1 Introduction

With rapid industrial and economic growth, global fossil fuel consumption has increased significantly, resulting in severe resource depletion and environmental concerns [1,2,3]. Therefore, to optimize energy generation systems and achieve sustainable industrial and economic growth, it is essential to develop and utilize advanced efficient energy storage and conversion systems. As efficient electrochemical energy storage devices, batteries are based on the conversion between chemical and electrical energy and can play important roles in the optimization of energy utilization and the improvement of environmental suitability. And aside from widespread applications in portable electronics, batteries are also expected to be applied in smart grid electricity storage and release applications because their electrochemical and energy characteristics are suitable for the storage and release of excess energy as required [1, 4]. However, with growing application requirements, there is an urgent need for advanced battery systems with higher energy densities, longer cycling lives and improved cost-effectiveness [5, 6]. Therefore, to meet these requirements, the focus must turn to the exploration of advanced battery systems with higher energy densities and longer life spans beyond current conventional battery packs. This is because lithium ion battery (LIB) technologies are approaching theoretical energy densities and cannot satisfy the demands of hybrid electric vehicles (HEVs), electric vehicles (EVs) and stationary electricity storage devices [7,8,9,10]. Therefore, tremendous efforts have been undertaken to find alternative advanced battery systems.

The lithium sulfur (Li–S) battery system is one such advanced battery system that is regarded as a promising next-generation energy storage system because of its higher theoretical energy density in comparison with LIBs [11,12,13,14]. The investigation of Li–S batteries can be traced back to the 1960s with breakthroughs made in 2009 by Nazar et al. [15,16,17] who synthesized mesoporous carbon sulfur composites as cathode materials in Li–S batteries and reported high specific capacities and stable cycling performances. In general, Li–S batteries are composed of lithium metal anodes coupled with sulfur cathodes and the energy conversion mechanism of Li–S batteries is based on the electrochemical reactions between Li and S8. This is different from the intercalation and deintercalation mechanisms of lithium ions in LIBs whose theoretical energy density is limited to ~420 Wh kg−1 or ~ 1400 Wh L−1 [18,19,20].

In terms of cathode materials for lithium batteries, sulfur cathodes possess high theoretical specific capacities of up to 1672 mAh g−1 and are nearly an order of magnitude higher than that of traditional cathode materials in LIBs such as LiCoO2, LiMn2O4 and LiFePO4 [14, 21,22,23,24]. And because of this, Li–S battery cells using lithium metal anodes and sulfur cathodes with an average cell voltage of 2.15 V versus Li+/Li can provide theoretical specific energy densities of ~2500 Wh kg−1 and volumetric energy densities of~2800 Wh L−1 [17, 25,26,27]. Furthermore, sulfur is an abundant resource that is inexpensive, making Li–S batteries cost-effective and a competitive candidate in the field of energy storage devices.

In terms of anode materials for lithium batteries, metallic lithium possesses both a high theoretical specific capacity reaching 3860 mAh g−1 and a low electrochemical potential of − 3.04 V versus standard hydrogen electrode [22, 28,29,30]. These advantages make lithium metal one of the most attractive candidates for the anode materials of next-generation high-energy density lithium batteries such as Li–S batteries and lithium air batteries [28, 31,32,33,34]. However, although the concept of lithium second batteries has been proposed since the 1970s, the instability and safety issues of metallic lithium anodes have prohibited their widespread application, resulting in the greater success of graphite anodes which were introduced in the 1990s [35,36,37,38]. Despite this, with the rapid development of sulfur cathode, recent research has increased focuses on the lithium metal anodes, main factors hindering the practical application of Li–S batteries.

Carbon materials and their composites have also attracted increasing attention from researchers because of their good conductivity, high surface specific area and mechanical strength, allowing their applications as potential matrices for both sulfur cathodes and metallic lithium anodes [39,40,41,42,43,44,45,46]. Based on this, numerous reviews summarizing progresses and introducing diversified electrode materials in Li–S batteries have been conducted recently [31, 47,48,49]; however, in this review, more attention will be paid to the multifunctionality of carbon-based frameworks with their advancements in the field of Li–S batteries being highlighted, including sulfur cathode materials and protective strategies for lithium metal anodes. Overall in this review, the electrochemical reactions of Li–S batteries will first be presented and the challenges existing in sulfur cathodes and lithium metal anodes will be discussed. The advantages of carbon-based frameworks as electrode materials in Li–S batteries will subsequently be highlighted, and effective methods to utilize carbon-based materials for both sulfur cathodes and lithium metal anodes will be provided. Finally, perspectives for the future research directions of high-performance and practical Li–S batteries will be presented.

2 Electrochemical Reactions and Challenges of Li–S Batteries

The typical configuration of a Li–S battery consists of a lithium metal anode, a sulfur composite cathode, an electrolyte and a porous separator isolating the electrodes [10, 12, 50]. And as the active material in the cathodes of Li–S batteries, sulfur exists as multiple types of allotropes with octa-sulfur (S8) being the most stable [6, 51]. Here, the most common sulfur composite cathodes are fabricated through mixing S8 and conductive carbon materials together by using a binder [52]. Unlike traditional LIBs, carbonate-based electrolytes are unsuitable for Li–S batteries because of the chemical instability of polysulfides in these types of electrolytes. Therefore, carbonate-based electrolytes are replaced by inert ether-based electrolytes that are composed of dimethoxyethane and dioxolane by using lithium bis-trifluoromethane sulfonimide as an electrolyte conductive agent [13, 14, 53, 54]. However, although numerous Li–S battery studies employ ether-based electrolytes, carbonate-based electrolytes do possess comparative advantages including the limited solubility of polysulfide intermediates as compared with ether-based electrolytes, reducing shuttle effects [55, 56]. As a result, researchers have reported that the utilization of microporous carbon or sulfur selenides can avoid nucleophilic reactions between polysulfides and carbonate-based electrolytes [57,58,59].

2.1 Electrochemical Reaction

The complete reaction of Li–S batteries is as follows:

$$ {\text{S}}_{8} + 16{\text{Li}} \leftrightarrow 8{\text{Li}}_{2} {\text{S}} $$

Despite being seemingly simple, the actual electrochemical process of Li–S batteries is complex with multistep transformations between S8 and Li2S, generating soluble and insoluble polysulfide intermediates. Figure 1 displays the charge/discharge voltage profile of a conventional Li–S battery using an ether-based electrolyte [51], with two discharge voltage plateaus being presented in the voltage profile. Here, the ring opening reaction of the cyclic S8 molecule occurs in the initial steps, forming S82−which is gradually reduced to S62− and S42− with an average voltage of 2.3 V versus Li+/Li [6]. During these steps, long-chain polysulfide intermediates are generated such as Li2S8, Li2S6 and Li2S4 that exhibit high solubility in ether electrolytes [15]. These high-order polysulfides can be further reduced to insoluble short-chain suifides such as Li2S2 and Li2S, corresponding to the long discharge voltage plateau at 2.1 V versus Li+/Li [25]. However, because of the insolubility of Li2S2 and Li2S, reaction kinetics are sluggish. In addition, the solid–liquid–solid transition process makes the mechanisms of the Li–S battery electrochemical reaction complex and difficult to understand and detailed mechanisms need to be further studied to gain deeper understanding.

Fig. 1
figure 1

Schematic illustration of the charge/discharge voltage profile of a conventional Li–S battery. Reproduced with permission from Ref. [51]

2.2 Challenges

Although Li–S batteries possess high theoretical capacities, high energy densities and better cost efficiencies, several key challenges impede their commercialization.

The challenges of sulfur cathodes are as follows:

  1. (1)

    Insulating nature of sulfur and lithium sulfide

    The electrical conductivity of sulfur is 5 × 10−30 S cm−1, and the electronic resistivity of lithium sulfide is larger than 1014 Ω cm [17, 60]. In addition, both sulfur and lithium sulfide possess low lithium ion diffusivity [48]. Because of these properties, sulfur and lithium sulfide are electronically and ionically insulating, leading to low electrochemical kinetics during charge and discharge cycling. Furthermore, this insulating nature increases the internal resistance of batteries, limits the adequate utilization of active materials and lowers the rate performance [17, 61]. Moreover, the deposition of lithium sulfide on cathode surfaces can hinder the contact between electrolytes and active materials, leading to decreased specific capacities. And because of all of this, it is difficult for Li–S batteries to achieve theoretical capacities in practical situations. Therefore, to enhance the conductivity of sulfur cathodes and improve the performance of Li–S batteries, it is necessary to construct conductive frameworks and disperse sulfur uniformly to achieve intimate contact within the cathode structure.

  2. (2)

    Volume expansion of sulfur

    The density of sulfur is 2.03 g cm−3, and the density of lithium sulfide is lower, 1.66 g cm−3 [25, 62, 63]. After full lithiation, the volumetric expansion of sulfur can reach about 80% [64]. And through repeated charge and discharge cycles, the large volume changes of the electrodes can result in the pulverization of cathodes and the exfoliation of active materials from the conductive framework, leading to the permanent losses of capacity and cycling life. To avoid this, cathodes should have appropriate spaces to accommodate electrode volume changes in the electrochemical reaction process. However, although this additional void space can improve the performance of sulfur cathodes, it can also decrease the volumetric energy density of batteries. Therefore, it is essential to balance the volumetric energy density and the porosity of the cathode.

  3. (3)

    Shuttle effect of polysulfides

    Long-chain polysulfide intermediates can be generated during cycling and are soluble in ether-based electrolytes [49, 65]. These dissolved polysulfide intermediates subsequently travel through the separator and diffuse from the sulfur cathode to the metallic lithium anode because of the concentration gradient between the electrodes [6]. And because of the reactivity between these polysulfides intermediates and lithium metal, the polysulfide intermediates can be reduced by lithium, generate short-chain polysulfides and even insoluble Li2S2 and Li2S, the precipitation of which on the surface of anodes can result in capacity fading and lithium metal destruction. In addition, short-chain polysulfides can also migrate back to the cathode and oxidize to long-chain polysulfides, leading to low Coulombic efficiencies [6, 15]. This overall phenomenon is known as the “shuttle effect,” which is a serious problem that lowers capacity, decreases Coulombic efficiency and increases self-discharge. However, the dissolution of polysulfides is a key factor in the adequate utilization of sulfur; therefore, the optimal designs of Li–S batteries must balance the high utilization of sulfur and the restraining of the shuttle effect.

The challenges of metallic lithium anodes are as follows:

  1. (1)

    Growth of lithium dendrites

    The uncontrollable growth of lithium dendrites on the surface of lithium metal is caused by the distribution in the homogeneity of surface current density and the uneven distribution and deposition of lithium [37, 66]. These lithium dendrite formations can cause serious safety problems, including the risk of fire and explosion, because lithium dendrites can penetrate separators and come into contact with cathode materials directly, leading to the short circuiting of batteries. This short circuiting in lithium metal batteries in turn severely impedes commercialization [32]. However, the breaking of these dendrites can result in the formation of “dead Li,” which can decrease battery capacities and Coulombic efficiencies [33]. Moreover, accompanying the growth of lithium dendrites, metallic lithium anodes can form porous lithium deposits, leading to the large polarizations and the volume changes of anodes [32, 67, 68].

  2. (2)

    Cracking of the solid electrolyte interface (SEI) layer

    The repeated plating/stripping process on the surface of lithium anodes during the cycling induces large volume changes in the electrode as well as the iterative destruction and restoration of the SEI layer [69], causing cracking. After the cracking of the SEI layer, fresh lithium metal surfaces are exposed to the electrolyte and can react to form new SEI layers due to the high reactivity of lithium metal, causing the constant consumption of electrolytes and metallic lithium anode materials. Because of this, unstable SEI layers on anodes decrease lithium metal battery Coulombic efficiencies and cycling lives [39, 70].

Currently, there are several major challenges impeding the practical application of Li–S batteries; however, numerous strategies have been utilized in the construction of sulfur cathodes and the protection of metallic lithium anodes to fabricate battery systems with high performances and cycling stabilities. In the subsequent sections of this review, the multifunctionality of carbon-based frameworks in Li–S batteries including sulfur cathodes and lithium metal anodes will be discussed and highlighted.

3 Carbon-Based Materials for Sulfur Cathodes

Carbon-based materials possess desirable multifunctionality and have been widely applied in applications such as electrochemical catalysis [71, 72], lithium batteries [73, 74], solar cells [75, 76] and other energy conversion systems [77,78,79]. In addition, carbon-based materials have also shown great potential for application in advanced energy storage and conversion systems [80]. Currently, carbon-based materials have been utilized in the majority of battery types and especially for Li–S batteries because of their high conductivity, high specific surface area and controllable functionality. These materials are suitable for fabricating Li–S battery sulfur cathodes with good performance and long life spans [47, 48]. In addition, the porous structure of carbon-based materials can effectively encapsulate sulfur, accommodate volume changes and trap polysulfide intermediates during cycling [51]. In this section, the advancements of carbon-based materials for sulfur cathodes will be discussed in detail, including the porosity and conductivity of carbon-based frameworks, polarity with functional groups and the affinity in carbon-based composites.

3.1 Porosity and Conductivity of Carbon-Based Frameworks

3.1.1 Porous Carbon Materials

In an example of porous carbon materials, Nazar et al. [16] synthesized a highly ordered mesoporous CMK-3 as the host of sulfur, taking advantage of the desirable properties of mesoporous carbon including the high surface area, homogeneous porosity structure and high conductivity. And as a result, the capacity of the resulting sulfur/mesoporous carbon composite reached 1005 mAh g−1 at a current density of 168 mA g−1 at room temperature. It is worth noting in this study that sulfur infiltrated into the channels of the carbon thoroughly at 155 °C at which temperature the viscosity of sulfur is lowest. This study significantly improved the performance of Li–S batteries and broadened the horizons of design strategies for sulfur/carbon composite cathodes.

To further enhance the sulfur content encapsulated in carbon hosts and minimize the shuttle effects of polysulfides, Jayaprakash et al. [81] synthesized mesoporous hollow carbon capsules using a hard template approach and reported that after the loading of sulfur into the carbon cavity and shell, the resulting hollow carbon@sulfur composite cathode maintained a specific capacity of 974 mAh g−1 after 100 cycles at 850 mA g−1. Similarly, Schuster et al. [82] reported the spherical ordered mesoporous carbon with a mean diameter of 300 nm that possessed a high specific surface of 2445 m2 g−1 and mesopores of 6.0 and 3.1 nm. Here, the mesoporous structure facilitated the homogeneous dispersion of sulfur and charge transfer during cycling and the resulting sulfur/carbon cathode produced an initial capacity of 1070 mAh g−1 and a high capacity retention of 700 mAh g−1 after 100 cycles at 1675 mA g−1.

Although these studies provide structures that can significantly improve Li–S battery performances, they only focused on single pore structures that cannot completely confine high-order polysulfides. Therefore, the use of hierarchical porous carbon materials that can combine the desirable properties of microporous, mesoporous and macroporous structures is a more effective method to deliver Li–S batteries with robust cycling performances [83]. An example of a hierarchical porous carbon (HPC) with mesopores and macropores surrounded by micropores is shown in Fig. 2 [84]. In comparison with mesopores and macropores, micropores possess strong interactions with lithium polysulfides, whereas the other types of porous structures can provide reservation spaces for elemental sulfur [85]. And as for the HPC material shown in Fig. 2, the distribution of pore sizes proves that this HPC material possessed a hierarchical porous structure with ~ 39% of the total pore volume being micropores. And at a current density of 4020 mA g−1, the capacity of the HPC-S achieved 539 mAh g−1 after 500 cycles with a capacity retention of 77%, exhibiting favorable cycling stability.

Fig. 2
figure 2

a Schematic illustration of the electrode structure and electrochemical process of the HPC particle, b SEM image and c TEM image of HPC spheres, d HRTEM image of the outer shell of an HPC sphere. Reproduced with permission from Ref. [84]

As for the preparation of porous carbon materials, MgO is often used as a template. For example, Zheng et al. [86] used MgO as a template to prepare porous carbon rods constructed with vertically oriented porous graphene-like nanosheets (HPCR) using chemical vapor deposition (Fig. 3). Here, the surface area of the resulting HPCR reached 2226 m2 g−1 with a high total pore volume of 4.9 cm3 g−1 along with a hierarchical porous structure, suggesting that HPCR can be used as a suitable matrix to encapsulate sulfur with high loading contents. And after the infiltration of sulfur into the carbon host by using a melting diffusion method, the researchers in this study reported that the sulfur loading of the HPCR-S composite reached 78.9 wt% and it was able to produce a high initial discharge capacity of 1430 mAh g−1 at 335 mA g−1 and a discharge capacity of 970 mAh g−1 at 1675 mA g−1 with a retention capacity about 700 mAh g−1 over 300 cycles, displaying good rate capability and cycling stability. In another example, Lyu et al. [87] also used MgO as a template to synthesize hierarchical porous structured carbon nanocages (hCNC) to achieve high sulfur loadings in Li–S batteries. In this study, the resulting hCNC possessed a high surface area of 1276 m2 g−1 and a total pore volume of 4.178 cm3 g−1, providing large spaces to accommodate sulfur. And because of this, the sulfur loading of the S@hCNC composite reached 79.8 wt%, providing a capacity of 558 mAh g−1 at 1000 mA g−1 after 300 cycles.

Fig. 3
figure 3

a The synthetic process for HPCR, b SEM image of HPCR, c TEM image of HPCR. Reproduced with permission from Ref. [86]

3.1.2 1D Carbon Materials

Carbon nanotubes (CNTs) are a type of typical one-dimensional (1D) carbon material that possess conductive characteristics and provide electron pathways for Li–S batteries with increasing performances [88,89,90,91]. CNTs can be used as conductive networks and mixed with sulfur to fabricate carbon–sulfur cathodes and aligned CNT/sulfur electrode materials with high sulfur loadings of up to 90 wt% through ball milling [92]. And although these aligned CNT/sulfur electrode materials can achieve higher energy densities than electrodes possessing lower sulfur loadings, their stability is still an issue. Moreover, the specific surface of CNTs is small and lacks porosity, leading to a lack of void spaces to encapsulate sulfur and restrain the shuttle effects of polysulfides. This issue can be resolved, however, by using CNT networks or 3D structures [93,94,95].

One method of overcoming the drawbacks of CNTs is to use multi-walled carbon nanotubes (MCNTs). For example, Zhao et al. [93] used hollow porous carbon to encapsulate MCNTs. (Preparation and synthesis methods are displayed in Fig. 4a.) They used MCNTs as templates and decorated the templates with SiO2 and a porous SiO2 layer. The researchers in this study reported that the SiO2 shell can occupy spaces and separate inner MCNTs from outer organosilicon compounds which play a role as carbon precursors and pore-forming agents. And after carbonization and the SiO2 etching process, the result tube-in-tube carbon nanostructure (TTCN) was acquired (Fig. 4b). The researchers reported that this resulting TTCN possessed many advantages such as the ability to provide adequate spaces to accommodate sulfur and the ability to prohibit the dissolution of polysulfides using the porous structures of the carbon layers. And as a result, the researchers reported a sulfur loading of 71 wt% and a discharge capacity of 918 mAh g−1 at 500 mA g−1 using this carbon/sulfur composite. Additionally, Jin et al. [90] reported that small CNTs can be encapsulated into larger CNTs following a process shown in Fig. 4c. In this study, anodic aluminum oxide was first used as a template to prepare a CNT layer with a large diameter of about 200 nm. CuO nanoparticles inside the larger CNT were subsequently reduced to Cu during annealing with Cu acting as a catalyst and promoting the growth of small CNTs inside the larger CNT (CNT@CNT) (shown in Fig. 4d). This resulting CNT@CNT was reported to possess large void spaces in the interior of the structure which can contain active materials and trap large amounts of sulfur up to 85.2 wt% inside the carbon host. (The S-CNT@CNT composite is shown in Fig. 4e.) In addition, the small CNTs in the inner tube can build electronically conductive networks to improve the conductivity of the composite and enhance the utilization of sulfur. And because of these beneficial properties, the CNT@CNT in this study provided high stability, high rate performances, and effectively impeded the dissolution of polysulfide intermediates, with the discharge capacity of a CNT@CNT-based cathode being 1193 mAh g−1 at 167 mA g−1 after 100 cycles. And even at increased current densities of 1670, 3340 and 8350 mA g−1, the discharge capacities of the electrode reached 1146, 1121 and 954 mAh g−1, respectively.

Fig. 4
figure 4

a Schematic illustration for the preparation of S-TTCN composite, b TEM image of the TTCN. Reproduced with permission from Ref. [93]. c Schematic illustration for the template growth of S@CNT and S-CNT@CNT, d TEM image of CNT@CNT, e TEM image of S-CNT@CNT. Reproduced with permission from Ref. [90]

Another viable approach to achieve high sulfur loadings and construct conductive pathways by using CNTs is to fabricate electrodes with CNT networks. For example, Fang et al. [96] assembled a single-wall carbon nanotube (SWCNT) network electrode that was capable of achieving an ultrahigh sulfur content reaching 95 wt%. In this study, the theoretical sulfur content in the SWCNT composite was first investigated based on SWCNTs that we recoated with sulfur possessing core–shell structures. And in accordance with the calculations, the sulfur content varied with the thickness of the sulfur shell. In addition, the researchers reported that the sulfur layers can uniformly wrap SWCNTs with a thickness of around 6 nm and that the resistance of the electrode was similar to that of the SWCNT network. Moreover, the SWCNT network can promote the transportation of electrons and ions as well as adsorb polysulfides during cycling, allowing this carbon matrix-based sulfur cathode to exhibit good cycling performances at high sulfur loading masses. And in practical results, this study reported that with an areal sulfur loading of 4.8 mg cm−2, the initial discharge capacity of the electrode was about 1212 mAh g−1 and the capacity retained 842 mAh g−1 at 170 mA g−1 after 140 cycles.

3.1.3 Graphene

Graphene is a typical 2D carbon material with a high theoretical specific area of 2630 m2 g−1 [80], a high electrical conductivity whose charge carrier mobility is > 2 × 105 cm2 V−1 s−1 [97] and a high Young’s modulus of > 0.5~1 TPa, suggesting high mechanical strength [97, 98]. These advantages of graphene make it a promising substrate to be used in energy storage and conversion systems, especially for Li–S batteries.

As an example, Chen et al. [99] utilized reduced graphene oxide as a substrate and monodispersed sulfur nanoparticles with the average diameters of 5, 10, 20, 40 and 150 nm onto it. Here, it is worth noting that the electrochemical performance of Li–S batteries can be significantly affected by the size of sulfur particles. And with increasing active substance sizes, sulfur utilization becomes limited because of the electronically and ionically insulating nature of sulfur and lithium sulfide. As a result, in the case in which 5 nm sulfur was dispersed on the graphene substrate, the graphene/sulfur cathode composite was reported to exhibit a high initial capacity of 1672 mAh g−1 at 167.2 mA g−1 and a capacity of 1089 mAh g−1 at 6688 mA g−1. The researchers attributed this performance to the small size of the dispersed sulfur, indicating that minimizing the size of sulfur can facilitate electrochemical kinetic processes during cycling.

In addition, the template method is also an effective method to prepare high-quality Li–S battery graphene materials. For example, Zhao et al. [100] used MgAl-layered double hydroxide (LDH) nanoflakes as templates to prepare unstacked double-layer template graphene (DTG) using chemical vapor deposition (CVD). (Process is shown in Fig. 5a.) In this process, LDHs were transformed into mesoporous MgAl-layered double oxide (LDO) flakes in the calcination process and the utilization of these flakes can not only separate the two graphene layers growing on the top and the bottom sides of the nanoflakes without stacking, but also induce graphene with a porous structure (Fig. 5b–c). Because of this, the specific surface area of the resulting DTG reached 1628 m2 g−1 with a mesoporous structure and after sulfur was loaded into the DTG to fabricate sulfur cathodes, the initial capacity reached 1084 mAh g−1 at 1672 mA g−1 and decreased to 701 mAh g−1 after 200 cycles. Additionally, this resulting cathode provided a high rate performance with a capacity of 1034 and 734 mAh g−1 at a current density of 8360 and 16720 mA g−1, respectively. Similarly, Tang et al. [101] reported that CaO, an alkaline-earth metal oxide, can act as a template and a catalyst to promote the growth of porous graphene by using chemical vapor deposition. Here, the resulting graphene also possessed a hierarchical porous structure and exhibited a high cycling stability and high rate performances in Li–S batteries with a specific capacity of 656 mAh g−1 at 8360 mA g−1.

Fig. 5
figure 5

a Schematic synthesis process of unstacked DTG, bc HRTEM images of graphene nanosheets cast onto LDO flakes. Reproduced with permission from Ref. [100]

3.1.4 3D Carbon Materials

3D carbon framework sulfur cathodes of Li–S batteries are mainly synthesized through fabricating substance transportation channels and conductive networks [102, 103], providing many advantages. These include: (1) large void spaces that can accommodate large amounts of sulfur, providing high active material loading masses in the cathode; (2) large quantities of transportation channels inside the framework, leading to rapid transportation rates of electrons and ions; and (3) conductive networks that can decrease composites resistances and promote sulfur utilization. Therefore, assembling Li–S battery cathodes by using 3D carbon frameworks can enhance electrochemical performances.

One typical 3D carbon framework is based on carbon nanotubes. As an example, Yuan et al. [104] studied 3D carbon frameworks based on carbon nanotubes and reported that 10~50 μm in length MWCNTs can act as both the sulfur host and the conductive pathways and that 1000~2000 μm in length vertically aligned CNTs (VACNTs) can act as both the long-range conductive frameworks and the binders in CNT paper electrodes (Fig. 6a–b). In this study, the researchers did not adopt the synthesis process of adding sulfur into VACNT/MWCNT paper which would not have favor the uniform distribution of sulfur. Instead, they dispersed sulfur onto MWCNTs first and subsequently mixed this with VACNTs to fabricate MWCNT/VACNT paper electrodes, avoiding the aggregation of sulfur and allowing for the uniform dispersion of sulfur inside the matrix. And in terms of utilization as a Li–S battery cathode, the resulting VACNT/MWCNT paper electrode was found to possess a high areal sulfur loading mass of up to 6.3 mg cm−2, an initial discharge capacity of 995 mAh g−1 and a residual discharge capacity of ~ 700 mAh g−1 at 0.38 mA cm−2 after 150 cycles. In another example, Li et al. [105] were able to achieve an ultrahigh areal sulfur loading of 19.1 mg cm−2 by fabricating a 3D CNT foam that provided a discharge capacity of 1039 mAh g−1 at a current density of 167.5 mA g−1.

Fig. 6
figure 6

a Schematic illustration of the VACNT/MWCNT paper sulfur cathode, b SEM image of VACNT/MWCNT paper. Reproduced with permission from Ref. [104]. c Schematic illustration of three current collector prototypes for Li–S batteries. Reproduced with permission from Ref. [106]

In addition to the previously mentioned advantages of 3D carbon frameworks, Peng et al. [106] also reported that these carbon architectures can prohibit the electrochemical corrosion of Al foil electrode current collectors, leading to improved cycling stabilities. There are three types of current collectors utilized in Li–S batteries, including 2D Al foils, 2D graphene films (2D GF) and 3D CNT current collectors (Fig. 6c). Of these, 2D Al foils cannot resist the corrosion of electrolytes and are often badly damaged over time. As for 2D GF, although they are stable, they still generate passivation layers on cathode surfaces. The construction of a 3D CNT film, however, is effective in preventing both the corrosion of the current collector and the formation of passivation layers because of its chemical stability and porous structure framework. Therefore, the protection of current collectors by using 3D carbon frameworks is essential in the development of Li–S batteries with enhanced cycling stabilities.

Another synthesis process to obtain 3D graphene frameworks through the growth of graphene on Ni foam can use CVD methods [107]. Cheng et al. [108] grew graphene on Ni foam and coated the resulting graphene foam (GF) with poly(dimethylsiloxane) which can enhance the flexibility of the electrode. Here, the GF was used as both the current collector and the sulfur host and possessed a sulfur content of up to 70 wt%, providing a residual capacity of 448 mAh g−1 at 1500 mA g−1 after 1000 cycles at an areal sulfur loading of 10.1 mg cm−2.

3.2 Polarity with Functional Groups

Before summarizing the advancements of functional groups in carbon materials, it is essential to understand the mechanisms of chemical adsorption between polysulfides and functional carbon matrixes. Based on this, Cui et al. [109] used ab initio simulations based on density-functional theory to analyze the interactions between lithium sulfides and functional groups. Here, the researchers used frameworks based on vinyl polymers –(CH2–CHR)n– in which R represented functional groups including amine, ester, amide, ketone, imine, ether, disulide, thiol, nitrile, sulide, luoroalkane, chloroalkane, bromoalkane and alkane to understand the binding energies between R groups and lithium sulide or lithium polysulides [15, 109]. Here, the researchers reported that the binding energy between alkane and LiS which represented end groups in lithium polysulfides intermediates was the lowest, exhibiting a weak affinity, whereas oxygen and nitrogen groups such as amine, ester and amide exhibited high binding energies with LiS species because of the stable coordination interactions in Li–O and Li–N. These calculations were further confirmed by experimental results and have been commonly used in heteroatom-doped carbon-based materials [110,111,112]. To further understand the mechanisms of chemical adsorption of sulfur in cathodes, Song et al. [113] used a combination of X-ray adsorption near-edge structure spectroscopy (XANES) and theoretical simulations to analyze interactional mechanisms and reported them based on the adsorption of sulfur examined by using XANES (Fig. 7a–b). It was noteworthy that there were no distinct changes in the coordination structure of oxygen without nitrogen doping. However, if the carbon matrix was doped with nitrogen, oxygen coordination structures changed significantly after the sulfur loading, suggesting strong interactions between oxygen functional groups and sulfur. And based on this, this study demonstrated that the nitrogen doping in carbon matrixes can enhance the chemical adsorption effects of oxygen functional groups, especially for nearby oxygen atoms. Furthermore, Hou et al. [114] investigated the lithium bond chemistry of Li–S batteries by combining quantum chemical calculations with 7Li nuclear magnetic resonance (NMR) spectroscopy (Fig. 7c–e). Here, the theoretical predictions were consistent with obtained 7Li NMR spectra of lithium polysulfides in which the chemical shift can quantitatively describe Li bond strengths. And based on this study, Li bonds have a dipole–dipole interaction with electron-rich donors, effectively promoting chemical adsorption interactions between lithium polysulfides and polarized sulfur hosts.

Fig. 7
figure 7

a Oxygen K-edge XANES spectra of MPNC and MPNC-S nanocomposites, b oxygen K-edge XANES spectra of MPC and MPC-S nanocomposites. Reproduced with permission from Ref. [113]. c Schematic diagram of a Li bond, d theoretically calculated and e experimentally obtained 7Li NMR spectra of Li2S8 before and after interacting with pyridine. Reproduced with permission from Ref. [114]

3.2.1 Porous Carbon Materials with Functional Groups

Conventional porous carbon materials can absorb polysulfide intermediates through physical interactions; however, the affinity between non-polar carbon hosts and polar lithium polysulfides is weak [115,116,117,118]. Therefore, carbon surfaces must be polarized to generate stronger chemical interactions between modified carbon matrixes and polysulfides. Here, doping heteroatoms into carbon materials is a good approach to achieve this functionality on matrix surfaces. As an example, Xia et al. [115] prepared N-doped porous carbon microspheres (NPCMs) using the heat treatment of microalgae (Fig. 8). And based on the Barrett–Joyner–Halenda (BJH) analysis, the resulting NPCMs material, a type of hierarchical porous carbon, possessed a pore size distribution of 5 nm and a distribution ranging from 10 to 50 nm. As a result, through combining the physical and chemical adsorption of polysulfide intermediates, the resulting NPCMs exhibited a high capacity of 1030.7 mAh g−1 with 91% retention capacity at 100 mA g−1 after 100 cycles. To further strengthen the interactions between the cathode and polysulfides, Chen et al. [119] decorated the surface of a carbon nanocube-sulfur cathode (P@CNC-S) using poly(3,4-ethylene-dioxythiophene) (PEDOT), resulting in a carbon nanocube with a large surface area of 2425 m2 g−1 and a pore volume of 3.72 cm3 g−1, providing high void spaces to encapsulate sulfur. And because of this, the resulting P@CNC-S cathode provided a high reversible capacity of 1086 mAh g−1 at 1673 mA g−1 and 842 mAh g−1 at 8365 mA g−1 after 1000 cycles, displaying long cycling stability.

Fig. 8
figure 8

a Schematic representation of the fabrication of NPCMs, b SEM image and c HRTEM image of NPCSMs, d cycling stability and Coulombic efficiency of NPCSMs electrodes. Reproduced with permission from Ref. [115]

Metal–organic frameworks (MOFs) also possess porous structures with high surface areas, and carbon materials derived from MOFs have been extensively applied in energy storage and conversion systems, such as fuel cells [120, 121], supercapacitors [122] and batteries [123,124,125,126]. Hao et al. [127] further applied MOFs to prepare MOF-derived 2D nanocarbon materials using thermal exfoliation to synthesize effective Li–S battery cathodes. In their study, the researchers selected copper 4,4′-bipyridine MOFs (Fig. 9a) composed of 2D building units stacked with each other through π−π interactions. Here, the van der Waals interactions between the stacked units of the 2D layer-structured MOF were relative weak, allowing the MOFs to easily exfoliate to 2D layers with a size of 10 μm and a nanoscale thickness of about 1~2 nm. And through annealing these exfoliated MOFs units, they converted them to nanocarbon stacks. The researchers reported that the exfoliation step of the MOFs is necessary in their study to obtain nanocarbons with the desired structure as discussed before. Finally, the nanocarbon stacks are put through a thermal process at 500 and 900 °C to obtain samples (UHCS-500, UHCS-900) with surface areas of 900 and 937 m2 g−1, respectively, and both samples possessed similar narrow pore size distribution ranges at about 0.7 nm. (The TEM images of UHCS-900 are shown in Fig. 9b–c.) In the subsequent testing, the researchers reported that the polarized surfaces of the nanocarbons exhibited hydrophilicity with high heteroatom dopings including N and O, and as a result, the synergistic functions of the microporous structure and the polar surface in the nanocarbons produced strong affinities with polysulfide intermediates, leading to nearly 100% cathode Coulombic efficiency.

Fig. 9
figure 9

a Structure of the layered MOF crystal and its building blocks, b, c TEM images of UHCS-900. Reproduced with permission from Ref. [127]

The sulfur loading of cathodes in Li–S batteries must reach 4~6 mg cm−2 to compete with LIBs; therefore, high sulfur loading is required to meet practical application demands [13, 128, 129]. However, because of the insulating nature of sulfur, the high sulfur loading results in low rate capabilities and rapid capacity fading. To achieve high volumetric capacity, Hu et al. [128] synthesized monodispersed porous nitrogen-doped carbon nanospheres (NCNSs) with a diameter about 350 nm using a template method. Here, the structure of the resulting carbon material was found to be interconnected with closely packed nanosphere clusters which are beneficial to the fabrication of sulfur/carbon cathodes with high sulfur loadings and good conductivity. Moreover, the surface area of the resulting NCNS reached 2900 m2 g−1 with a total pore volume of 2.3 cm3 g−1 and a pore size distribution dominated by microporous structures. And because of this, the sulfur loading density of the resulting NCNSs reached 5 mg cm−2. And at such high sulfur loading masses, the initial discharge capacity of the resulting battery was 1196 mAh g−1 with a retention capacity of 672 mAh g−1 at 836 mA g−1 after 500 cycles.

3.2.2 1D Carbon Materials with Functional Groups

Conductive CNTs are suitable hosts for sulfur in cathode materials, but their lack of surface polarity sites is not favorable for the enhancement of further polysulfide interactions. Based on this, Peng et al. [130] separately investigated pristine CNTs, O-doped CNTs and N-doped CNTs as sulfur hosts to construct stable Li–S batteries. In this study, the researchers revealed that although functional groups containing oxygen can interact with sulfur, oxygen doping can lead to poor conductivity and faster capacity decay. And according to calculations, pyridinic-N and quaternary-N can exhibit higher binding energies with Li2S/Li2S4, in which pyridinic-N can act as electron donation sites to Li in Li2Sx (x = 1~4), whereas sulfur-containing species can possibly interact with the neighboring carbon sites of quaternary-N. In addition, the researchers in this study reported that nitrogen doping in CNT skeletons can not only enhance the interactions between N-doped CNTs and sulfur species, but also provide high conductivity. Therefore, N-doped CNTs can provide the highest cycling performance in comparison with pristine CNTs and O-doped CNTs. And in the tests, at a current density of 167.2 mA g−1, the initial capacity of N-doped CNTs was found to be 937 mAh g−1 with a residual capacity of 645 mAh g−1 after 200 cycles.

Other types of 1D carbon materials including carbon tubes and carbon fibers have also been utilized as sulfur hosts in Li–S batteries, and several studies have used anodic aluminum oxide templates to synthesize carbon hollow tubes [131, 132]. Here, the hollow structure of the carbon tubes can accommodate sulfur in its empty central part but cannot trap polysulfides effectively because of the non-polar feature of the carbon host. To overcome this problem and enhance the cycling stability of such sulfur cathodes, carbon hollow tubes derived from anodic aluminum oxide templates can be modified with functional groups to possess strong chemical interactions with polysulfides. As an example, Zheng et al. [118] modified an electrode using polyvinylpyrrolidone (PVP) with functional groups containing oxygen atoms to effectively trap polysulfides, and the resulting PVP-modified electrode demonstrated better electrochemical performances with cathode capacities of 1180, 920 and 820 mAh g−1 at current densities of 334.6, 836.5 and 1673 mA g−1, respectively. Other than carbon hollow tubes, carbon nanofibers can also improve the cycling performances of Li–S batteries and electrospinning is a method that has been extensively used to prepare carbon fiber precursors [133, 134]. For example, Xu et al. [133] prepared carbonized fibers containing polyacrylonitrile and iron(III)(acac)3 and after etching and activation steps, the resulting carbon nanofiber not only possessed a hierarchical porous structure but also achieved a nitrogen doping of 7.5 at%, suggesting that carbon fibers can effectively immobilize polysulfides and provide better electrochemical performances. In this study, the composite sulfur loading reached 71 wt%, and the results showed an initial capacity of 945 mAh g−1 and a retention capacity of 765 mAh g−1 at 1675 mA g−1 after 200 cycles.

3.2.3 Graphene with Functional Groups

The graphene oxide is an important derivative of graphene and possesses surface functionality and performance controllability. And because of the existence of oxygen functional groups on graphene oxides, it also possesses strong interactions with sulfur and lithium polysulfides [135]. However, although graphene oxide can effectively immobilize polysulfides, its poor conductivity limits further applications. Because of this, several studies have investigated mildly oxidized, reduced and modified graphene oxides [136,137,138]. For example, Wang et al. [137] reported that poly(ethylene glycol) and mildly oxidized graphene oxide layers can be utilized to buffer volume changes and immobilize polysulfides during cycling. Additionally, Fang et al. [139] presented an interesting method to enhance the electrochemical performance of graphene substrates (Fig. 10a) through a combination of the conductive characteristics of graphene and the surface functionalities of graphene oxides. Here, graphene with high conductivity can be used as a current collector to enable the rapid transportation of electrons and partially oxidized graphene can provide anchoring sites for polysulfides. Ultimately, this all-graphene cathode can provide high cycling performances even at the sulfur loading of up to 5 mg cm−2, with an initial cathode discharge capacity of 1500 mAh g−1 and a residual capacity of 841 mAh g−1 after 400 cycles at 340 mA g−1.

Fig. 10
figure 10

a Schematic illustration of the all-graphene sulfur cathode. Reproduced with permission from Ref. [139]. b Schematic illustration of the stabilization of quinonoid imines by using phytic acid, and the reversible transition between protonated and deprotonated states with polysulfide desorption and adsorption. c, d TEM images of the NPGO. Reproduced with permission from Ref. [140]

As for modified graphene materials, Wang et al. [116] utilized ethylene diamine-functionalized reduced graphene oxide (EFG) as a platform to trap sulfur and lithium sulfide and reported stable cycling performances. In their study, ethylene diamine was grafted with the graphene oxide at 75 °C, allowing the graphene oxide to be partially reduced and conductive to a certain extent. And based on density-functional theory calculations, the binding energies between lithium sulfides and the resulting EFG were 1.13~1.38 eV, indicating strong interactions between them. And as a result, the sulfur cathode based on the resulting EFG provided stable cycling performances with a capacity of 650 mAh g−1 at 836 mA g−1 after 350 cycles. Exploiting modified graphene oxides consisting of new functional groups is also an efficient method to obtain sulfur cathodes with high performances. Based on this, Chen et al. [140] designed the quinonoid imine-modified graphene oxide (NPGO) by polymerizing aniline and phytic acid on the graphene oxide (Fig. 10b–d). Here, phytic acid played an important role in facilitating the formation of quinonoid imine groups and stabilizing graphene oxide-based composites. In addition, phytic acid can also promote the protonation of quinonoid imine groups, increasing the ratio of –NH+=. And because of this, the formation of quinonoid imine onto the graphene oxide can effectively adsorb polysulfides and favor the precipitation of Li2S. And in this study, at a sulfur loading of 3.3 mg cm−2, the modified graphene oxide provided an area capacity of 4.4 and 3.7 mAh cm−2 at 335 and 1675 mA g−1 respectively.

Additionally, like other carbon materials discussed before, doped heteroatoms can functionalize graphene and enhance the adsorption of polysulfides. There are many types of heteroatom-doped graphene, including boron-doped [141, 142], nitrogen-doped [143], phosphorus-doped [144] and sulfur-doped graphene [145], and the synthesis of doped graphene with heteroatoms such as B [146], N [138], P [147] and S [148, 149] has been reported to enhance electrochemical performances and provide trapping sites for polysulfides, indicating that the heteroatom doping is an effective method to obtain high-performance Li–S batteries. Of these heteroatom-doped materials, nitrogen-doped graphene materials have been widely explored in Li–S batteries. In one example, Hou et al. [150] investigated the interactions between doped nitrogen atoms in graphene and polysulfides by conducting various theoretical analyses in the field. And according to their computational results, pyrrolic-N and pyridinic-N both possess strong interactions with polysulfides, with pyrrolic-N possessing a slightly higher binding energy to polysulfides than pyridinic-N. In addition, the binding energy of quaternary-N to polysulfides was found to be weaker in comparison with that of pyrrolic-N and pyridinic-N, because quaternary-N cannot provide an extra lone-pair of electrons to interact with lithium [15]. Overall, this study provides valuable guidance for researchers to design methods to adjust the concentration of functional nitrogen doping sites to obtain high-performance Li–S batteries. And in accordance with these theoretical results, Song et al. [151] synthesized nitrogen-doped graphene with 6.53 wt% of doped nitrogen including pyrrolic-N and pyridinic-N and reported that at an areal sulfur loading of 5 mg cm−2, stable cycling performances and a high rate capability of 650 mAh g−1 at 1580 mA g−1 can be obtained.

3.2.4 3D carbon Frameworks with Functional Groups

Electrospinning can also be applied to fabricate three-dimensional freestanding electrodes in Li–S batteries. Here, carbon nanofibers interconnected with each other can construct freestanding electrodes without the addition of binders or current collectors, and are beneficial to the assembly of high-energy density batteries [152]. In one example, Yao et al. [153] constructed C–S bonds to restrain polysulfides using chemical bonding through the thermal treatment of electrospinning nanofibers and S0.6Se0.4. And because of the chemical interactions between Se and S, as well as between S0.6Se0.4 and the carbon nanofiber substrate, the shuttle effects of polysulfides were effectively suppressed, and at a current density of 1000 mA g−1, a retaining capacity of 346 mAh g−1 after 1000 cycles was reported, demonstrating stable cycling performances.

As a typical 3D carbon framework, 3D graphene frameworks can be fabricated by using solvothermal methods [154, 155]. In one study, Wang et al. [156] synthesized 3D nitrogen-doped graphene (3D-NG) derived from the graphene oxide and the ammonia solution as a nitrogen source in a solvothermal process. This resulting 3D-NG was reported to be capable of accommodating large amounts of sulfur, with the sulfur content reaching 87.6 wt%. In addition, because of the conductive configuration and polarized surface on the 3D-NG, it also improved cathode conductivity and effectively stabilized polysulfides.

Furthermore, in a study by Hu et al. [157], 3D hybrid graphene with graphene foam as the substrate was grown on the Ni foam via CVD and interconnected with the reduced graphene oxide. Here, the introduction of the reduced graphene oxide which still contained some oxygen functional groups can stabilize sulfur and enhance interactions with polysulfides. In addition, because of the large void spaces inside the graphene hybrid, the resulting 3D hybrid graphene was capable of accommodating large amounts of sulfur, with an areal sulfur loading reaching 9.8 mg cm−2. And based on this areal sulfur loading, a sulfur content of 83 wt% was achieved, with the 3D hybrid graphene exhibiting good rate capability and cycling performance. Furthermore, the researchers in this study reported that at an increased current density of 3350 mA g−1, the discharge capacity of the 3D hybrid graphene was higher than 538 mAh g−1 and returned to over 800 mAh g−1 at 83.75 mA g−1, comparable to the initial capacity during the rate testing. In addition, at 335 mA g−1, the graphene hybrid sulfur cathode provided an initial capacity of 1000 mAh g−1 and a retention capacity of 645 mAh g−1 after 350 cycles.

3.3 Affinity in Carbon-Based Composites

3.3.1 Carbonaceous Carbon Composites

Carbonaceous carbon composites are the mixtures of two or more types of carbon materials that can combine the advantages of each component and provide high performances due to synergistic effects [158,159,160]. In terms of Li–S batteries, carbonaceous carbon composites can be divided into two sections: conductive frameworks in sulfur cathodes and functional sections [161, 162]. For example, the conductivity of the electrode can have powerful influences on the performance of Li–S batteries. And because of the insulating nature of sulfur and lithium sulfide, the construction of beneficial conductive frameworks in the electrode is essential to provide conductive pathways for electrons and enhance the utilization percentage of sulfur. Therefore, the fabrication of networks with good conductivity is an essential aspect to obtain high-performance Li–S batteries. Unfortunately, although various carbon materials possess high conductivity, such as carbon nanotubes and graphene [47, 80, 161], they lack porosity configurations and possess non-polar surfaces. Here, poor porosity indicates a lack of void spaces to accommodate large amounts of sulfur, limiting the sulfur content of cathodes and decreasing the energy density of Li–S batteries. Additionally, the interactions between non-polar carbon materials and polysulfides are weak, resulting in the inability of carbon substrates to effectively restrain the shuttle effects of polysulfides, leading to poor cycling stability. Therefore, optimal conductive carbon substrates should be combined with functional carbon materials possessing porous structures or functional groups to enhance sulfur loadings and act as trapping sites for polysulfides.

For example, Sun et al. [163] prepared carbon hybrid cathodes based on mesoporous carbon CMK-3 and CNTs and reported high electrochemical performances. Here, mesoporous carbon can provide void spaces to accommodate sulfur and its porous structure can inhibit the diffusion of polysulfide intermediates. In addition, the aligned CNTs can not only provide conductive networks for the transportation of electrons and ions, but also interconnect the sulfur host based on mesoporous carbons. And as a result of the combination of these two materials, the researchers reported that at an areal sulfur loading of 1 mg cm−2, this laminated carbon hybrid exhibited high electrochemical performances with a discharge capacity of 1226 mAh g−1 and a retention capacity of 919 mAh g−1 after 100 cycles at 167.5 mA g−1. In addition, at an ultrahigh areal sulfur loading of 20 mg cm−2, the capacity maintained at 900 mAh g−1 at 167.5 mA g−1.

Doping heteroatoms into functional carbon sections is also an efficient method to increase the cycling performance of carbonaceous carbon nanotube composites. For example, Song et al. [164] dispersed CNTs inside nitrogen-doped micrometer-sized carbon spheres (MNCS-CNT) which act as sulfur hosts and provide chemical adsorption sites for polysulfides (Fig. 11a), and reported that at an areal sulfur loading of 5 mg cm−2, the resulting carbon sphere hybrid exhibited high cycling performances with an initial discharge capacity of 1438.3 mAh g−1 and a retaining capacity of 1200 mAh g−1 after 200 cycles. In addition, Pan et al. [165] also reported that CNTs can be modified using imidazolium-based ionic polymers acting as the source of nitrogen-doped porous carbons. And in this study, the researchers reported that the modified CNTs demonstrated significantly enhanced specific surface areas, increasing from 93 to 387 m2 g−1. And because of this enhanced specific surface area, the resulting carbon hybrid exhibited good electrochemical performances with the corresponding carbon hybrid sulfur cathode producing a discharge capacity of 1065 mAh g−1 at 837.5 mA g−1 and a retaining capacity of 817 mAh g−1 after 300 cycles. Furthermore, Cai et al. [166] reported that porous carbon–CNT hybrids can be prepared through a novel method of combining template synthesis and catalysis growth. In this study, nanosized CaCO3 was used as a template to synthesize nitrogen-doped porous carbon by using melamine–formaldehyde (MF) as a precursor. And because of the existence of a Fe-Co catalyst, these nitrogen-doped CNTs grew on the porous carbon as well, resulting in CNTs inside such carbon hybrids being bifunctional, capable of acting as conductive additions and blocking some large pores to reduce shuttle effects.

Fig. 11
figure 11

a Synthesis process of the MNCS/CNT composite. Reproduced with permission from Ref. [164]. b Schematic illustration of S/(G-GCNs) composite, c TEM image of G@FexOy, d TEM image of G-GCNs. Reproduced with permission from Ref. [173]. e Schematic illustration of ICFs Li–S cathode. Reproduced with permission from Ref. [174]

Aside from amorphous carbon with porous structures, carbonaceous carbon composites existing as CNTs can also be integrated with graphene, and these graphene/CNT cathode materials can be fabricated through a one-step catalytic approach [167]. For example, Zhao et al. [167] used FeMgAl-layered double hydroxide (FeMgAl LDO) as both a template and a catalyst and achieved the growth of graphene and SWCNTs in one step. In this study, it was reported that the co-growth of graphene and SWCNTs only occurred at temperatures above 950 °C in the CVD process with FeMgAl LDO as an iron source, and that the acquired graphene/CNT hybrid possessed a high conductivity of 3130 S cm−1 and a surface area of 806.2 m2 g−1. Additionally, at a sulfur content of 60 wt% in the resulting carbon hybrid, the capacity was reported to be 928 mAh g−1 at 1672 mA g−1. And even at 8360 mA g−1, the capacity could still reach 650 mAh g−1 after 100 cycles. Similarly, nitrogen-doped aligned carbon nanotube/graphene hybrids can be synthesized with template methods [168]. For example, Tang et al. [168] used metal oxide flakes as templates for the growth of graphene and CNTs, which was catalyzed by metal nanoparticles. Here, nitrogen doping was achieved by introducing NH3 as a nitrogen source during the CVD process. And as a result, the researchers reported that the resulting carbon hybrid after the nitrogen doping possessed increased sulfur utilization and exhibited high electrochemical performances with a high initial capacity of 1152 mAh g−1 at 1672 mA g−1. In another example, Zhang et al. [169] used the pre-prepared graphene oxide as a substrate and catalyzed the growth of CNTs onto it using a cobalt catalyst. In addition to the graphene oxide, these researchers also used urea as a carbon and nitrogen source to prepare CNTs and achieved nitrogen doping in the resulting carbon hybrid (3D GN-CNT). And as a result, the 3D configuration of the 3D GN-CNT matrix allowed for formation of rapid transport channels for electrons and electrolytes, and the heteroatom doping allowed for the effective immobilization of polysulfide intermediates, resulting in a composite with a high initial discharge capacity of 1049.6 mAh g−1 and a residual capacity of 639.1 mAh g−1 after 200 cycles at 167.5 mA g−1 with slow capacity fading rates.

Because of the large amounts of oxygen functional groups in the graphene oxide, it can be considered as a type of functional graphene. Based on this, Chen et al. [170] used MWCNTs as sulfur hosts and mixed an MWCNT/sulfur composite with the graphene oxide to construct a graphene-based hierarchical carbon hybrid. Here, although the graphene was reduced to graphene sheets during the synthesis process, the researchers can still detect the existence of oxygen because of the existence of residual functional groups. In addition, the configuration of the carbon composite can provide 3D conductive networks, buffering volume changes during electrochemical reaction processes and restraining solvable polysulfides. And as a result, the resulting carbon composite produced a high initial capacity of 1396 mAh g−1 at 334.4 mA g−1, demonstrating increased utilization percentages of sulfur.

Yu et al. [171] also reported that nitrogen-doped graphene, if used as a component for carbonaceous carbon composites, can not only improve the conductivity of the composites, but also inhibit the shuttle effects of polysulfides. In their study, the researchers constructed a carbon composite possessing a carbon configuration based on interconnected nitrogen-doped graphene/sulfur composites dispersed in a CNT/nanofibrillated cellulose framework. And because of the hierarchical architecture and functional groups in the 3D carbon framework, the transport of electrons and electrolytes was enhanced, accommodating increased sulfur loadings and confining the shuttle effects of polysulfides. And at a sulfur loading of 8.1 mg cm−2, this carbon composite provided an initial capacity of 935.6 mAh g−1 and a retaining capacity of 303.2 mAh g−1 after 1000 cycles at 837.5 mA g−1, demonstrating slowed capacity fading rates. Aside from the nitrogen doping to improve the cycling performances of carbon composites in Li–S batteries, Yuan et al. [172] reported that the co-doping of heteroatoms such as boron and nitrogen can also have positive influences on sulfur cathode materials. In their study, graphene-based materials were functionalized with a nitrogen and boron co-doped carbon layer and produced enhanced performances. Here, negative nitrogen atoms and positive boron atoms can separately adsorb lithium ions and polysulfides, effectively enhancing cycling stabilities of carbon-based sulfur cathodes. And because of this, the graphene-based N/B co-doped carbon layer examined in this study exhibited a high discharge capacity and a stable cycling performance at 836 mA g−1.

Another method to construct graphene-based carbonaceous carbon composites is to combine porous carbon materials with a graphene matrix. For example, Zhang et al. [173] utilized graphene as a substrate that was decorated with graphitic carbon nanocages (Fig. 11b–d), which resulted in a graphene-based composite that possessed a high specific surface area of 1210 m2 g−1 with an average pore size of 3 nm. And because of the architecture of this graphene composite, corresponding electrochemical nanoreactors in Li–S batteries can demonstrate improved cycling performances and rate capabilities for sulfur cathodes. In this study, a tested sulfur cathode based on the resulting graphene composite provided an initial capacity of 900 mAh g−1 with a high retaining capacity of 706 mAh g−1 after 1000 cycles at a current density of 1675 mA g−1. In addition, Wu et al. [174] used zinc nanoparticles as templates to synthesize 2D “bubble-like” interconnected carbon fabrics (ICFs) that were encapsulated by the reduced graphene oxide after a hydrothermal process (Fig. 11e). Here, the reduced graphene oxide acted as a backbone for the composite and binder-free sulfur cathodes can be fabricated. And as a result, the resulting composite exhibited high cycling performances with an initial capacity of 1149 mAh g−1 and a retaining capacity of 892.3 mAh g−1 after 200 cycles.

3.3.2 Inorganic Carbon-Based Composites

Inorganic materials, such as metal oxides [175, 176], metal sulfides [177], metal carbides [178] and other polar inorganic substances [179, 180] that possess stronger affinities for polysulfides and provide adsorption sites in comparison with non-polar carbon matrixes, have been widely applied to Li–S batteries and have been demonstrated to effectively improve electrochemical performances and cycling stabilities of Li–S batteries [181, 182]. Moreover, the combinations of inorganic substances with carbon matrices can also promote electron transport, further improving the cycling performances of Li–S batteries [183].

Metal oxides, possessing strong polar surfaces, can anchor polysulfide intermediates more tightly and significantly improve the electrochemical performances of Li–S batteries [184]. However, metal oxides also possess low electrical conductivities, hindering their utilization as active materials in Li–S battery cathodes. Therefore, the combination of metal oxides with strong polar surfaces and carbon materials with high conductivities is an effective method to improve the electrochemical properties of Li–S batteries [185, 186]. For example, the utilization of MnO2 in Li–S batteries has been reported to provide interaction sites with lithium polysulfides [175]. Based on this, Li et al. [187] filled hollow carbon nanofibers with MnO2(MnO2@HCF) (the reaction process shown in Fig. 12a) using MnO2 nanowires as templates. In their synthesis, MnO2 nanowires as the templates were first decorated with SiO2 and resorcinol formaldehyde. And after annealing, the reaction products were treated with NaOH, causing the MnO2 nanowires to transform into 2D nanosheets (Fig. 12b). The researchers attributed this phenomenon to the fact that α-MnO2 was transformed to an amorphous phase after annealing which can subsequently convert to birnessite-type MnO2 easily. And as a result, the resulting MnO2 carbon composite not only provided adequate void spaces to accommodate the sulfur loading of up to 71 wt%, but also possessed large interaction areas with polysulfides. Furthermore, the carbon nanofibers in the composite provided conductive pathways and promoted the transport of electrons and ions. Because of these factors, the researchers in this study reported that this composite exhibited good cycling performances with a high initial capacity of 1147 mAh g−1 and a retaining capacity of over 800 mAh g−1 after 100 cycles at a current density of 335 mA g−1, displaying high cycling stabilities. Hou et al. [188] also reported the use of birnessite-type MnO2 as trapping sites for polysulfides with the MnO2 being coated by carbon nanoboxes. Overall, these studies provide evidence that MnO2 has an affinity for polysulfides.

Fig. 12
figure 12

a Synthesis process and advantages of the MnO2@HCF/S composite, b TEM image of MnO2@HCF. Reproduced with permission from Ref. [187]. c Synthesis process of the TiO@C-HS/S composite, d TEM image of PS@TiO2@PDA spheres, e TEM images of TiO@C-HS, f cycling performance and Coulombic efficiency at 0.2 and 0.5 C of the TiO@C-HS/S electrode. Reproduced with permission from Ref. [190]

Titanium oxide materials belong to another type of metal oxides that are widely utilized in Li–S batteries [182]. For example, Zhou et al. [189] reported that TiO2 nanowire/graphene composites can be used as high-performance cathode materials in Li–S batteries, and according to their design strategy, graphene was chosen as a highly conductive building block and TiO2 can act as the functional section, possessing strong chemical adsorptions with polysulfide intermediates. And as a result, the resulting sulfur cathode in this study demonstrated a high cycling stability, with a residual capacity of 1053 mAh g−1 after 200 cycles at 335 mA g−1. To further enhance the electrochemical performance of titanium oxides in Li–S batteries, TiO@carbon hollow nanospheres can also be utilized as sulfur hosts (Fig. 12c–e) [190]. This is because TiO possesses high densities of oxygen and titanium vacancies, demonstrating theoretically higher electrical conductivities and stronger interactions with lithium polysulfides than TiO2. Therefore, TiO carbon composites can effectively restrain polysulfides and exhibit high electrochemical performances (Fig. 12f).

Other metal oxides have also been explored for sulfur cathodes, such as NiFe2O4 [191], Al2O3 [192], MgO [193], SiO2 [194], SnO [195], V2O5 [196], Fe3O4 [197], ZnO [198], CeO2 [199], and have demonstrated promising properties. For example, Fan et al. [191] synthesized a spinel carbon-based hybrid as the sulfur cathode in Li–S batteries which produced high performances by using CNTs to fabricate conductive networks and NiFe2O4 nanosheets to confine polysulfides. This resulting hybrid cathode, even after 500 cycles, retained a discharge capacity higher than 850 mAh g−1 at a current density of 1675 mA g−1. Furthermore, Hou et al. [194] prepared Si/SiO2@porous carbon spheres as a sulfur hosts to assemble high-performance Li–S batteries (Fig. 13). In this study, the researchers reported that Si/SiO2 possessed a strong affinity for lithium polysulfides and that the porous structure of the carbon composite can be adjusted by controlling the etching times with NaOH solution. And based on this study, a combination of physical and chemical adsorption with lithium polysulfides can effectively improve the cycling performance of Li–S batteries.

Fig. 13
figure 13

a Synthesis process of Si/SiO2@CS hybrid spheres, b schematic illustration of the concept of the “polysulfide reservoir” in the Si/SiO2@CS hybrid spheres. Reproduced with permission from Ref. [194]

In comparison with metal oxides, metal sulfides possess characteristically strong affinities for sulfur-containing species and some metal sulfides are metallic or half-metallic, meaning that they possess better conductivity than metal oxides [182, 200]. For example, CoS2/graphene composites can be synthesized as hosts for sulfur cathodes in which CoS2, being half-metallic, can provide a conductivity of 6.7 × 103 S cm−1 [201] and, as a result, can provide a strong affinity for lithium polysulfides and accelerate reaction kinetics. In addition, graphene sheets utilized as conductive substrates can further promote the transport of electrons and improve the electrochemical performance of Li–S batteries. In another example, Ye et al. [181] reported that because NiS possessed strong chemical interactions with polysulfides, high-performance NiS sulfur cathodes can be achieved through distributing NiS nanoparticles into carbon hollow spheres (Fig. 14a). And at an areal sulfur loading of 2.3 mg cm−2, the high-performance NiS sulfur cathode can provide a retaining capacity of 695 mAh g−1 after 300 cycles at a current density of 837.5 mA g−1 with a capacity retention of 96%. Chen et al. [202] also reported that a spinel-type metal sulfide, the Co3S4 nanoboxes, can deliver a high conductivity of 3.3 × 103 S cm−1 and can be used as the high-performance sulfur host in Li–S batteries as well. In this study, Co3S4 nanoboxes threaded by CNTs can be synthesized via a self-template approach in which the architecture is based on CNT 3D conductive pathways, and Co3S4 can serve as chemical adsorption sites for lithium polysulfides. And as a result, the resulting Co3S4-CNTs composite can exhibit long cycling stabilities and high rate performances even at 50 °C, delivering a high initial discharge capacity of 953 mAh g−1 and a retaining capacity of 718 mAh g−1 after 300 cycles at 335 mA g−1.

Fig. 14
figure 14

a Synthesis process of NiS@C-HS. Reproduced with permission from Ref. [181]. b Synthesis process of VN/G composite. Reproduced with permission from Ref. [206]

Metal carbides and metal nitrides can also deliver high conductivities and acceptable adsorption properties for polysulfides [178, 183, 203]. And as a typical type of metal carbides, MXene phases which combine the advantages of high conductivity and abundant chemical trapping sites for lithium polysulfides can be utilized as sulfur hosts in Li–S batteries [204]. For example, Liang et al. [205] synthesized MXene nanosheets, including Ti2C, Ti3C2 and Ti3CN, that were interconnected with CNTs which demonstrated high cycling performances. Here, the researchers suggested that the existence of CNTs in the composite can both fabricate the conductive networks of the sulfur cathode and guarantee the stability of the architecture, inhibiting the restacking of MXene nanosheets. The researchers also suggested that the high interactions between the MXene nanosheets of Ti2C, Ti3C2 or Ti3CN and lithium polysulfides were the result of a two-step process involving the formation of thiosulfate and the Lewis acid–base interaction of Ti–S bonds. And as a result, the MXene nanosheet/CNT composites exhibited high cycling performances with a retaining capacity of ~ 450 mAh g−1 at 837.5 mA g−1 after 1200 cycles. In terms of metal nitrides, Sun et al. [206] prepared a cathode for the Li–S battery based on a porous vanadium nitride/graphene composite that contained Li2S6 catholyte (Fig. 14b). And as a type of highly conductive metal nitrides, vanadium nitride, with a conductivity of 1.17 × 106 S cm−1 at room temperature, can combine the unique properties of strong chemical interactions with polysulfides and the catalytic functions of redox reaction kinetics, resulting in the vanadium nitride/graphene composite exhibiting a high discharge capacity of 1128 mAh g−1 at 1675 mA g−1 and a high cycling performance with a retaining capacity of 917 mAh g−1 after 200 cycles.

4 Carbon-Based Materials for Lithium Metal Anodes

Rapid developments have been made in recent years for sulfur cathodes, promoting the commercialization of Li–S batteries. In particular, lithium metal anodes, the primary cause limiting the widespread commercial application of Li–S batteries, have attracted increasing attention in an attempt to better understand lithium metal chemistry and pursue better performing metallic lithium anodes [207]. Deficiencies that hinder the practical application of lithium metal anodes include the uncontrollable growth of lithium dendrites and the cracking of SEI layers on the anode, leading to serious safety concerns and decreased lithium metal battery Coulombic efficiencies and cycling life spans. Therefore, to improve lithium metal anode performances, the interface between lithium and electrolytes must be stabilized and the lithium stripping/plating process must be optimized to suppress the formation of lithium dendrites and promote the uniform distribution of lithium during cycling.

Currently, several promising protective methods to stabilize lithium metal anodes have been explored, such as the stabilization of interfaces [39], the construction of lithium hosts [208] and the fabrication of nucleation sites [209]. For example, stable SEI layers must satisfy the requirement of mechanical strength to restrain the growth of lithium dendrites, taking advantage of the mechanical properties of the material. Another factor in the preparation of high-performance lithium metal anodes is the homogeneous deposition of lithium onto anodes to suppress dendrite growth during cycling. This factor includes several effective methods such as the construction of lithium hosts with high specific surface areas to decrease local current densities and the enhancement of the affinity between lithium and substrates to lower nucleation barriers. Among all these strategies, carbon-based materials have exhibited multifunctionality and their application in the protection of lithium metal anodes will be discussed.

4.1 Stabilization of SEI Layers

The formation of SEI layers on the surface of lithium anodes can prevent the unwanted side reactions between lithium metal and electrolytes during the electrochemical reaction process. However, SEI layers are generally brittle and prone to cracking during cycling [207]. This cracking of the SEI layer can result in the growth of lithium dendrites and the constant consumption of electrolytes. As a result, this phenomenon leads to low Coulombic efficiencies and rapid capacity losses [210]. Therefore, the SEI layer must be stabilized to suppress the growth of lithium dendrites and prevent the consumption of electrolytes, ensuring high-performance batteries.

Optimal SEI layers on lithium metal anodes should possess mechanical flexibility, chemical stability and ionic conductivity [207]. Based on this, Zheng et al. [39] designed a monolayer coating of interconnected amorphous hollow carbon spheres on the surface of a metallic lithium anode to favor the formation of a stable SEI layer and facilitate the uniform deposition of lithium metal beneath the carbon layer (The morphology of the hollow carbon spheres is shown in Fig. 15.). As a result, a stable SEI layer formed on the top surface of the carbon spheres with the Young’s modulus of the amorphous carbon layer being ~ 200 GPa, suppressing the growth of lithium dendrites. In addition, the researchers reported that the amorphous carbon layer did not inhibit charge transfer during cycling and the weak contact between the carbon layer and the copper foil allowed lithium to deposit in the void spaces between the carbon layer and the copper current collector. And as a result of this designed structure for a working electrode, uncontrollable growth of lithium dendrites after 50 cycles at a current density of 1 mA cm−2 was suppressed and the electrode exhibited a high Coulombic efficiency of ~ 99% for more than 150 cycles at 0.25 mA cm−2.

Fig. 15
figure 15

a Cross-sectional SEM image of the hollow carbon nanospheres, b TEM image of the hollow carbon nanospheres, c SEM image of the hollow carbon nanosphere thin film peeling off the Cu substrate. Reproduced with permission from Ref. [39]

Aside from amorphous carbon layers, Yan et al. [40] in their study reported that graphene can also be utilized as modified layers for copper substrates and that similar to amorphous carbon, graphene is also chemically stable and possesses a high mechanical strength with a Young’s modulus approaching 1.0 TPa. In addition, the pore diameter of the graphene hexagonal ring is too small to allow electrolytes or lithium metals to diffuse through, but the point and line defects in the plane of the graphene can allow for the penetration of lithium ions which can be sealed by SEI layers to prevent the infiltration of electrolytes. And because of this, lithium deposits smoothly between graphene and the current collector and a stable SEI layer can be generated on the top of the graphene. Based on this, graphene/copper anodes can provide a high Coulombic efficiency of 93% after 50 cycles at a current density of 1.0 mA cm−2. Kim et al. [211] also reported that the controllable growth of lithium can be achieved by combining the advantages of multilayered graphene-modified lithium foil with electrolyte additives. Here, the researchers suggest that multilayered graphene can inhibit direct contact between SEI layers and lithium metal to improve the Coulombic efficiency of the battery, and that the addition of Cs ions can enhance the diffusion of lithium ions and suppress the growth of lithium dendrites. And as a result, multilayered graphene with Cs ion additives displays synergistic effects and improved cycling stabilities.

4.2 Carbon-Based Lithium Hosts

Carbon materials possess high specific surface areas and electrical conductivities that can decrease local current densities and reduce the possibility of dendrite formation [37, 209]. Based on this, Mukherjee et al. [212] synthesized a porous graphene network using a thermal shock process with graphene oxide paper and reported that, the defect sites in the graphene can act as seed points and that lithium tends to deposit near divacancies forming high lithium content intercalated states. And because of the increased density of defects, significant improvements in lithium capacity of the porous graphene network were observed. In addition, the lithium metal was found to be encapsulated in the pores of the porous graphene architecture and the growth of lithium dendrites was effectively suppressed.

In another study, Zhang et al. [208] not only utilized unstacked graphene frameworks as the hosts for the deposition of lithium metal, but also used a LiTFSI-LiFSI dual-salt electrolyte to form a stable SEI layer. (Schematic diagrams are shown in Fig. 16a.) In this study, the resulting unstacked graphene produced a high specific surface area up to 1666 m2 g−1 and an electrical conductivity for the graphene of 435 S cm−1, decreasing surface current densities and restraining dendrite growth. The addition of the LiTFSI–LiFSI dual-salt electrolyte was also reported to facilitate the formation of a stable and flexible SEI layer, resulting in an anode with a high Coulombic efficiency reaching 93% at 2 mA cm−2.

Fig. 16
figure 16

a Schematic illustration of the Li depositing/stripping process on graphene flakes. Reproduced with permission from Ref. [208]. b Synthesis process of the layered Li–rGO composite film, c cycling performance of the symmetric Li–rGO electrode (blue) and bare Li foil (red) in the first 250000 at 3 mA cm−2.Reproduced with permission from Ref. [46]

Lin et al. [46] also synthesized sparked rGO films and infiltrated molten lithium into them (Fig. 16b) to obtain anodes with good electrochemical performances. In this study, after the spark reaction between GO films and molten lithium, the GO films were reduced to rGO films, generating uniform nanogaps that can accommodate the deposition of lithium simultaneously. Here, the synergetic effects between the lithiophilicity of rGO and the capillary forces of the nanogaps facilitated the uniform distribution and deposition of lithium. In addition, the rGO films were reported to be able to accommodate volume changes during lithium stripping/plating and can play an important role in the formation of a stable artificial interface to stabilize the SEI layer. And as a result, the resulting anode exhibited good electrochemical performances with a capacity of ~ 3390 mAh g−1 and a high cycling stability at 3 mA cm−2 (Fig. 16c).

4.3 Lithiophilicity of Carbon-Based Composites

The strength of the affinity between lithium and substrates can lower nucleation barriers and achieve the uniform distribution and deposition of lithium [46]. Therefore, to enhance the lithiophilicity of carbon fiber networks, Liang et al. [209] modified carbon fibers with silicon shells using CVD methods to synthesize a lithium composite with high gravimetric capacity and high cycling stability. The researchers here reported that compared with naked carbon frameworks, their silicon-decorated carbon framework displayed good wettability to molten lithium, allowing lithium to be uniformly deposited onto the surface of carbon fiber networks, with the 3D conductive pathways guaranteeing the rapid transport of electrons. And as a result, the resulting lithium composite exhibited a high gravimetric capacity of 2000 mAh g−1 and a high cycling stability with a low lithium plating/stripping overpotential of less than 90 mV at 3 mA cm−2 after 80 cycles. In addition, Yang et al. [213] reported that carbon nanofibers decorated with ultrafine silver nanoparticles acting as nucleation sites of lithium metal can facilitate the uniform deposition of lithium onto the surface of 3D architectures as well. The researchers here reported that such modified 3D carbon networks can effectively prevent the growth of lithium dendrites and generate smooth lithium deposition layers on the anode with a low lithium plating/stripping overpotential of ~ 25 mV at 0.5 mA cm−2.

In addition, nitrogen-doped graphene is also a lithiophilic material with dispersed lithiophilic functional groups that can facilitate the nucleation of lithium and prevent the growth of dendrites [214]. And according to calculations, pyrrolic-N nitrogen and pyridinic-N possess large binding energies with lithium, suggesting that not only are these nitrogen groups lithiophilic, they are also electrochemically stable and will not react with lithium atoms. The interactions between quaternary nitrogen and lithium atoms are weak however, because quaternary nitrogen cannot provide extra lone-pair electrons to adsorb lithium. Based on these findings, nitrogen-doped graphene can suppress dendrite growth and exhibit long-term cycling stability and high Coulombic efficiency, achieving a Coulombic efficiency of 98% and remaining stable near 200 cycles at a current density of 1.0 mA cm−2.

5 Conclusion and Outlook

This review has summarized the recent advancements of carbon-based materials in Li–S batteries and focused on the electrochemical performance, cycling stability, Coulombic efficiency and safety concerns of electrodes. Specifically, several rational design factors for carbon-based sulfur cathodes were systematically discussed, including porosity and conductivity of carbon-based frameworks, polarity with functional groups and the affinity in carbon-based composites. In addition, efficient protective strategies for carbon-based lithium metal anodes were also discussed, such as stabilization of SEI layers, carbon-based lithium hosts and lithiophilicity of carbon-based composites. In terms of sulfur cathodes, the creative studies discussed in this review provide significant design concepts and attempts to solve related problems such as the poor conductivity of sulfur and lithium sulfide, the volume changes of the cathode and the shuttle effects of polysulfides. In general, the conductive characteristics of carbon-based materials are beneficial to overcome the insulating nature of sulfur and lithium sulfide; and rational cathode nanostructures should possess void spaces to encapsulate active materials and accommodate sulfur volume changes, and possess strong interactions with polysulfide intermediates. Therefore, sulfur cathodes must possess high surface areas, hierarchical porous structures and chemical adsorption sites for lithium polysulfides. In terms of the configuration of lithium metal anodes, studies have demonstrated that carbon-based materials with high specific surface areas, mechanical intensity and lithiophilicity can facilitate the formation of stable SEI layers and suppress dendrite growth. And based on these reasonable design factors to stabilize lithium metal anodes, resulting lithium metal anodes can achieve high Coulombic efficiencies and long cycling stabilities. Moreover, because of the complexity of the solid–liquid–solid transition process during the charge and discharge process in Li–S batteries, it is essential to comprehensively understand the mechanisms of the electrochemical reactions of Li–S batteries and the mechanism details need to be extensively explored.

Although the development of Li–S batteries has made certain progress recently, a significant progress is still required to fulfill the demands of practical application. Currently, increasing amounts of reported studies have paid more attention to the enhancement of sulfur loadings in Li–S batteries to achieve higher energy densities, but the large excess of electrolytes decreases energy densities in Li–S batteries, making it difficult for Li–S batteries to become competitive with state-of-the-art LIBs [6]. Therefore, to promote the commercialization of Li–S batteries, several reference standards must be met, including a sulfur content of > 75 wt%, a sulfur utilization percentage of > 70%, an areal sulfur loading of > 5 mg cm−2, a ratio of electrolyte to sulfur of < 3 and a lithium excess of < 100% [215]. However, the high sulfur content and high areal sulfur loading requirements of high-energy density Li–S batteries will lead to the serious destruction of lithium metal anodes at high current densities, resulting in Li–S battery failures and safety issues. Moreover, because of the high electrochemical reactivity of lithium metal, electrolytes are consumed continuously during cycling. Because of these factors, several effective protection methods must also be put in place to enhance Coulombic efficiencies, stabilize lithium metal anodes and overcome safety issues. This is because if safety issues of lithium metal anodes cannot be properly addressed, Li–S batteries cannot achieve commercialization.

In conclusion, the utilization of carbon-based materials with high conductivity, high specific surface area and controllable functionality both in the sulfur cathode and in the lithium anode can provide effective methods to solve interrelated problems of electrodes in Li–S batteries. And although the commercialization of Li–S batteries is still far off, the emergence of great quantities of outstanding studies has broadened the horizons of researchers. Therefore, to further promote the commercialization of Li–S batteries, continued intensive efforts need to be devoted to the understanding and resolution of scientific and technological problems in Li–S batteries, with the rational design of advanced electrodes being able to significantly enhance the electrochemical performance of Li–S batteries and fulfill the ultimate goal of commercialization.