Influence of Germanium Sulfide on the Structure, Ag-Ion Conductivity and Stability of Glasses in the GeS₂-Sb₂S₃-AgI System

Abstract

This article discusses the superionic glassy GeS₂-Sb₂S₃-AgI system with mobile silver ions as a material for creating new energy-efficient solid-state ion emitters. The effect of replacing silver iodide with germanium sulfide on the structure of the electrolyte, activation energy of diffusion, and specific ionic conductivity was studied. Electrolytes (2.5 + x)GeS₂-27.5Sb₂S₃-(70 − x)AgI, x = 0, 5, 10, 15 were synthesized using the melt-quenching technique in evacuated quartz ampoules. The temperature dependence of conductivity and glass stability parameters (Hruby’s, Weinberg’s and Lu–Liu’s) were determined for them, and the mechanism for increasing glass-forming ability was clarified. It was shown that the presence of iodine in a germanium structural unit is more preferable than in an antimony structural unit; germanium structural units compete for iodine, reducing the number of SbI₃ crystallization centers and chain terminations, resulting in additional structural connectivity and stability. It was shown that when silver iodide was replaced by germanium sulfide, the decrease in conductivity due to the reduction in charge carriers was less than expected due to the expansion of the conduction channels.

1. Introduction

For the last 20 years, a class of solid-state ion emitters has appeared and has been actively developing [1]. These are all-solid-state devices where the emission of ions from a solid body occurs under the influence of an extraction potential. The emitter is a solid electrolyte deposited on a metal base, the ions of which carry out conductivity in the electrolyte. The emitter is placed into a vacuum chamber where, under the influence of an electric field from one to several tens of kilovolts, ions evaporate from the surface of the solid electrolyte. The use of such emitters as the main component of the propulsion systems of ultra-small satellites of the “CubeSat” format [2], which imposes the requirement to use the largest possible ion mass while maintaining high mobility characteristics, is promising.

One of the most popular types of electrolytes for such emitters is a silver-conductive one. Emitters with silver-ion electrolytes based on phosphate [1] glasses, silicate [3] ones, and crystalline RbAg₄I₅ [4] are known, but not with sulfide systems. In addition to ion mass, an important characteristic of the electrolyte for ion emission is ionic conductivity. A large number of sulfide silver-ion electrolytes with a conductivity of 10⁻³ S/cm and even 10⁻² S/cm are known: for example, AgI-Ag₃AsO₄ [5] materials based on AgI and Ag₂CrO₄, Ag₂MoO₄, and Ag₂WO₄ [6]. However, they lose their conductivity properties due to crystallization and degradation, which makes them unsuitable for practical use. Among crystalline superionic silver-conducting electrolytes, materials based on rubidium silver pentaiodide stand out [4]. However, they also have a problem with resistance to decomposition, react with atmospheric gases, and are unstable at low temperatures.

Of the large number of known highly conductive Ag⁺-ion electrolytes, the GeS₂-Sb₂S₃-AgI system, published by Lin et al. [7,8], stands out. Its main advantage, as described in the literature, is its low activation energy of the diffusion of silver ions (0.07 eV). The authors also declare this system to be resistant to atmospheric gases and to decomposition into phases over time. Such advantages make this system a good candidate in the search for stable glassy electrolytes with a conductivity of 10⁻³ S/cm for all-solid-state silver-ion batteries and silver-ion emitters, and the relatively low activation energy makes the conductivity properties more stable, which ensures the operation of such devices over a wider temperature range without significantly changing their characteristics. Potentially, a low value of activation energy of diffusion may have a positive effect on the threshold voltage of the silver-ion emission current. A flexible chalcogenide network may compensate for mechanical stresses arising from temperature changes during operation.

According to the available literature data, when silver iodide is introduced into non-vitreous Sb₂S₃, the structural chemical units—tetrahedra [SbS_3/2]—are transformed into structural units of the [Sb_3−xI_x] type, forming chains [8,9], which explains the region of stable glass transition in the range of 30–67 mol.% AgI [10]. Silver ions occupy the voids formed by tetrahedra, and the chains are interconnected by Van der Waals forces. With an increase in silver iodide content above 67 mol.%, the iodine content in structural units increases until the formation of SbI₃, which are chain terminators. That leads to a decrease in the average chain length and an increase in the tendency to crystallize due to a decrease in viscosity in the glass transition range.

However, the best composition, 2.5GeS₂-27.5Sb₂S₃-70AgI, obtained in [8], is still characterized by a tendency to crystallize, which is indirectly expressed in the production method by quenching in ice water. To clarify the effect of GeS₂ on the structure and understand the mechanism of increasing glasses’ stability in this system, a number of compositions, (2.5 + x)GeS₂-27.5Sb₂S₃-(70 − x)AgI, x = 0, 5, 10, 15, were chosen with the replacing of silver iodide as a source of iodine terminating the chain and reducing the glass-forming ability with germanium (IV) sulfide.

The aim of this research was to expand the range of stable solid Ag⁺-ion electrolytes with conductivity at the level of 10⁻³ S/cm and low activation energy for ion emission and other applications.

2. Materials and Methods

Glassy electrolytes (2.5 + x)GeS₂-27.5Sb₂S₃-(70 − x)AgI (x = 0, 2.5, 5, 10, 15) were synthesized from Ge, Sb, and S (all chemical grades), and AgI was mixed in a stoichiometric amount. Silver iodide was synthesized in advance from silver nitrate and potassium iodide (both chemical grade) and washed with distilled water and filtered 2 times. The mixture was placed into a quartz ampoule and then evacuated (residual gas pressure was 10⁻³ Pa). Synthesis was carried out in a rocking furnace. The synthesis temperature was 980 °C, the synthesis time was 5 h. Before quenching, the temperature was lowered to 750 °C. Quenching was carried out in water cooled to 1–2 °C. To eliminate internal stresses, annealing was carried out for 3 h at a temperature of 40 °C below the glass transition temperature.

The amorphous nature of the glasses was confirmed by X-ray diffraction (XRD) measurement (Bruker D8 Advance, Karlsruhe, Germany; 40 kV; 40 mA; CuKα).

Differential thermal analysis was carried out on a AnalitpriborTermoscan-2 device in heating mode, used to determine phase transitions of the first and second order at a heating rate of 5 °C/min; Al₂O₃ powder was used as a standard. The measurement error was no more than ±5 °C.

Raman spectra were recorded at a room temperature using a Bruker Senterra Dispersive Raman spectrometer with a confocal microscope in back reflection geometry. The radiation source was a semiconductor laser with a wavelength of 532 nm and a power of 10 mW. The radiation was focused using an X40 microlens. The measured spectra were normalized to 1 s.

The specific ionic conductivity of the glasses under study was determined using an MNIPI E7-30 immittance analyzer. Conductive layers of graphite (Graphit 33 spray) were deposited onto the resulting glassy electrolytes on both sides and were used as blocking electrodes. The measurements were carried out in the frequency range from 25 Hz to 3 MHz. The cell constant was determined as the ratio of the graphite layer area to the thickness of the electrolyte. Based on the data obtained, Nyquist plots were plotted, and the conductivity was determined as the reciprocal value of the real resistance obtained by modeling the equivalent circuit in the Z-View program.

The density of the samples d (g/cm³) was determined by hydrostatic weighing in toluene, the density of which, in turn, was controlled using a single-crystal germanium sample; lever scales of the VLR-200g-M brand were used with an error of 0.0001 g.

To assess the stability of the glass, the 3 most popular parameters were used: the Hruby’s parameter (K_H), the Weinberg parameter (K_W), and the Lu–Liu parameter (K_LL). They were calculated according to the following expressions [11]:

$$K_H={\displaystyle \frac{T_x-T_g}{T_m-T_x}},$$

(1)

$$K_W={\displaystyle \frac{T_c-T_g}{T_m}},$$

(2)

$$K_{LL}={\displaystyle \frac{T_x}{T_g+T_m}},$$

(3)

where T_x is the onset crystallization temperature, T_g is the glass transition temperature, T_m is the melting temperature, and T_c is the maximum crystallization peak temperature.

3. Results and Discussion

3.1. Glass Formation and Short-Range Order Structure

Figure 1 shows the extended ternary diagram for the GeS₂-Sb₂S₃-AgI system. The ternary diagram was originally presented by Lin et al. [7,8].

It can be seen that the samples of the system under study were obtained in the glass-formation region. Their amorphous characteristics were confirmed by X-ray diffraction. Figure 2 shows the X-ray diffraction patterns of the synthesized glasses. The figure shows that the glasses were amorphous and no crystalline phases were detected.

To establish the direct influence of germanium (IV) sulfide on the structure of short-range order, we recorded the Raman spectra of our samples when germanium sulfide replaced silver iodide in the structure. Despite the strong Rayleigh scattering in the low-frequency region, all the bands described in the literature were identified.

As it is known from the literature, with the addition of germanium sulfide (GeS₂), a new structural chemical unit [GeS_4−xI_x] is formed [12] which expands the conduction channels [8], reducing steric hindrances during the movement of ion in the structure, and reducing thereby the activation energy. However, the effect of GeS₂ has not been fully studied, in particular on the properties of ionic conductivity.

It is clear from Figure 3 that when silver iodide is replaced by germanium sulfide, the intensity of vibrations corresponding to the structural group SbI₃ [13] weakens (bands in the range of 106–163 inverse centimeters), the intensity of the band in the region of 250 inverse centimeters increases, corresponding to the structural unit [GeS_xI_4−x] [14], and the intensity in the region of 314 inverse centimeters, responsible for vibrations in Sb₂S₃, increases [15].

As can be seen from Figure 3, the bands overlap greatly, forming ranges. To obtain more information, the peaks were decomposed (Figure 4,

Table 1. Bands’ centers and their correspondence to structural elements.

Structural Element	Raman Shift, cm−1	Ref.
SbI3	38, 57, 70, 80,138, 165	[13]
SbSI	108, 115, 138, 139, 165, 290, 315	[16]
Sb2S3	107, 138, 155, 162,280, 290, 300, 308, 310, 315, 319	[15,17,18]
GeS2, S3Ge-GeS3, GeS4	101, 200, 255, 330, 340, 343, 365, 373	[15,16,18,19,20,21,22]

) into individual bands corresponding to those previously published in the literature.

Figure 4a,c shows that the intensity of the bands associated with SbI₃ 33, 70, 135, 161, and SbSI 105 cm⁻¹ is greatly reduced.

As can be seen from Figure 3 and Figure 4 and Table 1, with an increase in germanium content, the intensity of vibrations in the range of 200–450 cm⁻¹ increases significantly. In particular, the intensity of the bands in the region of 315, 290 cm⁻¹, associated with Sb₂S₃, increases, and the intensity of the bands in the region of 410, 370, 260, associated in literature with GeS₄, increases as well. In the publication [23], the band at 240 cm⁻¹ is attributed to Sb-S vibrations. The intensity of the band at 230 cm⁻¹ present in this range, attributed to Sb-S vibrations [24], decreases. It is worth noting that even without decomposing the peaks into components, with increasing germanium content the peaks shift to higher frequencies. These changes in the Raman spectra indicate an increase in the sulfur content (a decrease in x) in the structural units [SbS_3−xI_x] and [GeS_4−xI_x].

The results of Raman spectroscopy demonstrate a decrease in the intensity of the bands corresponding to SbI₃ and an increase in the intensity of the bands corresponding to Sb₂S₃, which occurs to a much greater extent than the decrease in iodine content in the glass composition due to the substitution of silver iodide by germanium sulfide. This finding suggests that iodine binds preferentially to germanium rather than to antimony. This hypothesis can be substantiated by the higher energy of the Ge-I bond in GeI₄ (219.4 kJ/mol of bonds) in comparison to Sb-I (196.8 kJ/mol of bonds) in SbI₃, as calculated in accordance with Hess’s law using thermochemical tabular data (data for calculation: ∆H⁰_formSbI_3gas = −7.9 kJ/mol, ∆H⁰_formSb_gas = 268.2 kJ/mol, ∆H⁰_formI_gas = 106.76 kJ/mol, ∆H⁰_formGeI₄ gas = −74 kJ/mol, ∆H⁰_form Ge_(gas) = 376.6 kJ/mol [25]).

Given the higher affinity of iodine for germanium as opposed to antimony, the incorporation of a germanium structural chemical unit between [SbS₂I] chains is likely to result in [GeS₂I₂] structural unit formation. This is due to the crosslinking of layers, which enhances stability and creates steric hindrances during the diffusion of ions in the conduction channel, as illustrated in Figure 5 and expressed in the following equation:

2[SbS₂I] (SbSI) + [GeS₄] → 2[SbS₃] + [GeS₂I₂]

It is also likely that the structure contains fragments where the chain termination by the [SbSI₂] terminator can be neutralized by the incorporation of a germanium tetrahedron. This should facilitate the diffusion of silver ions due to reduced steric hindrance and an increase in the average chain length, as illustrated in Figure 6 and expressed in the following equations:

2[SbSI₂] + [GeS₄] → 2[SbS₂I]+ [GeS₂I₂]

[SbS₂I] + [SbSI₂] + [GeS₄] → 2[SbS₂I]+ [GeS₃I]

Thus, replacing silver iodide with germanium sulfide leads to the following changes in the structural chemical units:

SbI₃ + [GeS₄] → [SbS_3−xI_x] + [GeS_4−xI_x]

This change in structure reduces the number of structural elements of SbI₃, which is a terminator of the antimony containing chains and also increases the average chain length, which provides greater viscosity and, as a consequence, greater glass-forming ability due to the greater number of bonds per structural unit.

Analyzing these data and comparing them with the literature on similar systems, we can conclude that the introduction of germanium sulfide increases glass stability not only due to cross-linking of [SbS_3−xI_x] chains, as described in literature [8], but also due to the reduction of the negative impact of iodine ions on the length of these chains and, as a consequence, the glass-forming ability and stability of the resulting glasses. There is a competition for iodine between the structural units [SbS_3−xI_x] and [GeS_xI_4−x]. Germanium chemical units attract part of the iodine, which reduces the average x in [SbS_3−xI_x] structural units; thus, these chains lengthen, further stabilizing the glass-forming framework.

In addition to Raman spectroscopy, we carried out differential thermal analysis of the samples under study.

3.2. Differential Thermal Analysis

The results of differential thermal analysis are shown in Figure 7. Thermograms have different scales since the 12.5Ge and 17.5Ge samples have weak thermal effects, so the thermograms of the 12.5Ge–17.5Ge samples were greatly enlarged, and to improve the signal quality, a Lowess smoothing filter and peak analyzer built into Origin Pro were applied.

An exothermic peak with a maximum of 166–182 °C can be identified in the thermogram, which practically does not change its position depending on the composition; it corresponds to the transition of β-AgI to α-AgI. This transition in crystalline AgI occurs at 147 °C, but according to the change in conductivity in [26], it is shown that in the Agl-Sb₂S₃ system, upon heating it occurs at around 180 °C, and upon cooling at 147 °C. Before the crystallization peak, there is an area of the glass transition. The temperature of this phase transition increases linearly with increasing germanium sulfide content, which may indicate a strengthening of the structure, which confirms the literature data on the cross-linking of structural chemical units of the [GeS_xI_4−x] type with chains of structural chemical units of the [SbS_3−xI_x] type.

With a germanium sulfide content increase from 2.5Ge to 17.5Ge, the intensity of the crystallization peaks decreases greatly, and in the 17.5Ge thermogram, the crystallization temperature and melting temperature of the resulting crystal are poorly distinguishable. To carry out the analysis, equal amounts of glass by mass were loaded into the ampoule, which may indirectly be evidence of the greater resistance of glasses to crystallization. For a more accurate assessment of the glass-forming ability, we calculated the Hruby’s parameter. The calculation results and phase transition temperatures are presented in

Table 2. Phase transition temperatures and the Hruby’s, Weinberg, and Lu–Liu parameters.

Sample	Tg, °C	Tx, °C	Tc, °C	Tm, °C	KH	KW	KLL
2.5Ge	199	219	232	329	0.18	0.10	0.41
7.5Ge	225	265	283	338	0.55	0.17	0.47
12.5Ge	235	280	294	343	0.71	0.17	0.48
17.5Ge	259	299	314	344	0.89	0.16	0.50

.

As can be seen from Table 2, the Hruby’s and the Lu–Liu parameters increase with increasing germanium sulfide content, while the Weinberg parameter for the 17.5Ge sample decreases. However, even the Weinberg parameter shows a tendency toward increasing glass stability. Using 3 coefficients is more objective, since different parameters are more applicable to different glassy systems [11].

3.3. Ionic Transport Properties

An example of frequency dependencies of complex resistance is shown in Figure 8.

As can be seen from Figure 8, the dependencies represent a semicircle and a polarization branch typical for cells with blocking electrodes. The temperature dependence of conductivity is presented in Figure 9; data on conductivity at a room temperature and activation energies are given in

Table 3. RT conductivity, activation energy and pre-exponent of samples under study.

Sample	σRT·10−3, S/cm	$\mathit{L}\mathit{n}({\mathit{\sigma }}_{\mathit{A}{\mathit{g}}^{+}})$	$\mathit{L}\mathit{n}({\mathit{\sigma }}_{\mathit{A}{\mathit{g}}^{+}}\cdot \mathit{T})$	$\mathit{L}\mathit{g}({\mathit{\sigma }}_{\mathit{A}{\mathit{g}}^{+}})$	$\mathit{L}\mathit{n}({\mathit{\sigma }}_0)$
		Ea, eV
2.5Ge	2.75	0.127	0.18	0.066	5.85
7.5Ge	1.96	0.15	0.18	0.064	6.33
12.5Ge	1.33	0.16	0.19	0.070	6.39
17.5Ge	0.67	0.20	0.22	0.085	6.99

.

It can be seen from Figure 9 and Table 3 that with an increase in germanium sulfide content from 2.5 to 17.5 mol.%, there is a smooth decrease in ionic conductivity. However, this decrease in conductivity from 2.5Ge to 17.5Ge for compositions with minimum and maximum germanium sulfide content is less than order, while the number of charge carriers and silver ions, according to the composition, decreased by 34%, from 25 to 18 at.%. Also, with an increase in germanium sulfide content, the activation energy for the diffusion of silver ions monotonically increases.

Ionic conductivity is an activation process, and it is described by the Arrhenius equation

$${\sigma }_{A{g}^{+}}\cdot T={\sigma }_0\cdot \exp \left({\displaystyle \frac{-E_a}{R\cdot T}}\right),$$

(4)

where ${\sigma }_{A{g}^{+}}$ is the Ag⁺-ionic conductivity, T is the absolute temperature, ${\sigma }_0$ is the pre-exponential factor, $E_a$ is activation energy of ion migration, and R is the universal gas constant.

It is worth noting that various authors use both ${\sigma }_{A{g}^{+}}\cdot T$ and ${\sigma }_{A{g}^{+}}$ variations of the Arrhenius equation [27], which may create inconsistencies in the results. The activation energy is determined from the slope of the temperature dependence of ionic conductivity in Arrhenius coordinates. It is also worth noting here that a large number of authors use the decimal logarithm instead of the natural logarithm, which creates an error in determining the activation energy since the Arrhenius equation contains an exponent. However, there are quite a lot of works using the decimal logarithm, and this should be taken into account when considering issues of determining the activation energy. Considering the above, Table 3 shows three activation energy estimates. The estimate of ${\sigma }_0$ is given for Arrhenius coordinates $Ln({\sigma }_{A{g}^{+}}\cdot T)$ − 1/T.

The activation energy values for the 2.5Ge sample in our work, 0.066 eV (Table 3), coincided within error with those published in the literature of 0.07 eV [8]. However, the conductivity is significantly different: 9.18·10⁻³ S/cm in [8] and 2.75·10⁻³ S/cm in the present study, which is probably due to our use of graphite electrodes, while the authors of [8] used gold ones.

3.4. Microscopic Model

To clarify the effect of germanium sulfide on conductivity, we used a microscopic model.

To understand ion transport at the microscopic level in glassy systems, it is possible to use the point defect model originally developed for ionic crystals but successfully applied to glassy materials [28], according to this model:

$${\sigma }_{A{g}^{+}}={\displaystyle \frac{F^2\cdot C_{A{g}^{+}}\cdot l^2\cdot {\nu }_0}{6\cdot R\cdot T}}\cdot \exp \left({\displaystyle \frac{-\mathsf{\Delta }G}{R\cdot T}}\right),$$

(5)

where F is the Faraday constant; $C_{A{g}^{+}}$ is the volume concentration of Ag ions; l is the jump distance between two stable states; ν₀ is the fundamental vibrational frequency; ∆G is the free energy change required for a diffusion jump, including both charge carrier formation and migration and expressed by the following equation:

$$\mathsf{\Delta }G=\mathsf{\Delta }H-T\cdot \mathsf{\Delta }S,$$

(6)

where ∆H and ∆S are the activation enthalpy and entropy of the ion migration. Combining equations, we obtain the following expression [29]:

$${\sigma }_{A{g}^{+}}={\displaystyle \frac{F^2\cdot C_{A{g}^{+}}\cdot l^2\cdot {\nu }_0}{6\cdot R\cdot T}}\cdot \exp \left({\displaystyle \frac{\mathsf{\Delta }S}{R}}\right)\cdot \exp \left({\displaystyle \frac{-\mathsf{\Delta }H}{R\cdot T}}\right),$$

(7)

Comparing (6) and (7), we obtain the expression for the theoretical estimate of the pre-exponential factor in the Arrhenius equation:

$${\sigma }_0={\displaystyle \frac{F^2\cdot C_{A{g}^{+}}\cdot l^2\cdot {\nu }_0}{6\cdot R}}\cdot \exp \left({\displaystyle \frac{\mathsf{\Delta }S}{R}}\right),$$

(8)

Substitute expression (8) into expression (4):

$$\begin{array}{c}\sigma T=\left[\left({\displaystyle \frac{v_0{L}^2{F}^2{C}_{A{g}^{+}}}{6R}}\right)exp\left({\displaystyle \frac{\mathsf{\Delta }S}{R}}\right)\right]exp\left(-{\displaystyle \frac{E_a}{RT}}\right),\end{array}$$

(9)

where $v_0$—the attempt frequency of silver ions [Hz], could be estimated according to [25] as 10¹³ Hz, $F$—the constant of Faraday [C/mol], $C_{A{g}^{+}}$—the volume concentration of silver ions [cm⁻³], $l$—the jump distance [cm], $T$—the absolute temperature [K], $E_a$—the activation energy [J/mol], $R$—the gas constant [J·(K·mol)⁻¹], and $\mathsf{\Delta }S$—the migration entropy change of silver ions [J·(K·mol)⁻¹]. From Equation (10), we calculated the change in migration entropy by taking the logarithm from Equation (8) and by using the values of the pre-exponential factor from Table 3. ${\sigma }_0$ was defined as a constant in the equation of a straight line in Arrhenius coordinates $Ln({\sigma }_{A{g}^{+}}\cdot T)$ − 1/T.

The results of the calculation are shown in

Table 4. Values of parameters of the pre-exponential factor calculated using the microscopic model in comparison with the experimentally determined ones.

Sample	ρ, g/cm3	Vm, cm3/mol	NAg+·10−22, cm−3	l, Å	Lnσ0exp	Lnσ0calc	ΔS, J·(K·mol)−1
2.5Ge	5.20	50.26	1.20	4.37	5.85	4.304	17.4
7.5Ge	5.10	50.29	1.20	4.37	6.33	4.297	21.4
12.5Ge	5.04	49.87	1.21	4.36	6.39	4.296	21.9
17.5Ge	4.87	50.60	1.19	4.38	6.99	4.280	26.9

, which indicated the migration entropy change of silver cations was increased with increasing GeS₂ content.

$$\begin{array}{c}\mathsf{\Delta }S=R·(\ln ({\sigma }_0)-ln\left(A\right))\end{array}$$

(10)

$$\begin{array}{c} A=\left({\displaystyle \frac{v_0{L}^2{F}^2{C}_{A{g}^{+}}}{6R}}\right)\end{array}$$

(11)

The volume concentration of charge carriers was calculated as follows:

$$C_{{Ag}^{+}}=\frac{\rho ·N_A}{M},$$

(12)

where ρ is the density of a sample; N_A is the Avogadro’s number; and M is the molar mass of a sample.

The distance between stable states is given by

$$l=\sqrt[3]{\frac{1}{C_{{Ag}^{+}}}},$$

(13)

As can be seen from Table 4, with equimolar replacement of AgI with GeS₂, the number of charge carriers and the average hop length remain virtually unchanged. Comparing the experimentally determined value of the pre-exponential factor in the Arrhenius equation with that calculated using the microscopic model, it is evident that the role of the entropy factor increases with increasing germanium sulfide content. Despite the decrease in conductivity with the introduction of GeS₂ over 2.5 mol.%, the conductivity of such glasses is higher than in similar systems not doped with GeS₂. This can be traced by comparing the conductivity of the 17.5Ge sample (17.5GeS₂-27.5Sb₂S₃-55AgI) σ_RT = 6.7 × 10⁻⁴ S/cm with the conductivity of the most similar glassy electrolyte not containing germanium 50Sb₂S₃-50AgI σ_RT = 5.0 × 10⁻⁵ S/cm [16], which indicates the positive effect of the entropy factor increases with increasing germanium sulfide on conductivity.

4. Conclusions

More stable compositions of (2.5 + x)GeS₂-Sb₂S₃-(70 − x)AgI glass system were obtained with an increased glass-forming ability, described with the Hruby’s coefficient that was more than 0.7 (K_W = 0.17, K_LL = 0.48), the ionic conductivity in the order of 10⁻³ S/cm for 2.5 ≤ x ≤ 12.5 samples, and activation energy not exceeding 0.1 eV.

The role of GeS₂ in enhancing the resistance of glasses to crystallization was clarified; GeS₂ promotes an increase of glass-forming stability due to both cross-linking of layers formed by [SbS_3−xI_x] structural units and to the competition for iodine between [SbS_3−xI_x] and [GeS_xI_4−x]. Based on Raman spectra and thermodynamic data, the formation of a bond between iodine and germanium is more preferable than with antimony. Germanium structural chemical units are able to include more iodine, up to x = 2, which does not make them a terminator of glass-forming chains. Although increasing the GeS₂ content above 2.5 mol.% negatively affects to the ionic conductivity and raises activation energy, it enhances the system’s suitability for conductivity by increasing the entropy and pre-exponential factors in the Arrhenius equation. Moreover, compared to similar glass systems without GeS₂ doping, the presence of GeS₂ has a positive effect on the ionic transport properties.