AlCl₃-NaCl-ZnCl₂ Secondary Electrolyte in Next-Generation ZEBRA (Na-ZnCl₂) Battery

Abstract

Increasing demand to store intermittent renewable electricity from, e.g., photovoltaic and wind energy, has led to much research and development in large-scale stationary energy storage, for example, ZEBRA batteries (Na-NiCl₂ solid electrolyte batteries). Replacing Ni with abundant and low-cost Zn makes the ZEBRA battery more cost-effective. However, few studies were performed on this next-generation ZEBRA (Na-ZnCl₂) battery system, particularly on its AlCl₃-NaCl-ZnCl₂ secondary electrolyte. Its properties such as phase diagrams and vapor pressures are vital for the cell design and optimization. In our previous work, a simulation-assisted method for molten salt electrolyte selection has shown its successful application in development of molten salt batteries. The same method is used here to in-depth study the AlCl₃-NaCl-ZnCl₂ salt electrolyte in terms of its phase diagrams and vapor pressures via FactSage^TM and thermo-analytical techniques (Differential Scanning Calorimetry (DSC) and OptiMelt^TM), and their effects on battery performance such as operation safety and charging/discharging reaction mechanism. The DSC and OptiMelt results show that the experimental data such as melting temperatures and phase changes agree well with the simulated phase diagrams. Moreover, the FactSage^TM simulation shows that the salt vapor pressure increases significantly with increasing temperature and molar fraction of AlCl₃. The obtained phase diagrams and vapor pressures will be used in the secondary electrolyte selection, cell design and battery operation.

1. Introduction

With fast-growing population in the world, the global energy consumption increases every year [1]. To cope with the challenges in rising energy demand and considering climate change by greenhouse gas emission, it is important to realize the energy transition by increasing the use of renewable clean energy from, e.g., photovoltaic (PV) and wind energy. Increasing share of these intermittent renewable energy resources in the energy system has promoted new research and development in batteries for large-scale stationary energy storage, such as ZEBRA battery (Na-NiCl₂ solid electrolyte battery) [2]. This commercial ZEBRA battery uses commonly available materials and safe reagents, and therefore has several advantages (e.g., higher safety, longer lifetime) over other batteries such as the lithium (Li)-ion battery [3] and Na-S battery [4].

The ZEBRA battery is a kind of high temperature molten salt battery, having an operating temperature around 300 °C, a voltage of approx. 2.58 V and an energy efficiency of >80% [3,4,5]. It has a high energy and power density of about 90 Wh/kg and 150 W/kg, respectively, and can withstand a high number of charging and discharging cycles (>5000 cycles), making it an attractive option in various energy storage applications including stationary energy storage [6]. FZSoNick, which is the global leading ZEBRA battery manufacturer and also one R&D partner in the EU project funding this work, reports that more than 40,000 ZEBRA battery systems with >500 MWh have today been installed globally for telecom, railways, energy storage, electrical mobility, etc. [7]. The electrochemical reaction in the ZEBRA battery is shown as follows:

$$2\mathrm{Na}+{\mathrm{NiCl}}_2\leftrightarrow \mathrm{Ni}+2\mathrm{NaCl}$$

(1)

In this battery (see Figure 1 left), a porous NiCl₂/Ni+NaCl electrode infiltrated with NaAlCl₄ is used as a positive electrode (cathode) and liquid Na produced during charging as a negative electrode (anode). The mixture of AlCl₃-NaCl (50-50 mol%), i.e., NaAlCl₄, is used as the main part of the secondary salt electrolyte in the positive electrode, which has a low melting point of about 156 °C and acts as a good Na⁺ conductor. A beta’’-alumina solid electrolyte (BASE) is employed as the primary electrolyte, which has a good Na⁺ conductivity of around 0.5 S cm⁻¹ at operating temperature. The primary BASE electrolyte is also used to physically separate the anode and cathode compartment to achieve low self-discharge for high energy efficiency [8].

In order to further improve the competitiveness of the ZEBRA battery on the energy storage market, much R&D effort was made in the past decades on next-generation ZEBRA battery to further increase its storage performance (e.g., efficiency, lifetime) and also to reduce its costs. The limited reserves and high costs of nickel are major factors impeding the widespread and large-scale adoption of ZEBRA technology [9]. In the material unit storage cost of the final cell (28 USD/kWh), Ni cost (18 USD/kWh) has the highest share with about 63%, and constitutes about 25% of the overall battery cost (73 USD/kWh) [10]. Replacing Ni with abundant and low-cost Zn, the material cost share of the cathode metal (4 USD/kWh) can be significantly reduced to about 23%, which can cut down the cell material and overall battery costs by around 46 and 20%, respectively. In addition, ZnCl₂ has a much lower melting temperature than NiCl₂ (290 °C, compared to about 1000 °C) [11]. While the electrochemical reactions in nickel electrodes take place via dissolved metal ions in the secondary electrolyte despite the high melting temperature of NiCl₂ [12], significant differences in the reaction mechanism and kinetics are expected in zinc electrodes. Bones et al. [13] previously reported that the Na-NiCl₂ ZEBRA battery operating at 300 °C experienced premature capacity loss due to the undesirable growth of Ni particles in the cathode. In contrast, Lu et al. [9] demonstrated that the Na-ZnCl₂ based battery, with its lower melting point of 240 °C, exhibited the formation of a liquid phase at 280 °C. This liquid phase effectively suppressed the growth of Zn and NaCl particles, presenting a distinct advantage for Na-ZnCl₂-based batteries. Furthermore, this battery showed excellent cycling stability for the Na-ZnCl₂ cell at 280 °C, displaying a four-step charging and discharging behavior. The conducted cycling tests at a C/2 rate (30 mA) between 48% and 90% state of charge (SOC) (65 mA h) for 56 cycles, revealing promising short-term cycling performance. However, it is essential to acknowledge that the long-term cycling performance of the Na-ZnCl₂ battery remains an area that has not been extensively explored in the literature. Another important factor is the vapor pressure, which might potentially affect the pressure management in solid-electrolyte batteries such as the ZEBRA batteries [14].

Lu, et al. [9,10,11,12,13,14,15] have studied next-generation Na-ZnCl₂ ZEBRA cells and their reaction mechanism, applying a porous Zn/NaCl cathode with AlCl₃-NaCl-ZnCl₂ secondary electrolyte (see schematic representation in Figure 1, right). The electrochemical reaction in the Na-ZnCl₂ ZEBRA battery is shown in Equation (2). These cells showed a voltage of round 2.1 V at 280 °C, and more stable cycling at 280 °C, compared to 240 °C, which was ascribed to formation of a liquid phase (Na₂ZnCl₄) in the binary NaCl-ZnCl₂ system above 250 °C [9]. The NaCl-ZnCl₂ phase diagram was further used to identify four stages occurring during reversible charge/discharge cycling—but in these explanations ignoring the presence of AlCl₃ [9]. The cathode is a key component for the performance of ZEBRA and Na-ZnCl₂ cells, as its microstructure largely determines the electronic and ionic cell resistance as well as life-time of the battery. With this in mind, Lee et al. [16] improved the cell performance significantly by using a carbon felt to reduce cathodic resistance. To analyze and optimize the influence of the microstructural properties of the cathodic region at different State of Charges (SOC) of the Na-ZnCl₂ battery, Godinez-Brizuela et al. [17] developed a synthetic geometric model for the cathodic region for microstructural analysis of its effective conductivity, which could be used to assist the cathode improvement.

$$2\mathrm{Na}+{\mathrm{ZnCl}}_2\leftrightarrow \mathrm{Zn}+2\mathrm{NaCl}$$

(2)

All the previous works above suggest that the phase content, in particular the presence of ZnCl₂-containing molten phases, has a significant effect on the cell performance of the Na-ZnCl₂ battery. In this context, it is important to note that the presence of AlCl₃ in the AlCl₃-NaCl-ZnCl₂ secondary electrolyte has significant effect on the battery behaviors. This affects phase formation, melting temperature, vapor pressure, and reaction mechanisms in this system, which can affect battery performance and safety significantly.

In our previous work, a simulation-assisted method based on thermodynamic simulation via FactSage^TM and thermal analysis via Differential Scanning Calorimetry (DSC), and a melting point apparatus (OptiMelt^TM) was successfully applied for selection of a molten salt electrolyte for Na-based liquid metal battery (Na-LMB), which is another type of molten salt battery without BASE solid electrolyte and uses the molten salt electrolyte as the primary electrolyte [18,19,20,21]. A low-melting-point NaCl-LiCl-KCl molten salt electrolyte was selected based on phase diagram and other experimental data. The Na-LMB test cell with the selected molten salt electrolyte shows an energy efficiency of about 80% and a coulombic efficiency of >97% [20,21]. Moreover, the Levelized Cost of Storage (LCOS) of the Na-LMB with the selected NaCl-LiCl-KCl molten salt electrolyte was estimated to be 0.027~0.029 USD/kWh for scaled application conditions, based on a 1MW/5MWh demo energy storage plant made of 100 Ah battery monomers [18].

In this study, the AlCl₃-NaCl-ZnCl₂ salt system (secondary salt electrolyte of Na-ZnCl₂ battery) was in-depth investigated with thermodynamic simulation via FactSage^TM, thermal analysis via Differential Scanning Calorimetry (DSC) and a melting point apparatus (OptiMelt^TM), used in our previous work. Formation of different phases at different compositions (i.e., variable SOC and AlCl₃ content) and their melting points were simulated and/or measured. Moreover, thermodynamic simulation on the salt vapor pressures in the operating temperature range was used to find the safety limit of the composition. Overall, this paper presents both simulated and measured data pertaining to the key thermal properties of the AlCl₃-NaCl-ZnCl₂ salt system, along with their application for battery design and enhancement. Notably, our study employs simulated phase diagrams to conduct a unique analysis of the phase formation within the battery during charging and discharging processes. This analysis offers insights into the battery’s composition and behavior at different states of charge (SOC), representing a novel contribution that has not been previously documented in existing studies.

2. Materials and Methods

2.1. FactSage^TM Thermodynamic Simulation

FactSage^TM 8.1 was used to simulate phase diagrams of the NaCl-ZnCl₂-AlCl₃ secondary salt electrolyte in the novel ZEBRA battery. The simulation was performed in the Phase Diagram mode of the FactSage^TM. In the simulation, two moles of NaCl and the other salt species (i.e., ZnCl₂, AlCl₃) as one mole were used for better use of the phase diagrams for battery design and optimization, because the charging and discharging reversible reaction involves two moles of NaCl to react with one mole of Zn as shown in Equation (2). Firstly, the ternary phase diagram of (NaCl)₂-ZnCl₂-AlCl₃ was simulated and studied. Moreover, after simulation of the ternary phase diagram, binary phase diagrams at some selected constant concentrations of AlCl₃ were plotted in the same module to achieve a better understanding of the ternary phase diagram. Additionally, the vapor pressures for different salt mixtures at different temperatures were also studied for the analysis of the maximum allowable temperature for the battery.

2.2. Sample Preparation

For the experimental studies, pure chloride salts including AlCl₃ (99% purity, anhydrous, Alfa Aesar, Kandel, Germany), ZnCl₂ (99% purity, anhydrous, Alfa Aesar, Kandel, Germany) and NaCl (99% purity, anhydrous, Alfa Aesar, Kandel, Germany) were used to synthesize the AlCl₃-NaCl-ZnCl₂ salt mixtures. Due to the hygroscopicity of chloride salts (particularly AlCl₃ and ZnCl₂), the storage, weighing, mixing and sample preparation of the salts were carried out in a glovebox (GS Glovebox System technik GmbH, Glovebox Mega 2, O₂ < 0.5 ppm, H₂O < 1 ppm, Malsch, Germany) swept with ultra-high-purity argon gas (Ar 5.0, purity > 99.999%). All salt mixtures (~5 g) for thermal analysis were grinded in a ceramic mortar for at least 30 min by hand, so that they were mixed well with acceptable small salt particle size for thermal analysis.

2.3. Differential Scanning Calorimetry (DSC) for Thermal Analysis

Differential Scanning Calorimetry (DSC) was used to determine exact melting points of the studied salt mixtures with different compositions. The measurements were carried out with a commercial equipment of DSC 204 F1 Phoenix^® (NETZSCH, Germany). The temperature calibration of the DSC equipment was conducted with standard materials (In, Sn, Zn, Bi, CsCl, Benzoic Acid, Biphenyl and KClO₄).

Sealed gold platted stainless steel crucibles were used for the DSC measurements due to high vapor pressure of AlCl₃ [19] and strong corrosiveness of the salt mixtures at high temperatures. About 10 mg of each prepared salt mixture was put into the crucible in the glovebox. Next, the crucible with the salt mixture was sealed with the gold platted steel lid by a toolkit in the glove box. Afterwards, all the crucibles were transferred out of the glovebox in sealed bottles with argon gas, then weighed and put into the DSC equipment (exposure time in air of <5 min).

The following parameters were set in the DSC measurements: Heating and cooling rates were 10 K/min, while protective gas flow around the furnace was N₂ (purity > 99.999%, Linde gas, Germany) at 20 mL/min. The reference crucible for the DSC measurement was an empty gold platted stainless steel crucible, which is identical to the test crucible with the salt. Three cycles of DSC (heating and cooling) between 30 °C to 410 °C were performed for each salt sample. The first cycle was not generally considered for the melting temperature determination because the salt mixture is non-homogeneous in the first cycle. One of the second and third cycles was selected to determine the melting temperature, when these two cycle curves were very close to each other.

In total, 11 salt compositions were selected for the DSC measurements as shown in

Table 1. Salt mixtures with selected compositions analyzed by DSC.

Sample No.	SOC (%)	AlCl3-ZnCl2-(NaCl)2 (mol%)
1	100	15:77.5:7.5
2	71	15:55:30
3	100	25:62.5:12.5
4	48	25:30:45
5	100	35:47.5:17.5
6	42	35:20:45
7	0	34.9:0:65.1
8	50	34.9:23.8:41.3
9	76.5	34.9:36.5:28.6
10	87.6	34.9:41.8:23.4
11	100	34.9:47.7:17.4

. Among them, samples No. 1–6 represent the salt compositions with different AlCl₃ concentrations (15, 25 and 35 mol% in AlCl₃-ZnCl₂-(NaCl)₂) at 100% SOC and a representative SOC. As NaAlCl₄ in the secondary salt electrolyte must not be consumed, 100% SOC of the battery is defined as 1:1 mol% of NaCl and AlCl₃ to maintain the basicity of the overall salt for compatibility with the BASE electrolyte [5]. Samples No. 7–11 (34.9 mol% AlCl₃ in AlCl₃-ZnCl₂-(NaCl)₂) correspond to realistic Na-ZnCl₂ cathode compositions with 30% active metal content (e.g., 65.1 wt% Zn, 34.9 wt% NaCl at 0% SOC) and ~50 vol% NaAlCl₄ at 280 °C (→ 45.9 wt% Zn, 24.6 wt% NaCl, 29.5 wt% NaAlCl₄ at 0% SOC). As sample No. 7 (0% SOC) and sample No. 11 (100% SOC) contain 0 mol% and 47.7 mol% ZnCl₂, respectively, the salt composition (i.e., mol% ZnCl₂ and (NaCl)₂) at various SOCs can be easily calculated from the SOC value and the constant AlCl₃ content $x_{AlC{l}_3}$:

$$x_{ZnC{l}_2}\left(mol\%\right)=(x_{ZnC{l}_2, 100\%SOC}-x_{ZnC{l}_2, 0\%SOC})·SOC=47.7·SOC,$$

(3)

$$x_{(NaCl{)}_2}\left(mol\%\right)=100-x_{AlC{l}_3}\left(34.9 mol\%\right)-x_{ZnC{l}_2}\left(mol\%\right).$$

(4)

$x_{ZnC{l}_2, 100\%SOC}$ and $x_{ZnC{l}_2, 0\%SOC}$ is the molar fraction of ZnCl₂ in the salt mixture of the fully charged (100% SOC) and discharged (0% SOC) battery, respectively.

2.4. Melting Point Measurement Apparatus (OptiMelt^TM)

The Automated Melting Point System—OptiMelt^TM (MPA 100, Stanford Research Systems) is a melting point measurement apparatus, which records the melting process using a camera. The melting can be observed from the camera and is recorded for further analysis. The salt mixture samples No. 1–6 were filled 3 mm high inside the capillary tube with about 1.5 mm inner diameter and finally inserted in the OptiMelt^TM. The following parameters were set: the temperature range was 30–400 °C, heating rate of 10 K/min was applied, which was consistent with the DSC measurements. For every experiment, two cycles were performed. In the same way as for the DSC measurements, the first cycle was used to homogenize the salt mixture and remove air bubbles in the salt. The images of the second cycle from the built-in camera were used to determine the melting temperature of the studied salt sample and to further study its melting process. The temperature at which melting starts in the salt is defined as the melting temperature of the salt. Since the results were obtained from visual inspection, the accuracy of measured melting temperatures is not as good as in the case of DSC. However, the OptiMelt^TM experiment is still a good supplement to determine/confirm the melting point and study the melting process of the salt.

3. Results and Discussion

3.1. Simulated Phase Diagrams

In order to in-depth investigate the equilibrium phases in the NaCl-ZnCl₂-AlCl₃ system, the phase diagrams were simulated with FactSage^TM. Since the overall charging and discharging reaction of Na-ZnCl₂ batteries involves the reversible consumption of two moles of NaCl to form one mole of ZnCl₂, as shown in Equation (2), all phase diagrams in this work were simulated for (NaCl)₂ instead of NaCl to ease their interpretation in terms of cell SOC.

The simulated ternary phase diagram of AlCl₃-ZnCl₂-(NaCl)₂ (1 atm, 300 °C) in Figure 2 shows that at a high concentration of AlCl₃ (>50 mol %) the melting point is <300 °C, and much salt vapor is present. However, this is not the case during normal operation of ZEBRA-type batteries, whose cycling is restricted to a content of (NaCl)₂:AlCl₃ > 1:2 (to avoid decomposition of NaAlCl₄ at 0% SOC). The salt contents of a realistic cathode composition with 30% active metal content and 50 vol% NaAlCl₄ in a test cell are marked exemplarily in Figure 2. As AlCl₃ does not participate in the battery charging and discharging reaction, the molar fraction of AlCl₃ is constant in the AlCl₃-ZnCl₂-(NaCl)₂ phase diagram at various SOC (0–100%). As assembled in the discharged state (0% SOC), the cell contains 34.9 mol% AlCl₃, 65.1 mol% (NaCl)₂, and Zn metal powder (not shown here); this corresponds to molten NaAlCl₄ and NaCl, with the latter being present as solid particles. Upon charge, the ZnCl₂ content increases and the NaCl content decreases. At 100% SOC, 34.9 mol% AlCl₃, 47.7 mol% ZnCl₂ and 17.4 mol% (NaCl)₂ is reached, and the salt mixture is liquid completely. From this phase diagram it is safe to conclude that sufficient AlCl₃ is needed to be present in the secondary salt electrolyte to avoid the high melting point of the salt mixture, since the high resistance of the secondary salt electrolyte due to large vol% of solid salts leads to bad battery performance. To display the influence of temperature, we present binary phase diagrams at different AlCl₃ contents (0–55 mol%) below.

The binary phase diagram of ZnCl₂-(NaCl)₂ (Figure 3a), i.e., absence of AlCl₃ (0 mol%) in the AlCl₃-ZnCl₂-(NaCl)₂ mixture, displays a eutectic point with a melting temperature of 254 °C at about 81 mol% ZnCl₂, which is in agreement with the literature [11]. As shown in Figure 3b–f, pseudo-binary phase diagrams were simulated with a stepwise increase in 10 mol% of AlCl₃ from 15 to 55 mol % in the salt mixture of AlCl₃-ZnCl₂-(NaCl)₂, corresponding to Na-ZnCl₂ cells with variable NaAlCl₄ contents between ~20 and 80 vol.% at the cathode. When the molar fraction of AlCl₃ increases, the salt mixture can be kept in the fully liquid state (i.e., very low resistance for better battery operation) over a wider compositional and SOC range. For instance, at 300 °C and for 15 mol% AlCl₃, a minimum 52 mol% of ZnCl₂ is required to keep the melt in liquid state as shown in Figure 3b, while for 55 mol% AlCl₃ only a minimum of 8 mol% ZnCl₂ is required (see Figure 3f). Thus, the AlCl₃ content (i.e., NaAlCl₄) at the cathode significantly influences the phase content of active Zn-species (i.e., ZnCl₂) at the cathode, and thus the reaction mechanism and kinetics of corresponding cells.

However, with the increasing content of AlCl₃, the cost of the secondary salt electrolyte (i.e., the battery) could increase, since AlCl₃ (about 0.9 USD/kg) has much higher price (more than ten times) than NaCl (about 0.06 USD/kg) [22]. Moreover, it is also evident in Figure 3a–f that the gas phase starts to form at lower temperatures with the increasing molar fraction of AlCl₃. It is defined in FactSage^TM that the gas phase starts to form when the vapor pressure of the salt mixture reaches 1 atm. For example, at 0 mol% AlCl₃, the gas pressure is above 1 atm at approximately 750 °C, i.e., much gas formation starts at this temperature. With the increase in the concentration of AlCl₃, the gas phase starts to form at lower temperatures, which is about 300 °C at 55 mol% of AlCl₃ in the salt mixture. From Figure 3, it can be stated that increasing molar fraction of AlCl₃ has both positive and negative effect on the salt electrolyte, on one hand it can decrease the melting point and could reduce the resistance with more existing liquid phase, but on other hand it can increase the vapor pressure and could affect the battery life span and safety. Moreover, increasing AlCl₃ can lead to more challenges such as materials corrosion [23]. Thus, it is suggested in future work to determine the optimal concentration of AlCl₃ in the secondary salt electrolyte regarding the performance and storage cost of the battery, based on phase diagrams, vapor pressure, charging/discharging range, compatibility with other materials in the battery, etc.

3.2. Differential Scanning Calorimetry (DSC)

To further check the simulated phase diagrams from FactSage^TM, DSC is used to determine the phase change temperatures at different salt compositions in Table 1. Some representative DSC curves are shown in Figure 4, Figure A2, Figure A3, Figure A4, Figure A5 and Figure A6. Performing three heating-cooling cycles, the second and third cycles of all the DSC curves were found to be almost identical with ±1.5 °C difference. The DSC curves for Salt Sample No. 1 in Table 1 is shown in Figure 4b. The red curves in the figures are showing the second cycle and the black curves represent the third cycle of DSC. The onset temperatures of heat adsorption peaks are recorded to be the starting temperatures of the phase change (i.e., melting temperatures). The DSC results of all the salt compositions in Table 1 are summarized in

Table 2. Table comparing the FactSage^TM and DSC results for 11 salt samples.

Sample No.	AlCl3	ZnCl2	(NaCl)2	FactsageTM	DSC	OptiMelt
	mol%	mol%	mol%	TTransition (°C)	TOnset (°C)	TStart (°C)
1	15	77.5	7.5	155 (Solidus), melting of NaAlCl4	152	about 150
				239 (Transition), melting of Na2ZnCl4	225	about 230
				299 (Transition), melting of ZnCl2	303	-
				395 (liquidus), mixing of NaAlCl4 and Na2ZnCl4	349	-
2	15	55	30	154 (Solidus)	152	about 150
				239 (Transition)	216	Continue melting
				290 (liquidus)	362
3	25	62.5	12.5	155 (Solidus)	151	about 150
				200 (Transition)	200	about 230
				260 (Transition)	-	Cont. melting
				300 (liquidus)	303
4	25	30	45	154 (Solidus)	153	about 152
				379 (liquidus)	360	Cont. melting
5	35	47.5	17.5	155 (Solidus)	152	about 150
				200–260 (Transition)	204	about 230
				300 (liquidus)	303.7	Cont. melting
6	35	20	45	154 (Solidus)	153	about 152
				379 (liquidus)	341.5	Cont. melting
7 (SOC 0%)	34.89	0	65.11	155 (Solidus)	152.6	Not measured
8 (SOC 50%)	34.89	23.83	41.28	155 (Solidus)	150.7
				230 (Transition)	204.5
				345 (liquidus)	349.5
9 (SOC 76.5%)	34.89	36.47	28.65	155 (Solidus)	145.3
				230–245 (Transition)	232.5
				345 (liquidus)	349.5
10 (SOC 87.6%)	34.89	41.76	23.35	155 (Solidus)	145.8
				235–260 (Transition)	230.2
				295 (liquidus)	290.1
11 (SOC 100%)	34.89	47.67	17.44	155 (Solidus)	152
				200 (Transition)	204
				300 (liquidus)	303.7

for comparison. The DSC results are presented after taking the average of the 2nd and 3rd cycle.

Moreover, the simulated results on the melting process obtained by FactSage^TM for sample No. 1 and 2 are shown in Figure 4a. FactSage^TM simulation shows that for high ZnCl₂ concentration (sample No. 1: 15.0-77.5-7.5 mol% AlCl₃-ZnCl₂-(NaCl)₂), the first compound melting is NaAlCl₄ at 155 °C followed by the melting of Na₂ZnCl₄ at 239 °C, ZnCl₂ at 299 °C, and finally mix of the two salt liquids (NaAlCl₄ and Na₂ZnCl₄) at 395 °C. The DSC diagram in Figure 4b (sample No. 1) shows four main peaks with the onset temperatures of 152, 223, 304 and 348.9 °C (FactSage^TM overestimates the last temperature) in the heating curve, which is in good agreement with the simulation.

Similarly, in Figure 4a for lower ZnCl₂ concentration (sample No. 2: 15-55-30 mol% AlCl₃-ZnCl₂-(NaCl)₂), the first compound melting is NaAlCl₄ at 154 °C followed by the melting of Na₂ZnCl₄ at around 239 °C and the mixing of liquid NaAlCl₄ and Na₂ZnCl₄ at around 290 °C. Figure 4c shows three main peaks with the onset temperatures of 152, 216 and 362 °C (FactSage^TM underestimates the last temperature) in the heating curve. Unlike in the case of the Na-NiCl₂ ZEBRA cell, the presence of multiple phases (some salts such as Na₂ZnCl₄ with melting temperatures <300 °C) creates different reaction products and leads to different reaction mechanism during charging and discharging cycles at 300 °C, as also shown by the earlier studies [9,10,11,12,13,14,15]. In Table 2, the results of OptiMelt^TM for samples No. 1–6 are summarized and compared with the DSC results. For all the salt compositions, the DSC results are in good agreement with the thermodynamic simulations. The deviation of the melting temperatures of the main species such as NaAlCl₄ and Na₂ZnCl₄ is smaller than 10 °C.

3.3. Comparison of DSC and OptiMelt^TM

In order to validate the results obtained from DSC, OptiMelt^TM analysis was conducted at three different AlCl₃ concentrations, i.e., samples No. 1–6. As OptiMelt^TM is a visual analysis technique (not as accurate as DSC) and samples No. 7–11 exhibit compositions similar to samples No. 5 and 6 (all contain about 35 mol% AlCl₃), the OptiMelt^TM analysis was performed on samples No. 1–6 only in this work. The results of the thermodynamic simulation and DSC show that the melting process of NaAlCl₄ starts at approx. 154 °C at different AlCl₃ mol% in the tested salt mixtures. Figure 5a,b show the solid salt mixtures below first transition temperature (154 °C). As soon as the temperature reached 154 °C, the salt mixture started to melt (see Figure 5c,d). Moreover, it can also be noticed at 160 °C (Figure 5c) that the amount of the solid salts decreased with the increase in AlCl₃ mol%. This can be explained with a larger amount of NaAlCl₄ formed when more AlCl₃ is present in the salt mixture. NaAlCl₄ has a low melting point of 150 °C as also confirmed by the phase diagram (see Figure 3). It can also be seen that samples No. 1, 3 and 5 in Table 1 have a large amount of ZnCl₂ present in the mixture and hence the solid salts are also in lower quantity at 300 °C (OptiMelt^TM images are not shown here) because of the lower melting point at a higher ZnCl₂ concentration (melting of Na₂ZnCl₄ at about 240 °C), similar to what we observe in the pseudo-binary phase diagrams in Figure 3. Overall, as summarized in Table 2, although the results of OptiMelt^TM experiments are not as accurate as DSC due to visual inspection, the results are comparable to those of DSC, and can give valuable information on, e.g., the amount of the remaining solid salt particles.

3.4. Vapor Pressure Simulation

First, to know how accurately the thermodynamic simulation of FactSage^TM can predict the vapor pressure of the AlCl₃-NaCl-ZnCl₂ system, the vapor pressures of AlCl₃-NaCl (58–42 mol%) mixture and pure AlCl₃ from the literature [19] were compared with the vapor pressures simulated with FactSage^TM (see Figure A1 in the Appendix A). The thermodynamic simulation shows a good correlation, with only a marginal difference of ±1.5% error for pure AlCl₃. Moreover, the vapor pressures of AlCl₃-NaCl and pure AlCl₃ increase significantly with increasing temperature, e.g., from 0.06 atm at 200 °C to about 0.2 atm at 250 °C for AlCl₃-NaCl.

For the AlCl₃-ZnCl₂-NaCl salt mixture, increasing AlCl₃ and ZnCl₂ content, and the temperature can significantly increase the vapor pressure of the salt mixture [24]. For safe operation of the Na-ZnCl₂ battery, a maximum allowed vapor pressure of 1 atm is assumed in this work for a preliminary study with the simulation. To estimate the maximum amount of AlCl₃ that can be used inside the battery at different temperatures, a worst-case scenario of overcharging is considered inside the cathode such that NaCl will be fully consumed (i.e., only AlCl₃ and ZnCl₂ will remain in the salt mixture). Figure 6 illustrates that increasing the AlCl₃ concentration in the electrolyte leads to a substantial increase in vapor pressure. This can be explained with the lower melting temperature and thus faster vapor pressure increase in AlCl₃-containing species such as NaAlCl₄ than other species (See Table 2 and Figure A1) [22]. It is also important to note that for 35 mol% AlCl₃ at a SOC of 137%, the salt vapor pressure exceeds the 1 atm limit at >500 °C. Moreover, when the secondary salt electrolyte contains 40 mol% AlCl₃, the simulated salt vapor pressure at 400 °C is about 0.9 atm, not exceeding 1 atm. This means that if only considering the salt vapor pressure, the battery with <40 mol% AlCl₃ in the salt electrolyte could safely work below 400 °C. Please note that due to strong hygroscopicity of the NaAlCl₄ salt, some hydrated water could exist in the secondary salt electrolyte. The hydrated water can lead to a higher vapor pressure than the simulated salt vapor pressure in Figure 6, and also corrosion problem. Thus, it is strongly suggested to consider the effect of the salt vapor pressure and presence of water in the salt on the battery safety and lifetime.

4. Conclusions

In this work, thermodynamic simulation of the phase diagrams with FactSage^TM and thermal analysis methods such as differential scanning calorimetry (DSC) and OptiMelt^TM were used to in-depth investigate the AlCl₃-NaCl-ZnCl₂ secondary salt electrolyte in the Na-ZnCl₂ solid electrolyte battery. The main results are summurized below:

• The simulation results (e.g., phase diagrams and vapor pressures of salt mixtures) via FactSage^TM are in good agreement with the experimental results from DSC and OptiMelt^TM in this work and from the literature. The deviation of the melting temperatures of the main species such as NaAlCl₄ and Na₂ZnCl₄ between simulations and experiments is smaller than 10 °C.
• From the phase diagrams it is concluded that AlCl₃ is needed to be present in the secondary salt electrolyte to ensure enough liquid salt existing during charging/discharging for low resistance of the battery, since the high resistance due to the secondary salt electrolyte leads to bad battery performance.
• With increasing concentration of ZnCl₂ at higher SOC, the melting temperature of the secondary salt electrolyte decreases significantly, since Na₂ZnCl₄ (about 240 °C) has much lower melting temperature than NiCl₂. As a result of replacement of Ni with Zn, active materials (ZnCl₂ and NaCl) should be present in the liquid state over a wider SOC range during battery cell cycling compared to classical Na-NiCl₂ ZEBRA cells.
• Increasing molar fraction of AlCl₃ has both positive and negative effect on the salt electrolyte. On one hand it can decrease the melting point and increase the ionic conductivity at lower temperatures, which could improve the battery performance, but on other hand it can increase the vapor pressure and could affect the battery costs, life span and safety (e.g., mechanical stability of the BASE electrolyte, cell pressure management). Moreover, increasing AlCl₃ can lead to more challenges such as corrosion of sealing.
• It is suggested in future work to determine the optimal concentration of AlCl₃ in the secondary salt electrolyte regarding the performance and storage cost of the battery, based on phase diagrams, vapor pressure, charging/discharging range, compatibility with other materials in the battery, etc.
• The salt vapor pressures vs. temperature and mol% AlCl₃ in AlCl₃-ZnCl₂-NaCl were simulated. Considering the rapid increase in salt vapor pressure at high temperatures and AlCl₃ content, it is strongly recommended to consider the impact of salt vapor pressure on battery safety and pressure management. The maximum allowable temperature is suggested to 400 °C for safe operation of the cell with the secondary salt electrolyte (AlCl₃ < 40 mol%).
• The reliable simulation results in this study, including phase diagrams and vapor pressures, show promise in assisting battery cell design and enhancement. Tests on a Na-ZnCl₂ test cell, based on the suggested electrolyte composition, reveal a three-step charging and discharging behavior consistent with our predicted phase diagram. Further insights into the Na-ZnCl₂ battery system’s performance will be published in an upcoming paper.