Journal ofMaterials Chemistry A

PAPER

Publ

ishe

d on

27

Sept

embe

r 20

19. D

ownl

oade

d by

SU

STec

h on

3/2

4/20

20 1

:37:

46 A

M.

View Article OnlineView Journal | View Issue

Promoting polys

aPillar of Engineering Product Development

Design, 8 Somapah Road, 487372, SingaporbDepartment of Physics and Engineering

Zhengzhou University, Zhengzhou 450052, CcEngineering Technology Research Centre

Devices and Systems of Guangdong Provinc

Shenzhen 518060, ChinadDepartment of Mechanical and Energy En

and Technology, Shenzhen 518055, China

† Electronic supplementary informa10.1039/c9ta06489c

Cite this: J. Mater. Chem. A, 2019, 7,25078

Received 17th June 2019Accepted 9th September 2019

DOI: 10.1039/c9ta06489c

rsc.li/materials-a

25078 | J. Mater. Chem. A, 2019, 7, 25

ulfide conversion by catalyticternary Fe3O4/carbon/graphene composites withordered microchannels for ultrahigh-rate lithium–sulfur batteries†

Meng Ding,a Shaozhuan Huang,a Ye Wang,b Junping Hu,a Mei Er Pam,a Shuang Fan,ac

Yumeng Shi,c Qi Ge ad and Hui Ying Yang *a

As a promising energy storage system, lithium–sulfur (Li–S) batteries are attracting increasing attention but

still limited by the sluggish reaction kinetics and shuttle effect caused by the dissolution of lithium

polysulfides. Herein, a significant improvement of conversion kinetics and areal sulfur loading is achieved

using an ordered microchannel graphene scaffold with incorporated catalytic Fe3O4 nanocrystals and

porous carbon as a multifunctional sulfur host. The synergy between the polar catalytic Fe3O4

nanocrystals and porous carbon frameworks enables a strong polysulfide anchoring effect and a fast

polysulfide conversion rate. Thus, the 3D ternary Fe3O4/porous carbon/graphene aerogel demonstrates

an ultrahigh rate performance of 755 mA h g�1 at 3C and a high areal capacity of 6.24 mA h cm�2 at

a sulfur loading of 7.7 mg cm�2. Moreover, the promoted reaction kinetics and reliable cyclability are

revealed by the visible evolution of polysulfides using in situ X-ray diffraction (XRD), and the enhanced

chemical anchoring of polysulfides is disclosed by density functional theory (DFT) calculations. This work

provides a promising approach to develop multifunctional ordered porous aerogels with metal oxide

nanocrystals for high-performance Li–S batteries, especially those which suffer from low sulfur loading

and inferior rate performance.

1. Introduction

Lithium–sulfur (Li–S) batteries are one of the most promisingrechargeable batteries to full the rapidly increasing demandfor modern smart grids and electric vehicles owing to their hightheoretical specic capacity (1672 mA h g�1), low cost andenvironmental friendliness.1,2 The natural abundance of sulfurand remarkable energy density (2500 W h kg�1 and 2800 W hL�1) of Li–S batteries have attracted signicant research effortson developing their commercial application.3,4 Nevertheless,there are also considerable challenges that limit Li–S batteriesfrom practical production. One of the major issues is the poorlifespan mainly caused by the dissolution of lithium

, Singapore University of Technology and

e. E-mail: yanghuiying@sutd.edu.sg

, Key Laboratory of Material Physics,

hina

for 2D Material Information Function

e, College of Optoelectronic Engineering,

gineering, Southern University of Science

tion (ESI) available. See DOI:

078–25087

polysuldes (LPSs) and large volume change during charge–discharge processes.5 Another major issue is that current sulfurloading is too low for practical energy storage devices (<2.0 mgcm�2).6,7 Moreover, the poor electronic and ionic conductivity ofsulfur and the discharge product Li2S results in sluggish reac-tion kinetics and large polarization, which further decreases thebattery performance.8

To pave the way for practical applications, designing reliableLi–S batteries with high areal sulfur loading, fast kineticmechanism and stable long-term cycling is essential. Variousporous carbon-based host materials for Li–S battery cathodeshave attracted much interest, which can improve the batteries'performance by reducing the electronic resistance and physi-cally suppressing LPS shuttling.9–11 Especially, compared withrandom pores, the ordered 3D porous structure with a highspecic surface area and abundant interconnected micro-channels is more effective to facilitate the interpenetration ofLPSs into the network, which is favourable for high areal sulfurloading and fast reaction kinetics.9,12 However, for conventionalnonpolar carbon materials, their bonding with LPSs is ratherweak and unstable, resulting in fast capacity decay.13–17 Alter-natively, polar nanomaterials, such as metal oxides,18,19

nitrides,8 suldes,20 carbides,21 and phosphides,22 havedemonstrated adequate bonding with LPSs, by experimental

This journal is © The Royal Society of Chemistry 2019

Fig. 1 Schematic illustration of (a) the preparation processes of Fe3O4/NC/G as the cathode host of Li–S batteries and (b) effective anchoringof polysulfides and promoting their conversion by Fe3O4/NC/Gcomposites.

Paper Journal of Materials Chemistry A

Publ

ishe

d on

27

Sept

embe

r 20

19. D

ownl

oade

d by

SU

STec

h on

3/2

4/20

20 1

:37:

46 A

M.

View Article Online

advances and theoretical perspectives. In this regard, con-structing 3D carbon hosts with polar nanomaterials is prefer-able in addressing the kinetics and shuttle issues. Recently,Fe3O4 has been explored as the cathode and interlayer materialsfor Li–S batteries due to its effective sulfur host ability. Forexample, Lu et al. proposed a carbon cloth composite cathodewith strongly coupled Fe3O4 and nitrogen-doped carbon toobtain a sulfur loading of 4.7 mg cm�2 and a capacity of 531 mAh g�1 at 4C.23 Liu et al. developed a Fe3O4 nanoparticle-deco-rated reduced graphene oxide (rGO) interlayer with a holeyplane for Li–S batteries to achieve a capacity of 589 mA h g�1 at2C.24 Despite the wide variety of materials being explored, therate performance is still not satisfactory with a high sulfurloading. Moreover, it is imperative to study explicit roles ofFe3O4 and carbon using advanced characterization tools forrational design of Li–S battery cathode materials.

Herein, we report a 3D ordered ternary Fe3O4/nanocarbon/graphene (Fe3O4/NC/G) composite host with a liquid polysuldecatholyte as the cathode active material for high rate and arealcapacity Li–S batteries. This hybrid 3D structure demonstratesseveral essential features. First, the ice-templated 3D scaffoldexhibits ordered aligned microchannels that can facilitateelectrolyte penetration and ion transportation. Second, thelarge surface area of the micro/meso-porous nanocarbonderived from a high surface area Fe-based metal–organicframework (Fe-MOF) can enhance LPS accommodation andallow a high loading of sulfur. Third, the highly-conductiveFe3O4 nanocrystals embedded in porous nanocarbon serve asthe “catalytic center” and foster strong chemical adhesion to theLPSs and fast conversion kinetics. Beneting from both thehighly conductive Fe3O4 nanocrystals and 3D graphene/nano-carbon aerogel, the electron transportation can be accelerated,which improves the overall reaction kinetics in the cells. Thesynergistic effect upgrades the ternary Fe3O4/NC/G composite toachieve a high specic capacity of 1250 mA h g�1 at 0.1C, anultrahigh rate capability (755 mA h g�1 at 3C), and a high arealsulfur loading of 7.7 mg cm�2 with stable cycling performanceover 100 cycles (capacity attenuation < 0.01% per cycle). More-over, in situ X-ray diffraction (XRD) reveals the polysulde redoxreaction mechanism and kinetics. The adsorption test andelectrochemical performance evaluation together with densityfunctional theory (DFT) calculations unveil that better LPSadsorption/conversion leads to higher sulfur utilization. Thecomprehensive investigation of the roles of the Fe3O4 nano-crystals and ordered carbon scaffold in adsorption and catalysiswill widen our horizons of rational cathode design with polarmetal oxide/carbon composites for practical high-performanceLi–S batteries.

2. Results and discussion

The Fe3O4/NC/G polysulde cathode was prepared by injectingliquid polysuldes into a MOF-derived composite. Fig. 1adepicts the synthesis process of the Fe3O4/NC/G host and pol-ysulde cathode. First, Fe-MOFs were synthesized via a hydro-thermal reaction and then functionalized with positivelycharged poly(diallyldimethylammonium chloride) (PDDA) and

This journal is © The Royal Society of Chemistry 2019

then easily anchored and assembled with negatively chargedgraphite oxide (GO). Second, the Fe-based MOF/GO aqueoussolution was mixed with ethanol (30 : 1, v/v) and lowered downinto liquid nitrogen at a given cooling rate and kept for 10minutes, and then transferred to a freeze dryer for 48 hours.During freeze-drying, numerous ice columns grew from thebottom and expelled the Fe-MOF/GO to replicate the unidirec-tional ice pattern. Third, the Fe3O4/NG/G composite was ob-tained aer annealing the MOF/GO composite in an argonatmosphere at 800 �C for 2 hours, via which the Fe-MOFs areconverted to Fe3O4 nanocrystals and porous carbon nanorods,and GO is reduced to graphene nanosheets. Finally, the cathodewas prepared by incorporating the Li2S6 catholyte into theFe3O4/NG/G composite. Through physical adsorption fromhighly porous nanocarbon/graphene and strong chemicalattraction between Fe3O4 and polysuldes, the Fe3O4/NC/Gcomposite is able to well entrap a high portion of the Li2S6catholyte, forming a stable cathode material for Li–S batteries.Moreover, the Fe3O4 nanocrystals together with the 3D inter-connected conductive architecture can promote the reactionkinetics, accelerate electron/ion transportation, and ease thenucleation of LPSs, expected to enhance the rate performance ofLi–S batteries (Fig. 1b).

The morphology and microstructure of Fe3O4/NC/G and rGOaerogels are characterized by eld emission scanning electronmicroscopy (FESEM) and transmission electron microscopy(TEM) as shown in Fig. S1† and 2. Fig. S1a and b† depict themorphologies of the as-prepared Fe-MOF and Fe-MOF/GO

J. Mater. Chem. A, 2019, 7, 25078–25087 | 25079

Fig. 2 Morphological and structural characterization of Fe3O4/NC/G composites: (a–c) FESEM images; (d and e) TEM images; and (f) HRTEMimages; the inset in (f) shows the corresponding SAED pattern.

Journal of Materials Chemistry A Paper

Publ

ishe

d on

27

Sept

embe

r 20

19. D

ownl

oade

d by

SU

STec

h on

3/2

4/20

20 1

:37:

46 A

M.

View Article Online

composites. As shown in the SEM images, the Fe-MOFs havea uniform rod-like shape with a length of �7.5 mm and a widthof �150 nm. Aer ice-templated freeze-drying and pyrolysis,abundant ordered microchannels with a diameter of 6–12 mmare created successfully in the Fe3O4/NC/G aerogel (Fig. 2a).From a closer FESEM image (Fig. 2b), it is obvious that theinterconnected network is composed of numerous Fe3O4/NCnanorods surrounded by rGO nanosheets. The rGO aerogel isprepared by the same freeze-drying method by ice-templating asFe3O4/NC/G so the ordered microchannel structure also can beobserved in the rGO aerogel as shown in Fig. S1c and d,† con-rming our reliable approach to prepare graphene aerogels withordered structures. The ordered microchannel structure notonly facilitates electron and lithium ion transportation, but alsoenhances the robustness of the aerogel structure.25,26 Thedetailed morphology of Fe3O4/NC/G is shown in Fig. 2c. Theaggregated Fe3O4 particles distributed on the surface of carbonnanorods and wrinkled rGO sheets are produced by high-temperature annealing. A large number of Fe3O4 nanoparticlescrystallized not only on the surface but also within the carbonnanorods as shown in the TEM image (Fig. 2d). A highermagnication TEM image (Fig. 2e) further determines that thesize of Fe3O4 particles throughout the carbon structure isaround 10 nm. The nanosized Fe3O4 can effectively reduce thediffusion length of lithium ions. The high-resolution TEM(HRTEM) image (Fig. 2f) shows well-resolved lattice fringes withan interlayer spacing of 0.25 nm corresponding to the (311)plane of magnetite Fe3O4. The inset of Fig. 2f shows the selectedarea electron diffraction (SAED) pattern, showing four brightdiffraction rings indexed to the (331), (440), (620), and (731)planes of Fe3O4 from inner to outer, respectively. Based on theHRTEM and SAED results, the pure phase of magnetite Fe3O4

nanocrystals can be conrmed. The corresponding energy-dispersive X-ray spectroscopy (EDX) mapping images (Fig. S2†)verify the main content of C, O, and Fe elements in Fe3O4/NC/Gcomposites and the uniform distribution of Fe throughout the

25080 | J. Mater. Chem. A, 2019, 7, 25078–25087

sample. The hybrid 3D structure of Fe3O4/NC/G with orderedmicrochannels and well-dispersed Fe3O4 nanocrystals is ex-pected to ease the electron transportation, shorten the iondiffusion pathway, and allow large volume change duringdischarge/charge processes.27

Additionally, the material composition and crystal structureof Fe3O4/NC/G are studied by X-ray diffraction (XRD). The XRDpattern of the Fe-MOF (Fig. S3†) is consistent with the reportedresults of hexagonal Fe-MOF (MIL-88).28 Aer pyrolysis at 800�C, the Fe-based MOF was converted to Fe3O4 with porousnanocarbon and GO was reduced to graphene. The major peakslocated at 30.1�, 35.5�, 43.1�, 57.0�, and 62.6� from the XRDpattern (Fig. 3a) are assigned to the (220), (311), (400), (511), and(440) planes of magnetite Fe3O4 (JCPDS card no. 88-0866),respectively. The specic surface area and porosity of sulfurhost materials play very important roles in the physicaladsorption of LPSs, which are investigated by nitrogen sorptiontests. In contrast to rGO with a surface area of 160.5 m2 g�1, theFe3O4/NC/G composite exhibits a higher surface area of 265.8m2 g�1 and a larger pore volume of 0.36 cm3 g�1 (Fig. 3b),rendering Fe3O4/NC/G a superior sulfur host with sufficientmicro/mesopores for promoting mass diffusion of the electro-lyte and accommodating dramatic volume change.29,30 Theincreased specic surface area of Fe3O4/NC/G can be ascribed tothe highly porous MOF precursor. To gure out the reason forthe increased surface area, we have studied pure porous carbonnanorods derived from the same Fe-MOF without rGO(Fig. S4†). It is shown that the specic surface area of Fe3O4/porous carbon composites directly derived from Fe-based MOFis 578.8 m2 g�1. Hence, it is concluded that Fe3O4/NC/G, derivedfrom Fe-MOF/GO composites, inherits a large surface area fromthe highly porous MOF structure.31 The content of Fe3O4 wasdetermined by thermogravimetric analysis (TGA) ina condensed air atmosphere in the temperature range of roomtemperature to 1000 �C. The Fe3O4 content in Fe3O4/NC/G iscalculated to be around 41.5 wt% (Fig. 3c).

This journal is © The Royal Society of Chemistry 2019

Fig. 3 (a) XRD patterns of Fe3O4/NC/G. (b) Nitrogen adsorption isotherms of Fe3O4/NC/G and rGO aerogels. (c) TGA curve of Fe3O4/NC/G. (d)The digital image of adsorption of Li2S6 by Fe3O4/NC/G and rGO aerogels. (e) XPS spectra of Fe 2p in Fe3O4/NC/G before and after adsorption ofLi2S6. (f) CV profiles of the symmetric cells with identical electrodes at a scan rate of 0.2 mV s�1.

Paper Journal of Materials Chemistry A

Publ

ishe

d on

27

Sept

embe

r 20

19. D

ownl

oade

d by

SU

STec

h on

3/2

4/20

20 1

:37:

46 A

M.

View Article Online

To verify the strong anchoring of Fe3O4 for polysuldes,adsorption tests are conducted to compare the adsorptionability of Fe3O4/NC/G and rGO, as shown in Fig. 3d. Theadsorption tests were performed by soaking 1.2 mg Fe3O4/NC/Gand 2 mg rGO aerogels (with a similar net surface area) in Li2S6solution for 24 hours, respectively. Compared with the blankLi2S6 liquid, the color of the remaining solution with Fe3O4/NC/G fades and nearly becomes transparent apparently while thesolution with rGO still remains yellow, intuitively demon-strating the improved chemical adsorption ability of Fe3O4/NC/G. Then the chemical interaction between Fe3O4/NC/G andpolysuldes is further studied by X-ray photoelectron spec-troscopy (XPS) as shown in Fig. S5.† The S 2p spectrum showsfour sulfur environments. The tting peaks at 162.1 and 164.0eV of the S 2p spectrum of Fe3O4/NC/G are related to theterminal sulfur (ST

�1) and bridging sulfur (S0B), respectively.32,33

In addition, the rise between 166 eV and 171 eV can be tted byanother two sulfur environments. The S 2p3/2 peaks at 166.8 eVand 168.4 eV can be ascribed to the binding energy of thiosul-fate and polythionate complexes, respectively.34 The formationof thiosulfate and polythionate complexes is due to the surfaceredox reaction between Fe3O4 and long-chain polysuldes. Bydecreasing the Fe oxidation state, the conversion of long-chainpolysuldes to thiosulfate and then to polythionate complexescan be realized. This conversion is proven to be reversible andbenecial to the battery performance because both thiosulfateand the polythionate complex can serve as mediators to restrainLPS dissolution.35,36 As evidence of the change of the Fe envi-ronment, Fig. 3e compares the XPS spectra of Fe 2p in Fe3O4

before and aer mixing with the Li2S6 electrolyte. Aeradsorption of Li2S6, the peaks of Fe 2p1/2 and Fe 2p3/2 becomebroadened and shi to a lower binding energy as a result of the

This journal is © The Royal Society of Chemistry 2019

bond formation between Fe3O4 and polysuldes that triggerselectron transfer from Li2S6 to Fe3O4.13,37 The overall redoxreaction can be attributed to the strong chemical interaction atthe interface of Fe3O4 and long-chain polysuldes.

In order to evaluate the catalytic activities of Fe3O4/NC/Gtowards the polysulde conversion reaction, symmetricbatteries are assembled and tested. The symmetric batteries areprepared using two identical aerogel samples with an identicalLi2S6 catholyte and anolyte. Cyclic voltammetry (CV) is per-formed for both Fe3O4/NC/G and rGO aerogels under a scan rateof 0.2 mV s�1 from �0.1 to 1.0 V. Because of the absence of thetransformation of S8 to Li2S6 in the rst cycle, their CV curves inthe second cycle are compared in Fig. 3f. Obviously, two pairs ofredox peaks at 0.7/�0.7 V and 0.42/�0.42 V are present in theFe3O4/NC/G aerogel sample yet only one pair of redox peaks at0.48/�0.48 V in the rGO aerogel are seen. The two pairs of redoxpeaks in CV proles of symmetric Li–S batteries are usuallyrarely perceptible for common carbon hosts, due to the weakinteraction between nonpolar carbon and LPSs and slow redoxkinetics. In the CV curve, the cathodic peaks at �0.7 and �0.42V are related to the reduction of S to Li2S6 and Li2S6 to Li2S,while the anodic peaks at 0.7 and 0.42 V are ascribed to theoxidation of Li2S to Li2S6 and then Li2S6 to S, respectively.22

Compared with rGO, Fe3O4/NC/G delivers a higher currentdensity and stronger redox peaks, indicating the acceleration ofthe electrochemical reactions and expeditious conversion ofLi2S6 by the outstanding catalytic effect of polar Fe3O4 nano-crystals.38 Electrochemical impedance spectroscopy (EIS) testsfurther conrm the reduced charge transfer resistance of theFe3O4/NC/G electrode, which is realized by the strong LPSaffinity of Fe3O4 nanocrystals and reduced ion diffusiondistance of porous nanocarbon (Fig. S6†). The high-frequency

J. Mater. Chem. A, 2019, 7, 25078–25087 | 25081

Journal of Materials Chemistry A Paper

Publ

ishe

d on

27

Sept

embe

r 20

19. D

ownl

oade

d by

SU

STec

h on

3/2

4/20

20 1

:37:

46 A

M.

View Article Online

intercept on the x-axis corresponds to the ohmic resistance ofthe cell (Ro). The semicircle in high-frequency region representsthe charge transfer resistance (Rct), and the inclined line at thelow frequency stands for the Warburg impedance for iondiffusion resistance (W).39 Fig. S6† reveals that both Fe3O4/NC/Gand rGO have reduced charge transfer resistances aer cycling,which can be attributed to the activation process along withbetter electrolyte inltration and Li+ ux ability. Even thoughaer cycling, Fe3O4/NC/G still demonstrates a smaller chargetransfer resistance compared with the rGO cathode. This can beexplained by the fact that Fe3O4 has a strong catalytic effect andstrong anchoring to sulfur so the conversion has been facili-tated to overcome the kinetic barriers and promote the overallcharge transfer.22 The outstanding reaction kinetics elucidatesthe indispensable functions of polar Fe3O4 nanocrystals andporous nanocarbon.

The improvement in the reaction kinetics and stability hasbeen further explored using Li–S batteries with a lithium anodeand Fe3O4/NC/G–Li2S6 cathode, in which the sulfur loading is3.85 mg cm�2. Fig. S7a† shows better CV performance of theFe3O4/NC/G cathode with much sharper and stronger redoxpeaks. The CV curves of Fe3O4/NC/G and rGO cathodes displaytwo obvious cathodic peaks, attributed to the reduction ofsulfur to long-chain lithium polysulde Li2Sx (4# x# 8) and thesubsequent conversion to short-chain polysuldes (Li2S2/Li2S),respectively. However, Fe3O4/NC/G exhibits two evident anodicpeaks at 2.39 V and 2.47 V, corresponding to the oxidation ofLi2S/Li2S2 to long-chain polysuldes and then to sulfur.22 rGOonly has one observable anodic peak. Two well-dened anodicpeaks in the Fe3O4/NC/G cathode are rarely visible for commonsulfur cathodes, indicating the enhanced polysulde conver-sion kinetics in the presence of Fe3O4. Even at higher scan rates,the two anodic peaks can be maintained as well, demonstratingsuperior electrochemical stability at higher scan rates(Fig. S7b†). The voltage proles of galvanostatic discharge–charge tests at 0.1C for Fe3O4/NC/G and rGO aerogels (based onthe mass of sulfur in the cell, 1C ¼ 1675 mA g�1) for the secondcycle are shown in Fig. 4a. There are two potential plateaus at2.35 and 2.1 V during discharging, corresponding to theconversion of S8 to long-chain LPSs and long-chain LPSs to Li2S,respectively, and the two plateaus at 2.2 and 2.4 V duringcharging indicate the transformation of Li2S to LPSs and then toS8. The longer and atter plateaus of the Fe3O4/NC/G aerogelwith a higher capacity of over 1200 mA h g�1 and lower polar-ization (vs. z0.18 V) suggest a more efficient redox conversionprocess. The conversion of Li2S to solid S8 in typical Li–Sbatteries generally requires a considerable overpotential to startowing to its natural inertness.40,41 Surprisingly, the Fe3O4/NC/Gcell greatly reduces the overpotential (48 mV) compared withthat of rGO (120 mV), further evidencing the higher activity dueto nanosize Fe3O4 and increased porosity. The rate capabilitiesof sulfur hosted in Fe3O4/NC/G and rGO are calculated based ona sulfur loading of 3.85 mg cm�2. The sulfur amount is deter-mined by the sulfur content in the Li2S6 catholyte adsorbed bythe free-standing sulfur host. Fig. 4b compares the rateperformance of Fe3O4/NC/G and rGO. And the discharge–chargeproles of Fe3O4/NC/G at different rates are shown in Fig. S8.†

25082 | J. Mater. Chem. A, 2019, 7, 25078–25087

Fe3O4/NC/G exhibits an average capacity of 1250, 1101, 973, 904,822, and 755 mA h g�1 at 0.1, 0.2, 0.5, 1, 2, and 3C, respectively.Even though at a large current density of 2C and 3C, the speciccapacity remains at 65.8% and 60.4% of the capacity at 0.1C. Incontrast, the rGO aerogel presents lower capacities under thesame conditions, especially at large current densities of 2C and3C. The capacity of rGO at 3C is only 150 mA h g�1 which is over75% less than that of Fe3O4/NC/G. The inferior rate perfor-mance of rGO can be ascribed to the worse reaction kinetics andmuch weaker interaction between graphene and LPSs. Corre-spondingly, the superior rate capability endowed by Fe3O4/NC/Gis consistent with the improved redox conversion kineticsdemonstrated in the CV tests. With both desirable LPSanchoring and catalytic effects, Fe3O4/NC/G enables expeditiousreaction kinetics.

The cycling performances of Fe3O4/NC/G and rGO aerogelswere evaluated at 0.1C (Fig. 4c). Fe3O4/NC/G exhibits a highreversible capacity of 1007 mA h g�1 and a high coulombicefficiency of 98% aer 100 cycles, with a small capacity decayrate of less than 0.01%. As for the rGO aerogels, the weaker LPSimmobilization and large charge transfer resistance lead toa rapid decay of capacity (707 mA h g�1) together with a reducedcoulombic efficiency (97%). The capacity of both Fe3O4/NC/Gand rGO cathodes increases for the initial few cycles, attributedto the reduced charge transfer resistance as shown in Fig. S6.†The cycling performance proves that as a catalytic ternarycomposite, Fe3O4/NC/G is able to effectively entrap LPSs andaccelerate their conversion simultaneously. To validate thelong-term cycling stability under more practical conditions, theareal sulfur loading of the Fe3O4/NC/G aerogel cathode isincreased from 3.85 to 5.78 and 7.70 mg cm�2. The relativecycling performance is presented in Fig. S9† at a current densityof 0.5C. Even at an ultrahigh areal sulfur loading of 7.70 mgcm�2, the robust 3D structure and polar Fe3O4 nanocrystalsenable the Li–S battery to maintain a high capacity with excel-lent capacity retention of 90.0% over 100 cycles. The arealspecic capacity reaches 6.24 mA h cm�2 with a sulfur loadingof 7.70mg cm�2 (Fig. 4d) at 0.5C and remains at 5.67mA h cm�2

over 100 cycles, demonstrating an outstanding areal capacityand cycling stability for Li–S cathodes. The encouraging cyclingperformance with high areal sulfur loadings further conrmsthe robustness and kinetic durability of the proposed Fe3O4/NC/G aerogel and its great potential in practical energy storagedevices.

In order to understand the advance of Fe3O4/NC/G in theconversion mechanism for Li–S batteries, in situ XRD is per-formed using a prototype Li–S battery depicted in our previousreports.18,35 The evolution of peak intensities of XRD patternsduring cycling is presented in a contour plot in Fig. 5a.Throughout the cycling process, there are two steady peaks ataround 30� and 35�, which correspond to the (220) and (311)planes of Fe3O4 (marked with small black triangles), indicatingthe electrochemical stability of Fe3O4 nanocrystals duringcycling. The in situ cell is cycled at a rate of 0.1C for the rst twocycles. For a regular Li–S battery, the rst discharge prolegenerally consists of two potential plateaus, including an upperone at around 2.35 V and a lower one at around 2.1 V. The upper

This journal is © The Royal Society of Chemistry 2019

Fig. 4 Comparison of (a) discharge–charge profiles at 0.1C, (b) rate performance from 0.1C to 3C and (c) cycling stability and coulombicefficiency at 0.1C between Fe3O4/NC/G and rGO aerogels with the Li2S6 catholyte. (d) Areal capacity of Fe3O4/NC/G with different sulfurloadings.

Paper Journal of Materials Chemistry A

Publ

ishe

d on

27

Sept

embe

r 20

19. D

ownl

oade

d by

SU

STec

h on

3/2

4/20

20 1

:37:

46 A

M.

View Article Online

plateau is assigned to the reduction of S8 to long-chain poly-sulde intermediates (Li2S8/Li2S6), and the slope region corre-sponds to the further conversion to Li2S4. The lower plateau isrelated to the subsequent reduction of polysuldes to insolubleLi2S2/Li2S.42,43 Unlike regular Li–S batteries, the rst dischargeprole of our Li–S batteries does not present the upper plateauat 2.35 V due to the absence of S8 at the beginning as shown inFig. S10.† The absence does not affect the performance of theas-fabricated Li–S batteries in the long-term and the upperplateaus occur in all the following discharging processes. Theas-fabricated lithium polysulde batteries starting with a sloperegion from the open circuit voltage of 2.35 V to 2.1 V experiencethe conversion from Li2S6 to dissoluble Li2S4, where no strongBragg peaks appear (known as “soluble species” region).44Whenthe potential reaches the lower plateau at around 2.1 V, a broadpeak at 27.1� comes into sight, representing the conversion ofLi2S4 to cubic Li2S (circle mark, JCPDS no. 23-0369). Duringfurther lithiation from 2.1 to 1.7 V, the peak intensity of Li2Skeeps growing until the end of the discharging process. Inter-estingly, two weak peaks at 25.6� and 28.3� are visible (four-pointed star marks) at the beginning of cycling and becomeweaker as the lithiation proceeds, and are ascribed to the pol-ysulde intermediates.45 These two peaks nearly disappear atthe end of discharge, indicating the conversion of polysuldesto Li2S.22,46 The observation of polysuldes by XRD is unusual,which may be due to the strong interaction between the Fe3O4

This journal is © The Royal Society of Chemistry 2019

nanocrystals and polysuldes. Once the polysuldes areattached to the surface of the Fe3O4/NC/G host, there would bea better signal detected by XRD. The ‘disappeared’ polysuldessuggest that the diffusion of long-chain LPSs to the Li anode isgreatly reduced and then the undesired anode corrosion can beeffectively suppressed. This result provides a reasonable expla-nation for the improved long-term cycling performance andenhanced coulombic efficiency of the Fe3O4/NC/G host asdemonstrated in Fig. 4c.

During the charging process, the Li2S peak (27.1�) becomesweaker, while the polysulde peaks (25.6� and 28.3�) becomestronger, suggesting the reversible conversion of Li2S to poly-suldes. Aer the complete decomposition of Li2S, there isa “soluble species” region between 2.35 V and 2.4 V, corre-sponding to the conversion of Li2S4 to long-chain polysuldes.Towards the end of the charging process, a group of intensivepeaks at 23.3�, 24.6�, 26.0�, 26.8�, and 28.3� (rhombus symbols)appears and grows stronger from 2.4 V, which can be assignedto the b-S8 (JCPDS 071-0137).46,47 A more general quanticationof LPS conversion is obtained from in situ XRD in the secondcycle (Fig. 5b). Instead of the absence of b-S8 at the beginning ofthe rst cycle, Fig. 5b shows a complete evolution of b-S8 overthe discharge–charge cycle. It is obvious that the peak intensityof b-S8 accumulated from the previous cycle starts to decreaseonce the discharging begins until the discharge capacity rea-ches 186 mA h g�1, indicating the conversion from b-S8 to long-

J. Mater. Chem. A, 2019, 7, 25078–25087 | 25083

Fig. 5 (a) The first and second discharge–charge curves of Li–S batteries with Fe3O4/NC/G at 0.1C and the corresponding in situ XRD patterns ina contour plot. (b) The evolution of b-S8 and Li2S during the second discharge–charge cycle at 0.1C based on the integrated areas of the (311) and(111) reflections, respectively. (c and d) Ex situ SEM images of the Fe3O4/NC/G aerogel after 100 cycles. (e) Element distribution mapping of C, Fe,and S for the Fe3O4/NC/G cathode after 100 cycles.

Fig. 6 (a–f) DFT-calculated molecular structures and adsorptionenergies of Li2Sn and S8.

Journal of Materials Chemistry A Paper

Publ

ishe

d on

27

Sept

embe

r 20

19. D

ownl

oade

d by

SU

STec

h on

3/2

4/20

20 1

:37:

46 A

M.

View Article Online

chain polysuldes. The following quantied evolution of Li2Sand b-S8 is consistent with the in situ XRD for the rst cycle.Unexpectedly, the nucleation starts earlier in Fe3O4/NC/G(45.8% depth vs. 60.9% depth of the total discharging process)compared with rGO (Fig. S11†), indicating the faster reactionkinetics of Fe3O4/NC/G. This phenomenon agrees well with thesuperior rate capacity of Fe3O4/NC/G as demonstrated previ-ously. The Fe3O4/NC/G host endows both faster conversionkinetics and easier nucleation due to the excellent catalyticproperties of Fe3O4 nanocrystals and the reduced electron/iontransfer resistance of the ordered interconnected porousstructure.

Regarding the cycling performance, the LPS dissolution/deposition process is inevitably retarded by the inherently poorelectrical/ionic conductivities and severe volume variation. Tohave an impartial evaluation of the cathode materials, ex situcharacterization is vital for Li–S batteries aer long-termcycling. Fig. 5c and d show the morphologies of the Fe3O4/NC/Gelectrode aer 100 cycles. No visible agglomerates or fracturescan be detected on the surface, suggesting that the 3D inter-connected porous structure can provide great protection of thecathode from pulverization caused by large volume variationduring cycling. The carbon nanorods with Fe3O4 nanocrystalsare still tightly embedded in the rGO sheets, conrming thedurable support for the polysulde conversion by the proposedrobust 3D structure of the Fe3O4/NC/G aerogel. Furthermore,the corresponding elemental C, Fe, and S maps in Fig. 5e revealthat the polysulde species are primarily adsorbed on Fe3O4/porous nanocarbon, rather than graphene sheets, proving the

25084 | J. Mater. Chem. A, 2019, 7, 25078–25087

efficiency of polar Fe3O4 nanocrystals and porous carbon intrapping LPSs by strong chemical anchoring and enhancedphysical adsorption. These results endorse the proposedsynergistic design of multifunctional cathode hosts, where 3Dgraphene is mainly aimed at creating robust supporting plat-form with high electronic conductivity, while the Fe3O4 nano-crystals and porous carbon are capable of capturing andimmobilizing the polysuldes effectively as well as facilitatingthe conversion processes simultaneously.

For a better understanding of the chemical interactionbetween Fe3O4 and polysuldes, we have simulated the molec-ular structures and determined the corresponding bondingenergies based on DFT calculations, as shown in Fig. 6. A

This journal is © The Royal Society of Chemistry 2019

Paper Journal of Materials Chemistry A

Publ

ishe

d on

27

Sept

embe

r 20

19. D

ownl

oade

d by

SU

STec

h on

3/2

4/20

20 1

:37:

46 A

M.

View Article Online

comprehensive comparison of the bonding energies for Fe3O4

and carbon as the polysulde host has been presented inFig. S12.† The results unambiguously reveal the inferiorbonding energies between LPSs and the carbon surface (0.10–0.52 eV), which are much lower than those between LPSs andFe3O4 (0.44–3.35 eV). Such a limited adhesion can result inloosening and detachment of polysuldes from the surface ofgraphene, which induce irreversible capacity loss and corrosionof the Li-metal anode. The favoured chemical adsorptionbetween polysuldes and Fe3O4 is predominantly ascribed tothe much stronger polar–polar interaction.48,49 From anotheraspect, the non-polarity of graphene in Fe3O4/NC/G aerogels isalso benecial to polysulde electrolyte penetration and ion/electron transfer with 3D interconnected porous carbonframeworks. A hetero-polar structure of Fe3O4/NC/G with bothpolar Fe3O4 and non-polar graphene facilitates the binding ofpolysuldes and the diffusion of polysuldes to the electrode.This result conrms the experimental observations that ternaryFe3O4/NC/G is able to offer effective anchoring and catalysis tothe polysuldes all along the cycling processes.

3. Conclusion

In summary, we have developed a free-standing 3D ternaryFe3O4/NC/G composite as an efficient host material to alleviatethe shuttle effect and promote polysulde conversion. The 3Dgraphene ordered structure provides a highly conductive andexible network to facilitate electrolyte penetration andaccommodate the volume expansion of sulfur during lithiation.And porous carbon, inheriting large porosity from the MOFprecursor, provides a large surface area to improve the physicaladsorption of polysuldes and decreases the ion diffusiondistance. Moreover, Fe3O4 nanocrystals derived from the Fe-based MOF serve as the “catalytic center” to effectively immo-bilize polysuldes and promote their redox reaction kinetics,beneting from the synergistic effects of 3D graphene, porouscarbon, and Fe3O4 nanocrystals (Fe3O4/NC/G).

The essential functions of 3D graphene, porous carbon, andFe3O4 nanocrystals are demonstrated by the outstanding Li–Sbattery performance of Fe3O4/NC/G, with an excellent rateperformance of 755 mA h g�1 at 3C, a stable cycling perfor-mance (coulombic efficiency$ 98%) over 100 cycles, and a highareal capacity of 6.24 mA h cm�2 at 0.5C with a sulfur arealloading of 7.7 mg cm�2. Using in situ XRD and DFT calculations,the mechanism of reaction and adsorption has been revealed,and the important role of Fe3O4 in the anchoring and catalysisof polysuldes has been disclosed. These promising resultshighlight the importance of polar metal oxides and synergisticengineering in designing high-performance multifunctionalsulfur host for Li–S batteries.

4. Experimental sectionSynthesis of Fe-MOF/GO composites

The Fe-MOF (MIL-88) was prepared based on a reportedprocedure.50 200 mg of FeCl3 (Merck) and 260 mg fumaricacid (Merck) were added to 20 mL of N,N-dimethylformamide

This journal is © The Royal Society of Chemistry 2019

(DMF, Merck) and then stirred for 10 minutes. The mixturewas transferred to a Teon-lined steel autoclave and treatedat 160 �C for 2 hours. The obtained orange product wascollected and washed using ethanol 3 times by centrifuga-tion. The as-prepared Fe-based MOF was dispersed in anaqueous solution with 1 g NaCl (Merck) and 5 mL PDDA (20wt%, Merck) by ultrasonication for 30 minutes. In themeantime, 6 mL GO suspension (10 mg mL�1) as prepared bya modied Hummer's method51 was diluted in 80 mLdeionized (DI) water by ultrasonication for 1 hour. Subse-quently, Fe-MOF/PDDA was collected from the aqueoussolution by centrifugation and washed by DI water to removePDDA residues. The above mixture was dispersed in 20 mL DIwater uniformly and introduced into the diluted GOsuspension drop by drop under vigorous stirring. Aer stir-ring for 6 hours, the dark brown composites of Fe-MOF/GOwere collected by centrifugation and washed with DI water 3times. Thereaer, the Fe-MOF/GO composites wereimmersed in liquid nitrogen for 10 minutes and then freeze-dried for 48 hours using a freeze dryer (VirTis BenchTop ProFreeze Dryer).

Fabrication of the Fe3O4/NC/G aerogel

60 mg dried Fe-MOF/GO powder was dispersed in a 20 mL GOsuspension (5 mg mL�1) together with 0.344 mL ethanol (99%,v/v ¼ 30 : 1) by ultrasonication for 30 minutes. Aer that, themixture was transferred into a polystyrene Petri dish witha diameter of 100 mm. Then the Petri dish was graduallysubmerged in liquid nitrogen at a given rate and kept for 10minutes to create an ordered interconnected porous structure.The ordered frozen sample was freeze-dried for 48 hours. Thedried sample was annealed in an argon ow at 200 �C for 1 hourand then at 800 �C for 2 hours with a temperature increase of3 �C min�1.

Adsorption test and symmetrical cell assembly

The Li2S6 (3 mmol L�1) electrolyte was prepared froma chemical reaction. Briey, Li2S and sulfur with a molar ratioof 1 : 5 were added to a certain amount of 1,3-dioxolane (DOL,Merck)/1,2-dimethoxyethane (DME, Merck) solution (1 : 1, v/v)and stirred at 60 �C for 12 hours. For the adsorption test, 20 mLLi2S6 electrolyte was diluted in 1 mL DME/DOL (1 : 1, v/v)solution, which then served as the test solution. 1.2 mg Fe3O4/NC/G and 2 mg rGO aerogel were immersed in 1 mL dilutedLi2S6 solution and aged for 24 hours, respectively. To assemblesymmetrical cells, the Fe3O4/NC/G (or rGO) aerogel was cutinto two circular plates with a diameter of 6 mm for theworking and counter electrodes, respectively. Together withthe electrodes, a Celgard 2400 separator and 34 mL Li2S6organic solution including 17 mL for the anode and 17 mL forcathode were used for cell assembly. The Li2S6 organic solu-tion contains 1/3 M Li2S6 and 1 M bis(triuoroethanesulfonyl)imide lithium (LiTFSI, in DME/DOL, v/v ¼ 1 : 1). CV tests wereperformed on an electrochemical workstation (VMP3, Bio-Logic) at 0.2 mV s�1.

J. Mater. Chem. A, 2019, 7, 25078–25087 | 25085

Journal of Materials Chemistry A Paper

Publ

ishe

d on

27

Sept

embe

r 20

19. D

ownl

oade

d by

SU

STec

h on

3/2

4/20

20 1

:37:

46 A

M.

View Article Online

Materials characterization

The morphology of the as-prepared aerogels was studied usinga FESEM (JSM-7600F, JEOL) and a TEM (JEM-2100F, JEOL). XRDpatterns were collected on a D8 Advance X-ray diffractometer(Bruker, Germany) with Cu Ka (l ¼ 0.154 mm) radiation. XPSspectra were collected on a PHI-5400 spectrometer with Al Ka X-ray (250W). TGA analysis (DTG-60, Shimadzu) was performed ina compressed air ow. The powder yield is determined as thepure phase of a-Fe2O3 (JCPDS 33-0664) by XRD (Fig. S13†).52,53

The Fe3O4 content is calculated based on the weight of the a-Fe2O3 yield accordingly. Nitrogen sorption measurements wereperformed using a gas sorption analyser (Autosorb-iQ-MP-XR,Quantachrome). The in situ XRD measurements were per-formed in a stainless-steel Swagelok-type cell connected toa battery tester (Neware, China).

Electrochemical measurements

The Li-polysulde coin cells (CR2032) were assembled in an Ar-lled glove box. The aerogels were punched into a small circularplate with a diameter of 6 mm and an average mass of 1 mg.Then, 17 mL of the catholyte containing 1/3 M Li2S6 and 1 MLiTFSI in DME/DOL (v/v ¼ 1 : 1) was added to the circular aer-ogel for the cathode (sulfur loading of 1.088 mg). The sulfurloading is determined to be 3.85 mg cm�2. The aerogel cathodeswith different sulfur loadings have been tested as well such as5.78 and 7.7 mg cm�2. A Celgard 2400 membrane was utilizedas the separator and placed on the top of the electrode. 34 mL ofthe blank electrolyte containing 0.3 M LiNO3 and 1 M LiTFSI inDME/DOL (v/v¼ 1 : 1) were placed on the separator, followed byplacing the lithium metal foil as the anode. A multi-channelbattery tester (Neware) was utilized for testing the galvanostaticcycling performances within a potential range from 1.7 to 2.8 Vat room temperature. The specic capacity was calculated basedon the mass of sulfur in the Li2S6 electrolyte.

Density functional theory calculations

The atomic conguration and binding energy were determinedby DFT calculations. All the calculations are based on densityfunctional theory (DFT) using the plane-wave pseudopoten-tials54 with the Perdew–Burke–Ernzerhof (PBE)55,56 exchange–correlation functional as implemented in the Vienna Ab initioSimulation Package (VASP).57 A cutoff energy of 450 eV isemployed for the plane wave expansion of the wave functions.The Brillouin zone is sampled with a 1 � 3 � 1 Monkhorst–Pack k-point mesh56 for structural optimization. The conver-gence criteria for the total energy and ionic forces were set to10�4 eV and 0.05 eV A�1, respectively. The construction witha 20 A vacuum zone is made in the z direction to minimize theinteractions between adjacent images.

The adsorption energy (Ead) of LixS8 on the Fe3O4 (311)surface is dened as Ead¼�(ELixSyFe48O64

� EFe48O64� ELixSy), where

ELixSyFe48O64andEFe48O64

are the total energies of the Fe3O4 (311)surface with and without adsorbates. ELixSy represents the totalenergy of S8, Li2S8 Li2S6, Li2S4 Li2S2, or Li2S molecules,respectively.

25086 | J. Mater. Chem. A, 2019, 7, 25078–25087

Conflicts of interest

There are no conicts to declare.

Acknowledgements

This work is supported by the Singapore University of Tech-nology and Design Digital Manufacturing and Design (DmanD)Center.

References

1 Y. Yang, G. Zheng and Y. Cui, Chem. Soc. Rev., 2013, 42, 3018–3032.

2 A. Manthiram, Y. Fu, S.-H. Chung, C. Zu and Y.-S. Su, Chem.Rev., 2014, 114, 11751–11787.

3 Z. W. Seh, Y. Sun, Q. Zhang and Y. Cui, Chem. Soc. Rev., 2016,45, 5605–5634.

4 H.-J. Peng, J.-Q. Huang and Q. Zhang, Chem. Soc. Rev., 2017,46, 5237–5288.

5 M. A. Pope and I. A. Aksay, Adv. Energy Mater., 2015, 5,1500124.

6 Z. Xiao, Z. Yang, L. Zhang, H. Pan and R. Wang, ACS Nano,2017, 11, 8488–8498.

7 H. J. Peng, J. Q. Huang, X. B. Cheng and Q. Zhang, Adv.Energy Mater., 2017, 7, 1700260.

8 Z. Sun, J. Zhang, L. Yin, G. Hu, R. Fang, H.-M. Cheng andF. Li, Nat. Commun., 2017, 8, 14627.

9 J. Schuster, G. He, B. Mandlmeier, T. Yim, K. T. Lee, T. Beinand L. F. Nazar, Angew. Chem., Int. Ed., 2012, 51, 3591–3595.

10 G. Zheng, Y. Yang, J. J. Cha, S. S. Hong and Y. Cui, Nano Lett.,2011, 11, 4462–4467.

11 R. Elazari, G. Salitra, A. Garsuch, A. Panchenko andD. Aurbach, Adv. Mater., 2011, 23, 5641–5644.

12 X. Ji, K. T. Lee and L. F. Nazar, Nat. Mater., 2009, 8, 500.13 Z. Li, J. Zhang, B. Guan, D. Wang, L.-M. Liu and X. W. D. Lou,

Nat. Commun., 2016, 7, 13065.14 H. J. Peng and Q. Zhang, Angew. Chem., Int. Ed., 2015, 54,

11018–11020.15 X. Ji, S. Evers, R. Black and L. F. Nazar, Nat. Commun., 2011,

2, 325.16 Z. Wang, Y. Dong, H. Li, Z. Zhao, H. B. Wu, C. Hao, S. Liu,

J. Qiu and X. W. D. Lou, Nat. Commun., 2014, 5, 5002.17 J. Song, T. Xu, M. L. Gordin, P. Zhu, D. Lv, Y. B. Jiang,

Y. Chen, Y. Duan and D. Wang, Adv. Funct. Mater., 2014,24, 1243–1250.

18 S. Huang, L. Liu, Y. Wang, Y. Shang, L. Zhang, J. Wang,Y. Zheng, O. G. Schmidt and H. Y. Yang, J. Mater. Chem. A,2019, 7, 6651–6658.

19 H.-E. Wang, K. Yin, N. Qin, X. Zhao, F.-J. Xia, Z.-Y. Hu,G. Guo, G. Cao and W. Zhang, J. Mater. Chem. A, 2019, 7,10346–10353.

20 T. Chen, L. Ma, B. Cheng, R. Chen, Y. Hu, G. Zhu, Y. Wang,J. Liang, Z. Tie and J. Liu, Nano Energy, 2017, 38, 239–248.

21 X. Liang, A. Garsuch and L. F. Nazar, Angew. Chem., Int. Ed.,2015, 54, 3907–3911.

This journal is © The Royal Society of Chemistry 2019

Paper Journal of Materials Chemistry A

Publ

ishe

d on

27

Sept

embe

r 20

19. D

ownl

oade

d by

SU

STec

h on

3/2

4/20

20 1

:37:

46 A

M.

View Article Online

22 S. Huang, Y. Von Lim, X. Zhang, Y. Wang, Y. Zheng, D. Kong,M. Ding, S. A. Yang and H. Y. Yang, Nano Energy, 2018, 51,340–348.

23 K. Lu, H. Zhang, S. Gao, H. Ma, J. Chen and Y. Cheng, Adv.Funct. Mater., 2019, 29, 1807309.

24 Y. Liu, X. Qin, S. Zhang, G. Liang, F. Kang, G. Chen and B. Li,ACS Appl. Mater. Interfaces, 2018, 10, 26264–26273.

25 Y. Wang, D. Kong, W. Shi, B. Liu, G. J. Sim, Q. Ge andH. Y. Yang, Adv. Energy Mater., 2016, 6, 1601057.

26 M. Yang, N. Zhao, Y. Cui, W. Gao, Q. Zhao, C. Gao, H. Bai andT. Xie, ACS Nano, 2017, 11, 6817–6824.

27 J. He, G. Hartmann, M. Lee, G. S. Hwang, Y. Chen andA. Manthiram, Energy Environ. Sci., 2019, 12, 344–350.

28 C. Serre, F. Millange, S. Surble and G. Ferey, Angew. Chem.,Int. Ed., 2004, 43, 6285–6289.

29 R. Ding, J. Zhang, J. Qi, Z. Li, C. Wang and M. Chen, ACSAppl. Mater. Interfaces, 2018, 10, 13470–13478.

30 J. Xu, H. Tang, T. Xu, D. Wu, Z. Shi, Y. Tian and X. Li, Ionics,2017, 23, 3273–3280.

31 Y. Z. Chen, C. Wang, Z. Y. Wu, Y. Xiong, Q. Xu, S. H. Yu andH. L. Jiang, Adv. Mater., 2015, 27, 5010–5016.

32 Y. Wang, R. Zhang, J. Chen, H. Wu, S. Lu, K. Wang, H. Li,C. J. Harris, K. Xi, R. V. Kumar and S. Ding, Adv. EnergyMater., 2019, 9, 1900953.

33 J. He, L. Luo, Y. Chen and A. Manthiram, Adv. Mater., 2017,29, 1702707.

34 G. Zhou, L.-C. Yin, D.-W. Wang, L. Li, S. Pei, I. R. Gentle, F. Liand H.-M. Cheng, ACS Nano, 2013, 7, 5367–5375.

35 S. Huang, Y. Wang, J. Hu, Y. V. Lim, D. Kong, Y. Zheng,M. Ding, M. E. Pam and H. Y. Yang, ACS Nano, 2018, 12,9504–9512.

36 X. Liang, C. Hart, Q. Pang, A. Garsuch, T. Weiss andL. F. Nazar, Nat. Commun., 2015, 6, 5682.

37 M. Descostes, F. Mercier, N. Thromat, C. Beaucaire andM. Gautier-Soyer, Appl. Surf. Sci., 2000, 165, 288–302.

38 H. Lin, L. Yang, X. Jiang, G. Li, T. Zhang, Q. Yao, G. W. Zhengand J. Y. Lee, Energy Environ. Sci., 2017, 10, 1476–1486.

39 Y. Wang, D. Kong, S. Huang, Y. Shi, M. Ding, Y. Von Lim,T. Xu, F. Chen, X. Li and H. Y. Yang, J. Mater. Chem. A,2018, 6, 10813–10824.

This journal is © The Royal Society of Chemistry 2019

40 Y. Yang, G. Zheng, S. Misra, J. Nelson, M. F. Toney andY. Cui, J. Am. Chem. Soc., 2012, 134, 15387–15394.

41 S. Meini, R. Elazari, A. Rosenman, A. Garsuch andD. Aurbach, J. Phys. Chem. Lett., 2014, 5, 915–918.

42 Y. V. Mikhaylik and J. R. Akridge, J. Electrochem. Soc., 2004,151, A1969–A1976.

43 D. Moy, A. Manivannan and S. Narayanan, J. Electrochem.Soc., 2015, 162, A1–A7.

44 S. Walus, C. Barchasz, R. Bouchet, J. C. Lepretre, J. F. Colin,J. F. Martin, E. Elkaım, C. Baehtz and F. Alloin, Adv. EnergyMater., 2015, 5, 1500165.

45 J. Conder, R. Bouchet, S. Trabesinger, C. Marino, L. Gublerand C. Villevieille, Nat. Energy, 2017, 2, 17069.

46 N. A. Canas, S. Wolf, N. Wagner and K. A. Friedrich, J. PowerSources, 2013, 226, 313–319.

47 S. Walus, C. Barchasz, J.-F. Colin, J.-F. Martin, E. Elkaım,J.-C. Lepretre and F. Alloin, Chem. Commun., 2013, 49,7899–7901.

48 H. J. Peng, G. Zhang, X. Chen, Z. W. Zhang, W. T. Xu,J. Q. Huang and Q. Zhang, Angew. Chem., Int. Ed., 2016, 55,12990–12995.

49 X. Tao, J. Wang, C. Liu, H. Wang, H. Yao, G. Zheng,Z. W. Seh, Q. Cai, W. Li and G. Zhou, Nat. Commun., 2016,7, 11203.

50 J. Troyano, A. Carne-Sanchez, J. Perez-Carvajal,L. Leon-Reina, I. Imaz, A. Cabeza and D. Maspoch, Angew.Chem., Int. Ed., 2018, 57, 15420–15424.

51 W. S. Hummers Jr and R. E. Offeman, J. Am. Chem. Soc., 1958,80, 1339.

52 K. Cao, L. Jiao, H. Liu, Y. Liu, Y. Wang, Z. Guo and H. Yuan,Adv. Energy Mater., 2015, 5, 1401421.

53 Z.-J. Zhang, Y.-X. Wang, S.-L. Chou, H.-J. Li, H.-K. Liu andJ.-Z. Wang, J. Power Sources, 2015, 280, 107–113.

54 J. Ihm, A. Zunger and M. L. Cohen, J. Phys. C: Solid StatePhys., 1979, 12, 4409–4422.

55 J. P. Perdew, K. Burke and M. Ernzerhof, Phys. Rev. Lett.,1996, 77, 3865.

56 G. Kresse and D. Joubert, Phys. Rev. B: Condens. Matter Mater.Phys., 1999, 59, 1758.

57 H. J. Monkhorst and J. D. Pack, Phys. Rev. B: Solid State, 1976,13, 5188.

J. Mater. Chem. A, 2019, 7, 25078–25087 | 25087