3D graphene framework supported Li₂S coated with ultra-thin Al₂O₃ films: binder-free cathodes for high-performance lithium sulfur batteries

Abstract

Lithium sulfide (Li₂S) has drawn special attention as a promising cathode material for emerging energy storage systems due to its high theoretical specific capacity and great compatibility with lithium metal-free anodes. However, Li₂S cathodes urgently require a solution to increase their poor electrical conductivity and to suppress the dissolution of long-chain polysulfide (Li₂S_n, 4 ≤ n ≤ 8) species into electrolyte. To this end, we report a free-standing Al₂O₃–Li₂S–graphene oxide sponge (GS) composite cathode, in which ultrathin Al₂O₃ films are preferentially coated on Li₂S by an atomic layer deposition (ALD) technique. As a result, a combination of high electron conductivity (from GS) and strong binding with Li₂S_n (from ultrathin Al₂O₃ films) was designed for cathodes. The newly developed Al₂O₃–Li₂S–GS cathodes are able to deliver a highly reversible capacity of 736 mA h g_Li2S⁻¹ (427 mA h g_cathode⁻¹) at 0.2C, which is much higher than that of corresponding cathodes without Al₂O₃ (59%). Also, the long-term cycling stability of Al₂O₃–Li₂S–GS cathodes was demonstrated up to 300 cycles at 0.5C with an excellent capacity retention of 88%. In addition, combined with density functional theory calculations, the promotional mechanism of ultrathin Al₂O₃ films was elucidated using extensive characterization. The ultra-thin Al₂O₃ film with optimal thickness not only acts as a physical barrier to Li₂S nanoparticles, but provides a strong binding interaction to suppress Li₂S_n species dissolution.

---

Introduction

The rapid development of portable and flexible electronic devices creates an increasing demand for an energy storage system with larger energy density, improved safety and lighter weight.^1,2 Rechargeable lithium/sulfur (Li/S) batteries are considered one of the most promising candidates due to their high theoretical energy density of 2580 W h kg⁻¹ and specific capacity of 1672 mA h g⁻¹, as well as the low cost and natural abundance of elemental sulfur.^3–5 However, sulfur cathodes need to be paired with lithium metal anodes, which impose serious safety issues owing to the possible formation of lithium dendrites.^6,7 To address this concern, many efforts have been made to search for alternative cathode materials.^1,8 Among these efforts, the fully lithiated Li₂S was selected because of its compatibility to pair with other attractive lithium metal-free anodes (such as silicon,^9–11 tin,^12,13 and graphite^14,15) and its high theoretical specific capacity of 1166 mA h g⁻¹.¹⁶ In addition, Li₂S could significantly minimize the structural damage of the cathode caused by the volume expansion of sulfur during the lithiation process because it occupies the maximum volume possible compared to octatomic sulfur.¹⁷ The higher melting point of Li₂S (∼938 °C) than sulfur also makes it possible to be modified by more treatments (such as chemical vapour deposition and atomic layer deposition (ALD)) without any loss of active materials.^18,19 Nonetheless, like its sulfur counterparts, the Li₂S cathode still suffers from poor electronic conductivity, the dissolution of Li₂S_n (4 ≤ n ≤ 8) species and its shuttle effect during charge/discharge cycling, which lead to inferior active material utilization and low cycling performance and coulombic efficiency.^20–23

In many strategies developed to circumvent the above issues, scientists have introduced various conductive materials, such as carbon-based materials,^24–33 polypyrrole³⁴ and transition metal disulfides.³⁵ Carbon-based materials with good electronic conductivity are commonly used to encapsulate and thus physically entrap Li₂S_n species; however the non-polarity of carbon-based materials leads to weak interaction with Li₂S_n species, reducing their ability to bind and confine these species during cycling. By contrast, polar Li₂S_n species can strongly bind with polar inorganic materials, such as TiS₂ and VS₂, to effectively suppress their dissolution, but electronic conductivity is not improved as much as with carbon-based materials.³⁵ Thus, there is an urgent need for a new class of encapsulation materials for the Li₂S cathode that can provide a good electronic pathway and possess strong chemical interaction with Li₂S_n species, which are the merits of carbon-based materials and polar inorganic materials.

The ALD technique is a promising approach for large-scale preparation of high-performance Li₂S cathodes by combining carbon-based materials and inorganic species because of its intrinsic versatility to form exceptionally conformal coatings with atomic-scale precision. It could be more successful to modify the Li₂S cathode owing to its high stability during the ALD process instead of volatilizing severely like sulfur.^36–39 Furthermore, to our best knowledge, to date there aren't any studies about improving Li₂S cathodes by using an ALD technique.

Herein, a free-standing Al₂O₃–Li₂S–graphene oxide sponge (Al₂O₃–Li₂S–GS) cathode with Al₂O₃ preferentially being coated on the Li₂S particles was fabricated by thermal treatment and an ALD technique. Li₂S–GS composites were in situ synthesized by the reaction of Li₂SO₄ and GS. During the ALD process, ultrathin Al₂O₃ films (∼1 nm) tend to grow on the surface of Li₂S particles rather than graphene sheets, which was verified by first-principles calculations, which is never studied in previous reports for sulfur cathodes improved by using an ALD technique. The graphene sheets without an Al₂O₃ coating endow Al₂O₃–Li₂S–GS composites with high electronic conductivity. In addition, the porous and flexible GS framework enables the composites to be used as cathodes directly without a binder and current collector, leading to easily prepared lightweight cathodes. To investigate the role of Al₂O₃ in cathodes, the discharge mechanism of Al₂O₃–Li₂S–GS cathodes is studied. Besides acting as a physical barrier to entrap Li₂S_n species, the Al₂O₃ coating on the surface of Li₂S particles also exhibits a binding interaction with Li₂S_n. The Li–O binding interaction will suppress the dissolution of long-chain polysulfide generated during the upper discharge plateau and promote the capacity release during the lower plateau, resulting in a well maintained capacity of the cathodes. Thus, the Al₂O₃–Li₂S–GS cathodes are able to achieve a competitive capacity and an excellent cycling stability over Li₂S–GS cathodes.

Experimental

Synthetic procedures

Preparation of Li₂S–GS composites. The graphene sponge (GS) was prepared by the hydrothermal reduction of graphene oxide (GO),⁴⁰ followed by vacuum freeze-drying and cutting into circular pellets. The Li₂SO₄–GS composites were prepared by vacuum impregnation of the GS pellets in Li₂SO₄ aqueous solution, followed by a quick-freezing technique via using liquid nitrogen, and then vacuum freeze-drying. The as-prepared composites were sealed in a quartz tube and annealed at 150 °C for 1 h to remove the crystal water, and then heated to 800 °C with a rate of 10 °C min⁻¹ and held at that temperature for 3 h under an argon (Ar) atmosphere with a flow rate of 200 sccm throughout the process. After cooling to room temperature (RT) naturally, the obtained Li₂S–GS pellets were directly transferred to a glove-box filled with Ar gas or a chamber of the ALD instrument. The pellets can be directly used as free-standing cathodes without the addition of a conductive additive, binder and current collector. The Li₂S content in the composites was investigated by utilizing the dissolution of Li₂S in anhydrous methanol, the Li₂S–GS (Al₂O₃–Li₂S–GS) composites were washed with methanol to remove Li₂S in a glove-box.⁴¹ The remaining GS (Al₂O₃–GS) were dried and weighed.

Thermal ALD of Al₂O₃ on Li₂S–GS composites. The ultrathin Al₂O₃ films were deposited on Li₂S–GS composites by thermal ALD (TALD-150 Kemicro Jiaxing) with different ALD–Al₂O₃ cycles (5, 10 or 30). Trimethyl aluminum (TMAl, Al(CH₃)₃) and ozone (O₃) (10 wt% in oxygen) were used as the precursors of the Al and O elements, respectively, which were placed in the chamber and purged with N₂ gas under the operation conditions of 120 °C and 2.0 Torr. The pulse sequence of Al₂O₃ was TMAl–N₂ purge–O₃–N₂ purge, corresponding to the dose time and the purge time was 0.02–40 to 0.5–40 s. The growth thickness per ALD cycle of Al₂O₃ was ∼0.1 nm per cycle.

Materials characterization

The specific surface area and electronic conductivity of GS were measured using the TriStar II 3020 and HALL 8800 measurement system, respectively. The morphologies of the composites were characterized by field-emission scanning electron microscopy (FESEM, Quanta 200FEG) with elemental mapping using energy-dispersive X-ray spectroscopy (EDX) and transmission electron microscopy (TEM, Tecnai G² F30) with elemental mapping using energy filtered TEM. X-ray photoelectron spectroscopy (XPS) was carried out by using an ESCALAB 250Xi with an Al Kα X-ray source. For SEM, TEM, and XPS, the samples were prepared and sealed in aluminum foil/poly-bags in an Ar-filled glovebox before being transferred into the chamber. X-ray diffraction (XRD) patterns were measured using a D/max-2000 with Cu Kα irradiation (λ = 1.54056 Å). Before testing, the samples were sealed using Kapton tape on a silicon wafer. Raman spectroscopy was performed with a DXRSMART Raman spectrometer using an excitation laser of 532 nm after the samples were sealed in a glass holder in an Ar-filled glovebox. UV-visible absorption spectroscopy was carried out by using a UV-8000 s with sealable quartz cuvettes (1400 μL, Hellma).

Electrochemical characterization

The Al₂O₃–Li₂S–GS and Li₂S–GS composites can be used as cathodes directly, and the mass loading of Li₂S in both cathodes was between 1.2 and 1.5 mg cm⁻². The electrolyte was 1.0 M lithium bis(trifluoromethane) sulfonamide (LiTFSI) dissolved in a mixture of 1,3-dioxolane (DOL) and 1,2-dimethoxymethane (DME) (1 : 1, v/v) and 1 wt% LiNO₃ as an electrolyte additive. CR2032-type coin cells were assembled with lithium metal foil as the anode and a Celgard 2400 as the separator in an Ar-filled glovebox (<0.5 ppm H₂O and O₂). The volume of the electrolyte in cells is 1 mg : 30 μL (mass of Li₂S : volume of the electrolyte). The electrochemical properties of the cathodes were investigated by galvanostatic measurement in a potential range of 1.8–2.6 V (vs. Li⁺/Li) with a Land CT2001A battery test system at RT, after being charged to 3.8 V at 0.05C (1C = 1166 mA g⁻¹). Cyclic voltammetry (CV) was conducted from open-circuit voltage to 3.8 V (vs. Li⁺/Li) at a scanning rate of 0.05 mV s⁻¹. Electrochemical impedance spectrum (EIS) measurements were measured on an electrochemical workstation (PARSTAT 4000) in a frequency range of 100 kHz to 0.01 Hz with a potentiostatic signal amplitude of 5 mV at room temperature.

Results and discussion

Materials synthesis and characterization

Unlike most of the Li₂S cathodes employing commercial bulk Li₂S,^30,42 and organic lithium reagents^43–46 as the starting materials, the lower cost, safer and more environmentally friendly GS and Li₂SO₄ were utilized as the raw materials to in situ form Li₂S in this work. The Al₂O₃–Li₂S–GS composites were synthesized as shown schematically in Fig. 1 (see the Experimental section for details). In brief, Li₂SO₄·H₂O–GS composites were initially fabricated via the vacuum infusion of GS (Fig. 1a) circular pellets in Li₂SO₄ solution followed by rapid freezing and vacuum freeze-drying methods (Fig. 1b). The as-synthesized samples were heated at 800 °C in an argon atmosphere to allow the in situ formation of Li₂S–GS (Fig. 1c).^41,47–49 Afterwards, ultrathin Al₂O₃ films are grown by using the ALD technique because of its intrinsic versatility to form exceptionally conformal coatings with atomic-scale precision. To predict the growth location of Al₂O₃, the first-principles calculations were performed to evaluate the heat of formation of graphene–O and Li₂S–O due to the growth mechanism of decomposition of O₃ bond with the substrate first and then react with trimethyl aluminum to form Al₂O₃ (ref. 50). The heat of formation of Li₂S–O was two times lower than that of graphene–O (ΔH = −51.13 kcal mol⁻¹) (Fig. 1d), indicating the preferential adsorption of oxygen radicals on the surface of Li₂S particles rather than graphene (Fig. 1e) leading to the preferential growth of Al₂O₃ on the surface of Li₂S particles. Meanwhile, it is commonly accepted that the oxygen radicals can be more easily absorbed by the polar Li₂S than the non-polar carbon of graphene.

Fig. 1 Schematic of Al₂O₃–Li₂S–GS composite synthesis: (a–c and f) steps involved in the synthesis process; (d) heat of formation of graphene–O and Li₂S–O calculated by the first-principles calculations; (e) schematic during Al₂O₃-ALD cycles.

As a result, Li₂S particles of Li₂S–GS composites can be coated preferentially by Al₂O₃ (Fig. 1f). The thickness of Al₂O₃ films is able to be controlled facilely by the ALD–Al₂O₃ cycles. This simple and controllable method is a promising approach for large-scale preparation of high-performance Li₂S cathodes. The GS framework with a well-defined and interconnected 3D porous network (Fig. S1†) is favorable for not only the infiltration of the electrolyte to accelerate the Li⁺ transport, but also the accommodation of the volume change of Li₂S during charge/discharge cycling. Additionally, the superior robustness of the GS framework endows the composites with the ability to act as the free-standing cathode. The uniform dispersion of Li₂SO₄·H₂O on the surface of graphene sheets (Fig. S2†), obtained by rapidly freezing the composite with liquid nitrogen and vacuum freeze drying, facilitates its complete reaction with carbon to generate Li₂S.

X-ray diffraction (XRD) pattern of Li₂S–GS composites (Fig. 2a) shows the characteristic diffraction peaks of Li₂S (JCPDS no. 26-1188). An XRD peak at 2θ = 25.7° is obviously detected, corresponding to the diffraction peak of the remaining reduced graphene oxide (rGO) sheets of the GS framework.⁴⁰ A peak at 2θ = 22° arises from the Kapton tape, which is used to protect the samples from moisture during the testing procedure (Fig. S3a†). The Raman spectrum of Li₂S–GS composites (Fig. 2b) presents the characteristic T_2g peak of Li₂S, the typical D-band and G-band peaks of rGO and the peaks of the glass holder (Fig. S3b†) used to protect Li₂S from moisture during Raman characterization.^34,35 The morphology analysis clearly shows that the Li₂S–GS composites maintained the 3D structure of the GS framework (Fig. S4†). Compared to Li₂SO₄·H₂O–GS composites, the distribution of Li₂S particles on graphene sheets is uneven (Fig. 2c), owing to volume shrinking when Li₂SO₄·H₂O converts into Li₂S. Elemental mapping from energy-dispersive X-ray spectroscopy (EDX) of Li₂S–GS composites (Fig. S5†) confirms the presence of carbon, sulfur and oxygen. The oxygen spectrum appears because of the oxygen of GS and the oxidation of Li₂S during transferring samples from the poly/foil bags to the scanning electron microscope (SEM) chamber. Transmission electron microscopy (TEM) of Li₂S–GS composites (Fig. 2d and e) and elemental mapping by TEM (Fig. S6†) are performed to verify the Li₂S particle distribution on the surface of graphene sheets. The high resolution TEM (HRTEM) image of the Li₂S in Li₂S–GS composites (inset in Fig. 2e) shows the inter-planar spacing of 0.33 nm coinciding with that of the (111) plane of Li₂S.

Fig. 2 Characterization of Li₂S–GS and Al₂O₃–Li₂S–GS composites: (a) XRD patterns of samples; (b) Raman spectra of Li₂S–GS and Al₂O₃–Li₂S–GS composites (the inset showing T_2g phonon mode of Li₂S together with their fitted peaks); (c) SEM image and (d, e) TEM image of Li₂S–GS composites (the inset showing the HRTEM image of the Li₂S in Li₂S–GS composites); (f) SEM image and (g) TEM image of Al₂O₃–Li₂S–GS composites; (h) TEM image of the remaining Al₂O₃–GS hybrids by removal of Li₂S from the Al₂O₃–Li₂S–GS composites. Scale bars: 5 μm in (c, f), 200 nm in (d, g), 10 nm in (e) and 5 nm in inset, and 100 nm in (h).

After ultrathin Al₂O₃ film coating, the phase composition of the as-prepared composites was characterized by XRD. As shown in Fig. 2a, the XRD peaks of Al₂O₃ are not observed due to the amorphous nature of Al₂O₃ by using an ALD technique.³⁷ Furthermore, there are no redundant XRD peaks, revealing no side reaction during Al₂O₃ growth by ALD. The Li₂S content of Al₂O₃–Li₂S–GS composites showed little change (∼58 wt%), indicating a small percentage of Al₂O₃ in the composites. The SEM and TEM images of Al₂O₃–Li₂S–GS composites (Fig. 2f and g) show the subtle structural alteration relative to Li₂S–GS composites. From the elemental mapping of Al₂O₃–Li₂S–GS composites (Fig. S7†), it is interesting to observe that Al₂O₃ exhibits a non-uniform distribution in composites. To clearly detect Al₂O₃, Al₂O₃–GS hybrids were obtained by the removal of Li₂S from Al₂O₃–Li₂S–GS composites. The ultrathin Al₂O₃ films are not visibly observed in the TEM image of Al₂O₃–GS hybrids (Fig. 2h) because of the ultrathin thickness (∼1 nm) and the amorphous nature. The robustness and elasticity of Al₂O₃–Li₂S–GS cathodes are also investigated, which can be compressed-fit into a small enclosure and then recovers almost instantly to their original shape after release (Fig. S8a†). In addition, the thickness of the cathode film is 30–40 μm, which was directly prepared from the Al₂O₃–Li₂S–GS cathode. The SEM images of the cross section of the Al₂O₃–Li₂S–GS electrode were newly recorded, as shown in Fig. S8b.†

Aiming to investigate the distribution of Al₂O₃, the EDX line profiles (Fig. 3a) of Al₂O₃–Li₂S–GS composites and elemental mapping (Fig. 3b) of Al₂O₃–GS hybrids strongly demonstrate that Al₂O₃ tends to grow on the surface of Li₂S particles. The Al line profile and elemental mapping illustrate a regional distribution implying the preferential growth of Al₂O₃ on Li₂S particles. This indicates that Al₂O₃–Li₂S–GS composites possess a high electronic conductivity stemming from the graphene sheets without an Al₂O₃ coating. The remaining Al₂O₃–GS hybrids achieve a competitive conductivity of ∼160 S cm⁻¹ compared to the remaining GS framework (∼162 S cm⁻¹). The ultrathin Al₂O₃ film coating on Li₂S–GS composites was also confirmed by Raman spectroscopy (Fig. 2b). For Al₂O₃–Li₂S–GS composites, the peak for Al₂O₃ has not been detected due to its non-response to Raman spectroscopy⁵¹ The intensity of the Li₂S T_2g Raman peak is much weaker compared to that of the rGO peak, indicating that the Li₂S particles are coated with Al₂O₃ films because of the surface-sensitive nature of Raman spectroscopy. By fitting Li₂S T_2g Raman peak (Fig. 3c), for Li₂S–GS composites, the peak occurs at 372 cm⁻¹, which agrees with the value in previous studies.^34,35,52 However, the T_2g peak position of Al₂O₃–Li₂S–GS composites shifts towards a lower wave number to 367 cm⁻¹. The decreasing frequency of the T_2g peak hints a decrease in the force constant of Li–S bonds because of the T_2g peak induced by the Li–S bond vibrations. To characterize the chemical interaction between Al₂O₃ and Li₂S, X-ray photoelectron spectroscopy (XPS) measurement (Fig. 3d and S9†) was carried out. Compared to Li₂S–GS composites, the Li 1 s spectrum of Al₂O₃–Li₂S–GS composites shows slightly asymmetric broadening in the higher energy and an additive fitted peak with a higher binding energy of 55.4 eV arises (Fig. 3d), owing to Li–O interactions between Al₂O₃ and Li₂S.^34,53,54 The results of XPS and Raman analyses suggest that the strong polarity of oxygen radicals during ALD–Al₂O₃ cycles interacted with Li atoms in Li₂S, verified by the first-principles calculations. Therefore, besides serving as a physical barrier to entrap Li₂S_n, the ultrathin Al₂O₃ films preferentially coated on Li₂S particles hold strong chemical interaction with Li₂S_n to limit their dissolution as well. The preferential growth of ultrathin Al₂O₃ films and the binding interaction between Al₂O₃ and Li₂S are favourable for improving the performance of cells since ultrathin Al₂O₃ films provide not only maintained conductivity which ensures the electron transfer pathway, but also a physical barrier and chemical interaction for suppressing the long-chain Li₂S_n species dissolution, enabling excellent cycling stability.

Fig. 3 Investigation of Al₂O₃ distribution and interaction between Li₂S and Al₂O₃ in Al₂O₃–Li₂S–GS composites: (a) EDX line profiles of Al₂O₃–Li₂S–GS composites; (b) elemental mappings of Al₂O₃–GS hybrids; (c) Raman spectra of the T_2g phonon mode and (d) XPS spectra of the Li 1s peak and of Li₂S, together with their fitted peaks. Scale bars: 5 μm in (a) and 100 nm in (b).

Electrochemical performance

To determine the role of ultrathin Al₂O₃ films in promoting the electrochemical performance, both Li₂S–GS and Al₂O₃–Li₂S–GS electrodes were used as free-standing cathodes directly without a binder and collector to assemble cells with Li foil as the anode. The mass loading of Li₂S is 1.2–1.5 mg cm⁻². The cells were charged to 3.8 V at 0.05C (1C = 1166 mA g⁻¹) in order to delithiate the Li₂S, and then were cycled in the voltage range of 1.8–2.6 V. The discharge cutoff voltage was set to 1.8 V in order to avoid the irreversible capacity of the LiNO₃ additive.⁵⁵ The capacity derived from graphene and Al₂O₃ was negligible due to their inactive nature for the electrode in this voltage window.

Electrochemical impedance spectroscopy (EIS) is employed to analyse the kinetics of the electrochemical reaction in our cell system. The semicircle in impedance spectra (Fig. 4a) of the as-assembled cells at the open-circuit potential corresponds to the charge transfer resistance (R_ct), which is associated with the apparent energies (E_a) of the electrochemical reactions (E_a ∝ ln R_ct, details seen in the ESI†).^19,56 The R_ct value of Al₂O₃–Li₂S–GS is 101.8 Ω, close to that of Li₂S–GS (90.4 Ω), indicating that ultrathin Al₂O₃ films have little impact on the kinetics of the electrochemical reaction before cycling. The first charge voltage profiles of the as-assembled cells charged to 3.8 V at 0.05C to activate Li₂S (Fig. 4b)^16,57–59 exhibit the close potential barrier at 3.5–3.55 V (inset in Fig. 4b), and a long sloping plateau at 3.4–3.6 V corresponding to the anodic peak at ∼3.5 V in cyclic voltammetry (CV) tests (Fig. S10†). The EIS spectra, first charge voltage profiles and initial charge CV curves all demonstrate the negligible impact of ultrathin Al₂O₃ films on electron and Li⁺ ion transmission in electrochemical reactions in this work.

Fig. 4 Electrochemical performance of Al₂O₃–Li₂S–GS cathodes compared to Li₂S–GS cathodes: (a) EIS spectra at open circuit voltage; (b) initial charge voltage profiles, (c) specific capacities with a voltage range of 1.8–2.6 V at 0.2C and (d) a typical discharge/charge profile showing the various DCVs and CCVs (points 1–9) and the corresponding percentage of sulfur in the electrolyte relative to the total sulfur in cathodes.

After the initial charging, the as-assembled cells were cycled by galvanostatic measurement in the voltage window of 1.8–2.6 V at 0.2C. As shown in Fig. 4c, the Al₂O₃–Li₂S–GS cathode exhibits a high initial specific capacity of 866 mA h g_Li2S⁻¹. The capacity at the 50^th, 100^th and 150^th cycles can achieve 752, 762, and 736 mA h g_Li2S⁻¹ corresponding to the capacity retentions of 87%, 88% and 85%, respectively. By contrast, the Li₂S–GS cathodes show a comparable initial specific capacity (824 mA h g_Li2S⁻¹), but much faster capacity decay at 0.2C (59% after 150 cycles). Thus, ultrathin Al₂O₃ films can suppress Li₂S_n species dissolution, contributing to the superior cycling stability.

To prove the confining effect of ultrathin Al₂O₃ films, the sulfur content in the electrolyte was measured by using inductively coupled plasma-optical emission spectroscopy (ICP-OES, see the ESI† for details). The cells were cycled at 0.2C and disassembled at various charge/discharge states during the 20^th cycle to measure the sulfur content in the electrolyte. As shown in Fig. 4d, points 1–9 represent the different discharge cutoff voltages (DCVs) and charge cutoff voltages (CCVs) during the 20^th cycle. Points 2 and 4 refer to 2.30 V DCV and 2.05 V DCV, corresponding to the two potential plateaus. Points 5 and 9 refer to 1.8 V DCV and 2.6 V CCV, corresponding to the maximum discharge and charge capacity. From the ICP-OES results (Fig. 4d) of Al₂O₃–Li₂S–GS cathodes, the sulfur percentages in the electrolyte are greatly decreased compared with Li₂S–GS cathodes, corresponding to 17%, 13%, 14% and 14% at points 2, 4, 6 and 8, respectively. The dissolution of polysulfide in the electrolyte is markedly alleviated owing to the presence of ultrathin Al₂O₃ films. Therefore, Al₂O₃–Li₂S–GS cells are able to achieve a more excellent stability.

In addition, the structure of the electrode and the kinetics of electrochemical reactions are well maintained attributed to the restriction effect of ultrathin Al₂O₃ films on the dissolution of polysulfide. The morphology of the electrodes after the 150^th discharge process was investigated, which were washed with dimethyl carbonate to remove the soluble species and dried naturally in an Ar-filled glovebox. In the case of Li₂S–GS cathodes, a thick film of insoluble lithium sulfides is clearly observed on their surface, as well as the increasing size of active material particles (Fig. 5a). Compared with the original Li foil (Fig. S11†), the surface of Li anodes from Li₂S–GS cells (Fig. 5b) exhibits a large number of lumps including the insoluble sulfur species and Li dendrites. The thick film and lumps, which are generated from the uncontrolled dissolution of polysulfide in the electrolyte, block the diffusivity of Li⁺ ions and passivate the surface of electrodes, leading to a large polarization and a low utilization of active materials.^36,37 However, the Al₂O₃–Li₂S–GS cathodes nearly preserve their original morphology without any obvious thick film on their surface (Fig. 5c). The surface of the Li anode is also smoother, which is ascribed to the effective confinement of polysulfide dissolution by ultrathin Al₂O₃ films (Fig. 5d), resulting in the high utilization of active materials. The morphology characterization of cathodes after repeated cycling strongly correlates with the results of EIS measurement (Fig. S12†). Compared with the results before cycling, Al₂O₃–Li₂S–GS cells show a lower R_ct value than that of Li₂S–GS cells after the 150^th discharge test at 0.2C indicating their better charge-transfer kinetics after 150 cycles. These results further confirm that the ultrathin Al₂O₃ films coated on Li₂S particles are able to greatly suppress the dissolution of Li₂S_n (4 ≤ n ≤ 8) species.

Fig. 5 Morphology of the electrodes after 150 cycles at discharge state: (a) cathode and (b) lithium anode from the Li₂S–GS cell; (c) cathode and (d) lithium anode from the Al₂O₃–Li₂S–GS cell. Scale bars, 10 μm in (a–d).

To further evaluate the electrochemical performance of Al₂O₃–Li₂S–GS cathodes, their long-term cycling properties and rate capability were measured, as shown in Fig. 6. The Al₂O₃–Li₂S–GS cathodes exhibited a stable cycling performance at 0.5C over 300 cycles (Fig. 6a), e.g. a high initial capacity of 728 mA h g_Li2S⁻¹ (∼1044 mA h g_sulfur⁻¹, 422 mA h g_cathode⁻¹), a reversible capacity of 643 mA h g_Li2S⁻¹ and a capacity retention of 88% after 300 cycles. The results are comparable to the work of Wang's group,¹⁹ which reported a capacity of 692 mA h g_Li2S⁻¹ after 145 cycles at 0.5C of nano-Li₂S/rGO paper (652 mA h g_Li2S⁻¹ in our work) under the similar conditions of active material loading; however, we achieved a capacity retention of ∼90%, much higher than 77.1% of nano-Li₂S/rGO paper. Due to their free-standing characteristics without a binder and collector, the Al₂O₃–Li₂S–GS cathodes achieve a high capacity of 373 mA h g_cathode⁻¹ based on the mass of the cathode after 300 cycles at 0.5C. The average coulombic efficiency of Al₂O₃–Li₂S–GS cathodes is ∼99% during cycling (Fig. 6a).

Fig. 6 Electrochemical performance of Al₂O₃–Li₂S–GS cathodes: (a) specific capacities based on Li₂S and cathode, respectively, coulombic efficiency during 300 cycles at 0.5C; (b) specific capacities and coulombic efficiency cycled from 0.05C to 6C; (c) charge/discharge profiles cycled from 0.05C to 6C and (d) specific capacity and coulombic efficiency upon prolonged cycling (1000 cycles) at 2C after activating at 0.05C.

Importantly, for evaluating the rate capability and electrochemical kinetics of Al₂O₃–Li₂S–GS cathodes, the cells were cycled at different C-rates from 0.05C to 6C and finally cycled at 0.2C (Fig. 6b). The Al₂O₃–Li₂S–GS cathodes deliver stable capacities of 889, 810, 748, 703, 635, 589, 573 and 546 mA h g_Li2S⁻¹ as the C-rates increase from 0.1C to 5C, respectively. Even at a high C-rate of 6C, the Al₂O₃–Li₂S–GS cathodes can attain a discharge specific capacity of ∼526 mA h g_Li2S⁻¹, delivering a competitive value compared with recently reported studies about Li₂S cathodes (Fig. S13†). The stable coulombic efficiency is about ∼99.5% at various C-rates. After testing at a high C-rate, the Al₂O₃–Li₂S–GS cathodes can still achieve a capacity of 778 mA h g_Li2S⁻¹ when the C-rate directly returns to 0.2C, demonstrating the excellent stability of cathodes. The discharge curve of Al₂O₃–Li₂S–GS cathodes still shows two plateaus at 2.25 V and 1.85 V even at 6C (Fig. 6c).

We attribute the good rate performance of Al₂O₃–Li₂S–GS cathodes to the following three reasons: first, the 3D interconnected porous structures of Al₂O₃–Li₂S–GS cathodes provide a fast Li⁺ transport channel. Second, due to the preferential growth of Al₂O₃ ultrathin films on Li₂S, rather than rGO, the high electrical conductivity of rGO sheets can be maintained, thereby facilitating electron transfer during the charging and discharging reactions. Third, the ultrathin Al₂O₃ films coated onto Li₂S greatly limits the dissolution of polysulfide into the electrolyte, which is able to mitigate the possible loss of active materials, enhance the reversible capacity, and maintain the structural integrity of Al₂O₃–Li₂S–GS cathodes. Therefore, considering the fast electron/ion transfer pathway, significantly suppressed polysulfide dissolution, and robust structural integrity of cathodes, the studied Al₂O₃–Li₂S–GS cathodes are capable of providing excellent rate performance. The long-term cycling performance of Al₂O₃–Li₂S–GS cathodes was tested at 2C after activating at 0.05C (Fig. 6d), it can exhibit a high initial capacity of 1028 mA h g_Li2S⁻¹ at 0.05C and 668 mA h g_Li2S⁻¹ at 2C. The reversible capacity at the end of the 100^th, 500^th and 1000^th cycles delivers 600, 509, and 438 mA h g_Li2S⁻¹ with a low average capacity decay rate of 0.028% per cycle and a high average coulombic efficiency of ∼99%.

We have presented the electrochemical properties of free-standing Al₂O₃–Li₂S–GS cathodes for Li/Li₂S cells with 58 wt% Li₂S and 10 ALD–Al₂O₃ cycles. For practical applications, the high loading of active materials with high electrochemical performance is desirable. Therefore, we altered the Li₂S content and the thickness of Al₂O₃ films. Three different Li₂S contents of Li₂S–GS composites were synthesized and their morphology and cycling performance at 0.2C were investigated (Fig. S14†). Meanwhile, we investigated three different ALD–Al₂O₃ cycles: 5, 10 and 30 cycles; the morphology and cycling performance (at 0.2C) of the as-prepared composites are shown in Fig. S15.† Thus, these experiments illustrate our choice of 58 wt% Li₂S and 10 ALD–Al₂O₃ cycles as the optimal composite structure.

Mechanism study

In order to understand the discharge mechanism of Al₂O₃–Li₂S–GS cells and the role of Al₂O₃ during the charge/discharge process, we tracked the different states of the discharge process with the help of ex situ characterization. First of all, the charge/discharge voltage profiles (0.2C) of Al₂O₃–Li₂S–GS and Li₂S–GS cells were collected, as shown in Fig. S16.† The Li₂S–GS cathodes exhibit two discharge plateaus corresponding to a two-step electrochemical process: ∼2.3 V and ∼2.05 V related to the formation of Li₂S_n (4 ≤ n ≤ 8) and the conversion from Li₂S_n to Li₂S₂ (or Li₂S), respectively. However, for Al₂O₃–Li₂S–GS cells, a supernumerary discharge plateau arises at ∼2.1 V, indicating a multistep sulfur reaction process. Aiming at a better understanding of the sulfur electrochemical reduction occurring in the Al₂O₃–Li₂S–GS cells, the intermediate states of the electrolyte and Al₂O₃–Li₂S–GS cathodes were investigated by ex situ measurements such as UV-vis spectroscopy,^60–62 Raman spectroscopy,^63–65 and XRD^66–68 (see the ESI† for details).⁶⁹ The corresponding results are shown in Fig. 7 and S17–S19.†

Fig. 7 Ex situ measurements of Al₂O₃–Li₂S–GS cathodes at (a) selected voltage states (3.8, 2.3, 2.1, 2.05 and 1.8 V); (b) UV-vis spectra, (c) Raman spectra and (d) XRD pattern.

It is commonly accepted that the strong UV-Vis absorption bands at 603, 538, 458 and 407 nm correspond to S₃˙⁻, S₈²⁻, S₆²⁻ and S₄²⁻, respectively.⁵⁹ It should be pointed out that S₃˙⁻ can be observed even at the end of discharge because of the disproportionation of polysulfide (S_n²⁻ = [(2n − 2)/5]S₃˙⁻ + [(6 − n)/5]S²⁻). From the UV-Vis spectra (Fig. 7b), it is evident that the chain of polysulfide in the electrolyte shortens gradually along with the increasing discharge depth, which is observed at the Al₂O₃–Li₂S–GS cathodes as well, according to the results of Raman (Fig. 7c) and XRD (Fig. 7d) measurements. Meanwhile, crystalline sulfur and Li₂S in the cathodes are formed at the end states of the charge and discharge processes, respectively. According to the measurements above, the probable and compendious discharge reaction mechanism (Fig. S20†) for Al₂O₃–Li₂S–GS cells was proposed as follows:

At the upper discharge plateau (2.3 V),

S₈ + 2Li⁺ + 2e⁻ → Li₂S₈

Li₂S₈ → Li₂S₆ + ¼S₈

At the middle discharge plateau (2.1 V),

2Li₂S₆ + 2Li⁺ + 2e⁻ → 3Li₂S₄

At the lower discharge plateau (1.8 V),

3Li₂S₄ + 2Li⁺ + 2e⁻ → 4Li₂S₃

2Li₂S₃ + 2Li⁺ + 2e⁻ → 3Li₂S₂

Li₂S₂ + 2Li⁺ + 2e⁻ → 2Li₂S

In terms of the additional discharge plateau, we propose that the primary reason for such a plateau stems from the interaction between Al₂O₃ and intermediate Li₂S_n (4 ≤ n ≤ 8) species. For this purpose, XPS measurement was conducted to study the Al₂O₃–Li₂S–GS cathodes at different charge/discharge states, as shown in Fig. 8 and S21–S23.† The fluorine and nitrogen elements stem from the electrolyte salt (LiTFSI) and additive (LiNO₃), respectively. The XPS activity over 168 eV is detected, which is induced by the electrolyte (LiTFSI) because of no washing for the studied cathodes (Fig. S20 and S21†). It is interesting to note that the binding energy of S 2p peaks is decreasing with the increasing depth of the discharge process (from 3.8 to 1.8 V, as shown in Fig. 8a–e). The main reason for this phenomenon is the decreasing amount of the S–S component and the increasing Li–S component in polysulfide as the polysulfide chain shortens with continuous discharge. Additionally, the S 2p XPS peaks of pristine sulfur and Li₂S are detected at 3.8 V and 1.8 V, respectively, revealing the formation of sulfur and Li₂S at the end of the charge and discharge processes and agreeing well with the results of ex situ measurements in the former text. In order to investigate the interactions between Li₂S_n species and Al₂O₃, the Li 1s XPS peaks of cathodes at different intermediate states are fitted (Fig. 8f and S23†). It is unsurprising that the Li–O binding stably exists throughout the charge/discharge process, indicating the persistent effect of Al₂O₃ on suppressing Li₂S_n species dissolution. The similar argument is demonstrated in a previous study that anchoring materials (Al₂O₃ in our work) could induce strong binding interactions with intermediate Li₂S_n species, and thus overcome their dissolution into the electrolyte, achieving long-term cycling stability and high-rate performance.⁷⁰

Fig. 8 Ex situ XPS measurements of Al₂O₃–Li₂S–GS cathodes at selected voltage states: (a) 3.8 V, (b) 2.3 V, (c) 2.1 V, (d) 2.05 V and (e) 1.8 V. (2p_3/2-yellow (S): 163.7 eV; 2p_3/2-cyan (Li₂S₈): 163.6 eV; 2p_{3/2-blueviolet} (Li₂S₆): 163.2 eV; 2p_3/2-magenta (Li₂S₄): 162.6 eV; 2p_3/2-green (Li₂S₃): 161.5 eV; 2p_3/2-blue (Li₂S): 160.3 eV) (f) Li 1s XPS spectrum of Al₂O₃–Li₂S–GS cathodes at 2.1 V involving the Li–S and Li–O binding.

As is well known, the Li₂S_n (4 ≤ n ≤ 8) species under the upper plateaus (the plateau at ∼2.3 and 2.1 V) are highly unstable and soluble in the electrolyte, and they cause the shuttle effect and with the electrolyte solvent, resulting in the loss of active materials and gradual capacity degradation. On the other hand, the capacity of the lower plateau (at ∼2.05 V), dominated by the capacity of electrode, is deeply dependent on the upper plateau. Therefore, stabilizing Li₂S_n (4 ≤ n ≤ 8) species and suppressing their dissolution are crucial for improving the performance of the electrode. Compared to Li₂S–GS cells, for which the upper plateau as well as the lower plateau shrink severely with prolonged cycling (Fig. S16†), the capacity of Al₂O₃–Li₂S–GS cells can be mainly attributed to the confinement effect of ultrathin Al₂O₃ films on long-chain polysulfides.

Conclusions

In summary, we have demonstrated the preferential growth of Al₂O₃ on the surface of Li₂S particles in free-standing Li₂S–GS composites by using an ALD technique. The porous GS framework without an Al₂O₃ coating provides excellent electronic and ionic conductivity. Besides acting as a physical barrier to entrap Li₂S_n, ultrathin Al₂O₃ films coated on the surface of Li₂S particles also exhibit strong binding interaction with polysulfides to effectively suppress their dissolution. Therefore, for Al₂O₃–Li₂S–GS cells, the capacity is maintained, resulting in an excellent stability. The free-standing Al₂O₃–Li₂S–GS cathodes are able to achieve a high capacity of 736 mA h g_Li2S⁻¹ with a high capacity retention of 85%, much higher than that of Li₂S–GS cathodes (59%). In addition, Al₂O₃–Li₂S–GS cathodes can deliver a capacity of 643 mA h g_Li2S⁻¹ (373 mA h g_cathode⁻¹) with a high capacity retention of 88% after 300 cycles at 0.5C and a low capacity decay rate of 0.028% per cycle during 1000 cycles at 2C. This work provides a new approach for coating Li₂S particles to obtain high-performance Li₂S cathodes.

Acknowledgements

This work was supported by the National Natural Science Foundation of China (No. 51572060 and 51502062), China Postdoctoral Science Foundation Funded Project (No. 2016M590279), the Fundamental Research Funds for the Central Universities (No. HIT.BRETIII.201224 and 201312) and Program for Innovation Research of Science in Harbin Institute of Technique (PIRS of HIT-No. 201506). Excellent Youth Foundation of Heilongjiang Scientific Committee (No. JC2015010). G. Wu acknowledges the start-up funding from the University at Buffalo (SUNY) along with the National Science Foundation (CBET-1604392).

References

1. B. Scrosati, J. Hassoun and Y. K. Sun, Energy Environ. Sci., 2011, 4, 3287–3295.
2. L. B. Hu, J. W. Choi, Y. Yang, S. Jeong, F. La Mantia, L. F. Cui and Y. Cui, Proc. Natl. Acad. Sci. U. S. A., 2009, 106, 21490–21494.
3. X. L. Ji and L. F. Nazar, J. Mater. Chem., 2010, 20, 9821–9826.
4. A. Manthiram, Y. Z. Fu, S. H. Chung, C. X. Zu and Y. S. Su, Chem. Rev., 2014, 114, 11751–11787.
5. Z. Li, H. B. Wu and X. W. Lou, Energy Environ. Sci., 2016, 9, 3061–3070.
6. R. Cao, W. Xu, D. Lv, J. Xiao and J.-G. Zhang, Adv. Energy Mater., 2015, 5, 1402273–1402295.
7. S. Urbonaite, T. Poux and P. Novák, Adv. Energy Mater., 2015, 5, 1500118–1500123.
8. M. S. Whittingham, Chem. Rev., 2004, 104, 4271–4301.
9. H. Jha, I. Buchberger, X. Y. Cui, S. Meini and H. A. Gasteiger, J. Electrochem. Soc., 2015, 162, A1829–A1835.
10. N. A. Liu, L. B. Hu, M. T. McDowell, A. Jackson and Y. Cui, ACS Nano, 2011, 5, 6487–6493.
11. M. Agostini, J. Hassoun, J. Liu, M. Jeong, H. Nara, T. Momma, T. Osaka, Y. K. Sun and B. Scrosati, ACS Appl. Mater. Interfaces, 2014, 6, 10924–10928.
12. J. Hassoun and B. Scrosati, Angew. Chem., Int. Ed., 2010, 49, 2371–2374.
13. J. Hassoun, Y. K. Sun and B. Scrosati, J. Power Sources, 2011, 196, 343–348.
14. G. Q. Ma, Z. Y. Wen, M. F. Wu, C. Shen, Q. S. Wang, J. Jin and X. W. Wu, Chem. Commun., 2014, 50, 14209–14212.
15. J. Bruckner, S. Thieme, F. Bottger-Hiller, I. Bauer, H. T. Grossmann, P. Strubel, H. Althues, S. Spange and S. Kaskel, Adv. Funct. Mater., 2014, 24, 1284–1289.
16. Y. Son, J. S. Lee, Y. Son, J. H. Jang and J. Cho, Adv. Energy Mater., 2015, 5, 1500110–1500123.
17. S. Jeong, D. Bresser, D. Buchholz, M. Winter and S. Passerini, J. Power Sources, 2013, 235, 220–225.
18. J. C. Guo, Z. C. Yang, Y. C. Yu, H. D. Abruna and L. A. Archer, J. Am. Chem. Soc., 2013, 135, 763–767.
19. C. Wang, X. S. Wang, Y. Yang, A. Kushima, J. T. Chen, Y. H. Huang and J. Li, Nano Lett., 2015, 15, 1796–1802.
20. S. Meini, R. Elazari, A. Rosenman, A. Garsuch and D. Aurbach, J. Phys. Chem. Lett., 2014, 5, 915–918.
21. R. Xu, X. F. Zhang, C. Yu, Y. Ren, J. C. M. Li and I. Belharouak, ChemSusChem, 2014, 7, 2457–2460.
22. J. T. Zhang, H. Hu, Z. Li and X. W. Lou, Angew. Chem., Int. Ed., 2016, 55, 3982–3986.
23. Z. Li, J. T. Zhang, Y. M. Chen, J. Li and X. W. Lou, Nat. Commun., 2015, 6, 8850–8857.
24. K. P. Cai, M. K. Song, E. J. Cairns and Y. G. Zhang, Nano Lett., 2012, 12, 6474–6479.
25. L. M. Suo, Y. J. Zhu, F. D. Han, T. Gao, C. Luo, X. L. Fan, Y. S. Hu and C. S. Wang, Nano Energy, 2015, 13, 467–473.
26. C. Y. Nan, Z. Lin, H. G. Liao, M. K. Song, Y. D. Li and E. J. Cairns, J. Am. Chem. Soc., 2014, 136, 4659–4663.
27. Y. Z. Fu, C. X. Zu and A. Manthiram, J. Am. Chem. Soc., 2013, 135, 18044–18047.
28. L. Chen, Y. Z. Liu, M. Ashuri, C. H. Liu and L. L. Shaw, J. Mater. Chem. A, 2014, 2, 18026–18032.
29. F. X. Wu, J. T. Lee, F. F. Fan, N. Nitta, H. Kim, T. Zhu and G. Yushin, Adv. Mater., 2015, 27, 5579–5586.
30. F. X. Wu, J. T. Lee, E. B. Zhao, B. Zhang and G. Yushin, ACS Nano, 2016, 10, 1333–1340.
31. M. Wu, Y. Cui and Y. Z. Fu, ACS Appl. Mater. Interfaces, 2015, 7, 21479–21486.
32. G. M. Zhou, E. Paek, G. S. Hwang and A. Manthiram, Adv. Energy Mater., 2016, 6, 1501355–1501363.
33. D. Sun, Y. Hwa, Y. Shen, Y. Huang and E. J. Cairns, Nano Energy, 2016, 26, 524–532.
34. Z. W. Seh, H. T. Wang, P. C. Hsu, Q. F. Zhang, W. Y. Li, G. Y. Zheng, H. B. Yao and Y. Cui, Energy Environ. Sci., 2014, 7, 672–676.
35. Z. W. Seh, J. H. Yu, W. Y. Li, P. C. Hsu, H. T. Wang, Y. M. Sun, H. B. Yao, Q. F. Zhang and Y. Cui, Nat. Commun., 2014, 5, 5017–5024.
36. H. Kim, J. T. Lee, D. C. Lee, A. Magasinski, W. I. Cho and G. Yushin, Adv. Energy Mater., 2013, 3, 1308–1315.
37. M. P. Yu, W. J. Yuan, C. Li, J. D. Hong and G. Q. Shi, J. Mater. Chem. A, 2014, 2, 7360–7366.
38. X. G. Han, Y. H. Xu, X. Y. Chen, Y. C. Chen, N. Weadock, J. Y. Wan, H. L. Zhu, Y. L. Liu, H. Q. Li, G. Rubloff, C. S. Wang and L. B. Hu, Nano Energy, 2013, 2, 1197–1206.
39. M. P. Yu, A. J. Wang, F. Y. Tian, H. Q. Song, Y. S. Wang, C. Li, J. D. Hong and G. Q. Shi, Nanoscale, 2015, 7, 5292–5298.
40. S. Lu, Y. Chen, X. Wu, Z. Wang and Y. Li, Sci. Rep., 2014, 4, 4629–4633.
41. Z. Li, S. G. Zhang, C. Zhang, K. Ueno, T. Yasuda, R. Tatara, K. Dokko and M. Watanabe, Nanoscale, 2015, 7, 14385–14392.
42. F. M. Ye, Y. Hou, M. N. Liu, W. F. Li, X. W. Yang, Y. C. Qiu, L. S. Zhou, H. F. Li, Y. J. Xua and Y. G. Zhang, Nanoscale, 2015, 7, 9472–9476.
43. Z. Lin, C. Y. Nan, Y. F. Ye, J. H. Guo, J. F. Zhu and E. J. Cairns, Nano Energy, 2014, 9, 408–416.
44. Y. Hwa, J. Zhao and E. J. Cairns, Nano Lett., 2015, 15, 3479–3486.
45. Y. Yang, M. T. McDowell, A. Jackson, J. J. Cha, S. S. Hong and Y. Cui, Nano Lett., 2010, 10, 1486–1491.
46. X. B. Meng, D. J. Comstock, T. T. Fister and J. W. Elam, ACS Nano, 2014, 8, 10963–10972.
47. Z. C. Yang, J. C. Guo, S. K. Das, Y. C. Yu, Z. H. Zhou, H. D. Abruna and L. A. Archer, J. Mater. Chem. A, 2013, 1, 1433–1440.
48. M. Kohl, J. Bruckner, I. Bauer, H. Althues and S. Kaskel, J. Mater. Chem. A, 2015, 3, 16307–16312.
49. S. Zhang, M. N. Liu, F. Ma, F. M. Ye, H. F. Li, X. Y. Zhang, Y. Hou, Y. C. Qiu, W. F. Li, J. Wang, J. Wang and Y. G. Zhang, J. Mater. Chem. A, 2015, 3, 18913–18919.
50. M. B. M. Mousa, C. J. Oldham and G. N. Parsons, Langmuir, 2014, 30, 3741–3748.
51. A. Mortensen, D. H. Christensen, O. F. Nielsen and E. Pedersen, J. Raman Spectrosc., 1991, 22, 47–49.
52. Z. W. Seh, H. T. Wang, N. Liu, G. Y. Zheng, W. Y. Li, H. B. Yao and Y. Cui, Chem. Sci., 2014, 5, 1396–1400.
53. J. F. Moulder, W. F. Stickle, P. E. Sobol and K. D. Bomben, Physical Electronics Division, Perkin-Elmer Corp., 1995, pp. 33–34.
54. F. X. Wu, J. T. Lee, N. Nitta, H. Kim, O. Borodin and G. Yushin, Adv. Mater., 2015, 27, 101–108.
55. S. S. Zhang, Electrochim. Acta, 2012, 70, 344–348.
56. K. Zhang, L. J. Wang, Z. Hu, F. Y. Cheng and J. Chen, Sci. Rep., 2014, 4, 6467–6473.
57. Y. Yang, G. Y. Zheng, S. Misra, J. Nelson, M. F. Toney and Y. Gui, J. Am. Chem. Soc., 2012, 134, 15387–15394.
58. L. Wang, Y. G. Wang and Y. Y. Xia, Energy Environ. Sci., 2015, 8, 1551–1558.
59. S. Walus, C. Barchasz, R. Bouchet, J. F. Martin, J. C. Lepretre and F. Alloin, Electrochim. Acta, 2015, 180, 178–186.
60. C. Barchasz, F. Molton, C. Duboc, J. C. Lepretre, S. Patoux and F. Alloin, Anal. Chem., 2012, 84, 3973–3980.
61. R. Dominko, M. U. M. Patel, V. Lapornik, A. Vizintin, M. Kozelj, N. N. Tusar, I. Arcon, L. Stievano and G. Aquilanti, J. Phys. Chem. C, 2015, 119, 19001–19010.
62. N. A. Canas, D. N. Fronczek, N. Wagner, A. Latz and K. A. Friedrich, J. Phys. Chem. C, 2014, 118, 12106–12114.
63. M. Hagen, P. Schiffels, M. Hammer, S. Dorfler, J. Tubke, M. J. Hoffmann, H. Althues and S. Kaskel, J. Electrochem. Soc., 2013, 160, A1205–A1214.
64. H. L. Wu, L. A. Huff and A. A. Gewirth, ACS Appl. Mater. Interfaces, 2015, 7, 1709–1719.
65. J. T. Yeon, J. Y. Jang, J. G. Han, J. Cho, K. T. Lee and N. S. Choi, J. Electrochem. Soc., 2012, 159, A1308–A1314.
66. S. Walus, C. Barchasz, J. F. Colin, J. F. Martin, E. Elkaim, J. C. Lepretre and F. Alloin, Chem. Commun., 2013, 49, 7899–7901.
67. N. A. Canas, S. Wolf, N. Wagner and K. A. Friedrich, J. Power Sources, 2013, 226, 313–319.
68. J. Nelson, S. Misra, Y. Yang, A. Jackson, Y. J. Liu, H. L. Wang, H. J. Dai, J. C. Andrews, Y. Cui and M. F. Toney, J. Am. Chem. Soc., 2012, 134, 6337–6343.
69. M. Wild, L. O'Neill, T. Zhang, R. Purkayastha, G. Minton, M. Marinescu and G. J. Offer, Energy Environ. Sci., 2015, 8, 3477–3494.
70. Q. F. Zhang, Y. P. Wang, Z. W. Seh, Z. H. Fu, R. F. Zhang and Y. Cui, Nano Lett., 2015, 15, 3780–3786.

---

Footnotes

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c6ta08039a

‡ These two authors contributed equally to this work.

---