Energy storage basics

Energy density is the product of the cell voltage (V) and the specific capacity (A h g^–1 or A h L^–1) based on the total mass/volume of the positive (“cathode”) and negative (“anode”) electrodes. In a Li-ion battery, Li⁺ ions shuttle between the positive electrode intercalation host, where they are stored upon discharge (i.e., a layered oxide LiMO₂ where M is composed of Co, Ni, Mn, and Al; or a phosphate such as LiFePO₄), and the graphitic carbon negative electrode, where they are stored on charging to a maximum content of Li_0.16C. The inherent atomic structure of both the negative and positive electrode host materials is maintained throughout cycling. Cell voltages are in the range of 3.4–3.8 V (versus Li), with theoretical capacities at the positive electrode as high as 280 A h g^–1. Theoretical energy densities are around 390 Wh kg^–1 for a full cell, which contains all components, but practical values approach 210 Wh kg^–1 and much less than that for “battery packs.”

The Li-S cell is quite different. The positive electrode consists of a porous, electronically conducting host, a vital component that delivers electrons to the insulating sulfur. Upon discharge, the Li⁺ ions oxidized from the lithium negative electrode migrate through the electrolyte to join up with the electrons passed via the external circuit. The sulfur is reduced to Li₂S, which is also an excellent insulator. The reaction is surprisingly reversible. The theoretical energy density is determined by the overall cell reaction, S₈ + 16Li ↔ 8Li₂S and its potential of 2.15 V versus Li/Li⁺. Although the voltage is lower than for commercial intercalation materials, it is compensated by the specific capacity that is the highest of any solid electrode material at 1,675 mA h g^–1. This correlates to theoretical energy densities of 2,500 Wh kg^–1 or 2,800 Wh L^–1 based on weight or volume respectively.^{Reference Bruce, Freunberger, Hardwick and Tarascon3}

Two types of Li-S cells have been explored; those in which the sulfur is contained within the positive electrode host to a large degree, and those where the intermediate, highly soluble sulfur cathode redox products (shown above as S_n^2–) are dissolved in the electrolyte to form a “catholyte.” Schematic representations of these cells are compared to a Li-ion cell in Figure 1. The contained cell concept benefits from higher specific energy per mass (gravimetric density) and much higher energy density per volume, owing to a lesser amount of electrolyte. Both greatly benefit from the fact that more Li⁺ ions and electrons are stored in Li₂S than in an intercalation cathode, and that Li metal—with a nine-fold higher energy density than graphitic carbon—serves as the negative electrode.

Figure 1.

Schematic representation of contained Li-S and catholyte-type Li-S cells, compared to Li-ion. The (a) Li-ion battery typically comprises two intercalation materials, here graphite and lithium cobalt oxide, separated by a nonaqueous electrolyte based on a lithium salt (X– = PF₆^–, ClO₄^–, TFSI^–…). In contrast (b–c), Li-S batteries usually use metallic lithium as the negative electrode. In the contained Li-S architecture (b), a porous conductive host physically retains the sulfur species within the positive electrode volume, whereas the catholyte (c) functions as both the electrolyte and positive electrode. Insulation from the negative electrode is realized by a Li⁺-selective membrane (represented by the white layer).

Nonetheless, the exciting theoretical values are mitigated by several factors in practical translation. The inactive conducting host required to “wire up” the insulating sulfur mass also must allow for space for electrolyte ingress and volume expansion encountered on forming Li₂S at the positive electrode. Most of the cathodes and sulfur cathode composites described in the literature present an active material fraction below 60 wt% with porosity regularly above 30%. The actual volumetric density of such a positive electrode would not exceed the current level of Li-ion technologies.^{Reference Hagen, Dörfler, Fanz, Berger, Speck, Tübke, Althues, Hoffmann, Scherr and Kaskel13}

Other factors include larger volumes/masses of electrolyte that are necessary for typical intercalation electrodes, and the need to protect or passivate the negative electrode and utilize an excess of Li because of its inefficient cycling. Therefore, compared to Li-ion technologies, gravimetric energy density stands more to win than volumetric energy density, but both still need to be improved. Specific energies of up to 600 Wh kg^–1 have been attained for laboratory-type contained cells, with energy densities of 150 Wh L^–1 demonstrated for catholyte cells. The former is more than sufficient for automobile driving ranges up to 250 miles,^{Reference Bruce, Freunberger, Hardwick and Tarascon3} and the latter is suited to redox flow batteries for electrical grid storage.^{Reference Yang, Zheng and Cui14} However, good cycling life is a challenge, especially for cells with low electrolyte volume and high cathode loadings suited for the real-world. Scale-up to large cells also requires different cathode architecture approaches than in the world of Li-ion technologies.

The catholyte cell

The development of Li-S cells started more than half a century ago,^{Reference Herbert and Ulam15–Reference Nole and Moss17} with relatively simple configurations that have since undergone dramatic improvements in the last half decade derived from new knowledge gained in materials and nanotechnology. Although early concepts encompassed both room- and high-temperature embodiments, problems with the high-temperature cells that utilized molten salt electrolytes started a quest for Li-S batteries operating at room temperature based on organic electrolytes, which have overtaken the field. The detailed mechanism of sulfur reduction is only beginning to be fully understood, as discussed later. Overall, reversible conversion of elemental sulfur (S₈) to lithium sulfide Li₂S takes place via highly soluble intermediate species, lithium polysulfides Li₂S_n, 3 ≤ n ≤ 8, formed by the reduction of solid S₈ or the oxidation of solid Li₂S.^{Reference Yamin and Peled5} The term “dissolution-precipitation” is used to describe the cell chemistry, in common with Li-O₂ cells. The fact that both end products are insulators means that the conductive additives used as host materials, such as carbon, dilute the active mass and reduce both volumetric and gravimetric energy density. Most literature reports quote the capacity in terms of only sulfur mass, though, masking this contribution.

Figure 2 shows a typical voltage profile of a Li-S cell obtained at constant current. In the first discharge region centered at 2.35 V, S₈ is reduced to soluble Li₂S_n. These sulfides are poorly structurally characterized and co-exist in rapid equilibrium in solution. The second plateau at 2.1 V marks the reduction of these polysulfides to insoluble Li₂S.^{Reference Akridge, Mikhaylik and White18} The two regions are demarcated by a “knee” (shown in red), which represents the overpotential necessary to nucleate solid products from the supersaturated polysulfide solution. The over-potential (i.e., energy barrier, shown in blue) on the charge profile reflects the activation energy needed to oxidize the insoluble and insulating Li₂S to LiPSs (primarily S₄^2–and S₆^2–). They are converted to S₈ at the last step at 2.4V.

Figure 2.

A typical voltage profile obtained for a Li-S battery showing the supersaturation point of lithium polysulfides in the electrolyte (red) and the overpotential (blue) encountered for the oxidation of Li₂S on charge. Structural evolution of the sulfur species upon discharge is represented, from elemental sulfur S₈ rings to chain polysulfides to solid antifluorite-type Li₂S.

Owing to poor cathode architecture in early Li-S cells, uncontrolled precipitation of solid products in the large pores of the cathode resulted in low capacities due to their redox inaccessibility. The development of catholyte cells, where Li₂S_n (n = 5–8) is dissolved in an organic electrolyte, was aimed at overcoming this problem.^{Reference Rauh, Shuker, Marston and Brummer19} The high mobility of LiPSs creates another problem, however; they readily migrate to the negative electrode, whereupon they are reduced and then diffuse back to the positive electrode and are oxidized on the next cycle. This internal “shuttle” gives rise to self-discharge and Coulombic inefficiency,^{Reference Ryu, Ahn, Kim, Ahn, Lee and Cairns20} while providing beneficial overcharge capability.^{Reference Rao and Shropshire21} The migration of active material in the electrolyte is also a cause of capacity loss during the first cycles, although recent studies suggest it does not participate in long-term capacity fading.^{Reference Nelson, Misra, Yang, Jackson, Liu, Wang, Dai, Andrews, Cui and Toney22,Reference Diao, Xie, Xiong and Hong23} The main problem is the uncontrolled precipitation of Li₂S/Li₂S₂. Over cycling, insoluble agglomerates form on the surface of the positive electrode that become electrochemically inaccessible.^{Reference Barchasz, Lepêtre, Alloin and Patoux24,Reference Cheon, Ko, Cho, Kim, Chin and Kim25} On the other side, the polysulfides that diffuse through the electrolyte to the Li metal surface eventually create a blocking surface of insoluble Li₂S on the negative electrode. Increased impedance and limited rechargeability result.

These problems plagued the catholyte cell for over a decade. But recently, the concept has been revisited, partly driven by difficulties encountered with the “contained” cathode design, but also because of developments with negative electrode protection (discussed next) and in positive electrode architectures.^26,27 Applications in redox flow batteries (RFBs) have been considered, where the catholyte is designed to cycle only in the range between S₈ and Li₂S₄, avoiding problems due to the formation and volume expansion of solid Li₂S₂/Li₂S.^{Reference Yang, Zheng and Cui14} This strategy results in good cycle life and compatibility with RFB design. For a 5 M Li₂S₈ catholyte, energy densities of 97 Wh kg^–1 and 108 Wh L^–1 have been reported without the need for expensive ion-selective membranes (Figure 3).

Figure 3.

(a) Schematic illustrating the structure of the lithium/polysulfide (Li/PS) battery. Polysulfides flow through the system during operation and are stored in the tank in downtime. Magnified scheme on the right: the solid electrolyte interphase passivating Li due to the presence of LiNO₃ in the catholyte. Consequently, the shuttle effect is significantly suppressed. (b) An experimental demonstration of a membrane-free Li/PS battery with Li and the carbon fiber electrode soaked in a Li₂S₈ solution. (c) Volumetric energy density of the Li/PS system versus concentration of sulfur in the catholyte (red). The capacity of Li is set to 2000 mA h g^–1 and 1000 mA h cm^–3 to take the extra Li needed for practical operation into consideration. The specific capacity between sulfur and Li₂S₄ is 418 mA h g^–1. The black line shows the energy density of a vanadium redox flow battery (VRB) based on the vanadium catholyte and analyte only (30 W h L^–1). The concentrations of both catholyte and analyte are supposed to be 1.7 M, which is the commonly used stability limit for vanadium flow batteries. From this assessment, a 10-fold increase in volumetric energy density from the VRB to flow Li/PS is conceivable. Reproduced with permission from Reference 14. © 2013 Royal Society of Chemistry.

The negative electrode

Passivation of Li metal electrodes is critical, not only to inhibit the precipitation of insoluble sulfides, but also to prevent the growth of dendrites on plating/replating of lithium. The latter creates a safety concern owing to the potential for short circuits and consumes Li via reaction of the electrolyte on each cycle by forming a continuously regenerated electrolyte interface layer. This issue has been a struggle for more than two decades,^{Reference Abraham, Brummer and Gabano28} although recent advances show that additives can achieve smooth, reliable dendrite-free Li deposition.^{Reference Ding, Xu, Graff, Zhang, Sushko, Chen, Shao, Engelhard, Nie, Xiao, Li, Sushko, Liu and Zhang29} This does not mitigate the LiPS shuttle and self-discharge though.

Two approaches have been explored to passivate the Li surface. One is the use of insulating protective ion-conductive membranes. Visco and coworkers at Polyplus^{Reference Visco and Chu30} created protective membranes from lithium superionic conductors such as Li_(1+x+y)Al_xTi_(2–x)Si_yP_(3–y)O₁₂, marketed by O’Hara Glass. Owing to the titanium in the composition, they are readily reduced by Li, meaning that an interlayer between the ceramic membrane and the lithium negative electrode is necessary. This typically constitutes a thin film of a gel electrolyte.^{Reference Zhang, Imanishi, Shimonishi, Hirano, Xie, Takeda, Yamamoto and Sammes31} A compliant seal around the assembly is also needed to accommodate the large swings in lithium volume experienced on cycling. New research in solid-state Li-ion conductors has demonstrated that lithium metal garnets are viable alternatives to phosphates. Based on the lithium-lanthanum-zirconium garnet, Li₇La₃Zr₂O₁₂,^{Reference Murugan, Thangadurai and Weppner32} these have the advantage of good stability to lithium metal and exhibit high conductivities when stabilized in the cubic phase.^{Reference Knauth33–Reference Shinawi and Janek35}

SION Power has developed another avenue for negative passivation by utilizing LiNO₃ as an electrolyte additive.^{Reference Mikhaylik36} Work by Aurbach et al. has elucidated the mechanism behind this process, showing that it creates a protective film of Li_xNO_y and/or Li_xSO_y on the negative electrode.^{Reference Aurbach, Pollak, Elazari, Salitra, Kelley and Affinito37} In fact, most academic researchers exploring “contained” Li-S cell chemistry (see later) also utilize LiNO₃ today for this purpose, thus routinely attaining Coulombic efficiencies that approach 100%. SION Power has demonstrated the production of Li-S cells based on catholyte redox chemistry with a stabilized Li metal anode that reportedly attains a specific energy of 350 Wh/kg, but with much lower (volumetric) energy density. These have been used to store energy for solar-cell powered unmanned aerial vehicles such as the Zephyr during night-time flight.

The problem of the lithium metal electrode has also led to efforts in recent years to explore non-Li negatives. A source of Li is needed. Li₂S as a cathode has been married with a tin anode,^{Reference Hassoun and Scrosati38} with a silicon anode,^{Reference Yang, McDowell, Jackson, Cha, Hong and Cui39} and a prelithiated silicon anode in combination with a sulfur cathode.^{Reference Elazari, Salitra, Gershinsky, Garsuch, Panchenko and Aurbach40} However, cycle life was limited. Most recently, prelithiated silicon-carbon and hard carbon anodes in conjunction with sulfur-infused porous hollow sphere cathodes have been shown to demonstrate long cycle life of over 1000 cycles at a high capacity of 1470 mAh g^–1.^{Reference Brückner, Thieme, Böttger-Hiller, Bauer, Grossmann, Strubel, Althues, Spange and Kaskel41}

The contained Li-S positive electrode

The alternative to the catholyte cell is to immobilize Li₂S_n in the cathode, simultaneously increasing the volumetric energy density. The porous substrate within which the Li₂S forms must be a good conductor and of low cost, hence the dominance of carbon. Because Li₂S has a lower density than elemental sulfur (1.66 g • cm^–3 versus 2.06 g • cm^–3, respectively), discharge causes an 80% increase in the positive electrode volume.^{Reference He, Ren, Wang, Pu, Jiang and Wan42} This has to be taken into account when designing the sulfur host, because it limits the maximum sulfur content and results in severe mechanical stress and decrepitation on cycling.^{Reference Elazari, Salitra, Talyosed, Grinblat, Scordilis-Kelley, Xiao, Affinito and Aurbach43} Inventive architecture is also critical to ensure that the solid end members remain in effective contact with the host framework.

A number of porous conductive carbons with 3D-accessible nanostructures have been developed in the last five years to “contain” intermediate Li₂S_n, starting with the mesoporous carbon CMK-3.^{Reference Ji, Lee and Nazar44} It is worth noting that the degree of actual containment is often not specified, and the cells may operate partly in catholyte mode. In principle, the conductive carbon framework constrains the sulfur within the framework and provides the requisite electrical contact for electron transfer to both S₈ and Li₂S at the nanoscale. Because the sulfur fraction must be adjusted to retain pathways for electrolyte/Li⁺ ingress and to accommodate the active mass volume expansion during cycling, large pore volumes are required if suitably high sulfur loadings are to be attained. This can trigger dissolution of Li₂S_n without clever approaches to overcome this. Strategies to retain polysulfides at the positive electrode have utilized hydrophilic polymers to provide a chemical gradient or physical barriers to retard diffusion of the polysulfides into the electrolyte: bimodal porous carbon architectures, oxide additives and shells, and carbon interlayers. These have been recently reviewed in detail,^{Reference Evers and Nazar8–Reference Wang, Zeng, Zhou, Yin, Li, Cheng, Gentle and Lu12} and a few selected advances are highlighted here.

Graphene/graphene oxide sheets are promising sulfur hosts because of their high conductivity and the wide range of compositions that can be produced through functional chemistry, utilizing the different architectures that can be constructed around sulfur particles or melt-diffused sulfur to form sandwiched/interleaved composites. Particularly important is that by controlling the ratio of reduced graphene to that of the native graphene oxide, bifunctional properties can be built in. These profit from the excellent electronic conductivity of the graphene combined with the Li₂S_n-absorption properties of the graphene oxide, while maintaining a scalable, low cost material.^{Reference Ji, Rao, Zheng, Zhang, Li, Duan, Guo, Cairns and Zhang45–Reference Chen, Zhao, Lu, Wu, Li, Chen, Tan, Ye and Amine47} The carbon host can also be coupled with a polymer coating to wrap particles and prevent any loss of active material in the electrolyte.^{Reference Wang, Yang, Liang, Robinson, Li, Jackson, Cui and Dai48} These materials also provide the opportunity to functionalize their surface with hydrophilic^{Reference Ji, Lee and Nazar44} or amphiphilic polymers.^{Reference Zheng, Zhang, Cha, Yang, Li, Seh and Cui49} This creates an even more favorable hydrophilic interaction with the soluble polysulfides and enhances contact with Li₂S, which reduces detachment of Li₂S from hydrophobic carbon surfaces. A multifaceted approach using a unique combination of surfactant-modified graphene oxide (Figure 4) with an elastomeric binder and an ionic liquid modified electrolyte has recently proven to be remarkably effective, achieving capacity fade rates as low as 0.039% per cycle over 1,500 cycles.^{Reference Song, Zhang and Cairns50}

Figure 4.

Schematic of the S−GO (graphene oxide) nanocomposite structure used in the multifaceted approach to Li-S cathode design. The presence of the cationic surfactant (CTAB) on the S−GO surface was shown to be critical for achieving improved cycling performance by minimizing the loss of sulfur; 1400 mAh/g_sulfur are delivered at a rate of C/5 (i.e., full discharge in five hours), representing a two-fold increase compared to state-of-the-art Li-ion technology (blue dashed line). Reproduced with permission from Reference 50. © 2013 American Chemical Society.

Porous hollow carbon spheres (PHCSs)^{Reference Jayaprakash, Shen, Moganty, Corona and Archer51,Reference Kim, Lee, Jung, Sun, Hassoun and Scrosati52} and hollow carbon nanofibers^{Reference Guo, Xu and Wang53,Reference Zheng, Yang, Chan, Hong and Cui54} have been explored recently. PHCSs, composed of a thin porous carbon shell with a hollow interior that can be partly filled with sulfur, offer particularly high sulfur utilization and good Li₂S_n retention. Many variations have been explored following the original report by Archer et al.,^{Reference Jayaprakash, Shen, Moganty, Corona and Archer51} including synthetic approaches to thicken the carbon shell and tailor the porosity.^{Reference Cuisinier, Cabelguen, Evers, He, Kolbeck, Garsuch, Bolin, Balasubramanian and Nazar55} Uniformly sized 220 nm hollow porous carbon nanospheres exhibit very steady capacity behavior in typical Li-S cell configurations with a capacity retention of 99.5% over 100 cycles^{Reference He, Evers, Liang, Cuisinier, Garsuch and Nazar56} and a much longer cycle life when coupled with a hard carbon anode.^{Reference Brückner, Thieme, Böttger-Hiller, Bauer, Grossmann, Strubel, Althues, Spange and Kaskel41}

The electrolyte

An optimized electrolyte solvent or solvent blend for the Li-S system is essential to improve Li-S battery performance, but not as much effort has been expended in this regard as in cathode design. Li-ion electrolyte solvents such as alkyl carbonates react with linear chain radical polysulfides formed on cycling and cannot be used. Typical solvents are linear and cyclic ethers, such as dimethyl ether (DME) and 1,3–dioxolane,^{Reference Peled, Gorenshtein, Segal and Sternberg57} with 1 M lithium bis(trifluoromethanesulfonyl)imide (LiTFSI) as the Li⁺ ion carrier.^{Reference Kim and Jeong58} Although this solvent has very high LiPS solubility and good conductivity, its low boiling point makes it impractical for most commercial applications. Less volatile, more viscous electrolytes, such as ethyl methyl sulfone^{Reference Ji, Lee and Nazar44} and oligomeric polyethers (glymes),^{Reference Yuan, Feng, Ai, Cao, Chen and Yang59} have been used to improve performance.

Room-temperature ionic liquids (RTILs), consisting of coordinated cations and anions, are of particular interest as electrolytes owing to their high stability, nonflammability, and nonvolatility. Recently, glyme-Li salt equimolar complexes were shown to exhibit RTIL-like behavior (e.g., non-volatility and high ionic conductivity).^{Reference Tamura, Hachida, Yoshida, Tachikawa, Dokko and Watanabe60} The oxygen atoms in a glyme coordinate with Li⁺ to form a complex cation, in which the Li⁺ cation and the ligand (glyme) serve as the Lewis acid and Lewis base, respectively. These electrolytes showed promising results for the Li-S system, in terms of Coulombic efficiency.^{Reference Park, Yamauchi, Takashima, Tachikawa, Ueno, Dokko and Watanabe61,Reference Suo, Hu, Li, Armand and Chen62} This has been explained by the suppression of the redox shuttle by minimizing Li₂S_n solubility and mobility in the electrolyte, similarly to other RTILs. Amine et al. have also used partially fluorinated ethers as a co-solvent with DME to greatly improve the Coulombic efficiency as well as cycling performance.^{Reference Weng, Pol and Amine63}

During the last few years, it was shown that Li-S batteries can be cycled very well in an all–solid-state configuration.^{Reference Sakuda, Hayashi and Tatsumisago64–Reference Lin, Liu, Dudney and Liang66} Solid-state electrolytes have many advantages, as they diminish safety concerns and completely inhibit self-discharge. Lithium sulfide glasses such as Li₂S-P₂S₅,^{Reference Nagao, Hayashi and Tatsumisago67} Li₂S-SiS₂,^{Reference Machida, Kobayashi, Nishikawa and Shigematsu68} or Li₂S-GeS₂-P₂S₅^{Reference Kobayashi, Imade, Shishihara, Homma, Nagao, Watanabe, Yokoi, Yamada, Kanno and Tatsumi69} form the foundation of the burgeoning field of all solid-state Li-S batteries. These glasses exhibit high ionic conductivity, but hypersensitivity to air and moisture. Arsenic-substituted Li₄SnS₄ recently demonstrated improved air stability,^{Reference Sahu, Lin, Li, Liu, Dudney and Liang70} although compatibility with metallic lithium is still problematic. To alleviate this issue, the same study reported on LiBH₄-LiI as a chemically stable interface between sulfide solid electrolytes and Li metal.^{Reference Sahu, Lin, Li, Liu, Dudney and Liang70} On the other hand, establishing an intimate interface with the solid active cathode material can also be a challenge. Solid electrolytes and all–solid-state batteries are emerging research areas to follow and will undoubtedly impact the development of high-energy batteries.

Operando studies

Achieving full, reversible reduction from elemental sulfur down to Li₂S, beyond the typically achieved specific capacity of 1200 mAh g^–1, is vital to improving the energy density of Li-S batteries. However, the factors that govern capacity and cycling stability are difficult to access owing to the amorphous nature of the intermediate species and their rapid equilibration.^{Reference Barchasz, Molton, Duboc, Leprêtre, Patoux and Alloin72,Reference Patel, Demir-Cakan, Morcrette, Tarascon, Gaberscek and Dominko73} Fundamental studies are required to understand sulfur speciation using in situ probes of redox processes within a working cell. Transmission x-ray microscopy imaging carried out operando, for example, suggests that loss of soluble Li₂S_n outside of the electrode may not be as problematic as previously thought.^{Reference Nelson, Misra, Yang, Jackson, Liu, Wang, Dai, Andrews, Cui and Toney22}Operando X-ray diffraction studies probe solid crystalline phases, showing that Li₂S precipitation starts at the beginning of the second discharge plateau (i.e., at the super-saturation point indicated in Figure 2) and is oxidized to a different phase of elemental sulfur on charge.^{Reference Walus, Barchasz, Colin, Martin, Elkaim, Lepretre and Alloin74,Reference Villevieille and Novak75}

In situ electrochemical impedance spectroscopy studies, in contrast, suggest that insoluble Li₂S forms at the very end of the low-voltage plateau,^{Reference Yuan, Qiu, Chen and Zhu76} corresponding to the voltage drop in the profile presented in Figure 2. Barchasz et al. investigated the complex set of equilibria involved in a catholyte cell with operando UV spectroscopy.^{Reference Barchasz, Molton, Duboc, Leprêtre, Patoux and Alloin72} Their studies identified S₆^2– and S₄^2– as the most probable stable polysulfides in solution and suggested that S₈^2– undergoes fast disproportionation via S₈^2– → S₆^2– + ¼S₈. These findings are in accord with X-ray absorption spectroscopy (XAS) measurements carried out at the sulfur K-edge. The first in situ sulfur XAS study revealed the reactivity of polysulfides with alkylcarbonate solvents,^{Reference Gao, Lowe, Kiya and Abruña77} showing its utility as a local-probe technique.

More recently, direct monitoring of sulfur speciation over the entire redox range in real time was enabled by operando XAS.^{Reference Cuisinier, Cabelguen, Evers, He, Kolbeck, Garsuch, Bolin, Balasubramanian and Nazar55} Spectra acquired at intermediate points every 50 mAh g^–1 in capacity were analyzed to achieve an understanding of sulfur speciation in the cell “on the fly,” as summarized in Figure 5. One finding is that high discharge capacities (>1200 mAh/g_S) are limited by conversion of the sulfur mass at high S content. Owing to rapid disproportionation of S₈^2–, sulfur is formed as it is consumed on reduction. Once Li₂S nucleation is initiated at the supersaturation point, reduction proceeds rapidly through the soluble Li₂S_n, which is kinetically favored versus reaction of the residual sulfur. Such information helps explain the “nano-sized” effect that maximizes sulfur reactivity (see above), and is critical for promoting improvements in Li-S battery cathode architecture and electrolyte design.

Figure 5.

Schematic summarizing the overall processes proposed to occur in the Li-S cell on discharge and charge based on operando x-ray absorption near-edge spectroscopy. On charge, surface oxidation of Li₂S—a kinetically hindered process—leads to species such as S₄^2– that are consumed gradually during the sloping plateau to form Li₂S₆, with final oxidation to elemental sulfur occurring at 2.4 V. On discharge, initial reduction of elemental sulfur results in unstable S₈^2–, which disproportionates to S₆^2– and sulfur. On further electron transfer, the higher reactivity of soluble, metastable Li₂S₆ leads to a reservoir of sulfur and a cascade of intermediate polysulfides, where Li₂S₄ was identified, although the formation of S₃^2– and S₂^2– may also occur as transient species. Their rapid reduction to S^2– leads to supersaturation in solution, followed by delayed and sudden onset of precipitation of Li₂S at the end of the 2.1 V plateau. Reproduced with permission from Reference 55. © 2013 American Chemical Society.

Summary and future prospects

Practical, high energy density storage using Li-S batteries that capitalize on the low-cost features of elemental sulfur is closer to realization than ever, but fundamental challenges remain. Achieving full, reversible reduction to Li₂S is vital to increasing energy density. To accomplish this, insights into the mechanisms of sulfur redox chemistry and speciation in the cell are further needed to determine how sulfur fraction and sulfide precipitation impact capacity. The last year has seen promising reports of capacity retention of +1000 cycles. While this is predicted to continue and improve through materials discovery for both cathodes and electrolytes, a more realistic, scalable approach to materials design that aims at cells with low electrolyte volume/high electrode loading is important. Li-S presents a wealth of opportunities in this regard. Future research should also center on the development of more effective negative electrodes through solid-state ion conducting protective membranes and alternate anode materials. By surmounting these obstacles, a leap forward in energy storage could indeed be achieved.