Nano-MgO/AB decorated separator to suppress shuttle effect of lithium–sulfur battery

Lithium–sulfur (Li–S) batteries are regarded as one of the most promising energy storage systems owing to their high specific energy, low cost and eco-friendliness. However, significant capacity fading caused by the shuttle of soluble polysulfides from the cathode to the anode significantly hampers their practical application. Here, we designed a nano-MgO/acetylene black (AB) decorated functional separator to suppress the shuttle of polysulfide intermediates, which can remarkably improve the electrochemical performance of Li–S batteries. Nano-MgO with the aid of the AB conductive network exhibits superior adsorption to polysulfides due to the synergistic effect of excellent chemisorption and improved electron conductivity. The electrochemical performance of the Li–S battery highly depends on the relative amount of nano-MgO and AB in the composite coating on the separator. A battery with the optimal decorated separator (MgO-25 separator, nano-MgO and acetylene black in the weight ratio 1 : 3) exhibits a high initial discharge capacity of 1238 mA h g−1 with high coulombic efficiency (∼97%) and retains a high capacity of 875 mA h g−1 after 100 cycles at 0.2 C. This study promotes the understanding of the synergistic effect of the polysulfide adsorbent and the conductive agent on the suppression of the shuttle effect, and provides a way to design polysulfide-blocking barriers for Li–S batteries.


Introduction
With the rapid development of electric vehicles and electronic mobile devices, energy storage systems with high energy density and long cycle life are highly required. However, the energy density of traditional lithium-ion batteries based on lithium intercalation electrochemistry is restricted by their low theoretical capacity. 1,2 Rechargeable Li-S batteries composed of a sulfur cathode and a lithium-metal anode are considered as one of the most promising candidates for next-generation energy storage devices due to their large theoretical specic capacity of 1675 mA h g À1 and high theoretical energy density of 2600 W h kg À1 . Additionally, sulfur as a cathode material is lowcost, naturally abundant and environmentally benign, [3][4][5] which contributes to the commercialization of Li-S batteries. Nonetheless, the practical application of Li-S batteries is still hindered by several major issues: (i) the poor conductivity of sulfur reduces the utilization of the active material; 6 (ii) the large volume change of the active material during the charge/ discharge process can destroy the morphology and structure of the electrode and leads to the shedding of the active material; 7,8 (iii) the shuttle phenomenon of dissolved polysulde intermediates (Li 2 S n , 4 < n # 8) from the cathode to the anode results in severe self-discharge, thus causing capacity loss and the short cycle life of the batteries. 9,10 Over the years, extensive efforts have been devoted to addressing the "shuttle effect" in the Li-S battery system. One of the most popular methods is to trap sulfur and polysuldes within various hosts, including carbon (porous carbon, [11][12][13][14][15] carbon nanotubes, 16,17 graphene 18,19 and elemental doped carbon 20,21 ), conductive polymers, 22,23 and metal-organic frameworks. 24,25 Carbon materials are the most widely employed host, however, the trapped polysulde still shows gradual loss following repeated cycles owing to the weak interaction of the physical adsorption and the open pore structure of the carbon material. Recently, researchers have proposed that a chemisorption interaction between sulfur and some hosts seems to be more efficient at trapping and immobilizing sulfur species. Compared to traditional carbon materials, metal oxides (such as TiO 2 , [26][27][28][29] 36,37 The hydrophilic nature of the nano-MgO nanoparticles and the presence of more electropositive Mg sites ensure that polysuldes can effectively be trapped on the cathode side of the separator. Moreover, Mg is more electropositive than the other metals, such as Ti and Mn, which promotes the increased electrostatic binding force between nano-MgO and the polysuldes. 36 Recently, researchers have found that inserting a polysulde barrier (so-called interlayer) between the cathode and the separator can adsorb polysulde species. [38][39][40][41][42] Directly coating a functional interlayer on the cathode side of the separator to block the shuttle of polysuldes is an effective strategy. Among various interlayer materials (carbon materials, metal oxides, and metal-organic frameworks), conductive carbon materials are particularly attractive, which can also act as expanded current collectors to reduce the cell resistance and to increase sulfur utilization. Similar to the sulfur cathode host, chemisorption can effectively suppress the shuttling of polysuldes. Due to the merits of nano-MgO, it is expected that the introduction of nano-MgO to the interlayer can effectively suppress polysulde shuttling. However, it has rarely been reported to date.
Here, we present a nano-MgO/acetylene black (AB) decorated functional separator to inhibit the diffusion of polysuldes in Li-S batteries, as shown in Scheme 1. Nano-MgO with the aid of the AB conductive agent can achieve a synergistic effect of excellent chemisorption and superior electronic conductivity, which facilitates the adsorption and transformation reaction of polysuldes. Moreover, the battery performance highly depends on the relative amount of nano-MgO and AB in the composite interlayer of the separator. Insufficient nano-MgO cannot adsorb polysuldes sufficiently, resulting in poor cycling stability. On the other hand, excess nano-MgO results in an increase in the non-conductive area of battery, and polysuldes are anchored on the separator permanently without being reused, leading to low sulfur utilization. According to our experiment, a Li-S battery with the optimal separator (MgO-25 separator, MgO and AB in the weight ratio 1 : 3) exhibits better performance than other investigated batteries, and shows a high initial discharge capacity of 1238 mA h g À1 and a high capacity of 875 mA h g À1 aer 100 cycles at 0.2 C.

Experimental
Preparation of the decorated separators A 2 wt% gelatin solution was prepared by adding an appropriate amount of gelatin into deionized water. Five slurries with different weight ratios of nano-MgO and acetylene black (AB) were prepared. Nano-MgO and acetylene black (AB) were rstly mixed in various ratios by ball-milling for 5 h at 25 C. The content of gelatin in all slurries is the same. The weight ratios of nano-MgO and AB in the rst three decorated separators are 1 : 3, 3 : 3 and 9 : 3, which are dened as the MgO-25 separator, the MgO-50 separator and the MgO-75 separator, respectively. The MgO-100 separator and the AB separator consist only of nano-MgO and AB, respectively. The slurries were then coated on Celgard 2400 separators with an applicator, and were subsequently dried in a vacuum chamber at 60 C for 24 h. Lastly, the decorated separators were tablet-pressed and punched into circles with a diameter of 19 mm. The loading mass of the decorated separators was controlled between 0.2-0.4 mg.

Fabrication of the cathode
The slurry of the common sulfur cathode was prepared by mixing sublimed sulfur, AB and gelatin (2 wt% gelatin solution) in the weight ratio 63 : 30 : 7, respectively. The slurry of the pristine MgO cathode was prepared by mixing nano-MgO, AB and gelatin in the weight ratio 20 : 60 : 20, respectively. Aer ball-milling for 5 h at 25 C, the slurries were evenly coated on aluminum foil with an applicator and dried in a vacuum chamber at 60 C for 24 h. Finally, the cathodes were tabletpressed and punched into circles with a diameter of 12 mm. The loading mass of sulfur in the sulfur cathode is 1.0-1.5 mg.

Preparation of the Li 2 S 6 -containing solution
In a glove box under an argon atmosphere, the sublimed sulfur and Li 2 S with a molar ratio of 5 : 1 were mixed in tetrahydrofuran solvent. Aer 24 hours of magnetic stirring, the Li 2 S 6containing solution was prepared.

Materials characterization
The surface morphologies of the differently decorated separators and the lithium-metal aer cycling were characterized by scanning electron microscopy (SEM, HITACHIS-4800) and high resolution transmission electron microscopy (HRTEM, HITACHIH-800). The elemental mapping check was performed using an energy dispersive spectrometer. The sulfur content of the differently decorated separator coatings was determined by thermogravimetric analysis (TGA, Hengjiu, China) in the temperature range 25-800 C under an air atmosphere, with a heating rate of 10 C min À1 and an air-ow rate of 100 mL min À1 . The crystalline phases of the nano-MgO particles and the MgO-25 separator slurry were observed by X-ray diffraction (XRD, D/max-2500/PC, Rigaku) from 10 to 90 with Cu Ka radiation (l ¼ 0.154 nm). UV-visual spectroscopy (UV/vis, SEC2000-DH) was used to analyze the adsorption effect of nano-MgO and AB on lithium polysulde. The conductivity and wettability of the decorated separators were measured by a four-probe meter and a contact angle meter, respectively.

Electrochemical measurements
Coin-type (CR2025) cells were assembled and sealed in an argon-lled glove box, where the oxygen and water contents were less than 1 ppm. Celgard 2400 separators with/without decorated interlayers were used as the separator, and lithiummetal was used as the anode. 30 mL of electrolyte was used, which consisted of 1 mol L À1 lithium bis(tri-uoromethanesulfonyl)imide (LiTFSI) and a solvent of 1,3dioxolane (DOL) and 1,2-dimethoxyethane (DME) (1 : 1, in volume) with 0.4 mol L À1 LiNO 3 additive. The galvanostatic charge-discharge cycling of the batteries was performed in a potential range of 1.7-2.8 V (vs. Li/Li + ) by a LAND-CT2001A instrument. Electrochemical impedance spectroscopy (EIS) was performed on a Solartron 1280 Z at an open-circuit voltage ranging from 100 MHz to 10 kHz at room temperature. Cyclic voltammetry (CV) tests were carried out with a Solartron 1280Z at a scan rate of 0.1 mV s À1 .

Results and discussion
From the SEM images of AB (Fig. 1a) and nano-MgO (Fig. 1b), the main particle size of AB and nano-MgO is $50 nm and $100 nm, respectively. The employed nano-MgO has obvious lattice fringes and the lattice spacing is estimated to be 0.2125 nm as shown in Fig. S1, † which is consistent with that of standard MgO (JCPDS no. 45-0946, d ¼ 0.21056 nm). Furthermore, the HRTEM image of the MgO-25 separator slurry is shown in Fig. 1c. The distribution of the nano-MgO particles in the slurry looks relatively uniform, demonstrating the sufficient blending of AB and nano-MgO in the composite interlayer. According to Fig. 1d, the nano-MgO and MgO-25 separator slurries both show characteristic peaks of MgO, including (111), (200), (220), (311), and (222) in the XRD patterns. The diffraction peak of the MgO-25 separator slurry at 2q ¼ 26 and is attributed to the existence of AB. The results indicate that nano-MgO can maintain the original crystal shape well when it has been prepared as decorated separator slurries. Fig. 1e shows photographs of the six separators involved in this study, which are (e 1 ) the Celgard 2400 separator, (e 2 ) the pristine MgO-coated (MgO-100) separator, (e 3 ) the MgO-75 separator, (e 4 ) the MgO-50 separator, (e 5 ) the MgO-25 separator and (e 6 ) the pristine AB-coated separator, respectively. The interlayer coating looks quite smooth, conrming that the functional slurries can uniformly coat on the Celgard 2400 separators. Fig. 1f and S2 † are cross-section SEM images of the separators, which reveal that the thickness of the commercial Celgard 2400 separator is about 20 mm. The nano-MgO/AB interlayer is around 5 mm for all of the decorated separators except the MgO-100 separator, for which the interlayer thickness is around 10 mm. Fig. 1g, S3 and S4 † are SEM images and EDS spectra of the MgO-25 separator, the MgO-50 separator and the MgO-75 separator, respectively, which further prove that nano-MgO is uniformly distributed in the composite interlayer of the separators. Moreover, the results indicate that the AB conductive network is established in the interlayer of the separators, which facilitates ion/electron transport and accelerates the transformation of the trapped polysuldes during electrochemical cycling.
Prior to making the Li-S battery with the nano-MgO decorated separator, we investigated the interaction between poly-suldes and nano-MgO by directly adding nano-MgO powder into a Li 2 S 6 solution. As shown in Fig. 1h, the solution became clear immediately, showing that nano-MgO can adsorb poly-suldes effectively. As a control experiment, the adsorption effect of pristine AB to polysuldes is very weak, which was further conrmed by UV-visual spectroscopy. The peak located at 330-350 nm (ref. 43 and 44) is ascribed to the existence of Li 2 S 6 and becomes weak when adding nano-MgO, showing the strong chemisorption interaction between nano-MgO and Li 2 S 6 . To further conrm the interaction between polysuldes and nano-MgO, we collected XPS spectra of the Mg 2p region of nano-MgO before and aer adsorbing Li 2 S 6 solution as shown in Fig. 1i. From the XPS spectra, the Mg 2p spectrum of MgO/ Li 2 S 6 shows a 1.02 eV shi to a lower binding energy, indicating that there is electron transfer from Li 2 S 6 to Mg ions. The wettability of the separator can affect the inltration of the electrolyte and the diffusion rate of ions during the charging and discharging process. Contact angle tests were carried out to evaluate the electrolyte wettability properties on the separator. As the organic electrolyte of Li-S batteries is seriously volatile, the contact angles of the six different separators were measured with deionized water, which possesses similar polarity to the electrolyte of Li-S batteries. From Fig. S5, † the water contact angle on the pristine Celgard 2400 separator is close to 90.0 . However, the water contact angles on the decorated separators gradually decrease with the increase in nano-MgO content, indicating that the wettability of the separators improves with the nano-MgO functional coating.
To study the electrochemical performance of the Li-S batteries with the MgO coating, the pristine MgO cathode (without sulfur) and the Li anode were rst assembled into a Li-MgO battery. As shown in Fig. S6, † there are no apparent redox peaks appearing in the cyclic voltammetry (CV) curves of the Li-MgO battery in the working voltage range 1.7 V to 2.8 V. Furthermore, Fig. S6b † shows that the discharge specic capacity of the Li-MgO battery is only 1.6 mA h g À1 and there is no change for the 100 charge-discharge cycles, indicating that nano-MgO has no extra capacity contribution to the Li-S battery in its operating voltage range. Fig. 2a and Table S1 † show the capacity and cycling performance of the Li-S batteries with the various separators. Among the Li-S batteries with the various separators, the battery with the MgO-25 separator has the largest initial discharge specic capacity of 1238 mA h g À1 with a high coulombic efficiency ($97%), and retains a specic capacity of 875 mA h g À1 aer 100 charge-discharge cycles at 0.2 C. The initial discharge specic capacity of the Li-S battery with the AB decorated separator is also very high, however, the capacity retention rate is only 61.82% aer 100 cycles. This result demonstrates that nano-MgO exhibits a synergistic effect and improves both the capacity and the cycling performance of battery considerably. The charge-discharge curves of the Li-S batteries at various rates are shown in Fig. 2b and S7. † All curves display the typical characteristics of Li-S batteries with two discharge plateaus and one charge plateau, corresponding to the electrochemical redox reaction of the Li-S battery. The rst discharge plateau (FDP) at about 2.35 V is attributed to the transformation of elemental sulfur into soluble polysuldes (Li 2 S n , 4 # n # 8) and the following discharge plateau at about 2.05 V is attributed to the transformation of the long-chain polysuldes into insoluble polysuldes (Li 2 S 2 or Li 2 S). The charge plateaus are ascribed to the transformation of Li 2 S 2 /Li 2 S into S 8 . With increasing rate, the voltage gap between the discharge and charge plateaus is extended, caused by sulfur that has not reacted at high rates. The plateaus are atter and the polarization is smaller for the Li-S battery with the MgO-25 separator than that of the batteries with other separators, suggesting a kinetically efficient reaction process. From Fig. 2c, the discharge specic capacity of the Li-S battery with the MgO-25 separator is 1252.2 mA h g À1 at 0.1 C, and is still 805.3 mA h g À1 at a rate of 2 C, exhibiting superior rate capability of all the batteries with decorated separators. Furthermore, it can recover 1121.9 mA h g À1 with a high capacity retention ratio of 95.9% when the discharge rate returns to 0.1 C. The ultrahigh reversibility demonstrates that the MgO/AB composite coating on the MgO-25 separator can effectively react with the trapped polysuldes during the charge-discharge process. The rst discharge plateau (FDP) analysis is a preassessment for polysulde diffusion and the well-retained FDP capacitance manifests good reversible electrochemical performance, which also demonstrates the effective inhibition of polysulde diffusion as the FDP relates to the formation of  highly soluble polysuldes. 45 The rst discharge plateau (FDP) capacity of various batteries has been studied at a rate of 0.2 C, as shown in Fig. 2d. When employing the MgO-25 separator, the FDP capacity is 380.2 mA h g À1 , which is approximately 90.7% of the theoretical capacity (419 mA h g À1 ). Moreover, the FDP capacity of the Li-S battery with the MgO-25 separator retains 84.82% of the theoretical specic capacity aer 100 cycles, while the Li-S batteries with the MgO-50 separator, the AB separator, the MgO-75 separator, the Celgard 2400 separator and the MgO-100 separator show a capacity retention of 81.8%, 73.5%, 70.1%, 52.0% and 40.9%, respectively. Although the initial FDP capacity of the battery with the AB separator is 386.5 mA h g À1 (nearly 92.2% of the theoretical value), the capacity shows continuous decay due to the serious "shuttle effect". The above results demonstrate that the MgO-25 separator can trap poly-suldes and react with the trapped species effectively due to the synergistic effect of nano-MgO and AB.
The charge/discharge proles of the Li-S batteries with the various separators at a rate of 0.2 C are presented in Fig. 3a-f. Among them, the prole of the Li-S battery with the MgO-25 separator exhibits the smallest potential difference (DE) between the discharge and charge plateaus due to the synergistic effect of strong chemisorption and the superior conducting network. However, the pristine Celgard separator and the pure MgO decorated separator show large DE values because their poor conductivity limits the secondary reaction of the adsorbed polysuldes. The pure AB coating cannot adsorb polysuldes effectively which also causes poor chargedischarge reversibility. The relative amount of MgO and AB inuences the performance of the Li-S batteries greatly. As illustrated in Fig. 3g 1 -g 3 , the insufficient nano-MgO cannot adsorb polysuldes sufficiently, which is proposed to result in poor cycling stability. However, excess nano-MgO (such as the MgO-50 and MgO-75 separators) causes an increase in the non-conductive area of the battery, and polysuldes will anchor on the separator permanently without being reused, leading to low sulfur utilization. In this case, excess MgO restricts the reaction kinetics and causes poor chargedischarge reversibility.
To further understand the synergistic effect of nano-MgO and AB on the reaction kinetic process of Li-S batteries, the CV and EIS of the batteries were investigated. Fig. 4a-d and S8 † show the rst four CV cycles of the Li-S batteries with the various separators at 0.1 mV s À1 . The reduction peaks at $2.35 V and $1.94 V are assigned to the transformation of sulfur to soluble polysuldes (Li 2 S n , 4 # n # 8) and the long-chain pol-ysuldes to insoluble Li 2 S 2 /Li 2 S, respectively. In the subsequent anodic scan, the main peak at $2.44 V relates to the transformation of Li 2 S to high-order polysuldes and sulfur. For all the assembled Li-S batteries, the reduction peak shis right  Fig. 4e shows Nyquist plots of the Li-S batteries with various decorated separators. The semicircle in the high-frequency region represents the charge transfer resistance (R ct ) at the interface between the cathode and the electrolyte, while the inclined line in the low-frequency region corresponds to the Warburg impedance produced from the diffusion of electrolyte ions in the electrode. The Nyquist plots of the Li-S batteries with the AB separator and the MgO-25 separator almost overlap in the high frequency region and show the smallest R ct , which conrms their excellent electronic conductivity. However, a second semicircle appears in the intermediate-frequency region for the battery with the pristine Celgard 2400 separator, which is attributed to battery self-discharge during the static process, indicating increased resistance. For all the tested batteries, another semicircle appears in the medium-to-high frequency region aer 5 cycles at a rate of 0.5 C, representing the deposition resistance (R d ) of Li 2 S/Li 2 S 2 (ref. 46) as shown in Fig. 4f. It is worth noting that the R ct and R d values of the battery with the MgO-25 separator are the smallest among all of the asprepared batteries that are presented in detail in Table S2. † The signicant reduction in resistance further conrms the synergistic effect of nano-MgO and AB, where the AB conductive network facilitates electron conduction during the reactivation of the adsorbed polysuldes.
Ex situ observation of the separator and the Li-metal anode before and aer charge-discharge cycling was performed to validate the polysulde adsorption effect. The surface morphology of the MgO-25 separator ( Fig. 5a and b) and the MgO-100 separator ( Fig. 5c and d) before and aer cycling were checked by SEM. The nano-MgO particles uniformly distribute on the surface of the separators before cycling. Aer 100 cycles at a rate of 0.2 C, polysuldes deposit on the surface of the MgO-25 separator due to the strong interaction between MgO and the polysuldes. In contrast, a thick polysulde aggregation layer on the surface of the MgO-100 separator is observed aer cycling, which is due to excess nano-MgO adsorbing many more polysuldes which cannot transform into sulfur sufficiently without the aid of the AB conducting network. Fig. 5e-j show photographs of the various separators aer 5 charge-discharge cycles at a rate of 0.5 C. The surface of the separators with nano-MgO in Fig. 5e-h are dark orange, implying more polysuldes are adsorbed on the surface. However, the color of the AB decorated separator and the pristine Celgard 2400 separator appears light orange, demonstrating a weak adsorption to pol-ysuldes ( Fig. 5i and j). Surface photographs (Fig. 5e 1 -j 1 ) and SEM images (Fig. 5e 2 -j 2 ) of the Li-metal anode of the batteries are also shown. From Fig. 5e 1 and e 2 , the surface of the lithium anode in the battery with the MgO-25 separator is the smoothest, meaning that the MgO-25 separator can effectively prevent polysuldes from depositing on the Li-metal anode. Thus, the ex situ characterization results further conrm that the MgO-25 decorated separator can effectively adsorb polysuldes and suppress the shuttle effect of Li-S batteries.

Conclusions
In summary, a nano-MgO/AB decorated separator is developed to suppress the shuttle of polysulde intermediates for the rst time. Nano-MgO with the aid of an AB conductive agent can achieve a synergistic effect of excellent chemisorption and superior electronic conductivity, which can adsorb polysuldes effectively and accelerate the reaction kinetics for the transformation of polysuldes into sulfur. Moreover, it is found that the battery performance highly depends on the relative amount of nano-MgO and AB in the composite interlayer of the separator. Insufficient nano-MgO cannot adsorb polysuldes sufficiently, resulting in poor cycling stability. On the other hand, excess nano-MgO results in an increase in the non-conductive area of the battery, and polysuldes would be anchored on the separator permanently without being reused, leading to low sulfur utilization. According to our experiments, the Li-S battery with the optimal separator (MgO-25 separator, MgO and AB in the weight ratio 1 : 3) exhibits a high initial discharge capacity of 1238 mA h g À1 with high coulombic efficiency (97%) and retains a high capacity of 875 mA h g À1 aer 100 cycles at 0.2 C. This study promotes the understanding of the synergistic effect of the polysulde chemical adsorbent and the conductive agent on the suppression of the "shuttle effect" of Li-S batteries, which is also helpful to guide the design of high-performance functional separators for practical high-energy-density Li-S batteries.

Conflicts of interest
There are no conicts to declare.