Unveiling the Advantages of Silicon Nitride Nanoparticle Anodes for Enhanced Cyclic Stability, Rate Performance, and Mechanical Robustness

Abstract

Silicon nitride, known as a convertible-type silicon-based anode material, has emerged as a promising alternative to pure Si anodes, featuring improved cyclic stability, rate performance, and kinetics. This study reports on the electrochemical properties of silicon nitride nanoparticles as an anode material. It demonstrates that this anode outperforms a pure silicon anode in terms of cyclic stability and kinetics. Silicon nitride undergoes conversion during initial lithiation, forming a matrix phase that facilitates charge carrier transport that enhances performance. As a result, silicon nitride retains 73% of its initial charge capacity, whereas only 55% for pure silicon, after 350 cycles. The in situ-formed ion-conductive matrix promotes Li-ion transport, yielding an improved rate performance. At 1 C rate, silicon nitride achieves 585 mA h g^–1 (38% of C/20 capacity) after 85 cycles, surpassing pure silicon’s 470 mA h g^–1 (22% of C/20 capacity). Electrochemical impedance indicates silicon nitride’s faster ionic conductivity and lower resistance compared to pure silicon. Electrochemical dilatometer findings show less electrode thickness increase in silicon nitride (29%) than in pure silicon (60%) during initial lithiation. Silicon nitride demonstrates potential as an attractive anode material for future Li-ion batteries due to improved cyclic stability, superior rate performance, and stable electrode geometry.

1. Introduction

Silicon (Si) has long been considered as a promising anode material in the development of high-performance next-generation lithium-ion batteries due to its high theoretical capacity (3579 mA h g^–1), which is over ten times higher than a conventional graphite anode (372 mA h g^–1).¹⁻³ However, its practical implementation has been hampered by poor cycling stability and limited high-rate performance. The rate limitation is linked to its intrinsic properties, while the severe cyclic stability challenge is caused by substantial volume expansion (∼300%) during charge–discharge processes, resulting in mechanical stress, pulverization, and eventual electrode degradation.⁴⁻⁶ These challenges are often met by nanostructuring and nanoengineering to minimize the strain during expansion⁷ and/or by surface modification to ensure good interparticle contact and conductivity.⁸ While these methods can yield promising results on the lab scale, they often are more difficult to implement in a large-scale industrial process. One thing is the added cost due to additional processing steps, but in the case of nanostructuring silicon anodes, the increased surface area of small particles only exacerbates the already problematic parasitic surface reactions, thus reducing the obtainable long-term Coulombic efficiency (CE).⁹

Recently, a group of silicon-based anode materials have emerged with some promise in resolving the long-term cycling instability of silicon.¹⁰ These materials include nonstoichiometric oxides, carbides, and nitrides of silicon (SiA_x, A = O, C, and N). The common feature of this family of materials is that they undergo in situ conversion reactions when they react with lithium ions during the initial lithiation process, generating active silicon with a new matrix phase (eq 1). In the subsequent cycles, the Li_zSi nanodomains reversibly alloy/dealloy with Li (eq 2), while the Li-conductive matrix ensures good stable cycling and rate performance.^11,12

1

2

Accordingly, these anodes have shown that by sacrificing some lithium in the early cycles to form stable matrixes, one can minimize the mechanical strain during lithiation and consequently ensures more stable cycling.¹³ For instance, nonstoichiometric silicon oxide (SiO_x) with different substoichiometries has been studied extensively and showed promising cyclic stability;^14,15 however, it suffered from low-rate performance.¹⁶⁻¹⁸ Because silicon-based materials are semiconductors, the low electrical conductivity and ionic transport significantly limit the rate of performance of these materials. Various approaches have been suggested to improve the conductivity through carbon coating/Ti doping¹⁹ and artificial solid–electrolyte interface (SEI) formation.²⁰ Despite the efforts that have been made, all these attempts are not only expensive due to imposing additional processing steps but also challenging for scale-up production.

On the contrary, SiN_x has garnered significant attention due to its demonstrated ability to enhance conductivity while exhibiting excellent cyclic stability.¹³ This performance is owed to the composition of the matrix generated from the SiN_x conversion reaction, which has been established to consist of components of ternary lithium silicon nitride (e.g., Li₂Si₂N) and lithium nitride (Li₃N), which have high Li-ion conductivity. Such matrix composition could improve the pathways for lithium ions, thereby increasing the potential charging rate as compared to other silicon-based anodes.^13,21,22 In this aspect, it is important to fully understand the effect of the conductive matrix on the desirable cyclability and an enhanced rate capability of the SiN_x anode. Recently, there have been some studies using silicon-rich silicon nitride nanoparticles integrated into carbon nanotube structures to improve rate capability.⁸ However, to the best of our knowledge, few studies have investigated the intrinsic transport properties of silicon nitride anodes. It is the aim of this study to provide electrochemical investigations on the kinetic properties and stability of the silicon nitride anode.

2. Materials and Methods

2.1. Synthesis of Silicon and Silicon Nitride Nanoparticles

Pure silicon (Si) and substoichiometric silicon nitride (SiN_x) nanoparticles were produced using a gas-phase free space reactor (FSR). The amorphous Si nanoparticles were synthesized by pyrolysis of silane gas (SiH₄) in the FSR at a temperature range of 575–675 °C as described elsewhere.²³ For the synthesis of amorphous and substoichiometric SiN_x nanoparticles, a similar procedure is modified by introducing ammonia (NH₃) as a nitrogen-containing precursor into the reactor, with a SiH₄:NH₃ flow rate ratio of 3:1 at 650 °C as described elsewhere.²⁴ After the completion of the pyrolysis reactions, the nanoparticles are collected and characterized.

2.2. Materials Characterization

Field-emission scanning electron microscopy (SEM, JSM-7901F, JEOL) analysis was conducted to investigate the morphology and particle size of the fabricated nanoparticles. Elemental maps were acquired using energy-dispersive X-ray spectroscopy (EDS, Aztec, Oxford Instruments) to analyze the composition of the nanoparticles and to estimate the nitrogen content and its distribution in the SiN_x nanoparticles. For low-voltage scanning transmission electron microscopy (LV-STEM), samples were prepared on lacey carbon TEM grids and examined at a 30 kV accelerating voltage. Fourier transform infrared (FTIR) spectroscopy analysis was recorded with a 532 nm laser source using a Bruker Alpha spectrometer in attenuated total reflectance mode for further qualitative analysis on the compositions of the synthesized nanoparticles.

2.3. Electrochemical Characterization

The slurries of Si and SiN_x were prepared by mixing the active materials, graphite (KS6L, TIMCAL), carbon black (C65, IMERYS), and sodium carboxymethylcellulose binder (CMC, M_W = 90 K, Sigma-Aldrich) with a weight ratio of 60:15:10:15 in aqueous buffer solvent using a planetary centrifugal mixer (Thinky). A buffer solution (pH = 3), prepared from citric acid and potassium hydroxide in deionized water, was used as a solvent. First, the CMC binder was dispersed in the buffer solution and mixed for 2 min at 1000 rpm, followed by 5 min at 2000 rpm. Then, the carbon black and graphite conductive additives were incorporated into the binder solution, and the above-mentioned mixing protocol was repeated. After achieving a well-dispersed conductive additives–binder solution, the active material (Si or SiN_x) was added and mixed for 2 min at 1000 rpm, followed by 5 min at 2000 rpm. To make sure the slurry is homogeneously dispersed, the mixing was carried out for an additional 20 min at 2000 rpm. The resulting slurry was coated onto a 20 μm thick copper foil current collector using a cylindrical film applicator with a resulting loading of active materials of 0.51 mg cm^–2 ± 0.01 mg cm^–2. The electrodes were allowed to dry under ambient conditions, followed by drying in a vacuum oven at 120 °C overnight to remove any remaining solvent and moisture. The dried electrodes were punched into 15 mm diameter discs, weighted and transferred into an argon-filled glovebox (O₂ < 0.1 ppm, H₂O < 0.1 ppm) (MBRAUN) through a vacuum chamber at 120 °C for 4 h. The electrodes were then assembled with CR2032 (Hohsen)-type coin cells in a half-cell configuration using the fabricated electrodes as the working electrode, a lithium metal foil (Gelon) of 15.6 mm diameter (∼50.4 mAh/cm²) as both the counter and reference electrode, and an 18 mm diameter Celgard 2400-type microporous membrane (25 μm thickness), together with an 18 mm diameter Viledon (FS 2207-25 DA WA) (250 μm thickness) as separators. 1.2 M lithium hexafluorophosphate (LiPF₆) in ethylene carbonate/ethyl methyl carbonate (EMC) (3.7 v/v) containing 10 wt % fluoroethylene carbonate and 2 wt % vinylene carbonate (Solvionic) was used as the electrolyte (85 μL). The electrochemical performance of the half-cells coin cells was tested using an Arbin MSTAT battery tester. The cells were cycled in the potential range between 0.05 and 1.0 V vs Li/Li⁺ at a current density of C/20 (132 mA g^–1) for the first three formation cycles and 0.2 C (528 mA g^–1) for the subsequent cycles. A constant-voltage (CV) step was imposed at the end of each cutoff potential until the measured absolute current reached to C/100 (26 mA g^–1) to ensure the completion of lithiation/delithiation. For the rate capability tests, a series of constant current (CC) was applied without CV steps. All the specific capacities values are calculated based on the active materials masses (Si or SiN_x, including the graphite additive).

The electrochemical impedance spectroscopy (EIS) measurements were conducted using a three-electrode cell setup (PAT, EL-CELL, GmbH, Germany). The construction of the cell followed the procedure outlined in the PAT, EL-CELL user manual (release 2.6). The working electrode had a diameter of 15 mm, while a 15.6 mm diameter lithium metal foil (Gelon) was used as the counter electrode. An insulating sleeve comprised of a polypropylene (PP), a lithium (Li) reference ring, and a borosilicate glass fiber (GF/A) separator was employed as the separator and reference electrode. 100 μL of the same electrolyte used for the coin cell was added onto the separator. Subsequently, the assembled PAT-CELL was attached to a PAT docking station, which was connected to a VSP-300 multichannel potentiostat from Biologic. The EIS test was conducted by applying an alternating current amplitude of 10 mV within a frequency range from 10 mHz to 200 kHz, in a climate chamber where the temperature was maintained at 25 °C.

Electrochemical dilatometry was performed using an ECD-nano-3 instrument (EL-CELL, Germany). The cell was assembled according to the instructions created by EL-CELL (manual v1.6). A 10 mm working electrode was punched by an EL-CELL hand punch. The counter and reference electrodes were made of Li metal. A 15.6 mm Li metal chip (Gelon) was cleaned with a scalpel, followed by cutting with a hand punch into a 12 mm diameter disc. The reference electrode was prepared by pressing the reference electrode cavity into excess Li metal from cutting. The reference electrode was then tapped until smooth, and the excess Li on the reference was trimmed by a scalpel, taking care to avoid touching the Teflon gasket with Li. A 12 mm Whatman GF/A glass fiber separator is used with the Li counter electrode. Midway through the assembly, the electrolyte is added to the glass frit after fitting in the reference electrode and the dead-volume valve but before placing the separator and working electrode. The electrolyte was dispensed in 50 μL increments or less via a micropipette. The electrolyte is added until a visible meniscus of liquid is formed between the glass frit and the PEEK T-joint, yielding an amount of approximately 350 μL of the same electrolyte as that used for coin cells. A Celgard 2400 separator with a diameter of 10 mm is used on the working electrode side. Finally, the working electrode (thickness ∼13 μm, excluding the current collector) is placed on top followed by a spacer and dilatometric membrane. The dilatometer cell is then secured to the frame, the sensor is mounted on top, and the signal connections are made according to the instructions created by EL-CELL (manual v1.6). Finally, the dilatometer measurement data are recorded upon cycling at C/20 (132 mA g^–1) rate with a BioLogic SP-300 potentiostat in a climate chamber where temperature was maintained at 25 °C.

3. Results and Discussion

It was of interest to investigate the SiN_x nanoparticle anode, while the pure Si nanoparticle anode was added for comparison. As demonstrated in the methodology section, both materials have been synthesized using an in-house chemical vapor deposition FSR using silane gas (SiH₄) as a silicon precursor and ammonia (NH₃) as a nitrogen source for the case of SiN_x synthesis. Figure 1 shows the SEM (a,c) and LV-STEM (b,d) images of the as-synthesized Si and SiN_x nanoparticles. One can see that the synthesized nanoparticles exhibit roughly a spherical morphology, with a primary particle size ranging from 50 to 100 nm. It is also noticeable from the LV-STEM images that the primary nanoparticles of both Si (1b) and SiN_x (1d) are somewhat agglomerated, forming clustered nanoparticle morphologies. EDS analysis revealed the presence of nitrogen in SiN_x of 18 at. % (Figure 1c, table inset).

Figure 1.

SEM (a,c) and LV-STEM (b,d) images of pure Si nanoparticles and SiN_x nanoparticles. The inserts on (b,d) are the respective higher magnification LV-STEM image, while the inset table on (c) is the EDS atomic percentage of SiN_x.

The FTIR measurements were performed in the mid-infrared region for further insights into the composition of the nanoparticles (Figure 2). The SiN_x composition was evidenced through a well-pronounced peak at around 854 cm^–1 attributed to Si–N asymmetric stretching.²⁵

Figure 2.

FTIR spectra of pure Si nanoparticles (black) and SiN_x nanoparticles (red) measured with a 532 nm laser source.

Figure 3 presents the summarized electrochemical performance of the pure Si nanoparticles and SiN_x nanoparticles, which began with three formation cycles at C/20 (132 mA g^–1) followed by long cycling at C/5 (528 mA g^–1) using a CC, CV (CC–CV) cycling protocol.

Figure 3.

Lithiation–delithiation voltage profiles in the 1st and 3rd cycles (a) and cyclic stability with Coulombic efficiency (b) of Si and SiN_x nanoparticle anodes.

Figure 3a displays the first and third lithiation (discharge) and delithiation (charge) voltage profiles of both pure Si nanoparticles and SiN_x nanoparticles. The initial lithiation profile of Si shows a higher-potential region before the typical lithiation potentials of Si corresponding to the formation of an SEI by the irreversible reduction of the electrolyte on the surface of the active materials. Following the SEI formation, a distinct plateau at potentials below 0.2 V corresponds to the Li alloying reaction with Si. In the case of SiN_x, a potential overshoot directly to around 0.27 V is noted in the early stage of the initial lithiation, indicating the overpotential requirement to activate the formation of a new phase (i.e., onset of conversion reaction). After the onset potentials of the conversion, a prolonged potential plateau follows, corresponding to the Li alloying/dealloying with the Si-based active material generated from the conversion reaction.¹¹

Noticeable differences could easily be distinguished between the two active materials regarding the initial irreversible capacity. SiN_x irreversibly consumed more lithium during its first formation cycle leading to lower initial Coulombic efficiency (ICE = 67.52%) compared to that of Si (ICE = 86.48%). This is associated with the creation of lithium-containing matrix phases, such as Li₃N and Li₂SiN₂.^21,22,26 Despite this initial penalty, the Coulombic efficiency (CE) increased rapidly in the subsequent cycles, reaching 96.67% for SiN_x and 94.57% for Si in the third cycle. This difference illustrates the benefits of the matrix phase as a result of the conversion reaction to allow SiN_x to outperform Si in terms of capacity retention (CR) (Figure 3b). The specific delithiation capacities based on the weight of active materials (Si or SiN_x, plus the graphite additive, Gr) for the first and third formation cycles are 2426 mA h g^–1 and 2279 mA h g^–1 for Si and 1859 mA h g^–1 and 1800 mA h g^–1 for SiN_x, respectively. The specific capacity values are calculated on the basis of total mass of active materials (Si/SiN_x + graphite). It is evidenced that SiN_x showed a drop in capacity by 3%, while Si experienced a 6% capacity drop between the first and third cycles. Further cycling demonstrates a faster drop in capacity for pure Si with its CR of only 55% after 350 cycles, while SiN_x maintained 73% of its capacity after 350 cycles. This difference elucidates how the incorporation of nitrogen into the silicon matrix imparts structural stability and mitigates the detrimental effects of the volume expansion. The unique composition of SiN_x results in a robust matrix, preserving the structural integrity of the electrode during repeated cycling processes. Notably, SiN_x shows a larger dip in capacity in the early cycles before slowly regaining much of the capacity in the later cycles and stabilizing. Such effects have previously been observed and been related to increased electrolyte penetration/electrode activation during operation.²⁷ For SiN_x, this electrode activation can, in addition to electrolyte penetration, be explained by increased ion conductivity from the formation of the Li₃N/Li₂SiN₂ matrix during the initial cycles. After some cycles, the capacities stabilize, while for pure Si, the decay continues. This stability suggests that through proper prelithiation techniques (or an overlithiated cathode) to compensate the initial irreversible capacity, one could benefit from the high capacity while still minimizing capacity loss during prolonged cycling for silicon nitrides. For silicon anodes, the advantage of prelithiation is more limited, as some degree of degradation occurs every cycle due to continuous cracking and SEI reformation. The differential capacity (dQ/dV) curves of Si and SiN_x for the first cycle are shown in Figure 4a. The conversion reaction of SiN_x can be seen in the dQ/dV during the first lithiation at a potential around 0.31 V. The SiN_x cathodic peak (0.31 V) represents an overpotential compared to the Si cathodic peak (0.37 V), which is consistent with the observed potential overshoot in the initial lithiation voltage profile (Figure 3a). Following the activation of the conversion reaction and SEI formation, two distinct cathodic (P₁ ∼ 0.2 V and P₂ ∼ 0.07 V) and two distinct anodic (P₃ ∼ 0.3 V and P₄ ∼ 0.5 V) peaks are observed, which is associated with the two-step lithiation and two-step delithiation processes for both materials. The two minor but distinct anodic peaks labeled “Gr” are observed at 0.13 and 0.16 V and are attributed to the delithiation of graphite which was used as a conductive additive in the electrode. The corresponding graphite cathodic peaks were not visible, likely overshadowed by the larger Si and SiN_x peaks in the overlapping potential range. During subsequent cycling (Figure 4b for Si and 4c for SiN_x), the dQ/dV curves for pure Si and SiN_x reveal notable differences. While the SiN_x peaks have maintained their positions upon cycling with only minor shifts, the Si peaks have shown significant shifts in position, indicating increased polarization and unfavorable cycling stability. The noticeable drop in the starting lithiation potential for pure Si especially after 10 cycles indicates its faster capacity degradation. These differences in potential shift can be inferred to mean that the SiN_x electrode exhibits improved structural stability.

Figure 4.

(a) Differential capacitance (dQ/dV) curves of Si (black dashed line) and SiN_x (blue solid line) for the 1st cycle. (b,c) dQ/dV curves at different cycle numbers for Si and SiN_x, respectively.

To investigate the rate capability of SiN_x, rate tests have been performed at C rates of C/20, C/10, C/5, C/2, 1C, and 2C (Figure 5). As seen from the initial rate test results, SiN_x preserved 32% (493 mA h g^–1) of its C/20 specific capacity at 1 C. On the other hand, Si maintained 30% (789 mA h g^–1) of its C/20 specific capacity at 1 C. When the rate measurement is repeated after 85 cycles, SiN_x preserved 38% (585 mA h g^–1) of its C/20 specific capacity. In contrast, Si maintained only 22% (470 mA h g^–1) of its C/20 specific capacity at 1 C, implying that SiN_x showed enhanced rate performance upon cycling, while the Si rate performance worsened (Figure 5C). It is well-documented that a silicon anode has a poor rate performance in addition to its poor cyclic stability. One of the key factors contributing to the superior rate performance of SiN_x is believed to be the formation of a conductive matrix containing lithium-nitride-like phases (Li₃N, a well-known ionic conductor) as a product of the conversion reaction between SiN_x and lithium ions, which acts as a conductive pathway for lithium-ion transport. This conductive network facilitates rapid charge and discharge processes, thereby enhancing the rate capability of the overall anode. A potential reason for the increased rate performance of SiN_x upon cycling could arise from the modification of the matrix phase toward a more conductive composition and/or the interconnection of the matrix domains over time, forming a continuous conductive pathway. The improved kinetic behavior reduces the overall impedance (Figure 6), leading to enhanced electrochemical performance, especially at high charge/discharge rates.

Figure 5.

Charge rate capability of Si and SiN_x anodes: initial rate (a, left) and rate measured after 85 charge–discharge cycles at C/5 (a, right). The charts demonstrate the impact of cycling on rate performance (b,c).

Figure 6.

Nyquist plots from EIS measured at (a) 0.2 V, (b) 0.1 V, (c) 0.06 V, and (d) 0.05 V vs Li/Li⁺ depict the behavior of both pure Si and SiN_x electrodes measured in three-electrode cells across various SOCs during the initial lithiation cycle.

EIS measurements have been carried out in three-electrode half-cells to further explore the electrochemical kinetics. Figure 6 displays the Nyquist plots of Si and SiN_x at different states of charge (SOC) in the initial lithiation. The EIS of both Si and SiN_x consists of two distinct semicircles at high-to-medium-frequency regions followed by a sloping line at a low-frequency region. The first semicircle in the high-frequency region corresponds to the typical SEI resistance (R_SEI), while the second semicircle in the medium–high-frequency region is typically attributed to charge transfer resistance (R_ct).²⁸ The sloping line in the low-frequency region is the reflection of Li-ion diffusion in the electrode. Comparative analysis of the impedance results provides valuable insights into the ionic transport properties and SEI stability difference between Si and SiN_x. The diameters of the two semicircles for SiN_x are lower than those of Si (Figure S1), which would explain the difference at higher C rate (Figure 5).²⁹⁻³¹ Additionally, the EIS analysis for the first five cycles (Figure S1) shows a difference in the stability characteristics between the two materials over cycling. The EIS spectra particularly in the high-frequency region remain relatively stable for SiN_x, with minimal changes observed, suggesting a relatively more stable electrode–electrolyte interface and consistent charge transfer kinetics. On the other hand, pure Si exhibits noticeable alterations, particularly, the disappearance or divergence of the second semicircle indicates a degradation, potentially attributed to structural changes or electrode–electrolyte interface instability. By fitting the EIS spectra with an equivalent circuit (Figure S2), the R_hf was estimated to be about 15.80 Ω for Si but 8.95 for SiN_x (Table S1) at 0.06 V in the third cycle. These lower resistances are due to (1) a stable SEI layer, which reduces impedance by minimizing side reactions, thus implying better protection of the electrode–electrolyte interface against undesired reactions and (2) a lower charge transfer resistance, which indicates a more facile movement of charge carriers across the electrode–electrolyte interface resulting in improved kinetics. Such important properties of SiN_x mainly arise because of the in situ-generated ion-conductive matrix (Li₂SiN₂ and Li₃N) during the initial lithiation process.

To evaluate the stability of SiN_x in terms of volume expansion, electrochemical dilatometry was performed to measure the thickness change during the lithiation/delithiation processes. Figure 7 shows the voltage versus time profile (bottom) and the associated thickness change (top) for both pure Si (black dashed line) and SiN_x (red solid line) anodes for the first, fourth, and fifth cycles. The estimated electrode thickness for both materials before cycling was approximately 13 μm (excluding current collector thickness). In comparison to the Si electrode, the SiN_x electrode displayed reduced thickness increase during all lithiation cycles. The thickness increase for pure Si during first cycle lithiation was measured to be about 7.74 μm (60% of initial electrode thickness), while only a 3.81 μm thickness increase was measured for SiN_x (29% of initial electrode thickness) (Table 1). Knowing these electrodes were not calendered, they are expected to possess significant porosity, which explains the low amount of dilation in the initial stages of lithiation³² and partly why the percentage expansions are lower than the theoretical volume expansion of pure silicon. Additionally, the binder content of these Si electrodes is higher than prior studies, which has been identified as a potential cause for less volume expansion when using Si nanoparticles with high surface area.³³ During the subsequent delithiation, Si contracted 5.19 μm in thickness, while SiN_x contracted 3.03 μm in thickness (Table 1). Comparing the first cycle thickness increase upon lithiation and thickness decrease on delithiation, Si showed 20% irreversible expansion, while SiN_x showed only 6% irreversible expansion. Si continued to expand irreversibly during the fourth and fifth cycles, while SiN_x remained stable, even appearing to slightly shrink after the initial irreversible expansion. The formation of the matrix phase for SiN_x could be the main reason for its lower expansion and its ability to maintain a reversible structure as compared to Si.

Figure 7.

Voltage–time profile (bottom) and associated thickness change (top) of pure Si (black dot line) and SiN_x (red solid line) electrodes during the 1st, 4th, and 5th C/20 cycles.

Table 1. Summary of Electrode Thickness Change Following the Lithiation/Delithiation Capacities of Pure Si and SiN_x Anodes in the 1st and 4th Cycles, as Measured by the Dilatometer.

anode type	lith_Q (mA h/g)		Δthickness (μm)		delith_Q (mA h/g)		Δthickness (μm)
	1st	4th	1st	4th	1st	4th	1st	4th
Si	2573	2107	7.74	5.47	2093	1965	5.19	2.64
SiNx	2345	1525	3.81	2.64	1501	1409	3.03	2.47

4. Conclusions

In summary, the comprehensive study comparing pure Si and SiN_x nanoparticle anodes has provided substantial insights into their electrochemical performance and structural stability. It is inferred from the results that the SiN_x nanoparticle anode exhibited improved cyclic stability, enhanced rate performance especially at higher C rates, and robust mechanical stability compared to its Si counterpart. The SiN_x nanoparticles displayed a CR value of 73%, while it was only 55% for pure Si after 350 cycles. The rate capability measurements evidenced the superior rate performance of SiN_x, emphasizing the role of the conductive matrix phase which formed during the conversion reaction of SiN_x. Furthermore, EIS results revealed a lower impedance for SiN_x, evidencing enhanced charge carrier kinetics. Additionally, dilatometer measurement showed a significantly reduced total expansion for SiN_x-containing electrodes (29%) compared to Si-containing electrodes (60%) during the initial lithiation, as well as a lower irreversible expansion in the first cycle and negligible irreversible expansion in the following cycles. These collective features highlight the potential of the SiN_x nanoparticle anode for the development of next-generation lithium-ion batteries, providing improved CR, enhanced rate performance, and mechanical robustness.

Supporting Information Available

The Supporting Information is available free of charge at https://pubs.acs.org/doi/10.1021/acsomega.4c07583.

• Additional experimental impedance spectra and analysis of Si and SiN_x, including equivalent circuit and fitted data (PDF)

The authors declare the following competing financial interest(s): AU is a co-inventor of several patents and patent applications related to the use of silicon nitride in batteries. While the authors retain no direct rights to these patents or applications, their assignee at the time of publication is IFE (the host institution of the authors).