Solid State Ionics 2 (1981) 315-320 North-Holland Publishing Company

PHOTOLYSIS OF MAg4I 5 SUPERIONIC FILMS

Suresh CHANDRA, Rakesh C. AGRAWAL and Narendra SINGH Department of Physics Banaras Hindu University, Varanasi-221005, India

Received 12 January 1981; in t'mal form 30 May 1981

Photolysis studies on MAg4Is t-rims are reported. The electrical conductivities of MAg4Is t'rims have been found to change when exposed to mercury light. This is explained on the basis of a model similar to Mott-Gumey theoxy of photol- ysis for silver halides. The formation of silver specks has been confirmed by photomicrographic studies.

1. Introduction

Recently we have developed an electrocodeposition technique for obtaining films of AgI-based superionic solids [ 1 ]. The electrical conductivity [2-4] and electrochemical cell voltage [5] studies on the elec- trocodeposited films of MAg4I 5 (where M = K, Rb, NH4) indicated that they predominantly contained high conducting MAg4I 5 with traces of low conduct- ing M2AgI 3 and AgI. Silver iodide is a photosensitive compound and a number of studies have been carried out on the photolysis of silver halides [6]. Kaneda and Muzuki [7] measured the Hall effect on RbAg4I 5 in the dark and in mercury light. They concluded that free electrons are present when RbAg4I 5 is excited in Hg light. (It may be mentioned here that the Hall ef- fect results of Kaneda and Mizuki have been disputed by Knotek and Seager [8], who have not found any measurable Hall effect.) The photoconductivity studies of Kennedy and Boodman [9] are more rele- vant to the present studies though their results are on Ag3SBr and Ag3SI superionic solids. They found that photoreaction yields an Ag-ion photocurrent. In view of the above, it is thought that the electrical conduc- tivity of photolysed f'rims annealed under Hg light may be different from those annealed in the dark. In the following we describe our results for MAg4I 5 Films annealed in the dark and in light.

0 167-2738/81/0000-0000/$ 0.75 © 1981 North-Holland

2. Experimental

MAg4I 5 fdms were deposited on a silver substrate by electrolysing an aqueous solution of MI at the op- timum electrolysis conditions (for details see refs. [2-4] ). The best films for KAg4I 5 and NH4Ag4I 5 were obtained at an electrolysis temperature of 45°C whereas electrolysis at 30°C gave the best film for RbAg4I 5. These fflrns were carefully washed and air dried. For annealing in the dark, these films were kept in a closed air oven at ~150°C for ~-3 h. For annealing under Hg light, a 300 W mercury lamp was fitted in- side the air oven. Electrical conductivity of the films as a function of temperature was measured using the experimental arrangement described earlier [2 -4] . The silver plate containing films acted as one electrode and the other electrode was prepared by applying Ag- colloidal paint in a known area on the film surface. The painted electrode was applied after annealing, so that light may fall on the film for photolysis reaction to proceed. Optical micrographs were snapped by a Leica camera attached to Pan Phot metallurgical microscope.

3. Results and discussion

Figs. 1 and 2 respectively give o versus 1/Tplots for low (~25-40/am) and high (~100 tam) thickness MAg4I 5 fflrns. The tailed points are the conductivity values when annealed in light whereas the other values

316 S. Chandra et al. / Photolysis o f Mag415 superionic films

2

10 0

5x1~'

2x1(~'

1 0 - I

2 ~

I 10 0

re ~ s~10-'

o

-- 2 xl(3'

RbAg4I 5 (Low thickness)

I I I

NH4Ag415 (Low thickness)

I i I I i

2, KAg4I 5 (Low thickness)

10 0

5x10-' a

2x10 -~ 2 . 4 2.6 2.8 3.0 3.2 3.6

t_9_x 3( *K -I ) .

Fig. 1. o versus l IT plots for low thickness (~-25-40 pm) MAg4I s (where M = Rb, Nil 4, K) electrocodeposited films obtained at 30, 45 and 45"C electrolysis temperatures re- spectively. The tailed point values are for photolysed films while other values are for films annealed in the dark.

are those for annealing in the dark. Table 1 gives the percentage change in 030 (the electrical conductivity at 30°C) and the change in AE (the activation energy) for MAg4I 5 Films. AE has been evaluated from the low-temperature range of the o versus lIT plot before

T q = o

o

2x10 -I

i

.10- '

5x10 -2

2x16 2

~ RbAg/,Z 5 ( High thickness)

~-~--~-

i I i L ~ 1 i

NH4AgtI 5 (High thickness)

2 -

5x10 -~ i i i i - i ~ I

3 KAgz.•5 (High thickness)

1 0 o

5 x 1 0 - f

?.xlO -j

2'.4 216 2'.8 3'.0 3',2 3'.4

IOa(,K-, i ~ T

Fig. 2. o versus I /T plots for high thickness (~100 pm) MAg4Is (where M = Rb, NH4, K) electrocodeposited Films obtained at 30, 45 and 45°C electrolysis temperatures re- spectively. The tailed point values are for photolysed films while other values are for films annealed in the dark.

a sudden jump in the conductivity starts appearing due to the 3 ~ a transition of AgI (for a detailed dis- cussion see refs. [2 -4 ] ), which is present in all our electrocodeposited Films. From figs. 1 and 2 and table 1, it is obvious that the presence of light during annealing does alter the behaviour of the films. The

Table 1 The change in electrical conductivity at 30* C (o30) and activation energy AE of MAg415 films of two thicknesses. The subscripts

D and L denote films annealed in the dark or in light, respectively

Sample Low thickness film High thickness film

lO0(a D - aL)/a D &E D - AE L lO0(o D - aL)/O D AE D - &E L (%) (eV) (%) (eV)

KAg4I s 0 0.01 -31 0.01

NH4Ag4I s - 7 0.02 -25 0.04

RbAg4I s - 11 0.02 -200 0.02

S. Chandra et al. / Photolysis of Mag415 superionic films 317

change is more apparent in conductivity values (par- ticularly conductivity values of high thickness Films) than in activation energy values, because the relative accuracy of our measurement for o is better. The fol- lowing paragraphs explain the change in conductivity on the basis of photolysis of the f'tim materials.

Photolysis of silver halides has been extensively studied in the past particularly because of their im- portance in photographic films (for a review see ref. [6]). The room-temperature band gaps of some silver halides, viz. AgBr, AgC1 and/~-AgI are 2.5, 3.0 and 2.8 eV respectively which are comparable to the band gap of MAg4Is, ~3-3 .2 eV [ 10,11 ]. Further, the preliminary band structure of RbAg4I 5 as proposed by Baur and Huberman [11] (shown in fig. 3) is anal- ogous to/3-AgI [12,13]. So, the MAg4I 5 photolysis reaction can be described similar to that for silver halide/~-AgI (or AgC1, AgBr).

The well known Mott-Gurney [14] theory of photochemical reaction in silver iodide occurs as fol- lows *:

* The modified defect notation of Kr/iger [ 15 ] is used. The sign of the charge (', ") indicates respectively the real neg- ative and positive charge in the ionic lattice.

Energy E teV)

• . . . . . . . ,

4 - - " . . . . . - . ' . - . ' ~ - ' i

1 E

J r ed ,p) ~S

Fig. 3. Room-temperature band structure of RbAg4I s. Light shaded region in I(5p) band shows the smeared I(5p) states [13].

AgAg ~ V'Ag + Ag i,

h p t 2I 1 ~ 2e + 2I i,

A~i + e' -+ Agmeta 1 (I)

Agmeta I + e' -~ Agmeta l , P

AgAg + Agmeta I 2(Ag)metal + V'Ag , I

2Ii + 2V'Ag -~ 2(VAgI )bulk -~ I2t,

hv 2AgI -~ 2(Ag)metal + 121'. (A)

Hamilton [16] suggested that the initial electron trapping centers in silver halides are interstitial silver ions. The silver metal specks formed by the above photoreactions further act as the trapping centers for new photogenerated electrons. These specks are spread all over the material. Thus, after photolysis, the material can be assumed to look like a system as shown in fig. 4. The photographs given in fig. 5 con- firm the presence of silver specks. Here the silver specks are embedded in a matrix of ionic solid. The amount of photolysed product is likely to be large on the surface because of the higher photon density available.

" / . ~ . . . .

?i.!. ::..;:

! i,::.:i:¢ i.) • . . .

- - M e t a l l i c p h o t o l y s i s product • ~:"~ - - S u p e r i o n i o solid

Fig. 4. Schematic illustration showing the manner in which metallic photolysis product (Ag) is dispersed in superionic solid matrix.

318 S. Chandra et al. / Photolysis of Mag415 superionic films

i iiii i¸ iii~

ii! ! iii l/ !i iil/ iii : iii ll i i! iii ' i!i li !!!i ¸ • 7 ' ~ ~ii ~ iill i !i ~!~i i l l :!i i ii i i ii i !! !ii iii~i~ii ii!! ii

Fig. 5. Photomicrographs (magnified 8800 times) of MAg4Is films photolysed for different times (t) with an incandescent 60 W lamp. (a) KAg4Is, (b) NHaAg4I s, (c) RbAg4I s,

Apart from the metal speck formation [photoreac- I

tion (I)] , the negatively charged Agmeta 1 specks [see the fourth step of the above photoreaction (A)] may attract positive holes h" (or I ' ) produced by the re- moval o f photoelectrons #. The result of this process

# I" (" representing the real positive space charge) is obtained by photon absorption (see pp. 1053-1056 of ref. [6] ) re- leasing an electron from the valence band (5p shell of I) in- to the empty conduction band.

photon e' I] -x ~, + I'.

S. Chandra et al. / Photolysis of Mag415 superionic films 319

would be "re-formation" of silver halide [6] :

Agmeta 1 + I" ~ AgI. (II)

Assuming MAg4I 5 photolysis to be similar to AgI (giving Ag-metal and iodine), the following reaction can be written:

hv xMAg4I 5 -+ MxAgvlx+ v + (4x - y)AgI

-~ MxAgylx+y + (4x - y)Ag + ½(4x - Y)I2t.

One of the more probable photodecomposition products of MAg4I 5 may be M2AgI 3 [17]. In that case the above photoreaction can be written as:

hv 2MAg4I 5 ~ M2AgI 3 + 7AgI, (Ill)

M2AgI 3 + TAg + 7121". (III')

Photoreaction (Ill) is a consequence of photoreac- tion (I) as applied to the superionic solid system MAg4I 5. Similarly, we can write a relevant photoreac- tion for our present system corresponding to photo- reaction (I1).

As discussed in our earlier papers [2 -4] , the elec- trocodeposited films contain traces of M2AgI 3 and AgI. With photoreaction (II) as an intermediate step, a rearrangement of ions of the above constituents may take place giving the following possible photo- reaction:

hv M2AgI 3 + 7AgI (or Ag' + I') ~ 2MAg4I 5. (IV)

change in electrical conductivity due to photolysis is also in the same sequence. Thus, silver speck forma- tion seems to have a direct relevance to the conduc- tivity results. It is imperative that the specks would be partially inside the surface of the film and may even interconnect few internal specks formed inside the film up to the photon penetration depth. As a consequence, the effective thickness of the films would decrease and hence the apparent o would in- crease as observed. The electronic conduction would still be normally negligible in the composite photol- ysed system because the specks are embedded in a medium of MAg4I 5 (see fig. 4). As such, no direct conduction channels/paths would be available for electrons.

Photoreaction (IV) may also increase the con- ductivity because the resulting product MAg4I 5 has a higher conductivity than the starting materials M2AgI 3 and AgI. But the final results have to be mediated via photoreaction (III'). Photoreaction (III') [or combined photoreactions (IV) + (III')] can explain our conductivity results given in table 1 where we see that {r L > o D for nearly all the films except for low thickness KAg4I 5 fdm in which the change in o was within the experimental error. The increase is more predominant in high thickness films. This was expected, because high thickness films con- tain more impurity compounds (for details see refs. [2-4] ) and photoreaction (IV) plays a more impor- tant role.

Now, we try to understand the changes in electri- cal conductivity of MAg4I 5 films in view of the above possible photoreactions.

The products of photoreaction (III) would de- crease o due to the formation of low conductivity photoproducts M2AgI 3 and AgI, which is contrary to our experimental results given in figs. 1 and 2 and table 1. Let us consider that AgI so formed further photolyses and gives (Ag) metal specks as in photo- reaction (ill'). Obviously, the specks would spread all over the exposed surface. If we examine carefully the photographs given in fig. 5, we note that the num- ber and size of Ag specks change with (i) different times of exposure and (ii) different compounds. The amount of photospecks in the three films is in the sequence KAg4I 5 < NH4Ag4I 5 < RbAg4I 5. The

Acknowledgement

The authors are thankful to Professor S. Ranganathan and Mr. G.M.K. Sarma of the Depart- ment of Metallurgical Engineering, Institute of Technology, Banaras Hindu University, for providing optical microscopic facilities.

References

[ 1 ] S. Chandra, Superionic solids: principles and applications (North-Holland, Amsterdam, 1981).

[2] S. Chandra and V.K. Mohabey, Phys. Star. Sol. 53a (1979) 63.

[3] V.K. Mohabey, Thesis, Ravishankar University, Raipur, India (1978).

320 S. Chandra et al. / Photolysis o f Mag4I 5 superionic films

[4] S. Chandra, J.N. Sharma, V.K. Mohabey and R.C. Agrawal, J. Phys. D13 (1980) 495.

[5] S. Chandra and R.C. Agrawal, in: Golden Jub. Comm. Vol. Nat. Acad. Sei., ed. U.S. Srivastava (Naya Prokash, Calcutta, 1980) p. 429.

[6] A.C.H. van Peski, in: Physics of electrolytes, Vol. 2, ed. J. Hladik (Academic Press, New York, 1972) p. 1051.

[7] T. Kaneda and E. Mizuki, Phys. Rev. Letters 29 (1972) 937.

[8] M.L. Knotek and C.H. Seager, Solid State Commun. 21 (1977) 625.

[9] J.H. Kennedy and E. Boodman, J. Phys. Chem. 74 (1970) 2174.

[10] S. Chandra and V.K. Mohabey, J. Phys. D8 (1975) 576.

[11] R.S. Bauer and B.A. Huberman, Phys. Rev. B13 (1976) 3344.

[12] J. Bohandy, J.C. Murphy, K. Moorjani and P.E. Fraley, Phys. Stat. SoL 49b (1972) K91.

[13] M. Caxdona, Phys. Rev. 129 (1963) 69. [14] N.F. Mott and R.W. Gurney, Electronic processes in

ionic crystals (Dover, New York, 1964) ch. 7. [15] F.A. KrUger, The chemistry of imperfect crystals

(North-Holland, Amsterdam, 1964) ell. 7. [16] J.F. Hamilton, Paper presented at 1966 Colloquium on

the Photographic Interaction between Radiation and Matter, Soc. Phot. Sei. Eng., Washington D.C.

[17] L.E. Topoi and B.B. Owens, J. Phys. Chem. 72 (1968) 2106.