Interlayer Design of Pillared Graphite by Na-Halide Cluster Intercalation for Anode Materials of Sodium-Ion Batteries.

Graphite is currently utilized as anode materials for Li-ion batteries, but it is well-known that graphite does not show good electrochemical performances as the anode material for sodium-ion batteries (SIBs). It was also reported that the low electrochemical performances of graphite originated from the larger ionic radius of the sodium ion due to the required higher strain energy for sodium-ion intercalation into graphite leading to an unstable sodium-ion intercalated graphite intercalation compound (GIC). In this work, using first-principles calculations, we introduce pillaring effects of Na n X (n = 3 and 4; X = F, Cl, or Br) halide clusters in GICs, which become electrochemically active for Na redox reactions. Specifically, to enable sodium-ion intercalation into graphite, the interlayer spacing of graphite is required to increase over 3.9 Å, and Na n X halide cluster GICs maintain an expanded interlayer spacing of >3.9 Å. This enlarged interlayer spacing of Na n X halide cluster GICs facilitates stable intercalation of sodium ions. Na3F, Na4Cl, and Na4Br halide clusters are identified as suitable pillar candidates for anode materials because they not only expand the interlayer spacing but also provide reasonable binding energy for intercalated sodium ions for reversible deintercalation. Based on the model analysis, theoretical capacities of Na3F, Na4Cl, and Na4Br halide cluster GICs are estimated respectively to be 186, 155, and 155 mA h g-1. These predictions would provide a rational strategy guiding the search for promising anode materials for SIBs.

Introduction

With rapidly increasing demands for portable electric devices, household electronics, and electric vehicles (EVs), high-efficiency battery energy storage systems are under extensive research and development efforts. In addition, large-scale energy storage systems (ESSs) have recently received great attention owing to the rapid progress of renewable energy production such as wind, tidal, and solar power generation, which require a grid-scale energy storage support. Among the suitable energy storage systems for these applications, lithium-ion batteries (LIBs) have been the leading energy storage systems with excellent electrochemical properties of high energy density, long life cycles, and high energy storage efficiency.[1,2]

Although the LIBs have been used as the primary energy storage devices with increasing commercial demands and performance advantages, there are concerns over the material limitations of LIBs due to rapidly growing demands and limited reserves of key materials (i.e., Li and Co), distributed in specific areas.[3] Furthermore, the basic battery structure of LIBs has remained the same as the initial Sony 1992 design of the graphite anode, organic liquid electrolyte, and layered oxide cathode (LiCoO2; LCO) with incremental improvement of cathode materials over the last three decades. The current commercial LIB uses Ni-rich layered oxide cathode materials (e.g., Li(Ni0.8Co0.1Mn0.1)O2 or Li(Ni0.8Co0.15Al0.05)O2; NCM811 or NCA) for higher capacity (∼200 mA h g–1) than LCO (∼140 mA h g–1), and this capacity increase reaches the material limits at an ∼90% Ni-rich NCM cathode. To overcome the material limits of LIBs, sodium-ion batteries (SIBs) are recently considered as a promising alternative to LIBs with great abundance of Na reserve and their similar electrochemical behavior to LIBs with Na+ replacing Li+.[4] For these reasons, many recent research studies have been focused on the development of high-capacity electrodes of SIBs by utilizing the mature system architecture of LIBs.

Initially, both cathode and anode materials used in LIBs are applied to SIBs with limited successes. Layered oxide cathode materials are shown to have a lower voltage of ∼3.5 V compared to ∼4 V in LIBs and also smaller charge storage capacity due to larger Na+-ion size. Alternative cathode materials are investigated for SIB applications.[5−8] Furthermore, graphite as the anode materials in SIBs has much more serious challenge due to negligibly small electrochemical activity with sodium ions. Graphite used as an anode material of LIBs stably exhibits a reversible high capacity of 372 mA h g–1 and long cycle life compared to the other anode materials in LIBs.[9] However, as an anode material of SIBs, graphite just delivers less than 35 mA h g–1 capacity even though Li ions and sodium ions are similar alkali ions. It has been reported that a graphite interlayer spacing of 3.4 Å is insufficient for a stable Na+-ion intercalation due to the relatively larger ionic radius of the sodium ion (1.02 Å) compared to the smaller ionic radius of the Li ion (0.76 Å), which requires an extra strain energy to expand interlayer spacing in accommodating sodium-ion intercalation.[9,11] In addition, Na+ intercalated in the graphite has also particularly lower electrochemical stabilization with the substrates than the other alkali ions (e.g., Li+ and K+) regardless of the ionic radius.[11]

To enable the Na+ intercalation into graphite as an active anode material for SIBs, there have been many attempts to modify Na+ electrochemistry with graphite. For example, thermodynamic stabilization by ion-solvent cointercalation in graphite intercalation compounds (GICs) is shown to intercalate a Na+-solvent complex primarily driven by large organic solvent molecule intercalation into graphite interlayers.[10] Similarly, cointercalation of sodium and ether-based electrolytes into graphite exhibits a relatively high capacity and stable cycle ability by facilitating thermodynamically stable sodium-ion intercalation without modifying pure graphite.[14−16] Even though the ion-solvent complex can be intercalated into graphite interlayers, the kinetics and capacity would be limited by large solvent molecules, and the solvent intercalation is also known to induce liquid-phase exfoliation by excessive solvent intercalation.[17] On the other hand, expanded graphite by modifying the atomic structures of graphite by oxidation into graphite oxide (GO) was introduced to enlarge interlayer spacing of GO by Wang et al.[9,13] The GO has shown significantly increased charge capacity, but the electrochemical voltage was also significantly increased up to 1.5 V (compared to <0.3 V for Li intercalation in graphite) due to stronger Na-ion interaction with oxygen functional groups in GO with 10 or higher at. % oxygen. Considering these previous approaches, a controlled increase in interlayer spacing, low electrochemical potential, and thermodynamic stabilization of intercalated layers are crucial for sodium-ion intercalation into graphite as a viable anode material for SIBs.[12,18]

In this study, we report a systematic and quantitative theoretical investigation on the possible pillaring strategy of increasing the graphite interlayer spacing, maintaining low electrochemical potential, and ensuring the stability of Na intercalated graphite against exfoliation. To avoid the graphite exfoliation induced by organic or neutral intercalation species,[20] possible roles of electropositive (e.g., Li, K, and Ca) and electronegative (e.g., F, Cl, and Br) cointercalants are investigated for Na intercalation in graphite. Electronegative cointercalants are found to form strongly bonded planar cation clusters with Na, NaX (n is the number of sodium ions and X = F, Cl, or Br), which can function as a pillar to increase the graphite interlayer spacing, suppress exfoliation, and facilitate Na intercalation in the pillared graphite as a promising anode material for SIBs. We note that previous studies on GICs have shown that a large amount of halide and alkali metal can cointercalate into graphite and form salt graphite with large interlayer spacing (6–10 Å or even larger), but these studies are limited to full occupation of interlayer space by multiple atomic layer salt compounds (e.g., Na2ClC13 and KClC16).[19−21] In addition to these ternary GICs, such as Na-halide-graphite and K-halide-graphite, other cointercalated alkali-anionic elements GICs (e.g., Na-O, K-O, and K-sulfur (S) GICs) and alkali-selenium (Se)/tellurium (Te) GICs (K-Se and K-Te GICs) have been also introduced.[22,23] Considering the cointercalation of alkali metal and halide into the GIC, herein, the potential for the GICs as an anode material was examined by first-principles calculation. Such fully occupied interlayer galley space has Na ions in strongly bonded salt layers and would not be able to provide reversible electrochemical storage of Na ions as an active anode material. Through careful atomic and electronic structure analyses, we found that an at least 3.9 Å interlayer is required for thermodynamically stable intercalation of sodium ions into graphite layers and that NaX halide clusters could increase interlayer spacing over 3.9 Å enabling the Na electrochemical activity. Specifically, we have confirmed that the pillared graphite with NaX halide clusters make it possible for sodium ions to thermodynamically intercalate into the graphite interlayers.

Computational Details

The atomic and electronic structures of the proposed model pillared graphite systems were systemically analyzed based on density functional theory (DFT) calculations.[24] We used the Vienna ab initio simulation package (VASP) with spin-polarized generalized gradient approximation (GGA) parameterized by the Perdew–Burke–Ernzerhof (PBE) exchange correlation functional.[25] For the calculation of Na-ion intercalation into pure graphite, we used a supercell of 3×3×1 unit cells of graphite with Monkhorst–Pack 4×4×4 k-point sampling, while a supercell of 6×6×1 unit cells and 2×2×4 k-point sampling were used for the Na-ion intercalation into graphite pillared by NaX halide clusters. The plane wave basis set cutoff energy of 450 eV was used. All examined materials were fully relaxed to thermodynamically stable atomic structures including van der Waals interactions.[26]

Results and Discussion

It is well-known that sodium ions do not intercalate into pure graphite even though other alkali metals, lithium and potassium, can intercalate into pure graphite.[11] The graphite’s initial AB stacking is known to transform to AA stacking by Li-ion intercalation,[26,27] and Na-ion intercalation into pure graphite is known to be thermodynamically unstable due to relatively high strain energy to expand the interlayer of graphite. The process of cation intercalation into graphite can be examined by two main reaction steps. The first step requires strain energy to expand the interlayer spacing according to the cation size, and the second step is electrochemical stabilization by the interaction between intercalated ions and expanded graphite (e.g., charge exchange between alkali ions and graphite). These two steps are endothermic and exothermic processes, respectively.[24] In other words, the GICs can be thermodynamically stable by stabilizing reaction with intercalated ions and graphite, which offsets the energy increase required for the interlayer expansion of graphite. First, we considered pre-expanded graphite without requiring strain energy and investigated the intercalation energy as a function of the pre-expanded interlayer spacing.

To examine the energetic features of sodium-ion (Na+) intercalation into graphite and the correlation between sodium-ion intercalation and interlayer expansion of graphite, we investigate the formation energies of Na intercalation into pre-expanded graphite layers as a function of the fixed interlayer distance. The formation energy is calculated as Eformation = Etotal(Na-GIC) – nEtotal(Na) – Etotal(graphite), plotted in Figure . Herein, Etotal(Na-GIC) is the total energy of the sodium-ion GIC, Etotal(Na) is the energy of the sodium atom, Etotal(graphite) is the energy for the 3×3×1 unit cell of graphite as each pre-expanded initial interlayer distance, and n is the number of sodium ions in each GIC. To consider the correlation between the number of intercalated sodium ions and required interlayer distances for intercalation, we examined three model compounds of Na1GIC, Na3GIC, and Na6GIC based on the AA intercalation mechanism.[25] We calculated the formation energy as the different initial interlayer spacing of graphite ranging from 3.2 to 5.0 Å as shown in Figure . Figure a shows that one sodium-ion intercalation is thermodynamically stable in pre-expanded interlayer spacing of at least 3.9 Å of the slabs while unstable for the interlayer spacing less than 3.9 Å. In addition to the case of one sodium ion, three and six sodium ions (per model supercell) intercalated into pre-expanded graphite were thermodynamically stabilized at the interlayer spacing over 4.2 and 4.4 Å, respectively. As the interlayer spacing further increases over 4.2 and 4.4 Å, the formation energy, Etotal(Na-GIC), is more stabilized as shown in Figure b,c.

Figure 1

Formation energy of sodium-ion intercalation into expanded graphite as variations of interspacing for 1 mol (a), 2 mol (b), and 3 mol (c) of sodium ions. Blue areas mean thermodynamically stable states.

Subsequently, these pre-expanded model structures are relaxed, and the GICs optimize the initial interlayer spacing of 3.2–5.0 Å to stabilized values for three model structures. The average values of stabilized interlayer spacing for three models (Na1G, Na3G, and Na6G) are 4.39, 4.48, and 4.49 Å as shown in Figure S1. All values of interlayer distances for relaxed structures tend to converse to the average values regardless of starting values of the initial interspacing (with some variations due to weak interactions across the van der Waals interlayer gap). The overall convergence of relaxed interlayer spacing to 4.4–4.5 Å regardless of sodium-ion contents (n = 1, 3, and 6 in the model structures) indicates that sodium-ion intercalated GICs have similar optimal interlayer spacing, which is ∼1 Å larger than intrinsic graphite interlayer spacing for thermodynamically stable intercalation of large sodium ions into graphite. These results are consistent with the expectation that sodium-ion intercalation could be facilitated by increasing interlayer spacing of pure graphite.[13]

To explore a design concept of interlayer pillaring for sodium-ion intercalation, we consider GICs with expanded interlayer spacing by preintercalation of other cations before sodium-ion intercalation. Lithium (Li), potassium (K), and calcium (Ca) are known to thermodynamically intercalate into pure graphite.[28−31] The preintercalated GICs by Li, K, and Ca exhibit enlarged interlayer spacing at thermodynamically stable states, as shown in Figure (note (b) n = 0). Based on these pre-expanded GICs as candidate anode materials for SIBs, sodium-ion intercalations were simulated. Pre-expanded GICs indicate different interspacings according to different cations: 3.59 Å for Li, 5.32 Å for K, and 4.40 Å for Ca. Based on the analysis in Figure and Figure S1, K and Ca seem to be promising preintercalants to expand the graphite interlayer spacing. Agreeing with this expectation, the interlayer spacings of K or Ca preintercalated graphite do not change by Na intercalation, but the interlayer spacing of Li preintercalated graphite increases toward 4.4 Å with Na intercalation (see Figure b). However, the sodium-ion intercalations into preintercalated GICs of Li, K, and Ca are all thermodynamically unstable as shown in Figure a with positive formation energies Eformation = Etotal(Na-M-GIC) – Etotal(Na) – Etotal(Na-M-GIC). In the equation, Etotal(Na-M-GIC) is the total energy of the M (M = Li, K, and Ca) ion intercalated GIC at varying sodium-ion contents. The Li and K preintercalated graphite interlayer spacings are too small or too large (3.59 or 5.32 Å compared to 4.4–4.5 Å, optimal spacing for Na) so that the subsequent Na intercalation cannot be stabilized by the expanded interlayer spacing determined by preintercalants. Even though Ca preintercalated graphite exhibits the lowest formation energy, it is still unstable to Na intercalation despite the optimal 4.40 Å interlayer spacing determined by the Ca preintercalant. This unstable formation energy is due to the increased repulsive Coulomb interactions of Ca2+ cations with Na+ cations.

Figure 2

Formation energies (a), interlayer distances (b), and atomic structures (c) of Li, K, and Ca ion preintercalated GICs as sodium-ion content changes. Blue areas mean thermodynamically stable states.

Considering these intrinsic properties of sodium intercalation into graphite and expansion of interlayer spacing by introducing preintercalating elements, there is an additional requirement of thermodynamically stable interaction between sodium ions and preintercalants. Graphite is well-known as an amphoteric material capable of converting to both positive and negative charge states according to types of intercalated species. Graphite becomes negatively charged by positive ion intercalation, while it could be positively charged by negative ion intercalation. Negative ion intercalation into graphite is thermodynamically possible and could form GICs similar to the GICs formed by intercalation of positive ions like Li, K, and Ca. Since the sodium ion is known to form stable ionic compounds with electronegative elements (such as chalcogen and halogen atoms),[32−34] such an anion preintercalant would avoid the repulsive Coulomb interactions found for Ca preintercalation. Based on these expected mechanisms, we now analyze anion intercalation into pure graphite, formation of anion preintercalated GICs, and their interactions with intercalated sodium ions. It is reported that halogen atoms could intercalate into graphite and form interlayer compounds with sodium ions.[19] We calculate the interlayer spacing and formation energy, Eformation = Etotal(Na-X-GIC) – Etotal(Na) – Etotal(Na-X-GIC), of sodium-ion intercalation into GICs preintercalated by a halogen element (X = F, Cl, and Br) to verify the thermodynamic stability as a function of sodium-ion contents.

In this equation, Etotal(Na-X-GIC) is the total energy of the halogen intercalated GIC as sodium-ion contents. Figure indicates that halogen elements expand graphite interlayer spacing to 3.97 Å for F, 4.46 Å for Cl, and 4.64 Å for Br (see Figure b, n = 0). These differences of expanded interlayer spacing for different halogen elements are originated from the size of anions (F– < Cl– < Br–). The larger anion size facilitates larger expansion of interlayer spacing. The total formation energies show negative values (shown in Figure a) indicating the formation of thermodynamically stable GICs. The initial sodium-ion intercalation into preintercalated graphite (Figure a, n = 1) shows very low formation energy, indicating a substantial stabilization interaction between sodium ions and halogen element GICs. For additional intercalation of sodium ions (i.e., n = 2), the formation energy decreases by the additional intercalation in all cases while maintaining thermodynamically stable intercalation. However, we note that these initial strong bindings of Na-halogen within the interlayer space would be an obstacle to a reversible deintercalation of sodium ions from the GICs leading to reduced electrochemical activity as anode materials for SIBs, similar to inactive salt compound GICs with fully occupied interlayer space by multiple atomic layers (e.g., Na2ClC13 and KClC16). The findings in Figure as well as the previous experimental data show that Na halide salt intercalation is possible but that all Na atoms would be strongly bonded in the interlayer halide.

Figure 3

Formation energies (a), interlayer distances (b), and atomic structures (c) of F, Cl, and Br ion preintercalated GICs as sodium-ion content changes.

To overcome the challenge of inactive Na halide formation by high-density halogen preintercalation, we examine the low-density halogen preintercalations using a larger model GIC system with 6×6×1 unit cells. Figure shows enlarged interlayer spacing in the vicinity of halogen atoms, 3.94 Å for F, 4.31 Å for Cl, and 4.43 Å for Br, close to the previous expanded interlayer spacings for small model systems. The interlayer spacing values are slightly reduced for the larger model systems by the gap closing van der Waals interactions, which are also shown by the reduced interlayer spacing between the low-density anion pillars (e.g., 3.58 Å for F, 3.61 Å for Cl, and 3.83 Å for Br in Figure ).

Figure 4

Atomic structures of halide ion preintercalated GICs of a 6×6×1 unit cell for F (a), Cl (b), and Br (c).

Next, we consider sodium-ion intercalation at different sites relative to the halogen preintercalant (12 positions indicated in Figure S2). The calculated formation energies are given in Table S1, and all Na sites show stable intercalation enabled by halogen preintercalants. For all halogen preintercalants, the nearest neighbor site (position 3 in Figure S2) is the most stable for sodium intercalation indicating NaX interlayer salt dimer formation. Further Na intercalation into NaX intercalated graphite leads to Na2X and Na3X planar salt cluster formation as shown Figures S3, S4, and S5 for X = F, Cl, and Br, respectively. These salt clusters are also very stable and would not participate in reversible dissociation reaction to provide Na ions during desodiation of electrochemical cycles in the SIB anode. The formations of stable Na3X clusters are verified by very low formation energies shown in Table . As sodium-ion content increases, formation energies of NaX become larger up to Na3X clusters.

Table 1

Formation Energies According to the Number of Sodium Ions (n) in NaX Cluster GICs

	formation energy (eV)
n	F	Cl	Br
1	–1.31	–1.67	–1.76
2	–1.56	–1.56	–1.62
3	–1.7	–1.66	–1.87
4	–0.58	–1.09	–1.1

However, Table also shows that formation energy decreases for an additional Na intercalation into Na3X cluster GICs. In addition, the 4th intercalated sodium ions are located away from the Na3X cluster as shown in Figure (most stable sites of the 4th Na ions). These atomic configurations and formation energies indicate that preintercalated halogen X would form stable Na3X clusters and that further intercalated sodium ions would be reversibly deintercalated from the Na3X cluster pillared GICs. Particularly, sodium-ion intercalation into Na3F halide cluster GICs shows thermodynamically stable and weak intercalation energy (−0.58 eV) compared to Na3Cl and Na3Br cluster GICs. The intercalation energy of the 4th Na ion is consistent with the interlayer spacings of Na3X pillared GICs as shown in Figure (4.30–4.67 Å for Na3F, 4.81–4.99 Å for Na3Cl, and 5.20–5.33 Å for Na3Br pillars) in comparison with Figure a. The formation energy of sodium-ion intercalation into Na3F halide cluster GICs and the expanded interlayer spacing indicate the most promising pillaring effects as an electrochemically active anode material for reversible Na intercalation/deintercalation with a reasonable potential (<0.6 V).

Figure 5

Atomic structures of an additional sodium-ion intercalation into Na3X (X = F (a,b), Cl (c,d), and Br (e,f)) cluster GICs. Red dashed circles are Na3X clusters.

To understand the origin of cluster formation energies, we have analyzed atomic and electronic structures for each NaX halide cluster GICs with increasing sodium-ion contents, n = 0–4. Tables S2–S4 list the net charges of X and Na ions for each NaX cluster GIC as sodium-ion contents (Nac = sodium ion in the NaX cluster, Na = sodium ion located far from the NaX cluster). Halogen elements are negatively charged by extracting electrons from the graphite for n = 0. When Na atoms are further intercalated forming NaX clusters (n = 1, 2, and 3), halogen ions are further negatively charged by extracting additional electrons from the Na cations: −0.62 to −0.88 for F, −0.34 to −0.83 for Cl, and – 0.09 to −0.79 for Br. However, halogen elements maintain the same charge states with one more sodium-ion (n = 4) intercalation into Na3-X-GICs. For all these cases, the charge states of Na cations are 0.87–0.90. These charge variations are consistent with changes of formation energy with varying sodium-ion contents. With increasing electronic charge of halogen by sodium-ion intercalation, the formation energy would also be stronger. On the other hand, formation energy becomes weaker for one more addition of sodium ion into Na3X cluster GICs, confirming that the charge states of halogen elements are saturated at Na3X clusters and that further intercalated sodium ions would not further transfer electrons to halogen elements.

Considering the electronegativity of halogen elements (F > Cl > Br), one can understand the initial charge states of intercalated halogen atoms (−0.62 for F, −0.34 for Cl, and −0.09 for Br) transferred from the graphite Fermi level (work function of 4.7 eV). As Na atoms are intercalated into the graphite interlayer, NaX clusters are formed, and further electrons are transferred from intercalated sodium ions to halogen elements as well as to graphite. These charge transfers facilitate to form thermodynamically stable sodium-ion intercalation GICs. Moreover, reversible sodium-ion intercalation/deintercalation (n > 3) after the formation of Na3X clusters is possible by relatively weak interaction between sodium ions and Na3X clusters with charge saturated halogen elements.[35] While halogen elements gain more electrons from Na atoms and form X-Na ionic bonds for n = 1–3, the ionic bonding lengths remain the same with an increasing number of ionic bonds as shown in Table S5. This result shows that the Na3X cluster formation is driven by the planar Na cation coordination geometry around the X anion and also indicates a possibility of an increased coordination number (n = 4) for larger halogen anions, Cl or Br.

Next, we examine the possibility for the formation of Na4X clusters by considering the differences of atomic size and charge states. We confirm that Na4Cl and Na4Br could be formed, while the Na4F cluster is unstable to be formed. Considering the larger ion sizes, Cl and Br could stably bind with more sodium ions. These Na4Cl and Na4Br exhibit interlayer spacing over 4.5 Å, as described in Figure .[36] Formation energies by intercalating one additional sodium ion into Na4Cl and Na4Br cluster GICs indicate −0.65 and −0.635 eV, respectively, comparable with −0.58 eV of the one additional sodium-ion intercalation into Na3F cluster GICs. Thus, Na4Cl and Na4Br cluster GICs could be also suitable as anode materials for SIBs.

Figure 6

Atomic structures of Na4X (X = Cl (a) and Br (d) cluster GICs) and an additional sodium-ion intercalation into Na4X (X = Cl (b,c) and Br (e,f)) cluster GICs. Red dashed circles: Na4X clusters.

Based on the design of Na3F, Na4Cl, or Na4Br cluster pillared graphite, we calculated the theoretical capacities for each cluster GIC (Figure ). Formation energies for each pillared graphite indicate that sodium-ion intercalations into each cluster GIC are thermodynamically stable as sodium-ion concentration increases until a total of nine sodium-ion intercalations into the model system (note Figure S2), while over 10 sodium-ion intercalations into each cluster are thermodynamically unstable. Considering that sodium ions are stacked in a stage mechanism between the slabs, the total number of sodium ions in Na3F cluster GICs is nine per layer (in the model system with 12 sites per layer as indicated in Figure S2), which consist of three sodium ions for the cluster formation and six for reversible intercalation/deintercalation.[11] From this analysis, one can estimate a theoretical capacity of 186 mA h g–1. In the same manner, the theoretical capacities are also calculated for Na4Cl cluster GICs (155 mA h g–1) and Na4Br cluster GICs (155 mA h g–1) in which the total nine sodium ions consist of four sodium ions in clusters and five for reversible intercalation/deintercalation. These theoretical capacities of the cluster GIC anode materials are competitive compared with the recently reported graphite anode materials using an electrolyte cointercalation system for SIBs (e.g., cointercalation of sodium ions with an ether-based electrolyte indicates ∼100 mA h g–1).[10,14,15]

Figure 7

Formation energies of Na-ion intercalation into Na3F, Na4Cl, and Na4Br cluster GICs. Blue areas mean thermodynamically stable states.

To examine the origins of these capacities, we calculated electronic structures of carbons (C) in Na3F, Na4Cl, or Na4Br cluster GICs (Figure ). Graphite consists of carbons and accommodates positive ions by receiving electrons to carbons from intercalated elements. Partial density of states (PDOSs) indicates variations of charge states of C depending on the number of intercalated sodium ions from non-Na intercalation to theoretically full Na intercalation in Na3F, Na4Cl, or Na4Br cluster GICs. Figure a shows that energy states of C in Na3F cluster GICs shift to the left side under the Fermi level by accommodating nine sodium ions, which is theoretically full Na intercalation, compared to the non-Na intercalation state. This shift means that C is reduced by receiving electrons by sodium-ion intercalations. In addition to Na3F cluster GICs, Na4Cl or Na4Br cluster GICs indicate also the same tendency for the reduction of C by retaining theoretically full Na intercalation. These mechanisms of charge storage in NaX halide cluster GICs are the same with the origins of charge storage in graphite as an anode material in conventionally used rechargeable batteries. It means that the NaX halide cluster GICs could be used as an anode in SiBs.

Figure 8

Partial density of states (PDOS) of (a) Na3F, (b) Na4Cl, and (c) Na4Br cluster GICs depending on Na intercalation. Red dashed lines are the Fermi level.

In addition, we also calculated the average potentials for the NaX halide cluster GICs to evaluate the rational process of sodium-ion intercalation. To examine the potential profiles of the GICs depending on the different stages, we constructed atomic models that the NaX halide clusters are located in each graphite layer (Figure S6) and the following equations for the average voltage (VAvg) that is VAvg = [μ(X-GIC) + nμ(Na) – μ(Na-X-GIC)]/n, where μ is the chemical potential. Herein, we adopt five sodium ions for Na3F cluster GICs and six sodium ions for Na4Cl/Na4Br cluster GICs for each layer, based on the calculated formation energies. Considering the stage mechanism, we confirmed that the calculated potentials of the GICs decrease during the intercalation from stage 2 and stage 1 (Figure ). The decreasing tendency of the potential variations is consistent well with the requirement as anode materials. For these reasons, the designed NaX halide cluster GICs could be considered as the suitable anode materials in SIBs.

Figure 9

Average potential of Na3F, Na4Cl, and Na4Br cluster GICs depending on the stages.

Conclusions

In this study, we analyzed the correlations between the energetic features and interlayer spacings of sodium-ion intercalation into graphite and confirmed that interlayer spacing is required over 3.9 Å for thermodynamically stable intercalation by DFT calculation. From the mechanism, we first consider preintercalated GICs of other cations (e.g., Li+, K+, and Ca2+) to secure enlarged interspacing of graphite. However, increased repulsive Coulomb interactions between the other cations (e.g., Ca2+) and Na+ cation hinder usage of preintercalated GICs of other cations as anode materials for SIBs despite procuring the optimal expanded interspacing. Considering the thermodynamically stable stabilization, we also examined sodium-ion intercalation into preintercalated GICs by halogen elements (e.g., F, Cl, and Br). We could confirm that the halogen elements form thermodynamically stable NaX cluster GICs as anode materials in SIBs. Among the alternatives for anode materials for SIBs, Na3F, Na4Cl, and Na4Br cluster GICs are suitable as anode materials due to securing expanded interspacing over 3.9 Å and reasonable binding with newly intercalated sodium ions by facilitating repetitive intercalation and deintercalation. Moreover, Na3F, Na4Cl, and Na4Br cluster GICs exhibit theoretically competitive capacities of 186, 155, and 155 mA h g–1. Therefore, we expect that this result could suggest a tactical intuition of developing promising anode materials beyond the conventional carbonaceous anode materials for SIBs.