Lithium–Sulfur Batteries Progress and Prospects

P R O G R E S S R E P O R T

L ithium–Sulfur Batteries: Progress and Prospects A rumugam M anthiram ,*S heng-Heng C hung ,a nd C henxi Z u

P rof. A. Manthiram, S.-H. Chung, C. Zu

M aterials Science and Engineering Program & Texas Materials Institute

T he University of Texas at Austin A ustin ,T X 78712 ,U SA E-mail: m anth@https://www.sodocs.net/doc/816020855.html, DOI: 10.1002/adma.201405115

1. I ntroduction A vailability of energy at an affordable cost without adverse environmental consequences is one of the major challenges to modern society. The increasing consumption and limited availability of fossil fuels, along with the environmental impact caused by burning fossil fuels, have prompted the development of alternate, sustainable, clean energy technologies. Renew-able energy sources, such as solar and wind, are appealing in this regard, but the effective utilization of these intermittent energy sources requires ef? cient and economical electrical energy storage (EES) systems. Rechargeable batteries are the most viable option for EES. Among the various rechargeable battery systems, lithium-ion batteries offer the highest energy density. H owever, the energy density of conventional lithium-ion batteries with insertion-compound cathodes (e.g., LiCoO 2,LiMn 2O 4 , and LiFePO 4 )

and anodes (graphite) is limited to ful? ll the demands of electric vehicles and smart grids. [ 1] D

evelopment of advanced energy-storage systems for portable devices, electric vehicles, and grid storage must fulfill several requirements: low-cost, long life, acceptable safety, high energy, high power, and environmental benignity. With these requirements, lithium–sulfur (Li–S) batteries promise great potential to be the next-generation high-energy system. However, the practicality of Li–S technology is hindered by technical obstacles, such as short shelf and cycle life and low sulfur content/loading, arising from the shuttling of polysul? de intermediates between the cathode and anode and the poor electronic con-ductivity of S and the discharge product Li 2 S . Much progress has been made during the past ? ve years to circumvent these problems by employing sulfur–carbon or sulfur–polymer composite cathodes, novel cell con? gurations, and lithium-metal anode stabilization. This Progress Report highlights recent

developments with special attention toward innovation in sulfur-encapsulation techniques, development of novel materials, and cell-component design. The scienti? c understanding and engineering concerns are discussed at the end in every developmental stage. The critical research directions needed and the remaining challenges to be addressed are summarized in the Conclusion.

Accordingly, new battery chemistries are

being intensively explored. R ecently, lithium–air (Li–air) and lithium–sulfur (Li–S) batteries with theo-retical energy densities of, respectively,

3500 W h kg ?1 and 2500 W h kg ?1,have

attracted much attention. However, Li–air batteries are met with insurmountable challenges due to several intrinsic prob-lems and low practical energy density. [ 2]

Compared with Li–air batteries, the chal-lenges of Li–S batteries are less severe, so Li–S batteries are believed to be more fea-sible for practical utility.

[ 3] The high energy density of Li–S batteries originates from the sulfur cathode having a high theoret-ical capacity of 1672 mA h g

?1 and the lith-ium-metal anode having a high capacity of

3860 mA h g ?1 . Also, sulfur is abundant

and environmentally benign. During dis-charge, Li

+ ions are produced at the lith-ium-metal anode and move through the electrolyte to the sulfur cathode, while the

electrons ? ow through the external circuit, producing Li 2S as

the ? nal discharge product at the cathode ( F igure 1).

D espite the advantages of high energy density and low cost,

the practical utility of Li–S batteries is challenging to realize. The main obstacles that prevent the commercialization of Li–S batteries are the low electrochemical utilization of sulfur and fast capacity fading. First, both sulfur and the discharge product Li 2 S are electronically and ionically insulating. There-fore, incorporation of sulfur into a conductive matrix (e.g.,

carbon, polymer, or metal) is usually required, which reduces the energy density. Second, unlike insertion-compound cath-odes, sulfur undergoes a series of complicated compositional and structural changes with the formation of soluble poly-sul? de intermediates, resulting in poor mechanical stability

and severe capacity fading. [ 4] Third, the dissolved polysul? de

species shuttle between the cathode and anode, which results

in active material loss from the cathode, a low Coulombic ef? -ciency, and passivation of the lithium-metal surface with insol-uble products, e.g., Li 2S /Li 2S 2 .

Fourth, the lithium-metal anode degrades due to surface passivation and an unstable solid-elec-trolyte interphase (SEI) that is formed with the organic electro-lyte, prohibiting the long-term cycling stability of Li–S batteries.

I n order to address the aforementioned problems with Li–S batteries, many efforts and improvements have been made in the past decade. [ 4a ,5] The improvements include: i) novel

cell component materials and structures (e.g., the cathode, binders, the electrolyte, and the anode); ii) mechanistic under-standing of the Li–S redox chemistries; and iii) innovation

PROGRESS REPORT

in cell con? guration. In this Progress Report, we present the

advances made regarding Li–S batteries in the past decade, with a focus on aspects that boost the sulfur utilization, cycle stability, and power density of Li–S batteries. Speci? cally, inno-vations in cathode structures and novel cell con? gurations are presented, e.g., use of sulfur–carbon and sulfur–polymer nanocomposites, porous polysul? de reservoirs, porous current collectors, binders, free-standing composite electrodes, inter-layers between the cathode and the separator, surface-coated separators, polysul? de catholytes, sandwiched cathode struc-tures, lithium-metal protection, and Li 2 S

activation. Based on the progress made with these strategies, future directions for further improvements and practical utility of Li–S batteries are presented. 2. L ithium–Sulfur Cells with a Sulfur Cathode

T he Li–S battery, fabricated with a sulfur cathode, is the most

promising high-energy-density system. [ 6] The conventional

cathode contains pristine sulfur mixed with conductive carbon

and binder.

[ 7] H owever, with a pure sulfur cathode prepared with sulfur–carbon binder mixtures, it is dif? cult to effectively utilize and stabilize the active material. The clusters or agglom-erates of insulating sulfur particles form inactive cores within the sulfur–carbon binder mixtures, which limits the redox reaction of the cathode. The dissolved polysul? des that form at intermediate discharge/charge states lead to low discharge/charge ef? ciency and loss of active material. The end discharge products (Li 2S 2

/Li 2 S mixture) that are formed by the reduction of the dissolved polysul? des redeposit on the surface of the elec-trodes, form inactive agglomerates/zones, and block ion and electron transport in the electrode. In addition, the structure of sulfur changes by 80 vol%, involving the phase transitions of sulfur–polysul? de–sul? de, which breaks down the interparticle contact among the insulating sulfur, the conductive carbon, the binder, and the current collector. This leads to rapid capacity

fading during cycling. [ 8]

T o solve these scienti? c challenges, modi? cation of the physical/chemical properties and the morphology of sulfur is

the ? rst step in the development of Li–S technology. Pristine sulfur in the cathode mixtures is replaced by various sulfur–porous-carbon nanocomposites or sulfur–conductive-polymer nanocomposites to increase the cathode conductivity and sup-press polysul? de migration.

[ 9] H owever, the addition of extra conductive additives may reduce the sulfur content in the com-posite cathode and increase the complexity for widespread use. Therefore, a balance between sulfur loading and battery perfor-mance is required. Moreover, since Li–S cells involve a conver-sion reaction, compared with conventional insertion-compound cathodes, modi? cation in the cell components along with the design of novel cathode con? gurations could be a promising

strategy for improving the performance of Li–S batteries. [ 4d ,10 ]

A rumugam Manthiram is the Cockrell Family Regents Chair in Engineering and the Director of the Materials Science and Engineering Program and the Texas Materials Institute at the University of Texas at Austin (UT-Austin). His research interests are in the area of materials for rechargeable batteries, fuel cells, and solar

cells, including novel synthesis approaches for nano-

materials. See https://www.sodocs.net/doc/816020855.html,/~manthiram/ for further details.

S heng-Heng Chung is currently a Ph.D. student in the Materials Science and Engineering Graduate Program at UT-Austin. He obtained his B.S. (2006) in Resources Engineering from National Cheng Kung University and his M.S. (2008) in Materials Science and Engineering from National Tsing Hua

University in Taiwan. His research is focused on process modi?

cation and cell-con? guration design for batteries.

C henxi Zu is currently a Ph.D. student in the Materials Science and Engineering

Graduate Program at

UT-Austin. She obtained her B.Eng. (2011) in Materials Science and Engineering from Beihang University in China. Her research interests are high-energy-density Li–S batteries, lithium anodes, and surface and interface studies.

F igure 1. Schematic and voltage pro? les of a Li–S cell. Reproduced with

permission. [ 6] Copyright 2012, Nature Publishing Group.

P R O G R E S S R E P O R T

2.1. C omposite Cathodes A high-performance Li–S cell relies heavily on the optimiza-tion within the cathode con? guration. A practical cathode design should include high cathode conductivity, outstanding polysul? de-trapping capability, and a robust electrode structure. The most promising approach is the encapsulation of sulfur within conductive additives to form a sulfur-based composite cathode. Porous carbon substrates and conductive polymers are essential to enable the redox accessibility of the insulating sulfur (conductivity of 5 × 10 ?30 S cm ?1 ). The high surface area and porosity of porous carbon materials, as well as the chem-ical gradient created by polymer coatings could satisfy the crit-ical requirements of good electronic and ionic conductivities, as well as retention of polysul? des within sulfur-based com-posite cathodes. Moreover, a porous carbon substrate or a soft polymer can buffer huge volume changes of the sequestered active material. [ 4b ,5c ] Surface functionalization techniques fur-ther modify the morphologies of the nanocomposites to limit the diffusion of polysul? de out of the nanocomposites or trap the migrating polysul? des.

[ 11 ] 2.1.1. S ulfur –C arbon Nanocomposites C onductive carbon acts well as an effective electronic conductor to enhance the utilization of the insulating sulfur. Its loose clusters serve as a porous framework to contain the redox prod-ucts. [ 12 ] However, poor links between the active material and the carbon matrix, as well as the unstable architecture of the carbon clusters result in rapid capacity fading and low ef? ciency during the initial several cycles. This results in unstable and poor cyclability with a cycle life of less than 50 cycles. Recent progress on sulfur–carbon composite cathodes has minimized these problems. Sulfur–carbon nanocomposites bene? t from their hierarchical micro-/mesoporous structural design, satisfying the

criteria for encapsulating sulfur into porous substrates.

[ 4a ,4c ,4d ,11a ] J i et al. ? rst presented a sulfur–mesoporous-carbon com-posite synthesized by a melt-diffusion process.

[ 13 ] The CMK-3 ordered mesoporous carbon, synthesized by using the SBA-15 silica template, has a high conductivity, uniform and narrow

mesopores (3 nm), a large pore volume (2.1 cm 3g ?1 ), and an

interconnected porous structure. As a result, at 155 °C, liquid

sulfur with the lowest viscosity achieved excellent active mate-rial encapsulation in the mesoporous space and hence, for the ?

rst time, exhibited high reversible capacity with good ef? -ciency in Li–S batteries. [ 4b ,13 ] Jayaprakash et al. utilized a sulfur-vaporization route to infuse gaseous sulfur into porous hollow

carbon with a mesoporous shell, as shown in F igure 2a . [ 14 ]

The use of mesoporous hollow carbon capsules can encapsu-late and sequester up to 70 wt% active material in their inte-rior and porous shell. As a reference, this sulfur content refers

to the content of active material in the nanocomposite. The carbon protection shell minimizes polysul? de dissolution and the shuttle effect. The utilization of sulfur vapor leads to molecular-level contact between the insulating active material and the conductive carbon shell. Moreover, the mesoporous

shell allows access of electrolyte and preserves fast Li

+-ion transport. Therefore, this scalable procedure produces ef? cient uptake of elemental sulfur with effective ion and electron trans-port for achieving outstanding cyclability. These porous hollow

carbon@sulfur nanocomposites, for the ?

rst time, provided a long cycle life of 100 cycles and a high reversible capacity

approaching 1000 mA h g ?1 at a high cycling rate.

F igure 2. SEM observation, schematic model, and cell performances of various sulfur-based nanocomposites: a) porous hollow carbon@sulfur nanocomposite, b) sulfur–polyaniline nanotube, and c) poly(ethylene glycol) (PEG)-wrapped graphene–sulfur nanocomposite. a) Reproduced with

permission.

[ 14 ] Copyright 2011, Wiley-VCH. b) Reproduced with permission. [ 31 ] Copyright 2012, Wiley-VCH. c) Reproduced with permission. [ 33 ]Copy-right 2011, American Chemical Society.

PROGRESS REPORT

Z hang et al. reported double-shelled hollow carbon spheres

synthesized by way of hollow SnO 2 spheres as the template. [ 15 ]

H owever, the larger core and thicker shell may decrease the redox accessibility of the encapsulated sulfur as compared with the single-shelled hollow carbon spheres reported by

Jayaprakash et al. [ 14 ] Recently, Peng et al. reported a hollow gra-phene nanoshell for sulfur encapsulation, which might be a

promising strategy to solve the concern of the non-conductive

sulfur core. [ 16 ] The resulting hollow graphene nanoshells had a

diameter of 10–30 nm with a high pore volume of 1.98 cm 3g ?1

for excellent sealing of the active material in the conductive carbon shell. The small diameter of the graphene nanoshell limits the particle size of the encapsulated sulfur well and, therefore, suppresses the formation of the non-conductive core. As a result, a high sulfur utilization of 91% was achieved at a C/10 rate. It is interesting to ? nd a reversible discharge

capacity of 419 mA h g ?1 and stable Coulombic ef? ciency of

95% without the addition of lithium nitrate after 1000 cycles at a high 1C rate. The corresponding capacity fading rate was only 0.06% per cycle. The reported cycle life and capacity fading rate were, respectively, longer and lower than the values of other Li–S batteries based on graphene in lithium-nitrate-free electrolytes. I n addition to heat-treatment procedures, a chemical-syn-thesis approach is another favorable process for synthesizing sulfur–carbon composites and creating a strong binding between the sulfur and the carbon substrate. Our group reported a carbon-wrapped sulfur nanocomposite synthe-sized in aqueous solution at room temperature, which is a

facile and nontoxic manufacturing process.

[17] The resulting core–shell structured sulfur–carbon nanocomposites encap-sulated the precipitated sulfur at the interspaces between the clusters of carbon nanoparticles. The strong chemical bonding between the sulfur nucleates and the dispersed carbon allowed this approach to be applicable to any carbon substrates. However, the formation of the sulfur core in the core–shell-structured sulfur–carbon nanocomposites may cause an inactive and non-conductive core in the agglom-erates of active material. [17] A similar problem can also be

found with the core–shell-structured carbon–sulfur composite prepared by a sulfur deposition method. In a case reported by Wang et al., the non-conductive sulfur shell on the nano-composite may limit the rate capability and utilization of

the active material.

[18] H ence, both nanocomposites showed a limited enhancement in high-rate cycle performance. The reversible discharge capacity of the former and the latter

after 50 cycles were, respectively, 697 mA h g

?1 at a C/4 rate and 336 mA h g

?1 at a C/2 rate. T o prevent sulfur aggregation, Ji et al. reported an appealing

method by using graphene oxide as a sulfur immobilizer.

[19] Graphene oxide has strong reactive functional groups on its surface to bond the nanosized sulfur particles (tens of nanom-eters).

[20] The nanocomposites were prepared by a two-step chemical deposition and thermal treatment process. The resulting graphene oxide–sulfur nanocomposites had a thin and uniform sulfur coating on the conductive graphene oxide sheet to avoid the aggregation of large sulfur particles. The nanocomposite exhibited good electrochemical reversibility

and capacity stability with a discharge capacity of >950 mA h g ?1 for 50 cycles. Inspired by this, our group prepared hydroxy-lated graphene–sulfur nanocomposites that achieved uniform

sulfur distribution on graphene and overcame the problem of non-conductive sulfur aggregation in the chemical prepa-ration stage. [21] This approach attaches the amorphous sulfur

onto hydroxylated graphene nanosheets by chemical bonding . The added hydroxyl functional groups limit the growth of bulk crystalline sulfur during composite preparation and inhibit polysul? de dissolution into the liquid electrolyte during cell cycling. However, the nanoscale sulfur coating on the carbon substrates may incur the concern of a low sulfur content in the composite cathode. On the other hand, although other solution-based synthesis processes also offer unique nanocom-posites with high performance, the use of the toxic CS 2solu-tion or the production of the toxic H 2

S byproduct may nullify the advantages of the environmentally friendly sulfur. [22]In general, the common merit of these chemical-synthesis pro-cesses is the use of an aqueous solution at room temperature to create strong chemical bonds between the sulfur and the carbon substrate. O verall, the advances in the development of carbon sub-strates (porous carbon, [ 13–15,23 ] carbon ? bers (CNFs),

[ 22c ,24 ] carbon nanotubes (CNTs), [ 25 ]graphene, [ 19, 21 ,25c ,26 ] and even carbide-derived carbon

[ 27 ] and active-material encapsulation techniques have boosted cell performance, with an increase

in discharge capacity from less than 500 mA h g

?1 to above 1500 mA h g

?1 and an enhancement in cycle life from 20 to ca. 1000 cycles.

[ 13 ,23c ,25d ,28 ] Although these sulfur–carbon nanocom-posites have shown progress in cell performance, a high spe-ci? c capacity (>1200 mA h g ?1 ) and an excellent cycle stability

during long cycle life (>80% capacity retention over 100 cycles) could not be realized simultaneously with any single composite

material. [ 5a ] T o develop high-capacity sulfur–carbon composite cathodes, porous carbon hosts with tunable micro-/mesopores are sug-gested to be combined with a suitable sulfur encapsulation process to achieve outstanding sulfur immobilization and ion/electrolyte transference. The combination of a highly conduc-tive carbon network with a porous carbon nanostructure is a novel method for a composited conductive/porous carbon host to effectively utilize the immobilized sulfur for Li–S batteries. In a conductive/porous carbon host, the conductive carbon materials (CNFs, CNTs), and graphene) effectively raise the redox capability of the encapsulated active material that are

protected by the nanoporous carbon substrate.

[ 29a ?d ]Moreover, heteroatom doping can chemically improve the reactivity and facilitate chemical sulfur-adsorption to the chemically stable

carbon host.

[ 23g ,29e ?i ] T o achieve practical applications, an increase in total acces-sible pore volume will be essential to enhance sulfur loading and the sulfur content. After impregnating sulfur into open pores, voids are necessary to ensure good electrolyte impreg-nation and fast lithium-ion transference to achieve an out-standing high-rate cell performance. On the other hand, these extra voids can also buffer volume changes of the active mate-rial. The graphitization level and the mechanical strength of the carbon hosts dominate the improvements in, respectively, cathode conductivity and the integrity of the nanocomposite

structure. [ 4a ,4d ,22a ]

P R O G R E S S R E P O R T

2.1.2. S ulfur –P olymer Nanocomposites P olymers are another type of frequently used additive in rechargeable Li–S batteries, especially the conductive polymer coating on sulfur particles. Conductive polymers can be tai-lored or used to modify the surface of cathodes to facilitate

ion and charge transport. [11b ] The corresponding synthesis

strategies aim at con? ning sulfur and its redox products in nanocomposites with controlled morphology. Core–shell structures, in which the insulating sulfur is the core and the conductive polymers are the shell, offer shorter path-ways for ion and electron transport and more freedom for compositional change. Wang et al. introduced the ? rst sulfur–conductive polymer nanocomposites synthesized with

polyacrylonitrile (PAN) and sublimed sulfur at ca. 300 °C. [11c ]

Following this, extensive efforts have targeted the synthesis of sulfur–conductive-polymer nanocomposites with various conductive polymers and structures to enhance the core–

shell structure. [4a,d ] Our group has reported several core–

shell structured sulfur–polypyrrole nanocomposites with various microstructure and morphologies as composite cath-odes.

[30] The polypyrrole coating functions as a stable inter-face between the liquid electrolyte and the polysul? de species, allowing the accessibility of ions and charge, but sequestering the diffusion of the active material. This demonstrates that sulfur–conductive-polymer nanocomposites containing dis-persed sulfur and conductive nanoparticles could become a viable approach to chemically overcome some of the persis-tent problems associated with rechargeable lithium–sulfur

batteries. [30b ] I n addition to the core–shell structure, conductive poly-mers have recently been used to mimic the structure of CNTs and CNFs. Xiao et al. reported a self-assembled polyaniline nanotube for enhancing sulfur encapsulation, as shown in Figure 2b .[31] Electropositive groups on the sulfur–polyaniline nanocomposites attract polysul? des through electrostatic forces, reducing the loss of active material during cell cycling and proving long-term cyclability (cycle life of 500 cycles). This progress implies that additional functionalities could be introduced into conductive polymers to further improve the performance. Thus, our group has investigated the effect of a mixed ionic/electronic conductor (MIEC) in sulfur–polypyrrole

nanocomposites for Li–S batteries.

[32] The MIEC of polypyr-role was synthesized with poly(2-acrylamido-2-methyl-1-pro-panesulfonic acid) (PAAMPSA). The resulting sulfur–MIEC nanocomposites reduced the overpotential and electrochemical impedance. The excellent electrochemical stability obtained at various cycling rates was attributed to the MIEC that facilitates ion and electron transfer and captures polysul? des within the cathode region. A lthough conductive or functional polymers have stronger chemical polysul? de-trapping capability than the chemically inert carbon, the relatively low electrical conductivity of the polymer additives (as compared with carbon) may limit the progress on both raising the active material utilization and

reutilizing the trapped active material. [ 10a ] In general, sulfur–

polymer nanocomposites can achieve outstanding cycle sta-bility, but still may not help to increase the practical speci? c capacity. 2.1.3. F unctional-Polymer-Supported Sulfur –C arbon Nanocomposites R ecent advancements in nanoscience and nanotechnology have offered exciting opportunities for the development of a mixed soft-polymer coating with conductive/porous carbon substrates for sulfur-based nanocomposites. Various polymer-coated sulfur–carbon nanocomposites that can effectively encapsulate the redox products and maintain a robust but porous electrode structure have been reported. These advantages are essential

for improving the electrochemical performance. [ 4b ,11a ,22a ] As a

pioneer in high-performance composite cathodes, Ji et al. ? rst demonstrated the enhanced discharge capacity and stability of a polyethylene glycol (PEG)-coated sulfur–CMK-3 nanocom-posite. [ 13 ] These PEG-coated sulfur–carbon nanocomposites physically contain the active material in their ordered CMK-3 mesoporous carbon substrates and further chemically retard polysul? de diffusion by creating a chemical gradient in the nanocomposites. In addition to improving the electrochemical reversibility, the soft-polymer coating can tolerate huge volume changes from the trapped active material. F igure 2 c depicts a PEG-wrapped graphene–sulfur nanocom-posite reported by Wang et al.

[ 33 ] The synthesized sub-microm-eter sulfur particles are coated with PEG surfactants and gra-phene sheets, which function as a chemical and physical pro-tection coating for trapping the polysul? de intermediates. The soft PEG coating further accommodates the volume change of the wrapped sulfur particles during cell cycling. The conduc-tive graphene coating offers the encapsulated sulfur particles a robust and electrically conductive shell. Therefore, PEG/gra-phene–sulfur nanocomposites show high and stable discharge

capacities up to 600 mA h g

?1 over more than 100 cycles. This depicts a signi? cant improvement in the cyclability of polymer-coated sulfur–carbon nanocomposites from 20 cycles (at a C/10 rate) to 140 cycles (at a C/2 rate). Wu et al. presented a polyaniline-coated sulfur–multiwall CNT (MWCNT) nanocom-posite.

[ 34 ] The polyaniline that was synthesized by a rapid in situ chemical oxidation polymerization covered the sulfur–carbon mixture to form a gel-like cathode composite. The resulting composite possesses a conductive MWCNT network to improve the cathode conductivity. The porous polyaniline shell reduces

the Li + -ion transfer pathway and prevents the dissolution of the active material. Thus, the synergistic effect between the porous carbon substrates and the functional polymer coatings greatly advances the cycle performance and extends the long-term cyclability of the composite cathodes from less than 50 cycles to 500 cycles. I n view of the results of the intensive research in this area, we may conclude that the encapsulation of sulfur within a nanoporous substrate is a good approach to make advanced Li–S batteries. Based on the physical adsorption/absorption or chemical anchoring/trapping capability, these nanocomposites in different contexts suppress the polysul? de dissolution and diffusion issue. It should also be emphasized that the integra-tion of conductive and porous substrates with a composited nanospace, heteroatom doping, and functional surface coatings shows even better electrochemical performance than the use of these individual approaches. H owever, these approaches may still be ineffective in stopping polysul? des that carry negative

PROGRESS REPORT

charges from constantly migrating toward the lithium anode during cell discharge. This may result in fast capacity fading during the initial several cycles, which usually happens in composite cathodes. Moreover, although conductive additives improve the capacity and cyclability, these electrochemically inert materials reduce the gravimetric energy density and limit the sulfur content and loading in Li–S batteries. In many cases, the total sulfur content in the cathode region is not more than 65 wt% by weight and the sulfur loading is not higher than

2 mg cm ?2 of the cathode. [ 35 ]

2.1.4. S maller Sulfur Molecule Encapsulation

T he developments regarding microporous carbon synthesis and sulfur-encapsulation techniques have opened another useful research direction to limit the rapid capacity fading by avoiding the formation of soluble polysul? de intermediates (Li 2S n

,n = 4–8). The concept is different from the conven-tional methods that mainly depend on chemical sulfur–carbon/

polymer bonding or physical polysul? de-absorption/adsorption capacity of the porous hosts. By using smaller sulfur molecules (S 2?4 )

, the aim is to con? ne them in the con? ned space of a conductive microporous carbon matrix as the starting active material. As a result, this approach may theoretically eliminate the formation of soluble polysul? des and improve the close contact between S 2?4 and the conductive carbon host. [ 4a ,10a ] Z hang et al. prepared sulfur–carbon spherical composites and encapsulated sulfur into the micropores of carbon. [ 36a ]An electrochemical stability up to 500 cycles and a superior high-rate performance were obtained, although the sulfur content was only 42 wt%. Xin et al. encapsulated smaller sulfur mol-ecules into a core/porous-sheath structure comprising a com-posited CNT@MPC matrix. The small sulfur molecules were shown to be loaded into the microporous carbon, which had a critical pore size of 0.5 nm. The encapsulated metastable small sulfur molecules (S 2–4 )

were tightly held in the con? ned nano-space of the conductive microporous carbon matrix. This lim-ited the formation of cyclo-S 8 molecules with the dimensions of 0.7 nm. The resulting small-sulfur-molecule–microporous-carbon-coated CNT (CNT@MPC) nanocomposites avoid the unfavorable transformation between cyclo-S 8 and S 4 2?.Elec-trochemical analyses provide convincing evidence that the small-sulfur-molecule–carbon nanocomposites show a single discharge plateau at 1.9 V (the lower-discharge plateau) and a single reduction peak in the cyclic voltammetry curves. This provides evidence that the discharge/charge process effectively avoids polysul? de formation and hence limits the loss of active material. As a result, a high reversible capacity of 1149 mA h g ?1 was obtained after 200 cycles. X in et al. continuously enhanced this concept by using a hollow porous carbon substrate to encapsulate the smaller

sulfur molecules. [ 37 ] However, they pointed out a challenge with this method. The previously used microporous carbon may not have suf? cient pore volume to host more than 50 wt% of the smaller sulfur molecules in the resulting nanocomposites. Thus, hollow micro-/mesoporous carbons are used as the new substrate. The hollow porous carbon has a microporous carbon shell to accommodate and sequester chain-like smaller sulfur molecules, which can prevent the loss of active material. In addition, the hollow carbon has a hierarchical porous struc-ture to facilitate electron and Li + -ion transport, which leads to

better and faster electrochemical reaction as compared with

the group’s previous work. [ 10a ,36b ] Excellent long-term cyclability

with a high reversible capacity of 730 mA h g ?1 after 600 cycles

at a C/10 rate and with a limited self-discharge for 30 days were achieved. Ye et al. further improved the application of smaller sulfur molecules by using a carbon host with a hierarchical

micro-/mesoporous structure. [ 23c ] The hierarchical carbon

matrix not only has abundant micropores to store chain-like sulfur molecules in the composite, but also has mesopores to

ensure fast Li + -ion transport. The limited active material loss and fast electrochemical kinetics show a long lifespan of 800

cycles with a reversible capacity of 600 mA h g ?1 at a 1C rate.

A series of research concludes that the concept of using chain-like small sulfur molecules is a promising approach to improve cell performance. The small sulfur molecules are con-? ned within the limited nanospace of the micropores. The lim-ited nanospace and the strong interaction between the active material and carbon substrate avoid the formation of unfa-vorable soluble polysul? des and subsequent polysul? de disso-lution and diffusion. Close contact with the conductive carbon host further improves the activity of the active material. Thus, an effective cooperation between the smaller sulfur molecule and the microporous carbon may overcome the severe poly-sul? de diffusion problem in conventional Li–S cells. The use of micro-/mesoporous carbon has been shown to facilitate elec-tron and Li + -ion transference, as well as increase the loading of smaller sulfur molecules to above 50 wt%. As a reference,

this sulfur content refers to the content of active material in the nanocomposite. We may conclude that this new research

concept may be suitably applied to different microporous sub-strates (e.g., carbon, polymers, or oxides) with hierarchically tunable mesopores for developing practical Li–S batteries. 2.1.5. I norganic-Compound –S ulfur Composites A s an alternative to conductive carbon and polymers, inor-ganic compounds such as TiO 2 or Ti n

O 2 n ?1 have been used as inherently polar substrates to adsorb the sulfur species. Seh et al. developed sulfur–TiO 2 yolk–shell nanoarchitectures with

an internal void space to accommodate the volume expansion

of sulfur.

[ 38a ] The capacity degradation was as small as 0.033% per cycle for 1000 cycles. In addition, the conductive Magnéli phase Ti 4O 7 was revealed to be an effective matrix for binding/interacting with sulfur species, resulting in better accommoda-tion of Li 2 S

, a higher reversible capacity, and improved cycling performances.

[ 38b,c ] 2.2. P orous Polysul? de Reservoirs

T he application of porous substrates in composite cathodes to contain soluble polysul? des leads to another new cell con-? guration. A new containment strategy is to embed porous additives directly in the cathode matrix as an internal poly-sul? de reservoir. This promising concept was ? rst reported by

P R O G R E S S R E P O R T

Ji et al., as shown in F igure 3a ,b. [ 39 ] An ordered mesoporous

silica substrate (SBA-15) is embedded within the sulfur–carbon composite cathode for trapping and then storing the poly-sul? des formed during cycling. The polysul? de-trapping capa-bility of the porous additive is through surface absorption. The suppressed polysul? de diffusion issue further limits the rede-position of the insulating Li 2S /Li 2S 2 mixtures onto the surface of the cathodes, which mitigates the rapid capacity fading and short cycle life. An appealing merit of embedding the poly-sul? de reservoir within the cathode is that a small fraction of the porous substrates (10 wt% in the cathode) can greatly immobilize the migrating polysul? des within the composite cathode. The same research group further investigated the understanding of the polysul? de-trapping mechanism. Evers et al. presented the use of a porous TiO 2 additive (less than 4 wt% in the cathode) as a polysul? de reservoir in composite cathodes, which further increased the capacity retention and extended the

cycle life to 200 cycles. [ 40 ] They demonstrated that the soluble

polysul? de species are majorly absorbed within the nanoporous additives and minorly trapped by surface binding. R ecent advances in porous polysul? de reservoirs include extended cycle life and limited capacity fading. Wei et al. used a pig-bone-based hierarchical porous carbon (named BHPC) that was derived from pig bones as another appealing

polysul? de reservoir.

[ 41 ] After heat and KOH alkali treat-ments, the BH PC that was subsequently activated at 850 °C

achieved a high surface area of 2156 m

2g ?1 and a large pore volume of 2.26 cm 3g ?1 . The highly porous BH PC possesses

interconnected pores with a broad pore-size distribution, containing macropores, mesopores, and micropores. The

meso-/macropores facilitate the transport of Li +

ions and the micropores help retain the capacity during cell cycling. As a result, a small amount of BHPC additive (6 wt% in the cathode) improves the cycle stability and the utilization of sulfur during cycling. Our group utilized a carbonized-eggshell membrane (CEM) as a polysul? de absorbent with a pure sulfur cathode, as

shown in Figure 3 c . [ 42 ] The CEM polysul? de absorbent derived

from a natural eggshell membrane has a unique porous struc-ture with a high porosity and surface area for accommodating the soluble polysul? des. The suppressed polysul? de diffusion mitigates the rapid capacity loss and the redeposition of the insulating Li 2S /Li 2S 2

. Thus, the porous CEM provides good long-term cyclability (150 cycles) with a high capacity retention rate of 85% and a low capacity fading rate of 0.10% per cycle for Li–S cells. This recon? rms the progress on the develop-ment of the porous polysul? de reservoir initiated by Nazar and

co-workers. [ 39,40 ]

T he intensive research provides convincing evidence that this new cell con? guration facilitates good cyclability. Such an enhanced battery performance is achieved with less than 10 wt% additive in the cathode, which satis? es the criteria of high-performance and high sulfur content/loading. Future work may involve the investigation and enhancement of the interaction between porous polysul? de reservoirs (metal oxides or carbon substrates) and polysul? des to determine whether the polysul? de absorption is due to the porous architecture or via a physical/chemical effect. In addition, the characteristics of the pore size, surface area, and trapping capability of the sub-strates on the battery performance can be studied in detail to achieve future long-term electrochemical cycling.

F igure 3. a) Schematic model of the polysul? de reservoir. Cell performances of various polysul? de reservoirs: b) SBA-15 polysul? de reservoir and

c) CEM polysul? de reservoir. a,b) Reproduced with permission.

[ 39 ] Copyright 2011, Macmillan Publishers Limited. c) Reproduced with permission. [ 42 ] Copyright 2014, American Chemical Society.

PROGRESS REPORT

2.3. P orous Current Collectors

T he conventional current collector used in Li–S battery research is a 2D aluminum foil, which is just a ? at supporter in the cathode. Moreover, the aluminum foil may encounter oxidation and corrosion during cycling, causing the sulfur to lose electrical contact with the current collector and increasing the internal resistance of the battery. [ 43 ]Therefore, appropriate, alternative current collectors are of great interest for long-term cycle stability and high energy density. A new cell con? gura-tion design that uses a 3D conductive/porous-metal current collector has demonstrated improvements with regard to cell cyclability for various rechargeable battery systems. [ 44 ]

I n the Li–S battery scenario, the conductive/porous matrix not only works as an inner conductive framework to guarantee fast electron transport, but also as an active material container to stabilize the active material mixtures in its conductive skel-eton. This design enables superior electrochemical stability of the sulfur cathodes with a capacity retention rate of 92% after 50 cycles. [ 45 ]Ballauff and co-workers conducted quantitative analysis on the capacity fading of Li–S batteries with different cell con? gurations. [ 46 ]According to the capacity-fading model, they concluded that porous-metal current collectors have posi-tive effects on the long-term cycle stability of batteries. The improved cycle stability may result from the stronger interac-tion between the metal substrates and sulfur species. [ 47 ]

H owever, the weight of the metal-foam substrate may incur concerns regarding reducing the low energy density. Therefore, lightweight porous carbon substrates are more promising cur-rent collectors for increasing the accessible reactive area and providing a stable morphology during cycling. [ 48 ]A nanocellular carbon consisting of carbon-nanofoam plates and interwoven carbon ? bers was used as a bifunctional current collector with a readily prepared pure sulfur cathode. [ 49 ]The nanofoam plates that were tightly attached onto the ?brous network served as reservoirs to store the active material and localize the dis-solved polysul?des. The coalescing carbon ? bers functioned as an embedded conductive network to improve the redox reaction. As a result, the bifunctional conductive/porous cur-rent collectors offer high discharge capacity and good cycle stability. Another example of progress regarding the develop-ment of porous current collectors is the use of a commercial H-030 carbon paper as the substrate with a pure sulfur cathode, as shown in F igure4a,b. [ 48a ]The highly porous H-030 carbon paper has a high porosity of 80% and low density of 0.4 g cm ?3. The abundant nanoporous space allows the pure sulfur cathode to achieve a high sulfur loading (2.3 mg cm ?2) and sulfur con-tent (80 wt%). The close connection between the active material and the conductive/porous current collector leads to a good rate capability with higher discharge capacity and longer cyclability, as compared with both the aluminum foil current collector and our previous porous current collector research. [ 45,49 ] T he advantage of embedding various conductive/porous substrates in sulfur cathodes is notable, which allows us to conclude that the porous current collector is a promising cell con?guration for Li–S batteries. Nevertheless, we would like to emphasize that porous carbon substrates offer better cycle performance than metal substrates. The difference between these selected substrates is dominated by their pore sizes and morphologies. We ? nd that porous metal substrates have large pores several tens of micrometers in size. After these large pores hold the active material, inactive cores may form agglom-erates of active material. On the other hand, porous carbon sub-strates have smaller pores less than 10 μm in size and also have

F igure 4. a) Schematic model of the porous current collector. b,c) Cell performances of various porous current collectors: porous H-030 carbon paper (b) and 3D aluminum foam/carbon nanotube scaffold (c). b) Reproduced with permission. [ 48a ]Copyright 2014, Elsevier. c) Reproduced with permis-sion. [ 51 ]Copyright 2014, Elsevier.

P R O G R E S S R E P O R T

a micro-/mesoporous structure, which decreases the forma-tion of large non-conductive agglomeration in their nanospace. Also, the nanospace of carbon substrates can swell the liquid electrolyte, which is a bene? t for stabilizing the electrochemical

reaction within the porous cathode. [ 48a ,49 ]

Z hang et al. investigated the effect of a nickel foam and a carbon-? ber cloth as the porous current collector in sulfur–

polypyrrole composite cathodes. [ 50 ] Although the carbon sub-strate has lower electrical conductivity than the metal one, the

cell with the carbon-? ber cloth has more stable electrochemical performance compared with that of nickel foam. The reported advantages of carbon ? bers result from its chemical, struc-tural, and morphological characteristics, which improves the af? nity toward sulfur–polypyrrole nanocomposites and there-fore enhances stable charge-transfer conditions over that of the metal counterpart. This result recon? rms the observations and ? ndings in our current collector research. C heng et al. designed a 3D aluminum-foam/CNT scaffold that integrates the advantages of both the metal substrate and

the CNT network, as shown in Figure 4 c .

[51a ] The combination of sulfur–CNT mixtures and aluminum foam not only builds long- and short-range electron pathways, but also creates a vast void space in the sulfur cathode. Thus, the composited metal-foam/CNT scaffolds possess conductive pathways for facili-tating ef? cient electron transport and an abundant porosity

for approaching a high sulfur loading of 7.0 mg cm

?2.The high-loading sulfur cathode delivers a high initial discharge

capacity of 860 mA h g ?1 . H owever, the cell could only cycle

for 15 stable cycles. Moreover, although the reported max-imum permitted sulfur loading in the cathodes approaches

12.5 mg cm ?2 , the initial discharge capacity drops down to 3.12 mA h cm ?2 , which is equal to less than 300 mA h g

?1.This implies that more efforts are needed to optimize the cyclability and feasibility of sulfur cathode with such a high active mate-rial loading. Zhou et al. utilized a graphene foam electrode to

host a high sulfur loading of 3.3–10.1 mg cm

?2,successfully realizing a high areal capacity of 13.4 mA h cm

?2.[51b ]The interconnected pores of the graphene-foam current collector not only store a high amount of the active material, but also prevent the loss of the encapsulated active material during cell cycling. As a result, the cell employing the S–graphene-foam electrode displayed a low capacity fading rate of 0.07% per cycle for 1000 cycles. T he bifunctional current collectors embedded within the pure sulfur cathode have shown that they can contain the active material and improve the electrolyte absorption of the resulting cathode. This ensures close contact between the insulating active material and the conductive matrix. The porous current collector inherently has a high mechanical strength to ensure the complete electrode structure is retained during cycling. As such, a stable and fast electrochemical redox reaction is guaran-teed. In addition, the use of a highly porous substrate further allows the sulfur loading in the porous carbon current collector

to approach 2–10 mg cm ?2 , which implies a promising gravi-metric energy density. However, the volumetric energy density

of the Li–S cell is strongly related to the thickness of the porous carbon current collector. Furthermore, the use of a rigid and fragile substrate may limit the ? exibility of the cell, which should be avoided in future developments.

2.4. B inders

I n addition to the current collector, it may be necessary to cus-tomize other cell components for Li–S batteries. Conventional

polytetra? uoroethylene (PTFE) or polyvinylidene di? uoride (PVdF) binders are used to link sulfur particles or sulfur-based nanocomposites with the conductive carbon and the current collector. Although they are chemically stable during cycling, conventional binders can neither effectively tolerate the huge volume changes occurring during cycling nor suppress the polysul? de dissolution and migration. Thus, alternative binders that can create robust electrode architecture and possess polysul? de-retention capability have been considered for Li–S

batteries. [ 4a ,11b ]

2.4.1. A lternative Binders

S him et al. first investigated sulfur cathodes with a

poly(ethylene oxide) (PEO) binder with different mixing

processes, ball milling, and mechanical stirring methods.

[7] The study indicated that the preparation methods affect the morphology of the PEO binder and the porosity in the sulfur cathodes, which influences the cycle performance. A cationic polyelectrolyte binder, poly(acrylamide- c o -diallyldimethyl-ammonium chloride) (AMAC), was used with high-loading sulfur electrodes (sulfur content of 80 wt%), as reported by

Zhang.

[52] AMAC is insoluble in organic solvents. There-fore, these electrodes with the AMAC binder retained a porous and complete structure during cycling and hence exhibited better cyclability than that with the PEO binder. Zhang further points out a problem of using the AMAC single-binder. The AMAC dissociates chloride ions and then creates galvanic cells between sulfur particles and the aluminum metal in the process of aqueous slurry coating. This leads to the pitting corrosion of the metal current col-lector. The solution is to introduce a dual-binder approach with poly(acrylonitrile–methyl methacrylate) (ANMMA) as the second binder to bind the cathode mixtures to the cur-rent collector, which has been proven to be effective in elimi-nating the corrosion issue. Sun et al. used a natural gelatin

polymer as the binder in cathode preparation. [53] A gelatin

binder has a high adhesion ability, ensuring the structure stability of the sulfur cathode, as well as a good dispersion ability, mitigating the aggregation of the active material during cathode preparation and cell cycling. Other alterna-tive binders that can enhance the adhesion among cathode mixtures, or suppress the agglomeration of cathode material

mixtures, have also been reported.

[53,54] A good binder in cathode preparation should not only be a high-adhesion agent but also be a strong dispersion medium. This is bene? cial for having a complete electrode structure and favors the uniform distribution between the insulating sulfur particles and the conductive carbon additives, which ensures a good electrical contact and results in high sulfur utilization. Although these newly developed binders seem to have better battery performance than the conventional binder, stable long-term cyclability cannot be found with them. This implies that good polysul? de-retention capability and long-term

PROGRESS REPORT

electrochemical cycling may be an essential factor in the binder design. 2.4.2. F unctional Binders

R ecently investigated functional binders include PEO, PEG, and F-127 block copolymer. Lacey et al. ? rst investigated the role of PEO and PEG in cathodes. [ 55 ] PEO and related deriva-tives of PEG were used as binders, coatings, or electrolyte addi-tives in Li–S cells to study the cell performance. This research clearly demonstrated that a common and similar performance improvement can be obtained with all three approaches. The enhanced electrochemical reversibility during cell discharge and the suppressed cathode passivation at the end of dis-charge result from the polyether chains, which can modify the solvent system at the interface. The PEO binder or the surface cathode coatings are found to dissolve or swell in the liquid electrolyte. The ether-modi? ed solvent interface delays the precipitation of insoluble discharge products and hence affords an improvement in the reutilization of the active mate-rial. [ 56 ] Our group used the Pluronic F-127 block copolymer in a sulfur–microporous carbon composite cathode preparation, as shown in F igure 5 . [ 57 ] The Pluronic F-127 partially replaces PVdF binder in the electrode to provide a polysul? de-retention capability. This is because Pluronic F-127 is an amphiphilic copolymer that contains a hydrophobic poly(propylene oxide) (PPO) middle block and two hydrophilic PEO end blocks. The F-127copolymer uses the hydrophobic PPO block to adhere to the hydrophobic sulfur in nanocomposites and utilizes the hydrophilic PEO blocks to create a chemical gradient, limiting the severe diffusion of polysul? des out of the cathode. As a result, the Pluronic copolymer not only maintains a uniform electrode structure during cycling, but also reduces the dissolu-tion of polysul? des, and further limits the formation of a dense

Li 2 S

inactivation layer on the cathode. [ 13 ] In addition, PEO can facilitate Li + -ion transference by an electrostatic coordination of

Li ions with ether oxygen atoms. [ 58 ] Therefore, cathodes using a

10 wt% Pluronic F-127 copolymer achieve an excellent electro-chemical stability at high 1C and 2C rates and even allow the rate capability to approach a 4C rate. T he progress of binder development has extended the cycle life to approach 100 cycles and improve the rate capability to a 4C rate. This emphasizes that the binder accounting for about 5–20% by weight in the cathode material mixtures should be well designed for Li–S batteries. In addition to guaranteeing a complete structure of the electrodes and a uniform distribu-tion of the cathode material mixtures, we may conclude that the functional binder in the future should be low cost and have a low electrical resistance, as well as have a strong chemical poly-sul? de-trapping capability to ensure a long cell cycle life.

2.5. F ree-Standing Sulfur -B ased Composite Electrodes I n addition to modifying the conventional cell components, another new approach is to remove inactive materials from the cell by directly employing a sulfur-based nanocomposite as a free-standing composite cathode. The sulfur-based com-posite electrode is a novel binder-free and current collector-free

electrode design. The direct application of the composite elec-trode not only eliminates the bulk resistance from the added binder but also decreases the net weight of the electrode. As a reference, the weight of the binder is about 5–20% by the weight in the cathode material mixture. The weight of the con-ventional aluminum foil current collector accounts for about

15–20% of a battery, by weight. [ 43a ] Progress on free-standing

composite cathode is focused on the development of highly

conductive substrates with a light weight and a high porosity. The most essential requirement is that the applied conductive/

porous substrate must have either a free-standing shape or a self-weaving characteristic, which is important to guarantee

its normal function as an electrode. Moreover, the ?

exible and robust substrate should retain its complete structure after impregnating the active material and during cell cycling.

2.5.1. T emplated Assembly of Composite Electrodes E lazari et al. presented the ? rst promising free-standing com-posite electrode, which is a sulfur-impregnated activated

carbon-? ber cloth ( F igure 6a ). [ 59 ] The activated carbon ? ber

cloth disks (Kynol 2000) have a high surface area of 2000 m 2g ?1

and a total pore volume of 1 cm 3g ?1 . The carbon disks are

F igure 5. Schematic model and cell performance of the Pluronic F-127

block copolymer. Reproduced with permission. [ 57 ] Copyright 2012, Amer-ican Chemical Society.

P R O G R E S S R E P O R T

overlaid with sulfur for pre-impregnation at 150 °C and then heated to 155 °C for 10–15 h to impregnate the microporous nanospace with melting sulfur. After the active-material-encap-sulation process, the activated carbon ? ber tightly absorbs the dispersed sulfur into its micropores (<2 nm) and enables excel-lent electrolyte penetration in its highly porous structure. As a result, the porous carbon-? ber cloth decreases the diffusion of the reduction products of sulfur and offers improved electro-activity and cyclability. Although the sulfur loading is as high

as 6.5 mg cm

?2 because of the binder-free electrode design, the weight of the applied substrate seems to be the main concern of this kind of free-standing composite electrode system. D or? er et al. had another appealing free-standing composite electrode system, which is a vertical aligned CNT (VACNT)/

sulfur composite cathode. [ 60 ] The VACNTs grown on a metal foil

via chemical vapor deposition were directly used as a binder-free porous carbon substrate for sulfur encapsulation. The

VACNTs have a density of 0.06–0.13 g cm

?3 and abundant free volume (94 vol%) for achieving high sulfur contents of 70 wt%. The direct connection between CNTs and the nickel metal foil offers a highly conductive and stable framework. In addition, the use of lighter aluminum or carbon substrates in the chem-ical vapor deposition enables improvement by decreasing the device weight and the cost of the resulting electrodes. H agen et al. achieved progress in the high sulfur loading of

VACNT electrodes.

[ 61 ] The sulfur content and sulfur loading in the electrode approach were, respectively, 90 wt% and

7.1 mg cm ?2

. The high electronic conductivity and a high integrity of the conductive network make the VACNTs an out-standing electrode substrate, which allows the use of a high-sulfur-loading composite electrode and leads to a high discharge capacity. Moon et al. designed a novel encapsulated sulfur array in CNTs by using anodic aluminum oxide (AAO) membranes

as the template, as shown in Figure 6 b .

[ 62 ] The accurate sulfur location in continuous CNTs with ultrathin wall thicknesses (10 nm) led to a high sulfur content of above 80 wt% with a

capacity of 863 mA h g

?1 and a high capacity retention rate of 76% after 1000 cycles. The rate capability also demonstrated a superior cell performance with a reversible capacity of

1223 mA h g

?1 after 150 cycles at a C/2 rate and 960 mA h g ?1 after 300 cycles at a 20C rate. After the initial cycle, the average Coulombic ef? ciency is above 98% without the use of LiNO 3 in the electrolyte. The outstanding cyclability (1000 cycles) and rate capability (C/2–40C rates) are attributed to the com-plete encapsulation of monoclinic sulfur, which physically limits the loss of active material. The encapsulated monoclinic sulfur further chemically increases the structural stability of the composite electrode by reducing the volume change during

cycling.

[ 63 ] This achievement greatly promotes the cycle stability of the sulfur cathode design. A lthough these free-standing composite cathodes boost the cell performance by developing a novel electrode architecture without the use of a binder or current collector, the weight of the substrates may need to be considered in calculating the sulfur loading and sulfur content. In addition, the sulfur-encap-sulation process and the template for the substrate will also dominate the feasibility of mass production of free-standing composite electrodes.

2.5.2. S elf-Supporting Composite Electrodes

U sing a self-supporting composite electrode is another approach for making free-standing cathodes. The preparation

F igure 6. SEM observation or schematic model and cell performance of various free-standing sulfur-based composite electrodes: a) sulfur-impreg-nated activated carbon ? ber cloth, scale bar = 10 μm, b) S@C NW electrode, and c) self-weaving sulfur–carbon composite cathode. a) Reproduced

with permission. [ 59 ] Copyright 2011, Wiley-VCH. b) Reproduced with permission.

[ 62 ] Copyright 2013, Wiley-VCH. c) Reproduced with permission. [ 65 ] Copyright 2012, Royal Society of Chemistry.

PROGRESS REPORT

process of these electrodes greatly depends on the physical mor-phology of the starting materials. For example, a free-standing graphene and a self-weaving MWCNT are appealing electrodes that allow no binder or current collector to be used in Li–S bat-teries. Free-standing ? exible sulfur–graphene composite elec-trodes extend the application of graphene from two dimensions to three dimensions. The use of sulfur vapor treatment or an in situ redox-reaction approach introduces active sulfur homoge-neously into graphene paper. The graphene framework not only offers fast electron transference and electrolyte immersion, but also functions as a supporting framework to store the active

material. [ 64 ] The self-weaving behavior originating from the

CNTs or the nanowire helps to form a free-standing composite electrode. Figure 6 c shows our self-weaving sulfur–carbon

composite cathode. [ 65 ] In the composite electrode, sulfur was

deposited onto MWCNTs by a water-based solution method. After ? ltration, the dispersed sulfur–MWCNT composite, with a sulfur content of 63 wt%, forms a free-standing thin ? lm that can be peeled off the ? lter paper after oven drying. Using a conductive MWCNT skeleton improves the active material utilization at high cycling rates. The excellent absorp-tion ability of the MWCNT framework localizes the electrolyte and suppresses the migration of dissolved polysul? des. The tortuous interspaces in the MWCNT network provide a stable accommodation of sulfur for rapid redox reactions and prevent sulfur/lithium sul? de (Li 2

S ) from forming extensive agglom-erates during cycling, leading to good cyclability.

[ 28,66 ] Jin et al. reported another free-standing sulfur–CNT composite elec-trode modi? ed by coating sulfur onto the oxidized CNF ? lm

via a heating process. [ 67a ] The resulting S/CNT ? exible cathode,

containing a high sulfur loading (5 mg cm ?2 ) and a high sulfur

content (65 wt%), exhibits a high reversible capacity after 100 cycles. The oxygenic groups on the oxidized CNTs form strong covalent bonds between sulfur and CNTs. This strong chemical bond ? rmly anchors the active material onto the conductive

CNT matrix, which enhances the cycle stability.

[ 19 ,67b ] Zhou et al. combined a thin-? lm coating technique and use of conduc-tive graphene to form an electrode with an integrated ? exible S–graphene–polypropylene separator. The integrated cathode–separator structure consists of an active material coating

(1.5–2.1 mg cm ?2 ) and a graphene coating (1.3 mg cm ?2 ) on a

commercial separator. The highly ? exible and robust composite electrode not only opens a new Li–S cell con? guration, but also

achieves long-term cycle stability. [ 68 ] It is noteworthy to say that

the calculated sulfur content approaches 50 wt% in the com-posited cathode–separator structure, implying the possibility of attaining a high sulfur content and loading by optimizing the manufacturing techniques. B oth free-standing graphene and self-weaving MWCNT can sequester high contents of sulfur in their nanoporous space and also guarantee a high utilization of the encapsulated active material. The sulfur content in the whole cathode region can approach 55–67 wt% (the calculation here includes the weight of electrode matrix). This implies that graphene and CNT membranes are expected to be practically useful for high-per-formance lithium–sulfur batteries. Direct sulfur encapsulation in the cathode has shown evidence to improve both the capacity and the cyclability. Over the past few years, researchers have reported several high-capacity composite cathodes approaching

reasonable sulfur content or sulfur loading. In the development of composite electrodes, the free-standing characteristics of these self-weaving or self-supporting papers make them novel electrodes. However, this characteristic also adds an extra engi-neering limitation for selection of the starting material. It has been indicated in the literature that only certain raw materials

may have the ability to form free-standing ? lms. [ 69 ]Moreover,

the free-standing thin ? lm has to take a balance between thick-ness/weight and mechanical strength. Therefore, follow-up developments should focus on the design of the substrate in terms of porous structure, morphology, and accessible pore volume. Optimization of the microstructure can increase the sulfur loading/content and improve the reaction kinetics during cycling. We may conclude that reducing the thickness of the free-standing matrix or the use of lightweight starting mate-rials as the ? exible and robust substrate can further improve the sulfur loading and sulfur utilization, leading to higher capacity, rate capability, and cycling ef?ciency. 2.6. I nterlayers

A dvanced cathode nanocomposites and novel composite elec-trodes utilize conductive and porous substrates in different ways to increase active-material utilization and suppress loss of the active material. H owever, theoretically, dissolved poly-sul? de anions will inevitably move toward the anode, driven by the chemical potential and concentration differences between the cathode and the anode during cell discharge. Accordingly, a polysul? de trap in between the sulfur cathode and the separator may be a suitable and an essential cell component for advanced Li–S cells to localize the polysul? de species at the cathode side of the cell. This concept was ? rst developed by our group as an

“interlayer”. [ 69 ]

T he free-standing interlayer has to be ? exible to provide smooth contact with the top surface of the cathode. In addi-tion, the interlayer needs to possess a porous structure or a large amount of accessible nanospace to store the shuttling polysul? de intermediates. As a result, the interlayer is enabled to work bifunctionally in the cell. First, the inserted interlayer functions as an upper-current collector, which can improve the ef? cient electron conduction by its high electrical conduc-tivity and fast ion transport through its abundant nanospace. Second, its nanospace further plays a more signi? cant role as the polysul? de-trapping site, which can effectively suppress the migration of dissolved polysul? des. A series of novel interlayer developments and relative analyses demonstrates that using an interlayer is a facile approach to provide Li–S cells with high electrochemical utilization and excellent cycle stability. In addi-tion, the interlayer further allows the use of the readily prepared

pure sulfur cathodes that contain high sulfur loading. [ 4a ,69 , 70 ]

2.6.1. C arbon Interlayers

A conductive MWCNT thin ? lm designed by our group was

? rst used as the bifunctional interlayer in Li–S cells, as shown in F igure 7a . [ 69 ] The free-standing MWCNT interlayer plays a key role in reducing the cathode resistance and in retaining

P R O G R E S S R E P O R T

the dissolved active material during cycling. First, the inter-woven MWCNTs create a superior electrically conductive

network.

[ 9a ,71 ] The increased electrical conductivity of the cathode promotes ef? cient active material utilization and stable high-rate cyclability. Second, the tortuous pores in the carbon paper can localize the migrating polysul? de species and retain the active material during repeated cycling. Furthermore, the ? exible MWCNT skeleton tolerates the volume changes of the trapped active material during cycling and ensures its contin-uous electron pathways for normally reactivating the trapped

active material.

[ 8c ,66 ] Thus, the MWCNT interlayer with only 40–50 μm thickness and around only 0.6–0.8 mg mass provides pure sulfur cathodes (sulfur content of 70 wt% in the whole cathode) with good high-rate performance from C/5 to 1C rates. This MWCNT-interlayer study, for the ? rst time, demonstrated that a novel cell con? guration can make impressive progress on pure sulfur cathodes for developing high-performance Li–S cells applying a facile preparation process. O ur group not only presented the new idea of the interlayer concept, but also extended the concept by studying the chem-ical polysul? de-trapping capability of an alcohol–alkaline/ther-mally treated carbon paper.

[ 72 ] The porous architecture of the treated Toray carbon paper interlayer also serves as a conduc-tive skeleton for trapping, localizing, and reactivating the dif-fusing polysul? des. In addition, the treated Toray carbon paper introduces hydroxyl functional groups onto the interlayer, enhancing its hydrophilicity. The hydrophilic functional groups create chemical gradients to chemically limit the diffusion of

polysul? des. [ 13 ] As a result, the improved cell performance

achieved a high initial discharge capacity of 1651 mA h g ?1.A high capacity retention of above 85% was obtained at various

cycling rates (C/5, C/2, and 1C rates). O n the other hand, the physical polysul? de-trapping capa-bility of the interlayer architecture was investigated through

a carbonized leaf interlayer.

[ 73 ] After carbonization treatment, the carbonized tree leaf maintained its unique structure. The inherent water-locking ? lm of natural leaves is now used to block the migrating polysul? des. The moisture reservoir of natural leaves is used to ensure proper electrolyte wetting of the inhibitor layer. The carbonized leaf interlayer covering on sulfur cathodes has the polysul? de-locking ? lm pointing toward the cathode to immediately suppress polysul? de diffu-sion and has the electrolyte reservoir facing the separator to absorb the electrolyte. The natural structural gradient and the micro-/mesoporous adsorption sites of the carbonized leaf interlayer not only intercept/trap the migrating polysul? des, but also facilitate their reutilization. Such a unique natural microstructure of the carbonized leaf interlayer leads to good long-term cyclability over 150 cycles with a reversible capacity

of 800 mA h g

?1 and a low capacity fading rate of 0.18% per cycle. A comparative analysis with the carbonized leaf interlayer placed in the reverse direction position (electrolyte reservoir pointing toward the cathode) shows a similar cycle stability but a relatively low reversible capacity. This is because the poly-sul? de locking-? lm is not in close contact with the cathode for trapping the dissolved polysul? des before they escape from the cathode. Thus, the dissolved polysul? des may freely diffuse out from the cathode and form agglomerates in the reservoir archi-tecture. This implies that the surface morphology of the inter-layer will physically in? uence the cell performance and their

F igure 7. Schematic model and cell performances of various bifunctional interlayers: a) microporous carbon interlayer and b) self-assembled polypyr-role nanotube ? lm. a) Reproduced with permission. [ 69 ] Copyright 2012, Macmillan Publishers Limited. b) Reproduced with permission. [ 82 ]Copyright 2015, Elsevier.

PROGRESS REPORT

relation; the corresponding mechanism needs to be examined in detail. A fter proving the effectiveness of the carbon interlayer, our group focused on systematically studying various inter-layers with speci? c microstructures for investigating the most essential structural parameters for promoting the interlayer design. The effect of the pore size was investigated by bimodal

micro-/mesoporous carbon paper interlayers. [ 70,74 ] A micro-porous carbon paper demonstrated that the micropores (less

than 1.5 nm) in the carbon interlayer enhance the polysul? de-trapping capability because the longest chain length esti-mated among the polysul? des (Li 2S 8 )

is around 2 nm. [ 70 ]Thus, after cycling, the trapped active material is found all over the microporous carbon interlayer and distributed homogenously. It is found that the uniformly distributed sulfur/sul? des do not form dense agglomerations on the surface of the cycled cathode. This ensures a good retention and reutilization of the

active material. [ 8a ,b ]

A comparative analysis with a bimodal micro-/mesoporous carbon paper (micropores: 1.5 nm, and mesopores: 6 nm) shows a relatively poor cycling performance compared with the microporous carbon paper (micropores: 1.5 nm) at 1C and 2C rates. This provides evidence that the microporous mate-rials may improve the polysul? de absorption because of the

similar dimensions of the micropores and the polysul? des.

[ 70 ] This analysis demonstrates that large mesopores cannot trap the migrating polysul? des well during cell cycling because of the size effect. Although mesopores may not trap the migrating polysul? des in their nanospace well, mesopores offer other

bene? ts, such as facilitating Li +-ion transport [ 22b ] and suf? -cient electrolyte accessibility during high-rate electrochemical

reactions.

[ 74 ] To demonstrate these concepts, we applied the bimodal micro-/mesoporous carbon paper in high-rate Li–S cells utilizing the sulfur/long-chain polysul? de redox couples. This research con? rmed that the bimodal micro-/mesoporous nanospace in the interlayer can hold the shuttling polysul? des and then offer suf? cient electrolyte accessibility. The intimate three-phase boundary, involving the active material, the conduc-tive network, and the electrolyte, ensures ef? cient sulfur utiliza-tion and high reversibility. Therefore, fast and highly reversible cathode redox reactions could be obtained in Li–S batteries at ultrahigh 10C and 15C rates over 250 cycles. O n the other hand, the thickness of the inserted interlayer was investigated by a modulable carbonized Kimwipe interlayer

system.

[ 75 ] This modulable interlayer offers a way to control the thickness by simply adjusting the number of folded layers (1-, 3-, or 6-layered architecture) in each carbonized Kimwipe inter-layer. As a reference, the weights of the 1-, 3-, or 6-layer sam-ples are 0.23, 0.68, and 1.3 mg cm ?2 . The cathode resistance/

impedance decreases with increasing number of folded layers. This provides evidence that the additional conductive pathways in the cathode are bene? cial for facilitating fast electron trans-port, enhancing the utilization of the active material, as well as

improving the discharge capacity and the cycle stability.

[ 69 ]In addition, the increase in polysul? de-locking ? lms and absorp-tion sites in the layered module leads to the electrochemical reaction being further con? ned within the cathode region. As a result, with an optimized thickness, the cell employing the lightweight carbonized Kimwipe interlayer exhibited a high discharge capacity, excellent rate capability, and long cyclability. It is noteworthy to mention that the microstructural analysis on the cycled hierarchical carbonized Kimwipe interlayer (6-lay-ered architecture) shows the decrease in the sulfur concentra-tion gradient from the cathode side to the separator side of the interlayer. This demonstrates that an optimum thickness of interlayer is important to achieve a good cell performance and recon? rms that the optimized surface morphology of the inter-layer is the key factor for improving the polysul? de-trapping

capability. [ 73,75 ]

T he comprehensive understanding of the effect of the inter-layer structural parameters (thickness, pore size, surface area, and conductivity) on battery performance was investigated

by utilizing free-standing binder-free CNF mats. [ 76 ] The free-standing CNFs prepared by the carbonization of the electro-spun nano? bers provide easier architectural tunability (pore

size and thickness) compared with previous interlayer mate-rials. The research ? ndings recon? rm that an optimum thick-ness and morphology of ? brous interlayers is a signi? cant factor to achieve good cell performance. The comparative anal-ysis demonstrates that the meso-/microporous CNF interlayer with an optimum thickness and a high surface area delivers

a high initial discharge capacity of 1549 mA h g

?1 at C/5 rate with 83% capacity retention after 100 cycles. At a low C/5 rate, the discharge capacity is majorly affected by the thickness of the ? brous architecture interlayer. The presence of mesopores in meso-/microporous CNFs may further facilitate the reac-tivation of the trapped active material by channeling the elec-trolyte via the mesoporous pathways to the microporous trap-ping sites and may also transport the electrolyte containing dissolved polysul? des to the microporous trapping sites. On the other hand, at a high 1C rate, the battery performance may depend on the accessibility and reactivation possibility of the trapped active material in the ? brous interlayer, which in turn is affected by the accessible porous structure and conductivity of the interlayer. T he development of other interlayers has made signi? cant progress based on the innovation by our group. These novel bifunctional carbon interlayers involve a reduced graphene oxide thin ? lm, an acetylene black mesh, a carbonized ? lter paper interlayer, and a cassava-derived carbon sheet. Wang et al. reported a reduced graphene oxide ? lm as the shuttle-inhibiting

interlayer. [ 77 ] The resulting reduced graphene oxide ? lm inte-grates reduced graphene oxide and Ketjen Black carbon. The

functional epoxy and hydroxyl groups on the reduced graphene oxide sheets strongly anchor the sulfur and polysul? des onto

the carbon–carbon bonds. [ 19 ] It is interesting to ? nd a gradual

increase in the discharge capacity, which indicates an activation process is required from the reduced graphene oxide interlayer. The solution is to create porous pathways in the reduced gra-phene oxide ? lm by the Ketjen Black carbon additive. The loose layer architecture is bene? cial for electrolyte and polysul? des to permeate between each reduced graphene oxide sheet, which facilitates an active reaction, and hence improves the utilization of the active material. J eong et al. reported a free-standing acetylene-black mesh

interlayer. [ 78 ] They conducted a comparative analysis between

an acetylene-black mesh and a graphitic mesocarbon microbead (MCMB) mesh prepared by mixing carbon particles and a PTFE

P R O G R E S S R E P O R T

binder. The resulting acetylene-black mesh possesses abundant macropores, while the porous structure of the MCMB mesh is blocked by the polymer binder. As a result, the cell using the acetylene-black mesh showed an improved battery perfor-mance, whereas the cell using the MCMB mesh could not be well cycled. This indicates that a critical engineering problem must be considered in the subsequent development of the carbon-binder-mixture interlayer. The porous structure of the porous carbon particle or clusters may be blocked by the con-glomeration of the binder. Other bifunctional carbon interlayers are a carbonized ? lter paper interlayer and a cassava-derived carbon sheet, which improve the cell performance based on the

unique microstructure from the starting materials. [ 79 ]

2.6.2. P olymer and Metal Interlayers

I n addition to carbon materials, polymer interlayers also

enhance the cyclability of Li–S batteries by a strong adsorp-tion effect between polymers and polysul? des. Ma et al. ? rst reported a polypyrrole functional interlayer that was fabricated

in situ onto the sulfur cathode. [ 80 ] As a ? rst polymer inter-layer, the polypyrrole interlayer was reported to inhibit the

dissolution and migration of the polysul? de species owing to the adsorption effect of polypyrrole nanoparticles to dissolved polysul? des. Moreover, the conductive polymers have good electronic and ionic conductivity for improving the cathode

conductivity and enhancing the rate capability of the cell.

[ 81 ] Ma et al. further made impressive progress on the polypyrrole

functional interlayer.

[ 82 ] They reported a novel self-assembled polypyrrole nanotube ? lm (Figure 7 b ), different from their polypyrrole nanoparticles interlayer. The new polypyrrole nano-tube ? lm has higher speci? c area and larger amount of pores, which signi? cantly improves the sulfur utilization and cycla-bility of the cathode with high sulfur loading (sulfur loading

of 2.5–3 mg cm ?2 ). Also, the addition of polypyrrole nanotubes

interlayer can suppress the corrosion reaction on the surface of the lithium anode. As a result, the cell applying the polypyrrole nanotube interlayer achieved a high initial discharge capacity of

1102 mA h g

?1 and a reversible capacity of 712 mA h g ?1after 300 cycles. The Coulombic ef? ciency also increased to 92% without the use of LiNO 3 in the electrolyte. Finally, an electri-cally conductive nickel foam was used as a metal interlayer. [ 83 ] Although the metal substrate has the highest electrical conduc-tivity, as compared with carbon and conductive polymers, as well as more robust porous structure for accommodating the sulfur species, the added weight of the metal foam may impact its feasibility. I t can be concluded that the bifunctional-interlayer con-cept developed in our laboratory is an inexpensive, effective approach for suppressing polysul? de migration during cell

cycling. [ 69,70,73?76 ] The use of these bifunctional interlayers to

suppress polysul? de migration has led to signi? cant progress in the performance of Li–S cells. Subsequent design of novel interlayers may focus on using micro-/mesoporous substrates with high conductivity and abundant accessible nanospace for trapping the migrating polysul? des and for subsequently reactivating the trapped active material. The surface functional groups may further improve the polysul? de-trapping capability. In addition to pursuing the cell performance, attention should

be paid such that the overall weight or volume of the cell is not increased with the introduction of the interlayer and the energy density is not decreased. Furthermore, the reported interlayers were prepared by using self-weaving ? bers, carbon-ized thin-? lm materials, or carbon-binder mixtures. The ? rst two approaches require a unique physical morphology for their starting materials, which currently limits their application and

mass production. [ 4a,c,d ] The third approach may reduce the elec-tronic and ionic conductivities or even block the open pore due

to the conglomeration of the binder on the carbon particles or

clusters. [ 4d ,78 ] Therefore, a breakthrough regarding these engi-neering problems will be essential for the commercialization of

interlayer-type Li–S cells.

2.7. S urface-Coated Separators

A ttempts to reduce the interlayer weight so that the impact in

energy density can be minimized could lead to a decrease in the mechanical strength and increase the complexity during manu-facturing. The loose structure of a lightweight interlayer cannot ensure its normal function in cells during repeated cycling and

may reduce the polysul? de-trapping capability.

[ 75,76 ] In addition, the current interlayers need to be constructed with interwoven ? ber networks or carbon-binder mixtures in order to be “free-standing.” To overcome these engineering problems, we ? rst present the concept of a carbon-coated separator for employing pure sulfur cathodes in high-performance Li–S batteries. The carbon-coated separator has a lightweight carbon coating attached onto the conventional separator. The carbon coating facing the sulfur cathode functions as a polysul? de trap and as an upper current collector to intercept the migrating active material and achieve an ef? cient reutilization of the trapped active material during long-term cycling. The conventional sep-arator serves as an electrically insulating membrane to facilitate

the ? ow of electrolyte and Li

+ ions but blocks electron trans-port. Furthermore, it functions as a highly robust substrate to support the coating layer, providing the carbon-coated sepa-rator with ? exibility and outstanding mechanical strength. In general, the carbon-coated separator exhibits the advantages of an interlayer and has been shown as an outstanding “contain-ment building” to suppress the “polysul? de leak.” In addition, the lightweight coating excludes weight and thickness concerns of the interlayer and the low sulfur content problems of sulfur-based nanocomposites. [ 35a,b, 84 ]

2.7.1. C arbon-Coated Separators

A Super P carbon-coated separator consisting of a layer of

Super P thin ? lm coated on one side of the commercial Cel-gard polymeric separator via the tape-casting method has been reported by our group, as shown in F igure 8a . [ 84 ] The carbon coating pointing toward the pure sulfur cathode is aimed at intercepting the migrating polysul? des immediately before they diffuse through the Celgard separator. The trapped poly-sul? des may be immobilized by the porous carbon coating, which consists of carbon nanoparticle clusters. In addition to

PROGRESS REPORT

suppressing the polysul?de migration, the conductive-carbon coating as an upper current collector facilitates fast electron transport in the insulating sulfur cathode. During long-term cycling, this upper current collector easily transports electrons into the trapped active material to reactivate them. Therefore, effective active material reutilization is accomplished during long-term cycling. It is also found that the carbon-coated sep-arator effectively limits the self-discharge effect and shows excellent capacity retention after a 3 month rest period. This is because the carbon coating, which has a smooth contact

F igure 8. Schematic model and cell performances of various surface-coated separators: a) Super P carbon-coated separator, b) Na? on-coated sepa-rator, c) PEG/MPC-coated separator, and d) Al 2O3ceramic-coated separator. a) Reproduced with permission. [ 84 ]Copyright 2014, Wiley-VCH. b) Repro-duced with permission. [ 91 ]Copyright 2014, Royal Society of Chemistry. c) Reproduced with permission. [ 94 ]Copyright 2014, Wiley-VCH. d) Reproduced with permission. [ 97 ]Copyright 2014, Elsevier.

P R O G R E S S R E P O R T

with the top surface of cathode, also functions as a protection layer to suppress the loss of active material during cell resting. Hence, calculations based on a mathematical model available in the literature provide evidence that the carbon-coated separator achieves the lowest self-discharge constant than other attempts

in Li–S technology. [ 45,85 ]

A MWCNT-coated separator provides a bundled/porous

MWCNT ? lter on the Celgard separator. [ 86 ] The MWCNTs,

with an outer diameter of 15–45 nm and a length of 5–20 μm, were coated onto the Celgard separator by vacuum ? ltration. During cell discharge, the MWCNT coating ? rst blocks the freely migrating polysul? des and then absorbs the intercepted polysul? de species in its nanospace (a high surface area of

410 m 2g ?1 ). The suppressed polysul? de diffusion effect is then analyzed by investigating the changes in the upper discharge plateaus and their corresponding discharge capacities ( Q H ,the-oretical capacity = 419 mA h g ?1). [ 87 ] It is well known in the lit-erature that this upper discharge plateau region corresponds to

the formation and existence of highly soluble polysul? des. [ 10a ,85c ]

Thus, the completeness of these overlapping upper discharge plateaus and the highly reversible upper plateau discharge capacity demonstrate that severe polysul? de diffusion has been suppressed by the MWCNT coating. In our subsequent studies on different carbon materials with different physical morpholo-gies, in general, the vacuum ? ltration process is suitable for ? brous nanosized carbon materials, such as MWCNTs, CNTs, and ? brous webs. The tape-casting method works for spherical carbon powders, such as Super P conductive carbon black and microporous carbon particles. The graphene-sheet-coated sepa-rator can be prepared by both approaches.

[ 88 ] Z hou et al. reported a thorough comparative analysis on dif-ferent novel cell con? gurations, including graphene current collectors, graphene-coated separators, and sandwiched gra-phene electrodes.

[ 89 ] Comparison between the two initial cell con? gurations clearly indicates that the cell with the aluminum foil current collector and the graphene-coated separator shows a better cycle performance than that with the graphene current collector and the commercial separator. This suggests that Li–S batteries may require a conductive and porous polysul? de trap to suppress the migration of polysul? des and thereby mitigate its side effects. On the other hand, a porous current collector may be majorly used for raising the sulfur loading due to its porous nanospace and electrically conductive framework.

2.7.2. P olymer-Coated Separators

A nother example of signi? cant progress with regard to the functional separator is the use of ion-selective membranes. The-oretically, a cation-selective membrane should solve the poly-sul? de shuttle issue by allowing the transport of Li + cations and stopping the diffusion of polysul? de anions. A lithiated Na? on ionomer ? lm was ? rst used by Jin et al. as a functional sepa-rator. [ 90 ] The lithiated Na? on ionomer ? lm can transfer Li +ions

and prevent the polysul? de anions from transporting through. The limited polysul? de diffusion improves the Coulombic ef? ciency up to 97% without the addition of a LiNO 3co-salt.

Although the high discharge/charge ef? ciency demonstrates the limited shuttle effect, the lithiated Na? on ionomer ? lm cannot effectively mitigate the fast capacity fading. The capacity

retention rates of cells using the lithiated Na? on ionomer ? lm with an ionomer electrolyte or a liquid electrolyte were, respec-tively, 69% and 47% for only 50 cycles. The low capacity reten-tion results from the formation of a passivation layer covering on the surface of the cathode, which creates the inactive zone

and shortens the cycle life. [ 8b ] Then, Huang et al. improved the

application of the cation-selective Na? on. [ 91 ] A Na? on-PP/PE/

PP composite separator consisting of a cation-shield on the commercial Celgard separator (Figure 8 b ) was combined with sulfur–CNT nanocomposites to prevent the polysul? de diffu-sion. The Na? on coating functions as the cation-selective mem-brane, which allows the free transportation of Li + cations and blocks the shuttling of polysul? des by Coulombic interactions. Therefore, the resulting cell shows signi? cant progress, not only on the Coulombic ef? ciency (high ef? ciency above 90%), but also on the capacity retention (a low capacity fading rate of 0.08% per cycle) for 500 cycles. The progress on the use of the Na? on ionomer ? lm or the Na? on coating results from the for-mation of an electrostatic shield for immobilizing polysul? de

anions within the cathode region of the cell. When Li + ions dis-sociate from the side chains, the SO 3 ? groups form an electric

force ? eld, allowing the transport of Li

+ cations and preventing the migration of polysul? de anions.

[ 92 ] Moreover, the channel and cluster structure of the ionomer coating layer serves as an

additional transport barrier. These SO 3 ?-group-coated channels

also block polysul? de anions by Coulombic interactions.

[ 93 ] A lthough an outstanding long-term cyclability and cycle sta-bility are presented by these studies, some engineering concerns may need to be solved in follow-up developments. The coating

layer on this surface-coated separator is 0.7 mg cm

?2,which may decrease the sulfur content of the cathode as we consider

its weight. The reported sulfur loading is 0.53 mg cm

?2with a sulfur content of 50 wt%. Thus, the calculated overall sulfur content is 30.1 wt%. In addition, the comparative analysis indi-cates that a lighter thin-? lm Na? on coating (0.15 mg cm

?2)cannot form a complete compact layer on the commercial sepa-rator, and hence shows a limited improvement in the battery

performance. [ 91 ] Also, the polarization and resistance from the

Na? on ionomer ? lm or coating may decrease the utilization of

the active material.

[ 90,91 ] 2.7.3. F unctional-Polymer -S upported Carbon -C oated Separators T he carbon coating layer involves various lightweight sub-strates with a highly tunable pore size. Thus, a carbon coating

possesses the essential morphology for physically ? ltering the dissolved polysul? des and suppressing the diffusion of poly-sul? des. A functional polymer coating has stronger chemical

polysul? de-trapping capability than the carbon counterpart.

However, the added weight and impedance from polymers may

decrease the sulfur content and impede the effective utilization of the active material. In addition, a non-conductive coating would restrict the reactivation and reutilization of the trapped active material, which causes continuous capacity fading.

A compromise is to create a polymer-supported carbon coating on a commercial separator, which integrates the phys-ical and chemical polysul? de-trapping agents. A composite

PROGRESS REPORT

separator with a PEG-supported microporous carbon coating as the polysul? de trap has been developed for Li–S batteries,

as shown in Figure 8 c . [ 94 ] The microporous-carbon/PEG-coated

separator thus facilitates the use of pure sulfur cathodes in high-performance Li–S batteries. The microporous carbon substrates function as a physical polysul? de ? lter to intercept the migrating polysul? de intermediates (chain length = 1.0 to

1.8 nm with 4 < x ≤ 8) [ 95 ] by their micropores (micropore size =

0.5 nm and 1.2 nm). The PEG polymers serve as a binder and a chemical polysul? de barrier. The polysul? des react preferen-tially with the PEG chains rather than with the bulk electrolyte

due to the hydrophilicity of the PEG. [ 13,56 ] In addition, the PEG

binder improves the adhesion between the microporous carbon particles and the commercial separator, which allows the thick-ness and the weight of the coating layer to be decreased. The microporous-carbon/PEG-coated separator offers Li–S cells utilizing pure sulfur cathodes as having a high electrochemical utilization (80%), long-term cyclability (500 cycles), and high cycle stability (0.11% capacity fade per cycle). This provides solid evidence that a synergistic effect between the physical and chemical polysul? de-trapping capabilities can suppress poly-sul? de diffusion and enable the use of high sulfur content. W e further compare the cycling performances of cells employing the microporous-carbon/PEG-coated separator with those employing composite cathodes. The microporous-carbon/PEG-coated separator shows better polysul? de-trapping capability and, therefore, has a great potential to be used in an advanced cell con? guration in Li–S cells. For example, an irre-versible and rapid decrease in capacity during the initial cycles is

commonly found in sulfur-based composite cathodes.

[ 23g ,25d ,96 ] This capacity fading is possibly due to loss of active material from nanocomposites or the redeposition of non-conductive

aggregates on the surface of the cycled composite cathode.

[ 13,49 ] The same phenomenon of severe polysul? de migration and the redeposition of the Li 2S 2/Li 2 S mixtures formed from the diffusing polysul? des could not be observed in the cell using the microporous-carbon/PEG-coated separator. This indicates the advanced composite separator not only solves this rapid capacity-fading problem but also may act as a better “contain-ment building” for suppressing the “polysul? de leak.”

2.7.4. C eramic -C oated Separator A nother modi? cation of the composite separator is the design

of a ceramic-coated separator. Zhang et al. reported the use of an Al 2O 3

ceramic absorbent layer on a Celgard separator, func-tioning as a physical obstacle to block the smooth migration

of the dissolved polysul? des, as shown in Figure 8 d .

[ 97 ]The Al 2O 3 ceramic coating, consisting of connected Al 2O 3

nano-particles, creates a porous barrier region between the sulfur cathode and the separator. These nanopores offer pathways

for Li + -ion transference but not for polysul? de diffusion. The polysul? de-trapping capability of the ceramic coating includes physical absorption during intermediate discharging and elec-trochemical deposition at the end of discharge. H owever, the Al 2O 3

ceramic coating offers a reversible capacity of less than 600 mA h g ?1 after 50 cycles. This phenomenon implies that

the ceramic coating may have the same challenges as those of

a polymer coating. The added weight and resistance may result

in a low utilization of the active material. I n light of these achievements, one may be tempted to con-clude that the surface-coated separator design opens up a new direction for scienti? c research. The key factor is to integrate a conductive and porous polysul? de blocking region in the commercial separator. The functional coating with physical or chemical polysul? de-trapping agents aims to intercept, absorb, and then trap the dissolved polysul? de species. The conduc-tive coating substrate also has high electrical conductivity and

porous channels to introduce electrons, Li + ions, and electro-lyte for reactivating the trapped active material. As a result,

an outstanding reutilization of the active material leads to an improved long-term cycle stability. Future efforts are directed to improving the adhesion between the lightweight coating layer and the supporting polymer separator by various thin-? lm coating techniques, such as tape-casting, vacuum ? ltration, spin-coating, or screen-printing. On the other hand, according to the reactivation capability of the coating layer, the use of ceramics or polymers may be less favorable because their insu-lating nature blocks ef? cient charge transfer and slows down the cathode reaction. H owever, the integration of conductive carbon and functional ceramic or polymer as the composite coating may offer an improvement to the cycling performance. Therefore, new coating materials possessing an optimized pore size, light weight, and additional physical/chemical functions may dominate the progress in battery performance.

3. L ithium–Sulfur Cells with a Polysul? de Catholyte

T he Li/polysul? de battery con? guration was ? rst reported by

Rauh et al. in 1979. [ 98 ] The con? guration is based on the com-bination of a carbon electrode and the dissolved polysul? de

(Li 2S n

,n ≥ 8) catholyte. The number of electrons transferred per sulfur molecule was 1.6 at that time; however, the cell experi-enced fast decay in capacity. Recently, Li/polysul? de batteries have regained popularity due to technological improvements in the polysul? de con? nement and lithium-metal anode protec-tion. For example, Zhang et al. used a LiNO 3 additive in the electrolyte to initiate the formation of a passivation layer on the anode side, and assembled Li/Li 2S 9

liquid cells with a stable reversible capacity of 500 mA h g ?1.[ 35c ]

3.1. L ithium/Polysul? de Cells with a Porous Current Collector R ecently, our group reported a highly reversible Li/dissolved polysul? de battery utilizing a self-weaving, free-standing

MWCNT paper as an electrode. [ 99 ] As shown in F igure 9a , the

electrode paper consists of abundant reaction chambers built by interwoven carbon nanotubes. The 3D current collector accommodates the polysul? de catholyte and provides path-ways for fast ion/electron transport, facilitating the reduction of sulfur and the oxidation of sul? de. An exceptional initial dis-charge capacity of as high as 1600 mA h g ?1 was obtained, cor-responding to a utilization of 96% of the theoretical capacity of Li–S batteries (Figure 9 b ). Moreover, the capacity retention was 87.5% after 50 cycles at a rate of C/10.

P R O G R E S S R E P O R T

F or a better understanding of the Li/polysul? de batteries, our group systematically investigated how the properties of the 3D current collector determine the electrochemical performance of the Li/polysul? de battery. In addition to the MWCNT electrode, our group prepared a binder-free CNF paper electrode with large interspaces between the ? bers. [ 100 ] In contrast to amorphous Li 2S in the discharged MWCNT electrode, crystalline Li 2

S was iden-ti? ed within the large interspaces of the discharged CNF elec-trode. This can be attributed to the absence of defects or func-tional groups on the surface of CNFs, providing an undisturbed environment for Li 2

S deposition. It was noted that the CNF elec-trode is free of a dense Li 2

S passivation layer during cell cycling, which was attributed to the 3D network of CNFs providing fast charge transfer. A reversible capacity of 1094 mA h g ?1at

a C/5 rate and a sulfur loading of 1.7 mg cm ?2 were obtained. Moreover, a stable reversible capacity of ca. 900 mA h g ?1was maintained with a high sulfur loading of 5.1 mg cm ?2 . The high sulfur utilization at the high sulfur loading makes this low-cost battery promising for practical applications. D espite the advantages of Li/polysul? de batteries such as the high sulfur utilization and superior reversibility, the “polysul? de shuttling” remains a persistent problem. In order to alleviate polysul? de shuttling, our group introduced graphene-based material in Li/polysul? de batteries. Graphene-based materials have been successfully applied in Li–S batteries because of their high electrical conductivity, mechanical strength, and ? ex-ibility. [ 19,21,33 ] Moreover, it has been reported that polysul? de species with a large electron density have intimate interaction with the conjugated π* states of graphene, resulting in an effec-tive trapping of polysul? des by the graphene matrix during cell

operation. [ 26a ] For example, Cao et al. used a synthetic graphene

sheet-sulfur nanocomposite (FGSS) with a Na? on coating on the surface and achieved a superior capacity retention of as

high as 84.3% over 100 cycles. [ 101 ] In addition, Evers and Nazar

used graphene to wrap sulfur particles and obtained a revers-ible capacity of 550 mA h g ?1 with a high sulfur content of ca. 78 wt% in the cathode. [ 102 ] Accordingly, our group reported the

synthesis of a TiO 2 -nanowire-embedded graphene hybrid mem-brane as a novel 3D current collector to control polysul? de shut-tling in Li/dissolved-polysul? de batteries. [ 103 ] The con? nement from both the graphene and TiO 2 nanowires effectively traps the polysul? de species and reduces the active-material loss to the electrolyte during the discharge and charge processes. Moreover, X-ray photoelectron spectroscopy (XPS) results pro-vided evidence for the up-shifting of the S 2p peaks to higher binding energies after charging the cell to 2.3 V. This suggests the catalytic effect of the hybrid membrane in promoting the kinetics of the polysul? de redox reactions. As a result, revers-ible capacity of 1053 mA h g

?1 was obtained over 200 cycles at a rate of C/5. Furthermore, a capacity of 850 mA h g

?1was achieved at a high, 2C rate. In addition, to enable the low-cost, large-scale synthesis of graphene-based materials as a cathode framework for Li–S batteries, one promising approach reported the production of multilayer graphene sheets with a worm-like structure by heating the commercially available expandable

graphite powders at 1000 °C in air. [ 104 ] Furthermore, function-alities and defects can be eliminated at such high temperatures,

bene? cial for the quality of graphene.

[ 105 ] I nstead of being the active materials in catholyte, polysul? de

species can also be used as an electrolyte additive to establish

the equilibrium between polysul?

de precipitation and dissolu-tion. In this way, the active material lost to the electrolyte was believed to be controlled. For example, Xu et al. prepared an electrolyte with Li 2S x (0.2 M) and LiNO 3 (0.5 M) and coupled it

with a sulfur–carbon composite. [ 106 ] Based on the polysul? de-containing electrolyte, the Li–S cells were capable of delivering a superior capacity of 1450 mA h (g sulfur) ?1.Furthermore, Chen et al. used polysul? des and LiNO 3 as a co-salt in ether-based electrolytes, and found that the polysul? de species in the

electrolyte helps increase sulfur utilization in the cathode. [ 107 ] Optimized performances were obtained after tailoring the elec-trolyte amount and polysul?

de concentration. Furthermore, the authors claimed that the excellent performance is due to the absence of Li 2 S

on the cathode framework in the presence of polysul? des. Elimination of insulating Li 2 S bulk particles from the cathode matrix using polysul?

de additive/catholyte was further con?

rmed by Demir-Cakan et al., [ 108 ] who investi-gated the Li–S cell performance either by using polysul? de as the active material or by placing sulfur power directly on the lithium-metal anode. Both approaches resulted in a greater per-formance than observed when con? ning sulfur into a conduc-tive matrix. The improved capacity retention was assumed to be related to the absence of severe Li 2

S passivation in the cathode

F igure 9. a) Schematic model and b) cell performance of the Li/poly-sul? de battery with a 3D current collector of MWCNTs. Reproduced with permission. [ 99 ] Copyright 2013, Wiley-VCH.

PROGRESS REPORT

and the unique SEI layer formed on the lithium metal. It is worth noting that lithium–polysul? de batteries could also be used for large-scale energy storage applications. A high energy

density of 170 W h kg ?1 and 190 W h L ?1 has been achieved

with proof-of-concept polysul? de semi-liquid batteries. [ 109 ]

3.2. L ithium/Polysul? de Cells with Sandwich Cathodes

I n order to combine the advantages of Li/polysul? de batteries

and the interlayer con? guration, our group adopted the sand-wich cathode structure in Li/polysul? de batteries. For example, we prepared novel sandwiched cathodes using two layers of free-standing carbonized sucrose-coated eggshell membranes (CSEMs) as a reservoir and with the dissolved Li 2S 6polysul?

de catholyte stabilized in between. [ 110 ] The CSEM derived from

natural chicken eggshell membrane possesses high porosity, high surface area, and long-range coalescing ? brous network. A bottom CSEM, functioning as the current collector, encapsu-lates the active material. An upper CSEM, working as the poly-sul? de trap, intercepts the migrating polysul? des. According to microstructural analysis, the CSEMs naturally possess abun-dant micropore arrays and a continuous macroporous network, which are essential factors for storing/trapping the active mate-rial and channeling the electrolyte in the cathode.

[ 22b ,23c ,36b ,110 ] The interwoven ? bers not only form the macroporous channels for improving the electrolyte immersion, but also function as continuous electron pathways for enhancing sulfur utiliza-tion.

[ 48,68,111 ] As a result, the design with sandwiched CSEM electrodes allows the dissolved polysul? des to be localized and the electrochemical reactions within the cathode region to be stabilized, resulting in a high discharge capacity, a long-term cycle stability, and a high sulfur loading. F or a more-? exible design of the sandwiched cathode, our group has developed an effective ? brous cathode framework consisting of two different functional layers: i) the upper cur-rent collector combined with a hybrid of hydrophilic polymers and microporous carbon to facilitate charge transfer and trap polysul? de, and ii) the bottom current collector with large inter-spaces between the low-cost CNFs to realize a high sulfur con-tent.

[ 112 ] Sodium alginate and poly(vinyl alcohol) were used as the hydrophilic polymers to improve the surface properties of the upper current collector and facilitate the electrolyte penetra-tion. Sodium alginate has relatively rigid polymer chains and is used to strengthen the membranes while the poly(vinyl alcohol) contributes to the membrane ? exibility. Reversible capacities

approaching 1000 mA h g ?1 were obtained with high sulfur

concentrations of >60 wt% in this free-standing sandwiched

cathode with a high sulfur loading of 5.5 mg cm

?2.Accordingly, we are likely to realize high-performance Li–S batteries pos-sessing high practical gravimetric energy densities through a simple, low-cost method.

4. L ithium-Metal Anodes

D espite the high capacity of the metallic lithium anode of ca.

3860 mA h g ?1 , practical utility of the lithium-metal anode is

hampered by its poor cycling ef? ciency and safety hazards.

The poor cycling ef? ciency and safety concerns are due to den-dritic redeposited lithium and unevenly distributed electrolyte

decomposition products on the lithium-metal surface. Further-more, corrosion of metallic lithium becomes more severe in Li–S batteries. In addition to lithium-dendrite formation and pulverization, which are common problems associated with lithium-metal anodes in organic electrolytes, dissolved inter-mediate polysul? de species in Li–S batteries can migrate to the lithium-metal anode and be reduced, leading to a passiva-tion by insoluble species, e.g., Li 2S 2/Li 2

S , on the surface of the lithium-metal anode, deteriorating the activity of the metallic

lithium. [ 113 ]

4.1. P assivation Layer on the Lithium-Metal Surface

R ecently, lithium-metal protection has attracted much atten-tion. Among various lithium-metal protection strategies for

Li–S batteries, most efforts are focused on building a stable passivation layer on the lithium-metal surface. For example, Mikhaylik et al. disclosed that an additive with N–O bonds, e.g., LiNO 3 ,

can suppress polysul? de shuttling and improve the reversible capacity of Li–S batteries.

[ 114 ] In addition, it was further explained by Aurbach et al. that LiNO 3 reacts with poly-sul? de species and lithium metal in the Li–S cell, forming effective passivation on the lithium-metal surface and blocking the direct contact between the lithium metal and the polysul? de

species. [ 115 ] In this way, polysul? de shuttling is controlled.

H owever, it is worth noting that satisfactory passivation by LiNO 3 alone is challenging to achieve and LiNO 3 is consumed gradually with cycling. Eventually, “polysul? de shuttling” could be prominent. Furthermore, LiNO 3 adversely decomposes at the cathode below 1.6 V, negatively impacting the cathode sur-face chemistry and deteriorating battery performance. Recently, Xiong et al. reported that polysul? de species play a critical part in forming a stable passivation layer on the lithium-metal sur-face.

[ 116 ] Two sublayers together form the self-limiting barrier to suppress further chemical reactions, where the top layer con-sists of oxidized products of polysul? des and the bottom layer is composed of the reduced products of polysul? des and LiNO 3.The improved lithium-surface chemistry due to the formation of a self-limiting barrier for chemical reactions in the pres-ence of polysul? de species was also reported by Demir-Cakan

et al. [ 108 ] However, the mechanism of the self-limiting effect is not fully understood. I n order to construct a reliable passivation layer on the lithium-metal surface, our group proposed using MX (where M is, for example, Cu, Fe, Ni, Co etc., and X represents the anions that control the release of M in the non-aqueous system) to initiate lithium-surface passivation layer formation. MX can interact with the lithium metal and polysul? de species, changing the physical/chemical properties of the lithium-metal surface. Our preliminary results show that copper acetate can be used as an electrolyte additive to initiate the formation of a protective passivation layer on the surface of the lithium-metal anode. [ 117 ] To demonstrate this concept, the CNF-paper/

dissolved-polysul? de cell was used. The CNF paper has macro-sized interspaces and allows direct contact between the corro-sive polysul? de species and the lithium-metal anode; therefore,