Gas Evolution Kinetics in Overlithiated Positive Electrodes and its Impact on Electrode Design

Abstract Increasing lithium contents within the lattice of positive electrode materials is projected in pursuit of high‐energy‐density batteries. However, it intensifies the release of lattice oxygen and subsequent gas evolution during operations. This poses significant challenges for managing internal pressure of batteries, particularly in terms of the management of gas evolution in composite electrodes—an area that remains largely unexplored. Conventional assumptions postulate that the total gas evolution is estimated by multiplying the total particle count by the quantities of gas products from an individual particle. Contrarily, this investigation on overlithiated materials—a system known to release the lattice oxygen—demonstrates that loading densities and inter‐particle spacing in electrodes significantly govern gas evolution rates, leading to distinct extents of gas formation despite of an equivalent quantity of released lattice oxygen. Remarkably, this study discoveres that O2 and CO2 evolution rates are proportional to 1O2 concentration by the factor of second and first‐order, respectively. This indicates an exceptionally greater change in the evolution rate of O2 compared to CO2 depending on local 1O2 concentration. These insights pave new routes for more sophisticated approaches to manage gas evolution within high‐energy‐density batteries.


Introduction
[14] This presents considerable challenges for the management of the internal pressure of batteries during operations due to the substantial accumulation of gas products in electrode systems over cycles.Consequently, this exacerbates safety concerns in batteries, such as swelling or rupture of cell casing.In the complex electrochemical environment of a battery's positive electrode, [15] the lattice oxygen is released as a reactive singlet oxygen ( 1 Δ g or 1 O 2 ).This reactive singlet oxygen can either evolve into inert triplet oxygen ( 3 O 2 ) through a selfquenching process or form CO 2 that initiates further electrolyte decomposition reactions. [13,14,16,17]The quantity of 1 O 2 released from each particle is critically determined by the intrinsic properties of the material during cycling. [18,19]However, the local consumption rate of 1 O 2 in each reaction pathway within composite electrodes, particularly in those producing O 2 and CO 2 , holds even greater significance.This rate is crucial in determining the extent of gas evolution reactions and the consequent formation of side-products, which, in turn, drive a variety of decomposition reactions. [13]Despite its paramount importance, the fundamental principle governing the reaction kinetics of gas evolution in composite electrodes characterized by complex and intertwined particle networks remains largely underexplored.This gap in knowledge necessitates a comprehensive understanding in consideration of two major aspects: the response of each gas evolution rates to changes in local 1 O 2 concentration, and the influence of composite electrode configuration (i.e.particle-particle distribution) on the local 1 O 2 environment and the overall gas evolution rates. [14]To precisely establish the relationship between gas evolution rates and 1 O 2 concentration in composite electrodes, it is essential to methodically vary 1 O 2 concentration at the electrode level.This can be achieved by using an overlithiated catthode designed to release a tunable amount of 1 O 2 across various cycling conditions.[22][23][24][25] Consequently, the kinetics of gas evolution should be scrutinized in a scenario where 1 O 2 is abundantly released, thereby minimizing the interference from other types of gas evolution processes.
In this study, we utilize Li 6 CoO 4 -a compound known for the substantial release of lattice oxygen at low potential due to its labile oxygen states [26] -as a model system to probe the kinetics of gas evolution as a function of 1 O 2 concentration, within a voltage range from ≈3.5 to 4.3 V vs Li/Li + that is compatible for conventional battery electrolytes.To discern the correlation between gas evolution rates and 1 O 2 concentration, we quantitatively analyzed the evolution of O 2 and CO 2 at various 1 O 2 levels using online electrochemical mass spectrometry (OEMS).We observed that O 2 evolution, resulting from the self-quenching of 1 O 2 , showed a quadratic relationship with increasing 1 O 2 concentration, indicative of a second-order reaction.In contrast, CO 2 evolution, arising from the interaction of the electrolyte with 1 O 2 , appeared to conform a first-order reaction relative to 1 O 2 concentration.Remarkably, our results revealed that the distribution of products from competing reactions can be influenced by the local 1 O 2 con-centration, which is affected by the distribution of 1 O 2 -releasing particles within a composite electrode, despite the equivalence in average amount.Therefore, we conclude that the extent of interrelated gas evolution reaction is intricately linked to both the 1 O 2 concentration and the configuration of composite electrodes.

Release of Lattice Oxygen in Li 6 CoO 4 During the Initial Charge
[29] Li 6 CoO 4 locally has a distorted LiO 4 tetrahedron in its anti-fluorite structure that leads to a large density of labile oxygen states below the Fermi level, and hence it exploits improved electrochemical activity of lattice oxygen. [26,30]The galvanostatic profile of Li 6 CoO 4 during the initial charge to 4.3 V versus Li/Li + showcases two distinct plateaus at ≈3.3 and 3.6 V vs. Li/Li + , as presented in Figure 1a.The first plateau is ascribed to the extraction of Liions concurrent with the oxidation of Co, which is confirmed by X-ray absorption spectroscopy (XAS) and X-ray absorption near edge spectroscopy (XANES) analyses that characterize the oxidation state of Co (Figure 1a,b).However, the subsequent plateau exhibits a reduction of Co, suggesting an O-to-Co charge transfer process that coincides with the release of lattice oxygen.This process likely contributes to the breakdown of Co−O coordination in Li 6 CoO 4 , as shown by diminishing intensities of peaks A and B that correspond to 1s→4p and 1s→3d transitions, respectively. [31,32]The release of lattice oxygen, which is inferred from these observations, can be deduced from its implication in gas evolution and structural degradation.
To confirm the release of lattice oxygen, we monitored the structural changes in Li 6 CoO 4 alongside gas evolution during the initial charging.Powder X-ray diffraction (XRD) revealed that Li 6 CoO 4 has an anti-fluorite structure with a tetragonal P42/nmc space group as shown in Figure S1 (Supporting Information). [26,27]In situ XRD analysis captured the progressive fading of the main diffraction peaks, which corresponds to (101) and ( 201) planes.These diffraction peaks gradually faded toward the end of the first plateau and completely vanished by the end of the second plateau as shown in Figure 1c.This loss of longrange ordering in the crystal structure aligns with the onset of O 2 and CO 2 (Figure 1d).Therefore, we deduce that the substantial gas evolution and associated degradation in the anti-fluorite structure of Li 6 CoO 4 is predominantly due to the release of lattice oxygen.

Release of Lattice Oxygen in Li 6 CoO 4 during the Initial Charge
Building on our findings that the release of lattice oxygen results at 3.6 V vs. Li/Li + , the quantity of lattice oxygen released per mole of Li 6 CoO 4 should theoretically remain constant under identical operational conditions.However, we found that the degree of gas evolution, which hinges on the reaction rates, appears to be substantially influenced by the concentration of reactants varying across different electrode configurations.To explore this further, we constructed electrodes with Li 6 CoO 4 and NCMA (LiNi x Co y Mn z A1 1-x-y-z O 2 , x >0.85) in two separate configurations: a double-layer composite with densely packed Li 6 CoO 4 , and a blend composite where Li 6 CoO 4 particles are dispersed throughout the electrode volume (Figure 2a; Figures S2 and S3, Supporting Information).The areal loading of the electrode composite was consistently maintained at 1.2 mg cm −2 for all electrodes.The active materials, constituting 97.5% of the electrode composite by weight, consisted of 3% Li 6 CoO 4 and 97% NCMA.The density of Li 6 CoO 4 was 0.077 g cm −3 for the simple blend electrode whereas it was 0.360 g cm −3 for the double-layer electrode, as estimated from the thickness obtained from cross-sectioned SEM images (Figures S2 and S3, Supporting Information).
The comparative analysis of O 2 and CO 2 evolution in these configurations during the initial charging (Figure 2b,c) showed that, despite an equal amount of Li 6 CoO 4 , the double-layer composite electrode exhibits significantly greater evolution of both O 2 and CO 2 .This phenomenon persisted regardless of the Li 6 CoO 4 layer's position within the double-layer structure, suggesting that the effect density of Li 6 CoO 4 -a reflection of particle aggregation -plays a critical role in gas evolution dynamics.With NCMA showing minimal gas evolution (Figure S4, Supporting Information), [14,33] its dilution effect of 1 O 2 from Li 6 CoO 4 is apparent.This allows to deduce that electrodes with varying effective densities and distribution of Li 6 CoO 4 provide distinct environments that influence the local 1 O 2 concentration, which consequently leads to changes in the rates of competing reaction pathways.Therefore, it is imperative to establish the kinetics of gas evolution reactions and comprehend their implications on the electrode design.
To establish a more concrete correlation between the gas evolution and 1 O 2 concentration, we quantitatively assessed the behavior of O 2 and CO 2 evolution across varying 1 O 2 concentrations.We could systematically modulate the 1 O 2 concentration by adjusting the Li 6 CoO 4 density within electrodes, while maintaining areal loading and electrode thickness constant (Figure 3a).Scan-ning electron microscopy (SEM) images confirmed the homogeneous distribution of Li 6 CoO 4 (Figure S5, Supporting Information).All electrodes were charged at a consistent current density of 40 mAg −1 to maintain uniform electrochemical conditions for Li 6 CoO 4 particles within composite electrodes.Our test across all electrodes showed consistent onset potential for gas evolution (O 2 and CO 2 ) and complete capacity utilization for both NCMA and  The relationship between O 2 evolution and Li 6 CoO 4 density during the initial charge is plotted in Figure 3f.A quadratic increase in O 2 evolution was observed with rising Li 6 CoO 4 density, fitting a second-degree polynomial trend.This trend is further substantiated by a linear correlation between the square root of detected O 2 and Li 6 CoO 4 density (Figure 3g).Even with modifications to density by adding conductive carbon instead of NCMA (Figure S6, Supporting Information), the quadratic relationship still persisted.This suggests that the O 2 evolution rate escalates with the Li 6 CoO 4 density while maintaining a consistent rate constant.Therefore, O 2 evolution, driven by the self-quenching of 1 O 2 , adheres to a second-order reaction in relation to the reactant concentration.
Figure 3h illustrates the CO 2 evolution during the initial charge, plotted as a function of the density of Li 6 CoO 4 .In contrast to O 2 evolution, CO 2 evolution exhibits a directly proportional increase with Li 6 CoO 4 density.The NCMA electrode without Li 6 CoO 4 only produces 0.027 μmol of the total CO 2 evolution (Figure S4, Supporting Information), representing a negligible amount in comparison to CO 2 from Li 6 CoO 4 .][36] Consequently, CO 2 evolution mediated by the reaction with 1 O 2 adheres to a first-order reaction.This linear increase also suggests that the initial interaction between 1 O 2 and electrolyte solvent may serve as the rate-determining step within the CO 2 evolution pathway, which involves several intermediate steps.
Throughout the course of CO 2 evolution pathway, a variety of side-products are invariably generated and incorporated in the electrolyte, involving a series of intermediate steps. [13,16]The identification of these side-products in the liquid phase is critical for delineating the prevailing reaction pathway for CO 2 evolution and for mitigating the confounding effects of concurrent side reactions on the measurement of gas evolution rates.Hence, we employed 1 H nuclear magnetic resonance (NMR) spectroscopy to detect and analyze these side-products in the liquid phase, thereby isolating and characterizing reaction rates of O 2 and CO 2 evolution without an interference from ancillary reactions.
Notable side-products produced in the presence of Li 6 CoO 4 were glycolic acid (3.98 ppm) and formic acid (8.16 ppm) as shown in Figure S7 (Supporting Information).Glycolic acid is presumably formed through the decomposition of ethylene carbonate (EC), where 1 O 2 deprotonates the carbon atom in the EC ring to produce water and an aldehyde-containing intermediate.This intermediate is then hydrolyzed, yielding glycolic acid and CO 2 . [13,16]Similarly, the formation of formic acid is initiated by 1 O 2 deprotonating the methyl group on dimethyl carbonate (DMC), leading to the formation of an intermediate that, upon hydrolysis, produces formic acid and CO 2 as the end products.
We quantified the amount of glycolic acid and formic acid by measuring their integrated peak area against an internal standard in 1 H NMR spectra, which correlates with the number protons in the sample volume. [37]The amounts of glycolic acid and formic acid rose linearly with increasing density of Li 6 CoO 4 (Figure 4a), establishing the direct relationships with 1 O 2 concentration.These findings substantiate that the observed liquid products and CO 2 are products of the same reaction pathway, which follows a first-order reaction pattern with respect to 1 O 2 .This further suggests that the initial interaction between 1 O 2 and electrolyte solvent molecules likely constitute the rate-determining step in the reaction pathway of CO 2 evolution.

Relative Amount of O 2 and CO 2 Depending on the Local 1 O 2 Concentration, Controlled by the Particle-Particle Distance and Distribution within an Electrode
The distribution of products from competing gas evolution reactions is governed by the proportion of 1 O 2 consumed in each reaction, which reflects their reaction rates.Notably, O 2 evolution follows a second-order reaction, as physical quenching of 1 O 2 into 3 O 2 requires two 1 O 2 molecules in (Equation 1). [18]On the other hand, CO 2 evolution follows a first-order reaction (Equation 2).By examining the ratio of evolved O 2 and CO 2 relative to the Li 6 CoO 4 density, a direct correlation between the rates of O 2 and CO 2 evolution and 1 O 2 concentration [ 1 O 2 ] can be deduced (Equation 3).The symbol "r" is the evolution rate, "k" is the rate constant, and "n" is the evolved amount.
Figure 4b presents the linear correlation between the [O 2 ]/[CO 2 ] ratio and Li 6 CoO 4 density during the initial charge, spanning the entire potential range from 3.8 to 4.3 V vs. Li/Li + , with a 4.2% deviation in their slopes.This persistence of distinct reaction orders for O 2 and CO 2 across various electrochemical biases confirms that the evolution rates of O 2 and CO 2 are governed by chemical kinetics rather than electrochemical processes, as long as 1 O 2 is released and participates in the chemical reactions.Additionally, these findings suggest that a greater fraction of 1 O 2 is utilized in O 2 evolution as opposed to CO 2 evolution, particularly when 1 O 2 concentration is locally high within the matrix of composite electrodes.
Overall, the extents of O 2 and CO 2 evolution may significantly vary depending on the electrode configuration (Scheme 1).While the amount of 1 O 2 released from an individual particle likely remains constant, the local environment pertaining to 1 O 2 is largely dependent on the loading density and inter-particle spacing within electrodes.As the evolution rates of O 2 and CO 2 are proportional to the concentration of 1 O 2 in quadratic and linear dependencies, respectively, the extents of O 2 and CO 2 evolution evolve contrastingly in response to changes in the local concentration of 1 O 2 .This, in turn, results in the distinctive product distribution depending on the electrode configuration.Thus, in the design of electrodes, it becomes imperative to employ more sophisticated approaches that consider not only the intrinsic properties of materials but also gas evolution kinetics based on local environments in conjunction with electrode configurations.

Conclusion
We comprehensively investigated the gas evolution kinetics in Li 6 CoO 4 -a system known to facilely release lattice oxygen-and their implications for electrode configurations.This study found that gas evolution rates significantly shaped by the design of the electrode, especially by the loading density and spatial arrangement of particles within it.By systematically establishing the relationship between the extents of gas evolution and the concentration of 1 O 2 , we discovered that the rates of O 2 and CO 2 evolution are correlated with the concentration of 1 O 2 in second and first order, respectively.It became evident that the extent of O 2 and CO 2 evolution may vary with the inter-particle spacing, even when the number of particles remains constant within the electrode.The different evolution rates in response to the local 1 O 2 concentra-Scheme 1.The contrasting product distributions based on the local concentration of 1 O 2 linked to electrode configuration, particularly in conjunction with the loading density and inter-particle spacing within an electrode.
tion underscore the impact of particle distribution on gas generation.This comprehensive examination of gas evolution kinetics in the context of overlithiated positive electrode configurations offers valuable insights for the strategic design of electrodes.Such designs aim to control the product distribution from competing reactions, optimizing the performance based on our understanding of reaction kinetics.

Experimental Section
Electrode Preparation and Electrochemical Operation: Pristine Li 6 CoO 4 , and LiNi x Co y Mn z A1 1-x-y-z O 2 (x > 0.85) powders were provided by LG Energy Solution.The powder X-ray Diffraction (XRD) pattern of pristine Li 6 CoO 4 was measured via in-house XRD instrument (D8 Advance, Bruker) in the 2 range of 15°-60°with a step size of 0.02°and step time of 3 s.Particles and electrode images were collected by FESEM (Apreo 2 S Hivac, Thermofisher Scientific) with an accelerating voltage of 15 kV.For electrode distribution analysis using FESEM, electrode samples were prepared using the ion-milling system (IM4000, Hitachi) with 6 kV accelerating voltage at 60°for 30 min at the Research Institute of Advanced Materials, Seoul National University.
Pristine Li 6 CoO 4 powder, polyvinylidene fluoride (PVDF), and Super P (TIMCAL) in the ratio of 97.5:1:1.5 wt.% were homogeneously dispersed in N-methyl-pyrrolidone (Acros Organics, 99%) in a planetary mixer (Thinky) at 2000 rpm for 15 min.The positive electrode composite was casted on an aluminum current collector with the thickness of 80 μm using a doctor blade and left to dry for 12 h under dry air condition of 80 °C.The areal loading of composite was 1.2 mg cm −2 after completely drying, and electrodes were pressed to 45 μm by calendaring.To prepare electrodes in varying densities of overlithiated materials, Li 6 CoO 4 was replaced with NCMA in different weight percentage while maintaining the areal loadings and thickness of electrode.All electrodes were transferred to an argonfilled glovebox ([O2] <1 ppm, [H2O] <0.1 ppm) for storage and assembly of any types of electrochemical cells used in experiments.2032-type coin cells were assembled with 12 mm positive electrode, 17 mm separator, 14 mm lithium metal counter electrode, and 1M LiPF 6 EC/DMC (1:1 volume ratio).For all electrochemical tests, a galvanostatic cycling was performed at a rate of 40 mAg −1 to 4.3 V vs. Li/Li + , followed by a constant voltage step until the current reaches 10 mAg −1 .
Operando Gas Analysis: Online differential mass spectrometry (HPR-40, Hiden Analytical) was modified to accommodate gas analysis on Li-ion batteries.OEMS experiments were performed with lithium metal counter electrodes and positive electrodes at varying densities of overlithiated materials, implemented in a custom-built OEMS cell platform.The volume of 1M LiPF 6 EC/DMC (1:1 volume ratio) electrolyte added to the cell was maintained at 1 mL.Electrochemical data were obtained with VSP-200, and signals from mass spectrometer were collected with Hiden analytical software.All OEMS experiments proceeded after 1 hour open-circuit-voltage to stabilize background signals from the OEMS instrument.For quantitative analysis of gas evolution, signals obtained from the mass spectrometry were quantified using the 3D calibration surface contour that was constructed based on signals and partial pressures of standard gases in various concentrations.
Operando X-Ray Diffraction Anaylsis: Operando XRD experiments were performed using pouch cells.Pouch cells were ≈5cm by 8cm, and the components consisting of pouch cells were analogous to coin cell as specified.The pouch cells were assembled with a lithium metal counter electrode, glass fiber separator, and Li 6 CoO 4 positive electrode with 1M LiPF 6 EC/DMC (1:1 volume ratio) electrolyte.Operando XRD patterns were collected by an in-house XRD instrument (D8 Advance, Bruker) in the 2 range of 15°-60°with an average measurement interval of 2 min per images.Data reduction from 2D to 1D was performed with MATLAB, and the sample-to-detector distance was calibrated based on diffraction peaks from the aluminum current collector as the reference.
Post-Mortem Characterization: For X-ray absorption spectroscopy experiments, electrodes were retrieved after disassembling coin cells at 5 different states of charge during the cycle with specified galvanostatic protocol.Retrieved electrodes were thoroughly washed with DMC to remove any electrolyte components that possibly hinder the accurate measurements, and electrodes were vacuum-sealed in aluminum pouches.The X-ray absorption near edge spectroscopy (XANES) at Co K-edge was measured at the 7D XAFS beamline of Pohang Light Source (PLS).The energy shift was calibrated based on the Co metal foil as the reference.After fitting the linear background in the pre-edge and post-edge regions and subtracting from each spectrum, the spectra were normalized using the Athena software.
1 H NMR spectroscopy (Ascend 500, Bruker) was incorporated for the analysis on byproducts formed in the electrolyte during the electrochemical operation under varying concentration of overlithiated materials.After the coin cells with electrodes in varying densities of Li 6 CoO 4 were cycled with a specified galvanostatic cycling protocol, glass fibers separators were retrieved from disassembled coin cells.Each of the four electrolyte samples cycled under varying concentration of overlithiated materials as well as a pristine reference sample were prepared by extracting electrolyte soaked in glass fiber separators.Each 1 H NMR spectra was acquired at 500 MHz frequency with 1024 scans, and dimethyl sulfoxide was used as the internal standard solvent for quantification of byproducts in the liquid phase.

Figure 1 .
Figure 1.a) Galvanostatic profile of Li 6 CoO 4 during charge to 4.3V vs. Li/Li + and changes in half-height position at different state of charge obtained from X-ray absorption near edge spectroscopy of Co K-edge.b) X-ray absorption near edge spectroscopy (XANES) of Co K-edge at different state of charge.c) in situ XRD contour maps of (101) and (201) diffraction peaks that indicates anti-fluorite structure.d) O 2 and CO 2 evolution measured by OEMS quantitative in situ analysis.

Figure 2 .
Figure 2. Change in extents of gas evolution depending on the configuration of electrode.a) The schematic of electrodes prepared in different configurations.b) Voltage profiles and corresponding O 2 and CO 2 evolution curves of the homogeneously distributed simple-blend composite electrode and the bilayer composite electrodes of Li 6 CoO 4 and NCMA.c) The change in amount of O 2 and CO 2 evolution depending on the configuration of electrode.

Figure 3 .
Figure 3. a) The schematic of electrodes in varying densities of Li 6 CoO 4 .Voltage profiles and corresponding O 2 and CO 2 evolution curves of the electrode with b-e) 3, 10, 50, and 100 weight % of Li 6 CoO 4 .f) The amount of O 2 evolution as a function of the density of Li 6 CoO 4 with a fitted polynomial curve that indicates a quadratic relationship.g) Square root of cumulative O 2 evolution as a function of the density of Li 6 CoO 4 with a linearly fitted line.h) Cumulative CO 2 evolution as a function of the density of Li 6 CoO 4 with a linearly fitted line.CO 2 evolution measured in the electrode solely composed of NCMA (open hexagon) was included in linearly fitted line to consider CO 2 evolved under absence of Li 6 CoO 4 .

Figure 4 .
Figure 4. a) Quantified amounts of glycolic acid and formic acid that readily form as side products of reactions between 1 O 2 and electrolyte solvents.The amount of glycolic acid and formic acid is plotted as a function of the Li 6 CoO 4 density.b) the ratio of evolved O 2 and CO 2 with respect to the density of Li 6 CoO 4 at six different points within the potential range from 3.8 to 4.3 V versus Li/Li + during the initial charge.