INFORMATION TO USERS
While the most advanced technology has been used to photograph and reproduce this manuscript, the quality of the reproduction is heavily dependent upon the quality of the material submitted. For example:
• Manuscript pages may have indistinct print. In such cases, the best available copy has been filmed.
• Manuscripts may not always be complete. In such cases, a note will indicate that it is not possible to obtain missing pages.
• Copyrighted material may have been removed from the manuscript. In such cases, a note will indicate the deletion.
Oversize materials (e.g., maps, drawings, and charts) are photographed by sectioning the original, beginning at the upper left-hand corner and continuing from left to right in equal sections with small overlaps. Each oversize page is also filmed as one exposure and is available, for an additional charge, as a standard 35mm slide or as a 17”x 23” black and white photographic print.
Most photographs reproduce acceptably on positive microfilm or microfiche but lack the clarity on xerographic copies made from the microfilm. For an additional charge, 35mm slides of 6”x 9” black and white photographic prints are available for any photographs or illustrations that cannot be reproduced satisfactorily by xerography. 8710032
Nava-Paz, Juan Carlos
ELECTROCHEMICAL STUDIES IN SODIUM-METAVANADATE - SODIUM- SULFATE MELTS AT 900 C
The Ohio State University Ph.D. 1987
University Microfilms
International300 N. Zeeb Road, Ann Arbor, Ml 48106
Copyright 1987 by Nava-Paz, Juan Carlos All Rights Reserved PLEASE NOTE:
In all cases this material has been filmed in the best possible way from the available copy. Problems encountered with this document have been identified here with a check mark V .
1. Glossy photographs or p a g e s ______
2. Colored illustrations, paper or print ______
3. Photographs with dark background
4. Illustrations are poor copy ______
5. Pages with black marks, not original co p y ______
6. Print shows through as there is text on both sides of page ______
7. Indistinct, broken or small print on several pages _
8. Print exceeds margin requirem ents______
9. Tightly bound copy with print lost in sp in e ______
10. Computer printout pages with indistinct print ______
11. Page(s) ______lacking when material received, and not available from school or author.
12. Page(s) seem to be missing in numbering only as text follows.
13. Two pages numbered . Text follows.
14. Curling and wrinkled pages
15. Dissertation contains pages with print at a slant, filmed as received
16. Other
University Microfilms International ELECTROCHEMICAL STUDIES IN NaVO^-NaJSO. MELTS AT 900 C d 2 4
DISSERTATION
Presented in Partial Fulfillment of the Requirements
for the Degree of Doctor of Philosophy
in the Graduate School of
The Ohio State University
By
Juan Carlos Nava-Paz, B.S., M.S.
*****
The Ohio State University
1987
Dissertation Committee: Approved by
R. A. Rapp
G. R. St.Pierre Adviser Department of Metallurical B. E. Wilde Engineering ©1987
JUAN CARLOS NAVA-PAZ
All Rights Reserved To my family, for their love, encouragement, and inspiration.
ii ACKNOWLEDGMENTS
The author expresses his sincere appreciation to
Dr. Robert A. Rapp for his support and guidance during
the course of this research, for many interesting and
stimulating discussions and for his assistance in "the
preparation of this manuscript.
A word of thanks is indeed in order to CEPET of
Venezuela for its invaluable support.
Aknowledgment is granted to my friends and colleagues
Yie Shing Hwang, Dianne Shi, Ravi Vilupanur, Yun Shu
Zhang and Chong Ook Park for their encouragement and moral support. VITA
September 26, 19b9 ...... Born - Maracaibo.Venezuela
1982 ...... B. S. , Universidad del Zulia Maracaibo, Venezuela
1983 ...... Scholarship Awarded by CEPET of Venezuela
198b ...... M.S. , The Ohio State University, Columbus, Ohio
FIELDS OF STUDY
Ma.ior Ptield: Metallurgical Engineering
Corrosion Drs. W. Johnson, S.Smialowska and R. Rapp
Electroohemi stry Dr. T. Kuwana
Chemical Metallurgy Drs. G. St.Pierre, W. Johnson, R. Rapp, J . Hirth
Physical Metallurgy Drs. P. Shewmon, G. Powell, G. Meyrick, W. Clark
Mechanical Metallurgy Drs. J. Hirth, R. Wagoner, R. Hoagland
Electron Microscopy Dr. W. Clark
Solidification Dr. Carroll Mobley
iv TABLE OF CONTENTS
ACKNOWLEDGMENTS ...... iii
VITA ...... iv
LIST OF TABLES ...... vi
LIST OF FIGURES ...... ix
CHARTER page
I . INTRODUCTION ...... 1
I I. ELECTROANALYTICAL TECHNIQUES ...... 12
Cyclic Voltammetry ...... 12 Chronopotentiometry ...... 14 Chronoamperometry ...... 18 AC Impedance ...... 20
111. EXPERIMENTAL PROCEDURE ...... 29
IV. RESULTS AND DISCUSSION ...... 37
Electrochemical Studies in Relatively Basic NaVOg-NagSO^ Solutions...... 37 Reaction Mechanism ...... 79 Electrochemical Studies in NaVO„-Na^SO. Solutions Under 0 ^ Gas ...... 94 Reaction Mechanism ...... 129 Electrochemical Studies in NaVO,-,-Na„S0. Solutions Under 0.1 % SOg-Og Gas ...... 137 Reaction Mechanism ...... 181 AC Impedance Results on The PtPainted WE 194 AC Impedance Results on The PtFoil WE .. 201 Discussion of AC Impedance Results ..... 208
CONCLUSIONS 213
LIST OF REFERENCES 215
v LIST OF TABLES
Table Page
1 . F'eak Potential As a Function Of Scan Rate For
a Pure Pt Foil WE Immersed In a NaVO^-Na^SO^
Solution Of Basicity -9.77 At 900 C ...... 41
2. Anodic To Cathodic Peak Current Ratio For
Various Scan Rates For a Pure Pt Foil WE
Immersed In a NaVO^-Na^SO^ Solution Of
Basicity -9.77 At 900 C ...... 41
3. Peak Potential As a Function Of Scan Rate For
a Pure Pt Foil WE Immersed In a Na^SO^ Melt
Of Basicity -6.66 At 900 C ...... 57
4. Chronopotentiometric Data For a Pure Pt Foil
WE Immersed In a NaVO^-Na^SO^ Solution Of
Basicity -9.77 At 900 C ...... 57
5. Variation Of Peak Potential With Scan Rate For
Cyclic Voltammograms On a Pure Pt Foil WE
Immersed In a NaV0o-Na,.SO. Solutions Of 3 2 4 Basicity -11.70 At 900 C ...... 97
6. Thermodynamic Data At 900 C ...... 82
7. Variation Of Anodic To Cathodic Peak Current
Ratio With Scan Rate For Cyclic Voltammograms
On a Pure Pt Foil WE Immersed In a NaVO,- o v i NaoS0/l Solution Of Basicity -11.70 At 900 C 2 4 8. Criteria For Determining The Reaction
Mechanism From Cyclic Voltammetry ...... 101
9. Variation Of Peak Potential With Scan Rate
For Cyclic Voltammograms On a Pure Pt Foil WE
Immersed in a Na^SO^ Melt At 900 C Under 0^ 121
10. Chronopotentiometric Data On a Pt Foil WE
Immersed In a NaVO^-Na^SO^ Solution
Of Basicity -11.70 At 900 C ...... 121
11. Variation Of Peak Potential With Scan Rate
For Cyclic Voltammograms On a Pure Pt Foil
Immersed In a NaVu,-Na,.S0. Solution At 900 C 3 2 4 Under Uncatalyzed 0.1% SOg-Or, Atmosphere ... 139
12. Variation Of Anodic To Cathodic Peak Current
Ratio With Scan Rate For Cyclic Voltammograms
On a Pure Pt Foil WE Immersed In a NaVO^-
NarS0. Solutions Of Basicity -11.72 At 900 C 139 4 13. Chronopotentiometric Data From a Pure Foil WE
Immersed In a NaVO,--Nar,SO. Solution Of 3 2 4 Basicity -11.72 At 900 C ...... 149
14. Variation Of Peak Potential With Scan Rate For
a Pure Pt Foil WE Immersed in a NaV0o-Nar.S0. 3 2 4 Solution of Basicity -13.15 At 900 C ...... 149
15. Chronopotentiometric Data From a Pure Pt
Foil WE Immersed In a NaVO^-Na^SO^ Solution
Of Basicity -13.15 At 900 C ...... 170 vii 16. Variation Of Peak Potential With Scan Rate For
a Pure Pt Foil WE Immersed In a NaVO^-NagSO^
Solution Of Basicity -13.85 At 900 C ...... 170
17. Chronopotentiometric Data From A Pure Pt
Foil WE Immersed In a NaVO^-NagSO^ Solution
Of Basicity -13.85 At 900 C ...... 176
18. Ohmic Resistance From AC Impedance
Measurements ...... 211
vi ii LIST OF FIGURES
Figure page
1. Cyclic Voltammetry ...... 13
2. Chronopotentiometry ...... lb
3. Chronopotentiometric Diagnostic Plot ... 17
4. Chronoamperometry ...... 19
b. AC Impedance Representation For a Purely
Activation Controlled Interface ...... 22
6. AC impedance Representation B'or a Mixed
Controlled Interface ...... 24
7. Bode Plot B'or An Activation Controlled
Interface Reaction ...... 26
8. Bode Plots For A System With Two Rate
Determining Steps ...... 26
9. Randles Plot ...... 28
1U. ZR vs ZI/W Plot For an Activation
Controlled Interface Equivalent To The
Circuit In Fig. b a ...... 28
11. Experimental Setup For Electrochemical
Studies At 900 C ...... 33
12. Apparatus Arrangement For Cyclic Voltam
metry and Chronopotentiometry ...... 34
ix 13. Block Diagram For AC Impedance
Measurements ...... 36
14. Cyclic Voltammograms On A Pure Pt WE
Immersed In A 10 m/o NaVOg-Na^SO^ Solution
Of Basicity -9.77 At 900 C ...... 38 lb. Cyclic Voltammograms At Various Cathodic
Switching Potentials Recorded On A Pt WE
Immersed In A 10 M/o NaVO^-Na^SO. Solution
Of Basicity - 9.77 At 900 C ...... 40
16. Graphical Application Of Equation f2] ... 43
17. Variation of Anodic to Cathodic Peak
Current Ratio With Scan Rate ...... 45
18. Variation of Peak Current With Square Root
Of Scan Rate ...... 46
19. Diagnostic Plot ...... 48
20. Cyclic Voltammogram From A Na^SO^ Melt Of
Basicity - 6.66 At 900 C ...... 50
21. Cyclic Voltammograms At Various Cathodic
Switching Potentials Recorded On A Pt WE
Immersed In A Na^SO^ Melt Of Basicity
- 6.66 At 900 C ...... 51
22. Cyclic Voltammogram From A Na^SO^ melt Of
Basicity - 6.66 At 900 C ...... 53
23. Cyclic Voltammograms From A Na^SO^ melt
Of Basicity - 6.66 At 900 C ...... 55
x Chronopotentiogram From a 10 m/o NaVO^-
Na^SO^ Solution Of Basicity - 9.77 At
900 C For An Applied Current Of 60 mA ...
Plot Of Chronopotentiometric Data
According to Eq. [3] For The Cathodic
Transition In Fig. 24 ......
Plot Of Chronopotentiometric Data
According to E q . [4] For The Cathodic
Transition In Fig. 24......
Chronopotentiogram From A Na^SO^ Melt Of
Basicity - 6.66 At 900 C For An Applied
Of 40 mA ......
Plot Of Chronopotentiometric Data
According To Eq. [3] For The Anodic
Transition A1 In Fig. 27 ......
Plot Of Chronopotentiometric Data
According To E q . [31 For The Cathodic
Transition Cl In Fig. 27 ......
Plot Of Chronopotentiometric Data
According To E q . [3] For The Anodic
Transition In Fig. 27 ......
Chronoamperograms On A Pure Pt WE
Immersed In A Na^SO^ Melt Of Basicity
- 6.66 At 900 C ......
Chronoamperograms On A Pure Pt WE
Immersed In A 10 m/o NaVOg-NagSO^ Solution Of Basicity - 9.77 At 900 C .... 72
Superimposed i vs E Curves From Chrono- amperometric Data From Basic NaVO^-Na^SO^
And NanS0. Melts At 900 C ...... 74 2 4 Chronoamperometric Data Plotted According
To Eq. (5| ...... 75
Superimposed Cathodic Polarization Curves
From Basic NaVO^-NagSO^ and NagSO^ Melts
At 9U0 C ...... 77
Anodic Polarization Curve From A Na^SO^ melt Of Basicity - 6.66 At 900 C ...... 78
Basicity Trace On A Pt WE Painted On A
Zirconia Tube Immersed In A N a r,S0. Melt 2 4 Of Basicity -6.66 At 900 C ...... 80
Cathodic Polarization Curve From A Basic
NaoS0. Melt At 900 C ...... 90 2 4 Basicity Trace On A Pt WE Painted On A
Zirconia Tube Immersed In A 10 m/o NaVO^-
Na.-.SO. Solution Of Basicity -9.77 at 900 C. 93 2 4 Cyclic Voltammograms On A Pure Pt WE
Immersed In A 10 m/o NaVO^-NagSO^
Solution of Basicity -11.7 Under 0^ Gas at 900 C ...... 96
Cyclic Voltammograms From a 10 m/o NaVO^-
Na^SO^ of Basicity -11.7 at 900 C For
Various Cathodic Switching Potentials . . . 99 x ii 42. Variation Of Anodic to Cathodic Peak
Current Ratio With Scan Rate For Cyclic
Voltammograms in Fig. 40 ...... 102
43. Peak Current Variation With Square Root
Of Scan Rate ...... 104
44. Mechanism Diagnostic Criteria For The
Anodic and Cathodic Peaks In Fig. 40 .... 105
45. Cyclic Voltammograms For Various Cathodic
Switching Potentials From a NagSO^ Melt
of Basicity -10.07 Under Gas At 900 C 107
46. Cyclic Voltammogram From A NagSO^ Melt
of Basicity -10.07 at 900..... C ...... 109
47. Chronopotentiogram From a Na^SO^ Melt
of Basicity -10.07 at 900 C...... Ill
48. Chronopotentiogram From a Na^SO^ Melt
of Basicity -10.07 at 900 C...... 114
4y. Chronopotentiogram From a 10 m/o NaVO^-
NagSO^ Solution of Basicity -11.7
at 900 C ...... ‘ 116
50. Diagnostic Plot For The Reaction Mechanism
From Chronopotentiometric Data ...... 118
51. Diagnostic Plot For Chronopotentiometric
Data From 10 m/o NaVO^-NagSO^
Solutions Of Basicity -11.7 at 900 C .... 119
52. Plot of Chronopotentiometric Data From
Fig. 49 According to E q . |_3] 120 xiii 53. Plot of Chronpotentiometric Data From
Fig. 49 According to E q . [4] ...... 122
54. Chronoamperograms From A Na^SO^ Melt
of Basicity - 10.07 at 900 C ...... 124
55. Chronoamperograms From a 10 m/o NaVOg-
Na^SO^ Solution of Basicity -11.7 at
900 C ...... 126
56. Superimposed i vs E Curves From Chrono -
amperometric Data From NaVO^-Na^SO^
Melts Under 0 Gas At 900 C ...... 127
57. Dynamic Polarization Curve From A 10 m/o
NaVOg-NagSO^ Solution Of Basicity -11.7
at 900 C ...... 128
58. Basicity Trace On a Polarized Pt WE
Painted On A Zirconia Tube Immersed
in A Na^SO^ Melt of Basicity -10.07 at
900 C ...... 130
59. Basicity Trace On A Polarized Pt WE
Painted on Zirconia Tube Immersed In a
10 m/o NaVOg-Na^SO^ Solution of
Basicity -11.7 At 900 C ...... 135
60. Cyclic Voltammograms From A 10 m/o NaVO,-
Na^SO^ Solution Of Basicity -11.72 Under
Uncatalyzed 0.1 % SO^-O^ Gas at 900 C ... 138
61. Cyclic Voltammograms For Same Experimental
xiv Conditions as In Fig. 60 For Various CSP 141
62. Variation Of Anodic to Cathodic Peak
Current Ratio With Scan Rate For Cyclic
Voltammograms On Fig. 60 ...... 142
63. Peak Current As a Function Of Scan Rate 144
64. Cyclic Voltammograms From A NagSO^ Melt
Of Basicity -13.46 Under Catalyzed 0.1% SOg
Og Gas at 900 C ...... 146
6b. Chronopotentiogram From a 10 m/o NaVO^-
Na^SO^ Solution of Basicity -11.72 at 900 148
66. Plot Of Chronopotentiometric Data From
Fig. 65 According to E q . [3] ...... 151
67. Plot Of Chronopotentiometric Data From
Fig. 65 According to E q . T4] 152
68. Plot Of Chronopotentiometric Data From
The Anodic Transition In Fig. 65 Ac
cording to Eq . L 3 J ...... 154
69. Chronopotentiogram From A Na^SO^ Melt
Of Basicity -13.46 at 900 C ...... 155
70. Chronoamperograms From A Vanadate-Sodium
Sulfate Solution Of Basicity -11.72 At
900 C ...... 157
71. i vs E Plot From Chronoamperometric Data
From a Vanadate-Sodium Sulfate Solution
Of Basicity -11.72 At 900 C ...... 159
72. Superimposed Polarization Curves From xv NaV03-Na2S04 And Na2S04 Melts At 900 C . . 160
73. Dynamic Polarisation Curve For Pt Foil WE
Immersed In a Na2S04 Melt Of Basicity
-13.46 at 900 C ...... 162
74. Cyclic Voltammogram From a 10 m/o NaVO^-
Na,SO. Solution Under Catalyzed 0.1% S0r
02 Gas at 900 C. Flow Rate 0.223 ml/sec . 164
75. Same as In Fig. 74 For Various CSP ...... 165
76. Chronopotentiogram From a 10 m/o NaVO^-
Na,,S0. Solution Of Basicity -13.15 At 900 168 2 4 77. Plot Of Chronopotentiometric Data From
Fig. 76 According to E q . [3] ...... 171
78. Cyclic Voltammograms From a 10 m/o NaVO^-
Na2S04 Solution Under Catalyzed 0.1% S02
02 Gas At 900 C. Flow Rate 2.223 ml/sec . 172
79. Chronopotentiogram From a 10 m/o NaVO.-- O Na2S04 Solution of Basicity -13.85 At 900 175
80. Plot Of Chronopotentiometric Data From
Fig. 79 According to E q . [3] ...... 177
81. Diagnostic Plot For Chronopotentiometric
Data From Vanadate-Sodium Sulfate Solutions
Of Basicity -13.85 at 900 C ...... 178
82. Cyclic Voltammograms From 10 m/o ^ 0 ^ -
Na2S04 Solutions Under U2 Gas At 900 C .. 180
83. Cyclic Voltammograms From 10 m/o V„0, - 2 5 Na.SO. Solutions Under a Catalyzed 0.1% 2 4 XV I SO^-O^ Gas At 900 C ...... 182
84. Basicity Trace On a Polarised Pt Painted
WE on A Zirconia Tube Immersed In a NaoS0. 2 4 Melt Of Basicity 13.46 at 900 C ...... 183
8b. Cyclic Voltammogram On a Pt WE Painted On
A Zirconia Tube Immersed In A Nar,SO. Melt 2 4 Of Basicity -13.46 At 900 C ...... 185
86. Basicity Trace On A Polarized Pt We
Painted on a Zirconia Tube Immersed In
a Vanadate-Sodium Sulfate Solution Of
Basicity -11.72 at 900 C...... 188
87. Basicity Trace On a Polarised Pt WE
Painted On a Zirconia Tube Immersed In
a Vanadate-Sodium Sulfate Solution Of
Basicity -13.85 At 900 C...... 192
88. AC Impedance Results On A Pt WE Painted
On a Zirconia Tube Immersed In A Vanadate-
Sodium Sulfate Solution Of Basicity -9.77 195
89. AC Impedance Results On A Pt WE Painted
On a Zirconia Tube Immersed in A Sodium
Sulfate Solution Of Basicity -6.66 196
90. Ac Impedance Results On A Pt WE Painted
On A Zirconia Tube Immersed In A Vanadate-
Sodium Sulfate Solution Of Basicity -11.7 198
91. AC Impedance Results On A Pt WE Painted
On A Zirconia Tube Immersed In A Sodium xvii Sulfate Solution Of Basicity -10.07 ..... 200
92. AC Impedance Results On a Pt Foil WE
Immersed In a Vanadate-Sodium Sulfate
Solution Of Basicity -9.77 202
93. AC Impedance Results On A Pt Foil WE
Immersed In A Sodium Sulfate Solution
Of Basicity -6.66 204
94. AC Impedance Results On A Pt Foil WE
Immersed In A Vanadate -Sodium Sulfate
Solution Of Basicity -11.7 ...... 205
95. AC Impedance Results On A Pt Foil WE
Immersed In A Sodium Sulfate Solution
Of Basicity -10.07 ...... 207
xv j j j CHAPTER I
INTRODUCTION
Serious corrosion problems are introduced upon the combustion of fuel oils containing vanadium and sodium in high temperature systems. Vanadium is a transition element that forms a variety of complexes in all oxidation states from +5 to -1, the +5 and +4 states generally being the most stable (1). Vanadium appears as an impurity in all crude oils of petroleum origin associated with the organic portions of the oil in the form of organometallic compounds called porphyrins (2).
Its concentration varies from 1 to as high as 1400 ppm, depending on the source of the crude oil.
During the combustion of fossil fuels, vanadium is undoubtedly present in oxidized form, but not necessarily as simple oxides. Vanadium pentoxide, the most likely oxide of vanadium upon normal industrial combustion operations, can combine readily with other metal oxides, particularly i-0 form vanadates. In ways that are not completely understood, the presence of 2
vanadates prevents oxide films from growing as a dense, adherent and protective coating. Therefore, catastrophic deterioration of engineering structures is observed. The abnormally short life times of boilers and gas turbines that burn vanadium-bearing fuels have long been an unsolved technical problem (2-5). Even a low concentration (a few percent) of vanadium in alloys decreases their resistance to oxidation, particularly with steels. The addition of vanadium to steels that normally form tight, adherent, and protective oxides converts the scale ( oxide layer) into a porous, bulky and friable substance that readily spalls off the steel surface (2).
Different approaches have been used to account for the catastrophic oxidation experienced by high- temperature alloys used in gas turbines, boilers, diesels, and other high temperature installations using vanadium-bearing crude oils as fuels. Because alkali metal compounds and vanadium pentoxide constitute the major components of residual oil ash, much of the research interest has focused on their properties and chemical reactions. It is generally accepted that the presence of molten oxides or complexes is necessary to promote accelerated oxidation at high temperatures, i.e., hot corrosion. 3
Monkman and Grant (6 ) studied the role of Na^SO^ in
vanadium-induced hot corrosion. They concluded that
Na^SO^ promotes accelerated oxidation below the melting
point of by forming a liquid phase with ^2^ 5 *
at temperatures above the melting point of , where
the effect of sodium sulfate on the viscosity of molten
mixtures is negligible, NagSO^ inhibits accelerated
oxidation presumably due in part to a dilution effect as
well as chemical effects. The concentration of
vanadium in the oxide scale increased as the metal-oxide
interface was approached
Leslie and Fontana (7) postulated that vanadium
pentoxide vapor is the main cause of vanadium-induced
catastrophic oxidation. Sykes and Shirley (3) carried
out tests to examine the degree of volatilization of vanadium pentoxide into the gas stream. They did not
observe appreciable volatilization up to 1000 C
However, V2^5 the furnace system, but out of direct
contact with the steel, increased the scaling rates at
750 and 850 C. They concluded that this effect might be
related to some catalytic activity which modifies the products of combustion to give higher rates of attack.
Vanadium oxide is used to catalyze SO^ (g) formation in commercial sulfuric acid plants. Fitzer and Schab (8) exposed specimens to pure oxygen as well as to oxygen saturated with vanadium pentoxide vapor, and compared their weight losses with those obtained when the specimens were dipped occasionally
into molten vanadium pentoxide. The results showed that vanadium pentoxide vapor greatly enhances the oxidation of copper and a chromium steel, and has some eftect on pure nickel, but very little effect on pure iron.
However, the presence of liquid vanadium pentoxide was more profound than the vapor in its influence on oxidation. It can be concluded that in the absence of a liquid phase, another mechanism, different from fluxing, must operate to accelerate oxidation. One proposed mechanism (9), based on the Wagner-Hauffe doping theory, postulates a profound increase in the number of cation vacancies in the oxide (for a metal-deficit semiconductor such as ferrous oxide or nickel oxide) owing to the substitution of the impurity cation with a high valence . However, the loose and spongy appearance of the scale formed on heat-resistant steels suggests that oxygen can penetrate to the metal-scale interface through fissures in the oxide layer, so that diffusion of cations through the oxide lattice is not the controlling rate, step and the doping theory is not applicable (10).
Accelerated corrosion can be encountered either when 5
metals are alloyed with alloying elements whose oxides
are low melting, such as molybdenum, vanadium, lead,
bismuth, etc., or when the surface of heat resistant
alloys are contaminated with low melting oxides or other
salts such as halides, sulfates, etc.(11). In either
case the growth of a thick corrosion product layer on
the surface is a common feature. Leslie and Fontana (7)
studied the unusual rapid oxidation at high temperatures
of alloys containing molybdenum. A rapidly formed
spongy scale with a characteristic "pie-crust"
appearance was observed. The phenomenon was attributed
to the accumulation of molybdic oxide (MoO^) vapor on
the metal surface. The spongy scale did not show
distinct layers consisting of different oxides, but a
single, highly porous layer. Inert markers applied to
the surface of the unoxidized specimen were retained on the outer surface of the scale layer, indicating that transport proceeds inward from the gas interface , and that the oxidation reaction takes place at the metal- oxide interface. The porous bulk oxide allows oxygen to penetrate to the metal surface. Besides an adequate supply of oxygen from the furnace atmosphere, a more powerful oxidizing agent must exist at the interface.
There must be some "oxygen carrier" that oxidizes the metal at the scale-metal interface more rapidly than the furnace atmosphere. Leslie and Fontana (7) proposed the dissociation of MoO^ with release of nascent oxygen which is assumed to increase greatly the oxidation of molybdenum-bearing alloys. Vanadium pentoxide and, to a
lesser degree, a variety of other oxides, accentuated and facilitated the catastrophic oxidation of alloys containing molybdenum.
Vanadium pentoxide increases the rate of oxidation of a variety of alloys free of molybdenum (6,8,11). The observation of a "pie-crust" scale for rapid oxidation suggests that the effect of the vanadium pentoxide is analogous to that observed for molybdenum-bearing alloys. Cunningham and Brasunas(ll) demonstrated that the severe corrosion of heat-resistant alloys contaminated by V.O,. was aggravated by addition of ^ J Na^SO^. The most severe mixture was in the range of 15-
20% Na^SO^. The increase in oxidation rate could not be explained in terms of changes in melting points but seems to be related to the oxygen solubility in the molten V - N a r.SO. mixtures. The most corrosive 2 b 2 4 solution also dissolves the greatest amount of oxygen .
Cunningham and Brasunas (11) proposed a mechanism involving a porous oxide scale which provides free access of atmospheric oxygen to a molten oxide film at the metal-oxide interface which is the zone of reaction. Brasunas(12) demonstrated that oxidation during the
initial stages of corrosion was much more severe in the
presence of liquid "than in a pure oxygen
atmosphere.
Several oxygen carrier mechanisms have been
postulated for vanadium-induced hot corrosion , ranging
from the disssociation of vanadium pentoxide to form
nascent oxygen at the interface, to a variety of complex
vanadates. Based on phase equilibrium studies, Foster et
al (13) reported NaVO^ and a NagO.SV^O^ complex as
potential corrosive compounds. Pure iron and chromium
vanadates can cause accelerated attack, and the enhanced
effect of the sodium vanadates is attributed to the
evolution of oxygen associated with the decomposition of
the vanadyl vanadate (14). The alternate formation and
decomposition of the vanadates may contribute to the
porosity of the scale, and thus to the easy access of
oxygen to the metal-oxide interface.
These studies presented evidence to support the
hypothesis that the accelerated corrosion by vanadium-
containing ash is initiated by molten complex compounds
and is continued by the reaction of the active metal
surface with oxygen that is absorbed in the molten material. 8
For alloys containing Mo, W, or V, catastrophic or self-sustained rapid oxidation can occur because of the dissolution of the oxides of these elements into Na~SO.. 2 4 Evolution of SOg from Na^SO^ occurs upon the addition of
V o0 c according to the reaction, 2 D
V2°b + Na2S04 > Na2V2°6 + S03 where
Na^O + SOg ^ Na^SO^
which shows explicitly that reduces the sodium oxide activity of the molten Na^SO^ (lb).
V.O. is an acidic oxide. According to the acid base 3 b definition by Lux,
2 - acid + 0 = base vanadium pentoxide has the property to complex oxide ions to form vanadates. Vo0. is present in the melt at 6 O higher concentration than dissolved SOg and; therefore, the basicity of the melt may be fixed by the acid-base equilibrium between and NaVOg.
Zhang and Rapp (16) have shown that oxide dissolution reactions increase the oxide ion concentration
(increase the activity of the alkali oxide) of the molten salts while producing anions of the protective oxide scales which are highly soluble.
The effect of vanadium pentoxide disappears below bbO 9
■to 600 C; its influence appears to be confined to the
acidic fluxing of protective oxides such as
3 + 2 Cr + 6 V03 3 V 2°5 + Cr2°3 or 3 + 3 NaVOg + Cr2°3 2 Cr + Na3 (V04) + 2 V04
with
2 - 2 - O
or
2 - 2 - O
At temperatures above the melting point of ^2^b’ m o s ‘*:'
alloys tested have shown a markedly accelerated
corrosion attack (14).
Fused salts generally exhibit predominant ionic
conduction, so that some electrochemical process is
necessarily involved in the accelerated oxidation
induced by a molten salt. The electrochemical process
must involve oxidation of the metal , ionic transference
in the electrolyte, reduction of an oxidant and perhaps
a chemical fluxing of the oxide scale (17). Rapp and
Goto (18) proposed a mechanism for the hot corrosion of
a pure metal consisting of an oxidation of the metal at the metal-oxide interface and a corresponding reduction of the oxidizing agents (0^ or SO^) taking place either at the oxide/salt interface or, under particular circumstances, at the salt/gas interface. Ordinarily, 10
oxygen or/and SOg molecules would need to dissolve and
diffuse through pure Na^SO^ to be reduced at the oxide-
salt interface. However, if transition metal (V, Mo,
Co, etc.) impurity ions exist with sufficiently high
concentrations in the salt film, the reduction of the
oxidizing agent may be shifted to the salt/gas interface
either by the counter — diffusion of multivalent
transition cations through the melt or else by the
hopping of electrons from an ionic species of one
valence to that of another. Clearly one needs further
information about the electrochemical reactions and
transport behavior in fused sodium sulfate containing
vanadium oxides and/or sodium vanadates.
Cyclic voltammetry, chronopotentiometry,
chronoamperometry and potentiodynamic polarization are
electrochemical analytical methods for the study of
electrode reaction mechanisms which permit the
investigation of the electroactive reactants,
intermediates,and products of the electrode reaction.
These electrochemical techniques have been applied in this research to study the electrochemistry of 10 m/o
NaVOg-Na^SO^ and 10 m/o VgO^-Na^SO^ solutions at 900 C.
Two high temperature selective ion electrodes were used in the course of the experiments. The thermodynamics of the system are presented and are compared with the 11
experimental results as an auxiliary tool in the elucidation of the reaction mechanism. CHAPTER II
ELECTROANALYTICAL TECHNIQUES
2. a Cyclic Voltammetry
Cyclic voltammetry is a controlled potential
electrochemical technique in which the potential of the
working electrode is swept linearly with respect to time
at a constant scan rate,// mV/sec, from an initial
potential, E^ , to a given potential E^ , called the
switching potential. Then the scan direction is reversed
to the initial potential , completing a triangular cycle
such as the one shown in Figure l.a(19). The potential
can be swept between these two potential limits in
continuous finite cycles. In this study the "European convention" is used, i.e., a positive E is an oxidizing potential and a positive i is an anodic current.
The resulting current is measured as a function of the applied potential as shown in Figure l.b. A redox reaction can be identified as a maximum in the current.
As the potential is swept from the initial value, the current flow (the rate of reaction) rises, because the 12 13
0 Switching time, A t — ►
A + e —► A
EH
A - e
Figure 1 . Cyclic Voltammetry. (a) Cyclic ootential sweep . (b) Resulting cyclic voltammogram. 14
reactant begins to be reduced , until the concentration
of the reactant at the surface of the working electrode
is depleted to zero by the reaction, then the current
begins to decrease and the reaction rate is limited by
the diffusion of more reactant from the bulk of the
solution to the electrode surface.
The peak current and peak potential are sensitive to
changes in scan rate. The peak current is also related
to the concentration of the electroactive species and
its diffusion coefficient in the bulk of the
electrolyte, as well as to the number of electrons
transfered in the redox reaction. Variation of the scan
rate, initial potential, switching potential, and bulk
concentration can be used to examine the electrode
reaction mechanism.
The evaluation with cyclic voltammetry of
complications in redox reactions due to coupled chemical
reactions and/or adsorption effects has been thoroughly presented elsewhere(20,21) .
2.b Chronopotentiometry
Chronopotentiometry is a controlled current electrochemical technique which involves the measurement of the variation with time of the potential at a working electrode during a short period of exhaustive 15
(a)
Input c a> i _ L_ D u
0
Time
(b) +
\<*— seconds-- ►{
O
c Q) O Q_
Time
Fiqure 2 . Chronopotentiometry. (a) Current vs. time input (b) Potential vs. time response. 16
electrolysis carried out at constant current under
linear diffusion conditions. Figure 2 shows a
representation of the input and output of the technique.
The resulting chronopotentiogram showing the variations
of potential with time can be used for a variety of
analytical purposes, including the measurement of
concentration of electroactive species, as well as
electrode or solution kinetics (22). The transition
time, 'T' , represents the time required to oxidize or
reduce completely all the electroactive species in the
immediate vicinity of the working electrode. This
transition time is of primary analytical importance,
because 'f is directly proportional to the bulk concentration of the electroactive species. The
transition time can be greatly changed by variation of the current. It is often necessary to choose an
appropriate current density so that the transition time
is relatively short. Otherwise, convection may disrupt the diffusion layer and upset the reproducible diffusion process. Convective interference is particularly likely in molten salts where fine thermostatic control is difficult at the high operating temperatures, so that temperature gradients occur in the solution (23).
For a purely diffusion controlled (reversible) redox process, the product i^^^ is independent of i, where Adsorption
Diffusion
Preceding reactions
i
Figure 3 . Diagnostic Dlot. 18 i is the applied current, Fig. 3. Deviations from a horizontal line give indications of preceding chemical reactions or adsorption effects accompanying the redox process.
Chronopotentiometry with current reversal can be used to study adsorption effects and proceding chemical reactions. In current reversal chronopotentiometry, the constant applied current is reversed at or before the transition time for the electroactive species so that it may be reoxidized or reduced, and a reverse transition time is recorded. The potential-time curve thus obtained can be treated quantitatively (22). If the current is reversed exactly at or before the transition time, then the ratio of the backward transition time to the forward transition time, ’ ‘*'s eclual "to one-third, independent of the experimental conditions. Deviations from this ratio indicates adsorption or/and kinetic complications. For a first-order chemical reaction, a relationship between t^ and t^ has been developed which allows the determination of the reaction rate constant
(24,25) .
2.c Chronoamperometry
Chronoamperometry is an electrochemical technique in which a potential step perturbation is applied to the working electrode and the current response is recorded 19
E, f
E i
E X:0
o 0 time 0 t im e
C.
nFAD/2Cb
, « ^ - * 1/2 it
t im e ---- ►
Figure 4 . Chronoamperometry. (a) Potential step vs. time input, (b) Current vs. time response (c) Diagnostic plot. 20
as a function of time. The applied potential step is chosen within a potential range for which a redox peak has been detected in a previously recorded cyclic voltammogram. According to the Cottrell Equation for a diffusion-control redox process, the current response is related to the diffusion coefficient and concentration of the electroactive species. For a diffusion 1/2 controlled redox process the it product is independent of time, Fig. 4 . Deviation from a horizontal line is an indication of a preceding chemical reaction or adsorption effects accompanying the redox process. Detailed derivations are given elsewhere
(19, 26,27).
2.d AC Impedance
In ac impedance measurements, the total impedance of the cell is measured from the response of the sinusoidal wave which has a small amplitude of ac potential superimposed on a dc potential. The total impedance
Ztotal °* ceH a complex variable,
Ztotal= ZK + ZIJ where ZR and Zlj are the real and imaginary components of the impedance, respectively. The total impedance is measured at various frequencies of the sinusoidal wave and it is usually plotted in the complex plane, which is known as a Cole-Cole plot, or Nyquist plot. 21
This electrochemical technique can provide kinetics and mechanistic information from a corroding interface.
Electrochemical interfaces are analogous to an electronic circuit consisting of an array of resistors and capacitors. Ac impedance measurements enable the researcher to characterize the electrochemical system in terms of its equivalent circuit based on established ac circuit theory.
The simplest impedance for a corroding interface may be represented by an equivalent circuit with only one time constant as in Fig. 5.a in the case of purely activation controlled corrosion, where R and C,-, are p dl the polarization and the double layer capacitance, respectively. The plot in Fig. b.b illustrates the expected response of the simple circuit in Fig. 5.a.
At high frequencies only the uncompensated resistance,
R q , contributes to the real portion of impedance.
Although R q is generally associated with the solution resistance, in reality it contains the resistances of the solution, electrical leads, surface films, etc. At very low frequencies, the polarization resistance, R , P also contributes to the real portion of impedance. In
Fig. 5.b the diameter of the semicircle is the polarization resistance. u
-AAAAr~ R p - V W r Rq = Uncompensated Resistance Rp=Polarization Resistance R(jl=Double Layer Capacitance
(b)
■Decreasing Frequency o g cn max= CRn, u> = 27rf o E
Kl “ 5
Z'(Real)
Rp=2|Z|tan 9 max NYQUIST PLOT High Frequency1 Z'^O , Z'-> R Low Frequency1 Z% 0, Z'-» R{}+Rp
Figure 5 . AC Impedance representation for a purely activation controlled interface, (a) Equivalent electrical circuit, (b) Nyquist plot. 23
The simple model in Fig. 5.a is not realistic for most
interfacial impedances because it does not take into
account rate control by diffusion of charged species,
which lead to the occurrence of the so-called Warburg
impedance. Considering the mass transport effect, the
faradaic impedance for a corroding system is the
summation of charge transfer resistance and mass
transfer impedance. The equivalent circuit and its
schematic impedance in a complex plane are shown in
Figs. 6.a and 6.b. The complex plane representation has
a semicircle at high frequencies dominated by charge
transfer resistance and the double layer capacitance.
The diffusional impedance begins to dominate the
interfacial impedance at frequencies below 1 Hz for most
systems with a straight line of 45° slope. The
polarization resistance in this case is the sum of the
charge and mass transfer resistance ( R + R ,). In the d limiting case of a purely diffusion-controlled interface
reaction, the Nyquist plot shows a straight line of
slope 45° for high frequencies. In the case of semi
infinite diffusion impedance (Warburg impedance) the
line does not bend even at very low frequencies.
The Bode plots constitute an alternate representation of the total impedance cell. The Bode format is desirable when data scatter precludes adequate fitting -VWAVW-
-»AWV—\A/- R
Decreasing o> vl
ZR
Figure 6 • AC Impedance representation for a mixed controlled interface, (a) Equivalent electrical circuit, (b) Nyquist plot . 25
of the Nyquist semicircle and, in general, provides a
clearer description of the frequency-dependent behavior
of the electrochemical system than does the Nyquist plot
Figure 7 shows the Bode plot for the same data of
Fig. 5.b . When the absolute impedance is evaluated as a
function of frequency, the values of R and R #-» are P “ obtained as the intercept with the ordinate at low and
high frequencies , respectively. At intermediate
frequencies, the break point of this curve should lie on
a straight line of slope -1. Extrapolation of this line to log W =0 yields the value for . The plot of the phase shift angle as a function of frequency yields a
frequency at which the phase shift of the response is maximum. From this maximum, C,-, can be evaluated. d 1
in some systems there are series of rate determining steps. Each step represents a system impedance component and contributes to the overall reaction rate constant. The ac impedance experiment can often distinguish among these steps and the Bode format is more sensitive to their detection than the Nyquist plot in the sense that each step can be characterized by the break points in the curve or by corresponding maxima in the phase shift angle, Fig. 8. For a diffusion- controlled step, the Bode plot will show a slope of -1/4 or -1/2 in the linear portion and a corresponding 26
|Z| = ■d.l.
CD
Figure 7 . Bode plot for an activation controlled interface reaction.
5.1
2
o> a> O CD
- 8 0 - 4 Log Frequency (Hz)
Figure 8 . Bode plots for a system with two rate determining steps. 27
maximum phase shift angle between 22.5° and 45°
When diffusion-control is suspected, a Randles plot
constitutes a reliable diagnostic criterion,Fig. 9. For
a diffusion-controlled reaction, a plot of ZR versus
- 1/2 W should yield a straight line indicating that ZR
- 1/2 and ZI are linear and equal functions in W . The
intercept with the ordinate yields the Warburg
impedance, Zw . For mixed control at the corroding
interface, a plot of ZR versus ZI/W at high frequencies
allows the determination of R , R and C,,. Such a II p dl plot should yield a straight line, as illustrated in
Fig. 10.
The object of an ac impedance measurement experiment
may be either to determine the values of the various
elements in the equivalent circuit or simply to confirm that a given electrochemical system fits a particular equivalent model. Detailed derivations and more practical applications of the technique are given elsewhere (28-30). CL Kl ZR(Ohm) Figure Figure Rr for an activation controlled controlled plot Zi/W activation vs. an ZR . for 10 Fiqure ici i Fg 5(a). Fig. in circuit interface equivalent to the the to equivalent interface u) l/ 9 2 Rnls plot. Randles . (Rad/Sec) (Rad/Sec) Zl/O ) 1/2 CHAPTER III
EXPERIMENTAL PROCEDURE
Electrochemical measurements were perfomed with a
three-electrode arrangement cell, i.e.,working electrode
(WE), reference electrode (RE) and counter electrode.
The working electrode was a 0.75 cm2 Pt foil spot-welded
to a Pt lead wire . The Pt lead wire was embedded in a
high-purity alumina tube and sealed with a high
temperature ceramic cement (Ceramabond 552). The Pt
foil was the only exposed area of the working electrode,
and it was completely immersed in the electrolyte. For the trace of basicity during electrode polarization, the working electrode was a painted Pt WE sintered on a 3.5 wt% CaO-stabilized zirconia tube immersed in the electrolyte. The WE area exposed to the electrolyte was approximately 0.75 cm2. Prior to an experiment, the Pt foil WE was ultrasonically cleaned in a methanol bath and then heated in a torch until red.
For all electrochemical techniques, the reference electrode was the Ag/Ag+ high-temperature electrode made
29 30
of a pure silver wire dipped into a 10 m/o Ag^SO^-NagSO^ mixture contained in a one-closed-end mullite (McDanel-
MV30) tube. The silver wire was spot-welded to a platinum lead wire and the mullite tube was sealed with high temperature ceramic cement to maintain a constant composition of the AggSO^-NagSO^ mixture in equilibrium with the SO^-SOg gas liberated upon melting of the salt.
Mullite has a nominal composition of SAlgO^^SiOg grains in a glassy boundary phase which provides exclusive sodium ion conduction. In cyclic voltammetry measurements a CaO-doped partially stabilized zirconia tube (Degussa ZR-23), was used as an additional reference electrode. The inside bottom of the tube was painted with platinum paste and sintered for 48 hours at
1100 C. A platinum lead wire was placed in contact with the sintered paint and the tube was left open to the air. The partially stabilized zirconia tube exhibits exclusive oxide ion conduction and is used to measure the partial pressure of oxygen in the melt at open- circuit potential.
The counter electrode was a 0.5 mm thickness Pt wire, shaped as a ring, spot-welded to a platinum lead wire.
The Pt lead wire was sealed into an alumina tube. The counter electrode was centered about the working electrode. 31
The electrodes and the electrolyte were placed in a
99.8% alumina crucible held at one end of a mullite support tube. The gas stream was dried and then directed to the surface of the electrolyte through a high-purity alumina tube. Heated Pt/Pd catalyst placed inside the alumina tube ensured the equilibrium between
SC>2 , Og and SO^. The gas outlet line was protected from the atmosphere by a bubbler bottle. - -The cell was placed in a high temperature mullite chamber which was closed by a water-cooled brass flange sealed by a Teflon o- ring. All the electrodes and leads were fitted through brass adapters(Cajon or Swagelok) containing rubber o- rings. A fan was used to cool the upper flange.
The electrolyte, NaoS0., and solute, Vo0 c or NaV0o , A H A o 3 were premixed and were maintained in an oven at 200 C in order to guarantee dryness. Once the system was assembled, the testing furnace was heated to 400 C for a period of 10 to 12 hours. Then the temperature was increased in intervals of 50 C up to the working temperature of 900 C. The working temperature was controlled to + /- 3 C by a Barber Colman 520 solid-state controller and a Barber Colman 621 B power supply. To measure the melt temperature, a type S, Pt-10%Rh, thermocouple was placed inside a closed-end mullite tube immersed in the melt. The depth of the melt was about 32
1.5 cm. Figure 11 shows the experimental setup.
In cyclic voltammetry experiments, an Aardvark
Model V potentiostat controlled the potential between
the working and reference electrodes. A Bioanalytical
Systems voltage generator Model CV-1B-120 was used to
generate the triangular waves to the potentiostat for
the cyclic potential scan. The current and potential
outputs were recorded with a Houston Instrument X-Y
recorder Omni-graphic 2000.
In chronopotentiometry measurements, a Wenking ST72 potentiostat in series with a resistance box was used to
impose a constant current to the electrochemical cell.
A Bascom-Turner 4120 recorder was used to record the potential-time response. Figure 12 shows the apparatus arrangement for cyclic voltammetry and chronopotentiometry.
Chronoamperograms and polarization curves were obtained using the Princeton Applied Research 350
Corrosion Measurement System.
In AC impedance measurements, a digital frequency analyzer (Schlumberger-Solartron 1170) was used to control the Ministat precision potentiostat (H. B.
Thompson & Associates) through which various frequencies of sinusoidal current waves were applied to the WE. All 33
Ag/Ag+/AAullitex 0 2/S 0 2 Z r0 2 Gas /WE RE CE Outlet
Furnace
Platinized Ceramic Catalyst—
T.C. a i2o 3 Crucible
~— Mullite K Furnace Tube
figure 11. Exoerimental setup for electrochemical studies at 900 C. Potentiostat folenliQl C E W E RE Output Input
Resistor Voltage Box Ramp
C E WE RE X-Y Cell Recorder
Potentiostat
CE WE RE
Resistor Ammeter Box Y-t Recorder
CE WE RE Cell
Figure 12. Apparatus Arrangement (a) Cyclic Voltammetry (b) Chronopotentiometry 35
data were collected and analyzed by an Apple lie
computer connected to the digital frequency response
analyzer. Figure 13 shows the block diagram for the AC
impedance measurements.
A Keithley Model 177 digital multimeter was used for the measurement of the electrode potentials.
The reagents used were:
Sodium Sulfate, anhydrous powder, Baker Analysed
Reagent, 3898-1
Vanadium Pentoxide, Certified, Fisher Scientific
Company, V-7
Sodium Vanadate, meta purified, Fisher Scientific
Company, S-455
99.99% silver wire, 2mm diameter Puratronic
24 gauge Pt lead wire.
0.020 " diameter Pt-10%Rh thermocouple wire
Pt Thermocouple wire, O.D= 0.015" Reference Electrode
Working Electrode
Counter Electrode
c Solatron
Ou'C'J *
Apple II Plus Mini - Computer
Figure 13 . Block diagram for AC imoedance measurements.
GJ cr> 37
CHAPTER IV
RESULTS AND DISCUSSION
Electrochemical Studies in Relatively Basic
NaVO^-Na^SO^ Solutions
A certain amount of NagO^ was added to a 10 m/o NaVO^-
NaoS0. solution to permit electrochemical studies in a 2 4 relatively basic melt. The amount of NagO^ added was calculated by applying the mass action law to the chemical equilibrium reaction,
Na2V 2°6 + 2 Na2°2 \ Na6V 2°8 + °2 1 3 The open-circuit potential indicated a basicity of -9.77 on the log a NagO scale.
Figure 14 shows a series of cyclic voltammograms recorded for various scan rates on a pure Pt foil WE immersed in this basic 10 m/o NaVO^-NagSO^ solution at
900 C . The scan rate was varied from 20 to 80 mV/sec.
All potentials were recorded against the Ag/Ag+/mullite high temperature reference electrode. One cathodic peak
(labelled Ic) and one anodic peak (labelled la) were 38
U-40
ui a, cc
PC T€ NIT IAL
figure 14 . Cyclic voltammogram for a pure Pt foil WE immersed in a 10 m/o NaV0 3 -Na2 S0 ^ solution of basicity
- 9.77 at 900 C. P = mV/sec obtained at about -1300 and -850 mV, respectively. No
appreciable variation of peak potentials with scan rate
was obtained. The redox relationship between peaks Ic
and la was established by progressively varying the
cathodic switching potential at a constant scan rate of
40 tnV/sec, Fig. 15. For a cathodic switching potential
more negative than -970 mV, an anodic peak is recorded
upon reversal of the potential, indicating the
reoxidation of the previously reduced electroactive
species. As the cathodic potential is made more
negative, the cathodic current increases and a better
defined anodic peak is obtained, indicating that more
reduced species are available for reoxidation in the
immediate vicinity of the working electrode. For a
cathodic switching potential of -1355 mV, a well defined
couple of cathodic and anodic peaks is obtained;
therefore, peaks la and Ic constitute a redox couple.
The reversibility of the redox reaction can be
evaluated from the change of the peak potentials with
scan rate. Table 1 lists the potentials associated with peaks la and Ic at various scan rates. The peak potentials are identical for scan rates of 60 and 80 mV/sec. The slight cathodic and anodic shifts for peaks
Ic and la at scan rates between 20 and 60 mV/sec can be associated with kinetic complications following the o*y-p
8-33 mA
500 mV POTE NT IAL
Figure 15 . Cyclic voltammogram for a pure Pt foil WE immersed in a 10 m/o NaVO^-Na^SO^ solution of basicity -9.77 at 900 C.
U = 40 mV/sec. 0.c.D(open circuit potential )= - 378 mV. 41
Table 1. Peak Potential As a Function Of Scan Rate For a
Pure Pt Foil WE Immersed In a NaVOg-NagSO^ Solution Of
Basicity -9.77 At 900 C.
E t la Ic (mV/sec) (mV) (mV)
20 - 865 - 1278 40 - 865 - 1290 60 - 853 - 1300 80 - 853 - 1300
Table 2. Anodic To Cathodic Peak Current Ratio For
rari ous Scan Rates Fo r a Pure Pt Foil WE Immersed In a
laV03-Na Solution Of Basic ity -9. 77 At 900 C 2S04
icp = icp iap isp iap/icp iap o o o
mV/sec) (m A ) (m A ) (mA) (mA)
20 53 . 50 15 . 50 46.00 0. 79 42.41 40 68 . 25 27 . 50 55. 50 0 . 88 60.29 60 81 . 25 34 . 25 65.00 0. 89 72. 76 80 90 . 00 40 . 00 74.00 0. 93 83. 62 42
redox process, as indicated by Nicholson and Shain (20).
The variation with scan rates of the ratio of anodic to
cathodic peak currents can be used to validate this
prediction. The cyclic voltammograms in Fig. 14 do not
allow an experimental base line from which the current
associated with anodic peak la could be measured.
Nicholson (31) has proposed the following theoretical
relationship which allows the determination of the
anodic to cathodic peak current ratio from the
experimental measurements:
iap/icp = (iap)Q/(icp)Q + .485(isp)Q/(icp)Q + .086 [2]
where (iap)Q,(icp)Q and (ips)Q are the currents
associated with the anodic peak, cathodic peak and with the switching potential, respectively, measured with
respect to the zero current axis, Fig. 16.
Table 2 lists the anodic to cathodic peak current ratios calculated from Eq. [2] for various scan rates.
Figure 17 shows a plot of the iap/icp current ratio as a function of scan rate. According to Nicholson and Shain
(20) the trend of the data in Fig. 17 suggests a reversible redox reaction followed by an irreversible chemical reaction which consumes the reduced species.
The plot tends to flatten to a constant ratio for sufficiently fast scan rates, at which the following chemical reaction has no significant effect on the tap.
ISD
POTE NTIAL
Fioure 16 • Grannical reoresentation of Equation 12 1 for the calculation of tiie iap/icp ratios. 44
recorded voltammogram, and the redox peaks should appear
at their normal potentials. This limiting case will
depend on the magnitude of the kinetic parameters.
A plot of peak current as a function of the square
root of the scan rate should yield a straight line
according to the Randles-Sevcik equation, when the redox
process is diffusion controlled ( Nernstian behavior).
Figure 18(a) and 18(b) are such plots for the cathodic
peak Ic and the anodic peak la, respectively. The
anodic peak current iap was calculated using the iap/icp
current ratios obtained from Eq. [2] and the value for
icp equals (icp)Q since the chosen baseline of zero
current corresponded to the cathodic process. The plots
in Fig. 18 suggest that the electroactive species are brought to the working electrode surface by semi
infinite one-dimension diffusion and that the electron transfer at the working electrode is relatively rapid, providing a reversible redox process.
A more careful examination of peak Ic reveals a sharp decay in current after the peak current has been reached, as well as an increase in the symmetry of the peak as the scan rate is increased, both features characteristic of adsorption. A smooth decay of the current proportional to the square root of time, as predicted by the Cottrell equation, would have been iap/icp 0.76 0.84 0.92 1.00 2 4 6 8 100 80 60 40 20 0 Figure Figure 17. ih cn rate. scan with Variation of anodic to cathodic peak current ratio ratio current peak cathodic to anodic of Variation v (mV/sec) 45 46
a. 1 0 0
< E CL O
4 6 8 10 1/2 V
b. 100
< E CL 0 3
1/2
Figure 18 . Variation of peak current with square root of .scan rate, (a) cathodic process (b) anodic process. 47
observed if the redox reaction had been totally
diffusion controlled. But no prepeak in the vicinity of
peak Ic is observed, which rules out the strong
adsorption of the reactant. However, weak adsorption of
the reactant could take place. Under this circumstance,
the diffusion-control peak and the adsorption-control
peak occur at nearly the same potential (21). The plots
in Fig. 18 can be used as diagnostic plots where a
deviation from a straight line would indicate adsorption
or a preceeding chemical reaction, Fig. 19. In the case
of weak adsorption of the reactant, Fig. 18(a) would
still indicate a diffusion-control reduction peak.
Since NagSO^ is used as the supporting electrolyte in
the present electrochemical studies, cyclic
voltammograms were recorded for a pure Pt foil immersed
in a molten NagSO^ bath containing an amount of NagOg
equivalent to the one added to the NaVOg-NagSO^
solutions. Under this condition, the effect of the
addition of NaVO^ to a basic supporting electrolyte can
be evaluated. The dimensions of the working electrode, the working temperature, as well as the current and potential scales were kept the same, so that one could distinguish the electrochemical response from the vanadate solutes. 48
Adsorption
Diffusion
Preceding reactions
1/2 V
Figure 19 . Diagnostic plot 49
Figure 20 shows a cyclic voltammogram for the pure Pt
foil immersed in a NagSO^ melt at 900 C of basicity
-6.66, at a scan rate of 40 mV/sec. Three cathodic
peaks (labelled 1’c, Il’c and III’c) are observed at
-22b, -412, and -1575 mV, respectively. Upon reversal of
the potential four anodic peaks (labelled I’a, II’a,
Ill’a and IV’a)are obtained at -1940, -1125, -560 and 40
mV, respectively. The cyclic voltammogram of Fig. 20
closely resembles the cyclic voltammograms in Fig. 14
for NagSO^ melts containing NaVO^, with peaks III’c and
Ill’a shifted in the cathodic direction by 272 mV with
respect to peaks Ic and la in Fig. 14. To differentiate
peak III’c from Ic, and peak Ill’a from la, cyclic
voltammograms were recorded upon varying the cathodic
switching potential with a fixed anodic switching
potential, Fig. 21. For a cathodic switching potential
of -1450 m V , no reduction or oxidation peak in the
potential interval -500 to -1350 mV was obtained for
these basic NagSO^ melts . For the same potential
interval, well defined cathodic and anodic peaks are obtained at -1300 and -880 mV, respectively, for basic
NagSO^ melts containing NaVOg, Fig. 15. From Fig. 21, a
significant cathodic current begins to be observed only for cathodic switching potentials more negative than
-1450 mV . For a cathodic switching potential of -1750 mV, well defined cathodic and anodic peaks III’c and 50
PQTEJmAi. i : i :u:tr.:: l ;:
Figure 20. Cyclic voltammogram for a pure Pt foil WE immersed in a Na^SO^ solution of basicity -.6.66 at 900 C. = 4 0 mV/sec. 51
8-33 m/i
5 D.O..mV|_ ___ POTE NTIAL
ir
Figure 21. Cyclic voltammogram for a basic Na0SO, melt at 900 C 2^ = 40 mV/sec. 4 52
III'a are obtained at -1600 and -1112 m V , respectively.
Therefore, peaks Ic and la in Fig. 14 correspond to the reversible reduction and reoxidation of a vanadium electroactive species in NagSO^ melts containing vanadates, without any electrochemical interference from the supporting electrolyte, NagSO^. Peaks III’c and
Ill’a constitute a redox couple.
For a fixed cathodic switching potential at -1850 mV, the anodic switching potential was made more positive to study the redox relationship, if any, between peaks I’a,
I ’c and II'c, Fig. 22. As the anodic switching potential is made more positive the current associated with peak III’c increases, as well as the current of the counter- part peak, Ill’a. The peak potentials for peaks III’c and Ill’a remain constant, so that the only effect of the changing anodic switching potential seems to be an increase in the concentration of the electroactive species reducible at -1600 mV, as indicated by the observed increase in the peak current.
The redox relationship between peaks III’c and Ill’a is again well established by the corresponding increase in the current associated with peak Ill’a. Anodic switching potentials more positive than 600 mV do not have any apparent effect on the current associated with peak III’c. 8-33 mA
O 500 mV POTENTIAL 4
Figure 22. Cyclic voltammogram from a Na„SCL melt of basicity - 6.66 at 900 C. 1 / = 40 mV/sec. 54
To investigate any possible direct redox relation
between peaks I’a and III’a, the cathodic switching
potential was reduced to -662 mV, Fig. 23. Making the
cathodic switching potential less negative causes a
decrease in the current associated with peak I'a and an
anodic shift of 30 mV in its peak potential. A steady-
state voltammogram was obtained after the second cycle,
excluding any direct redox relation with peaks III’c
and III’a. So the only effect of the cathodic switching
potential must be just chemical, and related to an
increase in the concentration of the reoxidizable
species at around 80 mV. No appreciable changes were
observed when the anodic switching potential was reduced
up to 200 mV. However, for an anodic switching
potential of 350 mV and a cathodic switching potential
of 75 mV, peak I’a almost disappears, and becomes more
evident as the cathodic switching potential is made more
negative than 75 mV, Figs. 23(h) and 23(i). The rise in
the cathodic current and the better defined anodic peak
I’a as the cathodic switching potential is made more
negative than 75 mV establishes the redox relation
between peaks I’a and I’c .
Cyclic voltammograms with anodic and cathodic
switching potentials at 750 mV and -2250 mV, respectively, were recorded at various scan rates ; the b b
fa
He
Figure 23. Cyclic voltammograms on a pure Pt foil WE immersed in a Na2S04 melt of basicity -6.66 at 900 C. 1 / = 40 mV/sec. 56
variation of peak potentials with scan rate is listed in
Table 3. Because the peak potential remains practically
constant, peak II’a is probably an adsorption peak of
the reoxidised product at -1100 mV, and peaks I’a and
I’c, and Ill’a and III’c, constitute reversible redox
couples. Peak IV’a corresponds to the reoxidation of
the product of the cathodic decomposition of Na^SO^ at
potentials more negative than -2000 mV. The chemistry
of the cyclic voltammograms presented up to this point
will be discussed later as more information on the redox
process is needed. So far it can be concluded that the
electroactive species of a NaVOg-NagSO^ melt as basic as
-9.77 undergoes a reversible redox reaction and that the
reduced species is unstable in solution. NagSO^ is a
suitable supporting electrolyte if the cathodic
switching potential is limited to -1450 mV in cyclic
voltammetry studies.
Chronopotentiometry has been used to assist cyclic
voltammetry in determining the redox mechanism of NaVO^-
NagSO^ solutions and pure NagSO^ containing NagOg.
Figure 24 shows a chronopotentiogram on a pure Pt foil
immersed in a 10 m/o NaVOg-NagSO^ solution at an open-
circuit potential basicity of - 9.77 at 900 C. When a cathodic current of 60 mA is applied between the counter
and working electrodes, a forward transition time, 57
Table 3. Peak Potential As a Function Of Scan Rate For a
Pure Foil WE Immersed In a Na^SO^ Solution Of Basicity
-6.66 At 900 C.
E I'a E I ’c E I11’a E III’c E IV ’ a mV/sec) (mV) (mV) (mV) (mV) (mV)
20 67 - 8 - 1070 - 1482 - 1920 40 92 17 - 1045 - 1495 - 1857 60 67 5 - 1095 - 1495 - 1845 80 92 17 - 1070 - 1470 - 1820 100 92 17 - 1095 - 1495 - 1820
Table 4. Chronopotentiometric Data For a Pure Pt Foil WE
Immersed In a 10 m/o NaVO^-Na^SO^ Solution Of Basicity
- 9.77 At 900 C.
i E n .. /tr, (mA) (V) 1 0
40 - 1.122 + 0.080 In {u} 1.26 4.92 60 - 1.142 + 0.078 In {u> 1.28 3.80 80 - 1.133 + 0.135 In {u} 0.7b 4.90
Note: In {u} = In [(' t 1/2 - t 1/2 )/ t 1/2 j 58 of 1.75 sec is obtained at potentials more negative than the open-circuit potential of -256 mV against a
Ag/Ag+/mullite reference electrode. A backward transition time, of .45 sec was obtained upon reversal of the current. If the current densities for the forward and reverse direction are equal, then the ratio of the forward to reverse transition times is a measure of the fraction of material reoxidized. If the oxidized and reduced species are stable in the solution, the ratio equals 3. If the product of the electrode reaction is unstable, the ratio will be increased because the reverse time will be shortened indicating that less reduced species are available in the vicinity of the working electrode to be reoxidized compared to the amount expected if the product of the reduction reaction were stable in solution. The exact value of this ratio will depend on the rate constant of the chemical reaction and on the forward transition time. If the product of the redox reaction is insoluble or adsorbed, then the reverse transition time will be lengthened and in the limit in which all product is adsorbed, the ratio equals unity. These criteria are strictly valid only if the reverse transition scan is recorded immediately after or before the forward transition time is completed. In the present case it was not possible to accomplish this because of the short Potential (V) -1.5 -0.5 - - 0.5 2.0 1.0 0 0 aS^ ouin f aiiy 97 a 90 . ple cret 6 mA. 60 current= Apnlied C. 900 at -9.77 basicity of solution Na^SO^ Figure Figure 24. 1 hoooetorm n pr P fi W imre i a 0 / NaVO-,- m/o 10 a in immersed WE foil Pt pure a on Chronopotentiogram 2 3 4 ie (sec) Time 5 6 7 8 9 10 cn co 60
extension of the forward transition time, and the
ratios in Table 4 may not be sufficiently accurate to
state a definite conclusion. However, the consistency
in the value of the ratios for various applied i b currents and for succeeding cycles at the same applied
current suggests that the reduced species are unstable
in solution, which is in agreement with the conclusions
from cyclic voltammetry studies.
For a reversible charge transfer reaction, the
following relationship between potential,E, and time, t,
is predicted when both, the oxidized and reduced species
are soluble,
E = E ^ / 4 + RT/nF { In [(TT1/2- t1/2)/ t1/2]} [3] where RT/nF has its usual significance, t is time in
seconds, rt' the transition time, and E ^ ,. is the quarter wave potential obtained when the second term on the
right in Eq. [3] is equal to zero, i.e. when t= 't' / 4 .
Eq. [3] indicates that a plot of E vs lnCCT?^2- t ^ 2)/ 1 /2 t ] should yield a straight line whose slope is RT/nF; from this slope, the number of electrons transfered can be obtained since the other terms are all well known.
Figure 25 shows such a linear relationship for the chronopotentiogram shown in Fig. 24. From the slope of this straight line the number of electrons transfered, n, of 1.28 was calculated. 61
The irreversibility of the redox reaction was also 1/2 1/2 evaluated by plotting E vs In ([t - t ) according
to the following equation,
E = E ’ + RT/ nF In (TT1/2 - t 1/2) [4]
according to Eq. [4] a plot of E vs In ( ' t f ^ 2 - t ^ 2 )
should yield a straight line whose slope will allow us
to calculate the product n. Figure 26 shows such a plot
for the same potential range shown in Fig. 25. The
adjustment to a straight line is definitely well
established for the reversible case in Fig. 25 and not
for the irreversible case in Fig. 26.
From the chronopotentiometry, the electroactive
species in NaVOg-Na^SO^ melts of basicity -9.77 (on the
log a scale) undergoes a reversible redox reaction
involving 1.28 electrons, nominally one, and the
product of the reduction reaction is unstable in
solution.
Figure 27 shows a chronopotentiogram for a pure Pt
foil immersed in a NagSO^ molten bath with an open-
circuit potential basicity of -6.66 at 900 C. When an
anodic current of 40 mA is applied between the counter
and the working electrode, a forward transition time
(labelled A1) is obtained. Upon reversal of the current, only one reverse transition time (labelled Cl)
is obtained. A second reversal of the current reveals Potential (V) -1.08 -1.16 0 0 . 1 - -1.24 - 1.2 o te ahdc rniin ie n i. 24. Fig. in time transition cathodic the for Figure Figure - I L I J 1.0 o Experimental Experimental o • Best fit Best • - 25. lt f hoooetoerc aa codn to according data chronopotentiometric of Plot 0.8 I l I I l I I I I l l I i I i l i I i 0. 0. 0. .2 -0 .4 -0 .6 -0 In 1/2t _ n i t 02 0.4 0.2 0 g 3 Eg. 0.6 CO 5 0 Potential - -1.08 4 2 \ - -1.16 1.00 12 10 0. 06 04 02 0 -0.2 -0.4 -0.6 .8 -0 -1.0 -1.2 e r u g i f q 4 i n Fi 24. . ig F in e tim n o i t i s n a r t c i d o h t a c e h t r o f 4 Eq. O Experimental Experimental O • Best fit Best • 6 2 L J . a accordi to g in d r o c c a ta a d c i r t e m o i t n e t o p o n o r h c f o t o l P In In ( t 7 - t1/2) t 172 - o I I L I I I I J 5 0 u Potential (V) - - 2.0 1.0 1.0 0 0 et f aiiy 66 a 90 . ple cret 4 mA. 40 current= Applied C. 900 at -6.66 basicity of melt Figure 27. 27. Chronopotentiogram on a pure Pt foil WE immersed in a Na?S0. Na?S0. a in immersed WE foil Pt pure a on Chronopotentiogram 10 20 Time (sec) Time 30 40 50 05 65 an additional anodic transition (labelled A2) at potentials more negative than the previously recorded A1 anodic transition time. The anodic transition times A1 and A2 do not correspond to the reoxidation in two stages of whatever species might have been reduced in the cathodic transition Cl, since the anodic transition time A1 was initially recorded without previously polarizing the working electrode in the cathodic direction. The anodic transition A1 obviously corresponds to the oxidation of the species recorded in the cyclic voltammogram as peak I’a in Fig. 20. Figure 1/2 28 shows a linear relationship between E and In [(^ 1/2 1/2 -t )/ t ] as predicted for a reversible redox process according to Eq. [3]. The number of electrons transfered, calculated from the slope of Fig. 28, equals
1.32 with Ej-/ a = -.124 V. The transition time corresponding to the reduction of the species oxidized at A1 could not be detected, because it was probably masked by the quick rise in potential upon reversal of the current. Figures 29 and 30 correspond to the plots of potential E vs the time relationship of Eq. [3] for the cathodic transition Cl and the anodic transition A2, respectively. The linear relationships suggest reversibility of the redox process. The number of electrons transfered, calculated from the slope of Fig. T -4 0
> S -80 o
fi I £ -120
O Experimental • Best fit -160
i L j L J L 1.2 -1.0 -0.8 -0.6 -0.4 -0.2 0 0.2
V/2 _ tV2 In t 1/2
t i g u r e 2 8 . Plot of chronopotentiometric data according to Eq, For the anodic transition time A1 in Fig. 2 7 .
CD CD Potential (V) -1.24 -1.48 -1.32 -1.40 f o r t h e c a t h o d i c t r a n s i t i o n tim e Cl in F ig . . ig F in Cl e tim n o i t i s n a r t c i d o h t a c e h t r o f e r u g i f - 1.0 o Experimental Experimental o • Best fit Best • 9 2 . - a accordi t< g in d r o c c a ta a d c i r t e m o i t n e t o p o n o r h c f o t o l P 0.8 - o 0.6 In t1/2_ 04 0.2 - -0.4 1/2 tU2~ 7 2 . 0 9 q 3 Eq. 0.2 02 -0 - 1.0 o Experimental - 8 • Best fit
- 1.1
03 5 - 1.2 £3 &o Ph
-1 .3
-1 .4 -3 .0 - 2.0 - 1.0 0 1.0
^ n - t in In t 1/z
..Figure 30 . Plot of chronopotentiometric data according to Eq. 3 for the anodic transition time A2 in Fig. 27 .
0 3 CO 69
29 equals .70 and E\^/A- -1.281 V. The slope of the line T r / 4 in Fig. 30 yields a value of n equal to 1.32 and E- 'e y 4 -1.282 V. The cathodic and anodic transitions, Cl and
A2, correspond to the reduction and oxidation of the
electroactive species of the redox couple labelled III’c
and III'a in the cyclic voltammogram in Fig. 20. This
second redox process involves a nominal number of
electrons transfered equal to one.
In earlier electrochemical studies by Park (32) on
pure Na^SO^ an autocatalytic reaction mechanism was
suggested for chronoamperometric results. The same
electrochemical technique was used in the present study
for Na^SO^ melts at various basicities. For NagSO^
melts of basicity -6.66, the open-circuit potential was
-707 mV ; a potential step of 100 mv more cathodic to
the open- circuit potential was applied to the working
electrode and the current-time response was recorded,
Fig. 31. The electrolysis time was set at 60 sec in
order to avoid significant effects from convection. The
response in Fig. 31(a) shows a continuous decay of the current with time with no apparent plateau in current as the one reported for an electrolysis time as short as 20 sec (32). According to an autocatalytic mechanism, the reduced species will be reoxidized via a chemical reaction creating a continuous supply of electroactive Current density (mA/cm2) 4 0 5 3 2 1 1 0 ia ptnil tp () 87 V () 10 mV. 1000 - (b) C. mV, 900 -807 at (a) : -6.66 step basicity of potential melt Final Na?S0, a in Figure 31. Chronoamperograms on a pure Pt foil WE immersed immersed WE foil Pt pure a on Chronoamperograms 20 Time (sec) 40 60 80 70 71
species in the immediate vicinity of the working
electrode. A plateau in current with time would be
observed since the electroactive species would not need
to diffuse from the bulk of the electrolyte but
rather they would be readily available at the working
electrode. Successive applied potential steps more
cathodic to the o.c.p revealed a similar decay in
current as predicted by the Cottrell equation, Fig.
31(b). Therefore, some other factors such as stray
currents because of inadequate shielding of the
electrochemical chamber, or convection, must have been
responsible for the observed current plateau.
Chronoamperometry studies were also conducted in NaVOg-
NagSO^ solutions of basicity -9.77 . A continuous decay
in current with time was again obtained. The final
potential steps used in NaVOg-NagSO^ solutions were the
same as the ones used in pure NagSO^. The maximum
currents for each potential were up to 5 times larger
than the ones obtained from NagSO^. A current-potential
curve drawn from chronoamperometric measurements is
shown in Fig. 33. The plot in Fig. 33 resembles a polarographic wave. The figure also shows that the electrochemical response for potential steps up to -1400 mV obtained from NaVOg-NagSO^ solutions is provided mainly by an electroactive vanadium species reducible between -600 and -1400 m V . The contribution from the Current density (A/cm2) -0.5 -0.4 - -0.3 - 0.2 0.1 Figure 32. Chronoamperograms on a pure Pt foil WE immersed immersed WE foil Pt pure a on Chronoamperograms 32. Figure ia ptnil tp () 15, b -30 () -1450, (c) -1350, (b) -1250, (a) step: potential Final n 1 mo ouin f aiiy 97 a 90 C 900 at -9.77 basicity of solution ^ O S ^ - ^ - O V a N m/o 10 a in 2 4 6 80 60 40 20 0 d -50 () 15, f -70 mV. -1750 (f) -1650, (e) -1550, (d) Time (sec) 72 73
background, i.e., Na^SO^, within the mentioned
potential range is proved to be negligible.
For a reversible redox reaction, the following
relationship between potential and current is expected
to hold,
E= E 1/2 + RT/nF In [(il - i )/ i ] [5]
where il is the limiting current and E ^ 2 *s the half
wave potential, and the other terms have their usual
significance. From Fig. 33, after substracting the
background, a limiting current of 0.2116 A/cm2 is
obtained for the reduction of the vanadium species.
Figure 34 shows a plot of E vs ln[(il - i)/i] according
to Eq. [5]. A straight line is obtained, and from its
slope the calculated number of electrons transfered of
.88, nominally one, is calculated. The intercept at
i=id/2, gives a half-wave potential, E ^/2’ “1096 ■
The half-wave potential is related to the standard redox
potential, E’o, and the ratio of the diffusion
coefficients of the reduced to oxidized species
according to Eq. [6],
E 1/2= E’o + RT/nF In DR / DO [6]
Since the ratio of diffusion coefficients in Eq. [6] is nearly unity in almost any case, E ^ 2 is usually a very good approximation to E ’o for a reversible couple (19). Current density (A/cm 2) 0.35 0.30 0.15 0.20 0.25 0.10 0.05 . -. -. -. -. -. -. -. -0.6 -0.8 -1.0 -1.2 -1.4 -1.6 -1.8 -2.0 2.2 t 0 C 24 2 C. 900 at data obtained from basic NaV0,-NaoS0, and basic Na„S0, melts melts Na„S0, basic and NaV0,-NaoS0, basic from obtained data figure 3 3 . Superimposed i vs E curves from chronoamperometric chronoamperometric from curves E vs i Superimposed Potential (V) Potential V0— — 2 a20 4—N 0 ^ a 03—N aV N □ aS04 02 20 a N 4- 0 Na2S o -c Potential (V) - - -0.9 - - 1.2 1.0 0.8 1.1 Figure Figure 08 04 04 . 12 . 2.0 1.6 1.2 0.8 0.4 0 -0.4 -0.8 34 34 Crnaprmti dt potd codn,t E. 5 Eq. to according, plotted data Chronoamperometric . I n [ ( i d — i ) / i ] 7b 76
Figure 35 shows potentiodynamic polarisation curves
for a pure Pt foil WE immersed in Na^SO^ and in
Na^SO^-NaVOg melts containing equivalent amounts of
Na^Og- The polarization curve for NaVOg-Na^SO^ melts
shows two stages of activation polarization . The first
stage, observed between -200 and -600 m V , corresponds to the reduction of electroactive species in the NagSO^
supporting electrolyte. A reduction reaction takes place in the supporting electrolyte at potentials more anodic to the open-circuit potential of NagSO^ melts of basicity -6.66, as indicated by the cyclic voltammograms previously discussed. A Tafel slope of -360 mV/decade suggests a number of electrons transfered of 1.29 for an assumed value of ot=.5. The second stage of activation polarization certainly corresponds to the reduction of vanadium species. This activation polarization stage is followed by a concentration polarization stage as indicated by a limiting current (diffusion-control) of
.1519 A/cm2 in the potential interval -1300 to -1800 mV.
A Tafel slope of -323 mV/decade suggests a number of electron transfered of 1.44 for an assumed value of
at - . 5 .
Figure 36 shows an anodic polarization curve for a pure Pt foil WE immersed in a NagSO^ melt of basicity
-6.66. A concentration polarization stage is observed in - 0.2 NaV03-N a2S04-N a20
- 0.6
-1.4
- 1.8
0.001 0.01 0.1 1 Current density (A/cm2)
Figure 35. Superimposed cathodic polarization curves from basic NaV03-Na2S04 and basic Na2S04 melts at 900 C. U = 1 mV/sec. 0.12
0.08
> 0.04
& 0.00
-0.04
-0.08 0.001 0.01 0.1 1 10
Current density (A/cm2)
Figure 36 • Anodic polarization curve from a basic NaoS0„ melt at 900 C. i/ = 1 mV/sec. 1 q
CD 79
the potential interval -300 to 600 mV with a limiting
current of .1433 A/cm2. This concentration polarization
stage corresponds to the same electroactive species that
is oxidized at a potential of 100 mV, labelled as peak
I’a in the cyclic voltammograms in Figs. 22 and 23. The
observed limiting current which is determined by the
diffusion of the electroactive species to the WE surface
suggests that peak I’a is a diffusion-control peak which
corresponds to the oxidation of an electroactive species present in a limited concentration in the NagSO^ melts ( disregarding the possibility of peak I’a being an adsorption-control peak). At potentials more positive than 600 mV, an anodic decomposition of the melt is observed.
Reaction Mechanism
1. NagSO^ melts of basicity -6.66 in the log aNagO scale
Figure 37 shows the stability phase diagram for the
Na-S-0 system at 900 C. Table 5 lists the thermodynamic data at 900 C . Unit activity was assumed for all condensed species in equilibrium . The diagram shows the stable, phases in equilibrium with an Og-SOg gas mixture as a function of the oxygen potential, log P Og, and the acidity of the melt, log P SO^. By applying the mass action law to the equilibrium decomposition reaction for sodium sulfate given as , 10 Log Pq2 -30 -5 Na \ Basicity trace o a w o a a i n o c r i z a on we t P d e t n i a p d e z i r a l o p a on e c a r t y t i c i s a B . e r u g i f ube i re i N2S4 0 C. 900 t a 6 6 . 6 - y t i c i s a b f o t l e m S04 Na2 a in ersed m im e b tu \ NaO \ 4 8 4 0 4 - \ \ N&20 \ N&20 2 °2\ /n I I K X_J L\_l J _ X I K I \ I » I I 5 -L o g a N a o \ 2 1 \ \ 20 15 10 0.5 -1.0 -1.5 -2.0 \ 7 ------Log P a2^2^7 N 0 4 N 24 20 0 S 7 -- \ \ 0.5
E(mV)Ag/Ag o a> 81
Na2S04 v Na20 + SOg [7]
a direct relationship between the partial pressure of
sulfur trioxide (PSOg) and the activity of sodium oxide
can be obtained for any given temperature through the
following expression,
log K= log P SOg + log a NagO [8]
where the activity of NagSO^ is considered to be unity
and the constant K is the equilibrium constant for
reaction [7] calculated from the standard Gibbs energy
change for reaction [7] at any temperature. Therefore,
a Na20 can be used as well to describe the phase
equilibria. The activity of NagO is used as the
criterion for basicity of the melt, as pH is used for
acidity in aqueous solutions. This diagram resembles
the Pourbaix diagram, E vs p H . As for a Pourbaix
diagram, the most thermodynamically stable compound can
be identified for a known oxygen potential and basicity
of the melt.
In the cyclic voltammetry experiments performed in
this system, the potential of the working electrode was measured with respect to the Ag/Ag+/mullite high
temperature reference electrode. This potential depends
upon the basicity of the melt; therefore, dashed
isopotential lines were added to the stability diagram.
If the basicity of the system is known, the chemistry of 82
Table 6. Thermodynamic Data At 900 C
Compound G ( KCal/
Na2S04 - 216.517
- 275.181 Na2S2°7 - 172.808 Na2S03 - 62.186 Na2°2 - 24.717 Na02 Na2° - 60.925
Na2S - 68.533
- 255.359 V2°5 - 247.428 V2°4 - V2°3 219.612 NaVOg - 197.528
Na.VO. - 3 4 303.859 VOSO. - 179.850 4 - 136.283 V5S8 so3 - 63.425 83
the potential scan in cyclic voltammetry measurements,
between limiting anodic and cathodic potentials, can be
traced on the diagram;this enables one to identify which
stable compounds may be used in the interpretation of
the probable reaction mechanism. However, this diagram
is just an auxiliary tool in the interpretation of the
results from electrochemical measurements because of the
assumptions involved in its construction and the lack
of thermodynamic data of some other compounds that may
be present. Nevertheless, the diagram gives a general
overview about which compounds could be present.
The basicity trace on Fig. 37, was simultaneously
registered as a cyclic voltammogram, was recorded on a
Pt painted WE on a zirconia tube immersed in a Na^SO^ melt of basicity -6.66. The stability areas for the
predominant anionic species in NagSO^ have been depicted on the diagram. According to the trace in Fig. 37 the
open circuit potential lies on the superoxide ( Og ) dominance field close to the boundary line with the
oxide (02 ) stability field.
In order to obtain a NagSO^ melt of basicity -6.66,
Na^Og is mixed at room temperature with Na^SO^ and then heated to the operation temperature of 900 C. Na^Og is expected to decompose at 460 C to increase the oxide concentration and oxygen content in the melt according 84
to the following reaction,
Na2°2 * Na2° + 1/2 °2 [9] Reviews of studies in nitrate melts (33,34) suggest the
ability of the melt anion NO^ to oxidize the oxide ions
to form peroxides and superoxides ions in the absence of
oxygen according to the following reactions,
o2~ + Non" o„2' + n o „~ rioi d *------2 2 0 22” + 2N03~ * 2 0 ~ + 2N02' [11]
The formation of significant amounts of nitrite ions is
suppressed if oxygen is present and even less oxide ions
remain at equilibrium due to the reactions,
0 2~ + 1/2 0 2 , "f 022' [12]
0 22" + 0 2 ; = ± 2 02' [13]
Such a mechanism is suggested to operate in other melts
e.g. halides, sulfates and carbonates (35). Stern et al
(36) and Deanhardt et al (37) support the stability of
Na2C>2 at a temperature of 839 C by chemical, manometric
and potentiometric techniques.
If such a mechanism could also operate in sulfate melts, superoxide and peroxide ions would be the predominant oxide species according to the following reaction scheme,
0 2~ + S042“ * S032~ + °22” [14]
022" + 2S042'^i= ± 2S032~ + 202~ [15] and in the presence of dissolved oxygen released by the 85
decomposition of NagOg,
0 22‘ + 0 2 v * 2 0 2" [16]
The trace in Fig. 37 suggests superoxide ions as the most stable anionic species at open-circuit potential.
Cyclic voltammetry and chronopotentiometry suggest a one electron reduction reaction in the vicinity of -1500 mV.
Oxide ions can not be further reduced and since 02 ions are suggested as the predominant species the following reduction reaction is proposed,
0 2" + e- > 022" [17]
Upon reversal of the potential in cyclic voltammetry, or upon reversal of the current in chronopotentiometry, a one electron oxidation reaction is detected in the vicinity of -1000 mV. Cyclic voltammetry and chronopotentiometry suggest that the redox reaction corresponds to the reoxidation of the species reduced at around -1500 mV. The redox couple was also concluded to be reversible,
0 22' ------* 0 2" + e- [18]
A second anodic peak in cyclic voltammetry ( a second transition time in chronopotentiometry) in the vicinity of 100 mV suggests a reversible one electron transfer reaction that is suspected to be the reoxidation of superoxide ions to oxygen,
0 2‘ ----- > 0 2 + e- [19] The presence of Og ions as the predominant anionic
species in the melt is further supported by the cyclic
voltammograms in Fig. 23, the chronopotentiogram in Fig.
27, and the polarogram in Fig. 36. From Fig. 23, peak
I’a at 100 mV is well defined at a cathodic switching
potential of -650 mV, indicating that it is essentially
independent of the redox process taking place at
potentials as negative as -1500 mV. As was previously
discussed, the only effect of a cathodic switching
potential more negative than -650 mV is to increase the
peak current. The peak current seems to be enhanced
when the cathodic switching potential is further
increased beyond -2000 mV, a potential at which the
cathodic decomposition of the melt takes place. The
cathodic decomposition reaction of NagSO^ has been proposed to increase the basicity of the melt by generating oxide ionic species of which could be produced. This reaction could account therefore for the
increase in the current of peak I’a, which is directly related to the concentration of the oxidizable species and the slight increase in basicity. The chronopotentiogram in Fig. 27 indicates that an electroactive species can be oxidized in the vicinity of
100 mV without the melt being previously polarized in the cathodic direction. The same transition time is lengthened in a second cycle after the WE has been 87
polarized up to -2000 mV. This observation is in agreement with the results from cyclic voltammetry . The polarogram of Fig. 36 shows an anodic limiting current in the potential interval -200 to 600 mV, indicating the oxidation of an electroactive species present in the melt before the anodic decomposition of the melt takes place.
Figure 22 established the relation between peaks I’a and III’c. Peak III’c was not isolated from peak I’a since the anodic switching potential was held at -175 mV, on the rising portion of peak I’a. However, the scan was always started from 0 mV, and a reduction peak occurred at -100 mV in the first scan, followed by peak
III’c ( Fig. 21). Succeeding cycles showed a less well defined peak III’c with a much lower peak current compared to the first scan. The anodic peak was reduced proportionately Steady-state voltammograms were obtained after four cycles. The presence of peak III’c,
even though some oxidation of Og took place, indicates that Og ions are present in sufficient concentration in the bulk of the solution. As discussed in Fig. 22, the current associated with peak III’c increases as the anodic switching potential is increased beyond the peak potential for peak I'a. As the anodic switching potential is extended, more 0 ^ oxidized at peak I'a can 8 8
be reduced at peak I ’c. Together with the 0^ ions
diffusing from the bulk of the solution, this results in
an increase in the electroactive species available to be
reduced at peak III’c, which in turns results in a
higher peak current at peak III’c. The peak current
associated with peak III’a is proportionally increased,
confirming the redox relationship between peaks III’c
and III’a and the identity of the species involved. A
steady-state voltammogram is obtained for an anodic
switching potential of 600 mV at which the anodic
decomposition of the melt begins .
The proposed reaction mechanism is as follows, CM i O + 1/2 0 2 °22 + s o 32- + s o / .... o 22- CM 1 0 1
CM peak III’c °2 + e " + e- peak III’a °22 °2 e- peak I ’a °2 °2 + peak I ’c °2 ' C °2
Fang and Rapp (38) performed similar electrochemical studies in basic melts at 900 C. Their cyclic voltammogram resembles the one in Fig. 21 for an anodic and a cathodic switching potentials of -175 and -1850 mV, respectively. In their voltammetric studies,the anodic switching potential was limited to -300 mV and 89
they did not detect the redox couple provided by peaks
I ’c and I'a in the present studies. Their
chronopotentiogram showed a reduction reaction in a
potential interval similar to the one at which the
cathodic transition Cl is recorded in Fig. 27. However,
the calculated number of electrons transfered varied
from 3 to 4 . As in the present study, the authors
proposed superoxide ions as the predominant reducible
ionic species in NagSO^, but their reaction mechanism
differs from the one here proposed.
Figure 38 is a dynamic polarization curve on a Pt foil
WE immersed in a Na2S0^ melt of basicity -6.66, at 900 C
and at a scan rate of 1 mV/sec. Activation polarization
is observed in the potential interval -1000 to -1500 mV with a Tafel slope of -500 mV/decade. Fon an assumed value of = 0.5,the number of electrons transfered equals 0.93, nominally one. This activation polarization stage corresponds to the reduction reaction labelled as peak III’c in Figs. 20 to 22 and the cathodic transition Cl in Fig. 27. The number of electrons transfered confirms the results from chronopotentiometry . Concentration polarization follows the activation stage. A second activation polarization stage is observed in the potential interval -2000 to
2250 mV with a Tafel slope of -250 m V . For an assumed Potential (V) - -3.0 - 2.0 1.0 2.0 1.0 mesd n No0 ml o bsct - .6 t 0 C. WE 900 at foil Pt 6.66 - pure a on basicity of curve melt NaoS0. polarization a in Cathodic . immersed 8 3 Figure = V x1" 1 0 001 .1 . 1 10 1 0.1 0.01 0.0011 1 4 x 10"5 x 10‘ m/e. 4 2 mV/sec. 1 C u r r e n t d e n s i t y (A/cm2 ) 90 91 value of a =.5, the calculated number of electrons transfered is 1.86, nominally 2. At this potential range the cathodic decomposition of the melt takes place. The large cathodic current at -2000 mV can only
2 - be generated from the reduction of the anion SO^ ,
S042" + 2e- * S02 + 2 0 2~ [20]
Isobars for the partial pressure of SOg has been added to the phase diagram in Fig. 37. According to the basicity trace, for a potential of -2000 mV, the partial pressure of SOg is close to 1 Atm and the generation of
SOg can be considered as a product of the cathodic decomposition of NagS04 .
The anodic peak labelled IV’a in Fig. 20 has been assigned in the literature (38,39) to the oxidation of the products of the cathodic decomposition of NagS04 .
2 - 2 - It can represent the oxidation of O to 0g
2 o 2_ ------* 0 22~ + 2e- [21]
2 - Og ions can be further oxidized to Og which in turns can oxidized to molecular oxygen. The generation of
2 - Og explains the enhancement experienced by peaks III’a and I’a as the cathodic decomposition of the melt is allowed.
2 - 2 - SOg can conbine with O g to form SO^ ,
S ° 2 + O g 2 ' --- k S O , 2 ' [22] 92
2. 10 m/o NaVOg-NagSO^ melts of basicity -9.77.
Zhang and Rapp (16) measured the basic and acidic
solubility of CeOg and other potential thermal barrier oxides in NagSO^-SO m/o NaVO^ solutions at 900 C. The vanadate anion significantly stabilized a Ceg(V04)^ acidic solute, raising the acidic CeO^ solubility as much as three orders of magnitude compared to the solubility of the same oxide in pure NagSO^. Contrary to the currently accepted mechanism for acidic hot corrosion (15), the oxygen anion is strongly complexed to form vanadates anions, but the effect is to increase the melt basicity , not reduce it. This result is expected because the anion of the acidic solute would be 2- 3- changed from SO^ to VO^ . Therefore, the melt is 3 _ buffered by an equilibrium between VO^ and VO^ ionic species. Figure 39 shows a phase stability diagram for the V-Na-S-0 system at 900 C. Dashed isopotential lines with respect to the Ag/Ag+/mullite electrode have been added. Isoactivity lines for the 10 m/o NaVOg-NagSO^ solution have also been drawn on the phase diagram to show the displacement of the stability fields for the various species in equilibrium with NaVO^. The basicity trace was recorded as a cyclic voltammogram was registered on a Pt painted WE on a zirconia tube immersed in a 10 m/o NaVOg-NagSO^ solution of basicity
-9.77. From Fig. 39, the basicity at the open-circuit voso
-2.0 -1.5 -1.0 -0.5
Figure 39. Basicity trace on a polarized painted Pt WE on a zircania tube immersed in a 10 m/o NaVOo-Na^SO, solution of basicity - 9.77 at 900 C. 3 2 4
CD CO 94
potential suggests VOg as the stable vanadium ionic
species in solution. From cyclic voltammetry and from
chronopotentiometry, the vanadium electroactive species
undergoes a one electron reversible redox process.
With NaVOg as the stable species and one electron
transfered, the reductiuon reaction must be as follows,
V03~ + e- k V032" [23]
from cyclic voltammetry and chronopotentiometry, the
reduction reaction is followed by a chemical reaction:
V032" + 0 2" “ V043" + 1/2 0 2 [24]
V043_ ---- * V03" + 0 2~ [25]
Since the redox process is suggested to be reversible,
in the anodic scan ,
v° 32" k VO," + e- [26]
Reactions [24] and [25] allows the basicity of the melt
to remain relatively constant.
Electrochemical Studies in NaV03~Na2S04 Solutions Under
02 Gas
Figure 40 shows a series of cyclic voltammograms, at various scan rates, on a pure Pt foil working electrode
immersed in a 10 m/o NaV03-Na2S04 solution under 02 gas
at 900 C. As the potential is scanned toward more negative values with respect to the open-circuit potential of -20 mV ( relative to a Ag/Ag+/mul1ite high 95 temperature reference electrode), one cathodic peak
(labelled Ic) is recorded at about -1350 mV. Upon reversal of the potential an anodic peak (labelled la) is recorded at about -600 m V . As the scan rate is increased, shifts for both peaks la and Ic in the cathodic direction are observed. A variation of the peak potential with scan rate is a criterion for irreversibility indicating a sluggish rate for the electrode transfer. However, kinetics complications arising from a chemical reaction following the redox process would also result in similar observations on the recorded voltammogram. In order to determine the redox relationship between peaks la and Ic, cyclic voltammograms were recorded at various cathodic switching potentials, Fig. 41. The anodic switching potential was fixed at 280 mV. At a cathodic switching potential as negative as - 825 mV a cathodic current response is observed. In cyclic voltammetry, a rise in the current is an indication of an electrochemical reaction taking place at the working electrode surface.
For a cathodic switching potential of -912 mV, upon reversal of the potential, a well defined anodic peak is observed at - 625 mV. As the cathodic switching potential is made more negative, a continuous rise in the cathodic current is observed as well as a better defined peak with a higher current for the associated 96
oc 2 ;• e m
5 00 mV
Figure 40. Cyclic voltammograms on a pure Pt foil WE immersed in a 10 m/o NaVO-j-Na^SO^ solution of basicity -11.7 at 900 C under 0^ gas. P = mV/sec. 9 7
Table 5. Variation Of Peak Potential With Scan Rate For
Cyclic Voltammograms On a Pure Pt Foil WE Immersed In a
10 m/o NaVO,-Na^SO. Solution Of Basicity -11.70 At 900 C
E t la E Ict (mV/sec) (mV) (mV)
10 - 588 - 1338 20 - 612 - 1350 30 - 625 - 1350 40 - 625 - 1375 50 - 625 - 1375
Table 7. Anodic To Cathodic Peak Current Variation With
Scan Rate For Cyclic Voltammograms On a Pure Pt Foil WE
Immersed In a 10 m/o NaVO^-NagSO^ Solution Of Basicity
- 11.70 At 900 C.
icp =icp iap isp iap/acp iap o o o (mV/sec) (mA) (mA) (mA) (mA)
10 24. 38 6. 25 23. 75 0. 82 19. 86 20 28. 75 8. 25 26. 87 0.83 23. 75 30 31. 25 10. 75 27. 50 0.85 26. 77 40 35. 62 13. 50 29. 12 0. 86 30. 68 50 38. 12 16.00 30 . 62 0.90 34. 13 98 anodic step, indicating that more reduced species are readily available in the immediate vicinity of the working electrode to be reoxidized. For a cathodic switching potential of -1400 mV, a well defined pair of redox peaks are recorded at -1375 and -612 m V . These peaks la and Ic clearly constitute a redox couple.
In cyclic voltammetry the variation of the anodic to cathodic peak current ratio with scan rate is used as a mechanistic criterion. In the present study it is not possible to experimentally determine a well defined current base line with respect to which the peak current associated with the anodic peak la could be measured.
As in the preceding discussion a theoretical model proposed by Nicholson was used to estimate the peak current ratios. Table 7 lists the results according to
Equation [2]. A plot of iap/icp as a function of scan rate is shown in Fig. 42. The trend of the data suggests a reversible redox reaction followed by a an irreversible chemical reaction which consumes and transforms the reduced species into a non electroactive species. At relatively high scan rates the currrent peak ratio will tend to unity and the effect of the following chemical reaction on the anodic response will become negligible and the redox peaks will appear at their normal potentials. a: 12 5 nr A ^ f .. ...
; '• ‘S O O i n V ^- POTENTIAL
figure 41 . Cyclic voltammoqrams on a pure Pt foil WE immersed in a 10 m/o N a V O ^ ^ S O ^ solution of basicity
-11.7 at 900 C under O^gas. V = 40 mV/sec. 1 0 0
The reversibility of the redox process can be
evaluated by plotting the peak current versus the square
root of the scan rate. According to the Randles-Sevcik
equation, for a diffusion-control redox process the peak
current varies linearly with the square root of the scan
rate. Deviation from linearity arises because of
adsorption of the reactant or because of a preceding
chemical reaction. Figures 43(a) and 43(b) show such
plots for peaks Ic and la, respectively. The linear
relationship suggests difussion-control, i.e,
reversibility of the redox process. As in the preceding
discussion, a close examination of peak lc reveals a
sharp decay in current once the peak current has been
reached and an increase in the symmetry of peak Ic with
increasing scan rate. Both features are characteristics of adsorption phenomena. However, since no prepeak is
observed, strong adsorption of the reactant is ruled out but weak adsorption of the reactant takes place, in which case, the adsorption-control peak is observed at the same potential at which a purely diffusion-control peak might be observed. Under this circumstance the diagnostic plot in Fig. 43(a) will not detect deviations from diffusion control because of adsorption phenomena.
Another reaction mechanism criterion is the variation of the ratio of peak current to the square root of the 1 0 1
Table 8. Criteria for determining the reaction mechanism from cyclic voltammetry
I. Reversible Charge Transfer:
0 + n e ~ — ? R II. Irreversible Charge Transfer: 0 7
.VMI 0 + n e” * R 06 VII III. Chemical Reaction Preceding 03 A Reversible Charge transfer;
04 Z 5 = 5 0 ill 03 0 + n e” R OOl •O IV. Chemical Reaction Preceding An V Irreversible Charge Transfer: Z 5 = 5 0 0 + n e" --- * R V. Charge Transfer Followed By A Reversible Chemical Reaction:
0 + n e" 5 =5 R R ■> Z VI. Charge Transfer Followed By An
I,VII Irreversible Chemical Reaction: 0 + n e" R 06 R ------' Z O Ol too VII. Catalytic Reaction With V Reversible Charge Transfer: 0 + n e~ R R + Z ----- * 0 vm. Catalytic Reaction With Irreversible Charge Transfer:
0 + n e" »• R R + Z------0 iap/icp 0.84 0.92 1.00 0.76 2 4 60 40 20 0 ih cn ae o cci vlamgas n i. 0 . 40 Fig. on voltammograms cyclic for rate scan with Figure Figure 42 42 . Variation of anodic to cathodic peak current ratio ratio current peak cathodic to anodic of Variation . v (mV/sec) 2 0 1 103
scan rate with scan rate, Table 8. The trend of the
data plotted in Figs. 44(a) and 44(b) suggests a
reversible redox reaction followed by an irreversible
chemical reaction. Since the trend was evaluated
considering a reduction reaction, the sequence of steps
for peak la is reverse: an irreversible reaction
precedes the reversible redox process. The analysis of
the cyclic voltammetry studies reveals that peaks la and
Ic constitute a reversible redox couple and that the
reduced species undergoes an irreversible chemical
reaction that produces an apparent non-electroactive
species.
To indisputedly state that the redox couple, la and
Ic, represents an electrochemical response from an
electroactive vanadium species present in NaVOg-NagSO^ melts under 0 ^, cyclic voltammograms were recorded on a pure Pt foil WE immersed in a pure NagSO^ melt under oxygen gas. The current sensitivity scale and the working electrode area were the same as the ones used in the NaVO^-NagSO^ studies. Figure 45 shows a series of cyclic voltammograms for various cathodic and anodic switching potentials. On cyclic voltammograms (a), (b), and (c), the cathodic switching potential (CSP) was fixed at -200 mV, and the anodic switching potential
(ASP) was varied from 575 to 975 mV to study its effect 104 b. 50
40
30
20
1011_____ i_____ i i_____ i_____ i_____ 2 4 6 8 V1/ 2
c. 40
32
24
16
i i i i j i __ 2 4 6 8 V 1/2
figure 43 . Variation of peak current with Square root of scan rate, (a) cathodic peak £b) anodic peak. 105
8.0
(VI 7.0
Q_ O 6.0
5.0 v b. 6.5
6.0
Q_ 5.5 CO 5.0
4.5 0 20 40 60 v
Figure 44. Mechanism diagnostic criteria for the cathodic and anodic peaks in the cyclic voltammograms of Fig. 40. on the peak labelled I’c. As the ASP was made more
positive, the current associated with peak Ic increased,
indicating that more oxidized species were available for
reduction in the immediate vicinity of the WE . An
increase in the ASP results in a higher anodic current.
Even though no anodic peak was detected, some
electroactive species present in the NagSO^ must be
oxidized at potentials more positive than 575 mV. The
anodic peak can be masked by the sharp rise in the
anodic current because of the decomposition of the melt
at such positive potentials. Peak I’c could correspond to the reduction of any electroactive species generated
during the anodic decomposition of the melt. In the
succeeding cyclic voltammograms, the ASP was fixed at
57 5 mV while the CSP was varied toward more negative values than -200 m V . For a CSP as negative as -850 mV,
an ill- defined anodic peak (labelled I’a) is observed at -575 mV. As the CSP is made more negative no appreciable cathodic peaks besides peak I’c are observed even though peak I'a becomes better defined and shifts toward more anodic potential values. However, no significant increase in the current of peak I’a is observed. Up to a CSP of -1625 mV no additional reduction reaction in the NagSO^ melt takes place.
Direct comparison of the cyclic voltammograms from pure
NaoS0. melts with those obtained from NaV0„-NaoS0. melts 2 4 2 2 4 107
ttr
Ira
Fiqure 45. Cyclic voltammograms for various cathodic switching potentials recorded on a pure Pt foil WE immersed in a pure melt of basicity -10.07 under 0£ gas at 900 C. V = 40 mV/sec. 108
undoubtedly established that peaks Ic and la in Figs. 40
and 41 corresponds to the reversible reduction and
oxidation of electroactive vanadium species.
For a CSP of -1800 mV in Fig. 45, a significant
cathodic response is observed. Upon reversal of the
potential, a second anodic peak (labelled II’a) at -1150
mV is obtained. Further increase in the CSP results in
a better defined peak II’a. No significant changes are
observed on peak I’a. For a CSP of -2200 mV, an
additional anodic peak (labelled III’a) is recorded at -
1900 mV, Fig. 46. A significant increase in the current
associated with peaks I’a and II’a is observed. After
repeated cycles, an ill-defined cathodic peak (labelled
II'c ) is recorded at - 1475 mV. The increases in the currents associated with peaks II’a and I’a are directly proportional to the concentration of the species
reoxidized at those potentials. These electroactive
species might as well be generated from the cathodic decomposition of the melt at potentials as negative as -
1750 mV. The electrochemical reaction mechanism for pure NagSO^ melts under O ^ gas is not quite clear. So far, two electroactive species are reoxidized in the potential interval -1200 to -200 mV with no apparent reduction reaction taking place to a potential of -1400 mV. The cathodic decomposition of Na^SO^ apparently 109
II a
5 00 m 3 O TE I IAL
Figure 46. Cyclic voltammogram on a pure Pt foil WE immersed in a Na2S04 melt of basicity -10.07 under 02 at 900 C. 1 1 0 generates reduced species that can be reoxidized in 2 steps or two electroactive species that can be individually reoxidized. Some other electrochemical technique is needed to assist cyclic voltammetry in determining the reaction mechanism.
Figure 47 shows a chronopotentiogram for a pure Pt foil immersed in a Na^SO^ melt under Og gas at 900 C. A negative current of 60 mA is applied between the counter and the working electrodes. In the forward direction, no apparent cathodic transition time is detected before the plateau in potential at -2000 mV is observed. This plateau in potential corresponds to the cathodic decomposition of the melt as in the cyclic voltammograms previously discussed. Upon reversal of the current after 7.5 sec of electrolysis, three anodic transition times are obtained. A second electrolysis scan reveals a cathodic transition in the potential range -1200 to
1400 mV. This transition time is followed by a potential plateau as mentioned earlier. Upon reversal of the current, three anodic transition times are observed. The durations of the transition times are too short to be used for analytical purposes. However, qualitative information can be drawn from this result.
Even though there is not a direct relationship between cyclic voltammetry and chronopotentiometry, a comparison Potential (V) - - 2.0 1.0 1.0 3 9 2 5 8 1 4 7 30 27 24 21 18 15 12 9 6 3 0 f aiiy 1.7 t 0 C ne 0 gas. 0^ under C 900 at -10.07 basicity of figure figure ______47. 1 ______Chronopotentiogram on a pure Pt foil WE immersed in a Na?S0. melt melt Na?S0. a in immersed WE foil Pt pure a on Chronopotentiogram i ______i _____ i ______Time (sec) i ______i ______i ______i ______i ______1 1 2
between both electrochemical responses can be
established. From cyclic voltammetry a cathodic peak
labelled II'c was obtained at -1475 mV, compared to a
cathodic transition time between -1200 and -1400 mV. The
second anodic transition time was obtained between -1400
and -1100 mV compared to peak II’a in cyclic voltammetry
obtained at -1200 mV. The third anodic transition time
between -300 and -100 mV compared to peak I ’a at -300
mV. The good correspondence in the response from the
two techniques is obvious. The second and third anodic
transition times are not related through a two-step
oxidation of a single species. If such were the case,
the third transition time would have been longer than
the second transition time because the current consumed
by the second process would have been smaller than the
total applied current and the transition time for the
second species would have been considerably lengthened.
Therefore, the second and third transition times
correspond to the reoxidation of two independent
electroactive species. From the cyclic voltammograms in
Fig. 45, a redox relationship between peaks II’c and
II’a is not clear. However, by analogy to the response
from the cyclic voltammograms and chronopotentiograms
obtained from NagSO^ melts of basicity -6.66 discussed
in the preceding section, the cathodic transition time might be electrochemically related to the second anodic 113
transition time and peaks II’a and II’c may constitute a
redox couple.
Figure 48 shows a chronopotentiogram for the same
experimental conditions but for an applied current of
200 mA. Initially for a negative current of 200 mA,
no cathodic transition time is detected in the forward
direction . Reversal of the current after 4.5 sec of
electrolysis reveals an anodic transition time within a
potential interval similar to the one at which the third
anodic transition time was observed in the
chronopotentiogram of Fig. 47. A second scan provides a
cathodic transition time and upon reversal of the
current three anodic transition times are revealed. The
appearance of the first and second anodic transitions,
as well as the appearance of the cathodic transition,
seem to be related to the extent to which the cathodic
and anodic decompositions of the melt are allowed to
proceed. Therefore, the electroactive species that are
reduced and oxidized at the cathodic transition and at
the second anodic transition , respectively, are
generated from the decomposition of the melt . The
apparent independence of the third anodic transition on the extent of cathodic decomposition of the melt
suggests that it is originally present for the melt
The corresponding cathodic transition may be taking Potential (V) - - 2.0 1.0 1.0 0 3 9 2 5 8 1 4 7 30 27 24 21 18 15 12 9 6 3 0 ______bsct -00 a 90 udr O under C 900 at -10.07 basicity f igure igure 48 i ______. Chronopotentiogram on a pure Pt foil WE immersed in a Na9S0. melt melt Na9S0. a in immersed WE foil Pt pure a on Chronopotentiogram . i ______i ______i ______Time (sec) 2 gas. i ______i ______i ______i ______i ______115 place at a potential at which it is masked by double layer charging effects upon changing the direction of the current. Its absence could also be related to the low concentration of the electroactive species which is directly related to the transition time according to the
Cottrell equation.
Figure 49 shows a chronopotentiogram on a pure Pt foil
WE immersed in a 10 m/o NaVOg-NagSO^ melt under Og gas at 900 C. When a current of -130 mA is applied between the counter and working electrodes a cathodic transition time of 3.75 sec is obtained at a potential which is cathodic to the open-circuit potential of -20 mV against a Ag/Ag+/mullite reference electrode. Upon reversal of the current, an anodic transition time of .6 sec is obtained. Table 10 lists the experimental results obtained at various applied currents. The forward to backward transition time ratio is used as a mechanism criterion to test for complications arising from adsorption of the reduction reaction product or from kinetic effects from a following chemical reaction that consumes the reduction reaction product. A ratio equal to 1 implies adsorption of the reduced product, while for a ratio greater than 3 a following chemical reaction is suspected. A ratio equal to 3 corresponds to a simple diffusion-controlled redox process. From Table Potential (V) - - 1.0 2.0 0 3 9 2 5 8 1 4 7 30 27 24 21 18 15 12 9 6 3 0 Na^SO^ solution of basicity -11.7 at 900 C under under C 900 at -11.7 basicity of solution Na^SO^ Figure Figure 49 . Chronopotentiogram on a pure Pt foil WE immersed in a 10 m/o NaVO^- NaVO^- m/o 10 a in immersed WE foil Pt pure a on Chronopotentiogram . Time (sec) 0^
gas. 6 1 1 117
10, for various applied currents, a ratio
greater than 3 is always obtained, confirming the
prediction from cyclic voltammetry of a following
irreversible chemical reaction. According to the Sand 1 /2 equation, a plot of i 'TT vs i should yield a
horizontal line whose intercept with the ordinate would
be a constant related to the diffusion coefficient and
concentration of the electroactive species. This
equation is valid for both reversible and irreversible
redox processes. Deviations from a horizontal line
indicates either a preceding chemical reaction or
adsorption of the reactant, Fig. 50. Figure 51 shows
such a plot for the data reported on Table 10.
Deviation from a horizontal line at relatively high
applied currents confirms the prediction concerning
adsorption of the reactant at the WE. The reversibility
of the redox process was tested by plotting the
potential, E, vs ln[('TT*/^ - t*1^ )/ t * ^ ] according to
Eq. [3]. Such a plot for the data gathered from Fig. 49
shows a straight line from whose slope the number of
electrons transfered was calculated to equal
0.63,nominally taken as one, Fig.52. Straight lines were
obtained when the same parameters were plotted using data for the various applied currents. Even though this trend may indicate reversibility of the redox process,
irreversibility can be implied if the calculated number Adsorption
Diffusion
Preceding reactions
Figure 50. Diagnostic Dlot. 118 ir1/2 (mAsec1/2) 250 350 300 1 10 5 170 150 130 110 Fiqure Fiqure 0 / NV^N^O sltos t 0 C ne 0 gas. 0£ under C 900 at solutions NaVO^-Na^SO^ m/o 10 51 1 ______. Diagnostic plot for chronopotentiometric data from from data chronopotentiometric for plot Diagnostic . I ______i(mA) I ______L 119 Potential (V) - - -1.3 - - 1.2 0.9 1.0 1.1 ng t q 3 . 3 Eq. to g in d r o c qure e r u iq F - -o 1.2 o • Best fit E x p e r i m e n t a l 2 5 . rm Fig. g i F from a t a d c i r t e m o i t n e t o p o n o r h c f o t o l P - 0.8
0.4 49
c a 0 2 1 1 2 1
Table 9. Variation Of Peak Potential With Scan Rate For
Cyclic Voltammograms On a pure Pt Foil WE Immersed In a
Pure NagSO^ Melt At 900 C Under O ^ Gas.
E r a E II ’ a E I11’a E II’c (mV/sec) (mV) (mV) (mV) (mV)
10 - 275 - 1275 - 1950 20 - 375 - 1175 - 1925 - 1450 30 - 375 - 1175 - 1925 - 1475 40 - 375 - 1175 - 1900 - 1475 50 - 375 - 1175 - 1900 - 1475
Table 10. Chronopotentiometric Data On a Pt Foil WE
Immersed In a 10 m/o NaVOg-Na^SO^ Solution At 900 C
Under 0 r Gas.
i E n i ^ b (mA ) (V) (mA/sec
120 - 1.052 + 0. 165 in (u> 0.61 5. 33 262.91 130 - 1.062 + 0 . 164 In {u> 0. 61 6. 25 251.74 140 - 1.047 + 0. 176 In {u} 0. 58 5. 44 258.15 150 - 1 .062 + 0 . 161 In {u} 0 . 63 5. 95 290.47 160 - 1.079 + 0. 169 In {u} 0. 60 5. 77 309.84
Note: In {u}= In [(TT1/2 - t 1/2 )/ t 1/2 ] Potential (V) - - -1.3 - - 1.2 1.1 0.9 1.0 . -. -. -. -0.2 -0.4 -0.6 -0.8 1.0 ng t q 4 Eq. to g in d r o c e r u q i F 53. o E x p e r i m e n t a l • Best fit P l o t o f c h r o n o p o t e n t i o m e t r i c d a ta from F ig . . ig F from ta a d c i r t e m o i t n e t o p o n o r h c f o t o l P In
1/ t ( 2 - t1/ - 2 ) 0 . 0.4 0.2 49
c a 2 2 1 123
of electrons transferee! decreases as the applied current
is increased. From Table 10, the calculated number of
electrons varies within an acceptable range of error,
and it averages 0.61 +/- 0.0162. A further check for
irreversibility was done by plotting the potential, E,
vs In ( ) according to Eq. [4]. If the
redox process were irreversible, such a plot would have
yielded a straight line. Figure 53 shows such a plot.
The adjustment to a straight line for the same potential
range as in Fig. 52, is comparatively poor. It is
concluded therefore, that the electroactive vanadium
species in a 10 m/o NaVOg-NagSO^ solution can undergo a
.60, nominally one, electron reversible redox reaction ,
followed by a chemical reaction that involves the
reduced species.
Figure 54 shows chronoamperograms for a pure Pt foil
WE immersed in a pure NagSO^ melt at 900 C under Og gas.
A continuous decay in current with time is observed when
potential steps of -1000 mV and -1200 mV are applied
between the reference and the working electrodes. The
decay in current does not agree with the earlier
observation by Park (32) of a constant current at a time
frame of 20 sec for potential steps as low as 100 mv cathodic to the open-circuit potential; for such potential steps, not much cathodic response is obtained Current density (mA/cm2) -50 -40 -30 -20 -10 potential step: (a) - 1000 mV, (b) - 1200 mV. mV. 1200 - (b) mV, 1000 - (a) step: potential a Figure Figure 0 Na„S0. melt of basicity -10.07 at 900 C under 0 9 gas. Final Final gas. 9 0 under C 900 at -10.07 basicity of melt Na„S0. 54. 10 hoomeorm o a ue t ol E mesd in immersed WE foil Pt pure a on Chronoamperograms 0 0 0 0 60 50 40 30 20 Time (sec) d 4 2 1 125 as indicated by the cyclic voltammograms in Fig. 46.
Figure 55 shows a series of chronoamperograms obtained for various applied potential steps for a 10 m/o NaVO^-
Na^SO^ solution . A continuous decay in current with time is also observed. The current responses for both systems after 3 sec of electrolysis at constant potential are compared in Fig. 56. From Fig. 56, for comparable potential steps, the current response obtained for the NaVO^-NagSO^ system is up to 5 times larger than the one obtained from pure NagSO^. The larger current response arising from any electroactive vanadium species reducible at potentials as negative as
-600 mV as indicated by cyclic voltammetry and chronopotentiometry. Once again, the electrochemical response from vanadium species can be clearly discriminated from that obtained from the supporting electrolyte, NagSO^. An apparent limiting (diffusion- control) current of .55 A/cm2 is calculated for vanadium species in Fig. 56. Figure 57 shows a cathodic polarogram at a scan rate of 1 mV/sec for a pure Pt WE immersed in a 10 m/o NaVO^-NagSO^ solution at 900 C under 0 An activation polarization stage is observed between -600 and -1000 mV, with a calculated Tafel slope of -316 mV. The calculated number of electron transfered was .68 for an assumed value of a =0.5. The activation polarization stage is followed by a mixed- Current density (A/cm2) -0.4 - -0.5 -0.3 - 0.1 0.2 ^ a. ia ptnil tp () 80 V () 90 V , mV 900 - (b) mV, 850 - (a) step: potential Final gas. O^ 1 mo ouin f aiiy 1. a 90 under C 900 at -11.7 basicity of solution ^ O S ^ - ^ O V a N m/o 10 a iue hoomeorm o a ue t ol E mesd in immersed WE foil Pt pure a on Chronoamperograms . 5 5 figure c - 00 V () 15 m, e - 10 mV. 1100 - (e) mV, 1050 - (d) mV, 1000 - (c) Time (sec) 0 0 70 60 50 6 2 1 Current density (A/cm2) 0.3 0.6 0.4 0.2 0.5 2 a. 2 24 2 4 2 J gas. 02 data obtained from NaV0--NaoS0. and NaoS0. melts at 900 C under under C 900 at melts NaoS0. and NaV0--NaoS0. from obtained data Figure 5 6 . Superimposed i vs E curves from chronoamperometric chronoamperometric from curves E vs i Superimposed . 6 5 Figure P o t e n t i a l ( V ) 127 Potential (V) - -1.60 -0.80 -0.40 1.20 0.00 t 0 C ne 02 a . gas 2 0 under C 900 at Figure 5 7 . Dynamic polarization curve on a pure Pt foil WE WE foil Pt pure a on curve polarization Dynamic . 7 5 Figure mesd n 1 mo a0 N2S4 ouin f aiiy -11.7 basicity of solution S04 -Na2 NaV03 m/o 10 a in immersed 04 0.001 1x 10"4 C u r r e n t d e n s i t y ( A / cm 2 ) U 1 mV/sec. 1 = 0.01 0.1 1 8 2 1 1 2 9
control stage. No apparent limiting current could be
detected.
Reaction mechanism
1. NagSO^ melts Under Og Gaseous Atmosphere
Figure 58 shows the basicity trace recorded during the
polarization of a Pt working electrode painted onto a
zirconia tube immersed in a NagSO^ melt at 900 C under
C>2 atmosphere. Prior to any polarization studies, the
condition of the melt was established in terms of the
partial pressure of oxygen and the activity of sodium
oxide. This initial state lay in the field for the
stability of superoxide ions close to the boundary line with the pyrosulfate ions stability field. The basicity trace does not traverse the stability field for oxide ions; therefore, the dominant minority ionic species in
2 - the melt should be and 0^ ions. Pyrosulfate ions are present in equilibrium with sulfate ions and sulfur trioxide according to the following reaction,
S042" + S03 S20 ?2' [27]
Further, sulfate ions are in equilibrium with SOg and oxide ions in the melt according to the equilibrium decomposition of sodium sulfate,
S042" S03 + 0 2" [28]
Since the melt is under an oxygen atmosphere, superoxide ions are also present in equilibrium with dissolved -10 ' 5 ' Log p0, ue mesd n N90 ml o bsct -00 a 90 udr . o 0 under C 900 at -10.07 basicity of melt Na9S0. a in immersed tube Figure V = 100 mV/sec. mV/sec. 100 = 58. 58. Basicity trace on a polarized Pt WE painted on a zirconia zirconia a on painted WE Pt polarized a on trace Basicity \^la2S q 2 q d 2. 1.5 -1 .0 -2 Log pso- a2S2°7 " 0.5 -0 - 0.5 1.0 0 < + < LU OJ CD
> E, 0 3 1 131
oxygen and oxide ions according to the reaction,
0 2~ + 3/2 0 2 ^ 2 0 2~ [29]
From the cyclic voltammograms in Fig. 45 (a), (b), and
(c), a well defined cathodic peak (labelled I’c) is
obtained at -50 mV. The cathodic current associated
with peak I’c increases as the ASP is made more
positive. This increase in ASP is responsible for a
continuous rise in the anodic current resulting from
the anodic decomposition of the melt. The high anodic
currents recorded during the anodic decomposition of the
melt can only be generated by the decomposition of
2 - SO^ ions according to the reaction,
S042" ------* S03 + 1/2 0 2 + 2 e"
Based on X-ray analysis on an anodically polarized WE,
Fang (38) proposed a parallel oxidation reaction to form
2 0 2 x 0 2 + 3 e
According to the phase diagram of Fig. 58, superoxide ions are stable at the potential of peak I’c and; therefore , peak I’c is related to the electrochemical reduction of molecular oxygen at the working electrode through the reaction,
0 2 + e- * 0 2" [30]
Peak I’c is only detected in the second cycle of CV after anodic decomposition of the melt is allowed to 1 3 2
take place, indicating a low solubility of molecular
oxygen in the Na^SO^ melt as reported by Andresen (40).
For CSP as negative as -1000 mV an anodic peak
(labelled I ’a) is recorded at -500 mV upon reversal of
the potential. The peak I’a potential is shifted to
more anodic potentials as the CSP is made more negative.
However, its peak current is not significantly affected
by the extension of the CSP up to -1850 mV, a potential
at which the cathodic decomposition of the melt begins
to take place. The peak I ’a potential stabilizes at
-350 mV, where accoding to the phase stability diagram of
Fig. 58, superoxide ions can be electrochemically
oxidized according to the reaction,
02" ----- * 02 + e- [31]
Peak I’a is observed in the first scan, indicating that
superoxide ions are originally present in the
equilibrated melt. But the relatively low peak currents
indicate a relatively low concentration of 0^ ions in
the melt.
For a CSP of -1800 mV an ill-defined anodic peak
(labelled II’a) is observed at -1150 mV. Further
increases in the CSP sharpens peak II’a, and for a CSP
of -2200 mV ( Fig. 46), peaks II’a and I’a are clearly
defined. An additional anodic peak (labelled III’a)
corresponding to the oxidation of some electroactive 1 3 3
species generated by the cathodic decomposition of
NagSO^ also appears. An ill-defined cathodic peak
(labelled Il’c) is seen at -1475 in succeeding cycles.
But peak Il’c becomes hidden by the rising of the
cathodic decomposition current . The extent to which
the cathodic decomposition is allow to proceed affects
the appearance of peak II’a and peak I’a. The cyclic voltammogram in Fig. 46 and the chronopotentiograms in
Figs. 47 and 48, closely resemble the cyclic voltammograms and chronopotentiograms for Na^SO^ melts of basicity -6.66. Therefore, the same species must be
involved in the redox process. For the cathodic decomposition of melts of basicity -6.66, the following two-electron transfer reaction was concluded from dynamic polarization:
S042" + 2e------* S02 + 2 0 2_ [32]
Isobars for the partial pressure of SC>2 added to the phase diagram in Fig. 58 sugget that SO^ formation is probable at the potential range at which the cathodic decomposition of the melt takes place.
A parallel decomposition reaction can be the reduction of dissolved molecular oxygen according to reaction,
0 2 + 2 e k 0 22' [33]
2 - 2 - 02 ions can further react with SC>2 to regenerate S04 ,
SOg + 022' * so42~ [34] 1 3 4
A significant increase in the peaks II’a and I’a currents upon reversal of the potential was observed when the cathodic decomposition of the melt takes place.
2 - At such peak potentials the oxidation of and 02 ions takes place, respectively. Therefore, the generation of such species can only be accounted by the oxidation of the oxide ions generated by the cathodic decomposition of the melt. Such an oxidation reaction takes place at -1900 mV according to reaction,
2 02" ---* 0 22' + 2 e [35]
2 - Og ions are further oxydyzed to Og ions at -1100 mV through a one electron redox reaction,
0 22'---* 0 2" + e [36]
0 2 ions are oxidized to molecular oxygen at a potential around -350 m V .
The reaction mechanism for the supporting electrolyte under 02 gas is the same as for NagSO^ as basic as
- 6 . 6 6 .
2. NaVOg-NagSO^ Solutions Under 02 Gas
Figure 59 shows the basicity trace recorded during polarization of a Pt WE painted on a zirconia tube immersed in a 10 m/o NaVOg-Na^O^ solution under Og gas at 900 C superpositioned onto the stability diagram for the system Na-V-O-S. At the open circuit potential, Log p02 900 C under 0^ gas. gas. 0^ under C 900 ue mesd n 1 mo a0-a 0 slto o bsct -17 at -11.7 basicity of solution S04 NaV03-Na2 m/o 10 a in immersed tube Figure Figure 59- 59- Basicity trace on a polarized Pt WE painted on a zirconia zirconia a on painted WE Pt polarized a on trace Basicity V 10 mV/sec. 100 = - 2.0 -1.5 0.5 -0 - 0.5 1.0 0 < LU + o>
O)
> , E 5 3 1 1 3 6
the stability diagram for the melt suggests VO^ ions as
the predominant ionic vanadium species in the melt. The
trace in basicity is essentially vertical (only +/- 0.1
shift) indicating that the basicity at the WE does not
change during polarization. In NaVOg-Na^SO^ solutions,
the vanadate ions (VO^ ) complexes with oxygen anions to
form orthovanadate anions, The acidic solute is changed
2 - 3 - from SO^ to VO^ and the basicity of the melt is
buffered by the VO^3 to VO^ activity ratio. The
buffering effect is achieved through the following
equilibrium,
VO “ + O 2" * VO.3- r371 3------*------4 Cyclic voltammetry and chronopotentiometry suggest that
the electroactive vanadium species undergoes a one electron reversible redox reaction and that the product of the reduction reaction is unstable and it undergoes an irreversible chemical reaction. Considering VO^ as the vanadium electroactive species, it reduces according to the reaction:
V03~ + e- * V032~ t38]
The product of this reduction reaction is expected to undergo an irreversible chemical reaction,
vo32' + 02- ---- * VO,3" + 1/2 0 2 [39]
Reactions [39] and [37] keep the basicity constant during polarization. 137
Electrochemical Studies In NaVOg-Na^SO^ Melts Under
0.1% SOg-O^ Atmosphere
Figure 60 shows a series of cyclic voltammograms at various scan rates on a pure Pt foil WE immersed in a 10 m/o NaVOg-NagSO^ solution at 900 C under an uncatalyzed
0.1% SOg-Or, atmosphere. The open-circuit basicity of the solution was calculated as -11.72 in the logarithmic scale for the activity of NagO. One cathodic peak
(labelled Ic) is obtained at around -1350 mV when the potential of the WE is scanned toward potential values cathodic to the open-circuit potential of 4 mV against a
Ag/Ag+/mullite reference electrode. Upon reversal of the potential, an anodic peak (labelled la) is obtained at around -550 mV. Table 11 lists the variation of peak potential with scan rate. The peak potentials do not vary appreciably as the scan rate is increased, indicating that the redox process is reversible. To establish any redox relationship between peaks la and
Ic, cyclic voltammograms were recorded at various CSP,
Fig. 61. For a CSP as negative as -900 mV, a well defined anodic peak is obtained at -587 mV. As the CSP is made more negative a continuous rise in the cathodic current is observed, indicating that more electroactive species are being reduced at the working electrode. As the availability of reduced species in the immediate 5 0 tvA
80
Fiqure 60. Cyclic voltammograms on a pure Pt foil WE immersed in a 10 m/o NaVO^-^SO^ solution of basicity -11.72 at 900 C under uncatalyzed 0.1% SO2 -O2 gas mixture. 1 3 9
Table 11. Variation Of Peak Potential With Scan Kate For
Cyclic Voltammograms Recorded On a Pure Pt Foil WE immersed In a 10 m/o NaVO^-NagSO^ Solution At 900 C
Under a 0.1% SOg-Og Uncatalyzed Atmosphere.
E lat E Ict (mV/sec) (mV) (mV)
20 - 575 - 1325 40 - 587 - 1350 60 - 600 - 1350 80 - 575 - 1400 100 - 575 - 1400
Table 12. Variation Of Anodic To Cathodic Peak Current
At Various Scan Kates From Cyclic Voltammograms Recorded
On a Pure Pt Foil WE Immersed In a 10 m/o NaVOg-NagSO^
Solution At 900 C Under An Uncatalyzed 0.1% SO,-0„ Gas.
ipc = ipc ipa ips ipa/ipc ipa o o o (mV/sec) (mA) (mA) (mA) (mA)
40 47. 50 19.00 32. 50 0.82 38. 85 60 56. 50 26.00 35.00 0. 85 47 . 83 80 63.00 32. 50 37 . 50 0.89 56.11 100 70.00 38. 50 42. 50 0 . 93 65. 14 1 4 0 vicinity of the working electrode increases, a better defined anodic peak is seen upon reversal of the potential. For a CSP of -1450 mV, one cathodic peak at
-1360 and one anodic peak at - 575 mV, respectively, are detected in the voltammogram. It can be concluded that peaks la and Ic constitute a reversible redox couple.
The variation of the anodic to cathodic peak current ratio with scan rate is commonly used as a diagnostic mechanism criterion. From the voltammograms in Fig. 61 it is not possible to clearly define a base to measure the anodic peak current. Therefore, once again the peak current ratio was calculated using Eq. [2] and the results are listed in Table 12. The trend of the data shown in Fig. 62 suggests that the reversible redox process is followed by an irreversible chemical reaction which involves the reduced species (the trend of the 1 /2 data in a plot ipc/// versus also supports this hypothesis). The reversibilty of the redox process can be evaluated by a plot of the peak current versus the square root of the scan rate. If diffusion of the electroactive species to the WE surface controls the redox proces, the peak current varies linearly with the square root of the scan rate according to the Randles- 1 /2 Sevsik equation. Fig. 63 shows a plot of ipc vs// for the cathodic peak Ic. A straight line is obtained 1 4 1
Figure 61. Ciclic voltammograms for the same experimental conditions as in Fig. 6 0 for various cathodic switching p otential. iap/icp 0.84 0.76 0.92 1.00 20 ih cn ae o cci vlamgas n i.60. 0 Fig. 6 on voltammograms cyclic for rate scan with Figure 6 2 . Variation of anodic to cathodic peak current ratios ratios current peak cathodic to anodic of Variation . 2 6 Figure 40 v(mV/sec) 80 0 0 1 0 2 1 2 4 1 1 4 3
indicating diffusion control. Deviation from a straight
line is an indication of either adsorption of the
reactant species or a preceding chemical reaction. A
careful look of peak Ic indicates a sharp decay in
current after the peak current has been reached and an
increase in the symmetry of the peak as the scan rate is
increased. Both features are characteristic of
adsorption. However, no prepeak indicating weak
adsorption of the reactant is observed. For this
circumstance, an adsorption-control peak would be
observed at the same potential as a reversible peak and 1 /2 a plot of ip vs would look like as if the redox
process were diffusion controlled . Indeed, a plot of 1 /2 ip v s If in Fig. 63 for the anodic and cathodic
peaks la and Ic show diffusion control.
Cyclic voltammograms were recorded on a pure Pt foil
immersed in molten Na^SO^ at 900 C under a catalyzed 0.1
% SOg-Og atmosphere. The calculated open-circuit potential basicity is -13.46 for the logarithmic scale for the activity of NagO. The cyclic voltammograms were recorded for various values of CSP, Fig. 64. These cyclic voltammograms do not indicate any cathodic or anodic redox process for a CSP as negative as -1700 mV.
Comparison of the response for pure Na^SO^ for such a
CSP with the one obtained for NaVO^-Na^SO^ solutions at 1 4 4
80
70 < E, 60 CL O 50
40 4 6 8 10 12 v1/2
70
60 < 50 Q. CCS 40
30 4 6 8 10 12 v1/2
Figure 63. Peak current variation with the square root of scan rate, (a) cathodic peak, (a) anodic peak. the same current sensitivity, indicates that the peaks
labelled la and Ic in Fig. 60 correspond to some
vanadium electroac.tive species present in NaVOg-NagSO^
solutions.
From Fig. 64, for a CSP of -2300 mV, three anodic
peaks are observed at -1850, -1087 and -200 mV,
respectively. The peaks labelled I’a and II’a seem to
correspond to the reoxidation of reduced species
generated from the cathodic decomposition of NagSO^.
Similar observations were drawn from cyclic
voltammograms obtained from more basic NagSO^ melts.
Such voltammograms, however, showed peak I’a as a self
sustained peak, in the sense that it could be detected
prior to any cathodic decomposition of the melt. The
reaction proposed for the cathodic decomposition of
NagSO^ suggests an increase in the oxide ions
concentration at the working electrode. Peak I'a may
correspond to the oxidation of superoxide ions whose
concentration is initially defined by the basicity of
the melt. The more basic the melt, the higher the
concentration of Og ions. Therefore, peak I’a is very well defined in relatively basic melts, e.g., -log aNa^O
= 6.66, and the dependence of its appearance on
cathodic decomposition increases as the basicity of the melt decreases because of the lower concentration of 1 4 6
O 5 0 0 mV , POTENTIAL^
Figure 64. Cyclic voltammograms on a pure Pt foil WE immersed in a pure Na^SO^ melt of basicity -13.46 under a catalyzed 0.1% SO2 -O2 mixture at 900 C. 147
oxide ions coupled to the low reported solubility of
molecular oxygen in molten Na^SO^.
Figure 6b shows a chronopotentiogram from a pure Pt
foil WE immersed in a 10 m/o NaVOg-Na^SO^ solution at
900 C under an uncatalyzed 0.1 % SOg-Og atmosphere for
an applied current of 60 mA. When a negative current of
60 mA is applied between the working and the counter
electrodes a transition time of 15.8 sec is recorded in
the forward cathodic direction . Upon reversal of the
current at the forward transition time, a reverse
transition time of 3 sec is registered. Table 13 list3
the chronopotentiometric data calculated for various
applied currents. The forward to backward transition
time ratios, are well above the value of 3
expected if the redox process were completely controlled
by diffusion of the electroactive species. As mentioned before, deviation from this value suggests a more complicated redox mechanism involving kinetic or
adsorption effects. In the present case, the 'tj./'t, f b ratios suggest that a chemical reaction which consumes the reduction reaction product follows the reduction reaction. The chemical reaction decreases the concentration of reduced species available to be reoxidized at the WE surface , resulting in a shorter backward transition time compared to the one expected if 1.0 -
______I______I______I______I______I______I______I______I______I______0 4 8 12 16 20 24 28 32 36 40 Time (sec) Figure 65 . Chronopotentiogram on a pure Pt foil WE immersed in a 10 m/o NaVO.,- Na^SO^ solution of basicity -11.72 at 900 C under uncatalyzed 0.1% SO^-O^ gas. 1 4 9
Table 13. Chronopotentiometric Data From a Pure Pt Foil
WE Immersed In a 10 m/o NaVOg-Na^SO^ Solution At 900 C
Under An Uncatalyzed 0.1% SO^-O^ Atmosphere.
i E n ' - Tf/^b
(m A ) (V)
2u0 - 0.947 + 0. 177 In {u} 0. 57 4. 81 150 - 0.956 + 0. 173 In (u> 0. 59 4.44 100 - 0.964 + 0. 175 In {u} 0. 58 3. 93 60 - 0.996 + 0. 188 In (u> 0. 54 5. 26 60* - 0.908 + 0. 148 In {u> 0.69
* Anodic Scan . /9 . /9 . „ Note: In {u} = In - tx/* )/ tx/*
Table 14. Variation Of Peak Potential With Scan Rate For
A Pure Pt Foil WE Immersed In a 10 m/o NaVOg-NagSO^
Solution At 900 C under a Catalyzed 0.1% SO^-Og
Gas Flowing At 0.223 ml/sec.
E lat eIc t elie tt (mV/sec) (mV) (mV) (mV)
16 - 500 - 1350 26 - 500 - 1400 - 1750 36 - 550 - 1425 - 1750 46 - 600 - 1450 - 1800 50 - 575 - 1400 - 1750 100 - 575 - 1425 - 1800 200 - 600 - 1425 150
the redox process were purely diffusion controlled.
The reversibility of the redox process was tested by
plotting the potential , E , versus In-t*^) 1 /2 /t ] according to Eq. L3]. Figure 66 shows such a
plot, a straight line is obtained from whose slope the
number of electron transfered is calculated as .54,
taken nominally as one. Similar trends of the data were
obtained for the other applied currents. For the
applied current range of 60 to 200 mA the calculated
number of electrons transfered does not decrease as the
applied current increases. A decrease in the calculated
number of electrons with increasing applied current is used as a criterion for irreversibility despite of the
indication of reversibilty from Eq. [3].
Irreversibility of the redox process was also checked by a plot of E vs In ('ZT - t*^) according to Eq.
[4]. Figure 67 shows such a plot. The poor fit of the experimental data to a straight line rules out the irreversibility of the redox process.
Equation [3] was applied for the backward transition time in Fig. 65. Figure 68 shows a plot of E vs ln[(
t^^^)/t^^^]. A straight line is obtained from whose slope the number of electrons transfered was calculated as .69. It can be concluded from Potential (V) - - -0.9 - - 0.8 1.2 1.0 1.1 cording t E. 3 Eq. to g n i d r acco Figure Figure 0.8 • Best fit o E x p e r i m e n t a l 66. lt aa rm . g i F from data c i r t e m o i t n e t o p o n o r h c f o Plot
In 0 /2-t1/2 t 1/ - 2 T t 1/2 0.40.4 0.8 65
1.2 1 5 1 Potential (V) - - - -0.9 1.2 1.0 0.8 ng to E. 4 Eq. o t g in d r o c c a e r u g i F -o 0.2 • Best fit o E x p e r i m e n t a l 67. a fo . g i F from ta a d c i r t e m o i t n e t o p o n o r h c f o t o l P 0.4
0.8 65 152 153
chronopotentiometry that the electroactive species in
NaVOg-Na^SO^ solutions undergoes a reversible redox
process with .57, nominally one, electron transfered.
Figure 69 shows a chronopotentiogram for a pure Pt
foil WE immersed in a pure NagSO^ melt at 900 C under a
catalyzed 0.1 % SOg-Og atmosphere. When a negative
current of 40 mA is applied between the counter and the working electrodes no cathodic transition time is
obtained. A plateau in current at -1950 mV is obtained
after 15 sec of electrolysis. This current plateau
corresponds to the cathodic decomposition of NagSO^ as
in cyclic voltammetry within the same potential range.
Upon reversal of the current three anodic transition times are observed at potentials comparable to the three anodic peaks obtained in cyclic voltammetry. In a second scan in the cathodic direction an ill-defined cathodic transition time is observed prior to the cathodic decomposition of the melt. These observations in the chronopotentiogram are in good agreement with the results of cyclic voltammetry. In cyclic voltammetry, no well defined cathodic peak was recorded in the vicinity of -1900 mV except in the second cathodic scan, which suggests that the electroactive species reducible at potentials between -1300 and -1500 mV is not originally present in the melt in sufficiently high Potential (V) - -0.9 0.8 0.7 t r a n s i t i o n in F ig . . ig F in n o i t i s n a r t .gr 68 6 F.igure - . 1.2 a for the anodic d o n a e h t r o f ta a d c i r t e m o i t n e t o p o n o r h c f o t o l P 65 - ng to E. 3 Eg. o t g in d r o c c a 0.8 o E x p e r i m e n t a l • Best fit 0
-0.4 4 5 1 - 2.0
- 1.0
0
0 05 10 15 20 25 30 35 40 45 50 Time (sec) Figure. 69. Chronopotentiogram on a pure Pt foil WE immersed in a Na2S04 melt of basicity -13.46 at 900 C under a catalyzed 0.1% SOo-O^ gas mixture. 156
concentration to be detected by either technique.
Therefore, these reducible species needed to be
generated, probably from the reoxidation of some species
form during the cathodic decomposition of NagSO^ or from
the reduction of some species generated during anodic
decomposition From electrochemical studies on more
basic NagSO^ melts, it was concluded that the cathodic
transition between -1300 and -1500 mV and the anodic
transition between -1300 and -1000 mV corresponds to a
reversible redox couple with one electron transfered.
Since similarities have been retained for cyclic
voltammograms and chronopotentiograms for the NagSO^
melts at the various basicities studied, it is
acceptable to assume that the electroactive species are
the same in all cases, the only difference being their
relative concentrations which would be fixed by the
basicity of the melt. The short transition times do
not provide additional information regarding the redox
process.
Figure 70 shows a series of chronoamperograms recorded
on a pure Pt foil WE immersed in a 10 m/o NaVOg-NagSO^ melt under an uncatalyzed 0.1% SOg-O^ gaseous atmosphere at 900 C. The applied potential step was varied from -600 to -1500 mV against a Ag/Ag+/mullite high temperature reference electrode. A decay in Current density (A/cm2) -0.5 -0.4 -0.3 - - 0.2 0.1 1 2 3 4 5 6 7 80 70 60 50 40 30 20 10 0 tp () 90 V () 10 m, c -10 V () 10 mV. -1200 (d) mV, -1100 (c) mV, -1000 (b) mV, 900 - (a) step: u nder an u n c a t a l y z e d 0.1% S0 o - 0 ? gas m i xt u r e . Fi n a l potential potential l a n Fi . e r u xt i m gas ? 0 - o S0 0.1% d e z y l a t a c n u an nder u Figure 7 0 . C h r o n o a m p e r o g r a m s on a pure Pt foil WE immersed immersed WE foil Pt pure a on s m a r g o r e p m a o n o r h C . 0 7 Figure n 1 mo oution o bas 1.2 t 0 C 900 at -11.72 y t i c si a b of n o i t solu ^ O S ^ a N - ^ O V a N m/o 10 a in Time (sec) 157 158
current with time was observed as described earlier.
The current response for 3 seconds of electrolysis is
plotted in Fig. 71 as a function of the applied
potential. A limiting current density of .4425 A/cm2
was obtained. Figure 72 shows the superposition of
dynamic polarograms on a Pt foil WE immersed in a 10 m/o
NaVOg-Na^SO^ solution and in a pure NagSO^ melt,
respectively, at 900 C, at a scan rate of 1 mV/sec. For
the same potential and current scale, the
electrochemical response from the NaVOg-NagSO^ melt is
totally different. An activation polarization stage is
obtained in the potential interval -200 to -600 mV for
the NaVOg-NagSO^ solution. For a =.5 the number of
electrons transfered is 1.16, nominally one, as
calculated from the Tafel equation. This activation
polarization stage is followed by a concentration
polarization stage characterized by a limiting current
density of 0.3846 A/cm2, a value in good agreement with
that obtained from chronoamperometric results. A
limiting current density is not observed in the
polarogram from a pure NagSO^ melt, thus the limiting
current arises from the diffusion-controlled reduction
of some vanadium species. This observation confirms the
results from cyclic voltammograms and
chronopotentiograms obtained under the same experimental conditions. Current density (A/cm) -0.5 - -0.3 -0.4 - 0.2 0.1 - 1.6 f i g u r e 71 . i vs E curve from c h r o n o a m p e r o m e t r i c data on a pure pure a on data c i r t e m o r e p m a o n o r h c from curve E vs i . 71 e r u g i f 1.2 t 0 C dr n .1 gs mixture. gas 2 0 " 2 0 S 1% 0. d e z y l a t a c n u an nder u C 900 at -11.72 t ol E mersed i a 0 / NaVO^-^SO^ sl f icity t i c si a b of n o i t solu ^ O S ^ - ^ O V a N m/o 10 a in d e s r imme WE foil Pt . -. -. -. -. -0.4 -0.6 -0.8 -1.0 -1.2 1.4 P o t e n t i a l ( V ) 159 Potential (V) - - - - -0.4 2.0 1.6 0.8 1.2 0
aS^ n N^O mls t 0 C. 900 at melts Na^SO^ and Na^SO^ Figure 7 2 . S u pe r i m p o s e d p o l a r i z a t i o n curves fr om NaVO^- NaVO^- om fr curves n o i t a z i r a l o p d e s o p m i r pe u S . 2 7 Figure I 1X10“4 ______C u r r e n t d e n s i t y (A / c m 2 ) .0 00 01 1 0.1 0.01 0.001 I ______NaV03-Na^O*- .1%S0 I ______I ______0 6 1 1 6 1
Figure 73 shows a dynamic polarogram from a Na^SO^
melt under a 0.1 % SOg-O^ atmosphere at 900 C for a
wider potential range than in Fig. 68. In the potential
interval of -2000 to -2250 mV an activation polarization
stage is obtained. For a =0.5 the calculated number of
electrons transfered equals 1.78, nominally 2, from the
Tafel equation. Within the same potential range, an
activation polarization stage is also observed from
polarograms obtained from NagSO^ melts of basicity
“6.66. Within this potential range the cathodic
decomposition of the electrolyte was observed in cyclic
voltammograms and chronopotentiograms obtained under
similar experimental conditions. Therefore, it could be
concluded that the cathodic decomposition of the NagSO^
electrolyte involves an overall 2-electron reduction
reaction .
Electrochemical studies were performed in a 10 m/o
NaVOy-Na^SO^ solution under a catalyzed 0.1 % SO^-Og
atmosphere to observe the possible effects of the gas
composition on basicity and electrochemical response.
The flow rate was also varied from .223 ml/sec to 2.23
ml/sec. Significant differences were obtained in melt
basicity, which in turn helps to define the species
involves in the reaction mechanism. The major role of
the catalyst is to accelerate the conversion reaction of
SO^ and Og to S O ^ . Dissolved SO^ is an acidic component - Potential (V) - 2.0 1.0 2.0 . % O-^ a mixture. gas SO^-O^ % 0.1 ol E mersed i a SO^ melt a 90 une catalyzed e z y l a t a c nder u C 900 at t l e m ^ O ^S a N a in d e s r imme WE foil Figure 7 3 . D y na m i c p o l a r i z a t i o n curve at 1 m V / s e c for a Pt Pt a for c e s / V m 1 at curve n o i t a z i r a l o p c i m na y D . 3 7 Figure X 4 C u r r e n t d e n s i t y ( A / c m 2 ) 2 6 1 1 6 3
which would compete with NaVO^ present in the melt to
fix the basicity of the solution. Indeed, the melt was
made more acidic than for the condition of uncatalyzed
SOg + Og gases. The acidity of the melt was further
increased by increasing the flow rate for the same
catalyzed gas composition . For uncatalyzed gases, no
increase in the melt basicity was observed with an
increase in the flow rate. The reason for this behavior will be explained later after a description of the experimental results.
Figure 74 shows a series of cyclic voltammograms on a pure Pt WE immersed in a 10 m/o NaVOg-NagSO^ solution at
900 C under a catalyzed 0.1 % SOg'Og gas mixture flowing at a rate of 0.223 ml/sec to give an open-circuit potential basicity -13.15 on the log a Na20 scale. The scan rate was varied from 16 to 200 mV/sec. At a scan rate of 46 mV/sec, one anodic peak (labelled la) and two cathodic peaks (labelled Ic ans lie) were obtained at
600, -1450 and -1800 mV, respectively. Table 14 lists the variation of peak potentials with scan rate. The peak potentials remain constant within a range of 25 mV which is considered within the experimental error for the magnitude of the measured potentials . The constancy of peak potential suggests reversibility for the redox process. From the preceding sections, the Figure 74. Cyclic voltammograms on a pure Pt foil WE immersed in a 10 rn/o NaVO^-Na^SO^ solution
of basicity -13.15 at 900 C under catalyzed 0.1 % S02-02 gas mixture. I* = mV/sec. 4 6 1 165
second cathodic peak lie corresponds to the reduction
of some electroactive species present in the supporting
sulfate electrolyte.
Figure 75 shows cyclic voltammograms at a scan rate of
66 mV/sec under same experimental conditions as in Fig.
74. In these voltammograms, the CSP was continuously
increased from -575 to -2250 mV. Upon reversal of the
potential at -800 mV a well defined anodic peak (la) was
obtained at -550 mV. As the CSP was further increased, the current associated with peak la increased,
indicating an increase in the amount of the species being reoxidized, because the peak current is directly proportional to the amount of electroactive species. As the CSP was increased , a continuous increase in cathodic current was seen until peak Ic was detected.
This increasing current indicates that more electroactive species are being reduced at the WE , readily available to be reoxidized upon reversal of the potential resulting in a well defined anodic peak la.
Peak Ic is obtained at -1475 mV. Upon reversal of the potential for a CSP of -2250 mV, two additional anodic peaks (labelled Ila and Ilia ) are obtained. Peak Ilia corresponds to the oxidation of the products of the cathodic decomposition of NagSO^. According to the previous section peak Ila at -950 mV corresponds to the Figure 75. Cyclic voltammograms at various cathodic switching potentials on a pure Pt foil W E immersed in a 10 m/o NaV02~Na2S0^ solution of basicity -13.15 at 900 C under a catalyzed 6 6 1 0.1 % SO 2 -O 2 gaseous atmosphere. 167
oxidation of some subproduct generated in the cathodic
decomposition of the melt. It can be concluded that
peaks la and Ic constitute a redox couple for an
electroactive vanadium species present in NaVOg-Na^SO^
solutions. The increase in the symmetry of peak Ic with
increasing scan rate suggests weak adsorption of the
reactant.
Figure 76 shows a chronopotentiogram for a pure Pt planar WE immersed in a 10 m/o NaVO^-NagSO^ solution at
900 C under a catalyzed 0.1 % SO^-Og gaseous atmosphere.
A forward transition time, , of 7.5 sec is obtained when a negative current of 125 mA is applied between the the working and counter electrodes. Upon reversal of the current, immediately after the forward transition time, a reverse transition time, . of 1.5 sec is recorded. If the ratio of the forward to backward transition times, is greater than 3, the product of the electrode reaction is unstable in the solution,i.e., the reverse time will be shortened because the amount of species to be reoxidized is decreased via a chemical reaction. Table 15 lists the chronopotentiometric data gathered for the present experimental conditions and various applied currents.
For all applied currents, the ratios are greater than 3, suggesting the occurrence of a following 1.0 -
______I______I______I______I______I______I______I______I______I______0 5 10 15 20 25 30 35 40 45 50 Time (sec)
Figure 76. Chronopotentiogram on a pure Pt foil WE immersed in a 10 m/o NaVO^-Na^SO^ solution of basicity -13.15 at 900 C under a catalyzed 0.1% S02-02 gas flowing at 0.223 ml/sec. 8 6 1 1 6 9
chemical reaction.
For a reversible redox reaction, a plot of potential,
E, versus the time relationship given by Eq. [3] should
yield a straight line whose slope would allow the
calculation of the number of electrons transfered in the
redox process. Figure 77 shows such a plot for an
applied current of 122 mA. A straight line is obtained,
confirming the reversibility of the redox process. The
same trend in the data was obtained for the other
applied currents and the number of electrons transfered
averages 0.63 +/- 0.04. Any possibility of
irreversibility was ruled out, based on the
insignificant change observed in the calculated number
of electrons transfered for increasing applied currents.
From the combined experimental results it can be concluded that the vanadium electroactive species in
NaVO^-Na^SO^ solutions of basicity -13.15 undergoes a
0.63, nominally one, electron reversible redox reaction followed by a chemical reaction that consumes the reduced species.
Figure 78 shows a series of cyclic voltammograms for a pure Pt WE immersed in a 10 m/o NaVOg-NagSO^ solution at
900 C under a 0.1 % SOg-Og gas mixture flowing at a rate of 2.223 ml/sec. Open-circuit basicity equals -13.85. 170
Table 15. Chronopotentiometric Data From a Pure Pt Foil
WE Immersed In a 10 m/o NaVOg-NagSO^ Solution At 900 C
Under a Catalyzed 0.1% SOg-Og Gas Flowing At 0.223
ml/sec.
i E n ^ f ^ b (mA) (V)
107 - 1.230 + 0. 155 In {u> 0.65 4.80 114 - 1.220 + 0. 144 In (u) 0. 70 4.88 122 - 1.233 + 0. 159 In {u} 0. 64 5.00 130 - 1.234 + 0. 176 In {u} 0. 57 4.00 137 - 1.250 + 0. 155 In {u} 0.65 4.04 145 - 1.260 + 0. 162 In (u> 0. 62 3. 74 152 - 1.258 + 0. 176 In {u} o. 57 4.00
Table 16. Variation Of Peak Potential With Scan Rate For
Cyclic Voltammograms Recorded On a Pure Pt Foil WE
Immersed In a 10 m/o NaVOg-NagSO^ Solution At 900 C
Under a Catalyzed 0.1% S02~02 Gas Flowing At 2.223 ml/sec.
E t Ic ^IIc la E Ilar T E Iliar r T (mV/Sec) (mV) (mV) (mV) (mV) (mV)
26 - 1250 - 37 5 - 600 - 1950 46 - 1250 - 1600 - 375 - 550 - 1950
66 - 1200 - 1600 - 370 - 550 - 1925 86 - 1350 - 1625 - 350 - 525 - 1900 106 - 1300 - 1625 - 325 - 500 - 1900 200 - 1700 350 - 1875 Potential (V) - - - -1.4 -1.3 1.0 1.2 1.1 ng to E. . 3 Eq. o t g in d r o c c a Lgr 77. Plot of chronopotentiometric dat rm Fi 6 7 . ig F from ta a d c i r t e m o i t n e t o p o n o r h c f o t o l P . 7 7 FLigure 12 08 04 04 0.8 0.4 0 -0.4 -0.8 -1.2 In [(T 3 2-t /2 1 /2)/t 1 /2] 172
Figure 78. Cyclic voltammograms on a pure Pt WE immersed in a 10 m/o NaVO^-Na^O^ solution of basicity -13.85 at 900 C under a catalyzed
0.1" S02-02 gas -Flowing at 2.23 ml/sec. P = m V / s e c , A. 66, B.46, C. 26, D. 200, E. 106, F. 86. 173
The scan rate was varied from 26 to 200 mV/sec. At a scan rate of 26 mV/sec, three anodic peaks (labelled la,
Ila, and Ilia) are obtained at -375, -600, and -1950 mV, respectively. Peaks Ilia and la corresponds to the reoxidation of products of the cathodic decomposition of the supporting electrolyte, Na^SO^. One cathodic peak
(labelled Ic) at -1250 mV is recorded. At a scan rate of 46 mV/sec a cathodic peak (lie) is observed at -1600 mV, as well as a broader peak Ic at -1250 mV. As the scan rate in further increased, peak lie becomes more obvious and peaks la and Ila approach each other. At a scan rate of 200 mV/sec, peaks Ila and lie are well resolved, while peaks Ic and la are not. What results as peak 11a at 200 mV/sec is the merger of the peak corresponding to vanadium species (at -500 tO -600 m V ) with the peak corresponding to the oxidation of electroactive species from the NagSO^. Peak Ic is interferred by peak lie at such relatively high scan rates. The behavior of peaks Ic and lie with changes in the scan rate is characteristic of the redox reaction involving strong adsorption of the reactant. As the scan rate is increased, the adsorption-control peak precludes the time-dependent diffusion-control peak. At sufficiently fast scan rates, the adsorption control peak is the only one resolved. Peak Ic is a diffusion- control peak while peak Il'c is an adsorption-control 174
peak. Peaks Ila and Ic constitute a redox couple of an electroactive vanadium species present in the solution.
Figure 79 shows a chronopotentiogram for a pure Pt planar WE immersed in a 10 m/o NaVO^-NagSO^ solution at
900 C under a catalyzed 0.1% SO^-O^ gas flowing at a rate of 2.223 ml/sec. A forward transition time ,'TT^, of
1.5 sec is obtained when a negative current of 120 mA is applied between the working and the counter electrodes.
Upon current reversal, a reverse transition time, of
.45 sec is recorded. Table 16 lists the chronopotentiometric data obtained at various applied currents. The ratio for all applied currents is greater than 3, implying that the reduction reaction product is unstable in solution and undergoes a chemical reaction. A plot of potential, E, versus the time relationship In ~ t*^ )/t^//^], according to Eq.
('31 yields a straight line, indicating that the redox reaction is reversible, Fig.80. From table 16, the number of electrons transfered averages 0.62 +/- 0.12, nominally one. A diagnostic plot i v s i reveals a diffusion-control process at low applied currents and adsorption of the reactant at applied currents higher than 140 mA, Fig. 81.
Under these experimental conditions, the vanadium Potential (V) - 2.0 0 a 0 slto o bsct -38 a 90 udr ctlzd .% S02~02 0.1% catalyzed a . c e s / under l C m 3 900 .2 2 at t a -13.85 g in w lo f basicity of gas solution S04 Na2 Figure Figure 79. hoooetorm n pr P W imre i a 0 / NV - NaVO m/o 10 a in immersed WE Pt pure a on Chronopotentiogram 3.0 6.0 Time (sec) 9.0 12.0
15.0 5 7 1 176
Table 17. Chronopotentiometric Data From A Pure Pt Foil
WE Immersed In a 10 m/o NaVO^-Na^SO^ Solution At 900 C
Under a Catalyzed 0.1% SO^-Og Gas Flowing At 2.223 ml/sec.
i E n
(mA) (V)
110 - 1.034 + 0. 14b In iu) 0. 70 5. 56 lib - 0.988 + 0 . 162 In (u> 0 . 62 5. 20 120 - 1.008 + 0. 199 In {u} 0 . bl 5.00 12b - 0.940 + 0 . 183 In {u> 0 . bb 6 . 66 130 - 0.973 + 0. 167 In (u> 0. 64 3. 50 13b - 1.043 + 0.113 In iu) 0.89 5. 50 140 - 0.977 + 0 . 15b In {u} 0 . 65 14b - 0.996 + 0. 177 In (u) 0. 57 Potential (V) - - -0.9 - -1.3 1.0 1.2 1.1 ng to E. 3 Eg. o t g in d r o c c a gure e r u ig F . -. -. -. -. 0 . 0.4 0.2 0 -0.2 -0.4 -0.6 -0.8 1.0 0 8 . a fo . g i F from ta a d c i r t e m o i t n e t o p o n o r h c f o t o l P In T ( [ 1 t - 2 / 1 t / ) 2 / 1 /2] 79 ______I______I______1______I______I______I______I______100 120 140 160 180 i (mA)
F i g u r e 81 . Diagnostic plot from chronopotentiometric data from 10 m/o NaVO^-Na^SO^ melts of basicity -13.85 at 900 C. 179
electroactive species undergoes a 0.62, nominally taken
as one, reversible redox process followed by a chemical
reaction that consumes the reduction reaction product.
Cyclic voltammograms were recorded from melts in which
the added solute species were Figure 82 shows such
cyclic voltammograms for 10 m/o V20^-Na2SO^ solutions
under oxygen at 928 C, basicity -12.05. The
electrochemical response is similar to the one described
for NaVO,.-Nar SO . solutions. The redox peaks are «J C* ** observed at nearly the same potentials and as in the
case of relatively acidic NaVOg-NagSO^ solutions ,
strong adsorption of the reactant accompanies the redox
reaction. From Fig. 82 at a scan rate of 100 mV/sec,
even though a high current is associated with peak Ic, a very shallow anodic peak is observed. If a following
chemical reaction consumes the reduction reaction
product, at slow scan rates and at a switching potential of a few hundred mV past both the adsorption and diffusion-control peaks, a very shallow anodic peak is expected since the chemical reaction decreases the amount of reduced species available to be reoxidized in the vicinity of the working electrode. For the same
CSP, increasing the scan rate causes an increase in the current associated with the anodic peak because the chemical reaction becomes less significant. 1 8 0
Figure 82. Cyclic voltammograms on a pure Pt WE immersed in a 10 m/o V205-Na2S04 melt of basicity -12.05 at 1200 K under 02 1 8 1
Similar features are observed when Vo0t-Na^SOmelts 2 o 2 4 are stabilized with catalyzed 0.1 % SO^-O^ gas mixtures.
The open-circuit basicity was -12.83, Fig. 83.
Reaction Mechanism
1. NagSO^ Melts Under Catalyzed 0.1 % SOg-O^
Figure 84 shows the superposition of the Na-S-0
stability phase diagram and the basicity trace recorded
during the polarization of a Pt WE painted onto a
zirconia tube immersed in a NagSO^ melt under a catalyzed 0.1% S O ^ - O ^ atmosphere. The trace is nearly a vertical line suggesting that the basicity of the melt at the WE remains essentially constant. However, for each succeeding polarization cycle the basicity shifts toward a more basic value. At the open-circuit potential, according to the phase stability diagram, the 2 - SgO,j ion is the most dominant ionic solute in the melt,
S 20?2" ; “ S042" + S03 [40]
Oxide and superoxideions from the equilibrium decomposition of the melt are likely to be in equilibrium with dissolved oxygen in the melt according to the reaction,
0 2~ + 3/2 0 2 v 2 0 2" [41]
From Fig. 64 three significant anodic peaks are recorded when the CSP is extended enough to allow the cathodic 1 8 2
Figure 83. Cyclic voltammograms on a pure Pt WE immersed in a 10 m/o V?0r-Na2S0. solution of basicity -12.83 at 1200 K under a catalyzed 0.i% jO^-O^ gas mixture. Figure 84. Basicity trace on a polarized Pt WE painted on a zirconia tube immersed in a Na2S04 melt of basicity -13.46 at 900 C under a catalyzed 0.1% S02-02 gas . decomposition of the melt. The appearance of peaks I’a and I’c at less negative CSP is masked by the lower sensitivity in current used during the recording of the voltammograms. For Fig. 64 the current sensitivity is
50 mA/inch compared to the sensitivity of 8.33 mA/inch
in Figs. 21 and 45. The sensitivity during these studies was determined by previous recorded voltammograms from NaVO^-NagSO^ melts. Figure 85 shows the cyclic voltammogram simultaneously recorded on the
Pt WE painted onto zirconia as the basicity trace in
Fig.84 was registered. In the first cycle peak I’c was absent and peak II’c was not well defined. Upon reversal of the potential at -2250 mV, peak III’a and very well defined peaks II’a and l’a were obtained. In the second cycle after reversal of the potential at 600 mV, peaks I’c and II’c were well defined. Peak II’c, as discussed in previous cyclic voltammograms and chronopotentiograms, occurs at potentials very close to the potential at which the cathodic decomposition of the melt takes place and it was usually masked by the large current that accompanied the electrochemical decomposition process. It has been also clearly established that the current associated with peaks I’c and II’c is enhanced in succeeding cycles as the cathodic and anodic electrochemical decomposition of the melt are allowed to take place. Thus the electroactive 185
M’a
Ilia
50 mV
500 mV POTENTIAL
Figure 85 . Cyclic voltammogram recorded on a Pt WE painted on a zirconia tube immersed in a Na^SO- melt of basicity -13.46 at 900 C under a catalyzed 0.1% SO2 -O2 gas. 1 8 6 species associated with both peaks were either generated or were secondary products from both decomposition processes. The chronopotentiogram in Fig. 69 closely resembles the chronopotentiograms from Na^SO^ melts of basicity -6.66 and -10.07 (log a Na^O scale). The change toward more basic values relative to the open- circuit basicity of the melt immediately adjacent to the WE between polarization cycles, indicates the local generation of oxide species and consumption of dissolved oxygen during deep cathodic polarization. Therefore, the reaction mechanism is the same as the one proposed for NagSO^ melts under an oxygen gaseous atmosphere and for more relatively basic conditions,
e ---- 1 peak I ’c °2 + °2 i + o 0 2" peak II’c
re 2 + 2e- ---> S02 + 2 o2- cathodic decomposition
2 02~ — 2 e peak III’a °Z + 1 1 0 + e - peak 1 1 'a °22 - + e- peak I ’a °2 o2
2. NaVO^-Na^SO^ Under 0.1% SOg-Or, Gaseous Atmosphere
According to the phase diagram, for the partial pressure of oxygen and the basicity of the melt at open- circuit potential, VOg are the ionic vanadium species stable in the NaVOg-NagSO^ solutions . The basicity trace recorded during polarization of a Pt WE painted onto a zirconia tube immersed in a 10 m/o NaVOg-NagSO^ 187
melt under uncatalyzed 0.1 % SO^-O^ gas suggests that
the basicity of the melt at the WE does not change
appreciably during polarization , Fig. 8 6 . As previously
discussed, NaVO^-Na^SO^ melts are buffered by a chemical 3- equilibrium in which VO^ constitutes the anionic
2 - solute resulting from the complexing of 0 with VOg
ions,
VOg" + 0 2~ . — — V043- [42]
The results from cyclic voltammetry and
chronopotentiometry suggest a one electron reversible redox process. Considering VOg ions as the vanadium electroactive species, the reduction reaction previously proposed is
VOg" + e------* V0 32' t43J
The product of this reduction reaction is suggested to
2 - be unstable in solution. VOg can undergo a chemical 3- reaction to produce VO^ and keep the basicity constant through equilibrium [42],
VOg2" + 02" ---- > VO,3" + 1/2 02 [44]
The redox reaction is reversible; therefore, for the anodic process, the following reaction holds,
VOg2" ---- * VOg" + e- [45]
For the NaV0g-Na2S04 solutions under an uncatalyzed 0.1%
S0,-,-02 gas mixture the basicity at the open-circuit potential and the basicity trace are essentially the Figure 86. Basicity trace on a polarized Pt WE painted on a zirconia tube immersed in a 10 m/o NaV03~Na2S04 solution of basicity -11.72 at 900 C under an uncatalyzed 0.1 % S02-02 gas mixture. .189 same as under 0 2 gas. For polarization studies where the 0.1 % SO^-Og gas mixture was passed through a platinized catalyst placed in the gas inlet tube just above the melt, the open-circuit potential basicity shifts toward more acidic values. A ten-fold increase in the flow rate for the same gas composition results in an decrease of one order of magnitude for the melt basicity. Under a catalyzed gas, the melt basicity for a gas flow rate of .223 ml/sec was between -12.8 to
-13.15. This basicity combined with the partial pressure of oxygen in the melt at the open-circuit potential suggests from the stability diagram that VO^ are the dominant ionic vanadium species in the melt.
The melt chemistry is close to the NaVO^-VgO^ line equilibrium. For a flow rate of 2.223 ml/sec the basicity of the melt was in the range -13.8 to -14.8, which is the basicity range for the boundary of equal activities of NaVOg and in the stability diagram.
The presence of a platinized catalyst in the gas phase inlet accelerates the reaction between SC>2 and C>2 to produce SO^. An increase in SO^ dissolved in the melt will result in the decomposition of NaVO^ to increase the activity of ^ 2^5 m e ^'t according to the reaction,
S03 + 2 NaVOg -----^ Na2S04 + V2 2 V(J3 ~ -----* V20 5 + O 2- [47] is a widely used catalyst in the production of sulfuric acid in to accelerate the reaction of SC>2 and Og to yield S O ^ . ^ 2^5 ‘*'s an n ~^ype oxide and it provides electrons in the catalytic process (41), V20 5 + S02 S03 + V204 Step 1 V204 + 1/2 0 2 V20 5 Step 2 S02 + 1/2 ° 2-;---1 SP3 Over All electrons passing from the catalyst to the absorbed species. Basically, the catalytic process is an electron exchange reaction at the active catalytic sites provided by the ^ 2^5 generally used in the molten state at temperatures ranging from 400 to 600 C. At our working temperature it can operate as a catalyst as well. Since a change in the flow rate is the only variable, while the composition of the gas is kept the same, reaction [46] should be limited by the arrival of SOg-Og (S03 is very dilute) at the melt surface. As the flow rate is increased, the rate at which the gaseous reactants are supplied is also increased and the activity of ^2^5 m e l’t increases through reaction [46]. The more the greater the number of active catalytic sites and the more S0 3 dissolved in the melt. Consequentely, relatively higher acidity. Electrochemical studies in NaV03~Na2S04 solutions of basicity -13 and lower reveal a one electron transfer 191 reversible redox process. The open-circuit basicities suggest VOg as the most probable vanadium ionic species as in melts as basic as -9.7. Therefore, the electroactive species are the same and they also undergo a one electron reversible redox reaction, VOg" + e- ----* V0 3 2' £48J For NaVOg-NagSO^ solutions of basicity - 13 and lower the equilibrium is shifted to the NaVOg-VgOg boundary line. Under this condition the basicity of the melt is buffered by some activity ratio of NaVOg and VgOg and NaVOg is present in the melt in higher concentration than SOg dissolved gas. Figure 87 shows the basicity trace during polarization of the Pt painted WE . The basicity at the WE increases drastically as the WE is polarized. During polarization of the WE the basicity trace goes over the stability field; therefore, solid ^ ^ 4 expected thermodynamically to be precipitated at the WE. Cyclic voltammetry and chronopotentiometry under these experimental conditions suggest that the reduction reaction product is not stable in solution and it decomposes according to the following reaction, VOg2" -----^ V02 + 02" [49] According to reaction [49] the basicity is locally increased at the WE. However, VOg has not been successfully detected by X-ray analysis of frozen Log p0j 0 1 - ue mesd n 1 mo aO-aS^ ouin f aiiy 1.5 at -13.85 basicity of solution NaVO^-Na^SO^ m/o 10 a in immersed tube 0 C ne a aaye 01 S^O gs itr foig t .3 ml/sec. 2.23 at flowing mixture gas SO^-O^ 0.1% catalyzed a under C 900 Figure Figure 87 •> . Basicity trace on a polarized Pt WE painted on a zirconia zirconia a on painted WE Pt polarized a on trace Basicity . 8 - L o g a Nap ' ' ' x 2 1 \ N \ 6 1 ' - 2.0 -1.5 \ - -0.5 1.0 0.5 0 E(mV)Ag/Ag CD ro 1 9 3 samples taken from these acidic melts after polarization of the WE. VOg has a dark gray color and stains of this color have been found in the bottom of the crucible around the WE. The color of the NaVO^-Na^SO^ solutions at basicity —13 and lower is brown-orange similar to the color observed when V ri0,- is added to Na^SO. , qualitative z o 2 4 evidence for the presence of ^ 2^5 unc^er this extreme conditions. The color of NaV0o-NaoS0. solutions of 3 2 4 higher basicity is similar to the color of NaVO^, light cream color. Previous studies in VgO^-NagSO^ solutions (42) indicated an increase in the solubility of SO^ with additions of V r,0,- t o Na,-,80. melts. Such results should 2 5 2 4 be expected because of the catalytic action of Vo0t in 2 5 the production of SO^. AC impedance measurements experiments were performed to characterize the WE/solution interface in terms of its equivalent electrical circuit. This technique provides information regarding to the reaction mechanism at the interface. The effect of vanadate addition on the conductivity of Na,SO. was also evaluated. 1 9 4 AC Impedance Resu1ts Un The Ft Painted Electrode 1. Basic NaVO^-Na^SO^ Solutions Figure 88 shows the compilation of AC impedance results on a Pt WE painted onto a zirconia tube immersed in a 10 m/o NaVOg-NagSO^ solution of basicity - 9.66 at 900 C. The Nyquist plot at high frequencies shows a straight line forming an angle of 57° with the real axis. The intercept with the real axis gives an ohmic resistance of 1.8 8 Q Figure 88 (b) shows the Bode plot for data gathered from Fig. 88 (a). A linear section of slope -.36 is obtained with a corresponding phase shift angle of 32.5°, indicating diffusion control at the electrode-solution interface. At frequencies lower than 0.1 Hz a second time constant contributes to the interfacial impedance. A Randles-type plot, Fig. 88 (c) for frequencies of 0.66 Hz and lower shows the expected linear relation for diffusion-control confirming the prediction from the Bode plot. A Warburg- type impedance of 1 0 fi is obtained for this diffusion controlled stage. For the diffusion-control stage predicted for frequencies of 1 Hz and higher a Randles plot gives a Warburg-type impedance of 1.5 fi 2. Na^SO^ Basic Melt Figure 89 shows the results from AC impedance on a Pt WE painted onto a zirconia tube immersed in a Na 2S° 4 19b 60 N 20 T (a) 20 40 60 80 ZR(Ohm) log log IZI 80 - ( b ) _____ i_ 60 6 2 40 cr N 20 (c) Fioure 88. AC impedance results on a Pt WE painted on a zirconia tube immersed in a N a V C ^ - N a ^ solution of basicity -9.77 at 900 C 19 b 4 0 (a) 60 8 0 30 INI -056 20 35 ( b ) log W 1210 14 16 >t WE painted on a zirconia )n of basicity -9.77 at 900 C. 2 0 0 150 E _ c o 100 — Kl 3 50 (a) 40 80 120 160 200 240 2 N ZR (Ohm) O' O -052 200 0 E CX N 100 0 2 4 6 8 10 12 14 16 -i/e l/W ,/f (Hz) Figure 89. AC impedance results on a Pt WE painted on a zirconia tube immersed in a Na^SO^ solution of basicity -6.66 at 900 C. 3 60 (a) ( b ) 50 120 160 200 240 2 40 ZR (Ohm) 30 -052 I 20 0 log W 8 10 12 14 16 ts on a Pt WE painted on a zirconia tion of basicity -6.66 at 900 C. 197 melt of basicity -6.66 at 900 C. The Nyquist plot in Fig. 89 (a) shows a straight line 45° with respect to the real axis. Figure 89 (b) shows the same data in the Bode format. A linear section of slope -.5 is obtained with a corresponding phase shift of 50° indicating diffusion-control at the electrode- solution interface as predicted by the Nyquist plot. A Randles plot in Fig. 89 (c) shows a linear relationship between the real component of the interfacial impedance and the inverse of the square root of the frequency. Such a linear relationship is expected for diffusion-control at the interface. A Warburg-type impedance of 9 is calculated from the intercept with the ordinate. From the Bode plot the ohmic resistance is calculated as 2.17 fi . 3. NaVO^'Na^SO^ So-Luti01"15 Under Oxygen Figure 90 collects the results from AC impedance measurements on a Pt WE painted onto a zirconia tube immersed in a 10 m/o NaVO^-NagSO^ solution under . The Nyquist plot in Fig. 90 (a) suggests electron transfer control at the electrode-solution interface as indicated by the semicircle obtained at high frequencies. Close examination of the semicircle on a more sensitive scale shows that the semicircle is a combination of two semicircles. The smaller semicircle is not centered in the real axes indicating gure 9. C mpe e results o a W painted o a a i n o c r i z a on d e t n i a p WE t P a on s t l u s e r ce n a ed p im AC 90. e r u ig F ube i re i NV--aS^ 7 at 90 C. 900 t a .7 1 1 - y t i c i s a b f o n o i t u l o s NaVO-^-Na^SO^ a in ersed m im e b tu log |Z| Zl (Ohm) 0 2 2 2 o - 6 - 0.25 10 HD- 0 14 ZR (Ohm) log W -030 18 2 26 (a) 30 4 20 30 40 CO QL O N JZ £ M r c O sz E 18 30 22 22 26 10 14 0 0 ' ____ L E JZ O cr M (a) 22 26 3018 0.2 0.4 0.6 3 (Ohm) Zl/W 40 30 30 26 E 20 _c O q: m - 0381 ( d ) og W 0 2 3 4 5 l/W 1'2 (Hz)'1'2 !:s on a Pt WE painted on a zirconia j solution of basicity -11.7 at 900 C. 1 9 9 irregularities at the surface of the electrode. The second semicircle is obtained because of adsorption at the WE. The interception of the first semicircle with the real axis provides the ohmic resistance of 2.6 fl . The polarization resistance is calculated from the diameter of the first semicircle as 8.89 fl . From a plot of ZR vs ZI/W, the capacitance of the double layer is estimated as 1.07 xlO-2 fj F. The data plotted in the Bode format do not support electron transfer control. but rather two diffusion-control stages with linear portions of slope -.38 and -.28, respectively. The phase shift angles are 32.5° and 24° . A Randles plot confirms the two diffusion-control stages, Fig. 90 (d). For frequencies smaller than 0.9 Hz a Warburg-type impedance was calculated as 13.25 fl . For frequencies higher than 0.9 Hz, an interfacial impedance of 13.75 fl is calculated from the intercept with the ordinate for the second linear relation in Fig. 90 (d). 4. Na^SO^ Melts Under Og Figure 91 shows the results of the AC impedance data obtained on a Pt WE painted onto a zirconia tube immersed in a Na^SO^ melt at 900 C under 0^. The Nyquist plot in Fig. 91 (a) shows a straight line making an angle of 52° with the abscissa. The intercept with the real axis gives an ohmic resistance of 2 fl The ube i re i a aS^ 0 C. 900 t a a i n o . c 0 r i 1 z - a y t on i c i s d e a t b n i a f p o WE t P n o i a t u l on o s s t Na^SO^ l u a s e r in ce n a ed ersed p m im im AC e b tu 91. e r u g i F 100 M O xz E ZR(Ohm) 200 0 0 3 100 60 80 20 0 0 200 / 1'2l/W (Hz)-1'2 ZR(Ohm) 2 400 ZR(Ohm) 3 600 4 2 -3 Oll/l 9 A O 6 0 200 400 600 2 4 0 ZR (Ohm) N -0.55 20 ( b ) 3 3 log W results on a Pt WE painted on a zirconia j solution of basicity -10. at 900 C. 2 0 1 data plotted in the Bode format shows a linear portion at intermediate frequencies with a slope of -.bb and a corresponding phase shift of 53.8b°. The trend of the data in Figs, 91 (a) and (b) suggest that diffusion controls the electrochemical process at the electrode solution interface. A Randles plot confirms this prediction. A Warburg-type impedance of 11.11 f) is calculated from the intercept with the ordinate. AC Impedance_ResuIts On The Ft Foi 1 _Ele ct_ro.de 1. NaVO^-Na^SO^ Basic Solutions Figure 92 shows the AC impedance data collected on a Ft foil WE immersed in a 10 m/o NaVO^-Na^SO^ solution of basicity -9.66 at 900 C. The Nyquist plot shows a straight line at high frequencies. Close examination on a more sensitive scale reveals a semicircle that is distorted by a straight line (diffusion-control). The intercept of the semicircle with the real axis gives an ohmic resistance of 1.27 O . The diameter of the semicircle gives a polarization resistance, Rp, of 0.41fl From a plot of ZR vs ZI/W, Fig. 92 (b), the capacitance of the double layer is calculated as b.3bxl0-3 fJF. The Bode plots in Fig. 92 (c) shows a linear portion at intermediate frequencies of slope -.5 and a corresponding phase shift angle of 46.94° , suggesting diffusion of the electroactive species as the aO-aS^ 0 C. 900 t a 7 7 . 9 - y t i c i s a b f o n o i t u l o s NaVO^-Na^SO^ 2 A i e results o Pt E mmesd n a 0 m/o 10 a in ersed m im WE l i o f t P on s t l u s e r ce n a d e p im AC 92. e r u g i F N o ' O Zl (Ohm) 140 100 120 80 60 20 0 50 3 2 -0.4 (a) 100 -0.5 ZR(Ohm) o W log 0 0 400 300 200 40 20 30 05 cr N E N r c _c E 2 0 2 400 200 300 100 2 0 2 1.6 1.5 E sz 1.4 c r N 1.3 ?0 ( b ) ZR (Ohm) 1.2 200 300 400 0.0002 0.0004 0.0006 0 0 0 0 8 ZR (Ohm) Zl/W 5 0 4 0 0 4 0 3 0 0 3 0 E xz CD 2 200 c r 20 N -0.5 100 ( d ) log W Its on Pt foil WE immersed in a 10 m/o city -9.77 at 900 C. 2 0 3 controlling process at the electrode-solution interface. At frequencies of 0.3 Hz and smaller, a second time constant is revealed by the trend of the data in the Bode plot because of diffusion-control. A Warburg-type impedance of 25 fl is calculated from the Randles plot in Fig. 92 (d). 2. Basic Nar.S0. Melts 2 4 Figure 93 shows the Ac impedance results on a Pt foil electrode immersed in a NagSO^ solution of basicity -9.66 at 900 C. The Nyquist plot in Fig. 93 (a) shows a straight line 45° with respect to the real axis. The Bode plot in Fig. 93 (b) reveals a linear portion with a slope of -.50. The corresponding phase shift plot shows two time constants at 48.33° and 47.22°. The value of the slope and the phase shifts confirms the pred< n of diffusion control at the electrode-so xon interface. A Randles plot for frequencies smaller than 1 /2 1 Hz reveals a linear relationship between ZR and W as expected for a diffusion control-process The Warburg-type impedance was calculated as 35.7 fl . 3. NaVUg-NagSO^ Under 0 r Figure 94 collects the AC impedance results on a Pt foil WE immersed in a 10 m/o NaVO^-NagSC)^ solution at 900 C under 0^■ The Nyquist plot at high frequencies shows a semicircle that is distorted by a straight line 2 0 4 8 0 0 600 E x z 2 400 N 200 (a) A 200 400 600 800 1000 ZR (Ohm) n 2 o> o 200 ( b ) cc M 100 0 2 3 4 1/W1/2 (Hz)'1'2 Figure 93. AC impedance results on a Pt foil WE immersed in a Na?SO. solution of basicity -6.66 at 900 C. 2 0 4 60 (a) 3 400 800600 1000 40 ZR (Ohm) 2 -05 ( b ) 20 0 3 3 5 log W 2 3 4 /W "2 (Hz)"'2 ts on a Pt foil WE immersed in a Na2S04 900 C. 80 1.4 (a) 60 E 6 40 3 M 20 2. 0 40 80 120 ISO 200 240 280 ZR(Ohm) 3 60 20 50 15 40 2 -0 30 E N JZ 30 O CT> CD 10 O or 20 Nl -0.45 jjgifP, 0 0 -10 0 log W Figure 94. AC impedance results on a Pt foil WE immersed in a NaV03~ Na2S04 solution of basicity -11.77 at 900 C. 1.4 3 ( b ) 2. 160 200 240 280 O 0.0002 0 0 0 0 4 0 0 0 0 6 Ohm) Zl/W 6 0 20 5 0 4 0 E 3 0 jC O CO c r 20 N ( d ) -10 0 0.1 0.2 0.3 0.4 l/W "2 (Hz)-1'2 a Pt foil WE immersed in a NaVO^- a t 900 C. indicating diffusion control. The intercept of the semicircle with the real axis gives an ohmic resistance of 1.23 fl The polarization resistance given by the diameter of the semicircle was calculated as .89 . The capacitance of the double layer was calculated from Fig. 94 (b) as 1.34x10-3/I F. The Bode plot in Fig. 94 (c) reveals two time constants as indicated by the slope of the linear portions and the corresponding phase shift angles because of diffusion control. A Randles plot for frequencies lower than 1 Hz suggests a Warburg- type impedance of 16 fl . For frequencies higher than 1 Hz, the Randles plot reveals a Warburg-type impedance of 1.44 n. 4. Na^SO^ Melts Under 0^ Figure 95 shows the results obtained from AC impedance measurements on a Pt foil WE immersed in a NaoS0„ melt at 900 C under 0o. The Nyquist plot in Fig. Cj H. Ca 9b (a) shows a straight line with a 4b° inclination respect to the real axis. A close examination of the Nyquist plot at higher frequencies shows a semicircle segment that becomes a straight line indicating diffusion control at the interface The intercept of the semicircle with the real axes gives an ohmic resistance of 1 . 35 fl . The polarization resistance given by the diameter of the semicircle is calculated as .90fl A double-layer capacitance of 2.12x10-4 F is 2 0 7 ( a ) 1000 E jC E O xz O N c r 500 N 0 4 ot 500 1000 1500 2000 ZR (Ohm) 3.4 7 0 6 0 0 "r~T 3.0 6 0 5 0 0 2.6 5 0 — 4 0 0 2.2 E E XI 4 0 - 0.6 O 2 3 0 0 05 cr N 3 0 N cn o 200 20 06 (c) -07 100 0.2 o. 0 0.4 log W Figure 95. AC impedance results on a Pt foil WE immersed in a NaoS0 solution of basicity -10.07 at 900 C. 2 4 2 0 7 E x z O OC N ZR(Ohm) D 150 0 2 0 0 0 0.001 0.002 0.003 0.004 D h m ) Zl/W 7 0 6 0 0 6 0 5 0 0 5 0 — 4 0 0 E 4 0 2 3 0 0 ce 3 0 N 200 20 -07 100 0.4 0.8 2.0 2.4 2.8 W i a Pt foil WE immersed in a NaoS0 3 C. 2 ' 2 0 8 calculated from Fig. 95 (b). The Bode plot in Fig. 95 (c) reveals two time constants with phase shift angles of 53.85° and 60.77°. The slope of the linear portions and the phase shift angles suggest two stages of diffusion control. A Warburg-type impedance of 40ft is calculated for frequencies smaller than 1 Hz. For frequencies greater than 1 Hz an interfacial impedance of 6 .6 6 ft was calculated from Fig. 95 (d). Discussion Of AC Impedance Results Results from AC impedance measurements indicate that for both types of electrodes diffusion of the reactants to the working electrode surface controls the redox process . For diffusion control a linear relationship i s expected between the real component of the impedance and the inverse of the square root of the angular frequency of the alternating current. Randles plots confirm this relationship. In particular cases, from both the Pt painted and Pt foil working electrodes, two stages of diffusion control were observed. Most of the Bode plots reveal phase shift angles higher than 45° for angular frequencies higher than 1 Hz. The corresponding Randles plots reveals the same linear relationship expected for diffusion-control. The Faradaic impedance (diffusion-control) of an electrode to a.c due to an electrochemical reaction as 2 0 9 Ox + e- ---* Red can be represented by a resistance, R ^ , and a capacitance, C , in series, which are related to the r concentration of the reactant species, their diffusion coefficients and the a.c angular frequency. The phase angle, 6 , must be less than 45° , this value only being approached as the heterogeneous rate constant of the electrode reaction approaches infinity. Phase angles exceeding 45° have been reported (43, 44) to arise from the presence of reactants adsorbed on the electrode surface. Under this condition a linear relationship as the predicted for purely diffusion control is obtained and the interfacial impedance must be corrected for this additional impedance contribution from adsorption . The effect of adsorption of the reactant also distorts the expected Nyquist plot for pure diffusion-control. For this particular condition the equivalent circuit is drawn as follows, Rfi Ra C a -VWv< VW*'------1 \- "W" where R and C are the resistance and capacitance from a a the contribution of the adsorbed species. Under particular experimental conditions the trend of the data in the Nyquist plot in the frequency range 1500 to 10,000 Hz suggests the possibility of electron 2 1 0 transfer control. Double layer capacitances were calculated and their values are three orders of magnitude lower than the double layer capacitance estimated for other sodium molten salts (4b). This stage of electron transfer-control was not supported by the trend of the same data in the Bode format disregarding a significant heterogeneous rate constant. In the absence of adsorption of the reactant the electrode-solution interface can be represented by a simple circuit composed of the ohmic resistance in series with a Warburg type resistance, -V'Aa I*--- Table 18 lists the ohmic resistance for the various melt conditions on the two working electrodes. For the painted Pt WE the ohmic resistance averages 2.16 +/-0.27fl The addition of vanadates does not change the ohmic resistance of the supporting electrolyte. B'or the Pt foil WE the ohmic resistance averages 1.26 +/- 0.06 Q No change was observed in the ohmic resistance because of the addition of vanadates to molten NaoS0.. This 2 4 result is expected because of the relatively low concentration of NaVO^ in the solutions. The difference in the average value of the ohmic resistance between the two electrodes is because of their relative position with respect to the reference electrode. The Pt foil was positioned 0 .1 b cm from the reference electrode 2 1 1 Table 18. Ohmic Resistance From AC Impedance Measurements. R n (Ohm) Painted Pt WE Pt Foil WE NaVO - 0 2 . 60 1 . 23 O o NaVOg- Basic 1 . 88 1 . 27 Na2 S ° 4 - 0 2 2.00 1 . 3b NaoS0 - Basic 2.17 1 . 18 2 4 2 1 2 while the Pt painted eletrode was placed approximately 1 cm from the reference electrode. The ohmic resistance is equivalent to the umcompensated resistance and increases with the distance between the working and reference 'electrodes. Ac impedance measurements for increasing amounts of NaVO, in NarSO. are encouraged in order to evaluate the O Lj fl effect of NaVOQ additions on the conductivity of Na^SO. 3 2 4 and on the double layer capacitance at the electrode solution interface. Such studies can reveal the predominant ionic species contributing to the double layer. CONCLUSION Vanadate ions, VOg , are the eiectroactive species in NaVOg-NagSO^ and V^O^-Na^SO^ solutions at 900 C. Vanadium pentoxide, V^,0g, and sodium vanadate, NaVOg, are acid compounds that have the ability to dissolved most of the protective oxide scales formed on high temperature alloys. Their detrimental effect was thought to be only confined to the fluxing of these oxides. Electrochemical studies in these melts revealed that vanadate ions undergo a one electron reversible redox reaction . Vanadium species in two oxidation states present in sufficient concentration in the melt can shift the reduction reaction site to the salt/gas interface by the counter diffusion of multivalent cations. Under this circumstance, the oxidizing agent does not need to diffuse through the film salt to be reduced at the oxide/salt interface and rapid oxidation of the metal can occur. Sodium sulfate is a suitable supporting electrolyte for the study of the electrochemistry of vanadates, providing that the cathodic switching potential in 213 cyclic voltammetry studies is limited to -1500 mV relative to the Ag/Ag mullite high temperature reference electrode. The sulfate ions are electrochemically stable in the potential range 500 to -1900 m V . Beyond these potential limits the anodic and 2 - cathodic decomposition of SO^ ions takes place. The 2 - cathodic decomposition of SO^ ions locally generates oxide ions and related species that can be reoxidized and interfere with the anodic response of some other solutes present in the solution. As concluded by ac impedance measurements, diffusion of the electroactive species controls the reaction at the electrode/solution interface for both WEs used in the present study. The Helmholtz-Stern model for the structure of the double layer does not apply to the NaVO^-Na^SO^ and pure NagSO^ solutions. Adsorption of the reactants accompanies the redox process. No significant change in the ionic conductivity of the electrolyte was obtained upon addition of NaVO^ to REFERENCES 1. R. J. Clark, "The Chemistry Of Titanium & Vanadium ", Elsevier Publishing Company, Amsterdam, 1968 2. National Academy Of Sciences, "Vanadium, Medical and Biological Effects Of Enviromental Pollutants", Wahsington D.C. 1974 3. C. Sykes And H. T. Shirley, "Simposium on High Temperature Steels and Alloys for Gas Turbines", Iron and Steel Institute. Spec. Rep. No. 43, 1952, 153-169 4. A. Preece, ibid, 1952, 149 5. ibid, 1952, 107 6 . F. C. Monkman and N. J. Crant, Corrosion, Vol. 9, December, 1953, 460-465 7. W. C. Leslie and M. G. Fontana, Trans. A.S.M., Vol. 41, 1949, 1213 8 . E. Filzer and J Schawb, Corrosion, Vol. 12, September, 19556, 459t 9. K. Hauffe, Wiss. Z. Univ. Greifswald, 1, 1951-1952, 1 10. K. Sachs, Metallurgia, Vol. 57, 1958, 167-173 11. G. W. Cunnigham and A. de S . Brasunas, Corrosion, Vol. 12, August, 1956, 35-51 12. A. de S. Brasunas, "Accelerated Oxidation of Metals at High Temperature". Massachusetts Institute of 215 2 1 6 Technology, Thesis (Sc.D) 1950 13. W. R. Foster, M. H. Leipold and T. S. Shevlin, Corrosion, Vol. 12, 1956, 539t 14. P. Schlapfer, P. Amgwerd and H. Preis, Schwizer Archiv Fur Angewandt Wissenschaft und Technik, October, 1949, 291-299 15. J. A. Goebel, F.S. Pettit, and G. W. Goward, Met. Trans., V o l . 4, 1973, 261 16. Y. S. Zhang and R. A. Rapp, "Solubilities of CeOg, HfOg and in Fused NagSO^-SO m/o NaVO^ and CeOg in pure NagSO^ at 900 C", to be published in Corrosion 17. M. Kawakami, K. S. Goto and R. A. Rapp, Trans. ISIJ, Vol.20, 1980. 646-658 j 18. R. A. Rapp and K. S. Goto, Proc. "2 International Symposium on Molten Salts", J. Braunstein Ed. The Electrochemical Society, 1979, 159-177 19. A. J. Bard and L. R. Faulkner, " Electrochemical Methods, Fundamentals and Applications", John Wiley & Sons, New York, 1980 20. R. S. Nicholson and I. Shain, Anal. Chem., Vol. 36, No.4, 1964, 706-723 21. R. H. Wopschall and 1. Shain , Anal. Chem., Vol. 39, No. 13, 1967, 1514-1527 22. D. G. Davis, "Electroanalytical Chemistry" Ed. A. J. Bard, V o l .1, 1966, 157 23. C. H. Liu, K. E. Johnson, H. A. Laitinen, "Molten 2 1 7 Salt Chemistry", Ed. Milton Blauder, John Wiley &. Sons, New York, 681 24. C. F. Furlani and G. Morpurgo, J. Electroanal. Chem. Vol.1, 1959/60, 351-362 25. W. H. Reinmuth, Amal. Chem., Vol.32., 1960, 1514 26. I. M. Kolthoff and P. J. Elving, "Treatise On Analytical Chemistry " Part 1, Vol.4, Section D, Chapters 42-44, Interscience Publishers, INC. New York, 1963 27. P. Delahay, " New Instrumental Methods In Electrochemistry", Interscience Publishers, INC, New York, 1954 28. F. Mansfeld, Corrosion, Vol. 36, No. 5, May 1981, 301-307 29. F. Mansfeld, M. W. Rending, and S. Tsai, Corrosion, Vol. 38, No. 11, November, 1982, 570-580 30. D. D. MacDonald,"Transient Techniques in Electrochemistry", Plenum Press, New York, 1977 31. R. S. Nicholson, Anal. Chem. Vol. 36, N.4, April, 1964, 1406 32. Chon Ook Park, "Electrochemical Reactions in Molten NagSO^ at 900 C" Ph. D dissertation presented at The Ohio State University, 1985 33. P. G. Zambonin, J. Electroanal. Chem. Vol. 45, 1973, 451 2 1 8 34. J. D. Burke and D. H. Kerridge, Electrochim. Acta, Vol. 19, 1974, 251 35. D. R. Flinn and K. H. Stern, J. Electroanal. Chem.,Vol. 63, 1975, 39 38. K. H. Stern, Rm Panayappan and D. R. Flinn, J. Electrochem. S o c ., Vol. 124, No. 5, 1977, 641-649 37. M. L. Deanhart and K. H. Stern, ibid. , Vol. 127, No. 12, 1980, 2600-2602 38. W. C. Fang and R. A. Rapp, ibid., Vol. 130, No. 12, 2335-2341, 1983 39. C. 0. Park and R. A. Rapp, ibid, Vol. 133, No.8 , 1986, 1636-1641 40. R. E. Andresen, ibid, Vol.126, 1979, 328 41. J. J. Carberry,"Chemical and Catalytic Reaction Enginnering", Me. Graw-Hill, Chemical Engineering Series, USA, 1976 42. W. C. Fang, "Reaction mechanisms in molten NagSO^ containing ^£^5 anc* NagSO^ containing CoSO^", Ph. D dissertation presented at The Ohio State University, 1982 43. H. A. Laitinen and J.E.B. Randles, Trans, of the Faradaic Soc., Vol 51, 1955, 54-62 44. S. K. Rangavajan and K.S.G. Doss, J. Electroanal. Chem., V o l . 4, 1962, 237 45. A. D. Graves, G. J. Hills and D. Inman, "Electrode Processes in molten Salts". Advances in Electrochemistry 2 1 9 and Electrochemical Engineering. Vol. 4. Paul Delahay Editor, -John Wiley & Sons, USA, 1966