Suppressing the Shuttle Effect in Lithium-Sulfur Batteries by a UiO-66-Modified Polypropylene Separator.

The lithium-sulfur battery is one of the most promising battery technologies with high energy density that exceeds the presently commercialized ones. The shuttle effect caused by the migration of soluble polysulfides to the lithium anode is known as one of the crucial issues that prevent the Li-S batteries from practical application. Modification of the separator is regarded as a convenient yet efficient strategy to alleviate the shuttle effect. In this report, we use a thermally stable and chemically robust metal-organic framework (MOF), UiO-66, as a physical and chemical barrier for soluble polysulfides to functionalize the commercial polypropylene separator. The Li-S cell assembled with such a separator shows a significantly improved cycling stability with an average specific capacity of ca. 720 mA h g-1 at a current rate of 0.5 C for 500 cycles. Experimental and theoretical investigations indicate that the cell performance enhancement results from the physical restriction of the MOF barrier layer and strong chemical interaction between UiO-66 and polysulfides. The excellent thermal stability and chemical robustness (in acid/alkali solutions, conventional organic solvents, and polysulfide electrolytes) of UiO-66 make it highly competitive among various materials developed for separator modification in Li-S batteries.

Introduction

The lithium–sulfur battery has been considered as one of the most promising battery technologies because of its high energy density and the low-cost and high abundant sulfur element.[1,2] However, there are still some crucial challenges toward the practical application of lithium–sulfur batteries, such as the shuttle effect of highly soluble polysulfide intermediates, which leads to internal short circuit inside the cell.[3−6] Many strategies have been proposed to alleviate the shuttle effect, for example, physical and/or chemical confinement of the soluble polysulfide intermediates in the cathode.[7−14] Alternatively, functionalizing the cell separator with an additional barrier layer also provides a straightforward approach.[15−21] In consideration of the compatibility of lithium–sulfur batteries with the existing battery technologies, it is preferred to use components well-developed for lithium-ion batteries such as the commercialized polypropylene (PP) or polyethylene separators. Such separators are advantageous in cost, mechanical strength, chemical stability, and electrochemical stability. However, they are incapable of restraining polysulfide diffusion. Accordingly, some functional materials are implanted on the commercialized separators as the blocking interlayer to prevent polysulfide migration from the cathode to the anode while maintaining the capability of ion conductivity.[22−24]

Metal–organic frameworks (MOFs) are a class of porous crystalline materials composed of metal ions (or clusters) linked by organic ligands.[25−27] Because of their large surface areas, regular pore sizes, and tailorable chemistry, MOF materials have shown great promise for a variety of applications.[28−34] Given their uniform (sub-) nanometer-sized pore structures, MOFs would be promising candidate materials to construct separators capable of alleviating the shuttle effect.[35−40] Three strategies have been developed for the preparation of MOF-based/-derived cell separators. For example, He et al. used HKUST-1 nanoparticles as assembly units and polyvinylidene difluoride (PVDF)–hexafluoropropylene as a binder to prepare the separator,[41] which served as a good physical barrier to confine the polysulfide intermediates in the cathode side and led to a stable Li plating/stripping even at high current densities. He et al. reported the in situ growth of ZIF-67 and its conversion into hollow Co9S8 arrays on a PP separator,[42] which functioned as an efficient polysulfide barrier for high-performance lithium–sulfur batteries. Wu et al. directly casted ZIF-8/carbon nanotube/PVDF slurry onto one side of a PP separator and improved significantly the reversible capacity and cycling stability of lithium–sulfur batteries.[43] Until now, however, little attention has been paid on the stability of MOFs on exposure to the electrolyte environment. In fact, the mismatched standard redox potentials of the transition metals in MOFs and polysulfides/lithium might raise the risk of structure collapse of the framework materials.[44]

In addition, most investigated MOFs are very sensitive to moisture or the acidity of the electrolyte, which adds on additional restriction to the electrolyte choice.[45] Therefore, it is crucial to search for proper MOF materials that can effectively alleviate the shuttle effect through physical and/or chemical interactions at the molecular level while maintaining good structural stability.[46,47] Herein, we report the UiO-66-modified PP separator that enables a significant improvement in cyclability of lithium–sulfur batteries. The UiO–PP separator was prepared by casting the slurry composed of MOF crystals, Super P carbon, and PVDF in 1-methyl-2-pyrrolidinone on a commercial PP separator. Electrochemical performance study indicates that the UiO–PP separator can significantly improve the cycling stability of the assembled cells over the conventional PP separator, with an average specific capacity of ca. 720 mA h g–1 at a current rate of 0.5 C for 500 cycles. Experimental results and first-principle calculations suggest that the chemical adsorption of polysulfides on UiO-66 frameworks combined with the physical restriction effect of the UiO–PP separator greatly facilitates the alleviation of shuttle effect and thus contributes to lithium–sulfur cells with improved stability and cyclability.

Results and Discussion

UiO-66 crystals were synthesized via the base/acid comodulated method.[49] This controlled synthesis strategy allows for the large-scale production of UiO-66 octahedral microcrystals with the uniform but tunable sizes. X-ray diffraction (XRD) analysis indicated that the as-synthesized product was crystalline and displayed a diffraction pattern (Figure a) that is identical to that of UiO-66.[48] The crystal structure was built with hexanuclear Zr6-octahedron clusters connected by terephthalic acid, featuring regular octahedral cages (11 Å) and tetrahedral ones (9 Å) connected by the 0.6 nm triangular window. Scanning electron microscopy (SEM, Figure b) revealed that the product consisted of octahedral crystals with an average size of ca. 400 nm and a relatively narrow size distribution. Energy-dispersive X-ray microanalysis confirmed the existence of C, O, and Zr and residual dimethylformamide (DMF) molecules in the as-prepared UiO-66 crystals.[50] Thermogravimetric analysis (TGA) was carried out to evaluate the thermal stability of the as-prepared UiO-66 sample. The TGA traces showed weight losses in the temperature ranges of 25–230, 230–390, and 390–550 °C, corresponding to the weight loss from volatilization of solvent molecules, the dehydroxylation and elimination of monodentate modulators, and the decomposition of 1,4-benzenedicarboxylic linkers, respectively (Figure c).[48,51] The porosity was investigated by N2-sorption measurement at 77 K. As shown in Figure d, the evacuated UiO-66 sample exhibited the type I isotherms with a rapid increase in N2 uptake at low relative pressure (<0.01), suggesting the microporous structure. The Brunauer–Emmett–Teller surface area was calculated as 1284.2 m2 g–1. The pore size distribution analysis indicated that the most populated pore widths were centered at 0.91 and 1.09 nm, which correspond to the regular tetrahedral and octahedron cages in UiO-66, respectively.[48]

Figure 1

Characterization of the as-prepared UiO-66 crystals. (a) XRD patterns of UiO-66 and crystal structures showing the Zr6-octahedron cluster, octahedron cage, and tetrahedral cage in the UiO-66 frameworks. (b) Representative SEM image of the as-prepared UiO-66 crystals and elemental mapping of C, O, N, and Zr elements. (c) TGA trace (normalized with the final weight to 100%) measured at a ramping rate of 10 °C/min in air. (d) N2-sorption isotherms (upper panel) and pore size distribution (lower panel) of the evacuated UiO-66 crystals.

The UiO–PP separator was prepared by mixing UiO-66 crystals with a conductive carbon and PVDF binder in N-methyl-2-pyrrolidone (NMP) and by casting the resultant homogeneous slurry onto a Celgard 2500 membrane. SEM investigation (Figures a and S1, Supporting Information) indicated that the MOF-containing layer on the PP membrane had a uniform thickness of ca. 20 μm and the UiO-66 crystals and carbon additive were distributed homogeneously in the layer. To probe the effectiveness of the MOF–PP separator for suppressing the diffusion of polysulfides, a glass vial was loaded with a crimson polysulfide electrolyte solution and installed with a cap with an open hole (Ø 4 mm) that was sealed with a piece of UiO–PP or PP separator. After the bottle was upturned into the blank electrolyte, the polysulfide diffusion across the UiO–PP or PP separators was visually investigated. As shown in Figure b, no obvious polysulfide diffusion into the blank electrolyte was observed for the MOF–PP separator within 24 h. In contrast, polysulfides diffused quickly through the PP separator and discolored the blank within 12 h. This result proved preliminarily that the modification of the PP separator with UiO-66 can restrain effectively the polysulfide diffusion.

Figure 2

Characterization of the UiO–PP separator. (a) Surface (upper) and cross-sectional (lower) structure of the UiO–PP separator. (b) Diffusion of the polysulfides in the visualized bottle using UiO–PP and PP separators. (c) Nyquist plots (upper) and corresponding equivalent circuit (lower) of a Li–S cell using UiO–PP and PP separators.

To verify whether such a capability brought down the Li+-ion conductivity across the separator in a Li–S cell, electrochemical impedance spectroscopy (EIS) was used to study the internal resistance of a CR2032 coin cell consisting of 1 M LiTFSI in 1,3-dioxolane (DOL)/1,2-dimethoxyethane (DME) as the electrolyte, S/Super P carbon as the cathode, Li metal as the anode, and UiO–PP or PP as the separator. The Nyquist plots and Nyquist equivalent circuit are shown in Figure c. In the equivalent circuit, Re, Rct, and Rf stand for the Ohmic resistances of the electrolyte, charge transfer, and the film electrode, respectively. CPEct and CPEf represent the constant-phase element. Zw is the Warburg impedance of the process of Li+ ions diffusing into the electrolyte. Both cells exhibited the similar Ohmic resistances of the electrolyte (6.2 Ω for the UiO–PP separator and 6.9 Ω for the PP separator, respectively). It is known that the soluble polysulfides can shuttle back and forth freely through the channels of the PP separator, causing the shuttle effect in a Li–S cell. However, for the UiO–PP separator, these channels were blocked, benefitting to suppress the migration of polysulfides to the Li anode. Therefore, the cell with a UiO–PP separator exhibited lower charge transfer (27.1 Ω) and film electrode resistances (15.2 Ω) than the PP separator counterpart (31.8 and 38.4 Ω, respectively). Besides, the open-circuit voltages of Li–S cells using the UiO–PP separator exhibited an antiself-discharge feature as shown in Figure S2 (Supporting Information). It was observed that the open-circuit voltage of a normal cell only with a PP separator remained at ca. 2.35 V versus Li+/Li with gradual decay in the measured 180 h, which indicates an immediate spontaneous reduction of sulfur into high-order polysulfides. Whereas this self-discharging was inhibited by the incorporation of the UiO–PP separator, of which the open-circuit voltage remained stable at ca. 2.7 V versus Li+/Li for more than 150 h because of the prohibition of soluble sulfur emigration into the anode. Nevertheless, cyclic voltammetry (CV) tests with different sweep rates (Figure S3, Supporting Information) indicated that the lithium-ion diffusion coefficient (Table S1, Supporting Information) was decreased by the introduction of a UiO interlayer.

The electrochemical performance of the UiO–PP separator was investigated and compared with that of the bare PP separator and that modified with SiO2 nanoparticles (denoted as SiO2–PP). Previous studies have demonstrated that the PP separators with and without SiO2 nanoparticle modification showed the similar stable electrochemical window and SiO2 nanoparticles could function as a physical barrier to alleviate polysulfide migration across the separator.[52]Figure a shows the CV profiles of the Li–S cells with UiO–PP, SiO2–PP, and PP as separators. All three cells showed typical reduction/oxidation reactions of elemental sulfur with two cathodic peaks in the potential range of 1.8–2.1 and 2.1–2.4 V (vs Li+/Li) and an anodic peak in the potential range of 2.4–2.7 V. The two cathodic peaks corresponded to the reduction of S8 to soluble polysulfide intermediates (Li2S, 4 ≤ x ≤ 8) and further to solid Li2S2/Li2S and the anodic peak was related to the conversion of Li2S2/Li2S back to Li2S and S8.[53,54] Compared with that using a PP separator, the cells with UiO–PP and SiO2–PP separators exhibited smaller polarization upon charging because of the improved electric conductivity caused by a conductive carbon additive in the barrier layer. Galvanostatic charge/discharge test (Figure b) showed that all cells exhibited one major plateau in charge processes and two major plateaus in discharge processes, consistent with one major anodic peak and two major cathodic peaks observed in CV profiles. The discharge capacity in the cell using the UiO–PP separator reached 1032 mA h g–1 at a current rate of 0.5 C (1 C = 1675 mA g–1), which was greater than those in the cells using SiO2–PP and PP separators (921 and 737 mA h g–1, respectively). In addition, the cells using the UiO–PP separator showed less overpotential compared with the other two. It is noteworthy that the cell with the SiO2–PP separator showed the increased discharge capacity and decreased overpotential compared with that using the PP separator, indicating certain capability of the SiO2 nanoparticle layer to alleviate the shuttle effect. Nevertheless, it is obvious that the UiO-66 barrier layer exhibits a more effective suppression of polysulfide migration to the anode side than that of SiO2.

Figure 3

Li–S cell performance using UiO–PP, SiO2–PP, and PP as separators. (a) CV profiles of the Li–S cells in the potential range of 1.5–3.0 V (vs Li+/Li) at a sweeping rate of 0.2 mV s–1. (b) Initial galvanostatic charge/discharge profiles at a current rate of 0.5 C. (c) Galvanostatic specific charge/discharge capacity (Cs) and corresponding Coulombic efficiency (ηCE) of the Li–S cells. (d) Rate performance and corresponding CE of the Li–S cells. In all panels, the colors of red, blue, and black are assigned to the cells using UiO–PP, SiO2–PP, and PP as separators, respectively.

The stability test of the Li–S cells was evaluated at a 0.5 C rate and shown in Figure c. All three cells showed a Coulombic efficiency of ca. 100%, which should result from the passivation of the Li anode by LiNO3.[55] Capacities at the end of the 500th cycle were observed as 586, 357, and 320 mA h g–1 and the corresponding average capacity retentions per cycle were calculated as 99.85, 99.75, and 99.78% for the cells using UiO–PP, SiO2–PP, and PP separators, respectively, indicating the higher cycling stability of the former. With higher sulfur loading and sulfur weight ratio in the electrode, the UiO–PP separator also exhibited better cycling stability and higher specific capacity (Figure S4, Supporting Information). XRD patterns recorded on the UiO–PP separator after cycling test demonstrated no obvious crystal structure (Figure S5, Supporting Information). Further examination on the UiO–PP separator that was stored in ambient condition for 3 days after cycling also suggested a well-maintained crystal structure of UiO-66. Besides, compared with the PP separator, the UiO–PP separator featured greatly decreased deposition of insoluble polysulfides and fluffy Li dendrites on the anode surface (Figure S6, Supporting Information), demonstrating in turn its ability in internal shuttle suppression and Li anode protection. The rate capability was evaluated by cycling the cells at different C-rates (Figure d). The average capacities at C-rates of 0.1, 0.2, 0.5, 1, and 2 C were 1120/937/848, 1016/781/745, 855/671/540, 645/618/326, and 461/514/191 mA h g–1 for the cells with UiO–PP, SiO2–PP, and PP separators, respectively. No doubt the cell with the PP separator showed the lowest capacity at all C-rates because of the uncontrolled shuttle effect. The cell with the UiO–PP separator exhibited higher capacity than that with the SiO2–PP separator at moderate C-rates (≤0.5 C), whereas they displayed the similar capacity at higher C-rates (≥1 C) as high rates would reduce the retention time of polysulfides within the electrolyte and thus weaken the shuttle effect.[56] Compared with previously reported MOF-modified separators (Table ), the UiO–PP separator should be competitive for the Li–S cell application with regard to the cycle life of cells and chemical stability of barrier materials.

Table 1

Comparison of Electrochemical Performance and Chemical Stability of MOF-Modified Separators

MOFs	cathode electrode	S loading	average capacity (mA h g–1)	number of cycles	ref	chemical stability
ZIF-7	S/Super P	63% wt ca. 1 mg cm–2	ca. 500@0.25 C	300	(44)	stable in basic solution (34)(37)(43)(44),
ZIF-8	S/carbon black	64% wt ca. 3.5 mg cm–2	ca. 650@0.2 C	100	(37)
ZIF-30	S/acetylene black	ca. 1.2 mg cm–2	ca. 950@0.2 C	100	(43)
ZIF-67	S/Super P	70% wt	ca. 900@0.1 C	200	(34)
ZIF-45	S/acetylene black	ca. 1.2 mg cm–2	ca. 840@0.2 C	100	(43)
HKUST-1	S/CMK-3	56% wt 0.6–0.8 mg cm–2	ca. 900@1 C	1500	(35)	stable in conventional organic solvent; unstable in aqueous phase; react with polysulfide (35)(41)(44)(57)(58),
HKUST-1	S/RGO	5.8 mg cm–2	ca. 750@2 C	2000	(41)
HKUST-1	S/Super P	50% wt	ca. 800@0.5 C	500	(57)
Zn-HKUST-1	S/CMK-3/Super P	56% wt	ca. 700 @1 C	1000	(58)
Y-FTZB	S/Super P	63% wt ca. 1 mg cm–2	ca. 650@0.25 C	300	(44)	stable in polysulfide electrolyte (37)(44),
Ni3(HITP)2	S/carbon black	64% wt ca. 3.5 mg cm–2	ca. 800@1 C	500	(37)	stable in polysulfide electrolyte (37)(44),
UiO-66	S/Super P	65% w.t. ca. 1.5 mg cm–2	ca. 720@0.5 C	500	this work	stable in acidic/basic solution, conventional organic solvent, and polysulfide electrolyte (48)

Density functional theory (DFT) simulation was performed to gain further insights into the roles of UiO-66 in alleviating effectively the shuttle effect in Li–S cells. According to previous investigations,[59] the S–S chain length in polysulfides (Li2S, 4 < n < 8) was in the range of 2.09 and 2.39 Å, and the Li–S bond length was approximately 2.6 Å. The lengths were all smaller than the pore size of UiO-66, indicating that Li2S could penetrate into the pores and interact with the inner walls of the pores. As a result, the pore size would be reduced to prevent further penetration of polysulfides across the pores. Figure a shows the optimized geometries of atomic model configurations between the UiO-66 framework and polysulfides. The intercalations were quantitatively illustrated by the adsorption energy calculation and reflected the immobilizing ability of polysulfides on the UiO-66 framework. The calculated adsorption energy was less than −2 eV for all the polysulfides except for S8 (Figure b) and was also less than those previously reported on materials derived from other MOFs,[60] indicating a strong chemical adsorption of polysulfides on the UiO-66 framework. Such a chemical adsorption should result from the strong binding ability between Li+ ions and O atoms of the ligands[61] and between the S and H atoms of the UiO-66 framework. We further used molecular surface ESP to verify the origin of the chemical absorption. The O atoms in the UiO-66 framework exhibited strong negative charge, whereas the Li+ ions in polysulfides exhibited strong positive charge (Figure c), which would lead to strong Coulomb attraction between Li+ ions and O atoms. Because of strong electrostatic attraction between Li+ ions and S2–, the polysulfides could be immobilized on the UiO-66 framework by Coulombic attraction. Such a chemical adsorption was observed through a visualized static adsorption test by dispersing UiO-66 crystals in the polysulfide solution (Figure S6, Supporting Information). After soaking the UiO-66 crystals in a polysulfide solution for 24 h, the color of the solution changed from yellowish brown to colorless, and the color of UiO-66 crystals changed from white to bright yellow, indicating that most polysulfides were captured by UiO-66 crystals. The chemical intercalation between UiO-66 crystals and polysulfides was further studied by X-ray photoelectron spectroscopy (XPS). Three contributions of O were found in the pristine UiO-66 crystals, which are assigned to the Zr–O, C=O, and H–O bonds, respectively (Figure S7, Supporting Information). After chemisorption, another contribution of O emerged at a binding energy of ca. 531 eV, which corresponds to the binding energy of Li–O bonds.[62] It could be concluded that the doped Li+ ions did interact with the 1,4-benzenedicarboxylic acid ligand (i.e., exchange the carboxylate protons) and form lithium alkoxide (i.e., lithium carboxylate group). Besides, S atoms with weak negative charges might also interact with the positively charged surface of the UiO-66 framework to further intensify the chemical adsorption. On the basis of the above analysis, the chemical adsorption of polysulfides on UiO-66 frameworks combined with the physical restriction effect of the UiO–PP separator greatly suppressed the shuttle effect and thus contributed to the improved stability and cyclability of Li–S cells.

Figure 4

DFT simulation of the chemical interaction between UiO-66 crystals and polysulfides. (a) Geometries of atomic model configurations and (b) corresponding adsorption energy between the UiO-66 framework and polysulfides. (c) Molecular surface electrostatic potential (ESP) of UiO-66 and polysulfides.

Conclusions

In summary, uniform UiO-66 crystals with narrow size distribution have been synthesized via solvothermal reaction. Because of its regular pore sizes and strong interaction with soluble polysulfides, UiO-66 can serve as both physical and chemical barrier materials for soluble polysulfides to suppress the shuttle effect in Li–S batteries. The Li–S cell assembled using a UiO–PP separator exhibits a long cycle life with a specific capacity of 586 mA h g–1 after 500 charge/discharge cycles at a current rate of 0.5 C and the nearly 100% Coulombic efficiency during the cycling, disclosing that the MOF–PP separator can prevent effectively the soluble polysulfides from migrating toward the Li anode while having negligible influence on lithium-ion transference. Moreover, the high thermal stability and excellent robustness against polysulfides, moisture, and oxygen make UiO-66 a highly promising candidate material for functionalizing the commercialized separators for Li–S battery applications. This work will also draw the attention toward the development of separators for energy storage devices with emphasis on the stability and efficiency.

Experimental Section

Preparation of UiO-66

1,4-Benzenedicarboxylic acid (0.60 mmol, 99.5%, Sigma-Aldrich) and triethylamine (0.60 mmol, 98%, J&K Chemicals) were charged into a round-bottle flask containing N,N-DMF (140 mL, 99.5%, J&K Chemicals). After the mixture was stirred at room temperature for 10 min, acetic acid (360 mmol, 99.5%, J&K Chemicals) was slowly added. The solution was heated to 120 °C in an oil bath and maintained at this temperature for 1 h. ZrCl4 (10 mL; 99.5%, Acros) solution in DMF (60 mM) was added, and the mixed solution was allowed to react at 120 °C for 6 h without stirring. After the solution was cooled to room temperature, the product was collected by centrifugation, washed several times with DMF and methanol, and soaked in methanol for 3 days.

Preparation of the UiO-66-Modified PP Separator (UiO–PP)

The as-prepared UiO-66 crystals, Super P carbon (Timcal), and PVDF (Fisher Scientific) with a weight ratio of 7:2:1 were added into NMP (anhydrous, Fisher Scientific) and thoroughly mixed in a dispensing system (ARE310, Thinky Corp.) at 2000 rpm for 20 min. The slurry was directly coated on one side of a commercial Celgard 2500 membrane using a doctor blade. The modified membrane was dried in a vacuum oven at 55 °C for 24 h and punched into a circular separator (Ø 19 mm). The typical loading amount was around 0.35 mg cm–2. For the preparation of the SiO2 nanoparticle-modified PP separator, the equal weight ratio of the SiO2 nanoparticle was used instead of UiO-66 crystals in preparation of the slurry while other procedures were kept unchanged. To prepare SiO2 nanoparticles, tetraethyl orthosilicate (TEOS, 2.08 mL, 95%, J&K Chemicals) was added to ethanol (50 mL, 95%, Sino Reagent) to form the precursor solution. Another solution composed of deionized water (3 mL), ammonia hydroxide (3.85 mL, 28 wt %, Sino Reagent), and ethanol (40 mL) was slowly added into the precursor solution and reacted at room temperature for 20 h with stirring. The product was collected by centrifugation, washed several times with ethanol and deionized water, and finally redispersed in ethanol. The as-prepared SiO2 nanoparticles were spheres with a diameter of ca. 400 nm and narrow size distribution.

Electrode Preparation and Cell Assembly

To prepare the sulfur electrodes, elemental sulfur (99.5%, Sigma-Aldrich), Super P carbon, and PVDF binder in NMP (5% weight ratio) were thoroughly mixed with a weight ratio of 65:30:5 in a dispensing system. The resultant slurry was cast onto an aluminum foil with a thickness of 100 μm using a doctor blade method. The thickness of the slurry was about 100 μm. The sulfur electrodes were dried in vacuum at 65 °C for 10 h and then cut into circular pellets (Ø 10 mm). The typical sulfur loading for each electrode was 1.5 mg·cm–2. The cell was assembled in an argon-filled glovebox (H2O and O2 concentration <1 ppm) using a standard CR2032 coin cell. The cell used the sulfur electrode as the cathode, metallic lithium as the anode (Ø 12 mm), UiO–PP as the separator, and 1.0 M bis(trifluoromethane)sulfonimide lithium salt (LiTFSI, battery grade, Sigma-Aldrich) in DOL (anhydrous, Sigma-Aldrich)/DME (anhydrous, Sigma-Aldrich) as the electrolyte (15 μL), respectively. The volume ratio of DOL/DME was 1:1 in the electrolyte solution to which a small amount of LiNO3 (1% weight ratio, anhydrous, Sigma-Aldrich) was added.

Calculation Method

The structure optimization was performed using Gaussian 09 software package with full electron basis set sto-3g[63−67] in the vacuum, and some atoms of UiO-66 had been frozen in order to keep the symmetry and periodicity of the structure. The basic unit of UiO-66 was selected to calculate the adsorption energy considering the periodicity. The adsorption energy calculation was conducted according to the equation Eabs = Etot – (E1 + E2), where Eabs is the absorption energy between UiO-66 and polysulfide species Li2S (n = 1, 2, 4, 6, 8), Etot is the total energy of UiO-66 with Li2S, and E1 and E2 are the energies of UiO-66 and Li2S, respectively.

Characterizations

Powder XRD patterns were recorded using a Philips X’per PRDMPD diffractometer with nickel-filtered Cu Kα radiation (λ = 1.5406 Å). SEM images were taken by a Zeiss supra 55 field-emission SEM with an accelerating voltage of 8 kV. Nitrogen-sorption studies were performed at 77 K up to 1 bar with an ASAP 2020 HD88 (Micromeritics). Before the sorption measurements, the samples were activated under 150 °C for 24 h. TGA was carried out using a simultaneous thermal analyzer apparatus (DSC1, METTLER TOLEDO) at a heating rate of 10 °C min–1 under air atmosphere. XPS measurement was performed on a Thermo ESCALAB 250XI system in vacuum (<10–9 Pa) with Al Kα source at 15 kV and 12 mA. Galvanostatic experiments were performed on a battery testing system (BT-2043, Arbin Instruments) in a potential range of 1.5–3 V (vs Li+/Li). The CV and EIS measurements were performed on an electrochemical workstation (CHI760D, Chenhua Instrument) at a sweeping rate of 0.2 mV s–1 and a frequency range of 10–105 Hz, respectively.