Achieving superb sodium storage performance on carbon anodes through an ether-derived solid electrolyte interphase

Abstract

High specific surface area carbon (HSSAC) is a class of promising high-capacity anode materials for sodium-ion batteries (SIBs). A critical bottleneck of the HSSAC anode, however, is the ultra-low initial coulombic efficiency (ICE) in commonly used ester-based electrolytes. This phenomenon further prohibits improving the specific capacity, long-term stability and rate capability of HSSAC anodes. This work reports the largely enhanced anode performance of several different HSSAC anodes in ether-based electrolytes. Very importantly, with the reduced graphene oxide (rGO) anode as one example, the ICE can be as high as 74.6% accompanied by a large reversible specific capacity of 509 mA h g⁻¹ after 100 cycles at a current density of 0.1 A g⁻¹. 75.2% of the capacity was retained after 1000 cycles at 1 A g⁻¹. Even at a high current density of 5 A g⁻¹, the specific capacity of the rGO anode can be obtained at 196 mA h g⁻¹. This extraordinary performance is ascribed to the stable, thin, compact, uniform and ion conducting solid electrolyte interphase (SEI) formed in an ether-based electrolyte. Fortunately, this SEI-modifying strategy is generic and is independent of the specific microstructures of HSSAC anodes, indicating a promising avenue for manipulating the SEI on HSSAC anodes through utilizing ether solvents to enable achievement of high ICE for large-capacity HSSAC anodes for practical applications.

---

Broader context

High specific surface area carbon (HSSAC) is a class of promising high-capacity anode materials for room-temperature sodium-ion batteries (SIBs). Unfortunately, two critical bottlenecks, namely the ultra-low initial coulombic efficiency (ICE) and large irreversible specific capacity severely impede the practical application of HSSAC anode materials. This work points out an easy and viable approach to tackle the above-mentioned two performance-limiting problems of HSSAC by utilizing an ether-based solvent. It has been revealed that the ether-based solvent can markedly manipulate the microstructure and composition of a solid electrolyte interphase (SEI) on HSSAC anodes, resulting in a stable, thin, compact, uniform and ion conducting SEI, which is the key to the extraordinary sodium storage performance and notably different from a SEI formed in the commonly used ester-based solvent. This approach is generically feasible for various HSSAC anode materials and will further promote the practical applications of ether-based electrolytes not only in carbon-based anodes but also in non-carbon anodes.

Introduction

Room-temperature sodium-ion batteries (SIBs), which are based on a similar ion-shuttle process to the mature lithium-ion batteries (LIBs), have developed rapidly in the last few years. Because of the advantages of sodium such as its natural abundance, widespread distribution, low cost, and excellent sustainability, SIBs are particularly appropriate for grid-scale energy storage systems.^1–6 In contrast to the large number of suitable positive electrode materials that have been investigated, possible materials for negative electrodes are more restricted. Among various possible anode materials, carbonaceous materials show better viability.^7–14 Recent reports have demonstrated that high specific surface area carbon (HSSAC) is thermodynamically feasible for sodium storage since sodium can not only fill the pores between turbostratic nano-domains (TNs) but can also intercalate into the TNs and reversibly react with surface defects.^15–17 Unfortunately, in the most commonly used ester-based electrolytes, such as ethylene carbonate (EC), diethyl carbonate (DEC), propylene carbonate (PC), dimethyl carbonate (DMC) and methyl ethyl carbonate (EMC), HSSAC anodes typically exhibit an ultra-low initial coulombic efficiency (ICE) in the range between 10% and 40%, which severely limits their industrial applications.^18–21

The ultra-low ICE of HSSAC anodes originates from the formation of a solid electrolyte interphase (SEI), which irreversibly consumes massive sodium ions on the large-area carbon surfaces.²² Moreover, since the Lewis acidity of sodium ions is weaker than that of lithium ions, the solubility of the electrochemical decomposition products of ester-based electrolytes in the sodium system is apparently higher than that for the lithium system.²³ Even worse, a few components in the SEI such as ROCO₂Na may cause degradation and instability of the electrodes during cycling.²⁴ All these features of the SEI result in poor electrochemical performance of SIBs when ester-based solvents are used as electrolytes in HSSAC anodes. Therefore, it is of vital importance to manipulate the microstructure and components of the SEI with the purpose of achieving high anode performance of HSSAC anodes in SIBs.

Recently, ether-based electrolytes have attracted extensive attention in SIBs since they can trigger the reversible co-intercalation of solvent molecules and sodium ions into graphite, a typical low specific surface area carbon (LSSAC), to form stable ternary compounds.^25–29 On the other side, the reaction of the ether-based solvent at the surfaces of HSSAC is intriguing and might play a significant role in the anode reacting process, particularly the formation of a solid electrolyte interphase (SEI). From this point of view, the role that the ether solvent plays at the interfaces with HSSAC and LSSAC anodes might be somewhat different. Besides, ether-based electrolytes can facilitate the transport of electrons and the diffusion of sodium ions,³⁰ as well as change the SEI formation kinetics on sodium metal.³¹ Our study reports the first observation of a stable, thin, compact, uniform and ion conducting SEI on HSSAC anodes in an ether-based diglyme electrolyte. We further show the greatly improved ICE of HSSAC anodes by utilizing ether-based electrolytes. Better yet, the reversible specific capacity, long-term stability and rate capability of HSSAC anodes can be markedly enhanced as well.

In this work, we took reduced graphene oxide (rGO) as a model anode to study the functions of ether-based solvents. The findings were also examined with another two typical HSSAC anodes: activated carbon (AC) and ordered mesoporous carbon (CMK-3). Amazingly, a rGO electrode cycled in an ether-based solvent can deliver much better electrochemical performance, with an initial coulombic efficiency of 74.6%, an impressive reversible specific capacity of 509 mA h g⁻¹ after 100 cycles at 0.1 A g⁻¹, a good rate capability (196 mA h g⁻¹ at 5 A g⁻¹), and a long-term stability (a capacity retention of 75.2% after 1000 cycles), which outperforms the majority of the present carbon anodes. Moreover, this is the common case with both AC and CMK-3, demonstrating that this SEI-modifying strategy is generic and is independent of the specific microstructures of HSSAC anodes.

Experimental methods

Preparation of different HSSACs

Graphite oxide was prepared from the graphite powder using a modified Hummers method as reported earlier.³² Thermally reduced graphene oxide, namely rGO, was obtained by a two-step thermal treatment of graphite oxide at 80 °C for 30 min and at 300 °C for 2 h in a vacuum, with a heating rate of 5 °C min⁻¹ from room temperature. For comparison, rGO-900 was obtained by a one-step thermal treatment of rGO from room temperature to 900 °C at a rate of 5 °C min⁻¹ and keeping still for 2 h. Two types of typical hard carbons, CMK-3 and AC (XFNANO, China), were directly used without any further treatment.

Characterization methods

Scanning electron microscopy (SEM) observation was performed using a Hitachi S-4800 (Hitachi, Japan). TEM and STEM observations were performed using a field emission transmission electron microscope (HR-TEM, FEI TecnaiG2 F30) at an accelerating voltage of 300 kV. XRD measurements were conducted at room temperature using the reflection mode (D8 FOCUS, Cu Kα radiation, λ = 0.154056 nm). Raman spectra were collected on a HORIBA Jobin Yvon LabRAM HR800 using a 532 nm laser source. Nitrogen cryo-adsorption was measured by using a BEL mini instrument, and the specific surface area was obtained by Brunauer–Emmett–Teller (BET) analyses of the adsorption isotherms. Fourier Transform infrared (FTIR) spectroscopy was performed on a Nicolet iS50. X-ray photoelectron spectroscopy (XPS) was conducted on an ESCALAB 250Xi (Thermo Fisher) at room temperature using a monochromatic Al Kα source. Atomic force microscope (AFM) observation was performed using a Bruker Mutimode8 (Bruker, Germany) in an argon-filled glovebox with contents of oxygen and moisture below 20 ppm.

Preparation of electrodes and post treatment of electrodes for ex situ characterization

All the active materials were ground into fine powder, and the ground active materials (70 wt%) were homogeneously mixed with conductive carbon black (15 wt%) and PVDF (15 wt%) in NMP to form a slurry. The slurry was stirred for 6 h and coated onto a carbon-coated Al foil with a constant thickness of 90 μm and a mass loading of approximately 0.8 mg cm⁻². After vacuum-drying at 110 °C for 10 h, the electrodes were cut into circular pieces with a diameter of 12 mm, which were used as anodes for the assembled cell. The electrolytes for comparison were 1 M sodium triflate (NaOTf) dissolved in a mixture of ethylene carbonate (EC) and diethyl carbonate (DEC) (volume ratio: 1 : 1), as well as 1 M sodium triflate (NaOTf) dissolved in diglyme. Using the above-mentioned electrodes and electrolytes, together with pure sodium foils as counter electrodes and Whatman GF/D glass fibers as separators, 2032-type coin cells were assembled in an argon-filled glove box with contents of oxygen and moisture below 1 ppm.

For further ex situ characterization, post treatment of electrodes was needed. It is worth noting that when using Whatman GF/D glass fibers as separators, few amounts of excessive glass fibers would unfortunately attach on the surface of electrodes after disassembly, causing an adverse impact on the characterization of morphology. Fortunately, polypropylene (PP) separators which could be easily separated from electrodes showed similar electrochemical characteristics to glass fiber separators, as shown in Fig. S15 (ESI†). Therefore, for all ex situ characterization, PP separators can be used as separators. After the disassembly, the recovered electrodes were carefully washed several times with a related solvent, namely DMC and DME. Then the washed electrodes were dried overnight to fully evaporate the pure solvent and then transferred to the vacuum box.

Electrochemical characterization

Galvanostatic charge–discharge tests were conducted on a LAND CT2001 battery program controlling system at different current densities with a voltage window of 0.005–3 V vs. Na/Na⁺. Cyclic voltammetry measurements were conducted using an electrochemical workstation (VMP3, Bio Logic, France) at various scanning rates also in the voltage range of 0.005–3 V vs. Na/Na⁺. Electrochemical impedance spectra (EIS) after different cycles were examined using the VMP3 multichannel electrochemical station in the frequency range of 10⁻²–10⁵ Hz by applying a 5 mV ac oscillation.

Results and discussion

Here, the rGO sample was prepared by a mild thermal reduction at 300 °C and was denoted rGO.^33,34 The well-developed pore structure of rGO is favorable for sodium diffusion and storage.²² Due to its unique surface charge storage mechanism and complex surface properties, rGO has exhibited extraordinary electrochemical characteristics in a conventional ether-based solvent, namely diglyme.

The widely accepted sodium storage mechanisms of rGO in ester-based electrolytes are sodium ad-/de-sorption at planar surfaces and pseudo-capacitive storage at oxygen functional groups on the edges.^22,35–37 As shown in Fig. 1 and Fig. S2 (ESI†), the electrochemical characteristics of rGO in EC/DEC (a conventional ester solvent) in this work conform to the above mechanism, while sodium salts had a negligible impact. In stark contrast, when substituting diglyme for EC/DEC, a huge difference was observed regarding both the sodium storage mechanism and the anode performance, while sodium salts still showed little impact, as shown in Fig. 1 and Fig. S3 (ESI†).

Fig. 1 Systematic comparison of the electrochemical performance of rGO electrodes in different solvents. (a) Galvanostatic charge–discharge profiles of the first two cycles for NaOTf (abbreviation of NaCF₃SO₃) in EC/DEC and diglyme at a current density of 0.1 A g⁻¹; (b) cycling performance of 100 cycles for NaOTf in EC/DEC and diglyme at 0.1 A g⁻¹; (c) rate performance of NaOTf in EC/DEC and diglyme at current densities ranging from 0.05 A g⁻¹ to 5 A g⁻¹; (d) cycling performance of NaOTf in EC/DEC and diglyme at 1 A g⁻¹; (e) CV curves for the first two cycles of NaOTf in EC/DEC and diglyme at a scan rate of 0.2 mV s⁻¹; and (f) EIS after the first cycle of NaOTf in EC/DEC and diglyme.

It is worth noting that the irreversible discharge capacity in the first cycle in EC/DEC is much higher than in diglyme. In Fig. 1e, there is an obvious broad cathodic peak between 0.1 V and 0.36 V during the first cycle in EC/DEC, which can be assigned to the severe decomposition of EC/DEC to form a solid electrolyte interphase (SEI), since an ester-based solvent has a higher reduction potential and is more easily reduced on the surface of the electrode than an ether-based solvent.²³ Therefore, large amounts of decomposed organic products precipitate to a thicker SEI that grows upon cycling, leading to a higher charge transfer resistance (Fig. 1f and Fig. S2, ESI†). Because the solvent plays a vital role in the formation of the SEI, the ultra-low ICE, which is the most annoying problem for rGO-based anodes, can be effectively addressed in an ether-based diglyme electrolyte. A dramatic improvement in the ICE from 39% to 74.6% is observed at a current density of 0.1 A g⁻¹ (Fig. 1a). The remarkable efficacy of diglyme in achieving larger ICE was also verified generically by examinations on AC and CMK-3 anodes, as shown in Fig. S4 (ESI†). The ICE for AC and CMK-3 increased from 13.7% to 59.6% and from 23.1% to 62.8%, respectively. After the first cycle, the coulombic efficiency of the rGO anode in diglyme rapidly increased up to nearly 100% compared to a very slow increase in EC/DEC.

Better still, the reversible charge capacity in diglyme (509 mA h g⁻¹) is almost twice that in EC/DEC (262 mA h g⁻¹) after 100 cycles, demonstrating the superiority of diglyme, not only because of the higher capacity but also the better cycling stability. The reversible charge capacity for AC increased from 59.7 mA h g⁻¹ to 217.2 mA h g⁻¹, while for CMK-3 from 272.1 mA h g⁻¹ to 344.1 mA h g⁻¹. The universal capacity increment suggests that the HSSAC anodes can be sodiated to a larger extent in the diglyme solvent, regardless of the diverse specific microstructural features. However, because of the varying microstructure and surface chemistry among rGO, AC and CMK-3, the individual performance of three HSSAC anodes varies. The rate performance of the rGO anode was tested at current densities ranging from 0.05 A g⁻¹ to 5 A g⁻¹ and is shown in Fig. 1c. When the current density decreased to 0.05 A g⁻¹, the charge capacity reached as high as 624.8 mA h g⁻¹ and became almost stable during the following ten cycles. Along with the increase in current density, no sudden capacity fade occurred, revealing the fast kinetics of sodium storage that takes place in diglyme. Even at the highest current density of 5 A g⁻¹, the capacity was as high as 195.8 mA h g⁻¹. This value is among the highest reversible capacities of carbon-based anodes for SIBs at such a high current density. This superior high-rate performance is mainly because of the optimized SEI which facilitates ultrafast electron transfer and ion transport. When the current density was switched back to 0.1 A g⁻¹, the capacity retention was as high as 97.7% of the previous value, demonstrating excellent reversibility. A long cycling test was also performed at a high current density of 1 A g⁻¹, as shown in Fig. 1d. The reversible capacity was 332.2 mA h g⁻¹ and even after cycling for 1000 cycles, the capacity retention was 75.2%, with only a small decay rate of 0.028% per cycle.

Cyclic voltammetry (CV) tests were conducted to understand the different electrochemical characteristics of the two solvents. As shown in Fig. 1e, there are sharper peaks at around 0.1 V during both the anodic and cathodic scans, which are presumed to be due to the insertion/extraction of solvated sodium ions into/from nano-voids formed by the disordered graphene nanosheets. Moreover, at a high voltage of around 2.5 V, a broad peak appeared, indicating the contribution of pseudo-capacitance that may originate from oxygen functional groups. For comparison, rGO-900 was produced by annealing rGO at 900 °C in an inert atmosphere to remove most of the oxygen functional groups (Fig. S5, ESI†). As a consequence, the broad peak centered at 2.5 V became much less apparent and the capacity became much lower, demonstrating its direct correlation with oxygen functional groups (Fig. S6, ESI†). Because the decomposition of EC/DEC is partially correlated with oxygen functional groups while this is not the case in diglyme, the pseudo-capacitance in diglyme is much higher than in EC/DEC.

In order to thoroughly understand the different reaction mechanisms of rGO in different solvents, CV tests at various scan rates were conducted, as shown in Fig. 2a and b. Charge storage due to double layer capacitance and pseudo-capacitance can be analyzed according to i = av^b, where the measured current i obeys a power law relationship with the scan rate v. Both a and b are adjustable parameters and the b-exponent values are determined from the slope of a log(i) versus log(v) plot. Two well-defined conditions are dependent on b: b = 0.5 and b = 1.0. When b = 0.5, the current response refers to a faradaic process including insertion, intercalation etc. When b = 1.0, the current response refers to a capacitive process.^38,39 Four conditions denoted C1, C2, C3 and D1 were chosen to compare the different mechanisms in different solvents. For C1, the b values are 0.764 for diglyme and 0.914 for EC/DEC, indicating that a considerable portion of the capacity in diglyme is contributed by extracting solvated sodium from nano-voids between crumpled graphene nanosheets, while in EC/DEC the desorption of sodium ions from graphene nanosheets contributed most of the capacity. For C2, the b values are 0.951 for diglyme and 0.864 for EC/DEC, indicating the easier desorption of sodium ions from graphene nanosheets and faster kinetics in diglyme than in EC/DEC. For C3, the b values are 0.803 for diglyme and 0.878 for EC/DEC, indicating that in both cases the existence of pseudo-capacitance resulted from the reversible reaction between oxygen functional groups and sodium ions but more evident in diglyme than in EC/DEC. For D1, the b values are 0.941 for diglyme and 0.806 for EC/DEC, indicating better adsorption of sodium on graphene nanosheets in diglyme than in EC/DEC and faster kinetics. As such, the possible mechanism of sodium storage on rGO-300 in diglyme is: (1) solvated sodium insertion into/extraction from nano-voids between crumpled graphene nanosheets at around 0.1 V (C1); (2) sodium adsorption on/desorption from graphene nanosheets (C2/D1); (3) in the high voltage region (C3), a reversible reaction between sodium and oxygen functional groups on the surface or on the edge sites.

Fig. 2 Analytical comparison of the electrochemical mechanism of rGO electrodes in different solvents. CV curves at various scan rates from 0.2 mV s⁻¹ to 2 mV s⁻¹ for (a) NaOTf in diglyme and (b) NaOTf in EC/DEC; (c and d) determination of the b values for various selected regions for (c) NaOTf in diglyme and (d) NaOTf in EC/DEC.

The intermediates formed during solvent reduction are distinct for different solvents, and hence we determined the components and microstructure of the SEI. The largely improved anode performance of rGO in diglyme is believed to be related to the morphology and structure as well as the chemical composition of the SEI. To probe and distinguish the SEI on the surface of rGO, we disassembled cells without cycling (denoted as pristine), after ten full cycles in diglyme (denoted as ether) and EC/DEC (denoted as ester). The electrodes from disassembled cells were preserved in a vacuum box and during the characterization were transferred as quickly as possible to avoid exposure to air or moisture.^31,40 As shown in Fig. 3, the pristine electrode still consisted of randomly agglomerated graphene nanosheets and surrounding conductive additives and binders. In contrast, after ten cycles in ether (diglyme), the surface of the rGO electrode showed fewer wrinkles and became less bright, indicative of thin SEI formation. In the ester (EC/DEC), the wrinkles almost disappeared and the surface of the electrode became even less bright, indicative of thicker SEI formation. TEM-EDS were conducted to analyze the elemental composition of the different electrodes as a function of depth. As shown in Fig. S7–S9 (ESI†), five elements, carbon, oxygen, sodium, fluorine and sulfur, were detected in electrodes. For the pristine electrode, trace amounts of sodium, fluorine and sulfur originating from a residual conductive salt in the electrolyte were detected, because they were not totally washed away by the solvent. For electrodes cycled in diglyme, less sodium and oxygen were detected. This phenomenon suggests that the diglyme solvent resists severe decomposition on the rGO anode surfaces and alleviates the irreversible sodiation processes. Apart from this, all elements in the three electrodes were uniformly distributed.

Fig. 3 Morphology comparison of rGO electrodes cycled in different solvents. SEM images of (a) pristine rGO electrodes, (b) rGO electrodes cycled in diglyme, and (c) rGO electrodes cycled in EC/DEC; low-resolution TEM images of (d) pristine rGO electrodes, (e) rGO electrodes cycled in diglyme, and (f) rGO electrodes cycled in EC/DEC; high-resolution TEM images of (g) pristine rGO electrodes, (h) rGO electrodes cycled in diglyme, and (i) rGO electrodes cycled in EC/DEC.

To further reveal the components of the SEI, XPS was conducted to analyze the element concentration and to deduce the possible compounds on the surface of the different electrodes.^40–42 Moreover, 60s Ar⁺ ion etching was conducted to depict the depth profile of composition. For convenience, we denoted ether-based electrodes after etching as ether-60s and ester-based electrodes after etching as ester-60s. As shown in Fig. S11 (ESI†), the newly appeared Na and S peaks are associated with the SEI. For electrodes cycled in ether, less sodium and oxygen were also found (Fig. S12, ESI†). The content of fluorine, which is from the binder and the salt NaOTf, was lower for ether. This could be related to the suppressed decomposition of NaOTf in diglyme solvent. In addition, after the 60s Ar⁺ ion etching, the contents of carbon, oxygen, sodium and fluorine in ether-60s and ester-60s dramatically changed, indicating a concentration gradient from the surface to sub-surface/bulk.

High-resolution XPS C 1s, O 1s, Na 1s, F 1s and S 2p_3/2 spectra were collected to reveal the detailed structure and component of the SEI of different samples. As shown in Table 1, for rGO electrodes cycled in ester, ether and both after etching, the most detected component for all samples is sp² carbon in rGO electrodes, indicating that the SEI thickness of all is within the XPS detection limit (ca. 10 nm). Moreover, the difference in concentration of sp² carbon between rGO electrodes cycled in ester and ether can be attributed to the difference of SEI thickness, namely the SEI is thicker when cycled in ester than in ether, which is consistent with the previous discussion. Interestingly, for both rGO electrodes cycled in ester and ether, the SEI is composed of both organics and inorganics. Organics are the dominant part of the SEI and mainly exist in the outer side. Specifically, polyethers and polyesters, which originated from the decomposition of the solvent, can be detected in both high-resolution XPS C 1s and O 1s, as shown in Table 1 and Fig. S13 (ESI†). According to the concentration variation before and after etching, the concentration of polymer species in rGO electrodes cycled in ether is lower in the exteriors and much lower in the interior of the SEI than rGO electrodes cycled in ester, demonstrating that a thinner and denser layer of decomposition products of solvents formed in the exteriors of the SEI. In addition, according to XPS O 1s, a sodium Auger electron peak clearly illustrates that a majority of the polymer species are sodium-based compounds. For inorganics in the SEI, Na₂CO₃/Na₂CO₂R can only be detected in ester according to XPS C 1s spectra, which results from the reduction of carbonate solvent and is unfortunately thermodynamically unstable. NaSO₃R can be detected in both solvents and can only be detected before etching according to XPS C 1s and S 2p_3/2 spectra, as shown in Fig. 4, indicating its existence in the exterior of the SEI. Besides, as a typical decomposition product from the conductive salts (NaOTf), CF₃ mainly exists in the exterior of the SEI in ether and almost disappears in the interior, while CF₃ exists in not only the exterior but also in the interior of the SEI in ester and the concentration is much higher than in ether, according to XPS C 1s and F 1s spectra in Fig. 4. As another typical decomposition product from the conductive salts (NaOTf), NaF can only be detected after etching, indicating its major existence in the interior of the SEI, according to XPS F 1s. Surprisingly, the existence of F–C (sp²) only in ether might result from a reaction between surface defects on rGO and reduction products of the conductive salt. Last but not least, as shown in Fig. S13 (ESI†), only one peak existed in XPS Na 1s and its intensity directly relates to the amount of sodium compounds. Therefore, higher intensity of the peak in ester again demonstrates a thicker SEI than in ether.

Table 1 Fitting results of high-resolution XPS of C 1s and O 1s

Elements	Position (eV)	Pristine (%)	Ether (%)	Ether-60s (%)	Ester (%)	Ester-60s (%)	Peak assignment
C 1s	284.5	62.9	62.9	69.6	50.8	61.0	sp^2-C
	285.5	14.1	14.3	20.9	20.1	20.3	–CH2–
	286.5	7.2	13.7	4.8	9.5	6.5	–C–O–
	288.0	6.6	8.0	4.7	11.2	8.5	–C=O–
	290.3	9.2	—	—	—	—	–CF2–
	290.8	—	0.9	—	5.5	3.4	Na2CO3/Na2CO2R
	293.2	—	0.2	—	2.9	0.3	CF3
O 1s	531.5	43.8	—	21.2	—	18.9	C=O/Na2CO3
	532.0	—	21.4	—	19.0	—	RSO3Na
	533.5	56.2	50.3	34.2	45.1	29.2	O–C/O–H/polyethers/polyesters
	536.8	—	28.3	44.6	35.9	51.9	Na KLL

---

Fig. 4 In-depth analysis of SEI components of rGO electrodes in different solvents. High-resolution XPS (a) F 1s spectra and (b) S 2p_3/2 spectra of pristine, ether, ester, ether-60s and ester-60s electrodes.

According to Fig. 5, it is postulated that the SEI formed in the ether-based electrolyte consists of a more compact and thinner organic layer at the exterior, which can significantly decrease the diffusion length of sodium ions and prevent extra decomposition of the electrolyte on rGO surfaces. Moreover, a relatively uniform mixture of organics/inorganics is located at the SEI interior, and this is excellent for sodium ion transport. Better still, some parts of the SEI, e.g. the sp² type F–C that does not exist in the ester SEI, can probably protect surface defects from irreversible reaction and promote the reversible electrochemical activity of surface defects, especially oxygen functional groups and edge sites. In contrast, the exterior part of the SEI formed in the ester-based electrolyte mainly consists of a thicker and non-uniform mixture of organics/inorganics, which is not stable during cycling. This is made even worse by the fact that some inorganics originating from the reduction of conductive salts are less conductive for sodium ion transfer and the thick SEI increases the diffusion distance. As a consequence, the diglyme electrolyte produces significantly better performance.

Fig. 5 Illustration of different components of the SEI in different electrolytes as well as its correlation with sodium storage.

Conclusions

We have demonstrated the extraordinary performance of an rGO anode for sodium storage in an ether-based diglyme electrolyte, which has been ascribed to the formation of a thin, uniform, compact, and ionic conducting SEI layer that is induced by the common ether solvent. An rGO electrode cycled in an ether-based solvent can deliver much better electrochemical performance, with an initial coulombic efficiency of 74.6%, an impressive reversible specific capacity of 509 mA h g⁻¹ after 100 cycles at 0.1 A g⁻¹, a good rate capability (196 mA h g⁻¹ at 5 A g⁻¹), and a long-term stability (a capacity retention of 75.2% after 1000 cycles), which outperforms the majority of the present carbon anodes. Our study showed the combined diffusion of sodium ions and capacitive sodium storage in rGO, which resulted from the unique properties of the ether-assisted improvement of the SEI. Moreover, this is the common case with other two typical HSSAC anodes, AC and CMK-3. This SEI-modifying strategy is generic and is independent of the specific microstructures of HSSAC anodes. Generally speaking, the low ICE of common high-capacity HSSAC anodes in ester-based electrolytes has likely limited their potential for practical application. The approach developed here for SEI modification indicates a promising avenue for manipulating the SEI on HSSAC anodes through utilizing ether solvents to enable achievement of high ICE for large-capacity HSSAC anodes for practical applications, inclusive of not only sodium-ion batteries, but also other metal-ion batteries.

Acknowledgements

We appreciate support from the National Science Fund for Distinguished Young Scholars, China (No. 51525204), the National Key Basic Research Program of China (2014CB932400), the National Natural Science Foundation of China (No. U1401243, 51372167 and 51502197), the Shenzhen Basic Research Project (No. ZDSYS20140509172959981), and the Australian Research Council Discovery Project (DP160103244).

Notes and references

1. S. Y. Hong, Y. Kim, Y. Park, A. Choi, N. Choi and K. T. Lee, Energy Environ. Sci., 2013, 6, 2067–2081.
2. D. Kundu, E. Talaie, V. Duffort and L. F. Nazar, Angew. Chem., Int. Ed., 2015, 54, 3431–3448.
3. H. Pan, Y. Hu and L. Chen, Energy Environ. Sci., 2013, 6, 2338–2360.
4. N. Yabuuchi, K. Kubota, M. Dahbi and S. Komaba, Chem. Rev., 2014, 114, 11636–11682.
5. W. Li, L. Zeng, Y. Wu and Y. Yu, Sci. China Mater., 2016, 59, 287–321.
6. R. C. Massé, E. Uchaker and G. Cao, Sci. China Mater., 2015, 58, 715–766.
7. E. Irisarri, A. Ponrouch and M. R. Palacin, J. Electrochem. Soc., 2015, 162, A2476–A2482.
8. H. Kim, H. Kim, Z. Ding, M. H. Lee, K. Lim, G. Yoon and K. Kang, Adv. Energy Mater., 2016 DOI:10.1002/aenm.201600943.
9. E. Castillo-Martínez, J. Carretero-González and M. Armand, Angew. Chem., Int. Ed., 2014, 53, 5341–5345.
10. H. Gao and J. B. Goodenough, Angew. Chem., Int. Ed., 2016, 55, 12768–12772.
11. S. Guo, P. Liu, H. Yu, Y. Zhu, M. Chen, M. Ishida and H. Zhou, Angew. Chem., Int. Ed., 2015, 54, 5894–5899.
12. L. Wang, Y. Lu, J. Liu, M. Xu, J. Cheng, D. Zhang and J. B. Goodenough, Angew. Chem., Int. Ed., 2013, 52, 1964–1967.
13. S. Wang, L. Wang, Z. Zhu, Z. Hu, Q. Zhao and J. Chen, Angew. Chem., Int. Ed., 2014, 53, 5892–5896.
14. H. Yu, Y. Ren, D. Xiao, S. Guo, Y. Zhu, Y. Qian, L. Gu and H. Zhou, Angew. Chem., Int. Ed., 2014, 53, 8963–8969.
15. C. Bommier, T. W. Surta, M. Dolgos and X. Ji, Nano Lett., 2015, 15, 5888–5892.
16. W. Luo, F. Shen, C. Bommier, H. Zhu, X. Ji and L. Hu, Acc. Chem. Res., 2016, 49, 231–240.
17. S. Zhang, W. Lv, C. Luo, C. You, J. Zhang, Z. Pan, F. Kang and Q. Yang, Energy Storage Mater., 2016, 3, 18–23.
18. Y. Li, L. Mu, Y. Hu, H. Li, L. Chen and X. Huang, Energy Storage Mater., 2016, 2, 139–145.
19. W. Luo, C. Bommier, Z. Jian, X. Li, R. Carter, S. Vail, Y. Lu, J. Lee and X. Ji, ACS Appl. Mater. Interfaces, 2015, 7, 2626–2631.
20. J. Xu, M. Wang, N. P. Wickramaratne, M. Jaroniec, S. Dou and L. Dai, Adv. Mater., 2015, 27, 2042–2048.
21. Y. Yan, Y. X. Yin, Y. G. Guo and L. J. Wan, Adv. Energy Mater., 2014, 4, 1301584.
22. Y. Wang, S. Chou, H. Liu and S. Dou, Carbon, 2013, 57, 202–208.
23. K. Xu, Chem. Rev., 2014, 114, 11503–11618.
24. M. Dahbi, N. Yabuuchi, M. Fukunishi, K. Kubota, K. Chihara, K. Tokiwa, X. Yu, H. Ushiyama, K. Yamashita and J. Son, Chem. Mater., 2016, 28, 1625–1635.
25. B. Jache and P. Adelhelm, Angew. Chem., Int. Ed., 2014, 53, 10169–10173.
26. A. P. Cohn, K. Share, R. Carter, L. Oakes and C. L. Pint, Nano Lett., 2015, 16, 543–548.
27. H. Kim, J. Hong, Y. U. Park, J. Kim, I. Hwang and K. Kang, Adv. Funct. Mater., 2015, 25, 534–541.
28. Z. Zhu, F. Cheng, Z. Hu, Z. Niu and J. Chen, J. Power Sources, 2015, 293, 626–634.
29. B. Jache, J. O. Binder, T. Abe and P. Adelhelm, Phys. Chem. Chem. Phys., 2016, 18, 14299–14316.
30. D. Su, K. Kretschmer and G. Wang, Adv. Energy Mater., 2016, 6 DOI:10.1002/aenm.201501785.
31. Z. W. Seh, J. Sun, Y. Sun and Y. Cui, ACS Cent. Sci., 2015, 1, 449–455.
32. W. S. Hummers and R. E. Offeman, J. Am. Chem. Soc., 1958, 80, 1339.
33. W. Lv, D. Tang, Y. He, C. You, Z. Shi, X. Chen, C. Chen, P. Hou, C. Liu and Q. Yang, ACS Nano, 2009, 3, 3730–3736.
34. C. Zhang, W. Lv, X. Xie, D. Tang, C. Liu and Q. Yang, Carbon, 2013, 62, 11–24.
35. G. Wang, X. Shen, J. Yao and J. Park, Carbon, 2009, 47, 2049–2053.
36. W. Lv, Z. Li, Y. Deng, Q. Yang and F. Kang, Energy Storage Mater., 2016, 2, 107–138.
37. J. Zhang, W. Lv, Y. Tao, Y. He, D. Wang, C. You, B. Li, F. Kang and Q. Yang, Energy Storage Mater., 2015, 1, 112–118.
38. E. Armstrong, D. McNulty, H. Geaney and C. O Dwyer, ACS Appl. Mater. Interfaces, 2015, 7, 27006–27015.
39. D. McNulty, D. N. Buckley and C. O'Dwyer, J. Solid State Electrochem., 2016, 20, 1445–1458.
40. S. Komaba, W. Murata, T. Ishikawa, N. Yabuuchi, T. Ozeki, T. Nakayama, A. Ogata, K. Gotoh and K. Fujiwara, Adv. Funct. Mater., 2011, 21, 3859–3867.
41. X. Bi, X. Ren, Z. Huang, M. Yu, E. Kreidler and Y. Wu, Chem. Commun., 2015, 51, 7665–7668.
42. M. Dahbi, N. Yabuuchi, M. Fukunishi, K. Kubota, K. Chihara, K. Tokiwa, X. Yu, H. Ushiyama, K. Yamashita and J. Son, Chem. Mater., 2016, 28, 1625–1635.

---

Footnotes

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c6ee03367a

‡ These authors are equal main contributors.

---