In situ conversion reaction of magnesium fluoride to boost the performance of the sulfide-based electrolyte Li₆PS₅Cl for all-solid-state lithium metal batteries

Abstract

Among the solid electrolytes of all-solid-state lithium metal batteries being pursued globally, Li₆PS₅Cl is one of the most promising candidates owing to its high ionic conductivity and easy processibility. However, Li₆PS₅Cl is vulnerable to the lithium anode because lithium can not only reduce Li₆PS₅Cl to generate passive interfaces but can also lead to the growth of lithium dendrites, which could penetrate the Li₆PS₅Cl bulk and eventually short-circuit the battery. Herein, we report that the electrochemical performance of Li₆PS₅Cl could be greatly enhanced by compositing it with MgF₂, which was the most effective metal fluoride among five studied materials. Specifically, critical current density was increased by 4.7 times, cycling durability in Li|electrolyte|Li symmetric cells was extended by 19 times, capacity retention in Li|electrolyte|LiNi_0.7Co_0.2Mn_0.1O₂ full cells was enhanced from 76% to 86%, and rate capability was boosted from 0.2C to 1C. Combination studies involving experimental characterizations and theoretical computations revealed that the performance-improving mechanism involved a sustained-release effect of capsule medicines, meaning during the charging/discharging cycles, MgF₂ could timely scavenge lithium dendrites to generate Li_xMg alloy and LiF, wherein Li_xMg could reversibly release/uptake Li and LiF could suppress the nucleation of lithium dendrites.

---

1. Introduction

With the increasing popularity of portable electronic devices and new-energy vehicles, the demand for advanced batteries to sustain growth is also increasing. Compared to traditional liquid-state lithium-ion batteries, all-solid-state batteries have received widespread attention for their higher safety and higher energy characteristics.^1–4 Solid electrolytes (SEs) can be divided into inorganic solid electrolytes (ISEs), solid polymer electrolytes (PEs) and composite electrolytes (CEs).^5–7 Sulfide-based solid electrolytes (SSEs) in the ISE category are advantageous in terms of their ionic conductivity, mechanical ductility, and interface contact.^8–11 In the big family of SSEs, Li₆PS₅Cl is particularly attractive, owing to its overall superiority in terms of high ionic conductivity, elemental abundance, and (electro)chemical stability.⁹ However, Li₆PS₅Cl still has some issues that need to be addressed. Like many other SSEs, because of the narrow electrochemical stability window, Li₆PS₅Cl in contact with lithium metal will be reduced to form a passive interface with high resistance.^12,13 During the charging/discharging processes, uneven lithium deposition/dissolution can generate lithium dendrites at the interface of and inside Li₆PS₅Cl, eventually causing a short-circuit of the battery. Consequently, Li₆PS₅Cl cannot be directly paired with lithium metal, run at high current densities, or cycled for a long duration. The above flaws impede the pace of utilizing Li₆PS₅Cl for practical applications but have stimulated researchers to explore mitigating strategies, mainly involving constructing interlayers and modifying Li₆PS₅Cl.

As widely used for oxide-based solid electrolytes,^14–16 the interlayer-construction strategy involves building heterogeneous interfaces between lithium and SSEs, which can be ion conductive, electron conductive, or mixed ion/electron conductive.^17–19 (1) Ion conductive interface layer: Yao et al. utilized the spontaneous reaction between pentafluorobenzamide and lithium to in situ produce a uniform LiF–Li₃N protective layer between SSEs and lithium metal.¹⁷ Such a composite layer could suppress side reactions of Li₆PS₅Cl, inhibit lithium dendrite growth, and facilitate the uniform deposition of lithium. The CCDs of Li₆PS₅Cl in lithium symmetric cells were increased by 2–3 times under the protection of an LiF–Li₃N layer. The corresponding Li|electrolyte|LiCoO₂ full cells exhibited remarkable cyclability for 500 cycles at 1C with a capacity retention ratio of 89%. (2) Electron conductive interface layer: Wang at al. designed a Mg₁₆Bi₈₄ interlayer at the Li/Li₆PS₅Cl interface to suppress the growth of lithium dendrites and a F-rich interlayer on the NCM811 cathode to reduce the interfacial resistance.¹⁸ The Mg₁₆Bi₈₄ interlayer increased the CCD from 0.4 mA cm⁻² to 2.6 mA cm⁻² and enabled the Li/Mg₁₆Bi₈₄–Li₆PS₅Cl–Mg₁₆Bi₈₄/Li cell to charge/discharge stably at 1.2 mA cm⁻² for over 2700 h. (3) Mixed ion/electron conductive interface layer: Wu et al. prepared a soft carbon (SC)–Li₃N interface layer between Li₆PS₅Cl and a lithium anode, in which the in situ lithiation reaction not only converted SC into LiC₆ with good electronic/ionic conductivity but also transformed the mixed-phase of α-Li₃N-and-β-Li₃N into β-Li₃N with a high ionic conductivity and diffusion rate of Li⁺-ions.¹⁹ The LiZrO₂@LiCoO₂|Li₆PS₅Cl/SC-Li₃N|Li full cell demonstrated a capacity retention ratio of >80% after 3000 cycles at a high current density (15C, 7.5 mA cm⁻²).

The second strategy includes the following three subgroups. (1) Elemental doping:^20,21 Liu et al. doped Li₆PS₅Cl with both Mg and F elements (Li₆PS₅Cl-MF) by adding MgF₂ into the precursors of Li₆PS₅Cl before the calcination process.²⁰ The in situ-generated LiF-rich interlayer could regulate the electron density distribution inside Li₆PS₅Cl to suppress lithium dendrite growth during the lithium plating/stripping processes. The CCD of the Li|Li₆PS₅Cl-MF|Li symmetrical cell reached 1.4 mA cm⁻², which was 233% higher than that of the pristine Li₆PS₅Cl. (2) Surface coating:^22,23 Sang et al. constructed a LiF@Li₂O shell on the surface of Li₆PS₅Cl, observing a “one stone, three birds” effect by realizing improved moisture tolerance, interface compatibility with the lithium metal, and stability against LiCoO₂.²² (3) Additive compositing:^24,25 Coskun et al. mixed Li₃TCA (prelithiated trithiocyanuric acid) with Li₆PS₅Cl in a mortar to obtain the Li₆PS₅Cl–Li₃TCA (2.5 wt%) electrolyte, which could establish a stable interface with lithium to achieve a long cycling-life.²⁴ The Li₃TCA additive enabled the Li₆PS₅Cl electrolyte to increase its CCD from 1.0 mA cm⁻² to 1.9 mA cm⁻² and to extend its cycling-life at 1.0 mA cm⁻² in lithium symmetric cells from 30 h to 750 h. Meanwhile, Zhang et al. ball-milled LiF together with Li₆PS₅Cl in a mass ratio of 11 : 41, and reported that the obtained composite electrolyte could effectively inhibit the growth of dendrites.²⁵ Also, the Li|Li₆PS₅Cl–LiF|Li symmetric cell could achieve an ultra-long cycling-life of 1800 h at 0.5 mA cm⁻², much higher than the cycling-life of 120 h for the Li|Li₆PS₅Cl|Li symmetric cell. The authors also designed a transparent all-solid-state Li|LPSC|LPSC-PTFE|LPSC-LiF|Cu cell encapsulated in a quartz mold, and used an optical microscope to observe the ability of LiF to inhibit the dendrites growth.

While the two types of mitigating strategies described above have greatly improved the performance of Li₆PS₅Cl, as shown above, each has some weaknesses. For the interlayer-engineering strategy, it is challenging to handle lithium anodes, while the functional surface area is limited due to its 2D feature, and the dendrites–suppressing interface is thin. Regarding the Li₆PS₅Cl-modification strategy, elemental doping needs high temperatures and is hard to control accurately, and it is difficult to apply multiple components in the additive-compositing processes. Combining the advantages of both strategies would be another interesting strategy. Consequently, we herein composited bifunctional materials of metal fluorides with Li₆PS₅Cl via a simple manual-grinding process and then focus our further studies on the best system MgF₂ among the five metal fluorides. Once lithium encounters MgF₂, the conversion reactions between them in situ produce LiF and Li–Mg alloy (eqn (1) and (2)).

MgF₂ + 2Li → 2LiF + Mg (1)

xLi + Mg ↔ Li_xMg (2)

Because MgF₂ particles were uniformly distributed at the macroscopic scale inside the entire matrix of Li₆PS₅Cl, the generation of LiF and Li–Mg occurred progressively during the charging/discharging cycles, analogous to the sustained-release effect of capsular medicines. Thus, under the synergistic effects of LiF to facilitate the uniform and fast dispersion of lithium ions, as well as to prevent breakdown of the electron-conducting network, and of Li–Mg alloys to scavenge the invading lithium dendrites, the bifunctional MgF₂ greatly enhanced the battery performance of Li₆PS₅Cl. The CCD of Li₆PS₅Cl–MgF₂(7 wt%) was 4.7 times that of Li₆PS₅Cl (i.e., 1.4 mA cm⁻²vs. 0.3 mA cm⁻²). Further, in Li symmetrical cells running at 30 °C and 0.2 mA cm⁻², the cycling lifetime of Li₆PS₅Cl was increased from 366 h to >7000 h, representing an over 19 times improvement. As for Li|electrolyte|NCM721 full cells, the capacity retention ratio was increased from 76% to 86% and the rate capability boosted from 0.2C to 1C.

2. Results and discussion

2.1 Screening metal fluorides

The conceptual design of the Li₆PS₅Cl-MF_x composites for achieving performance enhancements is illustrated in Fig. 1. In this design, in the vicinity of the lithium metal (gray), in situ conversion reactions occur between the metal fluorides (green) and lithium to produce LiF and lithium metal alloys Li–M (red), which together regulate the uniform deposition of lithium. Consequently, the synergistic effect of MF_x can inhibit the formation of lithium dendrites at the interface and “swallow” the invading lithium inside Li₆PS₅Cl, thereby avoiding perforation of the solid electrolyte.

Fig. 1 Conceptual illustration of the Li/Li₆PS₅Cl interface (A) and Li/Li₆PS₅Cl-MF_x interface (B) during the charge/discharge cycling, M = Mg, Zn, Al, Sn, and Ag.

Next, a proof-of-concept study was executed to select the best system for comprehensive investigation. It has been reported that alloy anodes with high solid solubility have low nucleation overpotentials, which would be conducive to achieving uniform lithium deposition and then inhibiting lithium dendrite growth.^26,27 With that in mind, alloys of Li–Ag, Li–Mg, Li–Zn, Li–Al, and Li–Sn were targeted in our study. Therefore, metal fluorides (MgF₂, ZnF₂, AlF₃, and SnF₂) powders were manually ground with Li₆PS₅Cl in an agate mortar to obtain composite electrolytes. Because the AgF powder itself decomposed when being ball-milled, changing from reddish brown to black, AgF was not further studied. The other electrolytes were assessed by measuring their CCDs in lithium symmetric cells. The CCD value increased from 0.3 mA cm⁻² for Li₆PS₅Cl to 1.4 mA cm⁻² for Li₆PS₅Cl–MgF₂(7 wt%), as shown in Fig. 2A. The CCD values of Li₆PS₅Cl–ZnF₂ and Li₆PS₅Cl–AlF₃ were 0.4 mA cm⁻² and 0.7 mA cm⁻², respectively (Fig. 2B and C). When grinding SnF₂ with Li₆PS₅Cl, the electrolyte changed from grayish-white to yellow, implying the generation of SnS₂ from some side reaction between SnF₂ and Li₆PS₅Cl (Fig. S1†). This fact may explain why the CCD value of Li₆PS₅Cl–SnF₂ did not increase (Fig. 2D). Consequently, MgF₂ was considered the best material among the four tested metal fluorides, probably due to the highest solubility of Mg in Li (68 at% at 100 °C).²⁸ In order to gain a deep understanding of the alloying effect, DFT simulations were performed to examine the charge density difference induced by a Li atom on the LiM_x surfaces (Fig. 2E). The localized charge accumulation (cyan) and depletion (yellow) on the additional Li atom suggest an alloying effect between Li and M. Based on quantitative calculation results, the dashed lines in Fig. S2† represent the contact position of the Li–M, while the most remarkable charge density difference for Li–Mg suggests the solubility of Li in Mg was the highest among the four cases studied; that is, since Li–Mg had the most delocalized charge density difference, the solubility of Li in Mg must be the highest among four cases. Furthermore, the calculated energy barriers for Li⁺ diffusion (Fig. 2F) revealed that the Li–Mg(100) had the lowest energy barrier of 0.066 eV, compared to 0.16 eV for Li–Al₃(100) and 0.20 eV for Li–Zn(100). This indicates that the Li⁺ diffusion rate in the Li–Mg alloy was the highest, which was consistent with the CCD results.

Fig. 2 CCD curves of Li|Li₆PS₅Cl-MF₂|Li cells at 30 °C, where MF₂ is (A) MgF₂, (B) ZnF₂, (C) AlF₃ or (D) SnF₂. (E) Charge density differences of Li–LiMg(100), Li–LiAl₃(100) and Li–LiZn(100), in which the green, orange, blue, and gray balls represent Li, Mg, Al, and Zn atoms, respectively. (F) Energy barriers for Li⁺ diffusion on LiMg(100), LiAl₃(100), and LiZn(100).

2.2 Assessing the electrolyte performance

After determining MgF₂ offered the best potential as the preferred metal fluoride, Li₆PS₅Cl–MgF₂ electrolytes were made via manual grinding and were then comprehensively evaluated. Fig. 3A illustrates the X-ray diffraction (XRD) results for MgF₂, MgF₂ after ball-milling (marked as MgF₂(B)), Li₆PS₅Cl, and Li₆PS₅Cl–MgF₂. There was no significant difference in the XRD patterns of MgF₂ before and after ball-milling treatment, indicating that the MgF₂ particles maintained good crystallinity. The optical images (Fig. S3†) showed that the MgF₂ particles after ball-milling were uniform. The XRD patterns of Li₆PS₅Cl were consistent with the XRD standard card (PDF #97-041-8490). From the electrochemical impedance spectroscopy (EIS) plots for the Li₆PS₅Cl and Li₆PS₅Cl–MgF₂ (4, 7, and 10 wt%) electrolytes (Fig. 3B), the ionic conductivities of the four electrolytes were determined to be 11.0, 4.6, 4.0, and 3.7 mS cm⁻¹, respectively. The electronic conductivity for Li₆PS₅Cl–MgF₂ (7 wt%) was two orders of magnitude smaller than that of Li₆PS₅Cl, that is 8.2 × 10⁻⁸ mS cm⁻¹vs. 1.6 × 10⁻⁶ mS cm⁻¹ (Fig. S4†), which would be conducive to suppressing the growth of lithium dendrites.²⁹ Next, CCD measurements in symmetric cells of Li|Li₆PS₅Cl–MgF₂|Li at 30 °C were used to determine the optimal mass fraction of MgF₂, as shown in Fig. 3C. At mass fractions of MgF₂ were 4, 7, and 10 wt%, the CCD values of the Li₆PS₅Cl–MgF₂ electrolytes were 0.8, 1.4, and 1.2 mA cm⁻², respectively, all much higher than that of the pristine Li₆PS₅Cl (0.3 mA cm⁻²). The best case was Li₆PS₅Cl–MgF₂ (7 wt%), which was 4.7 times higher. For comparison, we also prepared Li₆PS₅Cl–MgF₂ (7 wt%) via ball-milling, but the CCD value was only 0.5 mA cm⁻² (Fig. S5†), probably because of mechanical damage to the Li₆PS₅Cl lattices. In addition, two Li₆PS₅Cl–MgF₂ composite electrolytes with layered structures were further studied, i.e., with modifying MgF₂ only on the Li₆PS₅Cl surface (a sandwich structure of MgF₂/Li₆PS₅Cl/MgF₂) and with inserting a layer of MgF₂ inside Li₆PS₅Cl pellets (a sandwich structure of Li₆PS₅Cl/MgF₂/Li₆PS₅Cl). As shown in Fig. S6,† the CCD values for the as-obtained MgF₂/Li₆PS₅Cl/MgF₂ and Li₆PS₅Cl/MgF₂/Li₆PS₅Cl were 0.2 and 0.6 mA cm⁻², respectively. The inferiority of the former to pure Li₆PS₅Cl (0.3 mA cm⁻²) implies that the conversion reaction between Li and MgF₂ may have been too intensive to construct good Li/Li₆PS₅Cl interfaces. In contrast, the superiority of the latter indicates the positive role of loading MgF₂ inside the Li₆PS₅Cl membrane, which could otherwise be more easily penetrated by lithium dendrites. The galvanostatic cycling profiles of Li symmetric cells at 0.2 mA cm⁻² and 30 °C are displayed in Fig. 3D. The Li|Li₆PS₅Cl|Li cell got short-circuited after stably cycling for 366 h at an overpotential of approximately 7 mV. In stark contrast, the Li|Li₆PS₅Cl–MgF₂(7 wt%)|Li cell was able to cycle stably for >7000 h with a stable overpotential of 10 mV. The generated LiF and Li–Mg alloys at the Li/electrolyte interface could synergistically regulate the uniform deposition of lithium. Also, in this design, if some lithium dendrites happened to pierce into the electrolyte bulk, they would be continuously “swallowed” by MgF₂, therein presenting excellent stability to the lithium anode.³⁰ With the charge/discharge cycling, the composited MgF₂ particles function gradually via conversion reactions, like the sustained-release effect of capsular medicines. The excellent performance of Li₆PS₅Cl–MgF₂(7 wt%) reported here is comparable with some recently published works (Fig. S7†).

Fig. 3 (A) XRD patterns of MgF₂, MgF₂ after ball-milling (marked as MgF₂(B)), Li₆PS₅Cl and Li₆PS₅Cl–MgF₂(7 wt%). (B) EIS plots of Li₆PS₅Cl and Li₆PS₅Cl–MgF₂(4 wt%, 7 wt% and 10 wt%) electrolytes. (C) CCD curves of Li|Li₆PS₅Cl–MgF₂ (4 wt%, 7 wt% and 10 wt%)|Li cells at 30 °C. (D) Galvanostatic cycling of Li|Li₆PS₅Cl|Li and Li|Li₆PS₅Cl–MgF₂ (7 wt%)|Li cells at 0.2 mA cm⁻² and 30 °C; insets show the SEM and EDS analyses on the lithium anodes in two parallel cells after cycling for 100 h, i.e., (i–ii) SEM image and the corresponding EDS mapping of the O element for the Li|Li₆PS₅Cl|Li cell; (iii–iv) SEM image and the corresponding EDS mapping of the O element for the Li|Li₆PS₅Cl–MgF₂|Li cell.

Two additional cells in parallel to those in Fig. 3D were disassembled after cycling for 100 h to observe their lithium anodes by scanning electron microscopy (SEM) and energy dispersive spectroscopy (EDS), as shown in the insets therein. For the Li|Li₆PS₅Cl|Li cell, the lithium anode under SEM (Fig. 3Di) exhibited some moss-like black blobs. Since the lithium element was not detectable by EDS, the oxygen element was chosen for the EDS mapping (Fig. 3Dii), wherein some signal-intensive areas coincided well with the profiles of the black blobs in (i). Such results imply that those black blobs were originally lithium dendrites, which were instantly oxidized into hydroxides/oxides due to their extreme reactivity to water/oxygen during transferring the sample in air. In contrast, for the Li|Li₆PS₅Cl–MgF₂|Li cell, the lithium anode under SEM observation did not show lithium dendrites (Fig. 3Diii) and the EDS mapping of oxygen did not have concentrated regions (Fig. 3Div). Therefore, the difference in morphologies in these two cases confirmed that MgF₂ could, as expected, effectively regulate lithium plating/stripping and inhibit the growth of lithium dendrites.

Subsequently, Li₆PS₅Cl–MgF₂ (7 wt%) and Li₆PS₅Cl were further evaluated in Li|electrolyte|NCM721 full cells. The charge/discharge curves for Li₆PS₅Cl and Li₆PS₅Cl–MgF₂ (7 wt%) in the first cycle are shown in Fig. 4A. The discharge capacities in the first cycle for Li₆PS₅Cl and Li₆PS₅Cl–MgF₂ (7 wt%) were 140 and 159 mA h g⁻¹, respectively. As shown in Fig. 4B, the discharge capacities of Li|Li₆PS₅Cl|NCM721 and Li|Li₆PS₅Cl–MgF₂(7 wt%)|NCM721 were 107 and 137 mA h g⁻¹ after 100 cycles, corresponding to capacity retention ratios of 76% and 86%. The rate performances of Li|electrolyte|NCM721 full cells were measured, as shown in Fig. 4C. For the pristine Li₆PS₅Cl, the discharge capacity of the full cell was 146 mA h g⁻¹ at 0.1C and 123 mA h g⁻¹ at 0.2C, then the full cell got short-circuited at 0.5C. Instead, the discharge capacities of Li|Li₆PS₅Cl–MgF₂ (7 wt%)|NCM721 full cell at 0.1C, 0.2C, 0.5C, 1C were 161, 139, 114, and 94 mA h g⁻¹, respectively. When the current density returned to 0.1C, the discharge capacity recovered to 156 mA h g⁻¹. In addition, we also conducted a full cell performance test at 60 °C, as shown in Fig. S8.† The capacity retention ratio of the Li|electrolyte|NCM721 full cell increased from 77% to 85% and the rate capability was improved from 0.5C to 2C. Thus, a greatly enhanced electrochemical performance of Li₆PS₅Cl–MgF₂(7 wt%) in the full cell was demonstrated that could also be assigned to the in situ reaction between MgF₂ and lithium,^31–33 which could significantly reduce the interfacial side reactions between Li₆PS₅Cl and lithium, thereby improving the cycling stability and rate performance.

Fig. 4 (A) Charge/discharge curves of Li|electrolyte|NCM721 full cells for the first cycle at 0.2C and 30 °C, (B) cycling stability of Li|electrolyte|NCM721 full cells at 0.2C and 30 °C, and (C) rate performance of Li|electrolyte|NCM721 full cells at 30 °C.

2.3 Morphology characterizations of the electrolytes

The microstructures of the electrolytes were next investigated by SEM and EDS to observe the composite form of Li₆PS₅Cl and MgF₂. As displayed in Fig. 5A, the Li₆PS₅Cl electrolyte particles had a diameter of approximately 10 μm. Fig. 5B and C show the morphology of the Li₆PS₅Cl–MgF₂ (7 wt%) electrolyte. Compared with the pristine electrolytes, the Li₆PS₅Cl–MgF₂ (7 wt%) electrolyte displayed numerous small particles resembling MgF₂ with diameters of 1–2 μm, in addition to Li₆PS₅Cl. As shown in Fig. 5D–F, the EDS mappings confirmed that the elements P, S, and Cl were present and evenly distributed in Li₆PS₅Cl–MgF₂ (7 wt%). As indicated by the green ovals in Fig. 5G–H, the elements Mg and F exhibited localized point-like enrichment, corresponding to small particles marked by the green ovals in Fig. 5C. The SEM and EDS results indicated that the surface of Li₆PS₅Cl was decorated with MgF₂ small particles.

Fig. 5 SEM images of (A) Li₆PS₅Cl, and (B–C) Li₆PS₅Cl–MgF₂ (7 wt%). (D–H) EDS mappings of the elements P, S, Cl, Mg and F for Li₆PS₅Cl–MgF₂ (7 wt%).

Furthermore, TEM was utilized to investigate the electrolytes microstructures. As shown in Fig. 6A and B, consistent with the SEM results, the TEM images showed the presence of Li₆PS₅Cl particles with a diameter of about 10 μm, and the modification of the surface of the Li₆PS₅Cl–MgF₂(7 wt%) electrolyte by MgF₂, marked by green rectangles. As presented in Fig. 6C–G, the TEM image and corresponding EDS mappings of Li₆PS₅Cl–MgF₂ (7 wt%) support the presence of Li₆PS₅Cl decorated with MgF₂.

Fig. 6 TEM images of (A) Li₆PS₅Cl, and (B–C) Li₆PS₅Cl–MgF₂ (7 wt%). (D–H) EDS mappings of the elements P, S, Mg and F for Li₆PS₅Cl–MgF₂ (7 wt%).

2.4 Characterizations of the conversion reaction mechanism

The effectiveness of the in situ conversion reaction of metal fluorides in enhancing the stability of lithium metal was further verified through cyclic voltammetry (CV), EIS, and X-ray photoelectron spectroscopy (XPS). CV was performed on Li|Li₆PS₅Cl|Li₆PS₅Cl + C and Li|Li₆PS₅Cl–MgF₂(7 wt%)|Li₆PS₅Cl–MgF₂(7 wt%) + C semi-blocking cells with C as the cathode to investigate the redox processes of the electrolytes (Fig. 7A and B). In the scanning range of 0.65–5 V, two oxidation peaks were observed at 2.6 and 3.3 V, corresponding to the generation of S-Li₃PS₄–LiCl and S-P₂S₅–LiCl, respectively.³⁴ In the reverse scanning, two reduction peaks were observed at 1.9 and 1.6 V, attributed to the generation of Li₃PS₄–LiCl.³⁵ The Li|Li₆PS₅Cl–MgF₂(7 wt%)|Li₆PS₅Cl–MgF₂ (7 wt%) + C cell had smaller peak currents than that of Li|Li₆PS₅Cl|Li₆PS₅Cl + C cell, indicating that the Li₆PS₅Cl–MgF₂(7 wt%) electrolyte had stronger electrochemical stability and interface compatibility than the Li₆PS₅Cl electrolyte, which was consistent with the electrochemical performance of the full cells previously tested.³⁶ Next, time-resolved EIS measurements were conducted, as shown in Fig. 7C and D. Within the initial 24 h, due to the creep of the lithium anode, the total resistance of the Li|Li₆PS₅Cl|Li cell slightly decreased from 30 Ω (0 h) to 28 Ω (24 h). As the aging progressed, the total resistance of the symmetrical cell increased to 34 Ω (48 h) and 38 Ω (72 h), indicating the continuous deterioration of the interface between the electrolyte and lithium. On the contrary, the initial resistance of Li|Li₆PS₅Cl–MgF₂ (7 wt%)|Li decreased from 48 Ω (0 h) to 46 Ω (24 h) due to lithium creep and then remained at 46 Ω (48 h and 72 h), indicating the self-limiting interface side reactions between the electrolyte and lithium.^37–39 Furthermore, the composite electrolytes in the symmetric cells before and after cycling were characterized by ex situ XPS to explore the reaction mechanism during the cycling process, as shown in Fig. 7E and F. In the pristine state, the peak of F 1s at 686.1 eV matched well with Mg–F. After cycling for 100 h at 0.1 mA cm⁻², three peaks appeared at 685.5, 685.8, and 688.2 eV, corresponding to Li–F, Mg–F, and Mg–P–F, respectively.^40–43 Correspondingly, the peak of Mg 1s changed from the originally single peak at 1304.6 eV to three peaks at 1303.1, 1304.3, and 1306.4 eV after cycling (Fig. S9†), related to Mg/Li–Mg alloy, Mg–F, and Mg–P–F, respectively.⁴⁴

Fig. 7 CV curves of (A) Li|Li₆PS₅Cl|Li₆PS₅Cl + C, (B) Li|Li₆PS₅Cl–MgF₂(7 wt%)|Li₆PS₅Cl–MgF₂(7 wt%) + C semi-blocking cells. EIS plots of (C) Li|Li₆PS₅Cl|Li and (D) Li|Li₆PS₅Cl–MgF₂(7 wt%)|Li after aging for different durations. (E and F) XPS spectra for F 1s of Li₆PS₅Cl–MgF₂(7 wt%) in Li|Li₆PS₅Cl–MgF₂(7 wt%)|Li cell before and after cycling.

3. Conclusion

In summary, this work reports a simple method for improving the electrochemical performance of Li₆PS₅Cl by compositing it with MgF₂via manually grinding their powdery mixture. During the charging/discharging processes, lithium dendrites were inclined to grow at both the lithium/Li₆PS₅Cl interface and inside the Li₆PS₅Cl matrix. Owing to the presence of MgF₂, such lithium dendrites could be converted to LiF and Li_xMg alloy, like being “swallowed” by MgF₂. LiF could facilitate the uniform diffusion of lithium ions and prevent electron breakdown. Li_xMg could empower the alloying/dealloying process due to the high solubility of Li in Mg and then lower the plating/stripping overpotentials of lithium. Consequently, the bifunctional MgF₂ demonstrated a synergistic effect to suppress the harmful growth of lithium dendrites and to enhance the electrochemical performance of Li₆PS₅Cl. Because MgF₂ particles participated in “swallowing” lithium dendrites progressively in terms of the particle numbers and alloying depths, the mechanism is analogous to the sustained-release effect of capsular medicines. With the inclusion of MgF₂ particles, the Li₆PS₅Cl electrolyte displayed great performance enhancements, such as in the critical current density, the cycling-life in Li symmetric cells, and the rate capability in Li|electrolyte|NCM721 full cells. Additionally, the merits of MgF₂ in terms of its low cost and air stability can increase its potential for practical applications in all-solid-state lithium metal batteries. Nevertheless, since the MgF₂-composited Li₆PS₅Cl electrolyte herein is still far from practical applications, other performance-improving strategies need to be explored in the future, such as combining other materials with the MgF₂-coated Li anode, such as trying fluorides of lanthanum-group metals, and adding polymeric electrolytes.

Data availability

Data are available from the authors upon request.

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

We thank the Haihe Laboratory of Sustainable Chemical Transformations (24HHWCSS00009) and the Key Program of the National Natural Science Foundation of China (No. 52432005) for financial support. This work was also partially supported by the Graduate Top-notch Innovation Award Plan in Liberal Arts and Science of Tianjin University for the Year of 2023 (B2-2023-012).

References

1. A. M. Bates, Y. Preger, L. Torres-Castro, K. L. Harrison, S. J. Harris and J. Hewson, Are solid-state batteries safer than lithium-ion batteries, Joule, 2022, 6, 742–755.
2. C. Yu, F. Zhao, J. Luo, L. Zhang and X. Sun, Recent development of lithium argyrodite solid-state electrolytes for solid-state batteries: synthesis, structure, stability and dynamics, Nano Energy, 2021, 83, 105858.
3. S. Xia, X. Wu, Z. Zhang, Y. Cui and W. Liu, Practical challenges and future perspectives of all-solid-state lithium-metal batteries, Chem, 2019, 5, 753–785.
4. Y. Wang, X. Yang, Z. Zhang, X. Hu, Y. Meng, X. Wang, D. Zhou, H. Liu, B. Li and G. Wang, Electrolyte design for rechargeable anion shuttle batteries, eScience, 2022, 2, 573–590.
5. H. Liang, L. Wang, A. Wang, Y. Song, Y. Wu, Y. Yang and X. He, Tailoring practically accessible polymer/inorganic composite electrolytes for all-solid-state lithium metal batteries: a review, Nano-Micro Lett., 2023, 15, 42.
6. P. Lu, D. Wu, L. Chen, H. Li and F. Wu, Air stability of solid-state sulfide batteries and electrolytes, Electrochem. Energy Rev., 2022, 5, 3.
7. A. K. Yang, K. Yao, M. Schaller, E. Dashjav, H. Li, S. Zhao, Q. Zhang, M. Etter, X. C. Shen, H. M. Song, Q. Q. Lu, R. J. Ye, I. Moudrakovski, Q. Q. Pang, S. Indris, X. C. Wang, Q. L. Ma, F. Tietz, J. Chen and G. Olivier, Enhanced room-temperature Na⁺ ionic conductivity in Na_4.92Y_0.92Zr_0.08Si₄O₁₂, eScience, 2023, 3, 100175.
8. Y. Liang, H. Liu, G. Wang, C. Wang, Y. Ni, C. W. Nan and L. Z. Fan, Challenges, interface engineering, and processing strategies toward practical sulfide–based all–solid–state lithium batteries, InfoMat, 2022, 4, e12292.
9. Q. Zhang, D. Cao, Y. Ma, A. Natan, P. Aurora and H. Zhu, Sulfide-based solid-state electrolytes: synthesis, stability, and potential for all-solid-state batteries, Adv. Mater., 2019, 31, e1901131.
10. H. Liu, Y. Liang, C. Wang, D. Li, X. Yan, C. W. Nan and L. Z. Fan, Priority and prospect of sulfide-based solid-electrolyte membrane, Adv. Mater., 2023, 35, e2206013.
11. L. Ye and X. Li, A dynamic stability design strategy for lithium metal solid state batteries, Nature, 2021, 593, 218–222.
12. L. Jia, J. Zhu, X. Zhang, B. Guo, Y. Du and X. Zhuang, Li-solid electrolyte interfaces/interphases in all-solid-state Li batteries, Electrochem. Energy Rev., 2024, 7, 12.
13. Y. Wu, J. Xu, P. Lu, J. Lu, L. Gan, S. Wang, R. Xiao, H. Li, L. Chen and F. Wu, Thermal stability of sulfide solid electrolyte with lithium metal, Adv. Energy Mater., 2023, 13, 2301336.
14. S. J. Guo, Y. T. Li, B. Ling, N. S. Grundish, A. M. Cao, Y. G. Sun, Y. S. Xu, Y. L. Ji, Y. Qiao, Q. H. Zhang, F. Q. Meng, Z. H. Zhao, D. Wang, X. Zhang, L. Gu, X. Q. Yu and L. J. Wan, Coordination-assisted precise construction of metal oxide nanofilms for high-performance solid-state batteries, J. Am. Chem. Soc., 2022, 144, 2179–2188.
15. L. Zhai, K. Yang, F. Y. Jiang, W. B. Liu, Z. H. Yan and J. C. Sun, High-performance solid-state lithium metal batteries achieved by interface modification, J. Energy Chem., 2023, 79, 357–364.
16. L. Zhai, J. H. Wang, X. Y. Zhang, X. Z. Zhou, F. Y. Jiang, L. Li and J. C. Sun, Garnet-type solid-state Interface engineering of Li_6.75La₃Zr_1.75Ta_0.25O₁₂ via in situ built LiI/ZnLi_x mixed buffer layer for solid state lithium metal batteries, Chem. Sci., 2024, 15, 7144.
17. M. Wu, M. Li, Y. Jin, X. Chang, X. Zhao, Z. Gu, G. Liu and X. Yao, In situ formed LiF-Li₃N interface layer enables ultra-stable sulfide electrolyte-based all-solid-state lithium batteries, J. Energy Chem., 2023, 79, 272–278.
18. H. Wan, Z. Wang, W. Zhang, X. He and C. Wang, Interface design for all-solid-state lithium batteries, Nature, 2023, 623, 739–744.
19. M. Yang, K. Q. Yang, Y. J. Wu, Z. X. Wang, T. H. Ma, D. X. Wu, L. Yang, J. R. Xu, P. S. Lu, J. Peng, Z. W. Jiang, X. Zhu, Q. F. Gao, F. Q. Xu, L. Q. Chen, H. Li and F. Wu, Dendrite-free all-solid-state lithium metal batteries by in situ phase transformation of the soft carbon-Li₃N interface layer, ACS Nano, 2024, 18, 16842–16852.
20. C. Liu, B. Chen, T. Zhang, J. Zhang, R. Wang, J. Zheng, Q. Mao and X. Liu, Electron redistribution enables redox-resistible Li₆PS₅Cl towards high-performance all-solid-state lithium batteries, Angew. Chem., Int. Ed., 2023, 62, e202302655.
21. Z. Sun, Y. Lai, N. Lv, Y. Hu, B. Li, L. Jiang, J. Wang, S. Yin, K. Li and F. Liu, Insights on the properties of the O-doped argyrodite sulfide solid electrolytes (Li₍₆₎PS_(5−x)ClO_(x)x = 0–1), ACS Appl. Nano Mater., 2021, 13, 54924–54935.
22. J. K. Sang, B. Pan, Z. Tang, Y. Liu and Z. Zhou, One stone, three birds: an air and interface stable argyrodite solid electrolyte with multifunctional nanoshells, Adv. Sci., 2023, 10, 2304117.
23. Y. Z. Zhang, H. L. Chang, A. G. Han, S. J. Xu, X. Y. Wang, S. J. Yang, X. H. Hu, Y. J. Sun, X. Sun, X. Chen and Y. A. Yang, Synergistic Li₆PS₅Cl@Li₃OCl composite electrolyte for high-performance all-solid-state lithium batteries, Green Energy Environ., 2024 DOI:10.1016/j.gee.2024.07.001.
24. L. Braks, J. Zhang, A. Forster, P. Fritz, J. Oh, M. El Kazzi, J. W. Choi and A. Coskun, Interfacial stabilization by prelithiated trithiocyanuric acid as an organic additive in sulfide-based all-solid-state lithium metal batteries, Angew. Chem., Int. Ed., 2024, 63, e202408238.
25. H. Liu, W. Jiang, W. Chen, Q. Lin, S. Ren, Y. Su, R. Tong and Y. Zhang, Dendrite growth and inhibition in all-solid-state lithium metal batteries: in situ optical observation, J. Mater. Chem. A, 2024, 12, 3575–3579.
26. X. R. Chen, X. Chen, C. Yan, X. Q. Zhang, Q. Zhang and J. Q. Huang, Role of lithiophilic metal sites in lithium metal anodes, Energy Fuels, 2021, 35, 12746–12752.
27. R. Pathak, K. Chen, A. Gurung, K. M. Reza, B. Bahrami, J. Pokharel, A. Baniya, W. He, F. Wu, Y. Zhou, K. Xu and Q. Qiao, Fluorinated hybrid solid-electrolyte-interphase for dendrite-free lithium deposition, Nat. Commun., 2020, 11, 93.
28. M. Zhu, X. Zhao, R. Yan and J. Zhang, Recent research progress of alloy-containing lithium anodes in lithium-metal batteries, Curr. Opin. Solid State Mater. Sci., 2023, 27, 101079.
29. Z. Luan, P. Ren, X. Wang, H. Chen and Q. Zhou, Nitrogen doped sulfide solid electrolytes with enhanced air stability and lithium metal compatibility, Int. J. Appl. Ceram. Technol., 2024, 21, 3370–3377.
30. Z. Bi, R. Shi, X. Liu, K. Liu, M. Jia and X. Guo, In situ conversion reaction triggered alloy@antiperovskite hybrid layers for lithiophilic and robust lithium/garnet interfaces, Adv. Funct. Mater., 2023, 33, 2307701.
31. L. Tan, P. Chen, Q. Y. Chen, X. Huang, K. Y. Zou, Y. M. Nie and L. J. Li, A Li₃Bi/LiF interfacial layer enabling highly stable lithium metal anode, Rare Met., 2023, 42, 4081–4090.
32. F. Zhao, Q. Sun, C. Yu, S. Zhang, K. Adair, S. Wang, Y. Liu, Y. Zhao, J. Liang, C. Wang, X. Li, X. Li, W. Xia, R. Li, H. Huang, L. Zhang, S. Zhao, S. Lu and X. Sun, Ultrastable anode interface achieved by fluorinating electrolytes for all-solid-state Li metal batteries, ACS Energy Lett., 2020, 5, 1035–1043.
33. H. Lim, S. Jun, Y. B. Song, K. H. Baeck, H. Bae, G. Lee, J. Kim and Y. S. Jung, Rationally designed conversion-type lithium metal protective layer for all-solid-state lithium metal batteries, Adv. Energy Mater., 2024, 14, 2303762.
34. D. H. S. Tan, E. A. Wu, H. Nguyen, Z. Chen, M. A. T. Marple, J. M. Doux, X. Wang, H. Yang, A. Banerjee and Y. S. Meng, Elucidating reversible electrochemical redox of Li₆PS₅Cl solid electrolyte, ACS Energy Lett., 2019, 4, 2418–2427.
35. C. Yan, R. Xu, J. L. Qin, H. Yuan, Y. Xiao, L. Xu and J. Q. Huang, 4.5 V high-voltage rechargeable batteries enabled by the reduction of polarization on the lithium metal anode, Angew. Chem., Int. Ed., 2019, 58, 15235–15238.
36. B. W. Taklu, Y. Nikodimos, H. K. Bezabh, K. Lakshmanan, T. M. Hagos, T. A. Nigatu, S. K. Merso, H. Y. Sung, S. C. Yang, S. H. Wu, W. N. Su and B. J. Hwang, Air-stable iodized-oxychloride argyrodite sulfide and anionic swap on the practical potential window for all-solid-state lithium-metal batteries, Nano Energy, 2023, 112, 108471.
37. E. E. Ushakova, A. Frolov, A. A. Reveguk, D. Y. Usachov, D. M. Itkis and L. V. Yashina, Solid electrolyte interface formation between lithium and PEO-based electrolyte, Appl. Surf. Sci., 2022, 589, 153014.
38. W. Feng, Y. Zhao and Y. Xia, Solid interfaces for the garnet Electrolytes, Adv. Mater., 2024, 36, e2306111.
39. A. Banerjee, X. Wang, C. Fang, E. A. Wu and Y. S. Meng, Interfaces and interphases in all-solid-state batteries with inorganic solid electrolytes, Chem. Rev., 2020, 120, 6878–6933.
40. J. Jiang, Y. Ou, S. Lu, C. Shen, B. Li, X. Liu, Y. Jiang, B. Zhao and J. Zhang, In situ construction of Li-Mg/LiF conductive layer to achieve an intimate lithium-garnet interface for all-solid-state Li metal battery, Energy Storage Mater., 2022, 50, 810–818.
41. H. Lim, S. Jun, Y. B. Song, K. H. Baeck, H. Bae, G. Lee, J. Kim and Y. S. Jung, Rationally designed conversion–type lithium metal potective layer for all–solid–state lithium metal batteries, Adv. Energy Mater., 2024, 14, 2303762.
42. Z. Wu, C. Liu, X. Liu, Z. Jiang, C. Wei, Q. Luo, L. Li, L. Yu, L. Zhang, S. Cheng and C. Yu, AgPF₆ modified lithium interphases enable superior performance for Li_5.5PS_4.5Cl_1.5-based all-solid-state lithium metal batteries, J. Power Sources, 2024, 602, 234295.
43. G. Song, Z. L. Yi, F. Y. Su, L. J. Xie and C. M. Chen, New insights into the mechanism of LiDFBOP for improving the low-temperature performance via the rational design of an interphase on a graphite anode, ACS Appl. Mater. Interfaces, 2021, 13, 40042–40052.
44. Sahil, G. Natanasabapathi, S. Shyleshan, R. Kumar, M. K. Yadav and P. Kumar, Optically stimulated luminescence in LiF-MgF₂ system and its response as medical radiation dosimeter, Ceram. Int., 2023, 49, 16352–16362.

---

Footnotes

† Electronic supplementary information (ESI) available. See DOI: https://doi.org/10.1039/d4qi02717e

‡ These authors contributed equally to this work.

---