Electrolyte Optimization to Improve the High-Voltage Operation of Single-Crystal LiNi_0.83Co_0.11Mn_0.06O₂ in Lithium-Ion Batteries

Abstract

Single-crystal Ni-rich layered oxide materials LiNi_1−x−yCo_xMn_yO₂ (NCM, 1 – x − y ≥ 0.6) are emerging as promising cathode materials that do not show intergranular cracks as a result of the lack of grain boundaries and anisotropy of the bulk structure, enabling extended cyclability in lithium-ion batteries (LIBs) operating at high voltage. However, SC-NCM materials still suffer from capacity fading upon extended cycling. This degradation of capacity can be attributed to a reconstruction of the surface. A phase transformation from layered structures to disordered spinel/rock-salt structures was found to be responsible for impedance growth and capacity loss. Film-forming additives are a straightforward approach for the mitigation of surface reconstruction via the formation of a robust protection layer at the cathode’s surface. In this work, we investigate various additives on the electrochemical performance of single-crystal LiNi_0.83Co_0.11Mn_0.06O₂ (SC-NCM83). The results demonstrate that the use of 1% lithium difluoroxalate borate (LiDFOB) and 1% lithium difluorophosphate (LiPO₂F₂) additives substantially enhanced the cycling performance (with a capacity retention of 93.6% after 150 cycles) and rate capability in comparison to the baseline electrolyte (72.7%) as well as electrolytes using 1% LiDFOB (90.5%) or 1% LiPO₂F₂ (88.3%) individually. The superior cycling stability of the cell using the combination of both additives was attributed to the formation of a conformal cathode/electrolyte interface (CEI) layer, resulting in a stabilized bulk structure and decreased impedance upon long-term cycling, as evidenced via a combination of state-of-the-art analytical techniques.

1. Introduction

The first commercialization of LIBs was realized by Sony Corporation in 1991 [1]. Over the past three decades, LIBs have been widely used in portable electronics and electric vehicles due to their high energy density. Ni-rich layered oxide materials, LiNi_1−x−yCo_xMn_yO₂ (NCM, 1 − x − y ≥ 0.6), have been identified as one of the most promising cathode materials as a result of the synergistic effects of transition metals, such as their high specific capacity, high operating voltage, adequate rate capability, and reasonable cost [2]. Conventional polycrystalline NCM cathode materials consist of hundreds of nanometer-sized primary particles containing internal grain boundaries. These materials are prone to form intergranular cracks initiated along the grain boundaries during extended battery operations [3], due to anisotropic volume changes during lithium insertion and extraction [4]. Intergranular cracks may propagate from the particle’s bulk structure to its surfaces, especially during high-voltage operation or long-term cycling, ultimately leading to structural collapse [4]. The cracks effectively increase the electrode/electrolyte interface area, exacerbating surface-related degradations such as surface reconstruction, transition metal dissolution, and oxygen release [3], consequentially resulting in capacity loss.

Previous research suggested that single-crystalline NCM (SC-NCM) materials could address the issues of intergranular cracks whilst maintaining structural integrity by eliminating grain boundaries within the particles [4]. Qian et al. reported that SC-NCM622 exhibits a remarkable capacity retention of 94% after 300 cycles at 4.3 V operation, surpassing its polycrystalline counterpart (52%), as a consequence of the lack of intergranular cracks [3]. Similarly, Fan et al. observed that SC-LiNi_0.83Co_0.11Mn_0.06O₂ (SC-NCM83) outperforms its polycrystalline counterpart in terms of cycling performance at both 25 ℃ and 55 ℃ [4]. Furthermore, the morphology of SC-NCM83 particles remains mechanically intact, without any occurrence of cracks even after 600 cycles. This underscores the potential advantages of using SC-NCM over PC-NCM materials in LIB applications. Although SC-NCM83 efficiently removes intergranular cracks, they remain susceptible to surface layer reconstruction from a layered structure to spinel and/or rock-salt-like structures, a phenomenon that is known to worsen with increasing Ni content [5,6]. These structural changes lead to lower Li-ion conductivity, affecting lithiation and delithiation processes, which in turn result in impedance growth and capacity loss [5,6]. During the lithiation/delithation process, transition metals, especially Ni, may migrate to vacant Li sites due to the similar radii of Ni²⁺ and Li⁺ ions. This migration reduces Ni to lower valence states and its dissolution into the electrolyte [6]. The reduction of Ni is typically accompanied by oxygen loss [7]. Depending on the extent of Li loss, Ni migration, and oxygen loss, both spinel and rock-salt-like structures form on the surface of the cathode particles, degrading the capacity of the cells.

To improve the surface stability of NCM materials, several strategies have been proposed, including the formation of stable coating layers [8,9] and doping with cations or anions [10,11]. An alternative strategy involves the use of film-forming electrolyte additives to build a surface protective layer to suppress parasitic reactions on the cathode/electrolyte interface. Various additives have been explored to enhance the electrochemical performance of NCM cathodes, including lithium salt-type additives such as lithium difluoroxalate borate (LiDFOB) [12] and lithium difluorophosphate (LiPO₂F₂) [13], as well as boron-containing additives such as tributyl borate (TBB) [14] and tris(2-cyanoethyl) borate (TCEB) [15], and phosphorous-containing additives such as triphenylphosphine oxide (TPPO) [16] and 2-(2,2,2-trifluoromethoxy)-1,3,2-dioxaphospholane 2-oxide (TFEOP) [17]. Dong et al. reported that the addition of 1.5% LiDFOB into a carbonate-based electrolyte (1 M LiPF₆ in ethylene carbonate (EC)/dimethyl carbonate (DMC)/ethyl methyl carbonate (EMC) (1:1:1, v:v:v)) significantly improved the cycling stability of graphite||LiNi_0.83Mn_0.05Co_0.12O₂ cells from 59.9% to 83.1% capacity retention after 200 cycles in the voltage range 2.8–4.3 V [18]. X-ray photoelectron spectroscopy (XPS) analysis confirmed the formation of a stable CEI layer rich in boron and fluorine species, whilst avoiding the decomposition of LiPF₆ and organic electrolyte components. Shkrob et al. disclosed that the oxidation of DFOB⁻ ions generates difluoroborane dimers, which form strong B−O bonds and an oxide surface, effectively passivating the surface and preventing electrolyte oxidation [19]. Wan et al. confirmed that LiDFOB inhibited oxygen vacancy generation by building strong B-O bonds, effectively suppressing TM migration and surface reconstruction, thus resulting in improved cycling performance [7]. LiPO₂F₂ was identified as another promising additive used to decrease interfacial impedance and suppress electrolyte decomposition via the formation of a stable CEI layer. Zhao et al. reported that the addition of 1 wt% LiPO₂F₂ to a baseline electrolyte (1 M LiClO₄ in EC/EMC (3:7 wt%) enhanced the cycling stability of Li||LiNi_0.5Mn_0.25Co_0.25O₂ cells, achieving 95.3% capacity retention after 100 cycles in the voltage range of 3.0–4.3 V, superior to 40.8% of cells when using the baseline electrolyte [13]. They ascribed the improved cyclability to the decreased cell impedance and suppressed electrolyte decomposition because of the compact CEI layer. Li et al. showed that the incorporation of 1% LiPO₂F₂ in a baseline electrolyte (1 M LiPF₆ in EC/EMC (3:7 wt%)) resulted in a capacity retention of 98.3% after 150 cycles in the voltage range of 2.8–4.45 V, superior to the 86.2% of the cell using the baseline electrolyte [20]. The enhanced performance was attributed to the dense, stable LiF-rich CEI layer.

In practical applications of LIBs, single additives may not always fulfill the requirements; thus, we need to combine additives to harness potential synergistic effects [21,22]. While LiDFOB and LiPO₂F₂ additives have been individually investigated, possible synergistic effects resulting from their combination remain unclear. In this study, the performance of LiNi_0.83Co_0.11Mn_0.06O₂ (SC-NCM83) in a carbonate-based electrolyte (1M LiPF₆ in EC/DMC) with 1% LiDFOB and 1% LiPO₂F₂, as well as with both 1% LiDFOB + 1% LiPO₂F₂, was investigated. We reveal that cells containing both LiDFOB and LiPO₂F₂ additives exhibit superior cycling stability and rate performance compared to cells containing the baseline electrolyte or combined with a single additive, suggesting synergistic effects between LiDFOB and LiPO₂F₂. The underlying mechanism was examined via electrochemical impedance spectroscopy (EIS), X-ray photoelectron spectroscopy (XPS), and transmission electron microscopy (TEM). The results indicate that a combination of LiDFOB and LiPO₂F₂ additives efficiently suppresses impedance growth whilst maintaining bulk structural integrity, thus inhibiting electrolyte decomposition upon cycling via the formation of a robust and conformal CEI layer.

2. Materials and Methods

2.1. Materials

The synthesis followed a methodology previously reported in the literature [4]. The Ni_0.83Co_0.11Mn_0.06(OH)₂ precursor was prepared via a co-precipitation method. Aqueous solutions (2M) of NiSO₄⋅6H₂O, CoSO₄⋅7H₂O, and MnSO₄⋅5H₂O, with a molar ratio of Ni:Co:Mn of 83:11:06, were simultaneously introduced into a continuously stirred tank reactor under a nitrogen atmosphere. Subsequently, appropriate amounts of NaOH solutions (5M) and NH₃⋅H₂O solutions (4M) were fed separately to the reactor in order to adjust the pH (11.5) whilst stirring (500 rpm) at a temperature of 50 $\mathrm{℃}$. The co-precipitated Ni_0.83Co_0.11Mn_0.06(OH)₂ particles were subsequently washed, filtered, and dried in a vacuum oven at 110 $\mathrm{℃}$. Next, the precursor was mixed with LiOH⋅H₂O at a Li:TM ratio of 1.06:1 and annealed at 500 $\mathrm{℃}$ for 5 h, followed by calcination at 830 $\mathrm{℃}$ for 10 h in an oxygen atmosphere to obtain the SC-NCM83 particles.

2.2. Characterizations

XRD was employed to investigate the crystalline structure of SC-NCM83 particles. The Bragg angle (2$\mathsf{\theta }$) was scanned from 10°–90° at a rate of 1°/min. Scanning Electron Microscopy (SEM, Manufacturer: FEI, Hillsboro, OR, USA) was used to examine the morphology and microstructure of the SC-NCM83 particles. The CEI composition of cycled NMC811 cathodes was analyzed using XPS on a PHI Quantum 2000 (Manufacturer: Physical Electronics, Chanhassen, MN, USA) with a pass energy of 30 eV. The surface structure of the cycled cathode was performed on transmission electron microscopy (TEM, 2100F, JEOL). TEM samples were prepared by FIB (FEI Helios DualBeam, Manufacturer: ThermoFisher, Waltham, MA, USA). The HAADF-STEM imaging was performed on an JEOL ARM 200CF (Manufacturer: JEOL, Tokyo, Japan) with an operation voltage of 200 kV. Before any characterization, the cycled cathodes, harvested from fully discharged cells after cycling, were washed 3 times in DMC solvent to remove residual salts.

2.3. Electrochemical Measurement

To prepare the cathode, a homogeneous slurry was formed by mixing 90 wt% SC-NCM83, 5 wt% carbon black, and 5 wt% polyvinylidene fluoride (PVDF) in N-Methyl-2-pyrrolidone (NMP). The slurry was subsequently coated onto an aluminum current collector and dried for 12 h under vacuum conditions at 100 °C. Additionally, 2032-type coin cells were assembled in an argon-filled glovebox with Li metal as the anode and polypropylene as the separator. 1M LiPF₆ in EC/DMC (3:7 wt%) was chosen as the baseline electrolyte whilst exploring the addition of different additives, namely 1% wt LiDFOB, 1% wt LiPO₂F_2, and a combination of 1% wt LiDFOB + 1% wt LiPO₂F₂. The mass loadings of the active cathode materials were in the range of 5.4–9.2 mg cm⁻².

Galvanostatic charge/discharge tests were conducted at room temperature within a 2.7–4.4 V voltage range using a BioLogic electrochemical workstation. For the cycling test, all coin cells underwent two initial formation cycles at 0.1 C/0.1 C rate, followed by 0.2 C charging/0.5 C discharging rate (1 C = 200 mA g⁻¹). EIS was employed to investigate the impedance variation of the cells after charging to 4.3 V after the 1st, 10th, 50th, and 100th cycles. EIS measurements used the potentiostatic mode with a sinusoidal amplitude of 10 mV and a frequency range from 10 mHz to 10 kHz.

3. Results and Discussions

The SC-NCM83 cathode was synthesized by co-precipitation and subsequent calcination at 830 °C in an O₂ atmosphere (see Experimental section for details). SEM images at different magnifications of the SC-NCM83 particles are shown in Figure 1a,b, indicating an average particle with a size of 2–3 μm. In Figure 1c, the X-ray diffraction (XRD) pattern displays a higher degree of (003)/(104) facet intensity ratio along with apparent splitting of the (006)/(102) and (018)/(110) reflections, indicating a well-ordered layered structure. From Rietveld refinement, the lattice parameters are extracted and shown in

Table 1. Lattice parameters of the SC-NCM83 cathode from Rietveld refinement analysis of XRD data.

	a (Å)	b (Å)	c (Å)	c/a	Vol(Å3)	Ni in Li Layer (%)
SC-NCM83	2.875 (3)	2.875 (3)	14.187 (3)	4.934 (1)	101.580	2.42

. It is clear that the 2.42% intermixing of Li⁺ and Ni²⁺ in the structure (Ni_Li defects) further suggests well-ordered layered structure formation in SC-NCM83. The HAADF-STEM image in Figure 1d confirms the lack of internal porosity, which would benefit the bulk structural stability of the particles in long-term cycling. The highly ordered SC-NCM83 cathode material is corroborated from selected area in Figure 1d by high-resolution HAADF-STEM imaging (Figure 1e) and the corresponding fast Fourier transform (SAED, Figure 1f).

Figure 2a,b shows the cycling performance of cells using baseline electrolytes and baseline electrolytes with different additives. The coulombic efficiency exhibits a similar trend in all cases, as demonstrated in Figure 1a. All cells deliver similar initial specific capacity, while cells using additives show visibly improved cyclability compared to cells using only the baseline electrolyte. The cell using a combination of LiDFOB + LiPO₂F₂ additives shows the best performance, exhibiting 93.7% capacity retention after 150 cycles, in contrast to the cell using a baseline electrolyte, which retains only 72.7% of its initial capacity after 144 cycles. The cell with LiDFOB additives shows 90.5% capacity retention after 150 cycles, while the cell with LiPO₂F₂ additives shows only 88.3%. The cells containing additives show slightly lower initial discharge capacities but more stable cyclability compared with the cells using only the baseline electrolyte. The pronounced stability of cells containing additives may be attributed to the fact that the additives are preferentially oxidized to form a stable CEI layer along with the consumption of some active Li⁺, thus effectively improving the electrode/electrolyte interphase stability. In Figure 2b, the cell using baseline electrolyte displays inferior cycling stability compared to the cells using additives. Figure 2c shows a comparison of the rate capabilities of the investigated cells. All cells deliver similar specific capacity at current densities below 5C (1C = 200 mA g⁻¹) rates. However, as the current density increases, specifically at a rate of 5C, the cells using baseline electrolyte and LiPO₂F₂ additives exhibit an abrupt drop in capacity, while the cells using a combination of LiDFOB + LiPO₂F₂ additives or LiDFOB additives still maintain a comparable capacity. This suggests that the combination of LiDFOB + LiPO₂F₂ additives improves the Li-ion transport kinetics on the electrode surface, thus enabling better performances at high rates, e.g., at 5C.

To confirm the underlying reason leading to different cycling performances, the evolution of the cell’s charge-discharge profiles upon cycling was examined, as shown in Figure 3. All cells show a similar Coulombic efficiency of around 87% in the first cycle, as shown in

Table 2. Coulombic efficiency of the cells using baseline and different additives.

Coulombic Efficiency (%)
Electrolyte	Baseline	1% LiDFOB + 1% LiPO2F2	1% LiDFOB	1% LiPO2F2
1st cycle	87.7	87.5	87.6	87.4

. This indicates that irreversible reactions occur during the initial charging process, which may be attributed to the electrolyte’s involvement in the formation of a solid electrolyte interphase (SEI) and a CEI layer. The cells with additives show slightly lower Coulombic efficiencies in the 1st cycle, which may be attributed to the participation of the additives in the formation of SEI and CEI layers. Upon cycling, the cell using baseline shows rapid discharge voltage decay, indicating significant polarization as a result of the degradation of the interface. In sharp contrast, cells with additives—particularly those with LiDFOB + LiPO₂F₂—show considerably less voltage decay, suggesting a stable electrode/electrolyte interface contributing less pronounced polarization and, thus, higher capacity retention.

To elucidate the origin of the improved cycling stability and rate performance of the cells using additives, EIS measurements were conducted after a specific number of cycles, as shown in Figure 4. The Nyquist plots exhibit two semicircles and a tail. The semicircle in the high-mid frequency range represents surface film resistances (R_sf) during the charging/discharging of the electrochemical double layer, while the semicircle in the mid-low frequency range corresponds to the charge-transfer resistances (R_ct) [23]. The tail at low frequencies can be attributed to Li⁺ diffusion into the bulk electrode material. A corresponding circuit is shown in Figure 4c, where R_b signifies the equivalent series resistance, CPE1 and CPE2 denote constant phase elements to model the non-ideal capacitive behavior in the electrode, and Wo1 represents the Warburg element, accounting for Li⁺ diffusion in the low-frequency range. Fitting results for R_sf and R_ct are summarized in

Table 3. R_sf and R_ct of the cells using different electrolytes after a certain cycle number extracted from EIS measurements.

Cycle	10th		50th		100th
Impedance	Rsf (Ω)	Rct (Ω)	Rsf (Ω)	Rct (Ω)	Rsf (Ω)	Rct (Ω)
Baseline	42	22	67	141	75	255
LiDFOB	36	9	69	24	95	34
LiPO2F2	65	29	65	33	73	39
LiDFOB + LiPO2F2	28	6	37	21	51	24

. The cell using baseline electrolyte shows large R_sf and R_ct, which further increase upon cycling. The growth of R_ct is particularly noticeable, as it jumps from 22 Ω at the 10th cycle to 255 Ω at the 100th cycle. This rise in resistance may be attributed to the accumulation of a poorly conductive surface film on the electrode/electrolyte interface [24,25]. Compared to the cell using baseline electrolyte, the cell using LiDFOB or LiPO₂F₂ additives individually delivers reduced R_ct, although showing comparable R_sf. Thus, the overall impedance is decreased when utilizing LiDFOB or LiPO₂F₂ additives individually as additives. The cell using a combination of LiDFOB + LiPO₂F₂ additives shows significantly decreased R_sf and R_ct, as well as a suppressed increase upon cycling with respect to the other cells. These observations suggest that a combination of LiDFOB + LiPO₂F₂ additives contributes to the formation of a stable and conductive CEI layer, which accounts for the improved cycling performance and rate capability.

To understand the underlying mechanism of improved cycling performance of the SC-NCM83 cathode upon addition of additives, XPS measurements were performed and fitted to study the CEI layer composition on the cycled electrodes, as shown in Figure 5. The C 1s spectra of all cycled cathodes show characteristic peaks of C-C (284.8 eV), C-H (285.5 eV), C-O (286.7 eV), C=O (287.9 eV), and Li₂CO₃ (291.1 eV). The C-C and C-H species primarily originate from the conductive carbon (carbon black) and PVDF binder, while the C-O, C=O, and Li₂CO₃ species can be ascribed to electrolyte decomposition. Compared to the material cycled in the baseline electrolyte, cathode cycled in the electrolyte containing additives, particularly the combination of LiDFOB + LiPO₂F₂, shows much less intense C-O, C=O, and Li₂CO₃ signals, indicating a mitigation of the electrolyte decomposition. In all cases, O 1s spectra show the occurrence of species containing M-O (529.2 eV, M=Ni, Co, Mn), B-O (531.6 eV), C=O (532.2 eV), C-O (533.3 eV), and Li_xPO_yF_z (534.4 eV) on the cycled electrode. The B-O peak is only present in the cells containing LiDFOB as an additive. The cycled cathode with the electrolyte containing a combination of LiDFOB + LiPO₂F₂ additives shows intense B-O and weak Li_xPO_yF_z signals, indicating the formation of B-enriched CEI and the suppression of solvent decomposition. From a comparison of the F 1s spectra, the cycled cathode using additives shows the suppression of LiF (685 eV) and Li_xPO_yF_z (686.8 eV) signals compared to the cathode cycled in the baseline electrolyte. The formation of highly resistive LiF results in sluggish Li⁺ transport kinetics, ultimately resulting in capacity fading [26,27]. The C-F species (688.0 eV) originates from the PVDF binder, while the higher intensity indicates a smaller amount of decomposed electrolyte covering the cycled electrodes. The P 2p spectra exhibit characteristic peaks of P-O/P=O (134.3 eV) and Li_xPO_yF_z (136.4 eV) in all cells. However, the amount of Li_xPO_yF_z is substantially reduced when using a combination of LiDFOB + LiPO₂F₂ additives, further validating the suppression of LiPF₆ decomposition. B 1s spectra show that a considerable amount of B-O (191.6 eV) and Li-B-O (192.9 eV) species are formed on the cycled cathode surface, consistent with observations at the O 1s edge.

The possible impact of electrolyte additives on the CEI layer formation after 150 cycles was studied via high-resolution TEM to track changes in the surface morphology of the cycled cathodes. Figure 6 shows TEM images of the pristine and cycled SC-NCM83 particles in different electrolyte compositions. In Figure 6a,b, the pristine electrode shows a smooth surface, indicating no Li residual species formed during storage in a dry environment. The particles cycled in baseline electrolyte do not show any visible CEI layer, as judged from low (Figure 6c) as well as highly magnified (Figure 6d) high-resolution TEM images. As a consequence, surface corrosion is observed in Figure 6d, which is known to be one of the reasons responsible for the performance degradation in terms of cycling stability [12]. Thus, we were able to confirm the absence of CEI layer formation on the cycled NCM cathode particle in carbonate-based electrolytes via cryogenic electron microscopy to preserve the native state [28]. This finding helps to explain why a well-wrapped CEI layer is absent on the cycled cathode surface despite intense oxidative reactions occurring at the interface. In sharp contrast, the surface of the cycled particles exhibits a visible CEI layer with a thickness of approximately 10–30 nm, which is beneficial in maintaining an intact structure of the SC-NCM83 cathode particles by preventing corrosion as a result of contact with the acidic electrolyte. The particles cycled in the electrolyte containing LiDFOB or LiPO₂F₂ additives individually exhibit an inhomogeneous formation of a CEI layer (Figure 6e,g). This inhomogeneous thickness of the CEI layer is rationalized to hinder Li-ion transfer during battery operation, thus causing sluggish Li-ion transfer kinetics and resulting in high impedance. On the other hand, electrodes cycled in the electrolyte containing a combination of 1% LiDFOB + 1% LiPO₂F₂ additives, the formation of a mostly uniform CEI layer with a homogeneous thickness of ~10 nm, can account for the measured suppressed impedance increase, enhanced cycling stability, and superior rate performance compared to cells using single or no additives.

4. Conclusions

The Li||SC-NCM83 cells using an electrolyte containing a combination of 1% LiDFOB + 1% LiPO₂F₂ additives significantly improve capacity retention (93.7% after 150 cycles) and rate performance at 4.4 V operation, superior to the electrolyte without or with the individual use of additives. The improvement is attributed to a less-pronounced formation of LiF and B-enriched protective CEI layers on the surface of cycled SC-NCM83, as it suppresses charge transfer impedance (R_ct) growth and prevents cathode corrosion that is induced by the electrolyte decomposition. A more uniform and compact CEI layer of ~10 nm was found to form on cathode particles being cycled in the electrolyte containing a combination of LiDFOB + LiPO₂F₂ additives. Additionally, the formation of a robust and conductive CEI layer was found to be helpful to maintain faster Li-ion transfer kinetics and to achieve a good rate capability. Moreover, it alleviates the continuous decomposition of the baseline electrolyte at the particle interface via the formation of a protective layer, thus maintaining an intact and stable bulk structure of the SC-NCM83 particles. To summarize, we can safely claim that the use of a combination of LiDFOB + LiPO₂F₂ additives is a promising strategy to improve the overall performance of Ni-rich SC-NCM cathode LIBs. Thus, it may be worth putting more effort into the optimization of the combination of electrolytes in order to boost the cycling stability and rate capability at high-voltage operation of Ni-rich SC-NCM via the in-situ formation of a passivating CEI layer whilst maintaining Li⁺ diffusion properties and bulk structural integrity.