Vacuum-free spectroscopy reveals metastable oxygen redox intermediates in oxygen-redox cathodes

Vacuum-free spectroscopy reveals metastable oxygen redox intermediates in oxygen-redox cathodes | Research Square window.SnipcartSettings = { analytics: { enabled: false } }; (function() { var accessVector = localStorage.getItem('access_vector') || ''; window.dataLayer = window.dataLayer || []; if (accessVector) { window.dataLayer.push({ user: { profile: { profileInfo: { snid: accessVector } } } }); } })(); (function(w,d,s,l,i){w[l]=w[l]||[];w[l].push({'gtm.start':new Date().getTime(),event:'gtm.js'});var f=d.getElementsByTagName(s)[0],j=d.createElement(s),dl=l!='dataLayer'?'&l='+l:'';j.async=true;j.src='https://www.googletagmanager.com/gtm.js?id='+i+dl;f.parentNode.insertBefore(j,f);})(window,document,'script','dataLayer','GTM-K279D39R'); Browse Preprints In Review Journals COVID-19 Preprints AJE Video Bytes Research Tools Research Promotion AJE Professional Editing AJE Rubriq About Preprint Platform In Review Editorial Policies Our Team Advisory Board Help Center Sign In Submit a Preprint Cite Share Download PDF Article Vacuum-free spectroscopy reveals metastable oxygen redox intermediates in oxygen-redox cathodes William Chueh, Donggun Eum, Oscar Paredes Mellone, Eder Lomeli, and 11 more This is a preprint; it has not been peer reviewed by a journal. https://doi.org/ 10.21203/rs.3.rs-8285919/v1 This work is licensed under a CC BY 4.0 License Status: Under Review Version 1 posted You are reading this latest preprint version Abstract Capturing metastable reaction intermediates through spectroscopy is a fundamental challenge across many fields, as their intrinsic metastability often leads to transformation while being measured. This challenge is acute in oxygen redox in battery intercalation cathodes, where metastable intermediates, despite their long lifetime, remain difficult to capture reliably. Here, we demonstrate that vacuum conditions destabilize oxidized oxygen intermediates and drive their transformation to thermodynamically stable O 2 in Li 1.13 Ni 0.13 Mn 0.57 O 2 . To overcome these issues, we employ vacuum-free and cryogenic hard X-ray Raman spectroscopy, circumventing both vacuum and radiation-induced effects. This approach enables direct and reliable detection of metastable oxygen-redox intermediates in this system, while also capturing diverse oxygen-redox features across different cathodes and stoichiometries. Our findings demonstrate a robust method to identify redox intermediates, reconciling experimental observations and theoretical predictions and sharpening our understanding of oxygen-redox mechanisms. Physical sciences/Materials science/Materials for energy and catalysis/Batteries Physical sciences/Materials science/Techniques and instrumentation/Characterization and analytical techniques Physical sciences/Chemistry/Analytical chemistry Physical sciences/Chemistry/Electrochemistry/Batteries Figures Figure 1 Figure 2 Figure 3 Figure 4 Main text Reliably capturing metastable reaction intermediates has been central to uncovering fundamental reaction mechanisms across various fields such as photochemistry, biology, and radiolysis. 1 – 4 Short-lived and typically highly energetic intermediates are measured via time-resolved measurements. Another important class of intermediates are long-lived and metastable. Because these intermediates are usually close in energy to the thermodynamically stable states, they readily relax, making their direct detection challenging as they are easily perturbed by the measurement itself. Oxygen redox (OR) in battery intercalation cathodes is one such example. Although OR has garnered significant attention as a promising chemistry to enhance the energy density of rechargeable batteries, 5,6 its underlying redox mechanisms are under debate, plausibly arising from difficulty in capturing metastable oxidized oxygen states. Over the past decade, first-principles calculations have predicted numerous oxygen redox mechanisms. 6 – 14 Early studies emphasized transition-metal-driven reductive coupling, in which peroxo-like O–O species are formed alongside reduction of the transition metal. 6 , 7 Later reports challenged this view and predicted that different types of oxygen dimers can emerge during OR, with their bond lengths depending strongly on local structure, oxidation state, and states of charge (SoC). 8 – 11 Beyond dimer formation, alternative pathways have also been suggested, including high-valent Mn ( e.g. , Mn 4+/7+ ) redox and π -back-bonding interactions. 12 – 14 Extensive experimental efforts have been made to validate these theoretical mechanisms, particularly using soft X-ray resonant inelastic X-ray scattering (RIXS). 15 – 25 Through these measurements, a characteristic peak near ~ 531 eV (in excitation) and ~ 523 eV (in emission) is widely regarded as a fingerprint of oxygen redox. Recently, several studies have attributed this spectroscopic feature to trapped molecular oxygen (O 2 ). 17 – 23 This assignment is supported by the RIXS energy-loss features, which exhibit a vibrational structure characteristic of molecular O 2 . This spectroscopic signal has recently been used to support the claim that molecular O 2 is the sole oxidized-oxygen product in oxygen-redox reactions, and this interpretation has been applied across a broad range of cathodes, irrespective of their crystal structure, electronic configuration, or chemical composition, 17–23 even for compounds that are believed not to undergo OR. Yet, these observations stand in contrast to theoretical predictions that a variety of oxidized oxygen species should form during OR. 9–14 In addition, electrochemical profiles typically suggest multiple oxygen-redox peaks upon activation, indicating diverse redox environments. These inconsistencies have been recognized in numerous reports. 26 , 27 In this work, we demonstrate that these inconsistencies arise from the intrinsic metastability of oxidized oxygen intermediates, which are perturbed during the spectroscopic measurements owing to their energetic proximity to the thermodynamically stable state: O 2 (g). We show that vacuum environments can drive their transformation into stable O 2 in Li 1.13 Ni 0.13 Mn 0.57 O 2 . To overcome this challenge, we employed cryogenic hard X-ray Raman spectroscopy (XRS), which enables vacuum-free oxygen 1 s spectroscopy with bulk-sensitivity. 28 , 29 We observe not one but several distinct oxygen species (also across LiNiO 2 , Li 2 RuO 3 , disordered rocksalt), including superoxo, molecular oxygen, and π -interacting oxygen. Our findings reconcile theoretical prediction, electrochemical profiles, and spectroscopic measurements, highlighting the importance of vacuum conditions when gaseous products are the stable products. Our demonstration highlights critical experimental considerations when elucidating the complex nature of oxygen redox that involves metastable intermediates. Robust Bulk Spectral Measurements Hard XRS leverages inelastic X-ray scattering energy transfer to probe core-electron excitations (Supplementary Fig. 1). When operated in the low-momentum transfer regime, where dipole transitions dominate, it enables the acquisition of soft X-ray absorption-like spectral information using hard X-rays. 30 , 31 Owing to its high incident X-ray energy (~ 10.2 keV), this technique does not require vacuum and enables ambient pressure, bulk-sensitive measurements with a penetration depth of a few tens of µm, sufficient to average signals over hundreds to thousands of battery crystallites in a single spectrum (see Method section for experimental details). Here, we employ XRS under cryogenic conditions—preparing samples in liquid nitrogen and measuring them in a liquid-helium cryostat—to further preserve plausible metastable oxygen states by minimizing aging effects and largely suppressing radiation-induced damage. Using vacuum-free and cryogenic XRS, we characterize plausible oxygen redox activity across four cathodes: LiNiO 2 , Li 2 RuO 3 , disordered rocksalt (DRX)-Li 1.17 Ni 0.25 Ti 0.58 O 2 , and Li 1.13 Ni 0.3 Mn 0.57 O 2 . We selected these materials because previous RIXS measurements have consistently detected a single spectroscopic signature (~ 531 eV absorption peak) generally associated with dimerized oxygen, 19–22,32,33 suggesting a common redox mechanism. Figure 1 a–d presents cryogenic XRS spectra at the O K-edge collected at two voltages during the first charge, together with the difference spectra between the two voltages. The voltages were selected based on previous reports of plausible oxygen redox activity as to capture the evolution before and after. Strikingly, among the four materials examined, LiNiO 2 and Li 2 RuO 3 deviate substantially from prior reports. In LiNiO 2 , the absorption feature at ~ 531 eV is absent entirely, and the spectrum exhibits a blue shift (Fig. 1 a), likely due to enhanced Ni–O hybridization upon delithiation. 34 In Li 2 RuO 3 , by contrast, a strong single peak emerges at ~ 530 eV (Fig. 1 b), rather than at ~ 531 eV. Notably, neither of these materials shows the ~ 531 eV peak commonly associated with dimerized oxygen. On the other hand, both DRX and Li 1.13 Ni 0.3 Mn 0.57 O 2 consistently exhibit the ~ 531 eV absorption peak. However, whereas DRX shows only this single peak (Fig. 1 c), Li 1.13 Ni 0.3 Mn 0.57 O 2 exhibits multiple additional spectral features (Fig. 1 d and Supplementary Fig. 2), suggesting the presence of multiple oxidized oxygen species. We return later to assign these spectral features. The vastly diverse spectroscopic observations of oxygen across these four cathodes are consistent with their different electrochemical profiles (Fig. 1 e). In DRX, activation gives rise to only one new redox peak during discharge in the differential voltage curve (yellow curve in the third panel of Fig. 1 e), consistent with the presence of a single spectral feature. In contrast, Li 1.13 Ni 0.3 Mn 0.57 O 2 exhibits several distinct redox peaks after activation on discharge (yellow curve in the fourth panel of Fig. 1 e), consistent with the multiple features observed in vacuum-free and cryogenic XRS (Fig. 1 d). For LiNiO 2 and Li 2 RuO 3 , however, no new peaks emerge in the differential voltage curves; instead, only the intensities of the original peaks change after activation (first and second panels of Fig. 1 e). This one-to-one correspondence between electrochemical and spectroscopic signatures underscores the reliability of cryogenic XRS in capturing intrinsic oxygen-redox states or the lack thereof. Vacuum and Irradiation Effects on Spectral Measurements These distinct spectral features across different materials stand in clear contrast to the recent interpretation that molecular O 2 represents the sole oxidized-oxygen product in oxygen-redox reactions (Supplementary Note 1). We hypothesize that this inconsistency reflects the inherent metastability of oxidized lattice oxygen, which transforms into thermodynamically stable gaseous oxygen during measurement. To investigate this possibility, we employed the aging protocol shown in Fig. 2 a on Li 1.13 Ni 0.3 Mn 0.57 O 2 . Electrodes were charged to 4.8 V, where OR is fully activated, aged under vacuum for up to 100 hours, and subsequently discharged. Vacuum aging reflects the ultra-high-vacuum conditions required for soft X-ray oxygen spectroscopy (typically < 10 − 3 mTorr for several hours; Fig. 2 b). Notably, vacuum provides the chemical driving force for oxygen release, approximately − 0.1 eV/O 2 on the chemical potential scale. Additionally, cells were also aged at 30°C and 60°C as a baseline and to mimic plausible thermal effects during measurement, respectively. Across all storage conditions, the differential voltage curves during the first discharge exhibited a consistent trend (Fig. 2 c): high-voltage redox peaks diminished over time (red arrows), while a low-voltage peak at ~ 3.16 V progressively intensified (yellow arrow). Previous studies have linked the emergence of this low-voltage peak to Mn(III/IV) redox activation, triggered by oxygen gas release. 23 , 35 , 36 Notably, this transformation occurred more rapidly under vacuum and at 60°C than the baseline at 30°C. This trend was consistently observed at different states of charge once OR was activated above 4.4 V (Supplementary Fig. 4). These results demonstrate that these activated electrodes (as measured electrochemically) are highly sensitive to environmental conditions, which can drive the transformation of metastable oxygen species even before X-ray measurements. This effect is experimentally confirmed by comparing cryogenic XRS spectra of two electrodes (Fig. 2 d): one harvested immediately at 4.8 V and another exposed to vacuum for 24 hours prior to measurement—the time required in our setup to reach the vacuum level needed of < 10 − 3 mTorr. All electrodes were stored in liquid nitrogen until measurement to prevent aging prior to XRS measurement. After vacuum exposure, most of the newly resolved features shown in Fig. 1 d disappeared, and the ~ 531 eV peak weakened, highlighting the critical role of vacuum in perturbing the oxidized oxygen intermediates in Li 1.13 Ni 0.3 Mn 0.57 O 2 (see Supplementary Note 2). In particular, the reduction in the ~ 531 eV feature—reflecting an irreversible loss of detectable oxygen states—shows that even structurally confined molecular O 2 can be partially released, likely through diffusion of mobile O 2 . 37–39 We also examined the impact of beam irradiation on the spectra by varying the X-ray dose. While cryogenic XRS measurements exhibited minimal beam-induced effects, the beam effect is significant at room temperature (Supplementary Fig. 7). Furthermore, Supplementary Fig. 8 shows that vacuum exposure and increased X-ray flux produce comparable reductions in peak intensity, with the ~ 529 eV feature being particularly sensitive to both perturbations. These observations emphasize that both cryogenic temperature (at least in XRS) and vacuum-free condition are necessary conditions to detect OR intermediates in Li 1.13 Ni 0.3 Mn 0.57 O 2 . The necessity for XRS measurement at cryogenic conditions to capture intrinsic oxygen-redox states is further supported by comparing how the spectra evolve across different SoCs under cryogenic versus non-cryogenic conditions (Fig. 3 and Supplementary Fig. 9). Figure 3 a shows the voltage profile of Li 1.13 Ni 0.3 Mn 0.57 O 2 during activation. OR activation occurs above 4.4 V, giving rise to a high-voltage plateau that follows the Ni(II/IV) redox. Oxygen spectra were collected at three points along this plateau: prior to OR activation (point #1 in Fig. 3 a), at half activation (#2), and at full activation (#3). Under cryogenic conditions, multiple new oxygen-redox features appear and intensify from point #1 to point #3 along the high-voltage plateau (top panel in Fig. 3 b). In contrast, under non-cryogenic conditions, the ~ 531 eV feature remains overwhelmingly dominant, with no clear development of the additional oxygen-redox features (bottom panel in Fig. 3 b). Notably, in this voltage range, Mn L-edge and C K-edge spectra remained largely unchanged (Supplementary Fig. 10), indicating neither C–O byproducts form nor Mn participates in redox within this voltage range. Again, these results confirm that cryogenic measurement is required for XRS to capture the OR intermediates (see Supplementary Note 3 for further discussion). To gain further insight into these newly emerging features observed in cryogenic XRS, we simulated O K-edge absorption spectra using the Bethe-Salpeter equation for Li 2 MnO 3 , 40 a parent phase that hosts oxygen redox activity in Li 1.13 Ni 0.3 Mn 0.57 O 2 (Supplementary Fig. 12; see Methods for details). We systematically enumerated diverse oxygen configurations (3,237 structures in total) across different lithium contents in Li 2−x MnO 3 (see Supplementary Note 4). Based on bonding type and length, oxygen configurations were classified into five categories: peroxo ( \(\:{\text{O}}_{2}^{2-}\) ), ozonide ( \(\:{\text{O}}_{3}^{-}\) ), superoxo ( \(\:{\text{O}}_{2}^{-}\) ), molecular oxygen ( \(\:{\text{O}}_{2}\) ), and short Mn–O motifs with strong π -interaction ( d Mn–O < 1.7 Å). 9 Figure 3 c and Supplementary Fig. 14 present simulated spectra (dotted lines) for representative structures in each category, along with the summed spectra (scatter plots) within each panel. Each oxygen species exhibits a distinct spectral signature, with excitation energies markedly different from the ~ 531 eV peak. Only molecular oxygen corresponds to the ~ 531 eV feature. We then fit a linear combination of these simulated spectra to the experimental vacuum-free and cryogenic XRS data. For completeness, we also fit them to the non-cryogenic XRS (Fig. 3 d). In the cryogenic spectra, the multiple peaks observed in the differential plot (before and after OR activation) align well with the features of π -complex, superoxo and molecular oxygen. These spectral evolutions are further quantified in Supplementary Note 5. By contrast, the non-cryogenic XRS spectra are mostly dominated by a single molecular oxygen signature, indicating that all metastable intermediates resolved in cryogenic XRS convert into molecular oxygen under non-cryogenic conditions. This behavior is consistent with the vacuum-driven transformations observed in Fig. 2 d. Mechanistic Basis of Spectral Measurement Challenges The challenges of reliably capturing OR intermediates can be rationalized within a reaction framework involving two competing reaction pathways: one that is kinetically accessible and the other that is thermodynamically favored. According to the equilibrium Li–Mn–O phase diagram, 12 upon delithiation, Li 2 MnO 3 thermodynamically decomposes into oxygen gas and spinel-like phases (Supplementary Fig. 16). This decomposition process corresponds to the lowest-energy state (labeled as “Oxygen Loss” in Fig. 4 a). In contrast, the formation of π -complex and O–O dimers represent metastable pathways, which lie at higher energies across all lithium contents (Fig. 4 a). The energy differences among these metastable intermediates are relatively small (0.3–0.6 eV/Li), while the gap to thermodynamically stable state is ~ 1.5 eV/Li. The activation free energy controls whether the activation process proceeds via the kinetic or the thermodynamic pathway (Fig. 4 b). The equilibrium decomposition route involves extensive bond breaking and phase separation, resulting in a high kinetic barrier. In contrast, the formation of the metastable oxidized oxygen species requires only localized lattice distortions, making them kinetically more accessible with lower activation energies. 10 , 37 Crucially, the extent to which such distortions are accommodated strongly depends on the material’s structure, composition, and SoC, thereby governing the types and proportions of OR products captured by cryogenic XRS. Over time, these kinetic products gradually relax to the thermodynamic product through an energetically downhill reaction (Fig. 4 b), though the transformation is sluggish due to a high activation barrier. 11 , 37 – 39 This slow evolution is supported by experimental observations under 30°C aging, low-current cycling (Supplementary Fig. 17), and during extended cycling (Supplementary Fig. 18), each requiring prolonged time to induce noticeable change. By contrast, low oxygen partial pressure or elevated temperature ( e.g. , vacuum or 60°C) increases the driving force and accelerates the conversion (pink curve in Fig. 4 b), as revealed by the shaded region in Fig. 4 a (see Supplementary Note 6 for calculation details). Under such conditions, the energy of the “oxygen loss” pathway shifts downward by 0.4 eV/Li, widening its separation from the metastable intermediates to nearly 2 eV/Li. Such pressure-dependent stabilization provides a mechanistic basis for understanding the significant effect of vacuum on OR in Li 1.13 Ni 0.3 Mn 0.57 O 2 . Summary Our work establishes the importance of measurement conditions for capturing metastable oxygen-redox intermediates in intercalation cathodes. By eliminating vacuum environments and minimizing beam-induced effects through vacuum-free and cryogenic XRS, we resolve a diverse set of OR features across multiple cathodes. In Li 1.13 Ni 0.3 Mn 0.57 O 2 , multiple intermediates including superoxo and π -complex species were observed alongside molecular O 2 , whereas other cathodes such as LiNiO 2 and Li 2 RuO 3 exhibit markedly different features from prior reports, some absent of oxidized oxygen signatures altogether. Furthermore, we demonstrate that oxidized oxygen states are intrinsically metastable and can readily transform into molecular oxygen under vacuum or irradiation. Our work underscores the need for attention to measurement conditions when probing metastable oxygen-redox products and opens new avenues for capturing metastable products in this important class of redox reactions. Importantly, our findings place soft X-ray and non-vacuum-based measurements within a complementary framework, showing that each technique accesses mutually informative aspects of the oxygen-redox landscape. Methods Material synthesis. Li 1.13 Ni 0.3 Mn 0.57 O 2 was synthesized by mixing appropriate amounts of Li 2 CO 3 and (Ni 0.35 Mn 0.65 )CO 3 , followed by calcination at 900 °C for 20 hours in air. LiNiO 2 was prepared by thoroughly mixing LiOH·H 2 O and Ni(OH) 2 in a stoichiometric ratio and calcining the mixture at 650 °C for 10 hours in O 2 . Li 2 RuO 3 and DRX-Li 1.17 Ni 0.25 Ti 0.58 O 2 were synthesized following previously reported procedures. 6,32 Aging t est ing . For 30 °C aging, cells were first charged to the target voltage, held at open circuit at 30 °C for the desired duration, and then discharged. For 60 °C aging, cells were charged at 30 °C and aged inside a temperature-controlled chamber at 60 °C for the specified time. After aging, they were transferred to a 30 °C chamber and equilibrated for 1 hour before discharge. For vacuum aging, charged cells were disassembled in an Ar-filled glovebox, and the harvested electrodes were exposed to vacuum using a HiCube pumping system for the designated duration. The electrodes were then reassembled into fresh cells, rested at 30 °C for 1 hour, and discharged. X-ray Raman spectroscopy (XRS) . XRS were performed at Beamline 15-2 of the Stanford Synchrotron Radiation Lightsource. The radiation from the undulator was monochromated using a liquid nitrogen-cooled Si (311) double-crystal monochromator. X-rays were then focused onto the sample position to a beam size of 50 μm × 150 μm (H × V) using two Kirkpatrick-Baez (KB) mirrors. The XRS endstation at BL 15-2 consists of a multi-crystal Johann-type spectrometer with 40 Si (110) spherically bent diced analyzer crystals with a 1 m bending radius, operating in Rowland geometry. 30 The energy-loss measurements were performed in the so-called inverse geometry, i.e. , keeping the analyzed energy ω 2 ≈ 9.69 keV fixed while varying the incident X-ray energy ω 1 , such that the transferred energy (ω = ω 1 – ω 2 ) was scanned around the absorption edge. To optimize counting statistics and minimize the areal power density of the incident beam, all measurements were performed in reflection geometry, along the horizontal scattering plane, using a grazing angle of 2° to 4°. To minimize the effects of radiation energy deposition, all energy-loss scans were performed under cryogenic conditions (T ≈ 15 K). The XRS spectra presented in this work were obtained following a similar procedure as described in ref. 29, which allows for systematic control of radiation deposition on the sample during the experiment. For each sample, several energy-loss scans were acquired over numerous spots under the same experimental conditions. The acquisition time for each scan was 2.5s or 5 s per energy point with a total amount of 148 energy points per scan were measured for all experiments. For accurate energy calibration, periodic measurements of the elastic line were acquired during the experiments. Finally, by measuring over many fresh spots across each sample, the collected set of equivalent spectra was used to construct an averaged spectrum with improved statistical quality. The fitting procedure for background subtraction and normalization of the energy-loss scans followed a similar approach as discussed in refs. 28,29. The process involves simultaneous fitting of two components: (1) a second-order polynomial function accounting for the high-energy tail of the valence electron contribution, and (2) a model for the core-electron atomic background contribution, , to the total spectrum. For modeling the core-electron atomic background, Hartree-Fock tabulated atomic Compton profiles were converted to using the formulation of Ribberfors and Holm. 41,42 Additionally, asymmetry correction to the core-electron Compton profile was included. 41 The polynomial function was fitted to the total energy-loss spectrum for energy transfers ω 565 eV), the sum of the polynomial (valence-electron contribution) and the modeled atomic background (core-electron contribution) was fitted. All extracted XRS spectra were then normalized to have the same integrated intensity for energy transfers above 565 eV. Scanning t ransmission X-ray m icroscopy (STXM). Dispersed powder samples were prepared by sonicating in isopropyl alcohol (>99.5%) for 30 minutes. The resulting suspension was drop-cast onto copper TEM grids coated with a carbon film. STXM measurements were conducted at Beamline 7.0.1.2 of the Advanced Light Source, Lawrence Berkeley National Laboratory. A 53-energy STXM stack was collected for each sample with a pixel size of 50 nm. Total e lectron y ield (TEY); Total f luorescence y ield (TFY); Resonant i nelastic X-ray s cattering (RIXS). Soft X-ray TEY, TFY, and RIXS spectra were collected at the ultra-high-efficiency iRIXS endstation on beamline 8.0.1 of the Advanced Light Source (ALS), Lawrence Berkeley National Laboratory. All electrodes were harvested in an Ar-filled glovebox, double-sealed in aluminum pouches, and transferred into a custom vacuum-compatible transport kit to avoid air exposure. O K-edge RIXS spectra were acquired simultaneously with TEY and TFY modes, with two-dimensional RIXS maps constructed by recording emission spectra at each excitation energy. Emission energies were aligned to the elastic scattering line. To minimize beam-induced damage, samples were dithered laterally over a 0.6 mm range during measurement. Excitation energies were calibrated using a TiO 2 reference for the O K-edge. X-ray photoelectron spectroscopy (XPS) . Laboratory-based XPS measurements were performed using a PHI VersaProbe 4 instrument equipped with a monochromated Al Kα X-ray source (1486.6 eV, 47.9 W). Electrodes were mounted using double-sided tape, and charge compensation was applied using a dual ion and electron flood gun (1 V, 3 μA). All samples were loaded in an Ar-filled glovebox and transferred to the instrument with less than 1 minute of air exposure. The X-ray beam diameter, power, and accelerating voltage were 200 μm, 50 W, and 15 kV, respectively. Survey scans were collected with a pass energy of 224 eV and step size of 0.8 eV, while core-level scans used a pass energy of 55 eV and step size of 0.1 eV. Density f unctional t heory (DFT) c alculation. Density functional theory (DFT) calculations were performed using the Vienna Ab Initio Simulation Package (VASP, version 6.4.2), based on previous work by Chen et al . 9 Initial structural relaxations were carried out using the Perdew-Burke-Ernzerhof (PBE) implementation of the generalized gradient approximation (GGA), with a Hubbard U + J correction applied to Mn ( U = 5.0 eV, J = 0.5 eV), following the Liechtenstein formalism. 43,44 After identifying the lowest-energy configurations for each peroxide-type species, further relaxations were performed using the r 2 SCAN exchange-correlation functional to better account for the variable oxidation states of Mn in the system. 45 All calculations employed a plane-wave energy cutoff of 600 eV, a self-consistent field (SCF) convergence threshold of 1 × 10⁻ 5 eV, and a force convergence criterion of 0.02 eV/Å. A monoclinic unit cell containing four formula units of Li 2 MnO 3 was used, with a k -point mesh of 6 × 3 × 6. Spectroscopic OCEAN c alculation. Oxygen K-edge X-ray absorption spectra were computed using the Bethe-Salpeter Equation (BSE)-based OCEAN code (version 3.0.4), with DFT wavefunctions generated by Quantum ESPRESSO (version 7.0). Pseudopotentials were constructed using the scalar-relativistic Optimized Norm-Conserving Vanderbilt Pseudopotential (ONCVPSP) method. 46 The r2SCAN-relaxed structures obtained from the VASP calculations were used as inputs for all spectral simulations. Spectroscopic calculations were performed with a plane-wave cutoff of 140 Ry and an SCF convergence threshold of 1 × 10⁻ 10 Ry. A k-point mesh of 5 × 3 × 5 and 110 unoccupied bands were included in the BSE calculation. A uniform energy shift was applied to align all spectra at a given Li content to the post-edge continuum maximum, positioned at approximately 542.5 eV. Declarations *** Certain equipment, instruments, software, or materials are identified in this paper to adequately specify the experimental procedure. Such identification does not imply recommendation or endorsement of any product or service by NIST, nor does it imply that the materials or equipment identified are necessarily the best available for the purpose. *** Data availability All data are available in the main text or the supplementary materials. Additional data in this paper are available in the Stanford Digital Repository/Dryad. Acknowledgements The characterization aspect work was supported by the Assistant Secretary for Energy Efficiency and Renewable Energy, Office of Vehicle Technologies, Battery Materials Research Program, U.S. Department of Energy (DOE). E.Z.C.was supported in part by an ALS Doctoral Fellowship in Residence. Theoretical work is supported in part by the U.S. Department of Energy, Office of Basic Energy Sciences, Materials Sciences and Engineering Division. Calculations were performed on the Sherlock cluster at Stanford University and on resources of the National Energy Research Scientific Computing Center (NERSC), a Department of Energy Office of Science User Facility, using NERSC award BES-ERCAP0027203. We thank Dr. Sung-O Park for valuable discussions regarding the calculation results. This research used resources of the Stanford Synchrotron Radiation Lightsource (SSRL), SLAC National Accelerator Laboratory, which is supported by the DOE Office of Science, Office of Basic Energy Sciences under Contract No. DE-AC02-76SF00515. X-ray Raman analysis used resources of the National Energy Research Scientific Computing Center, a DOE Office of Science User Facility supported by the Office of Science of the U.S. Department of Energy under Contract No. DE-AC02-05CH11231 using NERSC award BES-ERCAP0031542. Use of the Advanced Light Source (ALS) at Lawrence Berkeley National Laboratory is supported by the DOE Office of Science User Facility under Contract No. DE-AC02-05CH11231. This work also utilized resources of the Stanford Nano Shared Facilities (SNSF), supported by the National Science Foundation under award ECCS-2026822. Author contributions D.E. and O.P.M. conceived the original idea. D.E., O.P.M., D. Sokaras, and W.C.C. designed the research. D.E. and O.P.M. carried out the XRS measurements with assistance from D. Skoien and D. Sokaras. O.P.M. performed the XRS data analysis. E.G.L. performed the DFT calculations and spectral simulations with constructive advice from J.V. and T.P.D. E.M. and H.R. collected the RIXS, TFY, and TEY datasets, and E.P.K.L.C. performed the XPS measurements. E.P.K.L.C. and N.K. conducted the STXM measurements, and E.Z.C. processed the data. S.L. and J.A.D. offered valuable comments for this project. D.E. and W.C.C. wrote the manuscript, and all authors contributed to its revision. D.E., D. Sokaras, and W.C.C. supervised all aspects of the research. Competing interests The authors declare no competing interests. References Zewail, A. H. Femtochemistry: Atomic-Scale Dynamics of the Chemical Bond. J. Phys. Chem. A 104 , 5660–5694 (2000). Chen, L. X. Probing transient molecular structures in photochemical processes using laser-initiated time-resolved X-ray absorption spectroscopy. Annu. Rev. Phys. Chem. 56 , 221–254 (2005). Williams, P. J. H. et al. New Approach to the Detection of Short-Lived Radical Intermediates. J. Am. Chem. Soc. 144 , 15969–15976 (2022). Lin, M. F. et al. Imaging the short-lived hydroxyl-hydronium pair in ionized liquid water. Science 374 , 92–95 (2021). Rahman, M. M. & Lin, F. Oxygen redox chemistry in rechargeable Li-ion and Na-ion batteries. Matter 4 , 490–527 (2021). Sathiya, M. et al. Reversible anionic redox chemistry in high-capacity layered-oxide electrodes. Nat. Mater. 12 , 827–835 (2013). Saubanère, M., McCalla, E., Tarascon, J.M. & Doublet, M. L. The intriguing question of anionic redox in high-energy density cathodes for Li-ion batteries. Energy Environ. Sci. 9 , 984–991 (2016). Chen, H. & Islam, M. S. Lithium Extraction Mechanism in Li-Rich Li 2 MnO 3 Involving Oxygen Hole Formation and Dimerization. Chem. Mater. 28 , 6656–6663 (2016). Chen, Z., Li, J. & Zeng, X. C. Unraveling oxygen evolution in Li-rich oxides: a unified modeling of the intermediate peroxo/superoxo-like dimers. J. Am. Chem. Soc. 141 , 10751–10759 (2019). Vinckeviciute, J., Kitchaev, D. A. & Van der Ven, A. A two-step oxidation mechanism controlled by Mn migration explains the first-cycle activation behavior of Li 2 MnO 3 -based Li-excess materials. Chem. Mater. 33 , 1625–1636 (2021). Kim, B. et al. A theoretical framework for oxygen redox chemistry for sustainable batteries. Nat. Sustainability , 5 , 708–716 (2022). Radin, M. D., Vinckeviciute, J., Seshadri, R. & Van der Ven, A. Manganese oxidation as the origin of the anomalous capacity of Mn-containing Li-excess cathode materials. Nat. Energy 4 , 639–646 (2019). Sudayama, T. et al. Multiorbital bond formation for stable oxygen-redox reaction in battery electrodes. Energy Environ. Sci. 13 , 1492–1500 (2020). Eum, D. et al. Coupling structural evolution and oxygen-redox electrochemistry in layered transition metal oxides. Nat. Mater. 21 , 664–672 (2022). Gent, W. E. et al. Coupling between oxygen redox and cation migration explains unusual electrochemistry in lithium-rich layered oxides. Nat. Commun. 8 , 2091 (2017). Hong, J. et al. Metal–oxygen decoordination stabilizes anion redox in Li-rich oxides. Nat. Mater. 18 , 256–265 (2019). McColl, K. et al. Transition metal migration and O 2 formation underpin voltage hysteresis in oxygen-redox disordered rocksalt cathodes. Nat. Commun. 13 , 5275 (2022). House, R. A. et al. Superstructure control of first-cycle voltage hysteresis in oxygen-redox cathodes. Nature 577 , 502–508 (2020). House, R. A. et al. The role of O 2 in O-redox cathodes for Li-ion batteries. Nat. Energy 6 , 781–789 (2021). House, R. A. et al. Covalency does not suppress O 2 formation in 4d and 5d Li-rich O-redox cathodes. Nat. Commun. 12 , 2975 (2021). Juelsholt, M. et al. Does trapped O 2 form in the bulk of LiNiO 2 during charging?. Energy Environ. Sci. 17 , 2530–2540 (2024). Ogley, M. J. W. et al. Metal-ligand redox in layered oxide cathodes for Li-ion batteries. Joule 9 , 101775 (2025). Marie, J. J. et al. Trapped O 2 and the origin of voltage fade in layered Li-rich cathodes. Nat. Mater. 23 , 818–825 (2024). Eum, D. et al. Voltage decay and redox asymmetry mitigation by reversible cation migration in lithium-rich layered oxide electrodes. Nat. Mater. 19 , 419–427 (2020). Jang, H. Y. et al. Structurally robust lithium-rich layered oxides for high-energy and long-lasting cathodes. Nat. Commun. 15 , 1288 (2024). Gao, X. et al. Clarifying the origin of molecular O 2 in cathode oxides. Nat. Mater. 24 , 743–752 (2025). Lebens-Higgins, Z. W. et al. Distinction between intrinsic and X-ray-induced oxidized oxygen states in Li-rich 3d layered oxides and LiAlO 2 . J. Phys. Chem. C 123 , 13201–13207 (2019). Paredes-Mellone, O. A. et al. Investigating the electronic structure of high explosives with X-ray Raman spectroscopy. Sci. Rep. 12 , 19460 (2022). Paredes-Mellone, O. A. et al. Deciphering decomposition pathways of high explosives with cryogenic X-ray Raman spectroscopy. PNAS. 122 , e2426320122 (2025). Sokaras, D. et al. A high resolution and large solid angle x-ray Raman spectroscopy end-station at the Stanford Synchrotron Radiation Lightsource. Rev. Sci. Instrum. 83 , 043112 (2012). Schülke, W. Electron Dynamics by Inelastic X-ray Scattering (Oxford, 2007). Li, B. et al. Capturing dynamic ligand-to-metal charge transfer with a long-lived cationic intermediate for anionic redox. Nat. Mater. 21 , 1165–1174 (2022). Li, N. et al. Unraveling the cationic and anionic redox reactions in a conventional layered oxide cathode. ACS Energy Lett. 4 , 2836–2842 (2019). Cao, R., Thomas, K. E., Ghosh, A. & Sarangi, R. X-ray absorption spectroscopy of archetypal chromium porphyrin and corrole derivatives. RSC Adv. 10, 20572–20578 (2020). Hu, E. et al. Evolution of redox couples in Li-and Mn-rich cathode materials and mitigation of voltage fade by reducing oxygen release. Nat. Energy 3 , 690–698 (2018). Yan, P. et al. Injection of oxygen vacancies in the bulk lattice of layered cathodes. Nat. Nanotechnol. 14 , 602–608 (2019). McColl, K., Coles, S. W., Zarabadi-Poor, P., Morgan, B. J. & Islam, M. S. Phase segregation and nanoconfined fluid O 2 in a lithium-rich oxide cathode. Nat. Mater. 23 , 826–833 (2024). Eum, D. et al. Electrochemomechanical failure in layered oxide cathodes caused by rotational stacking faults. Nat. Mater. 23 , 1093–1099 (2024). Csernica, P. M. et al. Substantial oxygen loss and chemical expansion in lithium-rich layered oxides at moderate delithiation. Nat. Mater. 24 , 92–100 (2025). Vinson, J. Advances in the OCEAN-3 spectroscopy package. Phys. Chem. Chem. Phys. 24 , 12787–12803 (2022). Holm, P. & Ribberfors, R. First correction to the nonrelativistic Compton cross section in the impulse approximation. Phys. Rev. A 40 , 6251 (1989). Sternemann, H. et al. An extraction algorithm for core-level excitations in non-resonant inelastic X-ray scattering spectra. J. Synchrotron Rad. 15 , 162–169 (2008). Perdew, J. P. et al. Generalized Gradient Approximation Made Simple. Phys. Rev. Lett. 77 , 3865 (1996). Liechtenstein, A. I., Anisimov, V. I. & Zaanen, J. Density-functional theory and strong interactions: Orbital ordering in Mott-Hubbard insulators. Phys. Rev. B 52 , R5467(R) (1995). Furness, J. W., Kaplan, A. D., Ning, J., Perdew, J. P. & Sun, J. Accurate and Numerically Efficient r2SCAN Meta-Generalized Gradient Approximation. J. Phys. Chem. Lett. 11 , 8208–8215 (2020). Hamann, D. R. Optimized norm-conserving Vanderbilt pseudopotentials. Phys. Rev. B 88 , 085117 (2017). Additional Declarations There is NO Competing Interest. Supplementary Files SupplementaryInformation.docx Supplementary Information Cite Share Download PDF Status: Under Review Version 1 posted You are reading this latest preprint version Research Square lets you share your work early, gain feedback from the community, and start making changes to your manuscript prior to peer review in a journal. As a division of Research Square Company, we’re committed to making research communication faster, fairer, and more useful. We do this by developing innovative software and high quality services for the global research community. Our growing team is made up of researchers and industry professionals working together to solve the most critical problems facing scientific publishing. Also discoverable on Platform About Our Team In Review Editorial Policies Advisory Board Help Center Resources Author Services Accessibility API Access RSS feed Manage Cookie Preferences © Research Square 2026 | ISSN 2693-5015 (online) Privacy Policy Terms of Service Do Not Sell My Personal Information {"props":{"pageProps":{"initialData":{"identity":"rs-8285919","acceptedTermsAndConditions":true,"allowDirectSubmit":false,"archivedVersions":[],"articleType":"Article","associatedPublications":[],"authors":[{"id":562136599,"identity":"d4f3299e-ebca-4b69-ba28-3334dbf35868","order_by":0,"name":"William Chueh","email":"data:image/png;base64,iVBORw0KGgoAAAANSUhEUgAAAZAAAAAyAQMAAABI0h/eAAAABlBMVEX///8AAABVwtN+AAAACXBIWXMAAA7EAAAOxAGVKw4bAAAA0UlEQVRIiWNgGAWjYFACHgYGxgYbMPMDAwMzkEogSksamJpBipbDJGgxOH72mMTHHeft7dnPHmz4uMeagZ89xwC/ljN5aZIzz9xO7OHJS2yc8SydQbLnDX4tZjd4zKR5224n8DDkmD/mOXCYweAGAVvAWv62nbPn4X9j2AzSYk+UFsa2A4w9EjkQLQYSBLTYn8kxtuxtS07sufHGsHHGgXQeiTPPCvBqkWw/Y3jjZ5udPXt/jmHDhwPWcvztyRvwasEAPKQpHwWjYBSMglGAFQAAqwNHUvjf9jYAAAAASUVORK5CYII=","orcid":"https://orcid.org/0000-0002-7066-3470","institution":"Stanford University","correspondingAuthor":true,"prefix":"","firstName":"William","middleName":"","lastName":"Chueh","suffix":""},{"id":562136600,"identity":"3b21dd46-7a48-484c-8326-d4c6157cba98","order_by":1,"name":"Donggun Eum","email":"","orcid":"https://orcid.org/0000-0002-3120-8790","institution":"Stanford University","correspondingAuthor":false,"prefix":"","firstName":"Donggun","middleName":"","lastName":"Eum","suffix":""},{"id":562136601,"identity":"3c4223dd-4a69-496f-8fd2-51188727a317","order_by":2,"name":"Oscar Paredes Mellone","email":"","orcid":"","institution":"SLAC National Accelerator Laboratory","correspondingAuthor":false,"prefix":"","firstName":"Oscar","middleName":"Paredes","lastName":"Mellone","suffix":""},{"id":562136602,"identity":"1f9f6166-47f2-4678-9846-8381e5bbb1c0","order_by":3,"name":"Eder Lomeli","email":"","orcid":"https://orcid.org/0000-0002-8141-2140","institution":"Stanford University","correspondingAuthor":false,"prefix":"","firstName":"Eder","middleName":"","lastName":"Lomeli","suffix":""},{"id":562136603,"identity":"c614e31e-9f44-4c28-837f-9bab03d1c602","order_by":4,"name":"Edward Mu","email":"","orcid":"","institution":"Stanford University","correspondingAuthor":false,"prefix":"","firstName":"Edward","middleName":"","lastName":"Mu","suffix":""},{"id":562136604,"identity":"bbdf862f-4929-4e6c-8c98-a0ea56d74169","order_by":5,"name":"Hari Ramachandran","email":"","orcid":"https://orcid.org/0000-0002-4469-7722","institution":"Stanford University","correspondingAuthor":false,"prefix":"","firstName":"Hari","middleName":"","lastName":"Ramachandran","suffix":""},{"id":562136605,"identity":"71420769-1e92-4495-8cc4-e62fea8b48c9","order_by":6,"name":"Dean Skoien","email":"","orcid":"","institution":"SLAC National Accelerator Laboratory","correspondingAuthor":false,"prefix":"","firstName":"Dean","middleName":"","lastName":"Skoien","suffix":""},{"id":562136606,"identity":"c5e4c91f-1183-430b-851b-6d180998e0e9","order_by":7,"name":"Emma Choy","email":"","orcid":"","institution":"Stanford University","correspondingAuthor":false,"prefix":"","firstName":"Emma","middleName":"","lastName":"Choy","suffix":""},{"id":562136607,"identity":"b2167e8c-ebb3-457e-b5be-737aa51827b1","order_by":8,"name":"Nidhi Kapate","email":"","orcid":"","institution":"Stanford University","correspondingAuthor":false,"prefix":"","firstName":"Nidhi","middleName":"","lastName":"Kapate","suffix":""},{"id":562136608,"identity":"15d2c25f-5d2b-43a5-9293-8bfa6faeb4d7","order_by":9,"name":"Serin Lee","email":"","orcid":"","institution":"Stanford University","correspondingAuthor":false,"prefix":"","firstName":"Serin","middleName":"","lastName":"Lee","suffix":""},{"id":562136609,"identity":"66cec3d5-fc2b-44c9-971c-e13808f8a780","order_by":10,"name":"Evan Carlson","email":"","orcid":"https://orcid.org/0000-0002-5577-6549","institution":"Stanford University","correspondingAuthor":false,"prefix":"","firstName":"Evan","middleName":"","lastName":"Carlson","suffix":""},{"id":562136610,"identity":"0ce82076-192d-4512-9f76-65b3995c61bb","order_by":11,"name":"John Vinson","email":"","orcid":"https://orcid.org/0000-0002-7619-7060","institution":"National Institute of Standards and Technology","correspondingAuthor":false,"prefix":"","firstName":"John","middleName":"","lastName":"Vinson","suffix":""},{"id":562136611,"identity":"b0fac2ec-ebad-448b-a764-b9ded443370a","order_by":12,"name":"Jennifer Dionne","email":"","orcid":"https://orcid.org/0000-0001-5287-4357","institution":"Stanford University","correspondingAuthor":false,"prefix":"","firstName":"Jennifer","middleName":"","lastName":"Dionne","suffix":""},{"id":562136612,"identity":"82d530e2-2c05-4618-bed6-73f8649f2163","order_by":13,"name":"Thomas Devereaux","email":"","orcid":"https://orcid.org/0000-0001-8072-9237","institution":"SLAC National Accelerator Laboratory","correspondingAuthor":false,"prefix":"","firstName":"Thomas","middleName":"","lastName":"Devereaux","suffix":""},{"id":562136613,"identity":"f7a8c395-826a-4a8f-9928-ab752ef53e9f","order_by":14,"name":"Dimosthenis Sokaras","email":"","orcid":"https://orcid.org/0000-0001-8117-1933","institution":"SLAC National Accelerator Laboratory","correspondingAuthor":false,"prefix":"","firstName":"Dimosthenis","middleName":"","lastName":"Sokaras","suffix":""}],"badges":[],"createdAt":"2025-12-05 09:16:27","currentVersionCode":1,"declarations":"","doi":"10.21203/rs.3.rs-8285919/v1","doiUrl":"https://doi.org/10.21203/rs.3.rs-8285919/v1","draftVersion":[],"editorialEvents":[],"editorialNote":"","failedWorkflow":false,"files":[{"id":102297658,"identity":"f7c67709-1b79-45af-bcfc-e8bacb9da778","added_by":"auto","created_at":"2026-02-10 10:28:41","extension":"png","order_by":1,"title":"Figure 1","display":"","copyAsset":false,"role":"figure","size":820478,"visible":true,"origin":"","legend":"\u003cp\u003eDistinct oxygen-redox signatures revealed by vacuum-free and cryogenic XRS.\u003c/p\u003e\n\u003cp\u003ea–d, Cryogenic XRS O K-edge spectra of various oxygen-redox cathodes at two charge cutoff voltages (top), with corresponding differential spectra (bottom). e, Initial-cycle differential voltage curves for all four cathodes charged to the specified cutoff voltages.\u003c/p\u003e","description":"","filename":"1.png","url":"https://assets-eu.researchsquare.com/files/rs-8285919/v1/21de4c8fb3f7ccaad7b6fa71.png"},{"id":102275703,"identity":"be57d77e-63dc-4a94-a5d5-26cebc1a7a25","added_by":"auto","created_at":"2026-02-10 05:46:18","extension":"png","order_by":2,"title":"Figure 2","display":"","copyAsset":false,"role":"figure","size":511480,"visible":true,"origin":"","legend":"\u003cp\u003eDestabilization of oxygen redox states under thermal and vacuum conditions.\u003c/p\u003e\n\u003cp\u003ea, Cycling protocol used to probe the oxygen-redox instability of Li\u003csub\u003e1.13\u003c/sub\u003eNi\u003csub\u003e0.3\u003c/sub\u003eMn\u003csub\u003e0.57\u003c/sub\u003eO\u003csub\u003e2\u003c/sub\u003e under three different aging conditions (vacuum, 30 °C, and 60 °C). Cells were charged to 4.8 V, rested under each condition, and then discharged. b, Schematic of conventional oxygen spectroscopy requiring ultra-high-vacuum conditions. The right graph shows the evolution of vacuum level as a function of evacuation time. The yellow shaded region indicates the typical vacuum threshold required for conventional oxygen spectroscopy. c, Corresponding discharge differential voltage curves as a function of aging time. d, Cryogenic O K-edge spectra of electrodes charged to 4.8 V, comparing one measured immediately after charging with another exposed to vacuum for 24 h before measurement.\u003c/p\u003e","description":"","filename":"2.png","url":"https://assets-eu.researchsquare.com/files/rs-8285919/v1/9b2b38efa27c3540c34269fc.png"},{"id":102297784,"identity":"d67ae161-0ecc-4d6f-899f-43bb5bd9aedb","added_by":"auto","created_at":"2026-02-10 10:29:10","extension":"png","order_by":3,"title":"Figure 3","display":"","copyAsset":false,"role":"figure","size":839271,"visible":true,"origin":"","legend":"\u003cp\u003ea, Initial charge curve showing the voltage points at which samples were harvested for spectroscopy measurements. b, Cryogenic and non-cryogenic XRS O K-edge spectra at the three sampling points in a, along with differential spectra between #1 and #3. c, Simulated oxygen K-edge X-ray absorption spectra for representative oxygen configurations, systematically enumerated across varying lithium contents in Li\u003csub\u003e2\u003c/sub\u003eMnO\u003csub\u003e3\u003c/sub\u003e. d, Fitting results of the differential spectra from cryogenic and non-cryogenic measurements using simulated spectra.\u003c/p\u003e\n\u003cp\u003eReliable detection and identification of oxygen-redox intermediates.\u003c/p\u003e","description":"","filename":"3.png","url":"https://assets-eu.researchsquare.com/files/rs-8285919/v1/40cfc5dc4a065f79b138a2f2.png"},{"id":102297844,"identity":"b3bbd8c6-6847-415f-a3ad-9e21366cbddc","added_by":"auto","created_at":"2026-02-10 10:29:21","extension":"png","order_by":4,"title":"Figure 4","display":"","copyAsset":false,"role":"figure","size":266785,"visible":true,"origin":"","legend":"\u003cp\u003eKinetic-thermodynamic pathways governing oxygen redox reactions.\u003c/p\u003e\n\u003cp\u003ea, Formation energies of various charging pathways in Li\u003csub\u003e2\u003c/sub\u003eMnO\u003csub\u003e3\u003c/sub\u003e, defined by the lowest-energy configurations in our enumeration. The O–O bond length ranges used to classify each pathway are: peroxo (1.4–1.8 Å), superoxo (1.3–1.4 Å), and molecular O\u003csub\u003e2\u003c/sub\u003e (\u0026lt;1.28 Å). All formation energies are plotted relative to that of pristine Li\u003csub\u003e2\u003c/sub\u003eMnO\u003csub\u003e3\u003c/sub\u003e. “Oxygen Loss” represents the thermodynamically favored pathway; all others correspond to kinetically accessible intermediates. The shaded region denotes how the formation energy shifts when the partial oxygen pressure is lowered to the vacuum level typical of soft X-ray oxygen spectroscopies across different temperature ranges. b, Schematic energy landscape for oxygen redox reactions. Kinetic and thermodynamic pathways are distinguished by their free energy and activation barriers. The numbers in the plot indicate the difference between the maximum and minimum lithium chemical potentials among the kinetically accessible intermediates.\u003c/p\u003e","description":"","filename":"4.png","url":"https://assets-eu.researchsquare.com/files/rs-8285919/v1/78082ad38dabccd69de46eee.png"},{"id":102397172,"identity":"ee446dfe-2c44-4f99-b1e7-5f0897f7b61b","added_by":"auto","created_at":"2026-02-11 10:06:27","extension":"pdf","order_by":0,"title":"","display":"","copyAsset":false,"role":"manuscript-pdf","size":3297271,"visible":true,"origin":"","legend":"","description":"","filename":"manuscript.pdf","url":"https://assets-eu.researchsquare.com/files/rs-8285919/v1/2cbbf75a-fd66-4b48-8324-666787f1fedf.pdf"},{"id":102275706,"identity":"8ac88796-919a-47e5-a00d-7f646a13e00b","added_by":"auto","created_at":"2026-02-10 05:46:18","extension":"docx","order_by":1,"title":"","display":"","copyAsset":false,"role":"supplement","size":3234326,"visible":true,"origin":"","legend":"Supplementary Information","description":"","filename":"SupplementaryInformation.docx","url":"https://assets-eu.researchsquare.com/files/rs-8285919/v1/a8172e1505cda7a3247a3d3b.docx"}],"financialInterests":"There is \u003cb\u003eNO\u003c/b\u003e Competing Interest.","formattedTitle":"Vacuum-free spectroscopy reveals metastable oxygen redox intermediates in oxygen-redox cathodes","fulltext":[{"header":"Main text","content":"\u003cp\u003eReliably capturing metastable reaction intermediates has been central to uncovering fundamental reaction mechanisms across various fields such as photochemistry, biology, and radiolysis.\u003csup\u003e\u003cspan additionalcitationids=\"CR2 CR3\" citationid=\"CR1\" class=\"CitationRef\"\u003e1\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR4\" class=\"CitationRef\"\u003e4\u003c/span\u003e\u003c/sup\u003e Short-lived and typically highly energetic intermediates are measured via time-resolved measurements. Another important class of intermediates are long-lived and metastable. Because these intermediates are usually close in energy to the thermodynamically stable states, they readily relax, making their direct detection challenging as they are easily perturbed by the measurement itself. Oxygen redox (OR) in battery intercalation cathodes is one such example. Although OR has garnered significant attention as a promising chemistry to enhance the energy density of rechargeable batteries,\u003csup\u003e5,6\u003c/sup\u003e its underlying redox mechanisms are under debate, plausibly arising from difficulty in capturing metastable oxidized oxygen states.\u003c/p\u003e \u003cp\u003eOver the past decade, first-principles calculations have predicted numerous oxygen redox mechanisms.\u003csup\u003e\u003cspan additionalcitationids=\"CR7 CR8 CR9 CR10 CR11 CR12 CR13\" citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e\u003c/sup\u003e Early studies emphasized transition-metal-driven reductive coupling, in which peroxo-like O\u0026ndash;O species are formed alongside reduction of the transition metal.\u003csup\u003e\u003cspan citationid=\"CR6\" class=\"CitationRef\"\u003e6\u003c/span\u003e,\u003cspan citationid=\"CR7\" class=\"CitationRef\"\u003e7\u003c/span\u003e\u003c/sup\u003e Later reports challenged this view and predicted that different types of oxygen dimers can emerge during OR, with their bond lengths depending strongly on local structure, oxidation state, and states of charge (SoC).\u003csup\u003e\u003cspan additionalcitationids=\"CR9 CR10\" citationid=\"CR8\" class=\"CitationRef\"\u003e8\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e\u003c/sup\u003e Beyond dimer formation, alternative pathways have also been suggested, including high-valent Mn (\u003cem\u003ee.g.\u003c/em\u003e, Mn\u003csup\u003e4+/7+\u003c/sup\u003e) redox and \u003cem\u003eπ\u003c/em\u003e-back-bonding interactions.\u003csup\u003e\u003cspan additionalcitationids=\"CR13\" citationid=\"CR12\" class=\"CitationRef\"\u003e12\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR14\" class=\"CitationRef\"\u003e14\u003c/span\u003e\u003c/sup\u003e\u003c/p\u003e \u003cp\u003eExtensive experimental efforts have been made to validate these theoretical mechanisms, particularly using soft X-ray resonant inelastic X-ray scattering (RIXS).\u003csup\u003e\u003cspan additionalcitationids=\"CR16 CR17 CR18 CR19 CR20 CR21 CR22 CR23 CR24\" citationid=\"CR15\" class=\"CitationRef\"\u003e15\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR25\" class=\"CitationRef\"\u003e25\u003c/span\u003e\u003c/sup\u003e Through these measurements, a characteristic peak near ~\u0026thinsp;531 eV (in excitation) and ~\u0026thinsp;523 eV (in emission) is widely regarded as a fingerprint of oxygen redox. Recently, several studies have attributed this spectroscopic feature to trapped molecular oxygen (O\u003csub\u003e2\u003c/sub\u003e).\u003csup\u003e\u003cspan additionalcitationids=\"CR18 CR19 CR20 CR21 CR22\" citationid=\"CR17\" class=\"CitationRef\"\u003e17\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e23\u003c/span\u003e\u003c/sup\u003e This assignment is supported by the RIXS energy-loss features, which exhibit a vibrational structure characteristic of molecular O\u003csub\u003e2\u003c/sub\u003e.\u003c/p\u003e \u003cp\u003eThis spectroscopic signal has recently been used to support the claim that molecular O\u003csub\u003e2\u003c/sub\u003e is the sole oxidized-oxygen product in oxygen-redox reactions, and this interpretation has been applied across a broad range of cathodes, irrespective of their crystal structure, electronic configuration, or chemical composition,\u003csup\u003e17\u0026ndash;23\u003c/sup\u003e even for compounds that are believed not to undergo OR. Yet, these observations stand in contrast to theoretical predictions that a variety of oxidized oxygen species should form during OR.\u003csup\u003e9\u0026ndash;14\u003c/sup\u003e In addition, electrochemical profiles typically suggest multiple oxygen-redox peaks upon activation, indicating diverse redox environments. These inconsistencies have been recognized in numerous reports.\u003csup\u003e\u003cspan citationid=\"CR26\" class=\"CitationRef\"\u003e26\u003c/span\u003e,\u003cspan citationid=\"CR27\" class=\"CitationRef\"\u003e27\u003c/span\u003e\u003c/sup\u003e\u003c/p\u003e \u003cp\u003eIn this work, we demonstrate that these inconsistencies arise from the intrinsic metastability of oxidized oxygen intermediates, which are perturbed during the spectroscopic measurements owing to their energetic proximity to the thermodynamically stable state: O\u003csub\u003e2\u003c/sub\u003e(g). We show that vacuum environments can drive their transformation into stable O\u003csub\u003e2\u003c/sub\u003e in Li\u003csub\u003e1.13\u003c/sub\u003eNi\u003csub\u003e0.13\u003c/sub\u003eMn\u003csub\u003e0.57\u003c/sub\u003eO\u003csub\u003e2\u003c/sub\u003e. To overcome this challenge, we employed cryogenic hard X-ray Raman spectroscopy (XRS), which enables vacuum-free oxygen 1\u003cem\u003es\u003c/em\u003e spectroscopy with bulk-sensitivity.\u003csup\u003e\u003cspan citationid=\"CR28\" class=\"CitationRef\"\u003e28\u003c/span\u003e,\u003cspan citationid=\"CR29\" class=\"CitationRef\"\u003e29\u003c/span\u003e\u003c/sup\u003e We observe not one but several distinct oxygen species (also across LiNiO\u003csub\u003e2\u003c/sub\u003e, Li\u003csub\u003e2\u003c/sub\u003eRuO\u003csub\u003e3\u003c/sub\u003e, disordered rocksalt), including superoxo, molecular oxygen, and \u003cem\u003eπ\u003c/em\u003e-interacting oxygen. Our findings reconcile theoretical prediction, electrochemical profiles, and spectroscopic measurements, highlighting the importance of vacuum conditions when gaseous products are the stable products. Our demonstration highlights critical experimental considerations when elucidating the complex nature of oxygen redox that involves metastable intermediates.\u003c/p\u003e"},{"header":"Robust Bulk Spectral Measurements","content":"\u003cp\u003eHard XRS leverages inelastic X-ray scattering energy transfer to probe core-electron excitations (Supplementary Fig.\u0026nbsp;1). When operated in the low-momentum transfer regime, where dipole transitions dominate, it enables the acquisition of soft X-ray absorption-like spectral information using hard X-rays.\u003csup\u003e\u003cspan citationid=\"CR30\" class=\"CitationRef\"\u003e30\u003c/span\u003e,\u003cspan citationid=\"CR31\" class=\"CitationRef\"\u003e31\u003c/span\u003e\u003c/sup\u003e Owing to its high incident X-ray energy (~\u0026thinsp;10.2 keV), this technique does not require vacuum and enables ambient pressure, bulk-sensitive measurements with a penetration depth of a few tens of \u0026micro;m, sufficient to average signals over hundreds to thousands of battery crystallites in a single spectrum (see Method section for experimental details). Here, we employ XRS under cryogenic conditions\u0026mdash;preparing samples in liquid nitrogen and measuring them in a liquid-helium cryostat\u0026mdash;to further preserve plausible metastable oxygen states by minimizing aging effects and largely suppressing radiation-induced damage.\u003c/p\u003e \u003cp\u003eUsing vacuum-free and cryogenic XRS, we characterize plausible oxygen redox activity across four cathodes: LiNiO\u003csub\u003e2\u003c/sub\u003e, Li\u003csub\u003e2\u003c/sub\u003eRuO\u003csub\u003e3\u003c/sub\u003e, disordered rocksalt (DRX)-Li\u003csub\u003e1.17\u003c/sub\u003eNi\u003csub\u003e0.25\u003c/sub\u003eTi\u003csub\u003e0.58\u003c/sub\u003eO\u003csub\u003e2\u003c/sub\u003e, and Li\u003csub\u003e1.13\u003c/sub\u003eNi\u003csub\u003e0.3\u003c/sub\u003eMn\u003csub\u003e0.57\u003c/sub\u003eO\u003csub\u003e2\u003c/sub\u003e. We selected these materials because previous RIXS measurements have consistently detected a single spectroscopic signature (~\u0026thinsp;531 eV absorption peak) generally associated with dimerized oxygen,\u003csup\u003e19\u0026ndash;22,32,33\u003c/sup\u003e suggesting a common redox mechanism. Figure\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003ea\u0026ndash;d presents cryogenic XRS spectra at the O K-edge collected at two voltages during the first charge, together with the difference spectra between the two voltages. The voltages were selected based on previous reports of plausible oxygen redox activity as to capture the evolution before and after.\u003c/p\u003e \u003cp\u003eStrikingly, among the four materials examined, LiNiO\u003csub\u003e2\u003c/sub\u003e and Li\u003csub\u003e2\u003c/sub\u003eRuO\u003csub\u003e3\u003c/sub\u003e deviate substantially from prior reports. In LiNiO\u003csub\u003e2\u003c/sub\u003e, the absorption feature at ~\u0026thinsp;531 eV is absent entirely, and the spectrum exhibits a blue shift (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003ea), likely due to enhanced Ni\u0026ndash;O hybridization upon delithiation.\u003csup\u003e\u003cspan citationid=\"CR34\" class=\"CitationRef\"\u003e34\u003c/span\u003e\u003c/sup\u003e In Li\u003csub\u003e2\u003c/sub\u003eRuO\u003csub\u003e3\u003c/sub\u003e, by contrast, a strong single peak emerges at ~\u0026thinsp;530 eV (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003eb), rather than at ~\u0026thinsp;531 eV. Notably, neither of these materials shows the ~\u0026thinsp;531 eV peak commonly associated with dimerized oxygen.\u003c/p\u003e \u003cp\u003eOn the other hand, both DRX and Li\u003csub\u003e1.13\u003c/sub\u003eNi\u003csub\u003e0.3\u003c/sub\u003eMn\u003csub\u003e0.57\u003c/sub\u003eO\u003csub\u003e2\u003c/sub\u003e consistently exhibit the ~\u0026thinsp;531 eV absorption peak. However, whereas DRX shows only this single peak (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003ec), Li\u003csub\u003e1.13\u003c/sub\u003eNi\u003csub\u003e0.3\u003c/sub\u003eMn\u003csub\u003e0.57\u003c/sub\u003eO\u003csub\u003e2\u003c/sub\u003e exhibits multiple additional spectral features (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003ed and Supplementary Fig.\u0026nbsp;2), suggesting the presence of multiple oxidized oxygen species. We return later to assign these spectral features.\u003c/p\u003e \u003cp\u003eThe vastly diverse spectroscopic observations of oxygen across these four cathodes are consistent with their different electrochemical profiles (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003ee). In DRX, activation gives rise to only one new redox peak during discharge in the differential voltage curve (yellow curve in the third panel of Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003ee), consistent with the presence of a single spectral feature. In contrast, Li\u003csub\u003e1.13\u003c/sub\u003eNi\u003csub\u003e0.3\u003c/sub\u003eMn\u003csub\u003e0.57\u003c/sub\u003eO\u003csub\u003e2\u003c/sub\u003e exhibits several distinct redox peaks after activation on discharge (yellow curve in the fourth panel of Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003ee), consistent with the multiple features observed in vacuum-free and cryogenic XRS (Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003ed). For LiNiO\u003csub\u003e2\u003c/sub\u003e and Li\u003csub\u003e2\u003c/sub\u003eRuO\u003csub\u003e3\u003c/sub\u003e, however, no new peaks emerge in the differential voltage curves; instead, only the intensities of the original peaks change after activation (first and second panels of Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003ee). This one-to-one correspondence between electrochemical and spectroscopic signatures underscores the reliability of cryogenic XRS in capturing intrinsic oxygen-redox states or the lack thereof.\u003c/p\u003e \u003cdiv id=\"Sec3\" class=\"Section2\"\u003e \u003ch2\u003eVacuum and Irradiation Effects on Spectral Measurements\u003c/h2\u003e \u003cp\u003eThese distinct spectral features across different materials stand in clear contrast to the recent interpretation that molecular O\u003csub\u003e2\u003c/sub\u003e represents the sole oxidized-oxygen product in oxygen-redox reactions (Supplementary Note 1). We hypothesize that this inconsistency reflects the inherent metastability of oxidized lattice oxygen, which transforms into thermodynamically stable gaseous oxygen during measurement. To investigate this possibility, we employed the aging protocol shown in Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003ea on Li\u003csub\u003e1.13\u003c/sub\u003eNi\u003csub\u003e0.3\u003c/sub\u003eMn\u003csub\u003e0.57\u003c/sub\u003eO\u003csub\u003e2\u003c/sub\u003e. Electrodes were charged to 4.8 V, where OR is fully activated, aged under vacuum for up to 100 hours, and subsequently discharged. Vacuum aging reflects the ultra-high-vacuum conditions required for soft X-ray oxygen spectroscopy (typically\u0026thinsp;\u0026lt;\u0026thinsp;10\u003csup\u003e\u0026minus;\u0026thinsp;3\u003c/sup\u003e mTorr for several hours; Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003eb). Notably, vacuum provides the chemical driving force for oxygen release, approximately\u0026thinsp;\u0026minus;\u0026thinsp;0.1 eV/O\u003csub\u003e2\u003c/sub\u003e on the chemical potential scale. Additionally, cells were also aged at 30\u0026deg;C and 60\u0026deg;C as a baseline and to mimic plausible thermal effects during measurement, respectively.\u003c/p\u003e \u003cp\u003eAcross all storage conditions, the differential voltage curves during the first discharge exhibited a consistent trend (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003ec): high-voltage redox peaks diminished over time (red arrows), while a low-voltage peak at ~\u0026thinsp;3.16 V progressively intensified (yellow arrow). Previous studies have linked the emergence of this low-voltage peak to Mn(III/IV) redox activation, triggered by oxygen gas release.\u003csup\u003e\u003cspan citationid=\"CR23\" class=\"CitationRef\"\u003e23\u003c/span\u003e,\u003cspan citationid=\"CR35\" class=\"CitationRef\"\u003e35\u003c/span\u003e,\u003cspan citationid=\"CR36\" class=\"CitationRef\"\u003e36\u003c/span\u003e\u003c/sup\u003e Notably, this transformation occurred more rapidly under vacuum and at 60\u0026deg;C than the baseline at 30\u0026deg;C. This trend was consistently observed at different states of charge once OR was activated above 4.4 V (Supplementary Fig.\u0026nbsp;4). These results demonstrate that these activated electrodes (as measured electrochemically) are highly sensitive to environmental conditions, which can drive the transformation of metastable oxygen species even before X-ray measurements.\u003c/p\u003e \u003cp\u003eThis effect is experimentally confirmed by comparing cryogenic XRS spectra of two electrodes (Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003ed): one harvested immediately at 4.8 V and another exposed to vacuum for 24 hours prior to measurement\u0026mdash;the time required in our setup to reach the vacuum level needed of \u0026lt;\u0026thinsp;10\u003csup\u003e\u0026minus;\u0026thinsp;3\u003c/sup\u003e mTorr. All electrodes were stored in liquid nitrogen until measurement to prevent aging prior to XRS measurement. After vacuum exposure, most of the newly resolved features shown in Fig.\u0026nbsp;\u003cspan refid=\"Fig1\" class=\"InternalRef\"\u003e1\u003c/span\u003ed disappeared, and the ~\u0026thinsp;531 eV peak weakened, highlighting the critical role of vacuum in perturbing the oxidized oxygen intermediates in Li\u003csub\u003e1.13\u003c/sub\u003eNi\u003csub\u003e0.3\u003c/sub\u003eMn\u003csub\u003e0.57\u003c/sub\u003eO\u003csub\u003e2\u003c/sub\u003e (see Supplementary Note 2). In particular, the reduction in the ~\u0026thinsp;531 eV feature\u0026mdash;reflecting an irreversible loss of detectable oxygen states\u0026mdash;shows that even structurally confined molecular O\u003csub\u003e2\u003c/sub\u003e can be partially released, likely through diffusion of mobile O\u003csub\u003e2\u003c/sub\u003e.\u003csup\u003e37\u0026ndash;39\u003c/sup\u003e\u003c/p\u003e \u003cp\u003eWe also examined the impact of beam irradiation on the spectra by varying the X-ray dose. While cryogenic XRS measurements exhibited minimal beam-induced effects, the beam effect is significant at room temperature (Supplementary Fig.\u0026nbsp;7). Furthermore, Supplementary Fig.\u0026nbsp;8 shows that vacuum exposure and increased X-ray flux produce comparable reductions in peak intensity, with the ~\u0026thinsp;529 eV feature being particularly sensitive to both perturbations. These observations emphasize that both cryogenic temperature (at least in XRS) and vacuum-free condition are necessary conditions to detect OR intermediates in Li\u003csub\u003e1.13\u003c/sub\u003eNi\u003csub\u003e0.3\u003c/sub\u003eMn\u003csub\u003e0.57\u003c/sub\u003eO\u003csub\u003e2\u003c/sub\u003e.\u003c/p\u003e \u003cp\u003eThe necessity for XRS measurement at cryogenic conditions to capture intrinsic oxygen-redox states is further supported by comparing how the spectra evolve across different SoCs under cryogenic versus non-cryogenic conditions (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003e and Supplementary Fig.\u0026nbsp;9). Figure\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003ea shows the voltage profile of Li\u003csub\u003e1.13\u003c/sub\u003eNi\u003csub\u003e0.3\u003c/sub\u003eMn\u003csub\u003e0.57\u003c/sub\u003eO\u003csub\u003e2\u003c/sub\u003e during activation. OR activation occurs above 4.4 V, giving rise to a high-voltage plateau that follows the Ni(II/IV) redox. Oxygen spectra were collected at three points along this plateau: prior to OR activation (point #1 in Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003ea), at half activation (#2), and at full activation (#3).\u003c/p\u003e \u003cp\u003eUnder cryogenic conditions, multiple new oxygen-redox features appear and intensify from point #1 to point #3 along the high-voltage plateau (top panel in Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eb). In contrast, under non-cryogenic conditions, the ~\u0026thinsp;531 eV feature remains overwhelmingly dominant, with no clear development of the additional oxygen-redox features (bottom panel in Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003eb). Notably, in this voltage range, Mn L-edge and C K-edge spectra remained largely unchanged (Supplementary Fig.\u0026nbsp;10), indicating neither C\u0026ndash;O byproducts form nor Mn participates in redox within this voltage range. Again, these results confirm that cryogenic measurement is required for XRS to capture the OR intermediates (see Supplementary Note 3 for further discussion).\u003c/p\u003e \u003cp\u003eTo gain further insight into these newly emerging features observed in cryogenic XRS, we simulated O K-edge absorption spectra using the Bethe-Salpeter equation for Li\u003csub\u003e2\u003c/sub\u003eMnO\u003csub\u003e3\u003c/sub\u003e,\u003csup\u003e40\u003c/sup\u003e a parent phase that hosts oxygen redox activity in Li\u003csub\u003e1.13\u003c/sub\u003eNi\u003csub\u003e0.3\u003c/sub\u003eMn\u003csub\u003e0.57\u003c/sub\u003eO\u003csub\u003e2\u003c/sub\u003e (Supplementary Fig.\u0026nbsp;12; see Methods for details). We systematically enumerated diverse oxygen configurations (3,237 structures in total) across different lithium contents in Li\u003csub\u003e2\u0026minus;x\u003c/sub\u003eMnO\u003csub\u003e3\u003c/sub\u003e (see Supplementary Note 4). Based on bonding type and length, oxygen configurations were classified into five categories: peroxo (\u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{\\text{O}}_{2}^{2-}\\)\u003c/span\u003e\u003c/span\u003e), ozonide (\u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{\\text{O}}_{3}^{-}\\)\u003c/span\u003e\u003c/span\u003e), superoxo (\u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{\\text{O}}_{2}^{-}\\)\u003c/span\u003e\u003c/span\u003e), molecular oxygen (\u003cspan class=\"InlineEquation\"\u003e\u003cspan class=\"mathinline\"\u003e\\(\\:{\\text{O}}_{2}\\)\u003c/span\u003e\u003c/span\u003e), and short Mn\u0026ndash;O motifs with strong \u003cem\u003eπ\u003c/em\u003e-interaction (\u003cem\u003ed\u003c/em\u003e\u003csub\u003eMn\u0026ndash;O\u003c/sub\u003e \u0026lt; 1.7 \u0026Aring;).\u003csup\u003e9\u003c/sup\u003e\u003c/p\u003e \u003cp\u003eFigure\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003ec and Supplementary Fig.\u0026nbsp;14 present simulated spectra (dotted lines) for representative structures in each category, along with the summed spectra (scatter plots) within each panel. Each oxygen species exhibits a distinct spectral signature, with excitation energies markedly different from the ~\u0026thinsp;531 eV peak. Only molecular oxygen corresponds to the ~\u0026thinsp;531 eV feature. We then fit a linear combination of these simulated spectra to the experimental vacuum-free and cryogenic XRS data. For completeness, we also fit them to the non-cryogenic XRS (Fig.\u0026nbsp;\u003cspan refid=\"Fig3\" class=\"InternalRef\"\u003e3\u003c/span\u003ed). In the cryogenic spectra, the multiple peaks observed in the differential plot (before and after OR activation) align well with the features of \u003cem\u003eπ\u003c/em\u003e-complex, superoxo and molecular oxygen. These spectral evolutions are further quantified in Supplementary Note 5. By contrast, the non-cryogenic XRS spectra are mostly dominated by a single molecular oxygen signature, indicating that all metastable intermediates resolved in cryogenic XRS convert into molecular oxygen under non-cryogenic conditions. This behavior is consistent with the vacuum-driven transformations observed in Fig.\u0026nbsp;\u003cspan refid=\"Fig2\" class=\"InternalRef\"\u003e2\u003c/span\u003ed.\u003c/p\u003e \u003c/div\u003e"},{"header":"Mechanistic Basis of Spectral Measurement Challenges","content":"\u003cp\u003eThe challenges of reliably capturing OR intermediates can be rationalized within a reaction framework involving two competing reaction pathways: one that is kinetically accessible and the other that is thermodynamically favored. According to the equilibrium Li\u0026ndash;Mn\u0026ndash;O phase diagram,\u003csup\u003e12\u003c/sup\u003e upon delithiation, Li\u003csub\u003e2\u003c/sub\u003eMnO\u003csub\u003e3\u003c/sub\u003e thermodynamically decomposes into oxygen gas and spinel-like phases (Supplementary Fig.\u0026nbsp;16). This decomposition process corresponds to the lowest-energy state (labeled as \u0026ldquo;Oxygen Loss\u0026rdquo; in Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003ea). In contrast, the formation of \u003cem\u003eπ\u003c/em\u003e-complex and O\u0026ndash;O dimers represent metastable pathways, which lie at higher energies across all lithium contents (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003ea). The energy differences among these metastable intermediates are relatively small (0.3\u0026ndash;0.6 eV/Li), while the gap to thermodynamically stable state is ~\u0026thinsp;1.5 eV/Li.\u003c/p\u003e \u003cp\u003eThe activation free energy controls whether the activation process proceeds via the kinetic or the thermodynamic pathway (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eb). The equilibrium decomposition route involves extensive bond breaking and phase separation, resulting in a high kinetic barrier. In contrast, the formation of the metastable oxidized oxygen species requires only localized lattice distortions, making them kinetically more accessible with lower activation energies.\u003csup\u003e\u003cspan citationid=\"CR10\" class=\"CitationRef\"\u003e10\u003c/span\u003e,\u003cspan citationid=\"CR37\" class=\"CitationRef\"\u003e37\u003c/span\u003e\u003c/sup\u003e Crucially, the extent to which such distortions are accommodated strongly depends on the material\u0026rsquo;s structure, composition, and SoC, thereby governing the types and proportions of OR products captured by cryogenic XRS. Over time, these kinetic products gradually relax to the thermodynamic product through an energetically downhill reaction (Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eb), though the transformation is sluggish due to a high activation barrier.\u003csup\u003e\u003cspan citationid=\"CR11\" class=\"CitationRef\"\u003e11\u003c/span\u003e,\u003cspan additionalcitationids=\"CR38\" citationid=\"CR37\" class=\"CitationRef\"\u003e37\u003c/span\u003e\u0026ndash;\u003cspan citationid=\"CR39\" class=\"CitationRef\"\u003e39\u003c/span\u003e\u003c/sup\u003e This slow evolution is supported by experimental observations under 30\u0026deg;C aging, low-current cycling (Supplementary Fig.\u0026nbsp;17), and during extended cycling (Supplementary Fig.\u0026nbsp;18), each requiring prolonged time to induce noticeable change. By contrast, low oxygen partial pressure or elevated temperature (\u003cem\u003ee.g.\u003c/em\u003e, vacuum or 60\u0026deg;C) increases the driving force and accelerates the conversion (pink curve in Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003eb), as revealed by the shaded region in Fig.\u0026nbsp;\u003cspan refid=\"Fig4\" class=\"InternalRef\"\u003e4\u003c/span\u003ea (see Supplementary Note 6 for calculation details). Under such conditions, the energy of the \u0026ldquo;oxygen loss\u0026rdquo; pathway shifts downward by 0.4 eV/Li, widening its separation from the metastable intermediates to nearly 2 eV/Li. Such pressure-dependent stabilization provides a mechanistic basis for understanding the significant effect of vacuum on OR in Li\u003csub\u003e1.13\u003c/sub\u003eNi\u003csub\u003e0.3\u003c/sub\u003eMn\u003csub\u003e0.57\u003c/sub\u003eO\u003csub\u003e2\u003c/sub\u003e.\u003c/p\u003e"},{"header":"Summary","content":"\u003cp\u003eOur work establishes the importance of measurement conditions for capturing metastable oxygen-redox intermediates in intercalation cathodes. By eliminating vacuum environments and minimizing beam-induced effects through vacuum-free and cryogenic XRS, we resolve a diverse set of OR features across multiple cathodes. In Li\u003csub\u003e1.13\u003c/sub\u003eNi\u003csub\u003e0.3\u003c/sub\u003eMn\u003csub\u003e0.57\u003c/sub\u003eO\u003csub\u003e2\u003c/sub\u003e, multiple intermediates including superoxo and \u003cem\u003eπ\u003c/em\u003e-complex species were observed alongside molecular O\u003csub\u003e2\u003c/sub\u003e, whereas other cathodes such as LiNiO\u003csub\u003e2\u003c/sub\u003e and Li\u003csub\u003e2\u003c/sub\u003eRuO\u003csub\u003e3\u003c/sub\u003e exhibit markedly different features from prior reports, some absent of oxidized oxygen signatures altogether. Furthermore, we demonstrate that oxidized oxygen states are intrinsically metastable and can readily transform into molecular oxygen under vacuum or irradiation. Our work underscores the need for attention to measurement conditions when probing metastable oxygen-redox products and opens new avenues for capturing metastable products in this important class of redox reactions. Importantly, our findings place soft X-ray and non-vacuum-based measurements within a complementary framework, showing that each technique accesses mutually informative aspects of the oxygen-redox landscape.\u003c/p\u003e"},{"header":"Methods","content":"\u003cp\u003e\u003cstrong\u003e\u003cem\u003eMaterial synthesis.\u0026nbsp;\u003c/em\u003e\u003c/strong\u003eLi\u003csub\u003e1.13\u003c/sub\u003eNi\u003csub\u003e0.3\u003c/sub\u003eMn\u003csub\u003e0.57\u003c/sub\u003eO\u003csub\u003e2\u003c/sub\u003e was synthesized by mixing appropriate amounts of Li\u003csub\u003e2\u003c/sub\u003eCO\u003csub\u003e3\u003c/sub\u003e and (Ni\u003csub\u003e0.35\u003c/sub\u003eMn\u003csub\u003e0.65\u003c/sub\u003e)CO\u003csub\u003e3\u003c/sub\u003e, followed by calcination at 900 °C for 20 hours in air. LiNiO\u003csub\u003e2\u003c/sub\u003e was prepared by thoroughly mixing LiOH·H\u003csub\u003e2\u003c/sub\u003eO and Ni(OH)\u003csub\u003e2\u003c/sub\u003e in a stoichiometric ratio and calcining the mixture at 650 °C for 10 hours in O\u003csub\u003e2\u003c/sub\u003e. Li\u003csub\u003e2\u003c/sub\u003eRuO\u003csub\u003e3\u003c/sub\u003e and DRX-Li\u003csub\u003e1.17\u003c/sub\u003eNi\u003csub\u003e0.25\u003c/sub\u003eTi\u003csub\u003e0.58\u003c/sub\u003eO\u003csub\u003e2\u003c/sub\u003e were synthesized following previously reported procedures.\u003csup\u003e6,32\u003c/sup\u003e\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e\u003cem\u003eAging\u003c/em\u003e\u003c/strong\u003e\u003cstrong\u003e\u003cem\u003e\u0026nbsp;\u003c/em\u003e\u003c/strong\u003e\u003cstrong\u003e\u003cem\u003et\u003c/em\u003e\u003c/strong\u003e\u003cstrong\u003e\u003cem\u003eest\u003c/em\u003e\u003c/strong\u003e\u003cstrong\u003e\u003cem\u003eing\u003c/em\u003e\u003c/strong\u003e\u003cstrong\u003e\u003cem\u003e.\u0026nbsp;\u003c/em\u003e\u003c/strong\u003eFor 30 °C aging, cells were first charged to the target voltage, held at open circuit at 30 °C for the desired duration, and then discharged. For 60 °C aging, cells were charged at 30 °C and aged inside a temperature-controlled chamber at 60 °C for the specified time. After aging, they were transferred to a 30 °C chamber and equilibrated for 1 hour before discharge. For vacuum aging, charged cells were disassembled in an Ar-filled glovebox, and the harvested electrodes were exposed to vacuum using a HiCube pumping system for the designated duration. The electrodes were then reassembled into fresh cells, rested at 30 °C for 1 hour, and discharged.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e\u003cem\u003eX-ray Raman spectroscopy (XRS)\u003c/em\u003e\u003c/strong\u003e\u003cstrong\u003e\u003cem\u003e.\u003c/em\u003e\u003c/strong\u003e\u003cem\u003e\u0026nbsp;\u003c/em\u003eXRS were performed at Beamline 15-2 of the Stanford Synchrotron Radiation Lightsource. The radiation from the undulator was monochromated using a liquid nitrogen-cooled Si (311) double-crystal monochromator. X-rays were then focused onto the sample position to a beam size of 50 μm × 150 μm (H × V) using two Kirkpatrick-Baez (KB) mirrors. The XRS endstation at BL 15-2 consists of a multi-crystal Johann-type spectrometer with 40 Si (110) spherically bent diced analyzer crystals with a 1 m bending radius, operating in Rowland geometry.\u003csup\u003e30\u003c/sup\u003e\u003c/p\u003e\n\u003cp\u003eThe energy-loss measurements were performed in the so-called inverse geometry, \u003cem\u003ei.e.\u003c/em\u003e, keeping the analyzed energy ω\u003csub\u003e2\u003c/sub\u003e ≈ 9.69 keV fixed while varying the incident X-ray energy ω\u003csub\u003e1\u003c/sub\u003e, such that the transferred energy (ω = ω\u003csub\u003e1\u003c/sub\u003e – ω\u003csub\u003e2\u003c/sub\u003e) was scanned around the absorption edge. To optimize counting statistics and minimize the areal power density of the incident beam, all measurements were performed in reflection geometry, along the horizontal scattering plane, using a grazing angle of 2° to 4°. To minimize the effects of radiation energy deposition, all energy-loss scans were performed under cryogenic conditions (T ≈ 15 K).\u003c/p\u003e\n\u003cp\u003eThe XRS spectra presented in this work were obtained following a similar procedure as described in ref. 29, which allows for systematic control of radiation deposition on the sample during the experiment. For each sample, several energy-loss scans were acquired over numerous spots under the same experimental conditions. The acquisition time for each scan was 2.5s or 5 s per energy point with a total amount of 148 energy points per scan were measured for all experiments. For accurate energy calibration, periodic measurements of the elastic line were acquired during the experiments. Finally, by measuring over many fresh spots across each sample, the collected set of equivalent spectra was used to construct an averaged spectrum with improved statistical quality.\u003c/p\u003e\n\u003cp\u003eThe fitting procedure for background subtraction and normalization of the energy-loss scans followed a similar approach as discussed in refs. 28,29. The process involves simultaneous fitting of two components: (1) a second-order polynomial function accounting for the high-energy tail of the valence electron contribution, and (2) a model for the core-electron atomic background contribution,\u0026nbsp;\u0026nbsp;, to the total spectrum.\u003c/p\u003e\n\u003cp\u003eFor modeling the core-electron atomic background, Hartree-Fock tabulated atomic Compton profiles\u0026nbsp;\u0026nbsp;\u0026nbsp;were converted to\u0026nbsp;\u0026nbsp;\u0026nbsp;using the formulation of Ribberfors and Holm.\u003csup\u003e41,42\u003c/sup\u003e Additionally, asymmetry correction\u0026nbsp;\u0026nbsp;\u0026nbsp;to the core-electron Compton profile was included.\u003csup\u003e41\u003c/sup\u003e The polynomial function was fitted to the total energy-loss spectrum for energy transfers ω \u0026lt; 527 eV. For the high-energy tail (ω \u0026gt; 565 eV), the sum of the polynomial (valence-electron contribution) and the modeled atomic background (core-electron contribution) was fitted. All extracted XRS spectra were then normalized to have the same integrated intensity for energy transfers above 565 eV.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e\u003cem\u003eScanning\u0026nbsp;\u003c/em\u003e\u003c/strong\u003e\u003cstrong\u003e\u003cem\u003et\u003c/em\u003e\u003c/strong\u003e\u003cstrong\u003e\u003cem\u003eransmission X-ray\u0026nbsp;\u003c/em\u003e\u003c/strong\u003e\u003cstrong\u003e\u003cem\u003em\u003c/em\u003e\u003c/strong\u003e\u003cstrong\u003e\u003cem\u003eicroscopy (STXM).\u003c/em\u003e\u003c/strong\u003e\u003cem\u003e\u0026nbsp;\u003c/em\u003eDispersed powder samples were prepared by sonicating in isopropyl alcohol (\u0026gt;99.5%) for 30 minutes. The resulting suspension was drop-cast onto copper TEM grids coated with a carbon film. STXM measurements were conducted at Beamline 7.0.1.2 of the Advanced Light Source, Lawrence Berkeley National Laboratory. A 53-energy STXM stack was collected for each sample with a pixel size of 50 nm.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e\u003cem\u003eTotal\u0026nbsp;\u003c/em\u003e\u003c/strong\u003e\u003cstrong\u003e\u003cem\u003ee\u003c/em\u003e\u003c/strong\u003e\u003cstrong\u003e\u003cem\u003electron\u0026nbsp;\u003c/em\u003e\u003c/strong\u003e\u003cstrong\u003e\u003cem\u003ey\u003c/em\u003e\u003c/strong\u003e\u003cstrong\u003e\u003cem\u003eield (TEY); Total\u0026nbsp;\u003c/em\u003e\u003c/strong\u003e\u003cstrong\u003e\u003cem\u003ef\u003c/em\u003e\u003c/strong\u003e\u003cstrong\u003e\u003cem\u003eluorescence\u0026nbsp;\u003c/em\u003e\u003c/strong\u003e\u003cstrong\u003e\u003cem\u003ey\u003c/em\u003e\u003c/strong\u003e\u003cstrong\u003e\u003cem\u003eield (TFY); Resonant\u0026nbsp;\u003c/em\u003e\u003c/strong\u003e\u003cstrong\u003e\u003cem\u003ei\u003c/em\u003e\u003c/strong\u003e\u003cstrong\u003e\u003cem\u003enelastic X-ray\u0026nbsp;\u003c/em\u003e\u003c/strong\u003e\u003cstrong\u003e\u003cem\u003es\u003c/em\u003e\u003c/strong\u003e\u003cstrong\u003e\u003cem\u003ecattering (RIXS).\u003c/em\u003e\u003c/strong\u003e\u003cem\u003e\u0026nbsp;\u003c/em\u003eSoft X-ray TEY, TFY, and RIXS spectra were collected at the ultra-high-efficiency iRIXS endstation on beamline 8.0.1 of the Advanced Light Source (ALS), Lawrence Berkeley National Laboratory. All electrodes were harvested in an Ar-filled glovebox, double-sealed in aluminum pouches, and transferred into a custom vacuum-compatible transport kit to avoid air exposure. O K-edge RIXS spectra were acquired simultaneously with TEY and TFY modes, with two-dimensional RIXS maps constructed by recording emission spectra at each excitation energy. Emission energies were aligned to the elastic scattering line. To minimize beam-induced damage, samples were dithered laterally over a 0.6 mm range during measurement. Excitation energies were calibrated using a TiO\u003csub\u003e2\u003c/sub\u003e reference for the O K-edge.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e\u003cem\u003eX-ray photoelectron spectroscopy (XPS)\u003c/em\u003e\u003c/strong\u003e\u003cstrong\u003e\u003cem\u003e.\u003c/em\u003e\u003c/strong\u003e\u003cem\u003e\u0026nbsp;\u003c/em\u003eLaboratory-based XPS measurements were performed using a PHI VersaProbe 4 instrument equipped with a monochromated Al Kα X-ray source (1486.6 eV, 47.9 W). Electrodes were mounted using double-sided tape, and charge compensation was applied using a dual ion and electron flood gun (1 V, 3 μA). All samples were loaded in an Ar-filled glovebox and transferred to the instrument with less than 1 minute of air exposure. The X-ray beam diameter, power, and accelerating voltage were 200 μm, 50 W, and 15 kV, respectively. Survey scans were collected with a pass energy of 224 eV and step size of 0.8 eV, while core-level scans used a pass energy of 55 eV and step size of 0.1 eV.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e\u003cem\u003eDensity\u0026nbsp;\u003c/em\u003e\u003c/strong\u003e\u003cstrong\u003e\u003cem\u003ef\u003c/em\u003e\u003c/strong\u003e\u003cstrong\u003e\u003cem\u003eunctional\u0026nbsp;\u003c/em\u003e\u003c/strong\u003e\u003cstrong\u003e\u003cem\u003et\u003c/em\u003e\u003c/strong\u003e\u003cstrong\u003e\u003cem\u003eheory (DFT)\u0026nbsp;\u003c/em\u003e\u003c/strong\u003e\u003cstrong\u003e\u003cem\u003ec\u003c/em\u003e\u003c/strong\u003e\u003cstrong\u003e\u003cem\u003ealculation.\u003c/em\u003e\u003c/strong\u003e Density functional theory (DFT) calculations were performed using the Vienna Ab Initio Simulation Package (VASP, version 6.4.2), based on previous work by Chen \u003cem\u003eet al\u003c/em\u003e.\u003csup\u003e9\u003c/sup\u003e Initial structural relaxations were carried out using the Perdew-Burke-Ernzerhof (PBE) implementation of the generalized gradient approximation (GGA), with a Hubbard \u003cem\u003eU\u003c/em\u003e+\u003cem\u003eJ\u003c/em\u003e correction applied to Mn (\u003cem\u003eU\u003c/em\u003e = 5.0 eV, \u003cem\u003eJ\u003c/em\u003e = 0.5 eV), following the Liechtenstein formalism.\u003csup\u003e43,44\u003c/sup\u003e After identifying the lowest-energy configurations for each peroxide-type species, further relaxations were performed using the r\u003csup\u003e2\u003c/sup\u003eSCAN exchange-correlation functional to better account for the variable oxidation states of Mn in the system.\u003csup\u003e45\u003c/sup\u003e All calculations employed a plane-wave energy cutoff of 600 eV, a self-consistent field (SCF) convergence threshold of 1 × 10⁻\u003csup\u003e5\u003c/sup\u003e eV, and a force convergence criterion of 0.02 eV/Å. A monoclinic unit cell containing four formula units of Li\u003csub\u003e2\u003c/sub\u003eMnO\u003csub\u003e3\u003c/sub\u003e was used, with a \u003cem\u003ek\u003c/em\u003e-point mesh of 6 × 3 × 6.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003e\u003cem\u003eSpectroscopic OCEAN\u0026nbsp;\u003c/em\u003e\u003c/strong\u003e\u003cstrong\u003e\u003cem\u003ec\u003c/em\u003e\u003c/strong\u003e\u003cstrong\u003e\u003cem\u003ealculation.\u003c/em\u003e\u003c/strong\u003e\u003cem\u003e\u0026nbsp;\u003c/em\u003eOxygen K-edge X-ray absorption spectra were computed using the Bethe-Salpeter Equation (BSE)-based OCEAN code (version 3.0.4), with DFT wavefunctions generated by Quantum ESPRESSO (version 7.0). Pseudopotentials were constructed using the scalar-relativistic Optimized Norm-Conserving Vanderbilt Pseudopotential (ONCVPSP) method.\u003csup\u003e46\u003c/sup\u003e The r2SCAN-relaxed structures obtained from the VASP calculations were used as inputs for all spectral simulations. Spectroscopic calculations were performed with a plane-wave cutoff of 140 Ry and an SCF convergence threshold of 1 × 10⁻\u003csup\u003e10\u003c/sup\u003e Ry. A k-point mesh of 5 × 3 × 5 and 110 unoccupied bands were included in the BSE calculation. A uniform\u0026nbsp;\u003c/p\u003e\n\u003cp\u003eenergy shift was applied to align all spectra at a given Li content to the post-edge continuum maximum, positioned at approximately 542.5 eV.\u0026nbsp;\u003c/p\u003e"},{"header":"Declarations","content":"\u003cp\u003e*** Certain equipment, instruments, software, or materials are identified in this paper to adequately specify the experimental procedure. Such identification does not imply recommendation or endorsement of any product or service by NIST, nor does it imply that the materials or equipment identified are necessarily the best available for the purpose. *** \u003c/p\u003e\n\n\u003cp\u003e\u003cstrong\u003eData availability\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eAll data are available in the main text or the supplementary materials. Additional data in this paper are available in the Stanford Digital Repository/Dryad.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAcknowledgements\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe characterization aspect work was supported by the Assistant Secretary for Energy Efficiency and Renewable Energy, Office of Vehicle Technologies, Battery Materials Research Program, U.S. Department of Energy (DOE). E.Z.C.was supported in part by an ALS Doctoral Fellowship in Residence. Theoretical work is supported in part by the U.S. Department of Energy, Office of Basic Energy Sciences, Materials Sciences and Engineering Division. Calculations were performed on the Sherlock cluster at Stanford University and on resources of the National Energy Research Scientific Computing Center (NERSC), a Department of Energy Office of Science User Facility, using NERSC award BES-ERCAP0027203. We thank Dr. Sung-O Park for valuable discussions regarding the calculation results.\u003c/p\u003e\n\u003cp\u003eThis research used resources of the Stanford Synchrotron Radiation Lightsource (SSRL), SLAC National Accelerator Laboratory, which is supported by the DOE Office of Science, Office of Basic Energy Sciences under Contract No. DE-AC02-76SF00515. X-ray Raman analysis used resources of the National Energy Research Scientific Computing Center, a DOE Office of Science User Facility supported by the Office of Science of the U.S. Department of Energy under Contract No. DE-AC02-05CH11231 using NERSC award BES-ERCAP0031542. Use of the Advanced Light Source (ALS) at Lawrence Berkeley National Laboratory is supported by the DOE Office of Science User Facility under Contract No. DE-AC02-05CH11231. This work also utilized resources of the Stanford Nano Shared Facilities (SNSF), supported by the National Science Foundation under award ECCS-2026822.\u003c/p\u003e\n\u003cp\u003e\u003cstrong\u003eAuthor contributions\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eD.E. and O.P.M. conceived the original idea. D.E., O.P.M., D. Sokaras, and W.C.C. designed the research. D.E. and O.P.M. carried out the XRS measurements with assistance from D. Skoien and D. Sokaras. O.P.M. performed the XRS data analysis. E.G.L. performed the DFT calculations and spectral simulations with constructive advice from J.V. and T.P.D. E.M. and H.R. collected the RIXS, TFY, and TEY datasets, and E.P.K.L.C. performed the XPS measurements. E.P.K.L.C. and N.K. conducted the STXM measurements, and E.Z.C. processed the data. S.L. and J.A.D. offered valuable comments for this project. D.E. and W.C.C. wrote the manuscript, and all authors contributed to its revision. D.E., D. Sokaras, and W.C.C. supervised all aspects of the research.\u003c/p\u003e\n\n\n\u003cp\u003e\u003cstrong\u003eCompeting interests\u003c/strong\u003e\u003c/p\u003e\n\u003cp\u003eThe authors declare no competing interests.\u003c/p\u003e"},{"header":"References","content":"\u003col\u003e\n\u003cli\u003eZewail, A. H. Femtochemistry:\u0026thinsp; Atomic-Scale Dynamics of the Chemical Bond. \u003cem\u003eJ. Phys. Chem. A\u003c/em\u003e \u003cstrong\u003e104\u003c/strong\u003e, 5660\u0026ndash;5694 (2000).\u003c/li\u003e\n\u003cli\u003eChen, L. X. Probing transient molecular structures in photochemical processes using laser-initiated time-resolved X-ray absorption spectroscopy. \u003cem\u003eAnnu. Rev. Phys. Chem.\u003c/em\u003e \u003cstrong\u003e56\u003c/strong\u003e, 221\u0026ndash;254 (2005).\u003c/li\u003e\n\u003cli\u003eWilliams, P. J. H. et al. New Approach to the Detection of Short-Lived Radical Intermediates. \u003cem\u003eJ. Am. Chem. Soc.\u003c/em\u003e \u003cstrong\u003e144\u003c/strong\u003e, 15969\u0026ndash;15976 (2022).\u003c/li\u003e\n\u003cli\u003eLin, M. F. et al. Imaging the short-lived hydroxyl-hydronium pair in ionized liquid water. \u003cem\u003eScience\u003c/em\u003e \u003cstrong\u003e374\u003c/strong\u003e, 92\u0026ndash;95 (2021).\u003c/li\u003e\n\u003cli\u003eRahman, M. M. \u0026amp; Lin, F. Oxygen redox chemistry in rechargeable Li-ion and Na-ion batteries. \u003cem\u003eMatter\u003c/em\u003e \u003cstrong\u003e4\u003c/strong\u003e, 490\u0026ndash;527 (2021).\u003c/li\u003e\n\u003cli\u003eSathiya, M. et al. Reversible anionic redox chemistry in high-capacity layered-oxide electrodes. \u003cem\u003eNat. Mater.\u003c/em\u003e \u003cstrong\u003e12\u003c/strong\u003e, 827\u0026ndash;835 (2013).\u003c/li\u003e\n\u003cli\u003eSauban\u0026egrave;re, M., McCalla, E., Tarascon, J.M. \u0026amp; Doublet, M. L. The intriguing question of anionic redox in high-energy density cathodes for Li-ion batteries. \u003cem\u003eEnergy Environ. Sci.\u003c/em\u003e \u003cstrong\u003e9\u003c/strong\u003e, 984\u0026ndash;991 (2016).\u003c/li\u003e\n\u003cli\u003eChen, H. \u0026amp; Islam, M. S. Lithium Extraction Mechanism in Li-Rich Li\u003csub\u003e2\u003c/sub\u003eMnO\u003csub\u003e3\u003c/sub\u003e Involving Oxygen Hole Formation and Dimerization. \u003cem\u003eChem. Mater.\u003c/em\u003e \u003cstrong\u003e28\u003c/strong\u003e, 6656\u0026ndash;6663 (2016).\u003c/li\u003e\n\u003cli\u003eChen, Z., Li, J. \u0026amp; Zeng, X. C. Unraveling oxygen evolution in Li-rich oxides: a unified modeling of the intermediate peroxo/superoxo-like dimers. \u003cem\u003eJ. Am. Chem. Soc.\u003c/em\u003e \u003cstrong\u003e141\u003c/strong\u003e, 10751\u0026ndash;10759 (2019).\u003c/li\u003e\n\u003cli\u003eVinckeviciute, J., Kitchaev, D. A. \u0026amp; Van der Ven, A. A two-step oxidation mechanism controlled by Mn migration explains the first-cycle activation behavior of Li\u003csub\u003e2\u003c/sub\u003eMnO\u003csub\u003e3\u003c/sub\u003e-based Li-excess materials. \u003cem\u003eChem. Mater.\u003c/em\u003e \u003cstrong\u003e33\u003c/strong\u003e, 1625\u0026ndash;1636 (2021).\u003c/li\u003e\n\u003cli\u003eKim, B. et al. A theoretical framework for oxygen redox chemistry for sustainable batteries. \u003cem\u003eNat. Sustainability\u003c/em\u003e, \u003cstrong\u003e5\u003c/strong\u003e, 708\u0026ndash;716 (2022).\u003c/li\u003e\n\u003cli\u003eRadin, M. D., Vinckeviciute, J., Seshadri, R. \u0026amp; Van der Ven, A. Manganese oxidation as the origin of the anomalous capacity of Mn-containing Li-excess cathode materials. \u003cem\u003eNat. Energy\u003c/em\u003e \u003cstrong\u003e4\u003c/strong\u003e, 639\u0026ndash;646 (2019).\u003c/li\u003e\n\u003cli\u003eSudayama, T. et al. Multiorbital bond formation for stable oxygen-redox reaction in battery electrodes. \u003cem\u003eEnergy Environ. Sci.\u003c/em\u003e \u003cstrong\u003e13\u003c/strong\u003e, 1492\u0026ndash;1500 (2020).\u003c/li\u003e\n\u003cli\u003eEum, D. et al. Coupling structural evolution and oxygen-redox electrochemistry in layered transition metal oxides. \u003cem\u003eNat. Mater.\u003c/em\u003e \u003cstrong\u003e21\u003c/strong\u003e, 664\u0026ndash;672 (2022).\u003c/li\u003e\n\u003cli\u003eGent, W. E. et al. Coupling between oxygen redox and cation migration explains unusual electrochemistry in lithium-rich layered oxides. \u003cem\u003eNat. Commun.\u003c/em\u003e \u003cstrong\u003e8\u003c/strong\u003e, 2091 (2017).\u003c/li\u003e\n\u003cli\u003eHong, J. et al. Metal\u0026ndash;oxygen decoordination stabilizes anion redox in Li-rich oxides. \u003cem\u003eNat. Mater.\u003c/em\u003e \u003cstrong\u003e18\u003c/strong\u003e, 256\u0026ndash;265 (2019).\u003c/li\u003e\n\u003cli\u003eMcColl, K. et al. Transition metal migration and O\u003csub\u003e2\u003c/sub\u003e formation underpin voltage hysteresis in oxygen-redox disordered rocksalt cathodes. \u003cem\u003eNat. Commun.\u003c/em\u003e \u003cstrong\u003e13\u003c/strong\u003e, 5275 (2022).\u003c/li\u003e\n\u003cli\u003eHouse, R. A. et al. Superstructure control of first-cycle voltage hysteresis in oxygen-redox cathodes. \u003cem\u003eNature\u003c/em\u003e \u003cstrong\u003e577\u003c/strong\u003e, 502\u0026ndash;508 (2020).\u003c/li\u003e\n\u003cli\u003eHouse, R. A. et al. The role of O\u003csub\u003e2\u003c/sub\u003e in O-redox cathodes for Li-ion batteries. \u003cem\u003eNat. Energy\u003c/em\u003e \u003cstrong\u003e6\u003c/strong\u003e, 781\u0026ndash;789 (2021).\u003c/li\u003e\n\u003cli\u003eHouse, R. A. et al. Covalency does not suppress O\u003csub\u003e2\u003c/sub\u003e formation in 4d and 5d Li-rich O-redox cathodes. \u003cem\u003eNat. Commun.\u003c/em\u003e \u003cstrong\u003e12\u003c/strong\u003e, 2975 (2021).\u003c/li\u003e\n\u003cli\u003eJuelsholt, M. et al. Does trapped O\u003csub\u003e2\u003c/sub\u003e form in the bulk of LiNiO\u003csub\u003e2\u003c/sub\u003e during charging?. \u003cem\u003eEnergy Environ. Sci.\u003c/em\u003e \u003cstrong\u003e17\u003c/strong\u003e, 2530\u0026ndash;2540 (2024).\u003c/li\u003e\n\u003cli\u003eOgley, M. J. W. et al. Metal-ligand redox in layered oxide cathodes for Li-ion batteries. \u003cem\u003eJoule\u003c/em\u003e \u003cstrong\u003e9\u003c/strong\u003e, 101775 (2025).\u003c/li\u003e\n\u003cli\u003eMarie, J. J. et al. Trapped O\u003csub\u003e2\u003c/sub\u003e and the origin of voltage fade in layered Li-rich cathodes. \u003cem\u003eNat. Mater.\u003c/em\u003e \u003cstrong\u003e23\u003c/strong\u003e, 818\u0026ndash;825 (2024).\u003c/li\u003e\n\u003cli\u003eEum, D. et al. Voltage decay and redox asymmetry mitigation by reversible cation migration in lithium-rich layered oxide electrodes. \u003cem\u003eNat. Mater.\u003c/em\u003e \u003cstrong\u003e19\u003c/strong\u003e, 419\u0026ndash;427 (2020).\u003c/li\u003e\n\u003cli\u003eJang, H. Y. et al. Structurally robust lithium-rich layered oxides for high-energy and long-lasting cathodes. \u003cem\u003eNat. Commun.\u003c/em\u003e \u003cstrong\u003e15\u003c/strong\u003e, 1288 (2024).\u003c/li\u003e\n\u003cli\u003eGao, X. et al. Clarifying the origin of molecular O\u003csub\u003e2\u003c/sub\u003e in cathode oxides. \u003cem\u003eNat. Mater.\u003c/em\u003e \u003cstrong\u003e24\u003c/strong\u003e, 743\u0026ndash;752 (2025).\u003c/li\u003e\n\u003cli\u003eLebens-Higgins, Z. W. et al. Distinction between intrinsic and X-ray-induced oxidized oxygen states in Li-rich 3d layered oxides and LiAlO\u003csub\u003e2\u003c/sub\u003e. \u003cem\u003eJ. Phys. Chem. C\u003c/em\u003e \u003cstrong\u003e123\u003c/strong\u003e, 13201\u0026ndash;13207 (2019).\u003c/li\u003e\n\u003cli\u003eParedes-Mellone, O. A. et al. Investigating the electronic structure of high explosives with X-ray Raman spectroscopy. \u003cem\u003eSci. Rep.\u003c/em\u003e \u003cstrong\u003e12\u003c/strong\u003e, 19460 (2022).\u003c/li\u003e\n\u003cli\u003eParedes-Mellone, O. A. et al. Deciphering decomposition pathways of high explosives with cryogenic X-ray Raman spectroscopy. \u003cem\u003ePNAS.\u003c/em\u003e \u003cstrong\u003e122\u003c/strong\u003e, e2426320122 (2025).\u003c/li\u003e\n\u003cli\u003eSokaras, D. et al. A high resolution and large solid angle x-ray Raman spectroscopy end-station at the Stanford Synchrotron Radiation Lightsource. \u003cem\u003eRev. Sci. Instrum.\u003c/em\u003e \u003cstrong\u003e83\u003c/strong\u003e, 043112 (2012).\u003c/li\u003e\n\u003cli\u003eSch\u0026uuml;lke, W. \u003cem\u003eElectron Dynamics by Inelastic X-ray Scattering\u003c/em\u003e (Oxford, 2007).\u003c/li\u003e\n\u003cli\u003eLi, B. et al. Capturing dynamic ligand-to-metal charge transfer with a long-lived cationic intermediate for anionic redox. \u003cem\u003eNat. Mater.\u003c/em\u003e \u003cstrong\u003e21\u003c/strong\u003e, 1165\u0026ndash;1174 (2022).\u003c/li\u003e\n\u003cli\u003eLi, N. et al. Unraveling the cationic and anionic redox reactions in a conventional layered oxide cathode. \u003cem\u003eACS Energy Lett.\u003c/em\u003e \u003cstrong\u003e4\u003c/strong\u003e, 2836\u0026ndash;2842 (2019).\u003c/li\u003e\n\u003cli\u003eCao, R., Thomas, K. E., Ghosh, A. \u0026amp; Sarangi, R. X-ray absorption spectroscopy of archetypal chromium porphyrin and corrole derivatives. RSC Adv. 10, 20572\u0026ndash;20578 (2020).\u003c/li\u003e\n\u003cli\u003eHu, E. et al. Evolution of redox couples in Li-and Mn-rich cathode materials and mitigation of voltage fade by reducing oxygen release. \u003cem\u003eNat. Energy\u003c/em\u003e \u003cstrong\u003e3\u003c/strong\u003e, 690\u0026ndash;698 (2018).\u003c/li\u003e\n\u003cli\u003eYan, P. et al. Injection of oxygen vacancies in the bulk lattice of layered cathodes. \u003cem\u003eNat. Nanotechnol.\u003c/em\u003e \u003cstrong\u003e14\u003c/strong\u003e, 602\u0026ndash;608 (2019).\u003c/li\u003e\n\u003cli\u003eMcColl, K., Coles, S. W., Zarabadi-Poor, P., Morgan, B. J. \u0026amp; Islam, M. S. Phase segregation and nanoconfined fluid O\u003csub\u003e2\u003c/sub\u003e in a lithium-rich oxide cathode. \u003cem\u003eNat. Mater.\u003c/em\u003e \u003cstrong\u003e23\u003c/strong\u003e, 826\u0026ndash;833 (2024).\u003c/li\u003e\n\u003cli\u003eEum, D. et al. Electrochemomechanical failure in layered oxide cathodes caused by rotational stacking faults. \u003cem\u003eNat. Mater.\u003c/em\u003e \u003cstrong\u003e23\u003c/strong\u003e, 1093\u0026ndash;1099 (2024).\u003c/li\u003e\n\u003cli\u003eCsernica, P. M. et al. Substantial oxygen loss and chemical expansion in lithium-rich layered oxides at moderate delithiation. \u003cem\u003eNat. Mater.\u003c/em\u003e \u003cstrong\u003e24\u003c/strong\u003e, 92\u0026ndash;100 (2025).\u003c/li\u003e\n\u003cli\u003eVinson, J. Advances in the OCEAN-3 spectroscopy package. \u003cem\u003ePhys. Chem. Chem. Phys.\u003c/em\u003e \u003cstrong\u003e24\u003c/strong\u003e, 12787\u0026ndash;12803 (2022).\u003c/li\u003e\n\u003cli\u003eHolm, P. \u0026amp; Ribberfors, R. First correction to the nonrelativistic Compton cross section in the impulse approximation. \u003cem\u003ePhys. Rev. A\u003c/em\u003e \u003cstrong\u003e40\u003c/strong\u003e, 6251 (1989).\u003c/li\u003e\n\u003cli\u003eSternemann, H. et al. An extraction algorithm for core-level excitations in non-resonant inelastic X-ray scattering spectra. \u003cem\u003eJ. Synchrotron Rad.\u003c/em\u003e \u003cstrong\u003e15\u003c/strong\u003e, 162\u0026ndash;169 (2008).\u003c/li\u003e\n\u003cli\u003ePerdew, J. P. et al. Generalized Gradient Approximation Made Simple. \u003cem\u003ePhys. Rev. Lett.\u003c/em\u003e \u003cstrong\u003e77\u003c/strong\u003e, 3865 (1996).\u003c/li\u003e\n\u003cli\u003eLiechtenstein, A. I., Anisimov, V. I. \u0026amp; Zaanen, J. Density-functional theory and strong interactions: Orbital ordering in Mott-Hubbard insulators. \u003cem\u003ePhys. Rev. B\u003c/em\u003e \u003cstrong\u003e52\u003c/strong\u003e, R5467(R) (1995).\u003c/li\u003e\n\u003cli\u003eFurness, J. W., Kaplan, A. D., Ning, J., Perdew, J. P. \u0026amp; Sun, J. Accurate and Numerically Efficient r2SCAN Meta-Generalized Gradient Approximation. \u003cem\u003eJ. Phys. Chem. Lett.\u003c/em\u003e \u003cstrong\u003e11\u003c/strong\u003e, 8208\u0026ndash;8215 (2020).\u003c/li\u003e\n\u003cli\u003eHamann, D. R. Optimized norm-conserving Vanderbilt pseudopotentials. \u003cem\u003ePhys. Rev. B\u003c/em\u003e\u003cstrong\u003e88\u003c/strong\u003e, 085117 (2017). \u003c/li\u003e\n\u003c/ol\u003e"}],"fulltextSource":"","fullText":"","funders":[],"hasAdminPriorityOnWorkflow":false,"hasManuscriptDocX":true,"hasOptedInToPreprint":true,"hasPassedJournalQc":"","hasAnyPriority":true,"hideJournal":false,"highlight":"","institution":"","isAcceptedByJournal":false,"isAuthorSuppliedPdf":false,"isDeskRejected":"","isHiddenFromSearch":false,"isInQc":false,"isInWorkflow":false,"isPdf":false,"isPdfUpToDate":true,"isWithdrawnOrRetracted":false,"journal":{"display":true,"email":"[email protected]","identity":"nature-portfolio","isNatureJournal":true,"hasQc":false,"allowDirectSubmit":false,"externalIdentity":"","sideBox":"","snPcode":"","submissionUrl":"","title":"Nature Portfolio","twitterHandle":"","acdcEnabled":false,"dfaEnabled":false,"editorialSystem":"ejp","reportingPortfolio":"","inReviewEnabled":true,"inReviewRevisionsEnabled":false},"keywords":"","lastPublishedDoi":"10.21203/rs.3.rs-8285919/v1","lastPublishedDoiUrl":"https://doi.org/10.21203/rs.3.rs-8285919/v1","license":{"name":"CC BY 4.0","url":"https://creativecommons.org/licenses/by/4.0/"},"manuscriptAbstract":"\u003cp\u003eCapturing metastable reaction intermediates through spectroscopy is a fundamental challenge across many fields, as their intrinsic metastability often leads to transformation while being measured. This challenge is acute in oxygen redox in battery intercalation cathodes, where metastable intermediates, despite their long lifetime, remain difficult to capture reliably. Here, we demonstrate that vacuum conditions destabilize oxidized oxygen intermediates and drive their transformation to thermodynamically stable O\u003csub\u003e2\u003c/sub\u003e in Li\u003csub\u003e1.13\u003c/sub\u003eNi\u003csub\u003e0.13\u003c/sub\u003eMn\u003csub\u003e0.57\u003c/sub\u003eO\u003csub\u003e2\u003c/sub\u003e. To overcome these issues, we employ vacuum-free and cryogenic hard X-ray Raman spectroscopy, circumventing both vacuum and radiation-induced effects. This approach enables direct and reliable detection of metastable oxygen-redox intermediates in this system, while also capturing diverse oxygen-redox features across different cathodes and stoichiometries. Our findings demonstrate a robust method to identify redox intermediates, reconciling experimental observations and theoretical predictions and sharpening our understanding of oxygen-redox mechanisms.\u003c/p\u003e","manuscriptTitle":"Vacuum-free spectroscopy reveals metastable oxygen redox intermediates in oxygen-redox cathodes","msid":"","msnumber":"","nonDraftVersions":[{"code":1,"date":"2026-02-10 05:46:13","doi":"10.21203/rs.3.rs-8285919/v1","editorialEvents":[],"status":"published","journal":{"display":true,"email":"[email protected]","identity":"nature-energy","isNatureJournal":true,"hasQc":false,"allowDirectSubmit":false,"externalIdentity":"nenergy","sideBox":"Learn more about [Nature Energy](http://www.nature.com/nenergy/)","snPcode":"","submissionUrl":"","title":"Nature Energy","twitterHandle":"","acdcEnabled":true,"dfaEnabled":true,"editorialSystem":"ejp","reportingPortfolio":"Nature Research","inReviewEnabled":true,"inReviewRevisionsEnabled":false}}],"origin":"","ownerIdentity":"c9c5f520-8152-47b4-8f0f-1385f6a811dc","owner":[],"postedDate":"February 10th, 2026","published":true,"recentEditorialEvents":[],"rejectedJournal":[],"revision":"","amendment":"","status":"under-review","subjectAreas":[{"id":59858736,"name":"Physical sciences/Materials science/Materials for energy and catalysis/Batteries"},{"id":59858737,"name":"Physical sciences/Materials science/Techniques and instrumentation/Characterization and analytical techniques"},{"id":59858738,"name":"Physical sciences/Chemistry/Analytical chemistry"},{"id":59858739,"name":"Physical sciences/Chemistry/Electrochemistry/Batteries"}],"tags":[],"updatedAt":"2026-02-10T05:46:14+00:00","versionOfRecord":[],"versionCreatedAt":"2026-02-10 05:46:13","video":"","vorDoi":"","vorDoiUrl":"","workflowStages":[]},"version":"v1","identity":"rs-8285919","journalConfig":"researchsquare"},"__N_SSP":true},"page":"/article/[identity]/[[...version]]","query":{"redirect":"/article/rs-8285919","identity":"rs-8285919","version":["v1"]},"buildId":"XKTyCvWXoU3ODBz1xrDgd","isFallback":false,"isExperimentalCompile":false,"dynamicIds":[84888],"gssp":true,"scriptLoader":[]}