Universal strategy of capacity compensation via the electrolyte for Li-ion batteries

Abstract

The initial irreversible capacity loss (ICL) during the first charging process greatly reduces the affordable energy and power densities of commercial Li-ion batteries (LIBs). Though various solid-phase prelithiation additives have been developed, they can hardly be adapted to the existing electrode production lines. Herein, we elucidate that the root cause of ICL is electron loss; therefore, the electron compensation is the most sustainable and economical capacity-compensation method rather than conventional prelithiation. Accordingly, we propose a universal capacity-compensation strategy via the electrolyte and establish general guiding principles for developing soluble capacity-compensation reagents. As a proof of concept, lithium polyphosphides (Li_xPPs) were synthesized and evaluated. Li_xPPs feature desirable solubility in various commercial electrolytes and undergo preferential oxidation at the cathode side, thereby not only compensating for ICL but also contributing to the construction of a stable cathode-electrolyte interphase (CEI). The feasibility of this capacity-compensation electrolyte was validated in multiple mainstream battery systems, including NCM811||graphite, LiFePO₄||graphite, and LiCoO₂||graphite full cells, highlighting both its great adaptability to the existing battery manufacturing processes and its potential to enhance the energy density of LIBs and beyond.

---

Broader context

Lithium-ion batteries (LIBs) have attracted significant attention for powering portable electronics and electric vehicles owing to their high energy density, long cycle life and low self-discharge rate. However, the initial irreversible capacity loss (ICL) during the first cycle considerably limits their practical energy densities. Conventional solid-phase prelithiation additives, meanwhile, face challenges in seamless integration into existing manufacturing processes. Herein, we identify electron loss as the fundamental cause of the ICL and propose that compensating for electrons, rather than relying on conventional prelithiation, represents a more sustainable and economical approach to address this issue. Building on this principle, we develop a universal capacity-compensation strategy via the electrolyte and establish general design criteria for soluble capacity-compensation reagents. As a proof-of-concept, lithium polyphosphides (Li_xPPs) undergo preferential oxidation at the cathode side, thereby not only compensating for the ICL but also facilitating the construction of a stable cathode–electrolyte interphase (CEI). When evaluated in terms of specific capacity, cost, electrolyte compatibility, operational feasibility, safety and lifespan, Li_xPPs demonstrate notable advantages over conventional prelithiation agents such as Li₂NiO₂ and Li₅FeO₄. The feasibility of using Li_xPPs has been validated in multiple mainstream battery systems, highlighting its great potential for enhancing the energy density of LIBs and beyond.

Introduction

Lithium-ion batteries (LIBs) have attracted significant attention for the development of portable electronic products, electric vehicles, and grid-scale energy storage due to their advantages of high specific energy density, long cycle life and low self-discharge rate. In the initial charging process, decomposition of the electrolyte on the surface of the electrodes leads to the formation of both a solid electrolyte interface (SEI) and a cathode electrolyte interface (CEI).^1–4 As shown in Fig. 1a, the side reactions that lead to the formation of a SEI at the anode side can be categorized into two types. One is the reduction reaction, relating to solvents (involving the consumption of electrons only, labeled as Q_e⁻ (I)), and the other is the reduction reaction, relating to the solvation of Li ions (involving the simultaneous consumption of both Li ions and electrons, labeled as Q_e⁻ (II)). Especially, in the latter case, the loss of electrons is generally more than the loss of Li ions.^5–7 Meanwhile, at the cathode side, the formation process of the CEI is mainly derived from the oxidation of solvents and lithium salts; the total electricity consumption of the CEI is denoted as Q_e⁻ (III). It is well known that the SEI induced by the reduction of the electrolyte is much thicker than the CEI, indicating that the reduction of solvents and lithium salts is way more than their oxidation at the cathode, resulting in the imbalance of side reactions between the number of electrons lost at the cathode side and the number of electrons gained at the anode side (Fig. 1a and Fig. S1a). The difference in capacity loss between the anode and cathode, denoted as Q_e⁻ (IV) and defined by Q_e⁻ (IV) = Q_e⁻ (I) + Q_e⁻ (II) − Q_e⁻ (III), must be compensated by active cathode materials that supply electrons. This process occurs alongside equimolar consumption of lithium ions, resulting in the initial irreversible capacity loss (ICL) and the consumption of limited active Li available^8,9 (Fig. 1a and Fig. S1a). In commercial batteries, 5–20% of the active Li ions from the cathode are involved in this process. Even worse, cracking and pulverization of the anodes will aggravate irreversible capacity loss. Therefore, various prelithiation technologies have been developed to alleviate ICL. As previously reported, Li₂NiO₂,^10,11 Li₅FeO₄,^12–15 Li₆CoO₄,^16,17 Li₂S/M,¹⁸ LiF/M¹⁹, Li₂O/M^20,21 (M = Co, Ni, Fe), Li₃N,^22–24 Li₂S,^25,26 Li₄SiO₄/S,²⁷ lithium powder and Li_xSi^28,29 are applied as prelithiation reagents. To date, the majority of reported prelithiation methods adopt solid prelithiation reagents either in the electrodes or on their surface, which may pose a threat to the structural stability of electrodes due to the volumetric shrinkage of the prelithiation reagents, or lead to gas release in the cases of Li₃N and Li₂O. In addition, it is difficult to achieve a 100% delithiation reaction with solid prelithiation reagents.

Fig. 1 Scheme of capacity-compensation additive. (a) Schematic of the initial side reactions and unnecessary irreversible capacity loss (the volume of a cylinder represents the gain or loss of electrons) in a normal battery. (b) Energy level diagram of NCM811, graphite and the capacity-compensation reagent in the electrolyte. (c) Theoretical specific capacity of the reported cathode prelithiation and the capacity-compensation reagent in this work (LiP₇, LiP₅, LiP₃) and their volume shrinkage ratios after delithiation (chemical formulas given in orange are the corresponding delithiation products). (d) Mass ratios of traditional prelithiation reagents or the capacity-compensation reagent to the NCM811 cathode material that are needed when the compensation capacity is required to be 10% of the cathode capacity. The right y-axis represents their redox voltages.

To select effective capacity-compensation reagents, it is necessary to clarify the irreversible reaction processes that happen at the anode and cathode sides, respectively. In the initial charging process, the irreversible electron consumption at the anode side should be balanced by that at the cathode side (Q_{e⁻ (anode)} = Q_{e⁻ (cathode)}). However, lithium-ion consumption differs for these two electrodes, including the main and side reactions (Q_{Li+ (anode)} ≠ Q_{Li+ (cathode)}). The number of electrons consumed by the reduction reaction of solvents and lithium salts at the anode outnumbers the number of electrons provided by the oxidation reaction derived from the electrolyte content at the cathode. In a normal battery, the difference in electric quantity caused by side reactions, Q_e⁻ (IV), should be balanced by cathode materials, resulting in the loss of an equal molar amount of active lithium (Fig. 1a and Fig. S1a). The utilization of cathode prelithiation agents effectively addresses the deficiency of the lithium source; however, it does not alleviate the constraint on electron supply. Hence, it solely compensates for the imbalance of electric quantities between the anodic electrochemical half-reaction and cathodic electrochemical half-reaction by employing equimolar active lithium sources, which are generally costly and susceptible to oxygen and humidity (Table S1). Therefore, a more economical and effective method urgently needs to be developed. One approach, named “capacity compensation”, can supply more electricity than the released lithium ions, thus precisely balancing the difference in electricity loss.

Based on the above discussion, we proposed a capacity-compensation strategy using a reagent that can be oxidized at the cathode side but does not necessarily contain lithium. However, if the reagent contains lithium, it would enable extra lithium to be released simultaneously, so that the released electrons and lithium ions can be fully utilized and exactly compensate for both electron and Li losses originating from the formation of the SEI on the anode side. Furthermore, it would be highly promising if the capacity compensation could be realized via electrolytes rather than with the common solid prelithiation reagents, based on the following considerations. (i) If the capacity-compensation additives are uniformly dispersed in the electrolyte, a homogeneous capacity-compensation reaction can be achieved at the electrode surface. (ii) If the capacity-compensation additive can be dissolved in the electrolyte, it will be compatible with the existing battery production process without increasing safety concerns. (iii) No adverse effects on the electrode stability should be introduced.

There are several difficulties in selecting an appropriate capacity-compensation reagent for application in the electrolyte. (i) It should have adequate solubility in the electrolyte and be compatible with solvents and lithium salts. (ii) It should preferably contain active Li, which means that it can be oxidized and release Li ions. (iii) As shown in Fig. 1b, the highest occupied molecular orbital (HOMO) energy of the capacity-compensation reagent should be higher than that of the cathode, so that it can be oxidized prior to the cathode. In the meantime, its lowest unoccupied molecular orbital (LUMO) energy should be much higher than that of the cathode, so that it will no longer take in lithium during subsequent cycles. (iv) The LUMO energy of the capacity-compensation reagent should preferably be higher than that of the electrolyte, ensuring its minimal consumption at the anode side, or its participation in the formation of a robust SEI. Crucially, it must not have a detrimental impact on the performance of the battery. Furthermore, the reagents will be even more appealing if they exhibit strong chemical adsorption on the cathode surface, thereby enhancing kinetic reactivity and achieving high capacity-compensation efficiency. Additionally, the oxidation products of the capacity-compensation reagents should function as a stable and robust CEI.

On the basis of the above discussion, we herein demonstrate polyphosphides (Li_xPPs), such as LiP₇, LiP₅ and LiP₃, as appropriate capacity-compensation reagents; these reagents are soluble in ester-based electrolytes³⁰ and meet all the above-mentioned requirements. As shown in Fig. 1c, Li_xPPs can be oxidized to P⁵⁺, and thus, they can provide high theoretical specific capacities (>4100 mAh g⁻¹). The traditional prelithiation reagents are usually solid, and their delithiation products are either solid or gas, accompanied by significant volumetric shrinkage. In contrast, Li_xPPs can dissolve in the electrolyte, resulting in a uniform solid film at the cathode surface, so that it will not compromise the integrity of the cathode. If the compensation capacity reaches 10% of the cathode capacity, the mass ratios of the required reagents to the cathode material (such as NCM811) are listed in Fig. 1d. Compared with traditional prelithiation reagents (e.g. Li₅FeO₄, 2.86 wt%), only 0.46 wt% Li_xPPs is needed. In addition, their relatively low oxidation potentials (calculated in the SI) enable the electron-loss reaction of Li_xPPs before the cathode, below the high cutoff potentials of conventional Li-ion batteries. The solubility empowers Li_xPPs to act as an electrolyte additive. When Li_xPPs was added to the electrolyte and applied in various commercial Li-ion batteries, an obvious capacity-compensation effect was achieved. In addition, Li_xPPs tend to be adsorbed at the cathode surface, contributing to the formation of a uniform and stable CEI, thereby achieving excellent long-cycle stability.

Physical and chemical properties of Li_xPPs in ester-based electrolytes

Inspired by our previous work, Li_xPPs can be fabricated by dissolving lithium phosphides in ester-based solvents, wherein the lithium phosphides can be converted into multi-component soluble Li_xPPs, such as LiP₇ and LiP₅.³⁰ The content of the soluble Li_xPPs can be up to 0.2 wt% in the electrolyte, based on ICP analysis (Fig. S2). The ultraviolet-visible (UV) spectrum of the Li_xPPs in EC/DEC (1/1) solvent was recorded by subtracting the background of EC/DEC (1/1) solvent. In Fig. 2a, two obvious peaks at 280.1 nm and 222.1 nm in the UV band were detected, indicating that the Li_xPPs are dissolved in the electrolyte. As shown in Fig. 2b, Fourier-transform infrared (FT-IR) spectroscopy of Li_xPPs in EC/DEC reveals peaks at 1420, 1221 and 1020 cm⁻¹, which are attributed to P–C, P=O and P–O–C bonds, respectively,^31,32 contributing to the solvation of Li_xPPs with EC or DEC. The formed Li_xPPs were recrystallized by the solvent vapor method and further studied with a high-angle annular dark-field scanning transmission electron microscope (HAADF-STEM). The Li_xPPs particles are about 3 nm in size, with the P element uniformly distributed throughout (Fig. S3a and b). The selected area diffraction pattern (Fig. S3c) reveals the existence of crystalline components, which are mainly LiP₇ and LiP₅. The above experimental results prove the successful synthesis of Li_xPPs. Mass spectrometric measurements were further performed to clarify the components of Li_xPPs. In Fig. 2c, a series of polyphosphides (PPs) such as P₅⁻, P₇⁻, and their solvated products can be detected. Similar dissolving behavior and the dissolved products are detected when using other solvents (Fig. S4), demonstrating the universal solubility of Li_xPPs in ester electrolytes.

Fig. 2 Characteristics of soluble Li_xPPs. (a) UV-vis spectrum, the inset shows digital images of the solutions. (b) FT-IR spectra. (c) Mass spectrum of Li_xPPs in the EC/DEC (1/1) solvent. (d) Anodic scan LSV of Li_xPPs in 1 M LiPF₆ EC/DEC electrolyte in the range of open circuit voltage (OCV) to 5.5 V using Pt as the working electrode and Li metal as the counter electrode. (e) High-resolution XPS spectra of P 2p on stainless steel using Li_xPPs in 1 M LiFSI EC/DEC electrolyte with a cutoff voltage of 4.0 V. (f) Cathodic scan LSV of Li_xPPs in 1 M LiPF₆ EC/DEC electrolyte in the range of OCV to 0 V using stainless steel as the working electrode and Li metal as the counter electrode.

The preferential oxidation potential of the Li_xPPs was estimated by linear sweep voltammograms (LSV) using the anodic scan (Fig. 2d). The blank electrolyte (1 M LiPF₆ in EC/DEC) shows an obvious peak at 4.5 V, corresponding to the decomposition of solvent and lithium salt, while the Li_xPPs-containing electrolyte is oxidized from about 3.2 V. This indicates the preferential oxidation of Li_xPPs before both the electrolyte and the commercial cathode³³ (NCM 3.5–4.3 V, LiCoO₂ 4–4.5 V, LiFePO₄, ≈3.4 V), so that the capacity-compensation can be realized in the initial charging process. Furthermore, the prior decomposition of Li_xPPs at the cathode surface is beneficial to building a CEI in order to suppress further detrimental decomposition of the electrolytes. X-ray photoelectron spectroscopy (XPS) was performed to study the valence state of the Li_xPPs after oxidation (Fig. 2e). To exclude the influence of P in LiPF₆, LiFSI was used instead as the lithium salt, with stainless steel as the working electrode and Li metal as the counter electrode. The results show that the majority of the P elements are oxidized to P⁵⁺, with other P remaining at the stage of P⁰ because of its low electronic conductivity and the increased over-potential. The residual small amount of P⁰ is beneficial for improving battery safety, as it is a widely used flame retardant.

The reduction potentials of Li_xPPs are very important to ensure that it would be consumed at the anode side. From the LSV cathodic scan performed from OCV to 0 V (Fig. 2f), the reduction potential of the Li_xPPs-containing electrolyte is found to be ∼0.75 V, which is much lower than the cutoff potential of cathodes (2 V). This result indicates that Li_xPPs will not be lithiated again in the subsequent cycles when Li_xPPs are applied in commercial batteries, such as the NCM811||graphite full cell. The adsorption energies (E_ad) of Li_xPPs on NCM811 and graphite, respectively, were calculated by DFT (Fig. S5). The E_ad of LiP₇ at the cathode (such as NCM811) is −8.021 eV, much stronger than that at the graphite anode (−2.920 eV). This indicates that Li_xPPs can be preferentially adsorbed and oxidized at the cathode side to compensate for the capacity consumption derived from SEI formation at the anode side.

Capacity-compensation effect of Li_xPPs in NCM811||graphite batteries

To confirm the capacity-compensation effect of Li_xPPs, electrochemical tests were carried out in NCM811||graphite full cells. As shown in the CV curves (Fig. 3a), the oxidative peak of the full cell utilizing the Li_xPPs-containing electrolyte starts from 3.4 V, which is earlier than that in the blank electrolyte (3.5 V). When the voltage was increased to 4.45 V, the NCM811 in the blank electrolyte decomposed and evolved oxygen tempestuously, while this phenomenon was prevented in the Li_xPPs-containing electrolyte. It is anticipated that Li_xPPs additives will contribute to the formation of a robust CEI that can withstand high voltages. The major reductive peak in the Li_xPPs-containing electrolyte shows a relatively high peak intensity and low overpotential, indicating the fast kinetic reaction process. As shown in Fig. 3b, the initial reversible discharge specific capacity of the full cell using the Li_xPPs-containing electrolyte is 221.4 mAh g⁻¹, which is higher than that of the blank electrolyte (196.5 mAh g⁻¹). The former's ICE (86.7%) is obviously higher than that of the blank electrolyte (79.6%), which is ascribed to the extra capacity compensation derived from Li_xPPs. As shown in Fig. 3c and Fig. S6, the cell using the Li_xPPs-containing electrolyte exhibits a better capacity retention of 84.2% upon cycling at 100 mA g⁻¹ after 400 cycles, much higher than that using the blank electrolyte (68.2%). When the high cutoff voltage is raised to 4.5 V, the Li_xPPs-containing electrolyte also exhibits a higher discharge capacity (Fig. S7), owing to the stable CEI on the surface of the NCM811 and the capacity-compensation effect of the Li_xPPs.

Fig. 3 Electrochemical performances of NCM811||graphite full cells using the Li_xPPs-containing electrolyte and the blank electrolyte. (a) First CV curves recorded at a scan rate of 0.1 mV s⁻¹ between 3 and 4.5 V. (b) Initial charge and discharge curves at 20 mA g⁻¹. (c) Cycling performance at 100 mA g⁻¹. (d) Rate performances of NCM811||graphite full cells. (e) Cycling performance and corresponding Coulombic efficiency of the 1.8 Ah NCM811||graphite pouch cell between 3 and 4.2 V at 1C. Mass loading of NCM811 is 12.4 mg cm⁻², and the N/P ratio is about 1.15.

The influence of Li_xPPs on high-rate performance was further investigated (Fig. 3d and Fig. S8), and the cell using the Li_xPPs-containing electrolyte showed low polarization at elevated rates, demonstrating the positive effects of Li_xPPs on the fast-charging kinetics. The 1.8 Ah pouch cells of NCM811||graphite are presented in Fig. 3e and Fig. S9; the pouch cell using the Li_xPPs-containing electrolyte displays an initial discharge capacity of 187.8 mAh g⁻¹ and a capacity retention of 91.9% after 100 cycles, which is higher than that with the blank electrolyte (178.8 mAh g⁻¹, 84.8% after 100 cycles). Interestingly, a small dose of Li_xPPs additive not only provides high-capacity compensation, but also significantly improves the cyclic stability.

Advantages of the Li_xPPs additive on the NCM811 cathode and graphite anode

In order to clarify the role of Li_xPPs in capacity-compensation and the formation of CEI, we have further studied the performance of NCM811 in a half-cell. As shown in the CV curves of the first two cycles (Fig. 4a and Fig. S10), the oxidation peak in the Li_xPPs-containing cell occurs at a lower potential than the control sample, which is attributed to the preferential oxidation of Li_xPPs. In addition, the reduction peak of the Li_xPPs-containing cell exhibits a larger peak area than the control cell, demonstrating a higher reversible capacity. Given the intricate mechanisms within the full cells, involving multiple side reactions occurring at both the anode and cathode interfaces, precise quantification of the capacity-compensating contribution derived from Li_xPPs remains a significant challenge. To address this issue, the initial charge curves of the NCM811 half cells are compared in order to assess the capacity contribution of Li_xPPs through electronic compensation because the excess lithium source in the Li metal can eliminate the influence of the anode side. In Fig. 4b, the initial charge capacity of NCM811 using the Li_xPPs-containing electrolyte is 263.4 mAh g⁻¹, which is 14.6 mAh g⁻¹ higher than the control cell (248.8 mAh g⁻¹), indicating that Li_xPPs can deliver a capacity-compensation effect. By calculating the actual capacity compensation and comparing it with the theoretical capacities supplied by Li_xPPs through Li⁺- and electron-compensation mechanisms, respectively, the results become more intuitive (Tables S2 and S3). The actual compensation capacity is 0.038 mAh, which exceeds the highest theoretical Li-compensation capacity of Li_xPPs (0.0118 mAh) but is lower than the highest theoretical electronic compensation capacity (0.423 mAh), demonstrating that the capacity compensation is mainly attributed to the electronic contribution. Therefore, it can be roughly quantified that the electron-compensation mechanism of Li_xPPs contributes approximately 5.9% additional capacity relative to the total capacity in the NCM system. The difference in their initial discharge capacities is 22.7 mAh g⁻¹, which is 8.1 mAh g⁻¹ higher than the difference in their initial charge capacities. This indicates that Li_xPPs not only provide additional capacity but also suppress the initial irreversible side reaction.

Fig. 4 Electrochemical performance and characterization of NCM811||Li and graphite||Li cells using the blank electrolyte and the Li_xPPs-containing electrolyte. (a) First cyclic voltammetry (CV) curve of the NCM811||Li cell at a scan rate of 0.1 mV s⁻¹. The mass loading of NCM811 is 2.3 mg cm⁻². (b) Initial galvanostatic charge and discharge curves of the NCM811||Li cell at 10 mA g⁻¹. (c) XRD patterns of the NCM811 electrodes after being soaked in blank and Li_xPPs-containing electrolytes for 12 h, as well as the XRD patterns of the NCM811 electrodes after 20 cycles in blank and Li_xPPs-containing electrolytes. (d) Structural representations of Li/Ni antisite disorder for bare NCM811 and NCM811 with the adsorption of P₇⁻. (e) and (f), High-magnification TEM of NCM811 electrodes after 20 cycles at 100 mA g⁻¹ in the blank (e) and Li_xPPs-containing (f) electrolyte. (g) Highest occupied molecular orbital (HOMO) energies and lowest unoccupied molecular orbital (LUMO) energies of the solvents, lithium salt and LiP₇. (h) Initial galvanostatic charge and discharge curves of the graphite||Li cell at 0.1 C. (i) Cycling performance of the graphite||Li cell at 0.2C. The mass loading of graphite is 2.1 mg cm⁻².

The cycling stability of NCM811 in the two electrolytes is compared in Fig. S11a. After 100 cycles at 40 mA g⁻¹, the discharge capacity is 186.9 mAh g⁻¹ in the Li_xPPs-containing electrolyte, higher than the control sample (177.9 mAh g⁻¹). The former exhibits much better performance at high rates than the latter (Fig. S11b). For example, the former delivers a capacity of 174 mAh g⁻¹ even at 400 mA g⁻¹, while the latter only provides 146 mAh g⁻¹. When the current density drops to 200 mA g⁻¹, the specific discharge capacity is restored in the cell using Li_xPPs-containing electrolyte. By contrast, the capacity decreased obviously in the control sample. Therefore, we infer that the Li_xPPs-derived CEI is also beneficial to the structural stability of NCM811, in addition to the capacity-compensation effect in the initial charging process. To study this further, X-ray diffraction (XRD) spectra (Fig. 4c, Fig. S12) were recorded to quantify the degree of Li⁺/Ni²⁺ disorder. The I₍₀₀₃₎/I₍₁₀₄₎ ratios of the NCM811 cycled in the blank and the Li_xPPs-containing electrolyte are 2.569 and 2.995, respectively. The high I₍₀₀₃₎/I₍₁₀₄₎ ratio means a low degree of Li⁺/Ni²⁺ disorder.^34,35

To clearly reveal the effect of Li_xPPs on the Li⁺/Ni²⁺ disorder, the NCM811 cathodes were soaked in both electrolytes, respectively, for 12 h. As shown in Fig. 4c, the I₍₀₀₃₎/I₍₁₀₄₎ ratios of the NCM811 soaked in the blank and the Li_xPPs-containing electrolytes are 5.420 and 5.095, respectively. The lower ratio of the NCM811 soaked in Li_xPPs-containing electrolyte demonstrates that the cation disorder happens to a certain extent before cycling. DFT calculations were applied to establish the exchange energies (E_ex) of a pair of Ni/Li. As shown in Fig. 4d, the bare NCM811 exhibits an E_ex of 1.282 eV in the superficial layer (named as superficial-E_ex), while the superficial-E_ex is −0.202 eV if the NCM811 was absorbed by P₇⁻, indicating that Li_xPPs can induce cation disorder in the superficial layer.

We further calculated the E_ex inside the NCM811 bulk (named as inside-E_ex). The inside-E_ex of NCM811 without and with adsorption of P₇⁻ are −1.322 and −0.935 eV, respectively. This reveals that the Li_xPPs adsorbed on NCM811 can restrain the deep cation disorder in the bulk, even if there is a slight cation disorder in the superficial layer. These results are consistent with the previous reports that a small amount of divalent nickel ion within the Li slabs can pillar the Li slabs and enhance the structural stability.^36,37 Thus, it is confirmed that Li_xPPs are conducive to maintaining the stable structure of NCM811 and forming protective CEI.

Transmission electron microscopy (TEM) was also carried out to investigate the morphology of the surface of NCM811 after 20 cycles (Fig. 4e and f). The thickness of the CEI formed in the blank electrolyte is more than 20 nm and uneven. In contrast, the CEI formed in the Li_xPPs-containing electrolyte is quite uniform with a thickness of about 4 nm. The chemical components of the CEI were further investigated by XPS. The P 2p_3/2 spectra of both electrolytes reveal a strong peak in the range of 132–138 eV, representing P–O/P=O and Li_xPO_yF_z (Fig. S13a and b). Additionally, the Li_xPPs-containing electrolyte exhibits an additional peak at 130.5 eV, which can be ascribed to the P⁰ derived from the oxidation products of Li_xPPs.^38–40 In the F 1s spectra, the sub-peaks can be assigned to LiF, Li_xPO_yF_z and C–F bonds (derived from PVDF) (Fig. S13c and d). As observed in scanning electron microscopy (SEM), the cathode cycled in the Li_xPPs-containing electrolyte exhibited barely any cracks (Fig. S14). In situ differential electrochemical mass spectrometry (DEMS) was performed to evaluate the gas generation originating from decomposition of the electrolyte solvents during the initial cycle (Fig. S15). The H₂ and CO₂ release rates are remarkably reduced in the half cell using the Li_xPPs-containing electrolyte, illustrating that the prior oxidation of Li_xPPs leads to the stable Li_xPPs-derived CEI and avoids continuous electrolyte decomposition.

It is necessary to study the effect of the Li_xPPs additive at the anode side to assess whether Li_xPPs are present in excess or cannot be consumed completely at the cathode side. As displayed in Fig. 4g, the solvation structure of LiP₇ shows higher HOMO and LUMO energies than that of LiPF₆, meaning that Li_xPPs will be preferentially oxidized but not reduced compared with LiPF₆ and its solvation structure. Graphite||Li half-cells were assembled to discover the effect of Li_xPPs on the graphite anode. As shown in Fig. 4h, the Li_xPPs-containing half-cell and control sample exhibit almost equal initial discharge capacity, while a pronounced capacity discrepancy is observed during the slope region extending from the open-circuit voltage to the intercalation plateau (OCV to 0.8 V vs. Li⁺/Li), corresponding to the SEI formation zone. The cell containing Li_xPPs exhibits a slight reduction within this region through their participation in this process. This can be approximately correlated to the diminished irreversible capacity loss observed at the anode, resulting in a 2.1% enhancement in the total capacity of the graphite anode. This ratio is lower than the initial charge capacity enhancement observed in the cathode half cell (5.9%), further suggesting that electron compensation on the cathode side is predominant. Furthermore, the cycling stability is improved by introducing Li_xPPs, which confirms that the Li_xPPs participates in the formation of the SEI (Fig. 4i). The contact angle decreases from 30.4° to 25.6° with the addition of Li_xPPs, demonstrating better affinity (Fig. S17). The lower R_ct indicates that the SEI formed in the presence of Li_xPPs possesses higher electronic conductivity (Fig. S18). XPS was employed to reveal the chemical composition of the SEI on graphite (Fig. S19). The P 2p and O 1s spectra of the Li_xPPs-derived SEI exhibit distinctly intensified P–O binding energy signals, suggesting that excess Li_xPPs facilitates the formation of a substantial amount of inorganic phosphate or Li_xPO_yF_z. In the C 1s spectra, the SEI formed in the Li_xPPs-containing electrolyte shows a relative decrease in C–C peaks representing the graphite substrate compared with that in the blank electrolyte. However, the organic carbon–oxygen components (e.g., ROCO₂Li, C–O, C=O) are increased, indicating that the additive promotes a more compact and fully covered SEI film in terms of spatial distribution. In the F 1s spectra, the SEI formed in the blank electrolyte after 20 cycles contains 26.5% of LiF. In contrast, the SEI formed in the Li_xPPs-containing electrolyte exhibits a higher content of LiF (53.9%), which is widely regarded as a reagent to inhibit the decomposition of LiPF₆ and ensure high cycling stability.

Universal application of Li_xPPs additives in commercial battery systems

In order to estimate the broad feasibility of the Li_xPPs capacity-compensation feature, Li_xPPs were added to LiFePO₄||graphite. The first charge–discharge profiles indicate that Li_xPPs can be decomposed at the LFP cathode and offer a charge specific capacity 14.2 mAh g⁻¹ higher than that using blank electrolyte (156.5 mAh g⁻¹) (Fig. 5a). Nevertheless, the discrepancy between the discharge capacity of LFP with the blank and Li_xPPs-containing electrolytes is 8.3 mAh g⁻¹, which is lower than the gap of the charge capacity (14.2 mAh g⁻¹) (Fig. 5a). For the LiCoO₂||graphite full cells, the battery using Li_xPPs exhibits an extra charge capacity increase of 5.4 mAh g⁻¹ and a discharge capacity improvement of 25.1 mAh g⁻¹ (Fig. 5b).

Fig. 5 Electrochemical performance of LiFePO₄||graphite and LiCoO₂||graphite full cells using the Li_xPPs-containing electrolyte. (a) and (b), First charge and discharge curves at 0.1C of (a) LiFePO₄||graphite and (b) LiCoO₂||graphite full cells. The mass loading of LiFePO₄ is 12.6 mg cm⁻². The mass loading of LiCoO₂ is 5.0 mg cm⁻². (c) Capacity improvement ratio of the first charge and discharge process of NCM811||graphite, LiFePO₄||graphite and LiCoO₂||graphite full-cells; the star represent for the enhancement in reversible specific capacity at the 100th cycle. (d) and (e), Cycling performance at 0.2C of (d) LiFePO₄||graphite and (e) LiCoO₂||graphite full cells. (1C = 170 mA g⁻¹ for LiFePO₄ and 1C = 274 mA g⁻¹ for LiCoO₂).

The variance of the charge and discharge capacity gaps between the LiFePO₄||graphite and LiCoO₂||graphite full cells can be ascribed to different cut-off voltages (Fig. 5c). The gaps of specific capacity between the blank and Li_xPPs-containing electrolytes during the first charge and discharge processes are marked as Q_charge gap and Q_{discharge gap}, respectively. Due to excessive anode materials in the full cell, the initial charging capacity hinges on the total electron-loss/oxidation reaction, including main and side reactions at the cathode side, labeled as Q_{total cathode charge}. According to eqn (1), the gap for the charge capacity (labeled as Q_charge gap) can be calculated as . The initial discharge capacity depends not only on the extra reversible capacity derived from Li_xPPs but also on the reduction of capacity loss originating from the side reaction at the anode, such as SEI generation. Therefore, the Q_charge gap and Q_{discharge gap} depend on both the high cut-off voltage and the type of cathode materials. NCM and LCO full cells exhibit higher charging voltages (4.3 V or 4.2 V), which is beneficial for the capacity-compensation agents to release more electrons, but also prone to causing more side reactions at both the cathode and the anode. In these two cases, the discharge gaps are larger than their charge gaps, indicating that, beyond the contribution of capacity compensation, the capacity-compensation agents participated in the formation of thin and stable SEI/CEI, thereby reducing capacity losses derived from the side reactions. In contrast, the charge gap in the LFP full cell is greater than that of LCO or NCM, indicating that the former experiences fewer SEI and CEI side reactions due to the narrow voltage window. Both LiFePO₄||graphite and LiCoO₂||graphite cells exhibit higher reversible capacities after 100 cycles than their control samples (Fig. 5d and e), signifying that Li_xPPs can participate in the formation of sturdy CEI and maintain long-term cyclability. To sum up, Li_xPPs is applicable to commonly commercialized battery systems and can balance the total electric quantity loss.

where represents the electricity consumption generated by CEI in the case of Li_xPPs-containing battery, Q_CEI represents the electricity consumption generated by CEI in the control battery, and represents the electricity provided by Li_xPPs.

The enhancement of reversible capacity plays a pivotal role in boosting the economic viability of batteries. For example, in the Chinese market, with the development of portable devices and electric vehicles, the shipment of mainstream cathode materials (LiCoO₂, LiFePO₄, and NCM) is increasing (Fig. 6a). In 2023, the total demand reached 248 × 10⁴ tons. Due to the fact that battery cost is evaluated based on the material system and its capacity (Fig. 6b), both the traditional prelithiation technology and the capacity compensation strategy can increase battery capacity. It can be understood that batteries with lithium supplementation or capacity compensation can provide extra value. Compared with the prelithiation reagent of Li₅FeO₄, under the same amount of addition, the capacity-compensation technology proposed in this work can provide more capacity and thus achieve additional profits (Fig. 6b). Furthermore, Li_xPPs can be dissolved in the electrolyte, which is compatible with the process of electrolyte injection. Therefore, the advantages of Li_xPPs, based on specific capacity, cost, compatibility, operability, safety, and life-span, are significantly higher than Li₅FeO₄ (Fig. 6c; detailed discussions are provided in the SI). The enhanced capacity retention directly contributes to increased battery longevity, which, in turn, supports the alleviation of resource constraints and fosters more environmentally sustainable practices.

Fig. 6 Economic benefit of the capacity-compensation electrolyte. (a) Shipments of cathode materials for lithium-ion batteries in China in 2014–2023 (data source: EVTank). (b) Cost of NCM811, LFP and LCO full cells (column) and the additional profit generated from the commercial cathode prelithiation reagent Li₅FeO₄ and Li_xPPs in this work (scatter). (c) Radar graph of Li₅FeO₄ and Li_xPPs for capacity-compensation applications (Scores of 2 to 10 indicate a range from low to high).

Conclusions

In this work, we proposed a capacity-compensation strategy via the electrolyte to pay off the irreversible ICL in the full-cell configuration that is compatible with the existing production process and, more importantly, does not sacrifice the structural stability of the electrode. Selection criteria for choosing the appropriate compensation reagents are also provided. As a demonstration, Li_xPPs, which is a low-cost and easy-to-synthesize additive that can be dissolved in electrolyte and oxidized prior to the cathode and electrolyte, can balance the difference of irreversible electric quantity losses between the cathode and anode. Using the Li_xPPs-containing electrolyte, not only are the ICL of various commercial full batteries greatly compensated, but their reversible capacity and cyclic stability are also significantly improved. Interestingly, besides the universal capacity-compensation effect, Li_xPPs participate in the formation of a thin, robust and uniform CEI, securing the structural integrity and cyclic stability. The compatibility of Li_xPPs with multiple commercial electrolytes allows its application in various battery systems. Furthermore, the addition of Li_xPPs can be readily adapted to existing battery production processes, without the need for extra equipment or procedures. We believe such a capacity-compensation strategy via electrolytes will encourage the development of novel prelithiation strategies for LIBs and beyond.

Methods

Synthesis of lithium phosphides

The lithium phosphides was synthesized by the method reported in the literature.^39,41 Biphenyl (Bp) was dissolved in tetrahydrofuran (THF) to prepare a uniform solution, then Li slices were added, followed by stirring to make 1 M Li-Bp/THF. Red P was pretreated with an NaOH solution to remove surface oxide before use. Then, it was ball-milled to reduce the particle size. Red P (100 mg) was added to 10 mL of 1 M Li-Bp/THF solution and reacted at room temperature in an Ar-filled glove box overnight. The suspension was centrifuged in a sealed tube, washed with THF three times, and then dried on a hot plate to produce lithium phosphides.

Synthesis of Li_xPPs in the electrolyte

1 M LiPF₆ in EC/DEC (1/1, in volume) (Canrd Technology Co. Ltd) was used as the blank electrolyte. Lithium phosphides was added to the blank electrolyte and stirred for 24 h. The solution was allowed to stand for 24 h, then filtered to remove the excessive insoluble substances. This electrolyte, containing soluble Li_xPPs, is named the Li_xPPs-containing electrolyte. Li_xPPs in the LiPF₆ EC/EMC/DMC electrolyte was prepared in the same way.

Electrochemical measurement

The NCM811 electrodes were prepared by mixing 90 wt% of NCM811 (obtained from Guizhou Zhenhua Co., Ltd), 5 wt% Super P carbon and 5 wt% polyvinylidene fluoride (PVDF, dissolved in N-methyl pyrrolidinone (NMP)). The resulting slurry was spread on the Al foil and then dried under vacuum at 70 °C for 12 h. The electrode was assembled in an argon-filled glove box as a CR2032-type coin cell with a graphite anode to form a full cell. Polypropylene (PP) separators were used.

Graphite electrodes were prepared by mixing 95 wt% of graphite, 1 wt% super P carbon, 1 wt% cellulose sodium (CMC), and 3 wt% styrene–butadiene rubber (SBR) in deionized water. The slurry was then spread on the Cu foil and dried under vacuum at 70 °C for 12 h. The NCM811||graphite pouch cell with a nominal capacity of 1.8 Ah (1C) was sealed in an aluminum-laminated polymer film bag in an argon-filled glove box. The ratio of negative-electrode capacity to positive-electrode capacity (N/P ratio) is 1.15 for the pouch cells. Approximately 10.0 g of electrolyte was added to each pouch cell, and the mass of the active material in the cathode was ∼10.0 g.

All the galvanostatic charge/discharge measurements were recorded with a Neware battery-test instrument. Linear sweep voltammetry (LSV) was conducted at a scan rate of 1 mV s⁻¹ with a CHI660E electrochemical workstation (Chenhua, China). The positive scan between OCV and 7.0 V was carried out using Pt as the working electrode and Li metal foil as the counter electrode in a three-electrode system. The negative scan between OCV and 0 V was carried out using stainless steel as the working electrode in coin cells. Electrochemical impedance spectroscopy (EIS) was performed with a CHI660E electrochemical workstation in coin cells utilizing Li metal as the counter electrode. Cyclic voltammetry (CV) tests were performed with an Ivium OctoStat.

Material characterization

UV-vis spectra of soluble Li_xPPs were collected with a U-3900 (Hitachi, Japan). FT-IR spectra were recorded with a Bruker Vertex 70. The mass spectra were recorded with a matrix-assisted laser desorption/ionization-ion trap-time-of-flight tandem mass spectrometer (Bruker Daltonics, Autoflex tof/tof III). X-ray diffraction (XRD) studies were performed on a Bruker D8-Focus instrument with Cu Kα radiation. FullProf software was used to carry out the structural refinement. X-ray photoelectron spectroscopy (XPS) was performed with a ThermoFisher Scientific, K-Alpha⁺. Scanning electron microscopy (SEM) was conducted with a Regulus 8100. Transmission electron microscopy (TEM) was carried out on a JEM-2100F.

Calculation methods

The adsorption energies (E_ad) of Li_xPPs on NCM811 and graphite are calculated by DFT within the Perdew–Berke–Ernzerh (PBE) of generalized gradient approximation (GGA) as implemented in the Dmol3 package. A vacuum space of 15 Å was placed. The adsorption energies (E_ad) are calculated from eqn (2).

E_ad = E_total − (E_LixPPs + E_NCM/graphite) (2)

where the E_total is the total energy of the Li_xPPs and NCM. E_LixPPs and E_NCM/graphite represent the energies of Li_xPPs and NCM or graphite, respectively.

The effect of Li_xPPs on the cation mixing of NCM811 was examined by DFT within the PBE of GGA, as implemented in the CASTEP package. The exchange energies (E_ex) are calculated from eqn (3).

E_ex = E_cm − E_or (3)

where the E_or is the original energy of the Li_xPPs adsorbed to NCM811. E_cm represents the energy of the Li_xPPs adsorbed on the cation mixing NCM811.

The DFT calculations of HOMO and LUMO energy levels were carried out within the generalized gradient approximation with the Perdew–Burke–Ernzerhof functional (GGA-PBE). Geometric optimization and energy calculations were performed utilizing the DMol3 module within Material Studio 2017.

Author contributions

X. W. and J. S. conceived the project and designed the experiments. X. W. synthesized the electrolyte additive and performed electrochemical measurements. S. Z. modified the figures. Y. C. carried out the TEM experiments. H. G. helped with the XRD refinement and graph optimization. Y. Z. assisted with DFT calculations. C. Z. and Q. X. provided the NCM material and performed the pouch cell measurements. X. W. wrote the draft. M. Z., J. L. and J. S. reviewed and edited the manuscript. All authors discussed the results and commented on the manuscript.

Conflicts of interest

There are no conflicts to declare.

Data availability

The data supporting this article have been included as part of the supplementary information (SI). Supplementary information is available. See DOI: https://doi.org/10.1039/d5ee07058a.

Acknowledgements

This work was supported by the National Key Research and Development Program of China (2023YFB2503600), the National Natural Science Foundation of China (22279089), the Natural Science Foundation of Tianjin (23JCJQJC00300), and the Municipal Key R&D Program of Ningbo (No. 2023Z109).

References

1. S. Ko, X. Han, T. Shimada, N. Takenaka, Y. Yamada and A. Yamada, Nat. Sustainability, 2023, 6, 1705–1714.
2. X. Yi, H. Fu, A. M. Rao, Y. Zhang, J. Zhou, C. Wang and B. Lu, Nat. Sustainability, 2024, 7, 326–337.
3. S. Tan, Z. Shadike, J. Li, X. Wang, Y. Yang, R. Lin, A. Cresce, J. Hu, A. Hunt, I. Waluyo, L. Ma, F. Monaco, P. Cloetens, J. Xiao, Y. Liu, X. Yang, K. Xu and E. Hu, Nat. Energy, 2022, 7, 484–494.
4. T. Liu, L. Lin, X. Bi, L. Tian, K. Yang, J. Liu, M. Li, Z. Chen, J. Lu, K. Amine, K. Xu and F. Pan, Nat. Nanotechnol., 2019, 14, 50–56.
5. B. L. D. Rinkel, D. S. Hall, I. Temprano and C. P. Grey, J. Am. Chem. Soc., 2020, 142, 15058–15074.
6. Y. Wang, Y. Liu, Y. Tu and Q. Wang, J. Phys. Chem. C, 2020, 124, 9099–9108.
7. E. W. C. Spotte-Smith, R. L. Kam, D. Barter, X. Xie, T. Hou, S. Dwaraknath, S. M. Blau and K. A. Persson, ACS Energy Lett., 2022, 7, 1446–1453.
8. K. Y. Zou, W. T. Deng, P. Cai, X. L. Deng, B. W. Wang, C. Liu, J. Y. Li, H. S. Hou, G. Q. Zou and X. B. Ji, Adv. Funct. Mater., 2021, 31, 2005581.
9. R. Zhan, X. Wang, Z. Chen, Z. W. Seh, L. Wang and Y. Sun, Adv. Energy Mater., 2021, 11, 2101565.
10. H. Park, T. Yoon, Y. Kim, J. H. Ryu and S. M. Oh, Electrochim. Acta, 2013, 108, 591–595.
11. T. Sangsanit, R. Songthan, S. Prempluem, W. Tejangkura and M. Sawangphruk, Nano Lett., 2025, 25, 9047–9053.
12. C. Liu, G. Xie, Z. Xu, C. Liang, P. Xu and H. Fang, ACS Appl. Mater. Interfaces, 2022, 14, 16117–16124.
13. W. M. Dose, C. Villa, X. Hu, A. R. Dunlop, M. J. Piernas-Muñoz, V. A. Maroni, S. E. Trask, I. Bloom, V. Dravid and C. S. Johnson, J. Electrochem. Soc., 2020, 167, 160543.
14. X. Xu, S. Chu, S. Xu, H. Li, C. Sheng, M. Dong, S. Guo and H. Zhou, Angew. Chem., Int. Ed., 2025, 64, 202415891.
15. Y. Zhu, R. Xu, Y. Zheng, Y. Chen, J. Yin, J. Xue, B. Zhang, L. Li, G. Zeng, H. Luo, X. Wu, K. Zhang, Z. Wu, S. Yang, S. Li, Y. Sun, D. Zhang, Y. Qiao and S. Sun, Angew. Chem., Int. Ed., 2025, 64, 202502126.
16. Y. Guo, X. Li, Z. Wang, H. Guo and J. Wang, J. Energy Chem., 2020, 47, 38–45.
17. W. Lee, H. Lee, Y. Byeon, J. H. Kim, W. Choi, M. Choi, M. S. Park and W. S. Yoon, Adv. Energy Mater., 2023, 13, 2302316.
18. Y. Sun, H. Lee, Z. W. Seh, G. Zheng, J. Sun, Y. Li and Y. Cui, Adv. Energy Mater., 2016, 6, 1600154.
19. J. Du, W. Wang, A. Y. Sheng Eng, X. Liu, M. Wan, Z. W. Seh and Y. Sun, Nano Lett., 2019, 20, 546–552.
20. Y. Sun, H. Lee, Z. W. Seh, N. Liu, J. Sun, Y. Li and Y. Cui, Nat. Energy, 2016, 1, 15008.
21. Y. Chen, Y. Zhu, W. Zuo, X. Kuai, J. Yao, B. Zhang, Z. Sun, J. Yin, X. Wu, H. Zhang, Y. Yan, H. Huang, L. Zheng, J. Xu, W. Yin, Y. Qiu, Q. Zhang, I. Hwang, C. Sun, K. Amine, G. Xu, Y. Qiao and S. Sun, Angew. Chem., Int. Ed., 2024, 63, 202316112.
22. Y. Sun, Y. Li, J. Sun, Y. Li, A. Pei and Y. Cui, Energy Storage Mater., 2017, 6, 119–124.
23. K. Park, B. Yu and J. B. Goodenough, Adv. Energy Mater., 2016, 6, 1502534.
24. J. Lu, W. Yuke, Q. Yan, Y. Siyu, C. Xinyu, Y. Ming, Z. Jing and Z. Fu, Energy Storage Mater., 2024, 66, 103204.
25. Y. Zhan, H. Yu, L. Ben, Y. Chen and X. Huang, Electrochim. Acta, 2017, 255, 212–219.
26. A. Shao, H. Wang, M. Zhang, J. Liu, L. Cheng, Y. Li, Y. Guo, Z. Wang, Q. Jia, X. Wang, X. Tang, X. Zhao and Y. Ma, Nat. Commun., 2025, 16, 4145.
27. M. Tian, Y. Yan, H. Yu, L. Ben, Z. Song, Z. Jin, G. Cen, J. Zhu, M. Armand, H. Zhang, Z. Zhou and X. Huang, Adv. Mater., 2024, 36, 2400707.
28. J. Zhao, Z. Lu, H. Wang, W. Liu, H. Lee, K. Yan, D. Zhuo, D. Lin, N. Liu and Y. Cui, J. Am. Chem. Soc., 2015, 137, 8372–8375.
29. J. Zhao, J. Sun, A. Pei, G. Zhou, K. Yan, Y. Liu, D. Lin and Y. Cui, Energy Storage Mater., 2018, 10, 275–281.
30. C. Liu, M. Han, Y. Cao, L. Chen, W. Ren, G. Zhou, A. Chen and J. Sun, Energy Storage Mater., 2021, 37, 417–423.
31. H. Liu, Y. Zou, L. Tao, Z. Ma, D. Liu, P. Zhou, H. Liu and S. Wang, Small, 2017, 13, 1700758.
32. Z. Sofer, J. Luxa, D. Bouša, D. Sedmidubský, P. Lazar, T. Hartman, H. Hardtdegen and M. Pumera, Angew. Chem., Int. Ed., 2017, 56, 9891–9896.
33. J. B. Goodenough and K. Park, J. Am. Chem. Soc., 2013, 135, 1167–1176.
34. J. Zhang, D. Zhou, W. Yang, J. Yang, L. Sun, G. Schumacher and X. Liu, J. Electrochem. Soc., 2019, 166, 4097–4105.
35. D. Seo, J. Lee, A. Urban, R. Malik, S. Kang and G. Ceder, Nat. Chem., 2016, 8, 692–697.
36. Y. Cho, P. Oh and J. Cho, Nano Lett., 2013, 13, 1145–1152.
37. J. Kim, H. Cho, H. Y. Jeong, H. Ma, J. Lee, J. Hwang, M. Park and J. Cho, Adv. Energy Mater., 2017, 7, 1602559.
38. M. Suyama, S. Yubuchi, M. Deguchi, A. Sakuda, M. Tatsumisago and A. Hayashi, J. Electrochem. Soc., 2021, 168, 60542.
39. X. Wang, C. Liu, S. Zhang, H. Wang, R. Wang, Y. Li and J. Sun, ACS Appl. Energy Mater., 2021, 4, 5246–5254.
40. S. Wenzel, D. A. Weber, T. Leichtweiss, M. R. Busche, J. Sann and J. Janek, Solid State Ion., 2016, 286, 24–33.
41. G. Wang, F. Li, D. Liu, D. Zheng, Y. Luo, D. Qu, T. Ding and D. Qu, ACS Appl. Mater. Interfaces, 2019, 11, 8699–8703.

---