1 The Electrochemical Behaviour of Zinc in Alkaline Solution

BY R. D. ARMSTRONG AND M. F. BELL

1 Introduction

This Report discusses the electrochemical behaviour of zinc in alkaline solution with particular emphasis on the electrochemical kinetics of this system. A typical anodic current-voltage curve for Zn in alkaline solution is shown schematically in Figure 1. The curve can be considered as exhibiting the active

Potential E

Figure 1 Schematic representation of the current-potential curve for the dissolution of zinc in alkaline solutions

region, AB, and the passive regions, BC and DE. These regions will be dis- cussed in separate sections. Related topics, i.e. soluble zinc species and the important forms of zinc oxide, will also be covered.

1

2

2 Soluble Zinc Species in Alkaline Solution

Electrochemistry

Substantial experimental evidence points to the fact that the predominant soluble zinc species in alkaline solutions is the tetrahedral ion Zn(0H): -.

Potentiometric experiments, carried out by Dirkse, gave results which he evaluated as follows. The initial assumption was made that the dissolution of zinc could be described by the equations

Zn+Zn2 + + 2e- (1)

xZn2+ +yOH- +zH,O-+Zn, (2) where Zn, is the complex Zn" species in solution. The potential for this reaction would be given by

which, because the activity of solid zinc can be assumed to be unity, reduces to

E,, = EzOnlzn2+ -0.0295 log azn2+ (4)

The equilibrium constant for reaction (2) is given by

Ezn then becomes

No attempt was made to evaluate z. From plots of Ez, against log Mzn* and considering several possibilities for the solution-soluble species [i.e. Zn(OHI2, Zn(OH)s, Zn(OH): -1 Dirkse obtained a value of 1 for x. He also obtained a value of 4 for y and gives evidence that water is a reactant rather than a product. The thermodynamic data, shown in Table 1, are derived from these results.

From i.r. and Raman studies, Fordyce and Baum* found absorptions for zincate solutions at the wavenumbers shown in Table 2. They attribute these to the Zn(0H)f - species and conclude that the experimental results can be explained by a single tetrahedral species in which there is some repulsion between the ligands and in which bond stretching is easier than is usually the case. They also suggest possible hydrogen-bonding in the structure.

* MZ, is the molar concentration of Zn2+. T. P. Dirkse, J. Electrochem. SOC., 1954, 101, 328. J. S. Fordyce and R. L. Baum, J. Chern. Phys., 1965,43, 843.

The Electrochemical Behaviour of Zinc in Alkaline Solution 3

In n.m.r. studies of the proton resonance in alkaline zinc solutions, Newman and Blongen3 found an average value of the chemical shift (dependent on the KOH concentration) which is consistent with a Zn(0H): - species. They also state that the evidence is in favour of the partial (3545%) covalency of the Zn-4 bond.

Table 1 Thermodynumic data for soluble zinc species

Quantity Reaction or Species Value E0 Zn + 40H- +Zn(OH)t- + 2e- 1.211 v AGO298 Zn+ 40H- +Zn(OH)i- + 2e- -- 233.68 kJ mol- AGO298 Zn( OH); - - 862.74 kJ mol- AGO298 Zn2 + 40H- +Zn(OH)Z- - 86.61 kJ mol- ' K Zn2 + 40H-+Zn(OH):- 1 . 4 ~ 1 0 1 5

(Formation constant)

Table 2 Infrared and Raman absorption data (cm- ') for zincate solutions

Raman I.r. (reflection) 1.r. (transmission) 484 430 430

3 The Active Dissolution of Zinc

Amstrong and Bulman4 studied the active dissolution of zinc in solutions having compositions in the range 3 x 10-2-2M-NaOH [of constant ionic strength (3 mol I -I) with NaCl J using the rotating-disc technique. They found that the current was dependent on rotation speed at constant potential and assumed this effect to be due to the simultaneous occurrence of deposition and dissolution. By extrapolating out diffusion from i-' vs. m-* plots, they obtained dissolution currents which showed a Tafel slope of 4 2 k 5 mV decade-'. This would give a corresponding cathodic Tafel slope for zinc deposition of 105 mV decade-', as the anodic and cathodic Tafel slopes are linked by the equation

1/30 = l/b,+ llb,

This work also showed that the reaction order with respect to zincate for zinc deposition is unity. These results are in substantial agreement with the work of Bockris et aL5 who found, by potentiostatic and galvanostatic techniques, an anodic Tafel slope of 49 mV decade-' and a cathodic Tafel

G . H. Newman and G. E. Blomgren, J. Chem. Phys., 1965,43,2744. R. D. Armstrong and G . M. Bulman, J. Electroanalyt. Chem. Interfacial Electrochem., 1970'25, 121. J. O'M. Bockris, Z . Nagy, and A. Damjanovic, J . Electrochem. SOC., 1972, 119, 285.

4 Electrochemistry slope of 1 13 mV decade -I. Bockris lists the reaction orders, calculated from the dependence of exchange current density on concentration. These are shown in Table 3 where they are compared with those found by Armstrong and Bulman. They agree with the orders measured from the concentration dependence of current density at constant potential. As Armstrong and Bulman studied only two solutions, not too much weight should be placed on their high reaction order with respect to hydroxide ion.

The results reported by Kabanov6 are in fair agreement with the work discussed above. He gives a value of 3 for the pH dependence of the anodic current density, @log iJapH). However, his reported anodic Tafel slope of 30 mV decade is rather low. This would imply that the cathodic Tafel slope was infinite.

Table 3 Reaction orders for deposition of zinc

Quantity Armstrong and Bockris' Bulman4

- 0.14

[s] E

ca. 3.5 2.6

Anodic Tafel slope/mV decade- Cathodic Tafel slope/mV decade-

,Z Calculated using equation (7).

42f 5 105"

0.67

49f 13 113+30

The Tafel slopes of approximately 40 mV decade are consistent with an overall mechanism

Zno & Zn' (8)

Zn' & Zn" (rate-determining step) (9)

where Zn' could be either an adsorbed or a solution-soluble intermediate. The reaction order with respect to hydroxide ion would suggest that two hydroxide ions are involved in the rate-determining step. However, it is difficult to be certain about the state of complexation of the zinc intermediates. Bockris favours a solution-soluble intermediate of the type Zn(0H)z.

Hampson and co-workers have studied the zinc system by a number of techniques. Using the faradaic impedance method' they fitted their results to a

B. N. Kabanov, Izvest. Akad. Nauk S.S.S.R., Ser. khim., 1962, 980. ' J. P. G . Farr and N. A. Hampson, Trans. Faraday SOC., 1966,62, 3493.

The Electrochemical Behaviour of Zinc in A lkaline Solution 5

Randles plot, with the intercept as 03 00 independent of both hydroxide and zincate ion concentrations. It was stated that this behaviour could be attribu- ted to slow adatom diffusion on the electrode surface. This cannot be true since Rangarajan has shown that, when adatom diffusion is considered, a ‘RandIes’ plot can be found if (i) Da/X; & kz and io c Da/X;, in which case the intercept is RT/nFio, or (ii) (o!/kz) .+ (kz/Da)+xo, when the intercept is (RTlnFio) [l +(k2/Da)+xo1

Thus in either case the exchange current density is an important component of the intercept and it is generally accepted that the exchange current depends on cOH- and ~ ~ ~ ( ~ ~ ) z -, as was found in the previously mentioned investigations.

Hampson reported from galvanostatic i-q t r a n s i e n t ~ ~ . ~ that io did not depend on zincate concentration, which is at variance with the results of other workers, as are his reported Tafel slopes and the value of the double-layer capacitance. l o Bockris’ has suggested that some of these discrepancies are due to the lack of correction for ohmic resistance in the solution.

The Dissolution of Zinc from a Zinc Amalgam in Alkaline Solution.-Gerischer reported that, whilst the overall reaction for the dissolution of zinc amalgam in alkaline solution is

Zn + 40H--+Zn(OH)42- + 2e- (10)

it can be broken down into two mechanistic steps, the rate-determining step being

Zn(Hg)+20H-+Zn(OH),+2e- (11)

followed by

The arguments results :

Zn(OH), +20H-+Zn(OH,)~- (12)

for this reaction depend on the folIowing experimental

61og i, 6Iog[Zn,2,+]

= 0.5 & 0.02

61og i , Slog[OH-] =

J. P. G . Farr and N. A. Hampson, J. Elertroanafyt. Chem. Interfacial Elcctroclienr., 1967, 13, 433; J . P. G . Farr, N. A. Hampson, and M. E. Williamson, ;bid., p. 462. N. A. Hampson, G. A. Herdman, and R. Taylor, J . Electroanolyt. Clicni. Iiiterjbcinl Electrochem., 1970, 25, 9.

l o D. S. Brown, J . P. G . Farr, N. A. Hampson, D. Larkin, and C. Lewis, J . Nectroonalyt. Chem. Inter-acial Electrochem., 1968, 17, 421. H. Gerischer, Z . phys. Chem. (Leiprig), 1953, 202, 302; Anafyt. Chetn., 1959, 31, 33.

6 Electrochemistry

According to the above mechanism, assuming that the difference between concentration and activity and also double-layer effects may be ignored, io can be expressed by

where Ki is the stepwise equilibrium constant, given by Kt = [Zn(OH):-']/ [Zn(OH)y -'][OH-]. On the basis of these results, a number of authors have

zFko , and a, of equation (K3K4) - a

reported values for the parameters ko, KO =

(16). These are summarized and compared in Table 4.

Table 4 Authors U Ko/cm s-I ko/mol cm-2 s-' a Ref. Gerischer 0.5 1 x 10-3 5 x 104 11 (Hampson et c ~ l . ) ~ Behr et al. 0.43 0.17-0.25 x 105 13 Matsuda and Ayabe 0.42 0.5 x 10-3 3 x 105 14

15 Kambara and Ishii - 0.7 x 10-4 -

ko was calculated using a value of 10'5*28 for K3K4 as suggested by Harnpson and Farr, and with which Matsuda and Ayabe agree. It should be noted that the slopes of the plots in the first figure of the paper by Hampson and F a n t 2 show the wrong dependence on the Zn(Hg) concentration, although the final results appear to be in agreement with the work of Gerischer.

4 Properties and Crystal Structures of Zinc Oxides, Hydroxides, and Peroxides

The zinc oxide-hydroxide system is complicated by the number of forms these compounds can take. The method of preparation of zinc oxide affects the catalytic activity of the product. l 6 For instance, the oxide prepared from the carbonate is more active than that from the nitrate. These both possess the same crystal structure but the active form is more disordered. (The amount of catalytic activity is associated with the degree of disorder.)

Inactive zinc oxide crystals usually exist as either short or long needles of hexagonal structure. In these crystals, the zinc and oxygen atoms form a hexagonal structure of the wrtzite type which from atomic radii must be rather open. Under certain conditions, zinc atoms may be lodged in interstitial sites giving rise to non-stoicheiometry. It is believed that this is the reason for zinc oxide's semiconductor properties (of the n type).

l 2 J. P. G. Farr and N. A. Hampson, J. Electroanalyt. Chem. Interfacial Electrochrm., 1968, 18, 407. B. Behr, J. Dojlido, and J. Malysko, Roczniki Chem., 1962, 36, 725 (Chem. Abs., 1963, 58, 2143d).

T. Kambara and T. Ishii, Rev. Polarog. Kyoto, 1961, 9,30(Chem. Abs., 1963,59, 9596e). l4 H. Matsuda and Y. Ayabe, 2. Elektrochem., 1959, 63, 1164.

l6 R. Faivre, Ann. Chim. (France), 1944, 19, 58.

T%e Electrochemical Behaviour of Zinc in Alkaline Solution 7

Each atom is surrounded by four oxygen atoms situated at the apices of a tetrahedron. Analysis by electron diffraction shows that the zinc atom is displaced from the centre of the tetrahedron by 11 pm in a direction parallel to the c-axis. Moreover, zinc oxide is partially covalent and the crystals are not purely ionic.

The ratio of the crystal axes, measured by X-ray diffraction, l 7 is a : c = 1.60200; a = 324.265 pm, c = 519.48 pm. The space group is 6 mm and there are two molecules per unit cell.

The zinc hydroxide system is much more complex. The solid exists in five crystalline forms, designated a, 8, y, 6, and e, and also in an amorphous form. At ordinary temperatures the only stable form is the &-hydroxide and all the other forms are converted into this form, although the transformations are not simple. Specific conditions are necessary for the preparation of any of the metastable forms.

A gelatinous form (amorphous zinc hydroxide) is obtained by adding a weakly alkaline solution to a solution of zinc nitrate. If the precipitation is not complete, this form is converted into the a-form.

The crystals of the a-hydroxide are usually only small and never well formed. Certain anions, such as carbonate, stabilize the a-hydroxide under certain conditions and the stabilized form possesses a hexagonal structure with alter- nate ordered and disordered layers. Anion incorporation into the lattice is the cause of the disordered structure. The length of the edge, a, is well defined and equal to 311 pm. The length of the other edge, c, depends on the anion in- corporated into the lattice. l 8

The 8-hydroxide can exist in two forms (Dl and / I2”). These both possess layer lattices in which the distance between the planes is 567 pm. These forms are converted slowly into the &-form. The y-hydroxide is found as protracted prismatic crystals but has been reported to crystallize in other forms, especi- ally needles. The unstable d-form, which is obtained by slow crystallization from supersaturated zincate solution, exists as rhombohedra1 flakes. It is converted quite rapidly into the &-form.

Large monocrystals of €-hydroxide are obtained if the solution from which it is obtained is only slightly supersaturated. This form belongs to the ortho- rhombic system and the lattice dimensions are a=516 pm, b=853 pm, and c =492 pm.20 Each zinc atom is tetrahedrally surrounded by hydroxide ions and the unit cell contains four molecules. The tetrahedra form zig-zag chains along the c-axis. Each hydroxide ion belongs to two tetrahedra and the lattice extends in three dimensions. The Zn-OH length is 195 pm and the OH-OH distance is 283 pm. The space group of this form is P21212L and z=4.*

* z is the number of molecules per unit cell. ’’ C. W. Bunn, Proc. Phys. SOC., 1935, 47, 835.

l 9 W. Feitknecht, Helv. Chim. Acta, 1949, 32, 2294. 2o R. B. Corey and R. W. G. Wyckoff, Z . Krist., 1933, 86, 8.

W. Feitknecht, Helv. Chim. Acta, 1938, 21, 766.

8 EZectrochemistry

The peroxide is formed from the hydroxide at low temperatures but loses oxygen rapidly at higher temperatures. It has been shown recently that this is a true peroxide, and a product corresponding to 4Zn02,Zn0,H20 has been isolated.

The stabilities of the hydroxide forms are in the order amorphous < a < /3 < y < S < & .

Table 5 shows the free energy and enthalpy changes for the conversion of the different forms into the &-form. Solubility products (cZn2+, a&-) and the enthalpy and free-energy changes for the conversion of the hydroxides into the oxide are shown in Table 6.

Table 5 Free energy and enthalpy of conversion of various forms of zinc hydroxide into the &-form

Transformation AG/kJ mol- AHlkJ mol- amorphous+& - 6.653 - 12.343 B1-e - 1.464 - 1.213 Y+& - 1.213 - 0.837

5 The Passive Regions of Zinc in Alkaline Solution

Despite intensive investigations, these regions remain the most complex and unresolved subjects in the electrochemistry of zinc in alkaline solution. Many authors have studied this system 21-29 using a wide range of concen- trations of both zincate and alkali and also numerous techniques. Because of this, correlation of their results is very difficult. Experimentally it is generally found that there are two main regions, BC and DE, in the i-E curve (Figure 1). In this Report, the active-passive transition will be taken to occur at the point where the first deviation from the active region occurs. This corres- ponds to the point B in Figure 1.

Most of the literature of the mechanism of passivation suggests an adsorp- tion model or a model involving the nucleation and growth of a two-dimen- sional layer, though a dissolution-precipitation mechanism has been suggested. 0--5 Thus, Kabanov et al.30-33 found that a quantity of charge

2’ R. Landsberg and H. Bartelt, 2. Elektrochem., 1957, 61, 1162. ” S. El Wakkad, A. El Din, and H. Kotb, J. Electrochem. Soc., 1958, 105, 47. 23 1. Sanghi and M. Fleischmann, Electrochim. Acta, 1959, 1, 161. 24 H. Fry and M. Whitaker, J . Electrochem. SOC., 1959, 106, 606. ’’ R. F. Ashton and M. T. Hepworth, Corrosion, 1968, 24, 50. 26 J. Euler, Electrochim. Acta, 1966, 11, 701. 27 J. P. Elder, J . Electrochem. SOC., 1969, 116, 757. 2 8 K. Huber, Helv. Chim. Acta, 1943, 26, 1037; J . Electrochem. SOC., 1953, 100, 376. 2 9 G. S. Vozdivenskii and E. D. Kochmann, Russ. J . Phys. Chem., 1965, 39, 347. 30 B. N. Kabanov, R. Burnstein, and A. N. Frumkin, Discuss. Furaduy SOC., 1947, N O . I ,

31 B. N. Kabanov and D. I . Leikis, Actu Physicochimica U.R.S.S., 1946, 21, 769. 32 T. I. Popova, V. S. Bagotskii, and B. N. Kabanov, Dokludy Akud. Narik S.S.S.R.,

33 B. N. Kabanov, Electrochim. Acta, 1962, 6 , 253.

p. 259.

1960, 132, 639.

Tabl

e 6

Free

ene

rgy

and

enth

alpy

for

conv

ersio

n of

zin

c hy

drox

ides

into

zin

c ox

ide

Com

poun

d Ty

pe

log(

solu

bi1i

typr

oduc

t)

AG

and

AH

in kJ

mol

-I f

or c

onve

rsio

n in

to o

xide

st

able

ZnO

ac

tive

ZnO

AG

A

H

AG

AH

Z

n(O

H),

amor

phou

s - 15

.95

- 2.

468

0 -

6.56

9 - 4.0

58

81 - 1

6.65

2.

678

11.0

87

1.42

2 7.

029

Y - 1

6.70

2.

970

11.4

64

1.13

0 7.

406

& - 1

6.92

4.

100

12.3

01

0 8.

242

ZnO

ac

tive

- 16.

89

-

-

-

-

stab

le

-16.

1 to

-16

.43

-

-

-

-

W

10 Electrochemistry

equivalent to 1 mC cm-* was sufficient to confer passivity on the electrode. This would seem to support the adsorption model. Hull et a1.34-37 are also in favour of this model because the current-voltage curves which they obtained on rotating-wire and rotating-disc electrodes would be difficult to explain in terms of a dissolution-precipitation model. These authors also carefully observed the colour changes seen on sweeping anodically (Figure 2).

c g 0 a U

( v s . H q 1 H g O l

Figure 2 Current-potential curve, showiiig the colours foiriid, of a ziiic wire electrode in unstirred SM-KOH at a sweep rate of 1.1 mV s- ' (after Hull et al.35)

The two-dimensional model of passivation was further substantiated by Armstrong and B ~ l m a n , ~ who from results obtained by potentiostatic cathodic reduction of the film on wire electrodes with coil- <0.3 mol 1-' suggest that passivity over at least a range of 500 mV is due to formation of no more than a monolayer of an anodic film. This monolayer film was stated to be analogous to the type I1 film of Powers and B r e i t e ~ ~ ~ even though the latter authors used 7M-KOH. One anomaly of this film, pointed out by Armstrong and Bulman, was that the calculated reversible potential of the ZnlZnO electrode is 100 mV more negative than the potential at which the active-passive transition occurs. Powers and Breiter suggest that this film starts to form earlier but that the active-passive transition is not seen until almost monolayer coverage has been achieved. It should be noted that the distinction between the two-dimensional nucleation model and the adsorption model is difficult to make because in order to do so it would be necessary to

34 M. N. Hull and J. E. Toni, Trans. Faradaay Soc., 1971, 67, 1128. 35 M. N. Hull, J. E. Ellison, and J. E. Toni, J. Electrochem. SOC., 1970, 117, 192. 36 M. N. Hull and J. E. Ellison, Electrochem. SOC. Abstracts (New York Meeting, 1969),

3 7 M. N. Hull and J. E. Toni, Electrochem. SOC. Abstracts (Detroit Meeting, 1969), p. 39. 3 0 R. W. Powers and M. W . Breiter, J . Electrochem. Soc., 1969, 116, 719.

p. 596.

The Electrochemical Behaviour of Zinc in Alkaline Solution 11 demonstrate the absencelpresence of two-dimensiocal nucleation on a polycrystalline solid metal surface.

Powers and Breiter,38 by in situ photomicroscopy, suggested that under quiescent conditions the type I1 film, which they reported, was formed by direct reaction beneath a loose, flocculent, precipitated film, designated type I. Numerous authors have attempted to describe the nature and composition of the 'passivating' film but it seems unlikely that they were studying the mono- layer film because it is so thin. It was postulated by Powers and Breiter that this type I film could consist of either Zn(OH)2 or ZnO although Powers39 later explained his X-ray diffraction results obtained using horizontal electrodes in terms of zinc oxide.

On the other hand, Huber,28 under ill-defined conditions, found a coating of this type but suggested that it consisted of y-hydroxide with only a small quantity of colourless zinc oxide. This would seem to be in reasonable agree- ment with the work of Nik i t i r~a ,~~ who made an extensive study by means of microscopy, as well as by X-ray diffraction, of the precipitate formed on a zinc anode. Working with 3-1OM-KOH solutions at temperatures between -20 and +20 "C, she identified prismatic y-hydroxide and rhombic E-

hydroxide as well as zinc oxide, depending on the KOH concentration and the temperature. At ordinary temperatures, the &-hydroxide was found to be precipitated in this region but the y-hydroxide could be observed at lower temperature and lower alkali concentrations.

Under much stronger anodic polarization, slightly below the potential at which oxygen evolution occurred in KOH solutions, Hampson et al. identified zinc oxide by electron diffraction, but also noted weak lines due to y-zinc hydroxide. In this region, K a b a n ~ v ~ ~ suggested a thickening of the dark film (type I1 film), which depends on alkali concentration, and that this film contains a certain amount of zinc peroxide as well as the oxide.

The darkening of this film has been ascribed to the incorporation of an excess of zinc within the structure.28 Thus it would seem that although the composition of loosely adherent films on zinc anodes has been determined in a number of particular circumstances, the composition of the thin passivating film is unknown, because of its nature.

Several groups of workers 27*41-46 have studied the passivation of zinc in concentrated alkali using the constant-current technique and have reported that equations of the type

il* = k (17)

39 R. W. Powers, J. Electrochem. Soc., 1969, 116, 1652. 40 Z. Ya. Nikitina, J. Appl. Chem. (U.S.S.R.), 1958, 31, 209. 41 T. P. Dirkse, J. Electrochem. SOC., 1955,102,9. 42 T. P. Dirkse, D. DeWit, and R. Shoemaker, J. Electrochem. Soc., 1968, 115, 422. 43 M. Eisenberg, H. F. Bauman, and D. M. Brettner, J . Electrochem. SOC., 1961,l08,909. 44 N. A. Hampson and M. J. Tarbox, J. Electrochem. SOC., 1963, 110, 95. 4s N. A. Wampson, M. J. Tarbox, J. T. Lilley, and J. P. G . Farr, Electrochem. Technol.,

46 T. P. Dirkse and N. A. Hampson, Electrochim. Acta, 1971, 16, 2049. 1964, 2, 307.

12 Ekctrochemistry

hold for a rather wide range of concentrations. This, however, cannot be interpreted as proof for a particular mechanism of p a ~ s i v a t i o n . ~ ~ As now recognized by Hampson and Di rk~e , '~ equation (17), which is a particular form of (1 8), is valid provided diffusion is the only mode of mass transfer and the actual mechanism of passivation is immaterial.

it6 = ( c , , ~ ~ - c ) I z F j - (3 In the two different regions BC and DE in Figure 1, the current varies only

slightly with potential. This behaviour is that which would be expected if (i) an anodic film of constant composition is present in the electrode surface and (ii) the thickness of the film varies linearly with potential."*

In the present situation, this does not seem to hold since there is consider- able evidence for a constant-thickness film. Therefore it is necessary to assume that the film composition/structure varies in such a way that there still exists a constant film-solution potential difference but that the free energy of activ- ation for the passage of zinc ions through the film varies linearly with p ~ t e n t i a l . ~ This is ascribed to the tightening of the monolayer film with in- crease in the anodic potential,

Kabanov et al. 4 8 - 5 2 attempt to distinguish between chemical dissolution of the passivating layer and direct metal dissolution. The Reporters consider that the measurements are not critical enough to allow this separation.

6 The Morphology of Zinc Deposited from Alkaline Solution

A major problem in the development of a secondary battery involving zinc as an active material has been the growth of zinc dendrites on recharging. Consequently, the morphology of zinc electrodeposited from aqueous alkaline solution has been the subject of extensive studies. Diggle et aLS3 proposed a theory to account for dendritic growth and suggested that the length of a dendrite will increase exponentially with time in the limiting- current case. The reason for this is that the tip of the dendrite will grow in a spherical diffusion layer, whereas the flat surface has a linear diffusion layer.

This is shown in Figure 3, where the flux of material to the tip will be governed by equation (19)

a c ax Flux = D-

47 R. D. Armstrong, Corrosion Sci., 1971, 11, 693. 48 A. Oshe and B. Kabanov, Zaschita Metallov, 1968, 4, 260. 49 E. Ivanov, T. Popova, and B. Kabanov, Elektrokhimiya, 1969, 5, 353. so T. Popova, N. A. Simonova, and B. N. Kabanov, Elektrokhimiya, 1967,3, 1419. 5 1 T. Popova, T. C. Vidovich, N. I. Simonova, and B. N. Kabanov, Elektrokhimiya, 1967,

5 2 B. N. Kabanov, Electrochim. Acta 1971, 16, 1321. s 3 J. W. Diggle, A. R. Despic, and J. O'M. Bockris, J . Electrochem. SOC., 1968, 115, 507;

3, 970.

1969, 116. 1503.

The Electrochemical Behaviour of Zinc in Alkaline Solution 13 and as the concentration of electroactive species will be zero at the tip, this can be modified to

C Flux = D Y (20)

r0

In the limiting case, the concentration at the height y will be related to the bulk concentration cb, thus

c, = cby/S (21)

Comparing (20) and (21), equation (22) can be derived.

It is clear from this equation that the flux of material increases linearly with the height, y, and thus the height increases exponentially with time.

cb Planar diffusion --------- 7- layer

Figure 3 The model for dendritic growth due to Diggle et al.s3

From the literature, it is not clear what causes the initial formation of dendrites, though the following causes of initiation have been suggested : (i) the emergence of screw dislocations; (ii) the presence of foreign particles; or (iii) two-dimensional nucleation. 5 4 The first seems unlikely since growth patterns obtained in Ref. 54 contained no spiral formation. Support for the other two comes from the experimental result that new growth centres can be formed by periodic anodic This could be explained if zinc oxide particles were important.

Diggle ef al. 53 presented a theory concerning the mechanism of dendritic electrocrystallization which treated the initiation of a dendrite in terms of

'' R. W. Powers, Electrochem. Technol., 1967, 5,429.

14 Electrochemistry

growing pyramids. This assumes that as the pyramid grows, the effective radius of curvature of the dendrite tip decreases until it reaches a critical radius, which is the condition for dendrite initiation. Diggle suggested that dendrites may be formed at the tips of pyramids. However, Gilman and Mansfeld,” working on single crystals using in situ light microscopy and scanning electron microscopy, found that the dendrites were more frequently initiated at the feet of the pyramids. They go on to suggest that it is unlikely that a dendrite will form at the tip of a pyramid which has reached the necessarily small radius and that pyramids and dendrites grow independently, the dendrites being nucleated at macroscopic defects or impurity centres.

On the other hand, it has been suggesteds6 that the mechanism of moss formation is due to an interplay of two factors: the kinetics of zincate deposition and the state of the electrode surface. The transport of zincate ions plays a part in the theories of both moss and dendrite propagation. However, during dendrite formation, as previously stated, zinc is deposited only at the tip, whereas it is deposited at both the sides and tip of the growing mossy whisker. It is well established that dendritic growth occurs under diffusion- limited current conditions, and certain experimental results can be explained on this basis. Thus Oxleys7 showed that dendritic deposits were formed at large overpotentials whereas at low overpotentials mossy deposits were found. Moreover, mossy deposits are formed at low current densities ( N 10 mA cm -’) from concentrated zincate electrolytes, whereas dendritic deposits are produced at higher current densities ( w 100 mA cm-2) from zincate-depleted electrolytes. It has been that a critical current density is necessary for the transition from mossy to dendritic growth and that this critical current density is dependent on temperature.

Other forms of deposit have been found under different conditions. These include smooth, bright deposits, which are produced on a rotating-disc electrode at low overpotentials and high Reynolds number or on a stationary electrode by an intermittent-charging t e ~ h n i q u e , ~ ~ a fine, black deposit obtained at very low current densities ( ( 5 mA cm-2) on a stationary electrodeYs8 and granular formed at current densities (- 20 mA cm -’) intermediate between those for moss and dendrite formation. A heavy dendritic sponge has been produced at overpotentials more cathodic than that for the formation of dendrites.60

5 s F. Mansfeld and S. Gilman, J . Electrochem. SOC., 1970, 117, 1521. 5 6 H. G. Oswin and K. F. Blurton, ‘Zinc-Silver Oxide Batteries’, ed. A. Fleischer and

J. J. Lander, Wiley and Sons, New York, 1971, Ch. 5. 5 7 J. E. Oxley, ‘Improvement of Zinc Electrodes for Electrochemical Cells’, NAS5-3908,

Final Report N67-30067 (1965). 5 8 J. E. Oxley and C. W. Fleischmann, ‘Improvement of Zinc Electrodes for Electro-

chemical Cells’, First, Second, and Third Quarterly Reports, N66-13568, N66-19656, and N66-26870 (1965-1966). ‘The Growth of Dendrites from Alkaline Zinc Solutions’, Extended Abstracts, The Electrochemical Society, Battery Division, 1965, 10, 3-5.

5 9 R. Yu. Bek and N . T. Kudryavtsev, Zhur. priklad. Khim., 1961, 34, 2613, 2020 (Chem. A h . , 1962, 56, 1284d, 1284j).

O 0 Z. Stachurski, ‘Investigation and Improvements of Zinc Electrodes for Electrochemical Cells’, December 1965, N67-26278; First Quarterly Report, June 1967, N67-39803.

The Electrochemical Behaviour of Zinc in Alkaline Solution 15

Naybour6'*62 has demonstrated the effect of electrolyte flow and substrate orientation on the morphology. He showed that the substrate orientation6' has an influence on the early stages of dendritic growth, there being fewer growths on the (OOO1) plane (base) of a single crystal, but once dendrites develop the substrate orientation has little effect. In .the subsequent paper62 he correlated the deposit morphology with the Reynolds number and found that on increasing the Reynolds number the morphology of the deposit undergoes the transitions from dendritic to mossy to flat (plate-like) structures.

From the above discussion, it is apparent that factors promoting the rapid transport of the zincate ions, such as high zincate concentrations, high temp- erature, or low electrolyte viscosity, tend to suppress zinc dendrite growth. In order to control the deposit morphology, it is necessary to control some or all of these factors, and methods of doing this have included restricting the charging current density to below the limiting current density58 and using intermittent-charging techniques such as current pulses, including both charging and discharging current 6 3 and half-wave rectified a.c. 64 As pointed out by Naybour,62 these limit the charging current density without obtaining an increase in the rate of achieving an acceptable zinc deposit.

Extensive work has been reported on the effect of additives on the morpho- logy of zinc deposited from alkaline solutions. Thus Gilman and Mans- feld65-67 have studied the effect of the inorganic additives lead, tin, and tetra- ethylammonium ions on the deposition characteristics of zinc on a zinc single crystal in alkaline solution. In agreement with earlier work by K u d r y a ~ t s e v , ~ ~ Oxley and F le i~chmann ,~~ and Bockris et a/ . ,68 Gilman and M a n ~ f e l d ~ ~ found that the addition of lead ions to the electrolyte had the beneficial effect of suppressing the monocrystal dendrites of zinc. They suggest that the deposit is characterized, in the presence of lead ions, by protrusions consisting of many small crystals. It was also shown, by light and scanning electron microscopy of the surfaces, that large parts of the surface are inactive for deposition. The addition of tin ions66 was similar to that of lead ions whereas the addition of tetraethylammonium ions66 affects the dendritic structure by making the stem and the side branches shorter. Earlier suggested that as the concentration of tetraethylammonium ions is increased, or higher members of these compounds are added, dendrites are completely suppressed. However, the effect of the tetraethylammonium ions is certainly smaller than

61 R. D. Naybour, Electrochim. Acta, 1968,13,763. 6 2 R. D. Naybour, J . Electrochem. SOC., 1969, 116, 520. 6 3 S. A. Roote, F. F. Blurton, and H. G. Oswin, Paper 34, presented at Electrochemical

64 V. V. Romanov, Zhur. priklad. Khim., 1961, 34, 2692. 6 5 F. Mansfeld and S. Gilman, J . Electrochorn. SOC., 1970, 117, 588. 66 F. Mansfeld and S . Gilman, J . Elecfrochem. Soc., 1970, 117, 1154. 6 7 F. Mansfeld and S. Gilman, J . Electrochem. Soc., 1970, 117, 1328. 6 8 J . O'M. Bockris, J. W. Diggle, and A. Damjanovic, First Quarterly Report to NASA,

69 J . W. Diggle, A. Damjanovic, and J . O'M. Bockris, Paper 18, presented at Electro-

Society Meeting, Chicago, October 1967.

Contract No. NAS5-9591, March 1966.

chemical Society Meeting, Detroit, October 1969.

16 Electrochemistry

that of lead or tin at a given potential and concentration. Gilman and Mansfeld66 suggested that tin would offer the advantage over lead of having an oxidation potential which is not much higher than that of zinc.

These workers seem to be in agreement that the effect of lead and tin additives may be attributed to codeposition of the metals with zinc, the co- deposited metals blocking further zinc deposition on to preferred growth sites and leading to additional nucleation.

In the literature, there is contradictory evidence on the effect of organic additives ; K u d r y a ~ t s e v ~ ~ has claimed that organic additives have no influence on the mossy zinc deposit, whereas Stachurski60 has suggested that surfac- tants reduce the rate of growth at constant potential and also change the morphology from mossy to dendritic. Similarly, surfactants reduce the rate of growth of dendrites, and the effect of separators in reducing the rate of dendritic growth is ascribed’ to dissolved surfactants. It would seem more likely, however, and as suggested by later that the major effect of separators is to reduce the rate of zincate mass transport.

7 The Behaviour of Porous Zinc Electrodes

In previous sections the discussion of the anodic behaviour of zinc has been restricted to smooth electrodes. However, in batteries, porous zinc electrodes are frequently used. Therefore, in this section, an attempt will be made to correlate the observations made on smooth electrodes with the behaviour of porous electrodes. As shown previously, it is necessary to be able to define the modes of mass transfer precisely in order to make any quantitative predictions about the behaviour of the zinc electrode. This is a major problem with porous electrodes, and is linked to the distribution of current and potential within the pores.

One of the first systematic investigations was made by B~eiter,~’ who studied the dissolution and passivation of vertical, porous zinc electrodes in solutions of 6M-KOH-0.25M-Zn0 using a potential-sweep technique. By comparing the current-potential curves for smooth and porous electrodes, he deduced that only a small fraction of the interior of the porous electrode participated in the electrochemical processes. Nagy and Bockris 7 3 have subsequently suggested that Breiter’s conclusion holds, but only at fast sweep rates. Nagy and Bockris discharged the porous zinc electrodes galvano- statically and studied the morphology of the film formed with a scanning electron microscope. The current distribution in the porous electrode was determined by chemical analysis of microslices of the electrode after discharge. They observed within the pores a porous ‘carpet-like’ film consisting of long,

7 0 N. T. Kudryavtsev, Zhur. $z. Khim., 1952, 26, 270. 71 T. A. Kryukova, ‘The Growth of Zinc Dendrites in some Swelling Polymers’, in

‘Soviet Electrochemistry’, Proceedings of the 4th Conference o n Electrochemistry, Consultants Bureau, New York, 1961, Vol. 111, pp. 147-151.

72 M. W. Breiter, Electrochim. Acta, 1970, 15, 1297. 7 3 2. Nagy and J. O’M. Bockris, J. Electrochem. SOC., 1972, 119, 1129.

The Electrochemical Behaviour of Zinc in Alkaline Solution 17

needle crystals with occasional side arms. They assumed that this film was produced by precipitation of zincate from solution. In order to fit the current distribution observed to that expected, it was necessary to postulate a thin, highly resistive film as being simultaneously present on the zinc surface.

The precipitated film was also found in the work of R u e t ~ c h i , ~ ~ who used galvanostatic, potentiostatic, and potentiodynamic techniques combined with optical and electron microscopy to investigate the electrochemical processes at porous zinc anodes. He found that the observable film was formed by precipitation from the solution and, by X-ray and electron diffraction studies, showed that this film consisted of zinc oxide. He also reported that the unit cell (hexagonal) dimensions (a0 = 321 pm, co = 516 pm) agreed with the literature values for ZnO (325 and 520.7 pm respectively).

Hampson et al.75 suggested that the contribution from the interior of the electrode must be considerable on the basis of galvanostatic measurements on vertical and horizontal electrodes. They agree that the film is formed by a dissolution-precipitation mechanism and suggest that the major component of the electrode polarization is due to the ohmic resistance of this film.

From the above discussion, it would seem that the behaviour of porous electrodes is very similar to that of a planar electrode under quiescent con- ditions. 74 P. Ruetschi, ‘Power Sources’, Vol. 4, Oriel Press, Brighton, 1972. 7 5 R. N. Elsdale, N. A. Hampson, P. C. Jones, and A. N. Strachan, J . Appl. Electrochem.,

1971, 1, 213.