<1->(AQ)
SI 03 <2-> (AQ)
SIH4 (G)
AREAS
SIF6<2->(AQ)
SIF4[H20]2(AQ)
SIF5[H20] <1->(AQ)
F<->(AQ)
HF(G)
HF (AQ)
HF2<1-> (AQ)
F20(G)
Pb
Pb<2+>(AQ)
PbO
Pb2 03
PB3 04
PB 02
PB02H2
PB02H<1->(AQ)
PB<4+>
PB F2 (S)
PBSI03
LINES
02(G)
H2(G)
UNITS
JOU
FILE
cpdalber
gibalber
cpdjandat
cpdnpldat
cpdnbsdat
cpdsgtdat
ACTIVITY
1.
ALL
COMPONENTS
2,1
AXES
ABSCISSA PH
H<+>(AQ)
ORDINATE EH
<-> (AQ)
LIMITS
abscissa -4.9.
ORDINATE-1.0 2.5
TEMPERATURE
298.15
PRESSURE
1.
AREAS
gibbs
check
all
OVERLAY
H2 (G) = 2 <-> (AQ) + 2 H <+> (AQ)
1..1..1.
OVERLAY
2 H20 = 02 (G) + 4 <-> (AQ) + 4 H <+> (AQ)
1.,1.,1.,1.
PLOT
VISUAL
mark lower abscissa 12
-3.,-2.,-1.,0.,1.,2.,3.,4.,5.,6.,7.,8.
MARK LOWER ORDINATE 6
-.5,0.,.5,1.,1.5,2.0
label lower abscissa 14
4.,4
-3.,-3
-2.,-2
0.0
1.0 1
2.0 2
3.0 3
4.04
5.05
6.06
7.0 7
8.08
9.09
LABEL LOWER ORDINATE 8
-1.-1
-.5-0.5
0. 0
0. 5 0.5
1. 1
1.51.5
2. 2
2.5 2.5
TITLE LOWER ABSCISSA
pH
TrTLE LOWER ORDINATE
Eh
TITLE UPPER ABSCISSA
SYSTEM (Pb-F)-SI-H20 at 25 C
LEGEND
linetype dashed .2 .2
OVERLAY
linetype DASHED .5.5
IDENT
linetype DASHED .2 .4
LIGAND
linetype full
Output File
[364]
(Sb-F)-Si-H 0 System
2
(Sb-D-Si-HaO System
Input Pile:
TITLE
Eh vs pH diagram for the (Sb-F^SWfiO system
Activities of afl Sb soluble components = 1e-6
TITLE
DIAGRAM
AXES
<1-> (AQ)
H <+> (AQ)
H2 0
LIGANDS
H2SI03
HSI03<1->(AQ)
SI03<2-> (AQ)
SIH4 (G)
AREAS
Sb
Sb0<+>
Sb02<1->
Sb4 06(c)
So2 04(c)
Sb2 05(c)
SbH3(g)
HSb02(aq)
Sb03H3(aq)
Sb04H4<1->
Sb F3 (c)
SbOF(aq)
Sb02H2F(aq)
Sb06H6<1->(aq)
Si F6 <2-> (aq)
SiF4[H20]2(AQ)
SIF5[H20] <1->(AQ)
F<1->(AQ)
HF(G)
HF (AQ)
HF2<1-> (AQ)
F20(G)
LINES
02(G)
H2(G)
UNITS
JOU
Gibbs
SbO<+>
-175640.
Sb02<1->
-339740.
Sb4 06 (C)
-1264400.
Sb2 04(c)
-795800.
So2 05(c)
-829300.
SbH3(g)
147740.
HSb02(aq)
407500.
Sb03H3(aq)
-644800.
Sb04H4<1->
-813200.
SbF3(c)
-836000.
SbOF(aq)
487400.
Sb02H2F(aq)
724700.
Sb06H6<1->(aq)
514093.
FIE
cpdalber
cpdjandat
cpdnpldat
cpdnbsdat
cpdsgtdat
ACTIVITY
1.1.1.
1.1.1.1.
1.10-610-61.1.1.1.1.1.1a-e 1.1.1.1a-e 1.1.1.1.1.
1.1.1.
1. 1.
ALL
COMPONENTS
2,1
AXES
ABSCISSA PH
H <+> (AQ)
ORDINATE EH
<1-> (AQ)
LIMITS
abscissa 0.4.
ORDINATE-0.51.
TEMPERATURE
298.15
PRESSURE
1.
AREAS
check
all
OVERLAY
H2(G) = 2<1->(AQ) + 2H<+>(AQ)
1..1..1.
OVERLAY
2 H20 = 02 (G) + 4 <1-> (AQ) + 4 H <+> (AQ)
1.,1.,1.,1.
PLOT
VISUAL 16.5.16.5.2..1.
mark lower abscissa 7
0.5,1.,1.5,2.,2.5,3.,3.5
MARK LOWER ORDINATE 5
-0.25,0.,0.25,0.5,0.75
label lower abscissa 9
0.0
0.5.5
1.0 1
1.5 1.5
Z02
2.52.5
3.03
3.5 3.5
4.04.0
LABEL LOWER ORDINATE 7
-.5-0.5
-0.25 -0.25
0. 0
0.25 0.25
0.5 0.5
0. 75 0.75
1. 1.
TITLE LOWER ABSCISSA
pH
TITLE LOWER ORDINATE
Eh
TITLE UPPER ABSCISSA
SYSTEM (Sb-F)-SI-H20 at 25 C
LEGEND
linetype dashed .2 .2
OVERLAY
linetype DASHED .5.5
IDENT
linetype DASHED .2 .4
LIGAND
linetype full
Output Pile:
[365]
(Bi-F)-Si-H 0
2
(Bi-D-Si-HaO System
Input File:
TITLE
EhvspH diagram for the (BI-F)-SWtiO system
Activities of Bi soluble components = 1e4
DIAGRAM
AXES
<> (AQ)
H<+>(AQ)
H20
LIGANDS
H2SI03
HSI03<1->(AQ)
SI 03 <2-> (AQ)
SIH4 (G)
AREAS
Bi
Bi <+3> (aq)
Bi 0 H <+2>
Bi6 012H12<+6>(aq)
Bi9 O20H20<+7>(aq)
Bid 021 H21 <+>(aq)
Bi9 022 H22 <+5> (aq)
Bi2 03
Bi205
Bi204
Bi407
Bi 0 <+>
Bi H3 (g)
SIF6<2->(AQ)
SIF4[H20]2(AQ)
SIF5IH20] <1->(AQ)
F<->(AQ)
HF(Q)
HF (AQ)
HF2<1-> (AQ)
F20(G)
LINES
02(G)
H2(G)
UNITS
JOU
Gibbs
Bi <+3> (aq)
+91820.
Bi 0 H <+2>
-136300.
Bi6 012H12<+>(aq)
-2293300.
Bi9O20H20<+7>(aq)
-3894400.
Bi9 021 H21 <+> (aq)
4113300.
BI9022H22 <+5>(aq)
4335600.
Bi2 03
493700.
Bi205
-382945.
Bi2 04
455838
Bi4 07
-973360
Bi 0 <+>
144446.
Bi H3(G)
231431.
FIE
cpdalber
cpdjandat
cpdnpldat
cpdnbsdat
System
ACTIVITY
1. 1.1.
1.1.1.1.
1.1e41e41e41e41e41e41.1.1.1.1e41.1.1.1.
1.1.1.1.1.
1.1.
Eh vs pH diagram for the (Bi-F)-Si-H20 system
==
1e4
Activities of all Bi soluble components
ACTIVITIES OF SPECIES
COMPONENTS
<->(AQ)
1.00000
2,1
H<+>(AQ)
1.00000
AXES
H20
1.00000
ABSCISSA PH
H2SI03
1.00000
H<+>(AQ)
HSI03<1->(AQ) 1.00000
ORDINATE EH
SI03<2-> (AO) 1.00000
<-> (AQ)
SIH4(G)
1.00000
Bi
1.00000
LIMITS
Bi<+3>(aq)
0.100000E-03
abscissa 0.4.
BiOH<+2>
0.100000E-03
ORDINATE -0.51.
Bi6 012H12<+>(a 0.100000E-03
Bi9 020 H20 <+7> (a 0.100000E-03
Bi9 021 H21 <+> (a 0.100000E-03
TEMPERATURE
Bi9 022 H22 <+5> (a 0.100000E-03
298.15
Bi203
1.00000
PRESSURE
Bi205
1.00000
1.
Bi204
1.00000
AREAS
Bi4 07
1.00000
check
BiO<+>
0.100000E-03
all
BiH3(g)
1.00000
SI F6 <2 > (AQ) 1.00000
OVERLAY
SI F4 [H20]2 (AQ) 1.00000
H2(G) = 2<->(AQ) + 2H<+>(AQ)
SIF5[H20] <1->(A 1.00000
1..1..1.
F<->(AQ)
1.00000
OVERLAY
HF (G)
1.00000
2 H20 = 02 (G) + 4 <-> (AQ) + 4 H <+> (AQ)
HF (AQ)
1.00000
1.,1.,1.,1.
HF2<1-> (AQ) 1.00000
F2 0(G)
1.00000
PLOT
02(G)
1.00000
VISUAL 16.5,16.5,2.,1.
H2(G)
1.00000
mark lower abscissa 7
MAIN SPECIES PER AREA 2
0.5,1.,1.5,Z,2.5,3.,3.5
LIGAND SPECIES PER AREA 1
MARK LOWER ORDINATE 5
-0.25,0.,0.25,0.5,0.75
TEMPERATURE OF SYSTEM 298.15 K
label lower abscissa 9
0.0
TOTAL LIGAND AREAS 2
0.5.5
1.0 1
TOTAL MAIN AREAS 6
1.5 1.5
A Bi + SIF6<2->(AQ)
2.0 2
B Bi + SIF5[H20J <1->(AQ)
2.5 2.5
C Bi <+3> (aq) + SI F5 [H20] <1 -> (AQ)
3.0 3
D BiO<+> + SIF6<2->(AQ)
3.5 3.5
E BiO<+> + SIF5[H20] <1->(AQ)
4.04.0
FBi + HF (AQ)
LABEL LOWER ORDINATE 7
Output File:
-.5-0.5
-0.25 -0.25
0. 0
0.25 0.25
0.5 0.5
0. 75 0.75
1. 1.
[366]
(As-F)-Si-H 0 System
2
( A 8 - F ) - S l - H a OS y s t e m
Input File:
TITLE
EhvspH diagram for the (AS-F>SI-H20 system
Activities of all As soluble components = 1M
DIAGRAM
AXES
<->(AQ)
H <+>(AQ)
H2 0
LIGANDS
H2SI03
HSI03<1->(AQ)
SI03<2-> (AQ)
SI H4 (G)
AREAS
As
As4 06
As 0 <+> (aq)
HAs 02
HAs04<-2>
H2As03<->
H2As04<->
H3As03<->
As205
AsH3(g)
SI F6 <2-> (AQ)
SI F4 [H20]2 (AQ)
SIF5[H20] <1->(AQ)
F<->(AQ)
HF(G)
HF (AQ)
HF2<1-> (AQ)
F20(G)
LINES
02(G)
H2(G)
UNITS
JOU
Gibbs
As406
-1151000.
As 0 <+> (aq)
-163800.
HAs 02
-402700.
HAs04<-2>
-707100.
H2 As 03 <->
-587200.
H2As04<->
-753300.
H3As03<->
-639900.
AsH3(g)
68900.
As205
-781400.
FILE
cpdalber
gibalber
cpdjandat
cpdnpldat
cpdnbsdat
cpdsgtdat
ACTIVITY
1.1.1.
1.1.1.1.
1.1.1*41.18419418-4 le-41.1.1.1.1.1.1.1.
1.1.
COMPONENTS
2.1
AXES
ABSCISSA PH
H<+>(AQ)
ORDINATE EH
<->(AQ)
LIMITS
abscissa 0.4.
ORDINATE-0.51.
TEMPERATURE
298.15
PRESSURE
1.
AREAS
check
all
OVERLAY
H2 (G) = 2 <-> (AQ) + 2 H <+> (AQ)
1..1..1.
OVERLAY
2 H20 = 02 (G) + 4 <-> (AQ) + 4 H <+> (AQ)
1..1..1..1.
PLOT
VISUAL 16.5,16.5,2.,1.
mark lower abscissa 7
0.5,1.,1.5,2.,2.5,3.,3.5
MARK LOWER ORDINATE 5
-0.25,0.,0.25,0.5,0.75
label lower abscissa 9
0.0
0.5.5
1.0 1
1.5 1.5
2.02
2.5 2.5
3.03
3.5 3.5
4.04.0
LABEL LOWER ORDINATE 7
-.5-0.5
-0.25 -0.25
0. 0
0.25 0.25
0.5 0.5
0. 75 0.75
1. 1.
TITLE LOWER ABSCISSA
PH
TITLE LOWER ORDINATE
Eh
TITLE UPPER ABSCISSA
SYSTEM (AS-FJ-SI-H20 at 25 C
LEGEND
linetype dashed .2 .2
OVERLAY
linetype DASHED .5.5
IDENT
linetype DASHED .2 .4
LIGAND
linetype full
Output File:
[367]