Calorimetric studies of AgI-doped tellurite glasses

Calorimetric studies of AgI-doped tellurite glasses

IOgRNALOF ELSEVIER Journal of Non-Crystalline Solids 190 (1995) 251-257 Calorimetric studies of AgI-doped tellurite glasses C.Y. Zahra, A.-M. Zahra...

371KB Sizes 0 Downloads 30 Views



Journal of Non-Crystalline Solids 190 (1995) 251-257

Calorimetric studies of AgI-doped tellurite glasses C.Y. Zahra, A.-M. Zahra * Centre de Thermodynamique et de Microcalorim~trie du CNRS, 26, rue du 141 e RIA, 13331 Marseille c~dex 3, France

Received 3 May 1994; revised manuscript received 22 March 1995

Abstract The heat capacity changes during the glass transition of TeOz-TIOo.5 glasses containing AgI or (AgI)o.75(TlI)0.25 up to mole fractions of 0.4 or 0.55 have been measured by differential scanning calorimetry. The glass transition temperatures of the binary compositions decrease with increasing T100.5/TeO2 ratio. With rising Ag ÷ concentration, the cohesion of the glass network is weakened and the structural contributions to the relaxation phenomenon as well as its activation enthalpies diminish. There is no interaction between the iodide in probably crystalline form and the host glass network. Compared with binary and ternary silver tellurite glasses, the corresponding thallium tellurite compositions are less stable.

1. Introduction New tellurite glasses based on the TeO2-AgO0, 5 system and incorporating AgI have already been described [1,2]. They possess interesting optical properties and their electrical conductivities attain values characteristic of fast-ion-conducting glasses. In the present paper, TeO2-T1Oo.5(-AgI) glasses synthesized recently [3,4] are further investigated by thermal measurements and compared with the silver tellurites of equivalent compositions. Some glasses containing the low melting eutectic mixture, (AgI)o.75(TlI)0.25, in the following designated by e, are also included in this study whose ultimate aim is to determine the effects of dopant iodides. With respect to pure AgI, the eutectic iodide mixture widens the glass-forming region and enables the fabrication of glasses containing still larger Ag ÷

* Corresponding author Tel: +33 91 28 20 50. Telefax: +33 91 50 38 29. E-mail: [email protected]

contents, hence displaying still higher electrical conductivities [3].

2. Experimental Glasses were prepared by Rossignol and coworkers [3,4] from TeO2, AgI, TII and TI2TeO 3. Their compositions were checked by weighing before and after the fabrication process. Disks approximately 1 mm thick were obtained by quenching the liquid between two blocks separated by a ring. The glasses are transparent and of yellow colour. They crystallize easily and become opaque, so that heating beyond the glass transition region and cooling at rates below 1 K / m i n have to be avoided. No other experimental problems were encountered owing to the high thermal conductivity of the glasses. Table 1 shows the 21 compositions studied which include six examined earlier [1], as additional data have been drawn from already existing and newlyperformed experiments. Mole fractions, x, as well as

0022-3093/95/$09.50 © 1995 Elsevier Science B.V. All rights reserved SSDI 0 0 2 2 - 3 0 9 3 ( 9 5 ) 0 0 2 7 7 - 4

C.Y. Zahra, A.-M. Zahra /Journal of Non-Crystalline Solids 190 (1995) 251-257


Table 1 DSC results obtained on silver and thallium tellurite glasses Symbol

Glass R



(kJmol -a)


0.231 0.233 0.227

1.53 1.55

0.35 0.35



194.6 195.6

0.230 0.225

1.57 1.55

0.36 0.35


(TEO2-0.66 AgOo.5)o.9o(AgI)o.lo

185.7 185.9

0.220 0.220

1.56 1.56

0.36 0.36


(TEO2-0.66 AgOo.5)o.85(AgI)o.15

172.0 172.2

0.207 0.205

1.54 1.53

0.35 0.34


Te02-0.86 AgOo.5


(TEO2-0.86 AgOo.5)o.9o(AgI)o.lo



(TEO2-0.86 AgOo.5)


138.5 138.1


(TEO2-0.86 AgOo.5)o.7o(AgI)o.3o


121.3 128.2


(TEO2-0.86 AgOo.5)o.6o(AgI)o.4o

450 488

91.0 89.5 90.8

0.148 0.157


177.0 177.0 163.0 164.0

201.5 201.8 202.5

ACp/Cp, 1

TEO2-0.66 AgOo.5

(xAgoo.~ = 0.46)

198.0 199.5 201.5 203.0




(xAgoo5 = 0.40)

Acp at Tg (jg-lK-1)

175.1 175.7

0.217 0.227 0.217 0.216


0.195 0.198 124.3

0.184 0.184

1.52 1.49 1.54 1.54


1.53 1.54 0.168 0.163

1.53 1.52

0.34 0.33 0.35 0.35 0.35 0.35

1.48 1.49

1.43 1.46

0.34 0.34

(X~oo ~ = 0.18)

254.2 255.2

0.24 0.23

TEO2-0.40 TIOo.5

(XTloo5 = 0.29)

212.2 213.0

0.204 0.213

TEO2-0.66 T1Oo.5

(xTloo 5 = 0.40)

171.5 172.5


150.0 151.8

0.196 0.201

140.7 141.1

0.195 0.194

131.1 134.9

0.194 0,195

1.64 1.65

0.39 0.39

121.6 121.5

0.187 0.187 0.184

1.65 1.66

0.39 0.38

(XTloo5 = 0.46)

TeO2-TIOo. 5

(XTnoo5 = 0.50)





630 640

578 530

0.20 0.20

1.62 1.65

0.33 0.31

0.30 0.32

TEO2-0.22 T1Oo.5

TEO2-0.86 TIOo.s


1.65 1.64

0.38 0.39

0.39 0.39

C.Y. Zahra, A.-M. Zahra /Journal of Non-Crystalline Solids 190 (1995) 251-257


Table 1 (continued) Symbol

Glass R

Haet (kJmo1-1)

Ts (°C)

Act, at Tg (jg-Z K - l )


Acp/Cp, 1




104.1 104.0

0.167 0.172

1.59 1.59

0.37 0.37




8 9 . 5 0.122


1.45 0.30



119.0 119.0

0.176 0.172

1.62 1.60

0.38 0.37



93.2 95.7

(0.142) (0.147)

(1.52) (1.55)

(0.34) (0.35)


(TeO2-0.86T1Oo.5)o.45(AgI o.75TIIo.25)o.55





AgOo.5/TeO 2 or TIOo.5//TeO2 ratios, R, are indicated. In general, two specimens of each glass were examined. Some compositions were prepared twice and the corresponding results are given in separate vertical columns. The specific heat capacity evolution was studied on specimens weighing less than 100 mg with the help of a thermal analyzer Perkin-Elmer DSC 4 which yields an uncertainty in the cp determination of ~< 2%. Each sample was first subjected to heating into the supercooled liquid state followed by cooling, also at 20 K/min, which made it adopt a firm position in the platinum crucible. Then its glass transition temperature, Tg, was measured during the



subsequent heating at 20 K/min. At times, the tictive temperature, Tf, was assessed according to the procedure proposed by Moynihan et al. [5]. Then the glasses were cooled at different rates, v, ranging from 1 to 20 K / m i n and the subsequent heating was invariably recorded at 20 K/rain. A mean activation enthalpy, //act, for structural relaxation may be deduced from the relation d In v / d ( 1 / T f ) = -Ha,:t/R.

3. Results

The cp curves of the glasses investigated never showed any discontinuity attributable to the 13 ~ o~















~ 0.20











> X Agl or


X e

Fig. 1. Plot of heat capacity differences at Tg against iodide concentration for series A (0), B (+), C (/, ) and D ( v ) glasses.

C.Y. Zahra, A.-M. Zahra / Journal of Non-Crystalline Solids 190 (1995) 251-257






-_ . . . . .

/ ////,"'
























T/oC Fig. 2. Evolution cp at 20 K / m i n of glasses belonging to series A.

summarizes the Tg values determined from each experimental curve as well as the differences, Ace, between the heat capacity values for the liquid (%.1) and the vitreous (Cp,g) state, taken at Tg. New sample

transition of pure AgI at 147°C or to the glass transition and/or crystallization of amorphous AgI, if present as separate phase. Only one glass transition region was observed before crystallization. Table 1


c2 O






\._. __


/,,//' / / --1





















~, T I ° C Fig. 3. Evolution of cp of glasses belonging to series C.

C.Y. Zahra,A.-M. Zahra/Journal of Non-Crystalline Solids 190 (1995) 251-257


XAgL or x~. The Acp diminish with increasing x and vanish for the hypothetical case of x = 1. A comparison of the heat capacity curves obtained on glasses heated at 20 K/rain after cooling at the same rate shows that the cp overshoot before the attainment of the supercooled liquid state diminishes with increasing AgI or e concentration, as illustrated in Fig. 2 for series A and Fig. 3 for series C. It is absent in glass C5 cooled at 5, 10 or 20 K / m i n (Fig. 4), as well as in series D. Other data given in Table 1 concern the activation enthalpies derived for series B [1] and C glasses, as well as the ratios Cp,l//Cp,g and Acp/Cp,] whose behaviour is discussed in the next section.

preparations yield slightly different results. Tg values for (TeO2) l_x(AgO0.5)x glasses with x = 0.18, 0.26, 0.33 and 0.50 are reported in Table 1 of Ref. [1] and some new determinations were added. Slight differences, greater than errors of measurement, exist with respect to Acp values given in Ref. [1], as they were measured at Te. In the case of binary glasses, a plot of Tg against the molar fractions of AGO0.5 or TIO0.s or against the ratios R shows a regular decrease of Tg with increasing x or R. Silver tellurite glasses have Tg higher than the corresponding thallium tellurites. Extrapolating both series to pure TeO2, a transition temperature of 320°C is obtained for a hypothetical TeO 2 glass. This value is in excellent agreement with the one published by Lambson et al. [6] for a (TEO2-1.5% Al20 3) glass. The glasses belonging to series A (AgO0.5/TeO 2 = 0.66), B (AgO0.5/TeO 2 --0.86), C and D (T100.s/TeO2 = 0.86) have decreasing Tg upon addition of AgI, the values for silver tellurites again being higher than for thallium tellurites. These results confirm earlier ones obtained on the same types of glass with various values of R [2-4]. The heat capacity differences between the supercooled liquid and glassy states are shown in Fig. 1 as a function of

4. Discussion

In binary (TeO2-R AGO0.5) and (TeO2-R TIO 2) glasses, an increase in R leads to falling glass transition temperatures, which is due to the growing instability of these materials. In fact, Rossignol et al. [4] have shown by infrared spectroscopy that thallium tellurite glasses depolymerize gradually. An addition of AgI or of (AgI)o.75(TlI)0.25 is also seen to decrease Tg, as the cohesion of the glass




















> T/°O Fig. 4. Evolution of % of glass C5 cooled at the rates indicated.




C.Y. Zahra, A.-M. Zahra /Journal of Non-Crystalline Solids 190 (1995) 251-257

network is weakened in the presence of increasing amounts of iodides. Table 1 shows that the ratios, CpjCp,g and AcJcp,i, remain approximately constant in all glass series up to an iodide mole fraction of 0.3. Following Ingram and co-workers [7,8], this fact may be taken as evidence that the iodide modifies neither the nature of the glass nor that of the supercooled liquid. Indeed, infrared and Raman studies on ternary silver and thallium tellurite glasses reveal only small interactions between silver iodide and the tellurite network [2,4]. Reverse Monte Carlo simulations [9] as well as experimental results cited therein and obtained on other glasses indicate also the coexistence of separate iodide and host glass networks which interact weakly or not at all. The A Cp data of the present study extrapolate to zero for XAg~ or x~ = 1, which we suggest shows that the iodide or iodide mixture is present as crystalline phase which does not contribute to the Acp jump during the glass transition. Hallbrucker and Johari [11] have already advanced this interpretation for the (AgPOa)l_x(AgI)~, glasses whose Acp values vanished in the hypothetical case of x = 1. Furthermore, Brillouin scattering data of these glasses show that the hypersonic velocities, the elastic constants and the moduli extrapolate to those of pure ot-AgI [10]. We think that the clusters resemble only a-AgI, as they do not show the ct ~ 13 transition. In the series A, B and C, the cp overshoot during reheating a sample after cooling at the same rate diminishes with increasing AgI content; the activation enthalpies also decrease (only their tendencies, but not their absolute values are taken into consideration). As illustrated in computer simulation studies [12,13] with varying activation energy and structure parameter, this decrease means that the structural contribution to the relaxation phenomenon becomes less and less important. On the contrary, the phosphate glasses studied by Hallbrucker and Johari [11] showed increasing tendency for structural relaxation with increasing AgI content, and increasing activation energy for the crystallization of these glasses was observed by Grosclaude et al. [14]. Glasses of composition C5 and series D resemble more closely the corresponding supercooled liquids, so that little relaxation and no cp overshoot occurs during cooling at rates of 5 to 20 K/rain. The ratios,


and ACp//Cp,l, are also lower when the amount of iodide exceeds a mole fraction of 0.3. The dopant salt may expand the glass network and render it more liquid-like upon cooling. In fact, the density of the tellurite glasses is seen to decrease with increasing AgI or e content [2-4]. Quite generally, substituting AGO0.5 by TIO0.5 leads to less stable and less conducting tellurite glasses. The decreased mobilities of the Ag + ions are due to higher activation energies for electrical conduction. The latter values are much lower than those for structural relaxation. This already well known fact confirms the decoupling of the motions of the conducting ions from those of the matrix atoms determining the glass transition. The structure of the dopant salt which assures the high ionic conductivity was re-examined lately by Rousselot et al. [15,16] who carried out neutron diffraction studies on (AgPOa)l_x(AgX)x glasses with X = I, Br, C1. They concluded that there is present a phase of type ct-AgI, the halide anion sublattice forming a bcc structure, over which the Ag + ions are randomly distributed in interstitial sites so as to build up A g X 4 units. A band attributable to distorted AgI 4 tetrahedra is also observed in the Raman and IR spectra of the present thallium tellurite glasses [4]. The apparition of a medium-range order in the phosphate glasses is ascribed to the presence of clusters of AgI in the glassy matrix, their sizes being 1 nm and less. Larger cluster sizes with increasing AgI may facilitate the diffusion of Ag ÷ ions via a percolation process [16,17]. The exact repartition of AgI in the host glass and the mechanism of ion conduction are, however, still open to debate (see Refs. [9,18,19]).

5. Conclusions

The effects of increasing AgI or (AgI)o.75(TlI)0.25 additions to binary (TeO2-R AGO0.5) and (TeO2-R T100. 5) glasses are manifold. (1) The glass transition temperatures diminish in each series, silver tellurite glasses possessing higher Tg than the corresponding thallium tellurite compositions. (2) During the glass transition, the structural contribution to the relaxation phenomenon and its acti-

C.Y. Zahra, A.-M. Zahra/Journal of Non-Crystalline Solids 190 (1995) 251-257

vation enthalpy decrease with increasing AgI concentration, indicating that the glass structure becomes more liquid-like. (3) There is no interaction between the iodide sublattice and the host glass network, as the heat capacity differences between the supercooled liquid and the vitreous states extrapolate to zero for pure iodide. The presence of ot-AgI-like clusters is inferred from these experimental results.

[5] [6] [7] [8] [9] [10]

This research was performed within the frame of a PIRMAT action. The authors are indebted to the CNRS for financial support.

[11] [12] [13]


[14] [15]

[1] A.-M. Zahra, C.Y. Zahra, M. Ganteaume, S. Rossignol, B. Tanguy, J.J. Videau and J. Portier, J. Therm. Anal. 38 (1992) 749. [2] S. Rossignol, J.M. R6au, B. Tanguy, J.J. Videau and J. Portier, J. Non-Cryst. Solids 155 (1993) 77. [3] B. Tanguy, J. Portier, S. Rossignol, J.M. R6au and J.J. Videau, J. Phys. (Pads) IV 2 (1992) C2-153. [4] S. Rossignol, J.M. R6au, B. Tanguy, J.J. Videau, J. Pottier, J.

[16] [17] [18] [19]


Dexpert-Ghys and B. Piriou, J. Non-Cryst. Solids 162 (1993) 244. C.T. Moynihan, A.J. Easteal and M.A. De Bolt, J. Am. Ceram. SOc. 59 (1976) 12. E.F. Lambson, G.A. Saunders, B. Bridge and R.A. E1-Mallawany, J. Non-Cryst. Solids 69 (1984) 117. M.D. Ingram, J.M. Hutchinson and A.J. Pappin, Phys. Chem. Glasses 32 (1991) 121. J.M. Hutchinson, M.D. Ingram and A.J. Pappin, J. Non-Cryst. Solids 131-133 (1991) 483. L. B6rjesson, R.L. McGreevy and J. Wicks, J. Phys. (Pads) IV 2 (1992) C2-107. L. B6rjesson, S.W. Martin, L.M. Torell and C.A. Angell, Solid State Ionics 18&19 (1986) 431. A. Hallbrucker and G.P. Johari, Phys. Chem. Glasses 30 (1989) 211. F.L. Cumbrera and A. Munoz, Thermochim. Acta 186 (1991) 69. A. Munoz and F.L. Cumbrera, Thermochim. Acta 196 (1992) 137. F. Grosclaude, J.P. Malugani, C. Rousselot, R. Mercier and M. Tachez, J. Phys. (Pads) IV 2 (1992) C2-215. C. Rousselot, M. Tachez, J.P. Malugani, R. Mercier and P. Chieux, J. Phys. (Paris) IV 2 (1992) C2-211. C. Rousselot, R. Mercier, J.P. Malugani, M. Tachez and P. Chieux, J. Phys. (Paris) IV 2 (1992) C2-219. S.W. Martin, Solid State Ionics 51 (1992) 19. K. Sebastian and G.H. Frischat, Phys. Chem. Glasses 33 (1992) 199. J.L. Souquet and D. Coppo, J. Phys. (Paris) IV 2 (1992) C2-75.