Solid State lonics 28-30 (1988 ) 1402-1405 North-Holland, Amsterdam

IMPEDANCE STUDIES ~,F THE CELL Ag/AgI/Ag [I "-ALUMINA/AgI/Ag

M.W. BREITER, H. DRSTAK and M. MALY-SCHREIBER lnstitut f ir Technische Elektrocherme, TU Wien, Vienna, Austria

Received 7 July 1987; in revised version 9 November 1987

The construction of the cell Ag/AgI/Ag I~ "-alumina/Agl/Ag is described. The impedance of this cell was measured between 10- 3 and 104 Hz at temperatures between 20 and 550°C. At temperatures below 100°C the cell impedance is determined to a large extent by the bulk resistance of the Agl layer and to a smaller extent by the impedance of the interface Ag/Agl. At temperatures between 160 and 350 ° C the impedance is controlled by the bulk resistance of the Ag [l-alumina and an impedance due to contact problems between Ag and Agl. The bulk resistance of the [l"-alumina becomes predominant between 350 and 550°C. A hindrance due to the transfer of silver ions from AgI to Ag I~"-alumina was not observable in the whole temperature range.

1. Introduction

The kinetics of the transfer of the conducting cat- ion from a solid electrolyte 1 into an electrolyte 2 has been mainly studied when electrolyte 2 was a liquid. Silver ion conductors, especially AgI [ 1-4] and A gaSI [5,6], have been investigated in aqueous electro- lytes. Organic electrolytes were used [7,8] for so- dium [3-aluminas.

Little work of a fundamental nature has been re- ported [9,10 ] for the case that both electrolytes are solids. Interest in the interface between two solid electrolytes with the same conducting ion exists for batteries [ 11 ] and sensors [9,12 ].

The symmetric cell Ag/AgI/Ag [~ "-alumina/AgI/Ag was chosen in this study. The two solid electrolytes are polyc~stalline silver ion conductors. The Ag [3 "- alumina is nat sensitive to water vapor [ 13 ]. Imped- ance spectroscopy was applied in an attempt to ob- tain info:mation on the ion transfer from one solid electrolyte m the other o n_e~

2. Experimental

Pieces ~,~ crux 1 cruX0.5 cm) were cut from bars of sodium [3 "-alumina obtained from Ceramatec Inc. as (etalyte material. They were ion exchanged in melton AgN©3 at 250°C. The specimens of Ag [3"-

alumina were repeatedly dipped into and pulled out from molten AgI. Layers of AgI were formed on them in this way. The AgI layers were sanded off on the small rectangular surfaces. They were evened on the square surfaces and had thicknesses d between 0.0 l and 0.15 cm. The contact between the AgI layer and the surface of Ag ~ "-alumina was checked by cross sectioning of a cell and looking at the interface by microscope. Good adherence was observed. Silver was sputtered onto the Agl surfaces. Then silver nets were attached by silver paint as leads. Rectangular pieces of aluminum oxide with flat surfaces were pressed against the silver nets on both sides. Tung- sten wire was wrapped around to hold the cell to- gether at higher temperatures.

The Ag/AgI/Ag ~ "-alumina/AgI/Ag cells were kept inside a small furnace in a glass vessel purged by pure nitrogen. The temperature was measured by a ther- mocouple attached to one of the aluminum oxide pieces and controlled by a temperature controller.

J

the temperature was increased step-wise from room temperature to 140°C and decreased after¢~'ards again. This is the temperature region in which the low temperature phases of AgI exist [ 14 ]. Indepen- dent measurements of the bulk conductivity of pieces of AgI, prepared in a similar fashion, proved [15] that a mixture of the [3 and T phase of AgI exists be- low 100 °C and that a transformation to the ~ phase

0 167-2738/88/$ 03.50 © Elsevier Science Publishers B.V. (North-Holland Physics Publishing Division)

M. W. Breiter et al./Cell Ag /AgI /Ag fl "-alumina/Agl /Ag 1403

occurs at temperatures between 100 and 140 °C. The measurements at temperatures between 160 and 550°C in the region of the ~ phase started at 550°C. The temperature was step-wise decreased at first and then increased again. Sufficient time was allowed be- tween individual runs for the temperature equilibration.

The measurements of the cell impedance were car- tied out between 10-3 and 104 Hz in the low tem- perature region and between 10-2 and 104 Hz in the high temperature region. An automated impedance meter produced and sold by Zahner-elektrik (West Germany) was used. The reader is referred to ref. [ 15 ] for details.

example for the low temperature region in a Bode plot. For comparison similar data are given in fig. ! b for ihe cell Ag/AgI/Ag with d=0.4 cm at 70°C. Squares denote the absolute value of Z~, and aster- isks the phase angle. Data at 300°C are presented in fig. 2 as an example for the high temperature region for the same cell.

A hysteresis phenomenon was observed in the re- gion of the low temperature phases and between 160 and 350°C in that of the high temperature phase. The absolute values of the impedances at a given fre- quency are smaller with decreasing temperature than with increasing one.

3. Results

The results of a measurement at 70°C and an A.gI layer thickness of 0.05 cm are shown in fig. la as an

9O vQ

A

a o O t 0 • . . . o • • . . o . . . . , . . . , D • • . . * • • . . ~ . ' ' 8 ? . ~

o °

L ' q ® . . . , , • • . . , . . . . . . . . " ~ • • " " ' " " e ¢ 5

. . . , ~ . . .

, - - 4 - ~ 0 . - , * . . . . ¢ " k . ' " * • • " ' * ~ . "~

" -~ -z - 1 ,~ 1 z .

l o g ~ { H e r t z )

o b

o

r - 4

0 - ~ , . . . . . . . . . . . /., . . . . . . . . . . .

- .~ -~ ~ ~ ~ .~

r ~

C, Z <

<

A

r 2

.g

0

log v ( Hertz )

Fig. 1. Impedance spectra at 70°C in Bode plots of the cells: (a) Ag/Ag~/Ag ~ "-aluminaJAgUAg and (b) Ag/AgI/Ag.

4. Discussion

Since the upper frequency limit is 104 Hz, the bulk impedances of AgI and Ag 1$ "-alumina may be con- sidered ohmic. The cell impedance is:

Zcel| = 2ZAr/A# + 2ZAgv~" + 2RA~ + R~,, . (1)

The interfaces were produced in the same way. Therefore a symmetric behavior is assumed. Com- putations of the bulk resistances from the conduc- t/v/ties, obtained by a four-probe measurement [ 15,16 ], and the geometric dimensions demonstrate for d> 0.05 cm:

r 20 ° T< ~00 °C, (a) R.~,gI> 10 R~.. ~o C< (b) R~.gx and R~,, comparable for

100°C< T< 147°C, (c) RAg~ < 10Ra. for the high temperature region. In principle the term (2ZAgAg~ + RAg~) might be

dete~Tmined experimentally from impedance mea- surements [ 15 ] of the cell Ag/AgI/Ag with the same thickness of the Agl layer as the cell Ag/AgI/Ag [~ "- alumina/AgI/Ag at a given temperature and the same frequencies. The bulk resistances Rgg~ and R~,, might be computed. The corrected impedance

Zco~ = Zc~ - 2 ZA~/Ag~ -- 2RAs~ --RV ,,

---- 2ZAg/A# ( 2 )

contains the desired information. However, this ap- proach only works in a qualitative fashion for two reasons:

(d) It is not certain that the mixture of the two low temperature phases of AgI is the same in the pro-

! 4 0 4 M.W. Breitrr et al. ~Cell Ag/Agl/Ag fl"-alumina/Agl/Ag

N m

~'~"~"5- i ~ I~

i3 ¢1 IN CI 13

, , , , l " , , ' , , t , , " i , , , , I

I ] rJ

ABSOLUTE VALUE a OF IMPEDANCE

m a _ m m " "

m ~ m ~ m M m m gl I~l gl

, 'i' 'i",'

PHASE ANGLE " ~ ": :~ :~ ~: .,:

) : I~I B'MI l ~ X ~ I " I ' ' ' I ' l ' l I~.~I 1~ ,L. b L . , ii~ . . I t L . I

Lo6 FREQUENCY (HERTZ)

Fig. 2. Impedance spectrum of the cell Ag/Agl/Ag ~ "-alumina/AgI/Ag at 300 ° C.

~ . ~

A

I l l

g.D till

U.I . .1

tlJ i f )

n

duction of the different cells. A compositional dif- ference will strongly affect R~,g~ since the T phase is a much better conductor than the J3 phase [ 14]. An effect of this type was confirmed by the measure- ments with cells of different th;.cknesses of the AgI layer. The cell impedance is nearly ohmic at 104 Hz and is approximately equal to RAgt. However, the values of RAg~ did not depend upon d in a 1L ~ear fash- ion because of the different compositior: of the mixtures.

(e) The existence [ 15 ] of poor contacts between

Ag and Agl in the high temperature region and the possibility of crack formation during the phase tran- sition from the low temperature phases to the high temperature phase introduce great uncertainty for a correction in this region. This is clearly seen between 160 and 350°C. The cell impedances at 104 Hz are practically ohmic, but differ by a factor of 3 for the three cells at 160°C. They should be equal to R~,,.

The frequency dependence of the absolute value and the phase angle may be compared in a qualita- tive fashion Ibr the cells Ag/AgI/Ag ~ "-alumina/AN/

llc 4 tic 9

R 5 . . . . . . . . . . RESISTANCE OF SOLID ELECTROLYTE

R 1, R 6 . . . . . REACTION RESISTANCE

R 3~ R 8 . . . . . DISCHARGE RESISTANCE

C 2, C 7 . . . . . REACTION CAPACITA~ICE

C 4, C 9 . . . . . DOUBLE LAYER CAPACITANCE

Fig. 3. N e t w o r k fo r t h e cell.

M. IV. Breiter et ai./Celi Ag/Agl/Ag fl"-alumina/AgI/Ag 1405

Ag and Ag/AgI/Ag. They are similar. It is suggested on the basis of this result that the frequency depen- dence is mainly due to ZA~'Agt in both cases. The in- terface Ag2/Ag [3 "-alumina does not contribute much to Z~n. The latter conclusion is confirmed by the impedance data between 103 and 104 Hz. The phase angle approaches relatively small values for the cell Ag/AgI/Ag 13"-alumina/AgI/Ag. The ohmic part of Z~n, computed in a series circuit, is close to the value of RAg1.

The impedance data in the low temperature region were analysed, using the network in fig. 3. The mean- ing of the components of this netwcck is given below the figure. The asymmetry of the network in fig. 3 originates from the way in which the computer pro- gram which leads to an opt imum fit between the ex- perimental data and the network is wrilten. If a symmetric system like Ag/AgI/Ag 13 "-alumina/AgI/ Ag is studied, the numerical values of corresponding components in the network have to turn out the same. This was the case in the present investigation. Only the interface Ag/Agl is considered. A software pro- gram, delivered together with the automated imped- ance meter, finds the values of the components which produce the optimal fit of the frequency dependence of the absolute value and the phase angle between the network and the experimental data. The results of the optimisation are represented by solid lines in fig. l a and l b, While it is possible to approximate the absolute values in a satisfactorT fashion, this is not achievable for the phase angle. It is suggested that the latter effect is due to the implicit assumption of homogeneous surfaces in the network of fig. 3. Many networks, pre, bably of the type in fig. 3, in parallel would approximate heterogeneous surfaces. The su- perposition of many time constants would yield a relatively fiat curve for the phase angle as function of frequency as it is seen in fig. l a and lb. Similar results were obtained at the other temperatures in this region.

A discussion of the data between i60 and 35u~2 is difficult because of the mentio~ed influence of the

7 Such an influence, should net exist AgI layer on ,~-~ce|l- according to the estimate given under ooint (c). The comribution of contact problems betwe~rl Ag and AN is considered responsible. (The reader is referred to the discussion in ref. [ 15 ].)

The impedance data between 350 and 550°C are relatively simple. The absolute value is nearly in- dependent of frequency between 10- 2 and 104 Hz and practically equal to R~,. The phase angle is very small. The contribution of the other components in Zcen (compare eq. (1)) is negligible relative to that of R~..

Acknowledgement

The investigations were made with partial support by the Office of Naval Research.

References

[11

[2]

[31

[41

[51

[61

[7]

[8]

[91

[101

[121

[131

[141

['51

[~61

J.H.A. Pieper and D.A. de Vooys, J. Electroanal. Chem. 53 (1974) 243. D.A. de Vooys and J.H.A. Pieper, J. Electroanal. Chem. 72 (1976) 147. K.J. Peverelli and H.P. van Leeuwen, J. Electroa,al. Chem. 110 (1980) 119 and 137. K. Kvastek and V. Horvat, J. Electroanal. Chem. 130 (1981 ) 67; 147 (1983) 83. Yu.V. Shirokov, V.S. Borovkov and P.D. Lukovtsev, Elek- trokhim. 5 (1969) 1211. Yu.V. Shirokov, P.D. Lukovtsev and V.S. Borovkov, EBek- trokhim. 7 (1971) 393 and 575. M. Voinov and H. Tar.nenberger, E~ecWochim. Acta 19 (1974) 959. G.C. Farrington, in: Electrocatalysis ca non-metallic sur- faces, ed. A.D. Franklin (NBS S0ecial Publ.) 455 11976) 341. P. Fabry, E. Sehouler and H. Kleitz, Electrochim. Acta 23 (1978) 539. A.M. Mikhailova and E.A. Uksche, Elektrokhim. 15 (1979) 1242. J.F. Kennedy, Thin Solid Films 43 (1977) 41. M. Gauthier and D. Fauteux, J. Electrochem. Soc. 131 (1984) 579. J. Garbarezyk, A. Jakubowski and H. Wasiucionek, Solid State Ionics 9/10 (i983) 249. T. Takabashi, K. Kuwabaraa and O. Yamamom, J. Elecwo-

H. Drstak and M.W. Breiter, Eleclrochim. Acta 33 (1988) 89. P. Linhardt, M. Maly-Schreiber and M.W. Breiter, in: Proc. 6th RISO Internat. Syrup. Metallurgy and Materials Sci- ence, eds. F.W. Poulser, N. Hessel Linderson, R. Claussen, S. Skaarup and O. Toil S~rensen (Ris 4 Nat. Lab., Reskilde, 1985) p. 475.