Xingmei Guo, Jinfeng Xie, Jing Wang, Shangqing Sun, Feng Zhang, Fu Cao, Yuanjun Liu,Xiangjun Zheng, Junhao Zhang,*, Qinghong Kong

1 School of Environmental and Chemical Engineering, Jiangsu University of Science and Technology, Zhenjiang 212003, China

2 Key Laboratory for Advanced Technology in Environmental Protection of Jiangsu Province, Yancheng Institute of Technology, Yancheng 224051, China

3 School of the Environment and Safety Engineering, Jiangsu University, Zhenjiang 212013, China

Keywords:TiO2/N-CNFs interlayer Composites Microstructure Adsorption Lithium-sulfur battery Shuttle effect

ABSTRACT ‘‘Shuttle effect”is detrimental for maintaining the high capacity and cycling reversibility of lithium-sulfur batteries (LSBs).To inhibit polysulfide migration, N-doped carbon nanofibers (N-CNFs) membrane comprising TiO2 nanoparticles (TiO2/N-CNFs) is fabricated using an electrospinning-calcination method and further applied as interlayer in LSBs.The TiO2/N-CNFs interlayer helps the battery to deliver a high specific capacity of 1155.2 mA?h?g-1 at 0.2 C with high Coulombic efficiency,good rate capability and stability.When cycling at 0.5 C,a capacity retention rate of 62.4%is achieved over 300 cycles,which is higher than that of CNFs and TiO2/CNFs counterparts.The excellent performance should mainly be attributed to the alleviated‘‘shuttle effect”deriving from high polysulfide trapping ability of TiO2 nanoparticles and N heteroatoms in interwoven CNFs.

1.Introduction

The surging power demands in modern life raises huge pressure on energy storage devices,including lithium/sodium-ion batteries,supercapacitors,fuel cells,etc[1-4].Lithium-sulfur batteries(LSBs)show great prospects due to the advantages of low price and environmental friendliness [5,6].More importantly, the theoretical specific capacity of sulfur is up to 1675 mA?h?g-1, which leads to extremely high theoretical energy density (2600 W?h?kg-1) of the device [7].During charging/discharging process, sulfur successively and reversibly transforms to soluble polysulfides and unsoluble Li2S on the cathode.Besides the insulation nature of sulfur and Li2S, the migration of polysulfides and their side reactions with lithium anodes also lead to low Coulomb efficiency and severe attenuation of capacity, which is termed as ‘‘shuttle effect” [8–10].Hence, solving the problems of low conductivity and ‘‘shuttle effect”is essential for promoting the energy storage performances of LSBs.

Previous efforts are mainly focused on the design and synthesis of host materials for advanced sulfur cathode.For example, carbonaceous materials, conductive polymers and metal compounds with various nanostructures have been developed as sulfur hosts and acted as cathode materials for LSBs [11–15].These materials can physically or chemically trap sulfur species and alleviate the‘‘shuttle effect” of polysulfides.However, the synthesis of nanostructured hosts and following sulfur loading usually involve tedious procedures.And, in many cases, merely cathode design can’t completely eradicate the diffusion of polysulfides and may even reduce the actual sulfur content in the cathodes.Hence,efforts on exploring additional methods to prevent polysulfide migration and avoid side reactions have never stopped, including anode protection [16], electrolyte optimization [17], separator coating,etc[18].

Suetal.[19]proposed a new configuration with an extra interlayer inserted between the cathode and the separator of LSBs,which has been proved to be a convenient and effective approach for immobilizing active species on the cathode side.Considering the structural flexibility and diversity, carbonaceous materials including carbon nanotubes, graphene and porous carbon paper have been synthesized and applied as LSB interlayers [20–22].Although good Li+conductivity and depressed‘‘shuttle effect”have been achieved, these interlayers inhibit polysulfide migration through the physical barrier of carbonaceous networks, which leaves a great space for improvement.So far, various methods including heteroatom doping and defect engineering have been employed to bring more active sites for adsorbing and trapping polysulfides[23,24].What’s more,as polar oxides can interact with polysulfides through more efficient chemical adsorption, it is a promising design to add oxides in carbonaceous interlayer to further promote the polysulfide adsorption capability and alleviate the detrimental ‘‘shuttle effect” in LSBs.Currently, attempts on applying SnO2,WO3,ZnO and TiO2as active species in LSB interlayers have been reported [25–29]; however, the approaches for further increasing their dispersity in carbonaceous matrix and coupling with other active sites (e.g.heteroatoms) to obtain more advanced LSB interlayers are still lack of investigation.

Herein, TiO2/N-CNFs was fabricatedviaelectrospinning followed by a two-step calcination process.The interwoven CNFs highly dispersed with TiO2and N heteroatoms provide sufficient active sites for trapping polysulfides, and meanwhile guarantee efficient transfer of Li+.When applying as interlayer for LSBs, the‘‘shuttle effect” was greatly alleviated, leading to excellent cycling stability and good rate capability.Although the application of TiO2based composites in LSBs has been reported in literature, most works focus on employing them as sulfur hosts in cathodes.It is a new attempt to disperse TiO2nanoparticles and N heteroatoms uniformly in CNFs through electrospinning and further apply them as additional interlayer in LSBs for enhancing energy storage behaviors.This work provides an exemplary and inspiring strategy for designing high-performance interlayers of LSBs through combining polar oxides and heteroatoms in carbon matrix.

2.Experimental

2.1.Synthesis of TiO2/N-CNFs interlayer

The TiO2/N-CNFs interlayer for LSBs was prepared through an electrospinning followed by calcination process.Before electrospinning, a precursor solution was prepared by following steps: Firstly,0.25 g urea was put into a 40 ml weighing bottle,followed by adding 6 mlN,N-dimethylformamide(DMF),4.8 ml ethanol and 0.9 g glacial acetic acid under constant stirring.After that, 2.25 g tetrabutyl titanate was added dropwise and stirred for 5 min.Then, 1.125 g polyvinylpyrrolidone (PVP, molecular weight 1300000) and 1.6 g acetic acid were added successively with 5 min stirring in each step.After sealing the weighing bottle and stirred at 40 °C overnight, a clear solution was obtained and ready for electrospinning.

In the electrospinning process, the as-prepared solution was poured into a 20 ml syringe with a needle diameter of 0.51 mm and fixed on the syringe pump of the electrospinning equipment.The propulsion speed was set as 0.6 ml?h-1and the rolling receiver was placed 15 cm away with a rolling speed of 400 r?min-1.Under a high voltage of 18 kV, a nanofiber membrane was formed on the receiver, which was peeled off and dried at 60 °C for 30 min.After that, the membrane was cut into 10 cm × 4 cm sheets and placed between two alumina plates for further calcination.Firstly,the membrane was pre-oxidized at 250°C for 2 h in air atmosphere.Then,the sample was transferred to a tube furnace, heated to 600 °C with a heating rate of 1 °C?min-1and kept for 2 h under N2atmosphere.After cooling to room temperature, the final product was obtained and named as TiO2/N-CNFs.For comparison, CNFs and TiO2/CNFs membranes were also prepared with the same procedures, except that only PVP or PVP with glacial acetic acid and tetrabutyl titanate was used as precursor solutions for electrospinning.

2.2.Preparation of C/S cathode

C/S composite was prepared by the melt diffusion method.Firstly,carbon black and sulfur were mixed and ground in a mortar with a mass ratio of 2:8.After that, the mixture was transferred into a 1 ml ampoule bottle and filled with inert gas before sealing.The sealed ampoule bottle was then placed in the oven and kept at 155 °C for 12 h to obtain C/S composite with a sulfur content of～80%, which can be quantitatively confirmed using thermogravimetric analysis(TGA)as shown in Fig.S1(in Supplementary Material).For cathode preparation, the C/S composite was mixed with polyvinylidene fluoride and acetylene black with a mass ratio of 8:1:1 and ground in a mortar for 30 min, after whichN-methyl-2-pyrrolidone as solvent was dropped into the mixture to obtain an uniform slurry after stirring for 3 h.The slurry was coated on an Al foil and vacuum dried at 80 °C overnight before cutting into circular cathodes for LSBs with a diameter of 13 mm.

2.3.Material characterization

Field emission scanning electronic microscope (FESEM, ZEISS Merlin Compact) and high-resolution transmission electronic microscope (HRTEM, JEOL JEM-2100F) with energy dispersive Xray (EDX) detector were used to observe the microstructure and elemental distribution of the samples.Powder X-ray diffractor(XRD, Shimadzu XRD-6000, Cu Kα radiation, λ=0.15418 nm) and Raman spectroscope(Thermo Fischer DXR,excitation wavelength:514 nm) were used to investigate the chemical composition and graphitization degree of the samples.For TEM and XRD characterization, the film was grinded into powder before testing.

2.4.Battery assembly and electrochemical test

2032 type button cells were assembled in an Ar-filled glovebox using 1 mol?L-1LiN(CF3SO2)2and 0.1 mol?L-1LiNO3in the solvent of 1, 3-dioxolane/1, 2-ethylene glycol dimethyl ether (DOL/DME,1:1 volume ratio) as electrolyte, and the amount of electrolyte in a button cell is about 15 μl?mg-1.Li foil with a diameter of 15.6 mm and thickness of 0.45 mm, C/S coated Al foil and Celgard 2600 micro-porous polypropylene membrane were employed as anode, cathode and separator, respectively.Moreover, an extra interlayer(CNFs,TiO2/CNFs or TiO2/N-CNFs)was inserted between cathode and separator as polysulfide inhibitor in this work.Cyclic voltammetry (CV) was conducted on CHI 670D electrochemical workstation with a scanning rate of 0.2 mV?s-1.Galvanostatic tests were carried out on a LAND CN-2001A battery system in the voltage range of 1.7–2.8 Vvs.Li+/Li.Electrochemical impedance spectroscopy (EIS) was performed at open circuit voltage with an amplitude of 10 mV and a frequency range of 105Hz to 0.1 Hz.

2.5.Polysulfide trapping measurement

The trapping capability for polysulfide was measured using a Hshaped glassware with two chambers divided by an interlayer.One chamber was filled with the deep brown solution of lithium polysulfide (Li2S6), which was prepared by mixing 400 mg Li2S and 23 mg S in 25 ml DOL/DME(1:1 volume ratio)followed by stirring at 50°C for 12 h.The other chamber was filled with clear DOL/DME solvent.CNFs,TiO2/CNFs and TiO2/N-CNFs were employed as interlayers to separate the two chambers.The color transition was compared after standing for 0, 0.5, 12 and 24 h, respectively.

Fig.1. (a)Fabrication process of TiO2/N-CNFs;(b,c)FESEM images of TiO2/N-CNFs;(d)-(g)HRTEM and lattice images of TiO2/N-CNFs;(h)Elemental mapping images of C,N,O and Ti in TiO2/N-CNFs.

3.Results and Discussion

Fig.1(a) shows the photographs of TiO2/N-CNFs, as well as its precursor and intermediate samples in the fabrication process.The black membrane as the final TiO2/N-CNFs product shows good flexibility and high strength with a thickness of ～30 μm (Figs.S2 and S3).As shown by the FESEM images in Fig.1(b) and (c), the membrane is composed of interwoven nanofibers with a diameter in the range of 50–200 nm and the surface of the fibers is not smooth due to the presence of TiO2particles.This feature is more clearly exhibited in TEM images, in which many ultra-small nanoparticles disperse in the fiber matrix (Fig.1(d) and (e)).The crystal lattices can be distinguished under higher resolution.As shown in Fig.1(f) and (g), the interplanar spacings of 0.352 and 0.243 nm match well with the (101) and (103) planes of TiO2,respectively,proving that the distributed nanoparticles are mainly TiO2crystallites.Fig.1(h) displays the elemental mapping images of TiO2/N-CNFs.It is obvious that C, N, O, Ti elements are evenly distributed in the fibric region.These results verify that TiO2/NCNFs were successfully fabricated with TiO2particles and N atoms uniformly dispersed in the CNFs matrix.However, for TiO2/CNFs without the addition of urea in the preparation process, although FESEM image shows similar fibrous structure with that of TiO2/N-CNFs(Fig.S4(a)),TEM image in Fig.S4(b)indicates that TiO2particles are bigger and not evenly distributed in the fibers for TiO2/CNFs.This indicates that urea plays an important role in preventing the coarsening and aggregation of TiO2particles in TiO2/N-CNFs,which is due to the coordination effect of urea with titanium species.

XRD was used to further characterize the composition of different samples.As shown in Fig.2(a),all XRD patterns exhibit a broad peak at around 23° which corresponds to the (002) plane of the partially graphitized region in amorphous carbon [30,31].For TiO2/CNFs and TiO2/N-CNFs, additional diffraction peaks at 25.4°,48.3° and 54.1° appear, which derive from the (101), (200) and(105) planes of anatase TiO2(JCPDS#21-1272), respectively.It is noteworthy that the intensity of TiO2peaks for TiO2/N-CNFs is lower than that of TiO2/CNFs.Based on the peak width at half height using Scherrer equation, the grain sizes in TiO2/CNFs and TiO2/N-CNFs are estimated to be 18.6 and 11.4 nm, respectively.This may be because the addition of urea coordinated and diluted titanium precursor,leading to smaller sizes and more uniform distribution of TiO2particles in TiO2/N-CNFs.This coincides well with the TEM images.To investigate the graphitization degree of carbon matrix, Raman spectra were fitted and compared in Fig.2(b).All curves have two typical bumps at 1360 and 1580 cm-1, which relate to the disordered structure related D-band and ordered/-graphitized structure related G-band, respectively.Hence, a lower integral area ratio of these two bands (ID/IG) generally means higher graphitization degree of carbonaceous materials [32-34].Compared to CNFs, theID/IGvalue of TiO2/CNFs decreases from 2.3 to 2.1.This is because the presence of titanium which have strong interactions with carbon atoms can catalyze and promote the graphitization process through carbide transformation mechanism [35].However, for TiO2/N-CNFs, the value increases back to 2.3, which may be because the presence of N heteroatoms causes more defects and disorder in the carbon structure.

Fig.2. XRD (a) and Raman (b) patterns of CNFs, TiO2/CNFs and TiO2/N-CNFs.

Fig.3(a) shows the CV curves of the as-assembled LSB with TiO2/N-CNFs as the interlayer between cathode and separator.As the electrode is not fully activated at the initial state, there is a small deviation between the plot of the first cycle and the following cycles.After fully activation,the good coincidence of CV curves indicating good reversibility and cycling stability of the battery with TiO2/N-CNFs interlayer.The two cathodic peaks at around 2.2 and 1.96 V correspond to the successive reduction of S8to long-chain (Li2Sn, 4 ≤n≤ 8) and short-chain sulfides (Li2Sn,1 ≤n≤4); while the anodic peak at about 2.5 V corresponds to the reverse process of polysulfides converting to S8[27].In addition,as shown in(Fig.S5),the CV current obviously decreases with the increase of cycle numbers for CNFs and TiO2/CNFs separated LSBs, indicating inferior reversibility and stability of the battery.This demonstrates that TiO2/N-CNFs is a better interlayer than CNFs and TiO2/CNFs to alleviate the ‘‘shuttle effect” of LSBs.Fig.3(b)shows the charge-discharge curves with the first cycle activated at 0.05 C(1 C=1675 mA?g-1)and the following cycles recorded at 0.2 C for TiO2/N-CNFs battery.The stable discharging and charging platforms are roughly consistent with the CV curves.A high specific discharge capacity of 2209.6 mA?h?g-1is exhibited at the first cycle of 0.05 C,and the subsequent three cycles are in good coincidence,which again verify the good reversibility and cycling stability.It is noteworthy that the initial specific discharge capacity at 0.05 C is higher than the theoretical specific capacity(1675 mA?h?g-1).This is mainly because side reaction happens in the first cycle, and the generation of cathode electrolyte interface(CEI)film on the surface of sulfur causes partial irreversible capacity.Moreover, the disordered carbon structure in TiO2/N-CNFs interlayer also causes some additional lithium storage intercalation [36–38].

To investigate the rate capability of LSBs with different interlayers, cycling tests were conducted at different current densities.As shown in Fig.3(c), the TiO2/N-CNFs battery delivers specific discharge capacities of 1250.9, 1113, 993.2, 875.5 and 734 mA?h?g-1at 0.1,0.2,0.5,1.0 and 2.0 C,respectively.More importantly,when the current density was changed back to 0.5, 0.2 and 0.1 C, the specific discharge capacity can still be restored to 813.1, 918.9,and 947.1 mA?h?g-1, respectively, indicating that the battery with TiO2/N-CNFs interlayer has excellent rate capability.As shown in(Fig.S6), the specific discharge capacities of CNFs and TiO2/CNFs are 908.1 and 496.0 mA?h?g-1at 0.1 C, respectively.However, at a high rate of 1 C, the values decrease to 406.5 and 15.2 mA?h?g-1,indicating worse rate capability than TiO2/N-CNFs battery.One important reason is that the TiO2/N-CNFs interlayer effectively inhibits the diffusion of polysulfides and ensures the stability of battery during cycling tests.However, as displayed in Fig.3(c),the capacity drops to 313.9 mA?h?g-1, when the current density further rises to 5 C.This indicate that the battery is severely polarized at 5 C and the anchoring effect of TiO2/N-CNFs on polysulfides may not fully take effect at excessive current densities.To allow smooth activation of the battery and ensure sufficient contact between electrode materials and electrolyte, batteries are often cycled at a relative low rate before high-rate operations.Fig.3(d)presents the cycling performance and Coulombic efficiency of the battery at 0.2 C.After 1 cycle at 0.05 C for activation,the battery’s initial specific discharge capacity at 0.2 C is 1499.8 mA?h?g-1and the capacity retains 1155.2 mA?h?g-1after 70 cycles with a retention rate of 77.0%.The Coulombic efficiency is 96.38%at the initial 0.2 C cycling and keeps almost 100%in the subsequent cycles,suggesting excellent cycling reversibility of the battery.This indicates that the‘‘shuttle effect”in the as-assembled battery is not serious,thanks to the inhibiting effect of TiO2/N-CNFs interlayer to polysulfides.Moreover, the cross-sectional SEM image of TiO2/N-CNFs after cycling test is shown in Fig.S7.The thickness of the interlayer swells to ～40 μm due to the infiltration of electrolyte during charging/discharging; whereas, the structure remains stable,which is critical for retaining the cycling stability of the battery.

To identify the real active sites in TiO2/N-CNFs, the long-term cycling performances of LSBs with CNFs, TiO2/CNFs and TiO2/NCNFs as interlayers were all tested and compared.As shown in Fig.4(a), neglecting the activating process in the first 3 cycles,the initial capacity for CNFs containing LSB at 0.5 C is 619.1 mA?h?g-1, which decreases to 369.5 mA?h?g-1with a capacity retention of 59.7% after 300 cycles.For TiO2/CNFs, the specific capacity and the retention ratio after 300 cycles is 653.1 mA?h?g-1and 60.8%,respectively, indicating great contribution of TiO2on promoting the capacity and stability of LSBs.This is because the presence of polar TiO2provides sufficient active sites for adsorbing polysulfides and hence greatly suppresses the ‘‘shuttle effect” [39].When N atoms are further doped into the CNFs matrix,much higher specific capacity and stability are achieved for the TiO2/N-CNFs containing battery(941.1 mA?h?g-1and 62.4%after 300 cycles),because N heteroatoms which induce defects and modify the electronic structures of carbon matrix could also act as active sites for trapping polysulfides [33].Photographs for the polysulfide trapping experiment are shown in Fig.4(b).When CNFs is used as separator in the H-shaped glassware, the color of electrolyte in the right chamber quickly turns to yellow, indicating easy diffusion of polysulfides through CNFs.When TiO2/CNFs is used as the interlayer, the clear electrolyte turns to light yellow after 24 h, suggesting improved trapping effect for polysulfides after adding TiO2.Notably,the color of electrolyte shows no obvious change after 24 h for TiO2/N-CNFs separated glassware, suggesting the strong blocking ability for polysulfides under the coupling effect of TiO2nanoparticles and N heteroatoms in CNFs.

Fig.3. (a)CV curves of LSB with TiO2/N-CNFs as interlayer;(b)Charge-discharge curves with the first cycle activated at 0.05 C and following cycles recorded at 0.2 C;(c)Rate capability test; (d) Cycling performance at 0.2 C after 1 cycle at 0.05 C for activation.

Fig.4. (a)Prolonged cycle life at 0.5 C of the battery with TiO2/N-CNFs as interlayer;(b)Photographs for the diffusion of polysulfides in H-shaped glassware with CNFs,TiO2/CNFs and TiO2/N-CNFs as separators.

Fig.5. Nyquist plots of CNFs, TiO2/CNFs, and TiO2/N-CNFs separated batteries.

As impedance is also an important factor influencing the performances of LSBs,Nyquist plots for CNFs,TiO2/CNFs and TiO2/N-CNFs separated batteries are compared in Fig.5,with the equivalent circuit and fitted results displayed in inset and Table S1.Re,RinandRctcorrespond to the resistance of electrolyte,Li+transport resistance inside interlayer&electrodes, and charge transfer resistance,respectively.CPE refers to the constant phase elements.All batteries show a similar electrolyte resistance in the range of 12–15 Ω.However, for Li+transport, the resistance of TiO2/CNFs (119.2 Ω)is decreased compared to that of CNFs(127.9 Ω),which proves that introducing TiO2can increase the Li+transport ability of the battery.Because the polar sites of TiO2hinders the migration of polysulfide anions, and thus promotes the transport of Li+[40].However, for TiO2/N-CNFs, the Li+transport resistance increases back to 126.7 Ω,which may be because N atoms have certain effect for Li+adsorption and decrease the Li+transport efficiency to a certain extent [41,42].Nevertheless, the value is still slightly lower than that of CNFs interlayer.As for the charge transfer efficiency,CNFs battery shows a highest resistance of 31.6 Ω;while the values for TiO2/CNFs and TiO2/N-CNFs decrease to 24.6 and 17.3 Ω,respectively, due to the alleviated ‘‘shuttle effect”.All in all, the above results prove that introducing TiO2/N-CNFs significantly depresses the ‘‘shuttle effect” of polysulfides without sacrificing the Li+transport efficiency in LSBs.What’s more, the Nyquist plot for TiO2/N-CNFs battery after cycling test was also recorded and shown in Fig.S8.The decreased Li+transport and charge transfer resistances are attributed to the thorough activation of interlayer&electrodes, as well as sufficient infiltration of electrolyte after long-term operation.

4.Conclusions

In summary, TiO2/N-CNFs membrane was successfully fabricated through an electrospinning-calcination method and applied as interlayer for LSBs.Benefiting from the numerous TiO2nanoparticles and N heteroatoms dispersed in the interwoven CNFs, the interlayer showed excellent capability for trapping polysulfides which greatly reduced the ‘‘shuttle effect” in the battery.The asassembled LSB delivered a high specific discharging capacity of 941.1 mA?h?g-1at 0.5 C after 300 cycles, which retained 62.4% of the initial value.Besides, excellent rate capability and low charge transfer resistance were also demonstrated in the presence of TiO2/N-CNFs interlayer.This work provides a simple and efficient strategy for preparing efficient LSB interlayers, which is of high value in the design and construction process of advanced energy storage devices.

Data Availability

Data will be made available on request.

Declaration of Competing Interest

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.

Acknowledgements

This work was financially supported by the National Natural Science Foundation of China (52102100, 52072330), Industry-University-Research Cooperation Project of Jiangsu Province(BY2021525), Guangdong Basic and Applied Basic Research Foundation (2020A1515110035).

Supplementary Material

Supplementary data to this article can be found online at https://doi.org/10.1016/j.cjche.2022.03.019.