LIU et al.

Abstract

It is essential to replace lithium‐ion batteries (LIBs) from the perspective of the Earth's resources and the sustainable development of mankind. Sodium‐ion batteries (SIBs) are important candidates due to their low price and abundant storage capacity. Hard carbon (HC) and graphite have important applications in anode materials of SIBs. In this review, the research progress in electrolyte and interface between HC and graphite anode for SIBs is summarized. The properties and performance of three types of widely used electrolytes (carbonate ester, ether, and ionic liquid) with additives, as well as the formation of solid electrolyte interface (SEI), which are crucial to the reversible capacity and rate capability of HC anodes, are also discussed. In this review, the co‐intercalation performance and mechanism of solvation Na⁺ into graphite are summarized. Besides, the faced challenges and existing problems in this field are also succinctly highlighted.

INTRODUCTION

The consumption of fossil fuels and the limited reserves of natural resources bring about environmental concerns and an energy security crisis. Substituting electricity energy with renewable and clean energy electric power is necessary to achieve sustainable development. Solar and wind powers are the most abundant and accessible sources of energy.^1,2 However, they are not stable and reliable, and the changing characteristics of the renewable energy sources pose major challenges to grid loading when renewable electricity energy is directly incorporated into the grid. Another worrying fact is locality of renewable resources, which is often far away from the load centers. Low-cost electrical energy storage is indispensable to eliminating the intermittency of production from renewable sources.³ Energy storage and transformation are particularly important in our life.⁴ Electrochemical energy storage has high efficiency, low cost, and strong adaptability to construct a smart grid, although the existing energy storage is mainly pumped to generate electricity.⁵ Lithium-ion batteries (LIBs) have been widely used in various intelligent electronic equipment. With the wide application of LIBs in electric vehicles and other industries in recent years, their further development is limited by a series of critical problems such as the resource reserve of lithium and the cost of LIBs. Although LIBs have a high energy density, the limited lithium reserves on Earth and high cost make it difficult to meet the large-scale electric energy storage (EES) for application in off-peak power grid regulation and smart grid construction.⁶ Sodium is rich in reserves on the Earth, accounting for 2.8% on the crust of Earth; therefore, sodium-ion batteries (SIBs) are another option for large-scale EES. It is convenient to store electricity, which is generated by solar cells and wind turbines (Figure 1).⁷

Figure 1. (A) Properties of organic electrolyte. (B) Applications of SIBs. (C) Components of SEI. (D) Characteristic of Na+ transport. SEI, solid electrolyte interface; SIB, sodium-ion battery

Sodium-based batteries have been the pioneer of “beyond lithium” technology since 1985, when high-temperature Na/NiCl₂ || Na/S battery was successfully developed by using Na⁺-β-alumina ceramic electrolyte.⁸ Similar to LIBs, SIBs consist of the cathode with sodium insertion materials and anode, which are electrically separated through the electrolyte. Different SIBs can be assembled to meet the performance of the battery by the selection of the battery components.⁹ The basic role of electrolytes is indispensable and independent in various electrochemical devices. The electrolytes provide an ion transferring channel between two electrodes of SIBs, which is the same as other electrochemical devices.¹⁰ For the study of solid electrolyte interface (SEI), which is closely related to electrode surface property, electrolyte composition, and electrochemical performance of the batteries, it started from LIBs.^11,12

The anode materials of SIBs are very important, and carbon-based materials are widely used in the anode of SIBs due to their excellent properties. With the appearance of various carbon-based anode materials, the performance of SIBs has been extensively studied. For example, hard carbon (HC), carbon black, graphite, nitrogen-doped carbon nanofibers,¹³ nitrogen-doped porous carbon fibers,¹⁴ biomass-derived carbon nanoparticles,¹⁵ nano HC,¹⁶ hierarchical 3D carbon-networks/graphene/Fe₇S₈,¹⁷ Fe₇Se₈ nanoparticles anchored on N-doped carbon nanofibers,¹⁸ and sulfur-doped carbon nanosheets¹⁹ have achieved good performance in SIBs. Among all these carbon-based anode materials, HC and graphite are rich in resources and very cheap, so they are very promising for large-scale application in SIBs.

Electrolyte and the interface between anode and electrolyte are closely related to each other, and both are extremely crucial to battery performance. For SIB, in which HC and graphite act as anode materials, there is no review to deeply discuss the electrolyte and interface so far. This review aims at filling this gap and mainly provides detailed and comprehensive research progress for SIBs electrolyte and interface to match HC and graphite anode in recent years. Herein, we first introduce carbonate-based electrolytes for HC anode and the influences of several factors (viscosity, ionic diffusion, and additives) on the performance of HC anode batteries. Next, ether- and ionic liquid (IL)-based electrolytes and their impacts on the performance of SIBs with HC anode are outlined, as well as some high-concentration ether-based electrolytes. In the following section, the formation and evolution process of SEI and its physicochemical properties are described, and moreover, the difference of SEI in SIB and LIB systems is also discussed. Then, we describe the ether solvation co-intercalations for graphite anode, compare the difference of Na ion intercalation into graphite structure in carbonate- and ether-based electrolytes, and analyze the mechanism of Na⁺-ether solvation co-intercalation into graphite. Finally, a personal perspective is presented on the currently faced challenges and existing problems in this field, which are expected to stimulate more exploitation and speed up the development of SIBs.

CARBONATE-BASED ELECTROLYTES FOR HC ANODE

The electrochemical performance of SIBs electrolyte relies on a multitude of parameters, including the composition and proportion of solvents and Na salts. It should be noted that the electrolyte suitable for other kinds of anodes, as a result of the complicated interface between electrode and electrolyte for specific anode materials, does not match the conditions required to achieve the best performance for HC anodes. To improve the electrochemical performance of SIBs, it is necessary to deeply understand the formation mechanism of the interface between HC anode and electrolyte with different compositions.

Solvents and Na salts for HC anode

Carbonate ester electrolytes are extensively applied in SIBs. For the research of electrolytes in HC anode, Komaba et al. achieved a high capacity and outstanding reversibility sodium-insertion performance by using propylene carbonate (PC) electrolyte compared with other carbonate electrolytes for HC anode and layered NaNi_0.5Mn_0.5O₂ cathode²⁰. Dahn et al. used accelerating rate calorimetry (ARC) analysis to demonstrate that dimethyl carbonate (DMC) and diethyl carbonate (DEC) are more reactive with sodium inserted HC than ethylene carbonate (ED)²¹. The same group also found that because cyclic EC has a preferential solvent effect on the high thermal stability of NaPF₆, the incorporation of NaPF₆ in EC/DEC or EC/DMC mixed solvents can enhance the reactivity.²¹ Passerini et al.²² proved that the reaction of Na_x–HC with EC/DMC and EC/DEC is faster than that with EC/PC by investigating various solvent mixtures of EC/DEC, EC/DMC, and EC/PC), as well as NaX salts (X = fluorosulfonyl-(trifluoromethanesulfonyl)imide [FTFSI], TFSI, ClO₄, PF₆, and FSI), and they also demonstrated that the SEI on Na_x–HC anode is mainly composed of double alkyl carbonates (Figure 2).²²

Figure 2. Measurement of CV on aluminum using a three-electrode Swagelok battery, where (A) (a) DEC/EC/NaFSI, (b) DEC/EC/NaFTFSI, (c) DEC/EC/NaTFSI, (d) DEC/EC/NaClO4, and (e) DEC/EC/NaPF6; (B) NaTFSI dissolved in (a) DEC/EC, (b) 5% NaPF6 + DEC/EC, and (c) Py14TFSI; (C) NaFTFSI dissolved in (a) DEC/EC, (b) 5% NaPF6 + DEC/EC, and (c) Pyr14FTFSI; (D) NaFSI dissolved in (a) DEC/EC, (b) 5% NaPF6 + DEC/EC, and (c) Py14FSI. Reproduced with permission: Copyright 2016, Wiley.22 CV, cyclic voltammetry; DEC, diethyl carbonate; EC, ethylene carbonate; NaFSI, sodium bis(fluorosulfonylimide); NaFTFSI, sodium fluorosulfonyl-(trifluoromethanesulfonyl)imide; NaTFSI, sodium bis(trifluoromethanesulfonyl)imide; Py14FTFSI N-butyl-N-methylpyrrolidinium fluorosulfonyl-(trifluoromethanesulfonyl)imide; Py14TFSI, N-methyl-N-butylpyrrolidinium bis(trifluoromethylsulfonyl)imide

Co-solvent candidates of DME, DMC, and DEC were used to improve the performance of EC/PC mixtures. As shown in Figure 3A, EC_0.45:PC_0.45:DMC_0.1 is the optimal composition with high ionic conductivity, low ion pairing degree, low viscosity, and at least a suitable SEI formed on the HC material, exhibiting good electrochemical performance (Figure 3D).²³ Electrolyte containing sodium bis(trifluoromethanesulfonyl)imide (NaTFSI) salt in PC solvent shows good cycling performance for HC/NaNi_0.5Mn_0.5O₂ cells.²⁰ Dahn et al.'s study also indicated that NaTFSI/PC electrolyte in HC/Na half cells exhibits the same capacity and cycle performance compared with NaClO₄/PC electrolyte.²⁵ Ponrouch et al.²⁴ demonstrated that binary EC:PC solvent mixture with NaClO₄ and NaPF₆ salts is the best electrolyte to improve the performance of Na/HC batteries by characterizing ionic conductivity (Figure 3B), viscosity (Figure 3C), thermal and electrochemical stability of different sodium salts (NaPF₆, NaTFSI, and NaClO₄) in single/mixture solvents of EC/PC, EC/DME, EC/DMC, and EC/triglyme. In their study, the good electrochemical performance of HC is ascribed to the high exothermic onset temperature, low reaction enthalpy of NaPF₆–EC:PC, and formation of an electrochemically and thermally stable SEI. The choice of electrolyte plays an important role in the interface stability and electrochemical performance of the HC electrode. Yuan et al.²⁶ found that DME-based electrolytes exhibited much thinner SEI on HC electrodes compared with the EC/DEC-based electrolyte. Mogensen et al.²⁷ used a variety of solvents and NaPF₆ salts to study the effects of electrolytes and additives on SIB performance. Among them, PC had the best performance, and its performance would be further improved when additives, especially vinyl carbonate (VC) additive, were added.²⁷

Figure 3. (A,C) Viscosity and conductivity of electrolyte composed of different solvent/mixtures and 1 M NaClO4.23,24 (B) Conductivity and viscosity of electrolytes composed of a variety of 1 M sodium salts and PC. Reproduced with permission: Copyright 2012, Royal Society of Chemistry.24 (D) Arrhenius diagram of the conductivity of electrolyte composes of different solvent/mixtures and 1 M NaTFSI. Reproduced with permission: Copyright 2013, Royal Society of Chemistry.23 DEC, diethyl carbonate; DMC, dimethyl carbonate; EC, ethylene carbonate; NaTFSI, bis(trifluoromethanesulfonyl)imide; PC, propylene carbonate

The concentration of electrolytes also plays an important role. As revealed by Geng et al., high-concentration NaTFSI-based carbonate ester electrolyte exhibits good performance for HC anode in special conditions, but the ionic conductivity sharply decreases when the temperature is lower than 5°C²⁸. Chen et al. found that 2 M NaTFSI/EC:DMC electrolyte enjoys the best performance for the HC anode compared with other concentration electrolytes (Figure 4).²⁹ The study of Patra et al. suggested that the medium concentration of 3 M NaFSI/PC:EC electrolyte exhibits low flammability, ionic conductivity of 6.3 mS cm⁻¹, CIP and AGG solvation structure, and viscosity of 23 cP, which is advantage to penetrating the separator and HC electrode. The stable SEI with the ability of fast Na⁺ transport, which is mainly composed of (CH₂)_n and NaF organic–inorganic components, is formed and therefore improves the Coulombic efficiency of the HC electrode.³⁰ Rangom et al.³¹ found that the SEI formed in carbonate ester electrolyte at a high charging rate during the initial sodiation can markedly improve the performance of the HC electrode. Hwang et al. designed Na[SO₃CF₃]-Na[N(SO₂F)₂]-Na[N(SO₂F)(SO₂CF₃)] ternary salt system in which a high-concentration electrolyte solution can be formed, and for the samples of different concentrations, 5 M PC shows good performance in HC/NaCrO₂ SIB.³²

Figure 4. Properties of hard carbon spheres electrode at 0.5–5 M NaTFSI electrolyte. (A) Initial discharge and charge diagram. (B) Cycle performance. (C) Rate capacity diagram. (D) Slope-platform capacity diagram. Reproduced with permission: Copyright 2020, Elsevier.29 DMC, dimethyl carbonate; EC, ethylene carbonate; NaTFSI, bis(trifluoromethanesulfonyl)imide

Electrolyte additives for HC anode

The research of Komaba et al.³³ indicates that fluoroethylene carbonate (FEC) additive can improve the electrochemical insertion reversibility for SIBs with HC and NaNi_1/2Mn_1/2O₂ electrodes, and they also demonstrate that the addition of ES and VC causes the detrimental effect to the SIBs. The mixed Monte Carlo/Molecular Dynamics method simulations of the atomic reaction indicate that the PC-based electrolyte with FEC additive promotes the formation of network organic species SEI composed of stable fluorine atoms with large electronegativity, thus exhibiting good capacity reversibility (Figure 5A,B).³⁴ The study on the electrochemical performance of HC anode in PC-NaCl₄ or PC-NaPF₆ salt with FEC additives indicates that PVDF binder in the latter exhibited excellent reversible capacity.³⁷ However, controversy exists in the FEC additive. Ponrouch et al.'s³⁵ study indicates that because more conductive SEI in the free-additive electrolyte is formed, the additive-free PC-/EC-based electrolyte shows excellent rate capability for HC anode compared with the electrolyte containing 2% FEC (Figure 5C,D).³⁵ However, Dugas et al.³⁸ argued that FEC additive was effective in improving the irreversible capacity of the Na-half battery but had few effects on the improvement of electrochemical performance in full SIBs with HC anode. Kim et al.'s³⁶ study confirms that the cycle capacity of symmetric batteries with HC electrodes is not increased by FEC additive, which just improves the cycle performance of Na metal anode and thus reduces the metallic sodium by-products harmful to HC anodes (Figure 5E). Their study also indicates that succinic anhydride (SA) additive improves the cycle performance of batteries with HC anode at 60°C owing to the formation of SEI with abundant sodium alkyl carbonate/Na₂CO₃ on the HC surface (Figure 5F–H).³⁶ In addition, Rb⁺ and Cs⁺ ion additives also can improve the stability and ionic conductivity of the SEI on HC anode for SIBs.³⁹ In the study of Bai et al.,⁴⁰ it was demonstrated that that ester additive in ether electrolyte can form a strong SEI, which therefore improves the electrochemical performance of HC anode. As stated above, there are still problems unsolved. More research is needed to fully reveal how additive affects battery performance, so larger improvement can be achieved.

Figure 5. Schematics of the SEI and PC–NaPF6 electrolytes on the surface of HC (A) without FEC and (B) with FEC. Reproduced with permission: Copyright 2015, American Chemical Society.34 (C) The relationship between potential and capacity and (D) the relationship between discharge capacity and cycle number for HC electrodes in 1 M NaClO4 in PC/EC mixture without (black curve) or with (red curve) FEC additive. Reproduced with permission: Copyright 2013, Elsevier.35 (E,F) Cycling performance of HC battery with SA, FEC, and base electrolytes at different temperatures (25°C and 60°C). Reproduced with permission: Copyright 2019, Elsevier.36 C 1s, F 1s, and P 2p XPS spectra for HC anode (G) before and (H) after 100 cycles at 60°C with succinic anhydride (SA), FEC, and base electrolytes.36 FEC, fluoroethylene carbonate; HC, hard carbon; PC, propylene carbonate; SEI, solid electrolyte interface; XPS, X-ray photoelectron spectroscopy

ETHER- AND IL-BASED ELECTROLYTES FOR HC ANODE Low-concentration ether electrolytes for HC anode

Except for carbonates ester electrolytes, ether-based electrolytes are alternative for HC anodes. The use of ether-based electrolytes was widely reported in LIBs,^41,42 especially Li–S batteries.^43–45 For SIBs, Dutta et al.'s⁴⁶ research suggests that NaClO₄–TEGDME ether-based electrolyte exhibits 97% reversibility after 100 cycles, which is higher than that in NaClO₄–PC electrolyte (70%), and the performance improvement is ascribed to the minimization of the unwanted sodium depositions in the ether electrolyte at low voltage (Figure 6A–D). Zhu et al.⁴⁸ found that the plateau capacity fading at high rates, which occurs in ester-based electrolytes, can be reduced in ether-based electrolytes for HC anode, possibly ascribable to the lower electrochemical polarization of ether-based electrolytes. As suggested by He et al.,⁴⁷ the electrolyte containing 1 M NaPF₆-ethylene glycol dimethyl ether (DME) enjoys higher ion conductivity and Na⁺ ion diffusion coefficient and also forms a thinner but more durable SEI on HC anode compared with ester-based electrolytes (Figure 6E–I), and the ether-based electrolyte in HC anode exhibits high initial Coulombic efficiency and high rate capacity at high mass load.⁴⁷ Morikawa et al.⁴⁹ revealed that ether solvents of DME and THF have better reaction kinetics than carbonate or sulfone, and they also demonstrated that NaBPh₄ salt electrolyte has a stronger ability to stabilize the SEI than other NaFSI, NaPF₆, and NaTFSI salts; therefore, HC anode exhibits excellent electrochemical performance in 0.5 M NaBPh₄/DME electrolyte.

Figure 6. (A–D) CV diagram of Super C-65 electrodes with different electrolytes. Reproduced with permission: Copyright 2017, Elsevier.46 (E) Nyquist diagram and fitting curves. SEM image of (F) ether and (G) ether-based electrolyte. (H) Wettability of ester and ether-based electrolytes to HC anode. (I) Changes in logarithm diffusion coefficient (log D) during charge and discharge. Reproduced with permission: Copyright 2018, American Chemical Society.47 CV, cyclic voltammetry; HC, hard carbon; NaTFSI, bis(trifluoromethanesulfonyl)imide; SEM, scanning electron microscopy; TEGDME, tetraethylene glycol dimethyl ether

High-concentration ether electrolytes for HC anode

To improve the electrochemical performance of the LIBs, a high-concentration electrolyte is often added to the LIBs.^50,51 For SIBs, high-concentration NaFSI-DME electrolyte (5 M) has been demonstrated to have outstanding electrochemical stability, high Coulombic efficiency, and good cycling performance for Na metal, and moreover, it also can improve the oxidation resistance durability of the aluminum substrate.⁵² In Schafzahl et al.'s⁵³ study, high-concentration NaFSI-DME electrolyte (DME:NaFSI = 2:1 mol/mol) shows the ability to improve both sodium metal and HC electrode electrochemical reversibility. Their results also reveal that when the ratio of NaFSI/DME was 0.5, the formed SEI is composed of polymer species, organic ethers, and degradation products of the NaFSI salt, which is different from Na and HC surfaces. Quantum molecular dynamics simulation in kinetic properties analysis of Na⁺ ions for SIBs demonstrates that the solvent/anion exchange reaction is another way for Na⁺ ions diffusion in high-concentration DME/NaFSI electrolyte.⁵⁴ As demonstrated in the study of Takada et al.,⁵⁵ most of the anions coordinate with multiple Na⁺ in a high-concentration electrolyte of 50 mol % NaFSI/butanonitrile (SN) (NaFSI:SN = 1 mol/mol), so the lowest unoccupied molecular orbital (LUMO) of the anions moves down, leading to preferential reduction of the anions which, on HC anode, form an anion-derived passivation film capable to effectively inhibit the further reduction of the electrolyte (Figure 7). So far, ether electrolytes and electrode interface researches are rarely reported; more relative studies are important to understanding and improving the electrochemical performance of the SIBs.

Figure 7. (A,B) Schematics of the equilibrium trajectories (10 mol %, 50 mol %). PDOS profiles on (C) 10 mol % and (D) 50 mol % solutions. Reproduced with permission: Copyright 2017, American Chemical Society.55 PDOS, projected density of states

IL-based electrolytes for HC anode

Traditional organic solvents cannot satisfy the wide-temperature application requirements of the SIBs, which therefore drives the development of a new type of electrolyte. Compared with organic solvents, ILs have many advantages: nonflammability, high ionic conductivity, wide electrochemical window, and electrochemical stability; therefore, they are considered to be a substitute for traditional electrolytes and have been widely studied as electrolytes for electrochemical devices, such as LIBs, electrochemical capacitors, dye-sensitized solar cells, and fuel cells.^56–58

For the research on ILs for SIBs, the cycle test showed that the reversible capacity of HC electrode was 250 mAh g⁻¹ and still maintained 95.5% after 50 cycles at 363 K in Na/NaFSI-[Py_1,3]FSI/HC battery.⁵⁹ Ding et al.⁶⁰ reported that the discharge capacity of HC anode in NaFSI-[Py_1,3]FSI IL electrolyte was 230 mAh g⁻¹ and the capacity retention ratio after 500 cycles was 84% at 90°C (Figure 8).⁶⁰ In another report, NaFSI-[Py_1,3]FSI electrolyte also maintained 90% of its initial capacity after 1000 cycles at 333 K and 363 K for SIBs composed of HC/NaCrO₂.⁶¹ N-propyl-N-methylpyrrolidinium bis(fluorosulfonyl)imide ILs electrolyte with NaFSI or NaTFSI salts could exhibit better cycling stability and higher Coulombic efficiency in three-dimensional carbon framework of graphene nanosheets and carbon nanospheres anode, in comparison with conventional organic carbonate electrolyte.⁶² Hwang et al.⁶³ mixed PC/Na[FSI]/Na[ClO₄] organic electrolytes with [C₃C₁pyrr][FSI] IL to obtain new electrolytes in SIB, which can not only reduce the flammability properties of organic electrolytes but also improve the ionic conductivity of the organic electrolyte and reduce the cost. Although ILs have many advantages mentioned above, the applications of ILs electrolyte at room temperature are limited. The interface capability between electrolytes and electrodes should be deeply studied.

Figure 8. Cycle property of the hard carbon electrode at different temperatures (25°C and 90°C). Reproduced with permission: Copyright 2015, Elsevier60

ELECTROLYTE AND ELECTRODE INTERFACE FOR HC ANODE

Since the SEI model was proposed in 1979,⁶⁴ the research on the morphology, composition, and electrochemical performance of SEI in LIBs has received extensive attention.⁶⁵ The electrolyte will be reduced if the electrochemical potential (μ_A) of anode is higher than the LUMO of the electrolyte. Therefore, a stable SEI is essential to improving rapid ionic conduction from the anode to the electrolyte and preventing the electron from passing through SEI and continuous electrolyte decomposition, if the electrolyte LUMO is not matched to μ_A.⁶⁶

Formation and evolution of SEI on HC anode

As indicated in the study on nano porous carbon–titanium anode, the amount of formed SEI increases with the surface areas of anode materials, and the composites materials, which have more sp² carbon atoms (C–C bonds), tend to produce thick organic oxygen-rich compounds SEI.⁶⁷ Bommier et al.⁶⁸ found that as the cycle progresses, SEI even can penetrate into the HC electrode. It was demonstrated by Li et al.⁶⁹ that precycling of cathode and anode led to preformation of SEI and therefore, improved the long cycling stability of Na_0.44MnO₂-HC full-cell.⁶⁹ Quantum chemical simulations suggest that the high reduction potential and lower barrier for the ring opening of EC make the main contribution to the continuous growth of SEI in SIBs.⁷⁰ The same report also gives the effects of VC and FEC additives: the presence of additive molecules can increase the barrier to EC decomposition; meanwhile, additive and EC molecules can form dimers, which changes the preferred reduction states of EC and leads to different decomposition paths, therefore resulting in the formation of different SEI compounds. The ex-situ characterizations with scanning electron microscopy (SEM), attenuated total reflectance (IR-ATR), and X-ray photoelectron spectroscopy (XPS) indicate that the main components of SEI are sodium ethanedicarbonate (SEDC) and NaF, as well as low-concentration sodium alkyl carbonates, sodium carboxylate, Na₂CO₃, and Na_xPF_yO_z on HC anode.⁷¹ It was also found that the SEI is in dynamic change. In the study from Liu et al.,⁷² the priority of single-electron reductive decomposition of VC, PC, and EC to form the organic SEI components of (CHCHCO₃Na)₂, (CH₃CH₂CH₂CO₃Na)₂, and (CH₂CH₂CO₃Na)₂ was demonstrated to follow the order: VC > PC > EC. However, for the two-electron reductive decomposition of EC, PC, and VC to form Na₂CO₃ products on the anode surface, it follows a different order: EC > PC > VC. Additionally, the two-electron reduction of ethylene sulfite (ES) and 1,3-propylene sulfite (PS) to form Na₂SO₃ is more difficult than the reduction of EC, PC, and VC, while the one-electron reductive decomposition of ES and PS to be organic SEI components is easier than that of carbonate esters. Density functional theory (DFT) calculations in this article show that the decomposition of FEC can yield VC and hydrogen fluoride (HF). Although the combination of F⁻ with Na⁺ can form NaF, it is easier to form NaF by the one-electron and two-electron reduction of FEC (Figure 9). Organic SEI components can be formed easily via the reductive decomposition of VC, ES, and PS additives, while there is a lower difficulty to form inorganic SEI components (NaF) by FEC decomposition.⁷² The reduction decomposition mechanism in SIBs is similar to that in LIBs.⁷³ In Eshetu et al.'s⁷⁴ study, XPS analysis indicates that the inner SEI near the HC anode is richer in inorganic substances and the outermost layer near the electrolyte is mostly organic compounds. The content of organic components in the SEI for sodiated HC electrode depends on the type of used Na salts and follows the order: NaFSI > NaFTFSI > NaTFSI > NaClO₄ > NaPF₆.⁷⁴ Impedance spectroscopy data (EIS), Fourier transform infrared spectroscopy (FTIR), transmission electron microscopy (TEM), and XPS analyses demonstrate that thanks to the large irreversible capacity in the first galvanostatic cycle, the SEI composed of porous organic–inorganic hybrid organic carbonate, Na₂CO₃, and NaF on HC electrode was formed in 1 M NaTFSI-PC electrolyte with 3% FEC additive.⁷⁵ Besides, the physical and chemical properties of SEI become stable after about 5–10 galvanostatic cycles at low current rates, which is attributed to the enrichment of inert inorganic components (Na₂CO₃ and NaF).

Figure 9. Potential energy diagrams of (A) EC, (B) FEC, (C) VC, and (D) ES reduction mechanism. Reproduced with permission: Copyright 2017, Wiley.72 EC, ethylene carbonate; ES, ethylene sulfite; FEC, fluorinated ethylene carbonate; VC, vinylene carbonate

Physical chemistry property of SEI on HC anode

For HC anode, accelerating rate calorimetry (ARC) results shows that sodium inserted in HC reacts with DMC to form sodium methyl carbonate and with EC or DEC to form sodium alkyl carbonates, which have a similar structure to sodium methyl carbonate (Figure 10A).²¹ As revealed by Raman spectra, the solvation shells of Na⁺ are mainly composed of EC, which is related to the fact that EC decomposition products are the main components of SEI (Figure 10A,B).²³ According to the solvation behavior in carbonate-based electrolytes, Cresce et al.⁷⁷ suggested the poor correlation between the Na⁺ solvation preference and the performance of HC electrodes in SIBs, and they also believed that Na⁺ solvation preference may play a secondary role in the SEI formation on HC anode surface. The non-SEI related Coulombic efficiency loss of carbon could be reduced for SIBs by designing carbon with different specific surface areas and synthesizing lower-level trapping defects pseudographitic carbon, as found by Lotfabad et al.⁷⁸ Lee⁷⁹ demonstrated that the SEI, which is composed of NaF, possesses higher shear modulus and the mechanical integrity due to the attraction between F⁻ and Na⁺ from NaO₂CO–R– and Na₂CO₃. Schafzahl et al.⁷⁶ synthesized a series of homologous SEI components of Na alkyl (from methyl to octyl) carbonate and they found that the room-temperature conductivity of these components was lower than 10⁻¹² S cm⁻¹. In addition, with the chain length increasing, the hardness of sodium alkyl carbonate becomes significantly softer (Figure 10C–E).⁷⁶

Figure 10. (A,B) Raman spectra of 1 M NaTFSI in a variety of solvents. Reproduced with permission: Copyright 2013, Royal Society of Chemistry.23 (C,D) The nanomechanical measurements stiffness of the material with AFM. (E) Young's moduli changes of NaACs (■) and LiACs (◆) with alkyl chain length. Reproduced with permission: Copyright 2018, American Chemical Society.76 AFM, atomic force microscope; NaTFSI, bis(trifluoromethanesulfonyl)imide

Comparison of SEI on the HC anodes for SIB and LIB

The SEIs of SIB and LIB systems have significant differences in compositions and properties. Iermakova et al.⁸⁰ demonstrated that the composition of SEI evolves with the Na electrode cycle and the SEI on HC electrode exhibits lower stability (partial solubility) for SIB compared with LIB. As shown in the study of Eshetu et al.,⁸¹ these factors (including high-solubility Na-SEI components, products from evolution, large size of Na⁺, lower Lewis acidity, and increase in Na⁺ reduction potential) may lead to completely different SEIs on HC anode for SIBs compared with LIBs. Mogensen et al.⁸² used XPS to demonstrate that the dissolution of a large amount of SEI leads to the increase of self-discharge for SIBs, and they also proved that since the SEI in SIBs is inferior to the SEI in LIBs in terms of self-discharge, the capacity loss of SIBs is very fast compared with that of LIBs (Figure 11A–F). The contrastive study from Eshetu et al.⁷⁴ shows that even when treated with the same electrochemical process, the SEIs in SIB are more disadvantageous than their counterparts in LIB: the former have more organic substances and Na-contained compounds (inorganic Na salt and Na-alkoxide intermediates) with higher solubility than that of similar Li compounds, thereby reducing the stability of SEI during long-term cycling of HC electrode. The different chemical properties of Na and Li may be related to ion radius, cation solvation, reduction potential, chemical/electrochemical reaction, solubility of SEI, and other parameters. Soto et al.'s⁸⁴ research on SEI formed on HC electrode indicates that Na-based SEI transports lithium ions more easily relative to Li-based SEI. Soto et al.⁸³ performed ab initio molecular dynamics (AIMD) simulation on HC–electrolyte interface and the transport mechanism of Na⁺, which demonstrates that followed by the formation of NaF via the decomposition of other compositions, the Na₂CO₃ components were formed due to the decomposition of EC at the edge of HC structure.⁸³ Besides, it was observed that owing to the larger ion size of Na compared with Li, the transportation of Na⁺ ions in the Li-based SEI component follows a cooperative exchange mechanism. However, the Li⁺ ions can be transported in the Na-based SEI component through interstitial, owing to their smaller radius (Figure 11G–J).

Figure 11. (A) Voltage changes over time of LIBs. (B) The charge and discharge capacities for LIBs. (C) Voltage changes over time of SIBs. (D) The charge and discharge capacities for SIBs. (E) Voltage increases for LIBs and SIBs during the first 100-h pause. (F) Calculation of the capacity loss of the batteries in (E). Reproduced with permission: Copyright 2016, American Chemical Society.82 Potential energy curves of ion migration mechanisms in (G) LiF, (H) NaF, (I) Li2CO3, and (J) Na2CO3. Reproduced with permission: Copyright 2018, American Chemical Society.83 LIB, lithium-ion battery; SIB, sodium-ion battery

ETHER SOLVATION CO-INTERCALATIONS FOR GRAPHITE ANODE Reasons for unfavorable Na⁺ intercalation into graphite

Intercalating lithium into graphite can yield binary graphite intercalation compounds (b-GIC) with a stoichiometry of LiC₆, which are mostly used in commercial LIBs as anode material, and moreover, this process is reversible.⁸⁵ Most alkali metals (e.g., Li, K, Cs, and Rb) can form b-GICs, but Na is an exception, which is possibly caused by the mismatch between the graphite interlayer spacings and the size of Na ions.^86,87 DFT calculations show that Na/Na⁺ has a higher redox potential than Li/Li⁺, so it can be explained why Na cannot form the stage I GICs.⁸⁸ As revealed by X-ray diffraction and Raman spectroscopy analyses in a previous report,⁸⁹ only low-stage Na-GIC near the graphite surface is formed when the cut-off potential is higher than the metal sodium deposition potential, and high-grade Na-GIC in the graphite body can be formed when reaching the deposition potential of sodium metal. In the same report, chronopotentiograms and potentiostatic intermittent titration analyses indicate that Na⁺ diffusion in graphite body is determined by thermodynamic limitation but not kinetic limitation. Lenchuk et al.⁹⁰ found that the thermodynamic instability of Na-GIC has no dependence on the concentration of Na atoms. From the perspective of the energetics and the structure of Li, Na, and K, Na-GIC has the largest binding energy. Because the size of Li is smaller than that of Na, the former has smaller binding energy. Moreover, van der Waals force between graphite layers will be interacted, thus improving the stability of Li-GIC. However, for Na, the required energy for structural deformation is high, and moreover, the van der Waals force between graphite layers is almost absent. Interestingly, Moriwake et al.⁹¹ used DFT calculations to study Cs-, Rb-, K-, Na-, and Li-GICs, and they thought that the ionic bond between carbon atom and alkali-metal ion leads to stability difference of these alkali-metal-GICs and weakens as the size of alkali-metal ion decreases. However, although Li-ion is the smallest in size, it can bond with C atoms via a covalent component, resulting in higher stability. Therefore, K and Li successfully form stable GICs, but Na cannot.

The origin of ether solvation co-intercalation for alkali metal

In 2014, Adelhelm et al.⁹² found that graphite can be effectively used as an electrode material for SIBs. It shows not only good reversibility under low overvoltage but also superior cycle life in ether electrolytes. The reversible ternary graphite intercalation compounds (t-GICs) are formed via solvation alkali ions “co-intercalation” presented in the following formula: [Image Omitted. See PDF]where C_n, A⁺, solv, and y represent the number of carbon atoms in graphite lattice, metal ion, solvent molecule, and solvent molecule number, respectively.^93,94 The stage-I ternary solvent co-intercalation compound of Na(diglyme)₂C₂₀ was estimated to be formed in diglyme-based electrolytes. The cathode reaction has superior performance in the following aspects: high energy efficiency, small irreversible loss in the first cycle, and the capacity of 100 mAh g⁻¹ after long 1000 cycles with Coulombic efficiency >99.87% for graphite anode. Asher firstly carried out the experiment of sodium graphite-layered compounds in history, and he believed that sodium intercalation compounds can be formed although the graphite intercalation compounds had low stability.^95,96 In 1996, XRD was employed to characterize the GIC structures by using Li, Na, K, Rb, and Cs alkali ions in five solvents of l-methoxy-butane (MB), 2,5-dimethyltetrahydrofuran (diMeTHF), DME, 2-methyltetra-hydrofuran (MeTHF), and tetrahydrofuran (THF). When using THF and DME solvents, regardless of the type of alkali metal, the ternary GIC via co-intercalation of alkali metal and organic solvent into graphite is obtained.⁹⁷ In 1998, the study on the ether and sodium cations co-intercalation effect in different graphitization degrees of poly (vinyl chloride) coke demonstrates that it is strongly controlled by the degree of graphitization compared with other alkali metals, depending on the bond strength between Na and ether solvent molecules and the volume of Na⁺ solvation complexes; besides, this report also suggests that co-intercalation of solvents with large molecule volume into graphite is harder than that of solvents with low molecule volume.⁹⁸ It was reported that the cokes degradation is caused by the co-intercalation of alkali metal ions and ether solvent molecules.⁹⁹ Mizutani et al.'s¹⁰⁰ study demonstrates that heavy alkali metals of Rb and Cs in cyclic ethers might be co-intercalated into graphite, while linear ethers solvation with light Li metals is easy to form ternary Li-solvent-GICs, for the solvation co-intercalation layer is affected by the solvation ion radius and the interaction between alkali metals and solvents. According to the study of Kim et al.,¹⁰¹ solvated alkali ions of Li, Na, and K are intercalated into graphite in DEGDME electrolyte (Figure 12), and the graphite interlayer distance is increased with the cation radius increase from Li to K, leading to the decrease of the repulsive force between graphene layer; that is, higher the alkali metal ions storage potential, higher discharged state stability. Zhu et al.'s¹⁰² study indicates that alkali metals (Li, Na, K) and intercalation graphite from reduced polyether and electrocatalysts can form new GICs. In addition, crown ether and Na (or K) co-intercalated t-GIC were synthesized by exchange with ethylenediamine or direct reaction with electrocatalyst and stage-1 co-intercalated double-layer GIC products are formed.¹⁰³ Yoon et al.¹⁰⁴ proved that the instability of Na-GIC is mainly caused by the disadvantageous local Na–graphene interaction, but stability can be effectively improved by screening Na ions with solvent molecules. Under the specific conditions of the electrolyte, in terms of the Na solvation energy and the LUMO level of the complex, it is possible to reversibly intercalate Na into graphite.

Figure 12. Charge and discharge curves and ex-situ X-ray diffraction spectra of graphite in (A) lithium-ion (B) sodium-ion battery, and (C) potassium-ion battery. (D) Schematic diagram of the process of alkali ion-solvent co-intercalating in graphite. Reproduced with permission: Copyright 2016, Royal Society of Chemistry101

Performance of Na⁺-ether co-intercalation graphite

Kim et al.¹⁰⁵ reported the Na storage and pseudocapacitive behavior of Na⁺-solvation co-intercalation in graphite by using high donor numbers ether-based electrolytes of DME, DEGDME, and TEGDME.¹⁰⁵ In their study, a reversible capacity of 150 mAh g⁻¹ was obtained for natural graphite electrodes, and its voltage can be varied between 0.6 and 0.78 V (vs. Na) by changing the chain length of linear ether solvent. Their research also shows that the type of electrolyte solvent markedly affects the oxidation/reduction potential and rate capability of graphite anode (Figure 13A–D). Zhu et al.'s¹⁰² study also indicates that Na/graphite battery using ether-based electrolyte showed excellent cycle performance within 6000 cycles at a high rate capability of 10 A g⁻¹. Glymes and the derivatives were used for exploring the effects of electrolyte solvents on Na⁺ solvation co-intercalation graphite electrodes of SIBs, and it is shown that ether-based electrolyte can activate graphite for SIBs and promote forming the Na⁺ solvation co-intercalation reversible t-GIC with high chemical stability. The results also show that redox potential changes with ether chain length, the mixing of ethers can realize the customization of redox behavior and that the unsatisfactory Na⁺ coordination triglyme may be the reason for its poor electrochemical behavior. Although graphene layer spacing is changed almost 250% during solvation ion co-intercalation, the mechanical stability is still maintained; chain length change slightly affects layer spacing, and a favorable charge transfer kinetics may be ascribed to the absence of SEI (Figure 13E).¹⁰⁶ The use of ether-based TEGDME-NaCIO₄ electrolyte in high-power SIB with graphite anode and layered P₂-Na_0.7CoO₂ cathode was reported, and the 45% capacity retention is obtained after 1000 cycles at the rate of 10 C, meaning that charging/discharging is just 6 min.¹⁰⁷ When homologs glymes (G_x) with O_{x + 1} atoms (x = 1–4) of linear ethylene glycol dimethyl ether was used as the solvent for the co-intercalation of Na⁺(G_x)_y-complexes in the graphite lattice, XRD analysis indicates that an intermediate stage 2Na–GIC (NaC₄₈) is formed for partial sodiation of the graphite electrode in the four glyme (G1–G4) solvents. In the complete sodiation stage, G1, G2, and G4 all form 1Na–GIC (NaC₁₈, 112 mA g⁻¹), but the G3 solvent system forms first-stage 1Na–GIC with less Na incorporated (NaC₃₀, 70 mAh g⁻¹); moreover, the formation of an SEI on graphite surface was also observed.¹⁰⁸ Maibach et al.¹⁰⁹ used soft X-ray photoelectron spectroscopy (SoXPES) to study the graphite anode co-intercalate in (TEGDME/NaFSI) electrolyte in SIBs, and they found that SEI formed in the electrode is thin and dynamic. Additionally, the SEI will break when forming Na-t-GIC. Goktas et al.¹¹⁰ studied the effect of temperature on co-intercalation of Na⁺ on graphite in linear ethers (monoglyme, diglyme, triglyme, tetraglyme, pentaglyme) in SIB, and they found that the performances of pentaglyme and triglyme are poor at room temperature but get better when the temperature rises. Compared with other glymes, triglyme is the worst in performance.¹¹⁰

Figure 13. (A) X-ray photoelectron spectroscopy spectra and (B) Raman spectra of sodium content in natural graphite, indicating that sp3-defects are formed in the natural graphite after the discharge and sp3 defect disappear in the later charging process. (C) High-resolution Raman spectra of natural graphite during charge and discharge process. (D) Ex-situ X-ray diffraction spectra of natural graphite in the cycle. Reproduced with permission: Copyright 2015, Wiley.105 (E) Electrode reaction model of solvent co-intercalating: diglyme molecules coordinate with alkali ion (Li or Na) through base ion–oxygen interaction. Reproduced with permission: Copyright 2016, Royal Society of Chemistry106

Mechanism of Na⁺-ether solvation co-intercalation graphite

In 2011, it was reported that Na⁺ and linear alkylamine co-intercalation graphite could form t-GIC.¹¹¹ A new GIC with intercalation arrangements and specified stages as follows was obtained: stage 1, monolayer (C3, C4); Stage 1, bilayer (C6, C8); and stage 2, bilayer (C12, C14). The new feature of the donor GIC is an intercalation double-layer arrangement, which parallels the alkyl chain surrounding the graphene layers.¹¹¹ The insertion of Na in graphite was reported to be a multi-stage reaction, in which the first stage GIC with excellent reversibility is formed when varying C/Na ratio from 28 to 21. As described by Kim et al.,¹¹² the repulsion between positive Na⁺ is screened by the longer linear solvent molecules, and Na intercalation potential increases with solvent chain length, indicating the possibility to adjust Na storage characteristics. Na(DME)₂C₂₆ prepared via Na⁺-DME co-intercalation in highly-oriented pyrolytic graphite (HOPG) was studied by Guan et al.,¹¹³ and they found that solvation co-intercalation t-GIC structure undergoes a series of evolutions from Phase 3, Phase 2, and then to Phase 1 transition. Their study also indicates that the interaction between the graphite layer and DME ensures the solvated Na⁺ co-intercalation graphite structural integrity. The t-GIC is partially degraded into sodium alkyl carbonate, Na₂CO₃, and graphite after being stored in the air for 14 days, and its structural degradation process is opposite to that of intercalation. The performance evaluation of solvated Na-DME co-intercalation graphite shows that solvated Na-DME can co-intercalate in the graphite layer without affecting the structural stability of co-intercalation graphite.¹¹³ Jung et al.¹¹⁴ studied what factors led to the good rate and cycle performance for solvation Na⁺ co-intercalation graphite and the poor performance in the case of Li⁺ solvent; their research shows that the diglyme–Na⁺ co-intercalation in graphite exhibits stable mechanical integrity owing to the van der Waals interaction between diglyme and grapheme. They also demonstrated that diglyme solvation-Na⁺ complex has a fast Na⁺ co-diffusion in the interlayer of graphite, which is equivalent to Li⁺ diffusion and five orders of magnitude higher than Li⁺-diglyme co-diffusion. The solvation co-intercalation graphite with dimethyl sulfoxide (DMSO) was studied by quantum mechanics meta-kinetics simulation, and the result indicates that the formation of 3 or 4DMSO-Na⁺ solvation complexes enjoys the minimum free energy. It is important for solvents to shield sodium in the insertion process of Na⁺ in graphite; meanwhile, adjusting the interaction between the solvent and the graphite sheets, as well as the anions and cyclic ether, can lead to a higher rate of anion exchange, which increases the mobility of Na⁺.¹¹⁵ Goktas et al.¹¹⁶ proposed that the electrode interface reaction of solvation Na⁺ co-intercalation in graphite is very different from the traditional graphite intercalation reaction, and the products from decomposed diglyme are volatile and/or soluble and do not form SEI in graphite anode interface. The mechanism of solvation Na⁺ co-intercalation in graphite was studied using different electrolytes with different ether chain lengths (DME, DEGDME, TriGDME, and TEGDME are named as O2, O3, O4, and O5, respectively) by in-situ XRD and in-situ dilatometry. It is indicated that the volume expansion of graphite electrode reaches 100%–130% in the first sodiation process and the volume change is related to the type of solvent, but the distance between graphite layers slightly decreases before complete sodiation to the cut-off for the O5 and O2 electrolyte. The solvation Na⁺ co-intercalation behavior for [Na-TriGDME]⁺ complex is hindered owing to the improper coordination geometry between TriGDME and Na⁺, but reversibility of graphite volume change and electrochemical performance for O4 TriGDME electrolyte is better than that of O3 and O5 electrolyte at a higher temperature of 60°C (Figure 14). The results also show that the morphology of the solid deposit layer on the electrode surface changes with temperature and electrolyte but has no direct correlation with the chain length. In addition, it is possible that the SEI after sodiation is dissolved in the subsequent desodiation process.¹¹⁷ Wang et al.¹¹⁸ found that the organic in SEI can adapt to the volume change of graphite during the cycle, and the inner inorganic in SEI can increase Young's modulus of SEI. Even though there is a large volume change in the ether electrolyte, the battery still has good performance.¹¹⁸ So far, the explanation of SEI formation/evolution in co-intercalation process remains controversial. Although some solvent co-intercalation mechanisms were proposed, more research about co-intercalation, interface reaction, and SEI formation should be focused in the future.

Figure 14. Performances of different electrolytes during the first five cycles at 25°C and 60°C: (A,D) Columbic efficiency; (B,E) sodiation capacity; (C,F) irreversible thickness change. Reproduced with permission: Copyright 2019, Elsevier117

OUTLOOK AND PERSPECTIVE

Nowadays, SIBs have been widely studied because of their excellent performance and low price for application in large-scale energy storage. In this article, we have summarized the research on the electrolyte and interface of HC and graphite anodes for SIBs in recent years. Although there have been more and more research on SIBs in recent years, there are many problems such as unsuitable electrolytes and additives, low capacity, low energy density, poor cycling life and rate performance, safety issue, and low initial Coulombic efficiency. It will not be far from the commercialization of SIBs if these problems are solved. HC and graphite, as excellent carbon-based anode materials, have received more and more attention. The compatibility of electrolytes, additives, and anode materials is a very important factor that cannot be ignored. Paying more attention to the study of match degree between electrolyte and electrode is crucial to improving battery performance. Many researchers have used ester electrolytes in SIBs, which exhibit good performance. Sometimes, when the concentration of the electrolyte increases, the performance of the battery can be improved; however, when the concentration is higher than a certain value, the performance will decrease instead. And the situation of a high concentration of ester and ether electrolytes is different. The safety of the battery is vital to its use and the popularity of the market. Although organic electrolytes (esters and ethers) have good performance, their safety still needs to be improved, especially when they are used at high temperatures. Therefore, developing new IL electrolytes with high thermal stability, electrochemical stability, and good performance is also a solution; but the low ionic conductivity and high viscosity of ILs restrict its further application, so further research is needed to overcome these problems. The SEI is also very important for SIBs, and the choice of electrolyte and additives will significantly affect the internal performance of the SEI. The properties (thickness, composition, uniformity, etc.) of SEI can affect the overall performance of HC anode and SIBs. Some additives can significantly improve battery performance, but in some cases, additives make the performance of the battery worse. Besides, the formation of binary graphite intercalation compounds (b-GIC) via intercalating sodium into graphite is irreversible, and b-GIC is mostly used in commercial LIBs as anode material. Ethers solvents Na⁺ co-intercalation graphite still exhibits good performance after thousands of cycles, which shows that graphite anode possesses great potential for application in SIBs. We believe that these problems will be solved under researchers' continuous efforts; commercial applications of SIBs will eventually come for large-scale energy storage.

ACKNOWLEDGMENTS

This project is financially supported by the International Science & Technology Cooperation of China under 2019YFE0100200, National Natural Science Foundation of China (Grant No. 51902024), Beijing Institute of Technology Research Fund Program for Young Scholars, the National Postdoctoral Program for Innovative Talents of China (BX20180038), China Postdoctoral Science Foundation (2019M650014), and Beijing Natural Science Foundation (L182022).

CONFLICT OF INTEREST

The authors declare no conflict of interest.