Transport properties in molten sodium sulfate at 1173 K

Solid State Ionics 45 ( 199 I ) 165-l 72 North-Holland

Transport Dong-Hyung

properties

in molten sodium sulfate at 117 3 K

Kim and George Simkovich

206 Steidle Building, The Pennsylvania State &tiversity, tJniversity Park, PA 16802, US4 Received I3 November 1990; accepted for publication 17 December 1990

In order to aid in further understanding hot corrosion processes, investigations of electrical behavior of molten Na,SO., have been undertaken. An ac impedance technique was employed to obtain the total electrical conductivity of a pure Na2S04 melt at 1173 K. The steady state polarization method of Wagner and Hebb was applied to study partial electronic conductivities of electrons and electron holes in a pure Na,SO, melt. The potentiostatic polarization technique was carried out to decide ionic transport numbers in the melt. These experiments are conducted as a function of Na,O activity in the melt by controlling the experimental gas atmosphere. It was observed that the total electrical conductivity of pure Na2S04 was of order of 2.32 X IO-’ (Q cm)-’ and varied only slightly with changes in the activity of NalO. From the Wagner-Hebb type dc polarization experiments on pure Na2S04, the electron conductivity was shown to be much greater than the electron hole conductivity over the entire range of NazO activities. The partial conductivity of electrons in NazSOs was about two orders of magnitude less than the total electrical conductivity. Thus, the transport number of electrons, tO,is of the order of lo-’ in a pure Na,SO, melt at 1173K. From the potentiostatic polarization technique, the cation transport number of a pure sodium melt was found to be about 0.985 at 1173 K. The self diffusion coefficient and the mobility of Na ions in the melt was estimated by using the above experimental data and their values were about 8.1 x IOe6 cm2/s and 8.0 x I Oe5 cm2/V s, respectively.

1. INTRODUCTION Hot corrosion is generaLly defined in broad terms as an accelerated or catastrophic oxidation of alloys and other materials. This form of attack is particularly severe in the temperature range of 1033 - 1273 K, and it has affected both aircraft engines and industrial gas turbines. There is a general agreement that condensed alkali metal salts, notably Na2S04, are a prerequisite to hot corrosion. The source of this salt may be (a) the direct ingestion of sea salt in a marine environment, (b) the formation of Na2S04 during combustion of fuels containing both sodium and sulfur, (c) the formation of Na2S04, during combustion, from sodium - contaminated, airborne dust and sulfur in the fuel [ 11. The exact mechanisms of hot corrosion are still uncertain, but from many studies on the hot corrosion mechanisms of metals and alloys, the various mechanisms that have been proposed can be broadly classified into two categories : (a) acidic - basic fluxing models [2-4) and (b) dissolution reprecipitation electrochemical model (51. The overall mechanisms of hot corrosion involve the dissolution of normally protective oxide layers and the formation of porous, nonadherent, and hence unprotective scales when alloy surfaces are covered by a thin fiim of liquid sodium sulfate. It appears that the initial formation of metal oxides is necessary for the reaction and the transport of oxygen through the molten salt phase is required to form such metal oxides. Little is known about the electronic transport properties in molten Na2S04. This study is, therefore, concerned with Elsevier Science Publishers

B.V. (North-Holland)

obtaining such information to aid in the elucidation of the mechanism of the process. The partial electronic conductivities by the Warner-Hebb tvue uolarization studies [6,7] as well as-the total &ctrical conduciivities utilizing an A.C. impedance technique were measured on molten Na2S04 as a function of Na20 activities at 1173 K since the proposed models describing the degradation behavior of alloys are strongly dependent 011Se, @, and /or SO3 gas pressure, i.e.. the NaZO activity in the Na2S04 deposit. In addition a potentiostatic polarization technique was employed to estimate the ionic transport numbers of a pure sodium sulfate melt at 1173 K. The relationship between the conductivity and the diffusivity, known as the NemstEinstein equation, provides the self diffusion coefficient and the mobility of Na ions in the molten Na2S04. The transport numbers of electronic species in the Na2S04 melt were also evaluated.

2. THEORETlCAL

CONSIDERATIONS

(i) Total Conductivity Measurement A.C. impedance measurements were conducted to obtain polarization free total electrical conductivity of molten sodium sulfate. In D.C. techniques, a space charge of either ions or electrons form in the vicinity of the electrodes which leads to a non-uniform field across the specimen. However, in A.C. impedance techniques, the use of small amplitude

D.-H. Kim, G. Simkovich /Molten sodium sulfate

166

sinusoidal potentials does not disturb the electrode properties and at higher frequencies polarization at the electrodes may be eliminated. Thus, an A.C. impedance technique was utilized to measure the resistance of a pure Na2S04 melt since the melts showed some polarization effects at the electrodes in our preliminary D.C. experiments. Impedance can be thought of as the resistance of a circuit to an alternating waveform as opposed to a pure resistance; it has not only magnitude but also direction - phase angle. One of the advantages of A.C. impedance techniques over conventional D.C. electrical conductivity efforts is the ability to separate the real and imaginary components of impedance. An impedance, 2, can be completely defined by specifying the magnitude, IZI, and the angle, 0, or alternatively by specifying the magnitude of its real , z’ , and imaginary, Z”, components [S]. There are a number of graphical interpretations available for impedance data analysis over a wide frequency range [9]. A plot of Z’ versus wz” was employed to evaluate the resistance of the molten salts. A plot of Z’ versus wz” shows a straight tine with a slope of -RpC and an intercept of RQ + Rp according to the following equation: z’ = RQ + R, - Rp C w Z” where

Z z”

is the real part of the impedance is the imaginary part of the impedance is the resistance of the electrolyte is the polarization resistance is the capacitance is angular frequency (= 27cf)

Rfi RP C W

The real and imaginary part of the impedance expressed by the following relationship [lo].

Z'

=

R,

RP

+

1 +w2C2R;

Z”

(1)

=

Ielect = I,

+ I, 1 - exp(- E)

exp (z)

- 1

(4)

where Ie.Ie: electron and electron hole currents, respectively oe”,o~4 electron and electron hole conductivity, respectively E : actual applied voltage F : Faraday constant R : gas constant T : temperature (K) L/A : cell constant. In the derivation of Eq. (4) it is assumed [l&19] that (i) excess electrons and holes follow the laws of ideal dilute solutions, (ii) their mobilities are independent of concentrations. (iii) the change in the concentration of atomic defects arising from thermal disorder with variation in the metal to nonmetal ratio is small, (iv) convection in the melt is negligible. The division rearrangement

of Eq.(4) gives

by [l - exp

(- EF/RT)]

and

can be

and a plot of the left hand side of I$. (5) versus exp (2)

w C R; 1 + w2 C2 R;:

cell is given by

(3)

Thus, as frequency increases, the straight line of a plot of Z versus wz” levels off,and the projection of this point onto the Z’ axis affords the sum of the resistances of the electrolyte and the circuit leads according to Eq. (2). (ii) Electronic Conductivity Measurement The idea that an appropriate choice of electrodes enables the suppression of either ionic or electronic transport in a ealvanic cell Drovides the basis for the oolarization technique. This technique has been extensively’employed to investigate electronic conductivity in ionic solids [ II- 161 and has also been applied to a few molten systems [17-l 91. Wagner [6] ‘h’as derived the appropriate relation for the polarization conditions from transport theory. This relation states that, under steady state conditions, the total current due to passage of electronic species through the polarization

(EF/RT) gives (~e’as the intercept and (J~Oas the slope. These values, combined with total electrical conductiviry results, permit the evaluation of the transport numbers of each electronic carrier in the molten salts. In the present work, D.C. current flowing through the polarization cell is measured at various applied voltages which are kept below the decomposition potentials of the sample to ensure that the measured current is only the electronic current. (iii) Potentiostatic Polarization Experiment The potentiostatic polarization cell technique was employed to determine ionic transport numbers for molten sodium sulfate at 1173 K. In a recent application of this technique, reasonable agreement was obtained with transport numbers from Tubandt and tracer diffusion techniques [20]. This technique is based on a two electrode cell. A constant D.C. potential is applied by a potentiostat and the current is monitored as a function of time. The electrode is chosen to be reversible with respect to the cation present in the electrolyte. Under the influence of a D.C. field cations migrate to the negative electrode and anions to the positive electrode. As the concentration of anions increases near the positive

167

D.-H. Kim, G. Simkovich /Molten sodium sulfate

electrode, local charge neutrality requires that a cation accompany each anion. This process establishes a salt concentration gradient across the electrolyte. As the cell polarizes the amount of current carried by the anion decreases but the amount carried by the cation remains constant. When the back potential created by the concentration gradient exactly opposes the applied potential, the anions no longer carry current, and the cell is completely polarized with respect to the anion. Thus, the ratio of final current due to the cation only to the initial current due to the cation and anion yields the cation transport numbers of molten Na2S04. (iv) Evaluation of the Self Diffusion Coefficient of Na Ions There is a definite relationship between the ionic component of the electrical conductivity and the self diffusion coefficient of corresponding component of the melt. This relation is known as the Nernst-Einstein equation [21] :

CT.

c. q!

1=1 D1

kT

(6)

where Oi is the conductivity of the i species, Di is the self diffusion coefficient of the ionic component corresponding to ai, ci is the concentration, qi is the charge, k is the Boltzmann constant and T is the absolute temperature. If (J is the total conductivity of the molten salt and ti is the transport number of the species under consideration, the relation holds for the following equation: Oi = ti

X (3

(7)

In other words, the total electrical conductivity measured by an AC. impedance technique and the transport number of Na ions determined by the potentiostatic polarization experiment provide the cationic conductivity of the melt. Thus, the self diffusion coefficient of Na ions in the melt can be estimated by the above equations under the assumption of complete ionization of a Na2S04 melt at the given experimental atmosphere. In addition, the self diffusion coefficient is related to the mobility of species i, ui. by the Einstein equation : Di = ui;T

(8)

where R is the gas constant and F is the faraday constant.

A: Quartz Tube B: Pt Wires ”

~~

E: Mullite Tube F: Au Wires G: Au Crucible

Figure 1. Schematic A.C. impedance cell arrangement

polarization free total electrical conductivity. As shown in Fig. 1, the three electrode system was utilized for the A. C. impedance measurements. The reference electrode was a silver wire immersed into a 10 m/o Ag2SO4/Na2SO4 melt contained in the Na ion conducting membrane, a mullite tube (0.7 cm O.D.) [22]. Pure gold wires serve as the working electrode and counter electrode. Platinum wires welded to these electrodes were employed as leads to connect to the EG & G Model 273 potentiostat coupled with Model 5208 Lockln amplifier. The schematic arrangements for the A.C. impedance measurements are depicted in Fig. 2. The total electrical conductivity of a pure Na2S04 melt and the melts containing various oxides was determined by measuring the resistance of the melts from the relation: R = ($)x(CellConstant)

3. EXPERIMENTAL

PROGRAM

Commercial SO2 balanced by oxygen gas from Liquid Carbonics Co. were utilized to fix the acidity of the melt at 1173 K. An anhydrous ultrapure sodium sulfate from Alfa Products was used in the present study. (i) A.C. Impedance Measurements AC. impedance measurements were performed to obtain

(9)

where R is the resistance measured in ohms, CJis the conductivity of the melt expressed in (ohm-cm)-l and the cell constant is given in terms of cm-l. The cell constant, which is a characteristic of the conductivity cell, depends on the length between the electrodes and the surface area of the electrode exposed to the melt. The cell constant is usually predetermined by measuring the resistance across an ionic solution of known specific conductivity. Since the specific conductivity values

D.-H. Kim, G. Simkovich /Molten sodium sulfate

168

1CRT Display

M 273 Potentiostat

M5208 Lock-In Amplifier

Figure 2. Schematic diagram of A.C. impedance experiment

of KC1 solution are well established [23], an 0.1 N KC1 solution was employed to determine the cell constant for the present work. The cell constant measurements were conducted at about 250C utilizing identical cell arrangements with the gold crucible to contain an 0.1 N potassium chloride solution instead of the melt. (ii) Wagner - Hebb Type Polarization Experiments The Wagner-Hebb type polarization technique was used to determine the partial electronic conductivities of a pure Na2S04 melt as a function of Na20 activity at 1173 K. A constant voltage was supplied to the polarization cell via the EC & G Model 273 potentiostat. A Solid State Electrometer, Keithley Model 61OC, was utilized to check the actual voltage on the polarization cell. The reference electrode was the same as that used in AC. impedance experiments. Platinum wires were employed to lead two gold electrodes into the potentiostat. Pure gold crucibles were used for these experiments. An A.C. impedance technique was utilized to determine the cell constant with the same cell geometry by measuring the conductivity of an 0.1 N KC1 solution instead of the molten salts.

cations under an electrochemical driving force with essentially no electronic conduction 124,251. 4. RESULTS AND DISCUSSION An A.C. impedance technique was employed to measure the resistance of the pure molten Na2S04 and subsequently to obtain the total electrical conductivity of the melt at I I73 K. Analysis of A.C. impedance data is commonly carried out over a wide frequency range and based on Eq. (1). A plot employing Fq. (1) is shown in Fig. 3 for a pure Na2S04 melt at an Na20 activity of lo-l2 at 1173 K. It is noted that the plot shows a straight line with a slope of -RpC

70

50

(iii) Potentiostatic Polarization Experiment Most molten salts are ionic conductors. Thus, it seems logical to assume that molten sodium sulfate is an ionic conductor. The relative contributions of the different carrier species ( cation vs. anion ) were investigated by this technique. A constant D.C. potential was applied via the potentiostat and the current was monitored automatically as a function of time. A symmetric cell configuration was employed, and the electrode was chosen to match the cation in the molten sodium sulfate. The reversible electrode consists of a mullite tube conductive to sodium ions containing a silver electrode immersed into a melt of Ag2S04- 90 m/o Na2S04. Mullite is a two-phase ceramic consisting of mullite grains (3Al203 - 2Si@) enveloped by silica. At high temperatures, dissolved alkali metal compounds in the silica film allow transport of alkali metal

N

30

10

-10

4e+4

6e+4

8e+4

WZ” Figure 3. 2’ versus wz” in a pure Na2S04 melt when log aNa ec@S t0 -12 at 1173 K

le+5

169

D.-H. Kim, G. Simkovich /Molten sodium sulfate

.”

I

6

;

lb

1’1 1; Ii

1’4 1’5

-Log (Sodium Oxide Activity) Figure 4. Lo&total conductivity) versus Na20 activity in a pure Na2S04 melt at 1173K

at low frequencies and Z’ , the real part of the impedance. approaches Rfi at high frequencies. The line starts to level off at approximately 8.0 x lo4 on the wz” axis which corresponds to a frequency of 5 kHz. The projection of this point onto the Z’ axis provides the resistance of the electrolyte, hence, the total electric.al conductivity of the molten salt by the relationship of Eq. (9). The mea&d total elect&al con&&ties of a pure NaTSOd at 1173 K are deuicted in Fie. 4 as a function of the activity of Na20 in the melt. The total electrical conductivities remain rather constant regardless of the changes in Na20 activities. The total electrical conductivity of a pure Na2S04 melt averaged about 0.232 (ohm-cm)-l which is about one order of magnitude less than the literature values [26-291. This discrepancy is most probably caused by the facts that the previous investigators had: (1) a relatively impure Na2S04, (2) a reaction between their quartz capillary and molten sodium sulfate. and (3) a reaction with their Pt electrodes. It was observed that there was significant deterioration of the quartz crucibles used to contain the Na2S04 melts in our preiiminary work, and that there was a reaction of Na2S04 melts with the Pt electrodes initially utilized. The Wagner-Hebb type polarization technique was used to determine the partial electronic conductivity of a pure Na2S04 melt as a function of Na20 activity at 1173 K. The theoretical basis for this experimental technique has been given by Wagner and Hebb. The appropriate equation from I

exp (EF/RT) Figure 5. Y versus exp (EF/RT) in a pure Na2S04 melt atlogaNa2Oequtisto-llat1173K

transport theory when the ionic currents were suppressed was given by Eq. (4) and Eq. (5). An example of the graphical interpretation of Eq. (5) for a pure Na2S04 melt at an Na20 activity of 10-l’ is shown in Fig. 5. MINITAB program at the Pennsylvania State University was utilized for regression analysis to obtain a slope and an intercept of the plot. This result shows an electron conductivity of 1.617 x 10m3 (ohm-cm)-] and an electron hole conductivity of 1.407 x 10e5 (ohm-cm)-’ in a pure Na2S04 melt when the Na20 activity equals 10-l *. Thus, the partial conductivities of electrons and electron holes were obtained as a function of Na20 activity in the melt and are depicted in Fig. 6. It can be seen that electron conduction in pure Na2S04 is considerably larger than that of electron holes over the entire Na20 activity range. Furthermore, it is noted that both electron and electron hole conductivities remain relatively constant regardless of the changes in Na20 activities. Thus. the total electrical conductivity as well as partial electronic conductivities of a pure Na2S04 melt are not dependent on the acidity and/or basicity of the melt. From the measured values of total electrical conductivities and electronic conductivities, the transport numbers of electronic species may be computed. These numbers are plotted in Fig. 7 for a pure Na2S04 melt. The transport numbers of electrons are of the order of 10-3 while those of

D.-H. Kim, G. Simkovich /Molten sodium sulJate

170

-2 -

-3 -

-4 -

.

. -5 -

.

-6 -6

9

.

l

.

10

-Log

11

(sodium

12

13

oxide

14

15

electron holes are of the order of 10e5. This indicates that the electronic conduction in a pure Na2S04 melt arises primarily via electron transport over the whole Na20 activity range. The transport numbers of electronic species in molten salts have not been measured extensively but the few that have been measured are similar to those determined in this study, e.g., tg = 3x10-3 in the molten eutectic of LiCl-KC1 at 4socc [17]. It is generally observed that Na+ ion conductivity prevails in solid Na2S04 ( 235-883 C) and that the partial electronic conductivity is negligible. Thus the electrode for the potentiostatic polarization cell was chosen to be reversible with respect to the cation ( Naf) present in the molten sodium sulfate. If the cation reversible electrode works properly, then complete cell polarization is expected when the anion diffusion, created by a salt gradient, exactly opposes the migration of this ion under the influence of the applied potential. The cationic transport numbers are displayed as a function of Na20 activity in the melt in Fig. 8. It is clear from this figure that the molten Na2S04 is an ionic conductor (tNa+= 0.985). However, this result shows that the cationic transport numbers are decreased in the low sodium oxide activity region of the melt. In another words, the contributions of the possible anion species such as SO4--, S2O7-- or 0.- are not negligible in such a region of the

activity)

Figure 7. Transport numbers of electronic species in a pure Na2S04 melt as a function of Na20 activity at 1173 K

1.1

70

q

50

T? 0

q

q

0

m q

x 8

30

q

f(-)

l

N+)

m Y

ca

10

-10 -I8

I

9

.

.

. .

I

.

10

-Log(sodium

I

11

. 7

I

12

oxide

.

. .

I

.

13

I

14

activity)

Figure 6. Electronic conductivities in a pure Na2S04 melt as a function of Na20 activity at 1173K

0.8 8

9

10

11

12

13

14

15

-Log (sodium oxide activity) Figure 8. Cationic transport numbers of pure pure Na2S04 melt as a function of Na20 activity at 1173K

171

D.-H. Kim, G. Simkovich /Molten sodium sulfate measurements the transport numbers of electrons. ta, and electron holes, te, were calculated as follows: b = 7.1 x 10-3 ;

q

q

q

q

Such experimental investigations show that pure Na2S04 melt had a somewhat low total electrical conductivity as compared to most ionic melts and that the electronic conduction occurs primarily via the transport of electrons. The cationic @ansport numbers of pure Na2S04 melt were obtained by utilizing the potentiostatic polarization technique. The results indicate that pure Na2S04 melt is a cationic conductor over a wide range of Na20 activities at 1173 K. From the above results, the diffusivity of the cation, one of the important kinetic parameters, can be estimated as 8.1 x 10-6 cm%ec.

0 q

1 a

9

-Log

te = 5.8 x 10-5

6. ACKNOWLEDGEMENT

IO

11

(sodium

12

oxide

13

14

15

The authors are pleased to acknowledge the office of Naval Research for the financial support of their project on contract No. N0014-86-K-0133. Discussions with our monitor Dr. A. J. Sedricks were most helpful.

activity)

Figure 9. Diffusion coefficient of Na ions as a function of Na20 activity in a pure Na2S04 melt af 1173 K

It is possible to determine the kinetic parameters of the molten salt with the known transport numbers of Na+ ions and the cationic conductivity of the melt. The self-diffusion coefficient of Naf ions in the melt was calculated from the experimental data’as described in the Eq. (6) and Eq. (7) and are depicted in Fig. 9. It is noted that the average selfdiffusion coefficient of Na+ ions is 8.1 x 10-e cm2/sec which is rather lower than that of sodium ions in other molten salts The literature shows that the measured tracer diffusivities of Na were 1.05 x low4 cm*/sec in NaF melt at 1295 K [30], 9.81 x lo-5 cm*/sec in NaCl melt at 1173 K [31], 2.92 x IO5 cm2/sec in NaN@ melt at 671 K [32], and 5.63 x 10e5 cm2/sec in Na2C03 melt at 1183 K [33]. The possible reason for the differences in diffusion coefficients between this work and the literature value is probably the relatively larger anions in Na2S04 than in other melts. From equation (8). the mobilities of Na+ ions are calculated and its average value is 8.0 x 10-S cm2/volt sec. 5. SUMMARY AND CONCLUSION The main thrust of this experimental program was to obtain some of the transport properties in the aggressive molten salt Na2S04. The total electrical conductivity measurements by an AC. impedance technique and Wagner-Hebb type polarization experiments provided the total electrical conductivity, electron conductivity, and electron hole conductivity of a pure Na2S04 melt at 1173 K. From these

7. REFERENCES [l] J. M. Quets and W. H. Dresher, Nut. Mat., 14, (1969) 583. 121 N. S. Bomstein, M. R DeCrescente, Met. Trans., 2, (1971) 2875. [3] J. A. Goebel and F. S. Pettit, Met. Trans., 1. (1970) 1943. [4] J. A. Goebel, F. S. Peuit, and G. W. Coward, Met. Trans., 4, (1973) 261. 151 R. A. Rapp and K. S. Goto, Proceedings of Second International Symposium on Molten Sal&, ed. J.Braunstein and J.R. Selman, (Electrochem. Sot. 1981) p. 159. [6] C. Wagner, Proc. of Seventh Meeting of the Inter. Comm. on Thermo. and Kinetics of Electrochem.. (Butterworth Scientific Publications, London, 1957) p. 361 [7] M. H. Hebb, J. Chem. Phys., 20, (1952) 185. [8] J. E. B. Randles, Dis. Faraday. Sot., 1, (1947) 11. [9] G. W. Walter, Corr. Sci., 26: (1986) 681. 1101 F. Mansfeld. Corr.-NACE. 36. (1981) 301 illj J. B. Wagner, Jr. and C. Wagne;, J. Chem. Phys., 26, (1957) 1597. [12] A. Lingras and G. Simkovich, J. Phys. Chem. Solids, 39, (1978) 1225. [13] A. V. Joshi and J. B. Wagner, Jr., J. Phys. Chem. Solids, 33, (1972) 205. [14] A. V. Joshi and i. B. Wagner, Jr., J. Electrochem. sot.. 122, (1975) 1071. [15] D. Raleigh; J. Phys. Chem. Solids, 26, (1965) 329. 1161 _ _ J. Schoonman. A. Wolfert. and D. F. Untereker. Solid State Ionics, li, (1983) IS?‘. [17] R. J. Heus and J. J. Egan, J. Phys. Chem., 77. (1973) 1989.

172

D.-H. Kim, G. Simkovich /Molten sodium sulfate

[18 ] R. J. Heus and J. J. Egan, Proceedings of International Symposium on Molten Saks, (Electrochem. Sot. ,Pennington, NJ, 1976) p. 523. [19] G. J. Reynolds, M.C.Y. Lee and R. A. Huggins, Proceedings of the Fourth International Symposium on Molten Salts, (Electrochem. Sot., Pennington, NJ, 1984) p. 519 [20] D.F. Shriver, S. Clancy, P.M. Blonsky. and L.C. Hardy, Sixth Ris$ International Symposium on Met. and Mat. Sci., 1985, ~353. .1211 W. Jost. Diffusion in Solids. Liauids. Gases. Academic Press, New York, i968, pp ‘142-143. [22] D. A. Shores and R. C. John, J. Appl. Electrochem. 10, (1980) 275. [23] D. Dobos, Electrochemical Data, Elsevier Scientific Publishing Co., New York, 1975, ~57. [24] R. J. Labrie and V. A. Lamb, J. Elecuochem. Sot 106, (1959) 895. 1

[25] H. Schmalzreid, Z. Phys. Chem. (Frankfurt), 38, (1963) 87. [26] A. Kvist, Z. Naturforschg.. 22a. (1967) 208. [27] A. Kvist, Z. Naturforschg., 22a, (1967) 467. [28] A. Josefson and A. Kvist, Z. Naturforsch, 24a, (1969) 466. [29] K. Matiasovsky, V. Danek, and B. LifflebJen, Electrochim. Acta, 17, (1972) 463. [30] K. Grjotheim, M. Malinovsky, M. Matiasovsky, K. Zuca, and H. A. Oye, J. Chim. Phys., 67, (1969) 145. [31] J. 0’ M. Bockris, S. R. Richards, and L. Nanis, J. Phys. Chem. 69, (1965) 1627. [32] S. Zuca and M. Contantinescu, Rev. Roum. Chim., 17, (1972) 385. [33] P. L. Spedding and R. Mills, J. Electrochem. Sot., 112, (1965) 594.