Grafting polyethyleneimine on electrospun nanofiber separator to stabilize lithium metal anode for lithium sulfur batteries

Journal Pre-proofs Grafting Polyethyleneimine on Electrospun Nanofiber Separator to Stabilize Lithium Metal Anode for Lithium Sulfur Batteries Mengfei Hu, Qingyan Ma, Yuan Yuan, Yankai Pan, Mingqi Chen, Yayun Zhang, Donghui Long PII: DOI: Reference:

S1385-8947(20)30249-7 https://doi.org/10.1016/j.cej.2020.124258 CEJ 124258

To appear in:

Chemical Engineering Journal

Received Date: Revised Date: Accepted Date:

19 November 2019 13 January 2020 27 January 2020

Please cite this article as: M. Hu, Q. Ma, Y. Yuan, Y. Pan, M. Chen, Y. Zhang, D. Long, Grafting Polyethyleneimine on Electrospun Nanofiber Separator to Stabilize Lithium Metal Anode for Lithium Sulfur Batteries, Chemical Engineering Journal (2020), doi: https://doi.org/10.1016/j.cej.2020.124258

This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition of a cover page and metadata, and formatting for readability, but it is not yet the definitive version of record. This version will undergo additional copyediting, typesetting and review before it is published in its final form, but we are providing this version to give early visibility of the article. Please note that, during the production process, errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

© 2020 Published by Elsevier B.V.

Grafting Polyethyleneimine on Electrospun Nanofiber Separator to Stabilize Lithium Metal Anode for Lithium Sulfur Batteries Mengfei Hu,a Qingyan Ma, a Yuan Yuan, a Yankai Pan, a Mingqi Chen, a Yayun Zhang, a* Donghui Long ab*

a. State Key Laboratory of Chemical Engineering, East China University of Science and Technology, Shanghai, 200237, China. b. Key Laboratory of Specially Functional Polymeric Materials and Related Technology, East China University of Science and Technology, Shanghai 200237, China

Corresponding author: Yayun Zhang, Email: [email protected]

Donghui Long, Email: [email protected]

1

ABSTRACT: High-energy-density lithium sulfur (Li-S) batteries are suffering several seemingly insurmountable barriers, including lithium dendrite formation and polysulfides shuttling. Functional separator, which bridges anode, electrolyte and cathode together, has the potential to offer a perfect solution to these concerns. Herein, we develop a functional ammoniated polyacrylonitrile (PAN) nanofiber separator (APANF) which can simultaneously inhibit Li dendrite formation and polysulfides shuttling. Branched polyethyleneimine (PEI) are fixed on the electrospun PAN nanofiber mat via a chemical grafting to provide amino groups. Such strongly polar separator can well regulate the uniform Li ions distribution and induce the formation of the Li3N-rich SEI layer, resulting in an interesting 3D spherical and dendrite-free Li deposit pattern. The coulombic efficiency of resulting Li anode can be improved up to 98.8% with a low overpotential of 15 mV. Meanwhile, the separator can also serve as a block for polysulfides shutting due to the strong chemical adsorption capability of PEI, thus facilitating the capacity retention of sulfur cathode. This work provides an easy and scalable alternative to conventional polyolefin separators for solving problems in both anode and cathode of Li-S battery. Keywords: Lithium dendrites, polysulfides shuttle, lithium-sulfur batteries, electrospinning, nanofiber separator

2

Highlights: 

APANF separator is fabricated by general electrospinning process and functional group grafting.



Li ion distribution can be well regulated due to the nanofiber structures and the polar groups on APANF.



APANF separator can simultaneously suppress Li dendrite formation and polysulfides shuttling.

3

1. Introduction Development of advanced energy-storage systems for electrical vehicles and grid storage must fulfill the requirements of high energy density and long cycle lifespan [1]. As the traditional Li-ion batteries are approaching the power limits, lithium-sulfur (Li-S) batteries have provoked an animated tide of research and become one of the most potential candidates for the next-generation energy-storage devices [2,3]. However, both anode and cathode are facing a lot of issues which hinder the practical applications of Li-S batteries. Lithium metal anode possesses the highest theoretical capacity (3860 mAh g-1) and lowest potential (-3.04 V vs. standard hydrogen electrode) [4]. However, dendrite formation is the most formidable challenge for the lithium metal anode because the risks of internal short circuit and rapid capacity fading. During repetitive plating/stripping of lithium, the uncontrolled growth is unavoidable due to the huge volumetric expansion, ununiform Li ions distribution and unstable solid electrolyte interface (SEI) layer [5]. Therefore, enormous efforts have been developed to mitigate the irregular growth of lithium metal against the possible reasons above. Modifying SEI layer via adjusting the composition of liquid electrolyte [6,7] or adding functional additives, [8-10] introducing “artificial SEI” layer [11,12] or polar interlayer [13], using solid / gel electrolytes [14-16], constructing composite lithium anode with a host [17-19], and employing three-dimensional current collectors [20-22], have alleviated the dilemma of Li dendrite growth in some ways. Meanwhile, the sulfur cathode, with a high capacity of 1675 mAh g-1, is also plagued by the insulation nature of sulfur and discharge product Li2S, and the “shuttle effect” of the soluble 4

polysulfides intermediates [23,24]. In order to improve the performance of sulfur cathode, on the one hand, conductive hosts such as carbon [25-28] or conductive polymers [29] are used to promote the conductivity of the whole cathode. On the other hand, a mass of method including various encapsulation of sulfur [30,31], chemical adsorption or electrocatalysis of polysulfides [32-34], or insertion of interlayer [35-37], selective ion separator [38] and other polar separator [39-40] as a functional block, have been studied to mitigate the “shuttle effect”. However, major of present researches are only for either anode or cathode, which can hardly simultaneously suppress both dendrite growth and polysulfides shuttling. Separator which plays a key role in ion transport and influences rate performance, cell life and safety, is given many requirements [42]. Besides the basic demands of mechanical/electrochemical stability and wettability of electrolytes, functional separators are expected to be able to obstruct polysulfides and provide regular pathways for Li ions. The most common strategy is direct surface modification of traditional polyolefin (PP or PE) separators by coating or dipping with polar materials to enhance the properties of the whole separator [43,44]. However, on account of the strongly non-polar nature of polyolefin materials, the coating layer usually exhibits poor adhesion with the polyolefin matrix of separator, especially against the repetitive charge/discharge. Besides, by simple coating, the polyolefin separators can hardly promote the thermal stability, still suffering from irreversible crispation or deformation at high temperature. Also, some alternative polymer separators, such as polyvinylidene fluoride (PVDF), chitin, poly (methyl methacrylate) (PMMA), polyacrylonitrile (PAN) and etc., are studied [45-47]. Typically, electrospinning and solvent evaporation are used 5

to fabricate these separator membranes, which contributes to a series of channels or pathways for Li ion transportation. The three-dimensional membranes with abundant polar groups can improve the uptake of electrolyte, forming a reservoir for Li+, which can regulate the Li+ flux. Herein, we demonstrate a solution to simultaneously suppress lithium dendrite formation and polysulfides shuttle using an ammoniated polyacrylonitrile nanofiber separator (defined as APANF, illustrated in Fig. 1). PAN is believed to be a suitable polymer for battery separator application, due to the easy processability and excellent resistance to oxidative degradation. To give the multifunctionality, polyethyleneimine (PEI) are further fixed on the PAN via a chemical grafting. With the cross-linked nanofiber structure and the polar surface groups, Li ion distribution could be well regulated, leading to more uniform 3D spherical Li deposition pattern. Meanwhile, Li3N-rich solid electrolyte interface (SEI) layer can be realized, further inhibiting the Li dendrite formation. With such a separator, the Li anode could deliver a high of coulombic efficiency of 98.8% and a low overpotential of 15 mV. On the other hand, APANF also has strong chemical adsorption ability towards polysulfides, therefore effectively blocking the shuttling effect. As a result, the bifunctional APANF can assist the Li-S cell achieving better cycling stability and extended life span. Moreover, APANF can also extend to other Li-metal batteries like Li/NCM523 with carbonate-based electrolyte. This work provides new alternative to conventional polyolefin separators for helping meet the greater demands for the development of practical energy storage devices. 2. Experimental Section Materials. 6

Polyacrylonitrile (PAN) (Mw=150 000, Sigma Aldrich), polyethyleneimine (PEI) (branched, Mw=25 000, Sigma Aldrich), dimethyl formamide (DMF) (Adamas), ethylene glycol (Greagent). Preparation of PANF. PANF separator was fabricated by electrospinning with PAN/DMF solution. To obtain the electrospinning solution, 1.6 g PAN was dissolved into 18.4 g DMF, followed by stirring at 60 oC

for 6 h. The electrospinning solution was then loaded into a 20 mL plastic syringe equipped

with a 21-gauge stainless steel needle with a feed rate of 1.0 mL/h. During the electrospinning process, the electrospinning solution was subjected to a high voltage of 12 kV and the obtained nanofibers were collected on an aluminum collector with the distance of 12 cm from the needle tip. After electrospinning for 8 h, the PANF mat was peeled from the collector and then dried at 110 oC for 8 h. Preparation of APANF. The obtained PANF was added into 50 mL of 10 mg mL-1 PEI/ethylene glycol solution (10 mg bPEI per 1 mL ethylene glycol). The reaction mixture was refluxed with stirring at 140 oC for 8 h to get PEI-grafted-PANF (APANF). Subsequently, the APANF was taken out, washed with deionized water and ethanol for several times, and then dried finally dried in a vacuum oven at 80 oC for 12 h. To improve the mechanical strength of APANF (and PANF), the obtained membranes were treated with hot press at 140 oC. Material Characterization. Morphology analysis was carried out using scanning electron microscopy (SEM, JEOL 7

7100F). The functional group information was collected on FTIR spectrometer (Spectrum 100, PerkinElmer, America) in the frequency range of 4000-650 cm-1. X-ray photoelectron spectroscopy (XPS) was conducted using an Axis Ultra DLD X-ray photoelectron spectrometer. The X-ray source was operated at 15 kV/10 mA. The working pressure was lower than 2 ×108 Torr (1 Torr = 133.3 Pa). Electrochemical Characterization. To testify the electrochemical performance of the separators, APANF, PANF and Celgard 2500 (control sample) are cut into 19 mm disks for the following cells. For Li/Cu half cell, LIR2016 coin cells are assembled in the glove box with Ar atmospheres, using Li foil (ɸ 16 mm) as reference and counter electrode and copper foil (ɸ 15 mm) as working electrode. For better comparison, a fixed amount of 20 μL electrolyte (1 mol L-1 LiTFSI in DOL/DME) is used. Galvanostatic tests are performed on a Land CT2001A Battery Testing System (Wuhan, CHN). A fixed amount (1 mAh cm-2) of lithium was plated on the working electrode and then stripped away until the voltage up to 1 V. Current densities of 1.0, 3.0 and 5.0 mA cm-2 are employed. Electrochemical impedance spectroscopy (EIS) was carried out on an electrochemical working station (Bio-Logic VSP, FR) from 100 kHz to 0.01 Hz with a sinusoidal excitation voltage of 5 mV. Cyclic Voltammetry (CV) profiles were tested between 3.0-0 V with a scan rate of 0.2 mV s-1. Linear Sweep Voltammetry (LSV) profiles were tested from 0-6 V with a scan rate of 10 mV s-1. For Li/Li symmetrical cells, LIR2025 coin cells were assembled with two identical Li foils (ɸ 16 mm) as anode and cathode. 20 μL 1 mol L-1 LiTFSI in DOL/DME is used as electrolyte. 8

Then the cells were cycled at 1.0 or 3.0 mA cm-2 for a total capacity of 1.0 mAh cm-2 in each half cycle. For Li-S full cells, sublimed sulfur and super C were mixed by strongly milling with a mass ratio of 2:1 as cathode material. The cathode material was then made into a slurry with PVDF as binder and NMP as solvent, followed by a doctor-blade casting and dried in a 50 oC vacuum oven for 12 h. And the cathode foil is cut in 15 mm disk. LIR2025 coin cells were assembled with 50% excess of Li foil and 20 μL electrolyte (1 mol L-1 LiTFSI in DOL/DME, with 1% LiNO3). The areal sulfur loading was about 2.2 mg cm-2. DFT Computational Study. In the present study, C22H57N11 and C16H34 are employed as the model structures of PEI on APANF and PP of Celgard separator, respectively, which are presented in Fig. 5b. The choice of model structures takes account of both reliability and time cost of calculation. The N atoms exist on bPEI as primary, secondary and tertiary amine. In order to cover all kinds of amino groups, the structural formula of branched-PEI monomer shown in Fig. 5b is used, which is big enough to construct adsorption to polysulfides. And we set n=2 to choose C22H57N11 as the model. The choose of C16H34 is in a similar way. DFT calculations at M06-2x/6-311+G(2d) level were carried out to optimize the geometries of model structures and polysulfide (Li2S4) [48-50]. The Polarizable Continuum Model using the integral equation formalism variant (SMD) was also used to account for the solvent effect [51], and the 1, 4-Dioxan was adopted as the solvent to simulate the electrolyte environment of the Li-S system. The correctness of all the optimized structures were confirmed through frequency analysis and all energies of structures 9

were obtained after zero-point energy (ZPE) correction. All calculations were conducted with Gaussian 09 quantum calculation package [52]. The adsorption energy can be defined as ∆E= E(x/Li2S4) - E(x) - E(Li2S4) In which E(x/Li2S4) stands for the energy of the model compounds of polymers with Li2S4, E(x) the energy of APANF or celgard polymer model compound, and E(Li2S4) the energy of Li2S4.

3. Results and Discussions The APANF separator is fabricated via two steps including electrospinning and grafting as Fig. 2a shows. Free-standing PAN mat is firstly obtained via the electrospinning (Fig. S1) and then react with PEI in 140 oC oil bath with refluxing process. Branched PEI (bPEI) as a common macromolecular initiator has reactive end groups on the one of the ends. At 140 oC, the cyano groups (-C≡N) on PAN will open, and the reactive end groups on bPEI can trigger free radical reaction on the main chain of PAN to form grafted branches as Fig. S2 shows [53]. The asprepared APANF separator show homogenously yellow color (inset in Fig. 2b) with a high tensile strength of 19 MPa (Fig. S3). SEM image (Fig. 2b) reveals that the APANF is consisted of uniform nanofibers with average diameters of ~350 nm. The whole thickness of the membrane is controlled to ~60 μm by adjusting the spinning time (Fig. 2c). Compared to the fresh PANF, APANF have nearly invariable diameters and morphologies of nanofibers (Fig. S4), but their surface chemistry shows obvious change. FTIR spectra indicate the intensity of the peak at 2242 cm-1 (C≡N) decreases after PEI grafting, and two peaks at 1644 cm-1 and 1559 10

cm-1 appear respectively, indicating the existence of C=N and N-H groups through grafting. Thermal stability is a prerequisite of separator. The conventional Celgard separator deforms and curls into a rod irreversibly after heat treatment at 150 oC for 1 h (Fig. 2e), while APANF and PANF barely change. The improved thermal stability of APANF separator should alleviate the risk of battery fire caused by the local heat spot generation and shrinkage of separator. The wettability of separators with electrolytes is another basic property. The contact angle is tested using typical ether-based electrolyte for lithium-sulfur batteries to inspect the affinity with electrolyte. As Fig. 2f shows, the contact angle of APANF is 18.9o, much smaller than Celgard with contact angle of 48.5o. In addition, the electrolyte uptake capacities of these separators are also compared. The maximum uptake capacity of APANF is about 280%, approximately twice higher than that of Celgard (Fig. S5). The improved the wettabiliety can be expected to benefit the interfacial compatibility between separator and electrodes and enhance the Li ion transportation. The ion conductivity is evaluated using electrochemical impedance spectroscopy (EIS). At room temperature, the ion conductivity of APANF is calculated to be 0.185 mS cm-1, slightly lower than Celgard of 0.343 mS cm-1. However, the Warburg impedance of APANF is much smaller than those of PANF and Celgard via comparing the slopes of spectra in the lowfrequency stage, indicating a better diffusion of Li+ approximal to the surface of Cu foil. This result should be due to the high affinity of APANF with electrolyte, which reduces the concentration gradient of Li+ on the surface of the electrode. To estimate the functionality of APANF for lithium anode, lithium deposition behavior is 11

investigated using Li/Cu cell. 1.0 mAh cm-2 of lithium metal is plated on the Cu foil under a current density of 1.0 mA cm-2. In a control experiment, the Celgard and PANF separators are also investigated. After 50th cycling, the lithium electrode cycled with Celgard separator exhibits the uncontrolled formation of tortuous lithium dendrites with loose and porous structure (Fig. 3a). This is a typical one-dimensional dendritic growth pattern without any modification with conventional PP separator. When PANF is employed, it induces the lithium deposits into a piled but porous structure, which could consume a lot of electrolyte to form SEI layer. By contrast, APANF separator leads to an interesting 3D morphology of lithium deposition, as Fig. 3c shows, which is dense, mellow and ball-liked. The cross-section views (Fig. 3d-f) indicate the thickness of lithium deposition for APANF is ~12.5 μm, much denser than that with PANF (~18.6 μm) and Celgard (~24.5 μm). Moreover, the APANF separator maintains its smooth fibric structure after repetitive cycling, without any dead lithium on the surface (Fig. 3i). On the contrary, the surface of Celgard (Fig. 3g) and PANF (Fig. 3h) have adhesion of dead lithium particles, which will cause the irreversible capacity loss. Such striking contrast verifies that the APANF separator with amine groups has the capability to turn Li deposition behavior into a 3D growth pathway. To further understand the behavior of lithium deposition on the APANF separator, different deposition conditions are investigated. As shown in Fig. 4, homogenously spherical structure is more obvious when a small amount of lithium of 0.2 mAh cm-2 is plated. And the size of particles gets smaller from 3 μm to 1.3 μm as the current density increases from 1.0 to 3.0 mA cm-2 (Fig. 3a and 3d). When 1.0 mAh cm-2 of lithium is plated, there is still spherical structure 12

for Li deposits. Even increasing the lithium plating capacity to 3 mAh cm-2, the spherical structure is still visible, without the presence of dendrites (Fig. 4c and 4f). The reaction process on the anode is explored by cyclic voltammetry (CV) method. As Fig. S6a shows, with the Celgard, the reduction peaks at 1.42-1.23 V and at 0.9-0.62 V are corresponding to the decomposition of TFSI- anions and the DOL/DME solvent, respectively. The CV curves of cell with PANF has the similar reduction peaks. However, as for cell with APANF, the reduction voltages of both TFSI- and DOL/DME are significantly lower than those of Celgard and PANF, which suggests a hysteresis in the reduction process. That is to say, when discharge process is run from high voltage to low, more scope or time is strived for the uniform Li ion distribution, which is beneficial for large amount and uniform Li nucleation. This multipoint nucleation might lead to the spherical morphology. The X-ray photoelectron spectroscopy (XPS) is used to analyze the chemical composition of the SEI (Fig. 4g and Fig. S6b). After Ar ion etching for 250 s, the high-resolution N 1s spectra of SEI layer on the surface of lithium deposits show the presence of Li2NxOy and Li3N only for the APANF separator. Li3N is among one of the fastest Li-ion conductors with ionic conductivity on the order of 10-3~10-4 S cm-1 at room temperature [54]. It can improve the Li ion transportation through the SEI layer, and mitigate the Li ion concentration gradient on the surface of the anode. The formation of Li2NxOy and Li3N should be contributed by the active amino group from PEI, which will contribute to higher Li ion conductivity and robust SEI layer, and then mediate effectively the growth path of Li deposits. As Fig. 4h shows, with the enhancement of the ion transportation on the whole surface of Li anode, the uniform Li growth 13

can be realized. To further verify the functionality of APANF for protection of lithium metal anode, the electrochemical performance of the separators is investigated. Coulombic efficiencies (CE) of the Li/Cu half cells with APANF separator are investigated. Celgard and PANF separators are also tested as contrast samples. As Fig. 5a shows, when the current density of 1.0 mA cm-2 is applied, the CE of cell with Celgard suffers a fluctuation for 40-50th cycle, following with a thorough decay, which indicates constant breakage and reformation of SEI layer. Pure PANF separator helps the cell achieve relatively stable CE of 98.1%, but tends to decay after 80th cycle. In addition, the initial CE for the first cycle is only 91.0%, suggesting that a certain amount of lithium metal sacrifices into forming SEI layer or dead lithium. On the contrast, APANF separator assists the cell to reach a high CE of 98.8% and stable CE of 98% after 120 cycles. When larger current densities are applied, the CE of cell with APANF is higher than 97% at 3 mA cm-2 (Fig. 5b) and 93% at 5 mA cm-2 for 60 cycles (Fig. S7), while Celgard and PANF show the unstable CEs with quick fading. Furthermore, the polarization profiles are measured to reflex the difficulty level of lithium plating/stripping. Celgard and PANF show the large nucleation overpotentials and the increasing voltage hysteresis (Fig. S8). In case of APANF, the voltage hysteresis is at ~120 mV with a slight rise, indicating small barrier for the heterogeneous nucleation and lithium growth. The impedance changes with different cycle numbers are then analyzed (Fig. 5c and 5d). The impedance is fitted in Table S1 with an equivalent circuit in Fig. S9b. After the 1st cycle of discharge/charge, the impedance of cell with Celgard increases dramatically from 2.43  to 14

660 , indicating the formation of troublesome SEI layer (Fig. S9a). PANF is a little better than Celgard, with the impedance change from 14.10  of 365  for the 1st cycle. And the impedance can maintain at ~180  for the subsequent cycles. By contrast, the impedance of cell with APANF is much lower, with a mild increase from 11.90  to 140 . As the cycle number increasing to 100th, the impedance is stabilized around 120~150 , which is mainly due to the surviving from insulated “dead lithium”. Li||Li symmetrical cells are further assembled to inspect the cycling stability. The voltagetime profiles of cells at 1.0 mA cm-2 are shown in Fig. 5e. For the initial cycles, Celgard shows low polarization of ~15 mV, due to its low impedance of nucleation. After 200 h stable cycling, the voltage profile shows a gradually increase up to 300 mV and a sudden drop of overpotential at ~260 h, responding to a local short circuit. Compared with Celgard, PANF lengthens the span life of cell up to ~380 h but keeps a high overpotential of ~30 mV. In the case of APANF separator, the cell can continuously run for more than 400 h without short circuit, and with a stable overpotential of ~15 mV. At a high current density of 3.0 mA cm-2, APANF also shows long cycle life against Celgard (Fig. S10). Except for suppressing the dendritic in lithium metal anode, capability of alleviating the shuttle of polysulfides in sulfur cathode by APANF is also evaluated. A home-made visualized set-up is employed to simulate the ability of blocking the polysulfide diffusion by separating Li2S4 solution (upper side) and DOL/DME solvent (bottom side) with different separators (Fig. 6a). With the Celgard separator, the colorless DOL/DME solvent turns into yellow within 1 h and gradually becomes brown within 12 h, indicating that Li2S4 can easily permeates through 15

the commercial PP separator. When fresh PANF separator is employed, the color change of bottom side is slow than that with Celgard, but the bottom side finally becomes yellow after 6 h. APANF separator shows the best performance in blocking the Li2S4 shuttling, as no color change is observed until 12 h. To further explore that the interaction between APANF and Li2S4, DFT calculation is also employed. The binding energies of Li2S4 to PP of Celgard and PEI branches on APANF are calculated to be 0.483 eV and 1.589 eV, respectively. This result demonstrates the strong anchoring of polysulfides by APANF with formation of Li-N bonds. Therefore, it is reasonable that the strong chemical adsorption could immobilize polysulfides on the APANF separator, thus effectively suppressing shuttling, which is consistent with the results of visualized adsorption test. Since the improved adsorption ability of APANF, good electrochemical performance of LiS full cell can be expected. Here the sulfur cathode material is mechanical mixture of elemental sulfur and conductive carbon (2:1, w/w) with an areal sulfur loading of 2.2 mg cm-2 (Fig. S11). At the rate of 0.5 C (Fig. S12), the cathode with APANF separator achieves an initial capacity of 858 mAh g-1, higher than that of PANF (767 mAh g-1) and Celgard (660 mAh g-1). There is a capacity loss of cells with all three kinds of separators, mainly due to the interaction between super C and S, which is quite weak to form the binding of active material and conductive network. Though APANF can realize relative mitigation of polysulfides shuttling against Celgard, the polysulfides will still pass through the separators. However, APANF helps to maintain a highest capacity retention of ~74% after 150 cycles. At the rate of 2 C (Fig. 6c), the cell with APANF also can exhibit a considerable initial capacity of 728 mAh g-1, and a capacity 16

retention of 70.62% after 500 cycles, much higher than Celgard (51.01%) and PANF (49.83%). The galvanostatic discharge-charge profiles of cells with different separators are shown in Fig, S13. All discharge curves at various rates exhibit two typical discharge plateaus, relating two typical reduction stages of Li-S batteries. At 0.1 C, the cell with APANF possesses a higher discharge plateau at ~2.31 V, corresponding to reduction of S to higher order intermediate Li2Sn (n > 4). The longer discharge plateau at ~2.14 V is attributed to the reduction of insoluble Li2S2/Li2S. The overpotential between charge and discharge plateaus is 0.148 V, smaller than those of cells with Celgard (0.174 V) and PANF (0.149 V), which indicates a lower polarization. This difference among three separators in discharge/charge process is more obvious at 0.5 C, as cell with APANF still shows the least polarization and capacity loss. Rate performance of Li-S cell is also studied (Fig. 6d). At current densities of 0.2, 0.5, 1.0, 2.0, 5.0 and 10.0 C, APANF delivers specific capacities of 834, 670, 608, 545, 395 and 250 mAh g-1, which are much higher than those of Celgard, especially at 5.0 C and 10.0 C. It can be deduced that, in the discharging state, the Sx2- will try to move towards anode side with the effect of concentration gradient, but partly trapped by the amino groups on APANF separator. This “trapping” can be considered as adsorption and the trapping by complicated 3-dimensional nanofiber structure. In the charging state, the absorbed Sx2- will return to the cathode side with the reversed electric field force. It should be noticed that the functionality of APANF separator is not to tightly grasp all polysulfides, but to retard their dissolution in the electrolyte. EIS is also analyzed for cells before and after cycling. The semicircles in the high- and medium-frequency regions correspond to resistance of the surface films (Rf) and the charge17

transfer at electrode-electrolyte interface (Rct), respectively (listed in Table S2). Before cycling, the impedance of cell with three different separators are similar around 140  (Fig. S14a). After 50th cycling, the Rf and Rct for cell with APANF are obvious smaller than that of Celgard and PANF (Fig. S14b), indicating improved interfacial properties, which is due to the reducing of corrosion of lithium by polysulfides. After the 50th discharge/charge cycles, the Li anode is disassembled from the Li-S cells (Fig. S15) to observe the surface morphology of Li. In the existence of Li2Sx, the Li anode with APANF is smooth plate-like, while the Li anode with Celgard was rough and mossy. It can be concluded that the effects of APANF on both anode and cathode promotes the performance of Li-S together. Besides, the APANF can also be applied in the carbonate-based system. When APANF separator is employed in the Li||NCM523 cells using EC/DEC/EMC electrolyte, the capacity retention after 150 cycles can be enhanced to 65% compared with that of Celgard (Fig. S16a). The rate performance with APANF is also improved against Celgard (Fig. S16b). These results reflect that the APANF separator has general applicability for lithium metal batteries.

4. Conclusion In summary, PEI grafted PAN nanofiber separator is demonstrated to play a role in both enabling dendrite-free lithium plating on anode and suppressing shuttling of polysulfides from sulfur cathode. The strong polar ammonia groups provided by PEI branches improve the affinity with electrolyte and uniform the Li ion distribution. And the free ammonia groups can contribute to the Li3N-rich SEI layer formation, thus leading to a spherical morphology of 18

lithium deposition. As a result, with APANF separator, the Coulombic efficiency can be stabilized at 98% for 120 cycles. And the symmetrical cell with APANF can reach a long lifespan over 800 h. Meanwhile, the intricate branched structure of APANF can also block polysulfides intermediates in Li-S battery, which profits by the strong adsorption effect of the PEI branches. With such positive influence on both anode and cathode, the cycling and rate performance of Li-S cell can be improved with APANF separator. This work illustrates an easy and win-win solution for lithium sulfur battery and could be also applicable for other lithium metal batteries.

5. Acknowledgements This work was partly supported by National Science Foundation of China (No. 21878091 and No. 21576090), and Fundamental Research Funds for the Central Universities (222201718002).

References [1] Manthiram, A.; Chung, S.; Zu, C. Adv. Mater. 27 (2015) 1980. [2] Goodenough, J. B.; Kim, Y. Chem. Mater. 22 (2010) 587-603. [3] Rosenman, A.; Markevich, E.; Salitra, G.; Aurbach, D.; Garsuch, A.; Chesneau, F. F. Adv. Energy Mater. (2015) 1500212. [4] Cao, R.; Xu, W.; Lv, D.; Xiao, J.; Zhang, J.-G. Adv. Energy Mater. (2015) 1402273. [5] Li, S.; Jiang, M.; Xie, Y.; Xu, H.; Jia, J.; Li, J. Adv. Mater. (2018) 1706375. [6] Yu, H.-L.; Zhao, J.-N.; Ben, L.-B.; Zhan, Y.-J.; Wu, Y.-D.; Huang, X.-J. ACS Energy Lett. 2 (2017) 1296. 19

[7] Carbone, L.; Gobet, M.; Peng, J.; Devany, M.; Scrosati, B.; Greenbaum, S.; Hassoun, J. J. Power Source 299 (2015) 460. [8] Li, W.; Yao, H.-B.; Yan, K.; Zheng, G.-Y.; Liang, Z.; Chiang, Y.-M.; Cui, Y. Nat. Commun. 6 (2015) 7436. [9] Zhang, X.-Q.; Cheng, X.-B.; Chen, X.; Yan, C.; Zhang, Q. Adv. Funct. Mater. (2017) 1605989. [10] Ye, H.; Yin, Y.-X.; Zhang, S.-F.; Shi, Y.; Liu, L.; Zeng, X.-X.; Wen, R.; Guo, Y.-G.; Wan, L.-J. Nano Energy, 26 (2017) 411. [11] Zhao, J.; Liao, L.; Shi, F.; Lei, T.; Chen, G.; Pei, A.; Sun, J.; Yan, K.; Zhou, G.; Xie, J.; Liu, C.; Li, Y.; Liang, Z.; Bao, Z.; Cui, Y. J. Am. Chem. Soc. 139 (2017) 11550. [12] Liang, X.; Pang, Q.; Kochetkov, I. R.; Sempere, M. S.; Huang, H.; Sun, X.; Nazar, L. F. Nat. Energy 2 (2017) 17119. [13] Cheng, X.-B.; Hou, T.-Z.; Zhang, R.; Peng, H.-J.; Zhao, C.-Z.; Huang, J.-Q.; Zhang, Q. Adv. Mater. 28 (2016) 2888-2895. [14] Han, X.; Gong, Y.; Fu, K.; He, X.; Hitz, G. T.; Dai, J.; Pearse, A.; Liu, B.; Wang, H.; Rubloff, G.; Mo, Y.; Thangadurai, V.; Wachsman, E. D.; Hu, L. Nat. Mater. 16 (2017) 572. [15] Zhao, C.-Z.; Zhang, X.-Q.; Cheng, X.-B.; Zhang, R.; Xu, R.; Chen, P.-Y.; Peng, H.-J.; Huang, J.-Q.; Zhang, Q. Proc. Natl. Acad. Sci. 114 (2017) 11069-11074. [16] Lu, Q.; He, Y. B.; Yu, Q.; Li, B.; Kaneti, Y. V.; Yao, Y.; Kang, F.; Yang, Q. H. Adv. Mater. 29 (2017) 1604460. [17] Zhang, R.; Chen, X.; Shen, X.; Zhang, X.-Q.; Chen, X.-R.; Cheng, X.-B.; Yan, C.; Zhao, C.-Z.; Zhang, Q. Joule 2 (2018) 764-777. [18] Hu, M.; Yuan, Y.; Liu, Y. Tian, L. Zhang, Y.-Y. Long, D.-H. Electrochim. Acta 302 (2019) 301-309. [19] Cheng, Y.-F.; Ke, X.; Chen, Y.-M.; Huang, X.-Y.; Shi, Z.-C.; Guo, Z.-P. Nano Energy 63 20

(2019) 103854. [20] Zhao, H.; Lei, D.; He, Y.-B.; Yuan, Y.; Yun, Q.; Ni, B.; Lv, W.; Li, B.; Yang, Q.-H.; Kang, F.; Lu, J. Adv. Energy Mater. (2018) 1800266. [21] Hu, M.; Yuan, Y.; Guo, M.; Pan, Y.; Long, D.-H. J. Mater. Chem. A 6 (2018) 14910. [22] Liu, S.; Xia, X.; Zhong, Y.; Deng, S.; Yao, Z.; Zhang, L.; Cheng, X.-B.; Wang, X.; Zhang, Q.; Tu, J. Adv. Energy Mater. (2017) 1702322. [23] Zhao, Y.; Liu, M.; Lv, W.; He, Y. B.; Wang, C.; Yun, Q. B.; Liu, B. H.; Kang, F. Y.; Yang, Q. H. Nano Energy 30 (2016) 1-8. [24] Wang, H. Q.; Zhang, W. C.; Xu, J. Z.; Guo, Z. P. Adv. Funct. Mater. 28 (2018) 1707520. [25] Bin, D. S.; Chi, Z. X.; Li, Y. T.; Zhang, K.; Yang, X. Z.; Sun, Y. G.; Piao, J. Y.; Cao, A. M.; Wan, L. J. J. Am. Chem. Soc. 139 (2017) 13492-13498. [26] Wang, Z.-J.; Gao, H.; Zhang, Q.; Liu, Y.-Q.; Chen, J.; Guo, Z.-P. Small 15 (2018) 1803858. [27] Luo, C.; Zhu, H.; Luo, W.; Shen, F.; Fan, X.; Dai, J.; Liang, Y.; Wang, C.; Hu, L. ACS Appl. Mater. Interfaces 9 (2017)14801. [28] Wei, Y.; Tao, Y.; Zhang, C.; Wang, J.; Qiao, W.; Ling, L.; Long, D. Electrochim. Acta 188 (2016) 385. [29] Fu, Y.; Manthiram, A. J. Phys. Chem. C 116 (2012) 8910-8915. [30] Long, B.; Qiao, Z.; Zhang, J.; Zhang, S.; Balogun, M.-S.; Lu, J.; Song, S.; Tong, Y. J. Mater. Chem. A 7 (2019) 11370-11378. [31] Yu, M.; Ma, J.; Xie, M.; Song, H.; Tian, F.; Xu, S.; Zhou, Y.; Li, B.; Wu, D.; Qiu, H.; Wang, R. Adv. Energy Mater. (2017) 1602347. [32] Tao, Y.; Wei, Y.; Liu, Y.; Wang, J.; Qiao, W.; Ling, L.; Long, D. Energy Environ. Sci. 9 (2016) 3230. [33] Wei, Y.; Kong, Z.; Pan, Y.; Cao, Y.; Long, D.; Wang, J.; Qiao, W.; Ling, L. J. Mater. Chem. A 6 (2018) 5899 21

[34] Lin, H.; Yang, L.; Jiang, X.; Li, G.; Zhang, T.; Yao, Q.; Zheng, G.; Lee, J.-Y. Energy Environ. Sci. 10 (2017) 1476. [35] Liu, M.; Li, Q.; Qin, X.; Liang, G.; Han, W.; Zhou, D.; He, Y.-B.; Li, B.; Kang, F. Small (2017) 1602539. [36] Fan, Y.; Yang, Z.; Hua, W.; Liu, D.; Tao, T.; Rahman, M. M.; Lei, W.; Huang, S.; Chen, Y. Adv. Energy Mater. 7 (2017) 1602380. [37] Zhang, J.; Ma, W.-Z.; Feng, Z.-Y.; Wu, F.-F.; Wei, D.-H.; Xi, B.-J.; Xiong, S.-L. J. Energy Chem. 39 (2019) 54-60. [38] Huang, J.-Q.; Zhang, Q.; Peng, H.-J.; Liu, X.-Y.; Qian, W.-Z.; Wei, F. Energy Environ. Sci. 7 (2014) 347-353. [39] Lai, Y.; Wang, P.; Qin, F.; Xu, M.; Li, J.; Zhang, K.; Zhang, Z. Energy Storage Mater. 9 (2017) 179. [40] Qiu, Z.-F.; Shi, L.-Y.; Wang, Z.-Y.; Mindemarkc, J.; Zhu, J.-F.; Kristina, E.; Zhao, Y.; Yuan, S. Chem. Eng. J. 368 (2019) 321-330. [41] Shaibani, M.; Akbari, A.; Sheath, P.; Easton, C. D.; Banerjee, P. C.; Konstas, K.; Fakhfouri, A.; Barghamadi, M.; Musameh, M. M.; Best, A. S.; Rüther, T.; Mahon, P. J.; Hill, M. R.; Hollenkamp, A. F.; Majumder, M. ACS Nano 10 (2016) 7768. [42] Zhang, W.; Tu, Z.; Qian, J.; Choudhury, S.; Archer, L. A.; Lu, Y. Small 14 (2018) 1703001. [43] Ryou, M.-H.; Lee, D. J.; Lee, J.-N.; Lee, Y. M.; Park, J.-K.; Choi, J. W. Adv. Energy Mater. 2 (2012) 645. [44] Zhuang, T.-Z.; Huang, J.-Q.; Peng, H.-J.; He, L.-Y.; Cheng, X.-B.; Chen, C.-M.; Zhang, Q. Small 12 (2016) 381-389. [45] Zhai, Y.; Wang, N.; Mao, X.; Si, Y.; Yu, J.; Al-Deyab, S. S.; El-Newehy, M.; Ding, B. J. Mater. Chem. A 2 (2014)14511. [46] Kim, J. K.; Kim, D. H.; Joo, S. H.; Choi, B.; Cha, A.; Kim, K. M.; Kwon, T. H.; Kwak, S. 22

K.; Kang, S. J.; Jin, J. ACS Nano 11 (2017) 6114-6121. [47] Zhu, J.; Chen, C.; Lu, Y.; Zang, J.; Jiang, M.; Kim, D.; Zhang, X. Carbon 101 (2016) 272280. [48] Krishnan, R.; Binkley, J. S.; Seeger, R.; Pople, J. A. J. Chem. Phys. 72 (1980) 650. [49] Mclean, A. D.; Chandler, G. S. J. Chem. Phys. 72 (1980) 5639. [50] Chen, J. J.; Yuan, R. M.; Feng, J. M.; Zhang; Q.; Huang, J. X.; Fu, G.; Zheng, M. S.; Ren, B.; Dong, Q. F. Chem. Mater. 27 (2015) 2048. [51] Marenich, A. V.; Cramer, C. J.; Truhlar, D. G. J. Phys. Chem. B 113 (2009) 6378. [52] M.J. Frisch, G.W. Trucks, H.B. Schlegel, G.E. Scuseria, M.A. Robb, J.R. Cheeseman, G. Scalmani, V. Barone, B. Mennucci, G.A. Petersson, H. Nakatsuji, M. Caricato, X. Li, H.P. Hratchian, A.F. Izmaylov, J. Bloino, G. Zheng, J.L. Sonnenberg, M. Hada, M. Ehara, K. Toyota, R. Fukuda, J. Hasegawa, M. Ishida, T. Nakajima, Y. Honda, O. Kitao, H. Nakai, T. Vreven, J.A. Montgomery Jr., J.E. Peralta, F. Ogliaro, M. Bearpark, J.J. Heyd, E. Brothers, K.N. Kudin, V.N. Staroverov, T. Keith, R. Kobayashi, J. Normand, K. Raghavachari, A. Rendell, J.C. Burant, S.S. Iyengar, J. Tomasi, M. Cossi, N. Rega, J.M. Millam, M. Klene, J.E. Knox, J.B. Cross, V. Bakken, C. Adamo, J. Jaramillo, R. Gomperts, R.E. Stratmann, O. Yazyev, A.J. Austin, R. Cammi, C. Pomelli, J.W. Ochterski, R.L. Martin, K. Morokuma, V.G. Zakrzewski, G.A. Voth, P. Salvador, J.J. Dannenberg, S. Dapprich, A.D. Daniels, O. Farkas, J.B. Foresman, J.V. Ortiz, J. Cioslowski, D.J. Fox Gaussian 09, Revision D.01 Gaussian, Inc., Wallingford CT 2013. [53] Zhao, R.; Li, X.; Sun, B.; Li, Y.; Li, Y.; Yang, R.; Wang, C. J. Mater. Chem. A 5 (2017) 1133. [54] Ma, G.; Wen, Z.; Wu, M.; Shen, C.; Wang, Q.; Jin, J.; Wu, X. Chem. Commun. 50 (2014) 14209. 23

Fig. 1. Illustration of Li-S batteries with different separators. (a) Conventional PP (Celgard) separator, (b) APANF separator.

24

Fig. 2. (a) A schematic of preparation of APANF. (b) Top-view and (c) cross-section SEM images of APANF. The inset in Fig. 1b is digital picture of APANF. (d) FTIR spectra of PANF and APANF. Comparisons of (e) thermal shrinkage and (f) contact angle of separators. (g) Ion conductivity information of separators.

25

Fig. 3. Top-view and cross-section SEM images of Li plating on Cu foil with (a, d) Celgard, (b, e) PANF and (c, f) APANF separators after 50 cycles. The morphology of (g) Celgard, (h) PANF, (i) APANF, after cycling. The insets are the high-resolution images.

26

Fig. 4. (a-f) SEM images of lithium deposits with APANF separator under current densities of 1.0 and 3.0 mA cm-2 for total capacity of 0.2, 1.0 and 3.0 mAh cm-2. (g) X-ray photoelectron spectroscopy (N 1s peak) of SEI layer after 250 s etching. (h) The illustration of APANF for regulating Li ion flux and uniform Li metal anode.

27

Fig. 5. Coulombic efficiencies of Li/Cu half cell with different separators at current densities of (a) 1.0 and (b) 3.0 mA cm-2. Electrochemical impedance spectra (EIS) of (c) PANF and (d) APANF separators at initial state and after cycling. (e) Voltage profiles of symmetrical cells with different separators.

28

Fig. 6. (a) Visualized test of polysulfides adsorption test with separators. (b) Molecular models of interaction between PANF and Li2S4 by DFT calculations. (c) Galvanostatic cycling performance of Li-S cells with different separators at a current density of 2C. (d) Rate performance of Li-S cells with Celgard and APANF.

29

Graphical abstract:

Highlights: 

APANF separator is fabricated by general electrospinning process and functional group grafting.



Li ion distribution can be well regulated due to the nanofiber structures and the polar groups on APANF.



APANF separator can simultaneously suppress Li dendrite formation and polysulfides shuttling.

Declaration of Interest Statement The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

30