Pulsed-laser-deposited LiMn2O4 thin film solid-state microbatteries with extended voltage window cycling

Pulsed-laser-deposited LiMn2O4 thin film solid-state microbatteries with extended voltage window cycling Juan Carlos Gonzalez-Rosillo*†1, Jędrzej Morzy*2, Yaroslav E. Romanyuk2, Moritz H. Futscher2, Albert Tarancón†1,3, Alex Morata†1. 1. Department of Advanced Materials for Energy Applications, Catalonia Institute for Energy Research (IREC), Jardins de les Dones de Negre 1, 08930, Sant Adrià del Besòs, Barcelona, Spain 2. Laboratory for Thin Films and Photovoltaics, Empa – Swiss Federal Laboratories for Materials Science and Technology, Überlandstrasse 129, 8600 Dübendorf, Switzerland. 3. Catalan Institution for Research and Advanced Studies (ICREA), Passeig Lluís Companys 23, 08010, Barcelona, Spain *: shared first co-authorship †: corresponding authors Abstract Thin film microbatteries provide on-chip and surface-mount energy storage for Si-based microsystems, where device area is the primary constraint. Commercial implementations, available for more than 20 years, have largely relied on LiCoO2 cathodes because they are straightforward to process and package. LiMn2O4 oers a cobalt-free alternative, but in conventional liquid-electrolyte Li-ion cells its use is constrained by Mn dissolution and capacity fade, especially when the voltage window is widened to access its theoretical capacity of ~119 µAh·cm-2·µm-1 (~296 mAh·g-1). Thin-film solid-state architectures can mitigate these limitations and are naturally aligned with footprint-limited applications, where areal capacity and areal energy are the relevant figures of merit. The focus of this study is to examine the device behaviour of LiMn2O4 thin film microbatteries operated in a wider voltage window, using a LiPON solid electrolyte and a Li metal anode. Polycrystalline ~850 nm LiMn2O4 cathodes were grown by Pulsed Laser Deposition with sequential Li2O enrichment during growth. X-ray diraction, Raman features, and depth-profiling glow discharge optical emission spectroscopy are consistent with the presence of a Li-rich spinel component formed during deposition. The resulting LiMn2O4/LiPON/Li cells, cycled between 2.0 and 4.5 V, deliver up to ~50 µAh·cm-2 at low rates; at higher rates, the wider window enables capacities up to ~4 times those obtained on the same devices in the conventional 3.5–4.5 V window. Impedance measurements are used to track evolution during conditioning and operation. Finally, we provide an overview of relevant LiMn₂O₄ solid-state thinfilm microbatteries and outline a tentative route to stabilize the LiMn2O4/LiPON interface under widerwindow operation. Introduction Miniaturized IoT, wearable, and implantable systems require energy storage that fits a fixed footprint, integrates on-chip with microelectronics, and remains compatible with standard workflows within the semiconductor industry.1 Thin-film microbatteries meet these requirements using solid electrolytes and micrometre-scale cathodes.2,3 They can be fabricated directly on Si chips or supplied as components for conventional assembly. The technology has been commercially available for more than two decades, predominantly with LiPON electrolytes and LiCoO2 cathodes. The device performance in this area-limited context is most meaningfully expressed as areal capacity (µAh·cm-2), areal energy (µWh·cm-2), and areal power density (mW·cm-2). With appropriate orientation control, LiCoO2 films have even been demonstrated to reach ≈1.2 mAh·cm-2 at ~30 µm thickness in solid-state cells.4 Within this context, LiMn2O4 poses an attractive, cobalt-free alternative to LiCoO2 . First reported in the 1980’s, LiMn2O4 combines a relatively high operating voltage (4.2V) with a spinel framework that supports three-dimensional Li diusion with the use of abundant manganese.5 In the conventional voltage range (3.5–4.5 V vs. Li/Li+), LiMn2O4 has a theoretical capacity of ~59 µAh·cm-2·µm-1 (148 mAh·g1), though practical capacities of approximately ~48 µAh·cm-2·µm-1 (120 mAh·g-1) are typically achieved.5 Extending the cycling range to 2.0–4.5 V allows access to an additional plateau near 3 V, corresponding to the formation of the Li2Mn2O4 phase. This extended range eectively doubles the theoretical capacity to ~119 µAh·cm-2·µm-1 (296 mAh·g⁻¹), enabling significantly higher energy density than, for instance, LiCoO2 (~138 mAh·g-1, 69 µAh·cm-2·µm-1). However, this comes at the cost of severe manganese dissolution (in liquid electrolytes) due to the formation of Mn2+, 5–8 as well as substantial volume changes, and particle cracking associated with the lithiation process.9 Only in very thin films (< 200 nm, low areal capacity), several studies with liquid electrolytes have shown that a wider voltage window can be accessed for few tens of cycles approaching the theoretical capacity of the Li2Mn2O4 phase.10–12 In particular, Zheng et al. reported that strain/orientation control can stabilize Li₁₊ₓMn₂O₄ and improve utilization; however, such approaches typically rely on epitaxial or single-crystal growth and are diicult to translate to scalable polycrystalline processes.13 A representative survey of solid-state LiMn2O4 thin film batteries spanning over 25 years is summarized in Table S1 (Supporting Information, Section I), including fabrication methods, cycling conditions and voltage range, film thicknesses, and achieved areal capacities. Early studies using e-beam evaporation and RF sputtering followed by high-temperature annealing established crystalline LiMn2O4 operation but frequently reported limited areal capacity in the conventional narrow voltage window, especially for films > 1 µm where diusion constraints dominate.14–17 Amorphous counterparts generally showed lower energy storage capabilities,14,18,19 while 3D architectures20,21 and composite cathodes22 have shown promising results in improved cycling stability and rate capability. Despite these advancements, the near-full capacity utilization reported in early studies in the field23 has proved diicult to be reproduced broadly, and even recent reports indicate that LiMn2O4/LiPON interface can evolve under cycling even within the narrower voltage window.24 Because thickness, porosity, current density, and voltage range dier across studies, device-to-device comparisons should be made cautiously. However, a consistent trend emerges from our survey: increasing the thickness above 500 nm of planar, polycrystalline LiMn2O4 while maintaining high cathode utilization is challenging, whereas exploration of wider-window operation in solid-state cells remains comparatively understudied and may oer a practical lever to improve areal capacity at the device level. In this work, we examine Pulsed Laser Deposition (PLD) with a multilayer protocol that introduces sequential Li₂O enrichment during growth to produce ~850 nm, polycrystalline LiMn2O4 cathodes for operation over a wider voltage window, in solid-state thin-film batteries. X-ray diraction, Raman features, and Glow Discharge Optical Emission Spectroscopy depth profiles are consistent with the partial formation of a Li-rich spinel Li2Mn2O4 component during deposition. In our LiMn2O4/LiPON/Li cells, the wider voltage window cycling shows promising stability and provides capacities of 30 - 50 µAh·cm-2 at current densities of 1-30 µA·cm-2 corresponding to an approximately two-fold increase in the capacity compared to the narrower voltage window. Methods Thin film fabrication: Commercially available ceramic targets of LiMn2O4 and Li2O (CODEX) were employed for the Large-Area Pulsed Laser Deposition (LA PLD-5000, PVD Products) system, equipped with a Coherent (Lambda Physik) COMPex Pro 205 KrF excimer laser (λ = 248 nm). Both targets were sequentially ablated following a multilayer deposition approach, similar to our previous studies.12,25–29 For this work, the deposition parameters were adjusted to enhance lithiation and optimize processing time. Specifically, a 4:3 ratio of ablation pulses was maintained between the LiMn2O4 and Li2O targets, with 2600 and 2000 pulses, respectively, and a total of 60 repetition cycles. The deposition was conducted at a temperature of 650 °C and a partial oxygen pressure of 20 mTorr, with a target-substrate distance of 90 mm. The laser was operated at a fluence of 1.3 J·cm-2 and a frequency of 20 Hz to reduce total deposition time while minimizing excessive thermal exposure. The substrates used were 1 x 1 cm2 Pt (80 nm)/Ti (10 nm)/Si3N4 (300 nm)/SiO2 (100 nm)/Si (0.5 mm) chips, fabricated at the Institute of Microelectronics of Barcelona (IMB-CNM-CSIC). These multilayer depositions resulted in films with a total thickness of approximately 850 nm, as determined by Spectroscopic Ellipsometry and crosssectional SEM imaging. This increased thickness and optimized deposition protocol were designed to achieve higher lithiation levels by incorporating additional lithium during the growth process. Lithium–phosphorus oxynitride (LiPON) solid-electrolyte of 1 µm in thickness was manufactured by RF magnetron co-sputtering of 2″ targets of Li3PO4 (99.95%, Kurt J Lesker Co., rate approximately 0.7 nm min–1) and Li2O (99.9%, Toshima Manufacturing, rate approximately 0.6 nm min–1) in a N2 atmosphere (flow set to 50 SCCM) at powers of 100 and 120 W, respectively, and a working pressure of 4 × 10–3 mbar. The target-to-substrate distance was set to 25 cm. The Li-metal anode for the solid-state configuration was thermally evaporated from lithium pellet (99+%, Thermo Fisher Scientific, Nexdep evaporator) at a rate of 25 Å s–1, to a total thickness of 2 µm. The Li deposition was patterned with circular dots of 0.0079 cm2 area as individual Li metal reservoirs that defined the separate cells. Structural characterization: Scanning electron microscopy (SEM) was conducted at IREC using a Zeiss Auriga instrument equipped with a 30 kV Gemini FESEM column and an in-lens detector, enabling highresolution imaging of the film’s cross-sectional and surface morphology. X-ray diraction (XRD) measurements were performed using a Bruker D8 Advance diractometer in the θ–2θ configuration, covering a range of 10–60° with a step size of 0.01°. To protect the detector, the region around the main reflection of the platinum substrate was excluded from the measurements. Glow Discharge Optical Emission Spectroscopy (GDOES) was performed using a Spectruma GDA 750 HR instrument equipped with a 2.5 mm anode. Measurements were conducted under optimized conditions of 700 V and 2.5 hPa Ar pressure to ensure a flat and uniform crater for accurate depth profiling. Raman spectroscopy was conducted using an Xplora Nano spectrometer (HORIBA) with 532 nm (green) and 633 nm (red) lasers. Spectra were acquired with an acquisition time of 2 seconds and 60 accumulations. Laser powers of 0.9 and 9 mW were used for the green laser, while 0.4 and 4 mW were used for the red laser, ensuring measurements below the threshold for sample degradation. Electrochemical characterization: For liquid electrolyte measurements, the films were characterized using a TSC Surface Cell (rhd instruments, Germany). The electrolyte consisted of 1 M LiPF6 in a 50/50 (v/v) mixture of ethylene carbonate (EC) and dimethyl carbonate (DMC), both of battery-grade quality (Sigma-Aldrich). Lithium metal served as both the reference and counter electrode, while the active area of the film was defined by the cell (and empirically measured under optical microscope) as 0.3 cm2. The mass of the cathode was estimated based on thickness measurements obtained via Spectroscopic Ellipsometry. The film consisted of a dense 730 nm layer and a porous 117 nm layer (50% porosity), corresponding to an equivalent LiMn₂O₄ loading of 0.27 mg·cm-2, estimated using the theoretical density of LiMn2O4 (4.02 g·cm-3). For the solid-state configuration measurements, the liquid cell capacities were used as a starting point for C-rate determination. Therefore, 1C rate was defined as 30 µA cm-2 regardless of the voltage window. Cycling and EIS measurements for the solid-state setups were performed on Squidstat Plus potentiostats. Briefly, EIS was done in potentiostatic mode with 50 mV excitation amplitude, 10 steps per decade, at 3.5 V. Frequency range was set to 1 MHz – 0.01 Hz. Results and discussion Structural characterization Cross-sectional SEM images (Fig. 1a.) of PLD-deposited 850 nm-thick LiMn2O4 films on top of Pt-coated silicon substrate reveal a rough surface morphology, characteristic of the multi-layer deposition process used for their fabrication.25,26,28,29 Using spectroscopic ellipsometry,27,28,30 the total thickness of the films was measured to be 847 nm, consisting of a dense 730 nm layer and a porous 117 nm top layer. For the solid-state thin-film batteries, LiPON was conformally deposited on the LiMn2O4 films (RF sputtering, see details elsewhere31,32 and in the Methods section). This conformal coverage (Fig. 1b) indicates that the surface roughness remains within acceptable limits for device fabrication. The final solid-state batteries were completed with an evaporated layer of Li metal (~2 µm) as the anode (not shown). XRD (Fig. 1c) confirms polycrystalline LiMn2O4 as the dominant phase. In addition to the expected reflections, extra peaks appear at 2θ ≈ 18º and 37º, consistent with the (101) and (211) planes reported for tetragonal Li2Mn2O4. These features suggest the coexistence of the Li-rich spinel within the film. We note that this response coincides with adjustments to the PLD fabrication sequence (higher repetition rate and increased pulses per layer) implemented to shorten deposition time and increase Li supply during growth (see Methods). The modified sequence enables thicker films (~850 nm) while preserving comparable surface roughness to our prior multilayer reports while shortening fabrication time.25,26,28,29 Minor Mn3O4 signatures are also observed (log-scale intensity), indicating limited surface reconstruction under the present conditions. Glow Discharge Optical Emission Spectroscopy (GDOES) was used to probe the compositional depth profiles of the films (Fig. 1d). Mn and O profiles remain relatively uniform throughout the film, while the Li signal shows a relative enrichment within the first ~150 nm from the surface and the last ~250 nm near the substrate. This means that ~40–50% of the total thickness exhibits a higher Li concentration than the middle region of the film. Given the semi-quantitative nature of the presented GDOES measurements, we interpret these trends as relative enrichment zones rather than absolute compositions. Taken together, the XRD and GDOES signatures support a non-uniform distribution of a Li-rich spinel component, with higher prevalence near the film boundaries. Additionally, carbon signal detected at the very surface (within the first 15–20 nm) likely corresponds to a thin surface carbonate layer formed during brief air exposure prior to the GDOES measurement. Figure 1: a) Cross-sectional SEM image of the as-deposited LiMn2O4 film.b) Cross-sectional image of a LiMn2O4/LiPON/Pt film. Please note that the top Pt was deposited for ease of SEM visualization and LiPON protection. c) XRD patterns of the asdeposited film and the Pt-coated chip as reference. The bottom pannel shows references for LiMn2O4 (JCPDS 00-035-0782), Li2Mn2O4 (JCPDS 00-038-0299) and Mn3O4 (JCPSD 01-080-0382). d) GDOES depth profile of the as-deposited film. Raman spectroscopy provides complementary evidence for a Li-rich spinel component and helps correlate compositional variations with depth. Spectra were collected with green (532 nm) and red (633 nm) lasers. The complementary use of the red laser enhances the detection of secondary Mn3O4 phases, a common feature usually occurring in LiMn2O4.29 The Raman spectra in Fig. 2 are dominated by the LiMn2O4 bands (fittings and peak assignments are provided in the Supporting Information),29,33 and show a feature at ~409 cm-1 that has been assigned to Li2Mn2O4 in literature.33 This band was not observed in earlier films grown by our standard sequence and is consistent with the increased lithiation associated with the modified PLD protocol. The Li2Mn2O4-related band is present at both low and higher laser powers (more prominent with 532 nm, see Supporting Information, Section II), suggesting that the Li-rich spinel extends to the near-surface region and is also present near the substrate, in agreement with the observed GDOES enrichment zones. Beyond the 409 cm-1 signature, we observe powerand wavelength-dependent changes in the Mn–O vibrational region (~600–650 cm-1). Because 532 and 633 nm excitations probe slightly dierent eective depths and resonance conditions, the altered band ratios are consistent with a depth-varying convolution of LiMn2O4and Li-rich spinel modes, i.e., lithium distribution inhomogeneities across the film thickness. Figure 2: Raman spectra acquired with a) green laser (532 nm) and b) red laser (633 nm), acquired at 0.9 and 0.4 mW powers, respectively. Electrochemical characterization To evaluate the electrochemical performance of the PLD LiMn2O4 films across the wider voltage range, both liquid electrolyte (1 M LiPF6 in 50:50 v/v EC:DMC, Li metal) and solid-state (LiPON/Li metal) configurations were tested. Figure 3a shows the results of cycling in organic liquid electrolyte within the narrower voltage range (3.5 – 4.5 V), which revealed the characteristic redox plateaus of LiMn2O4. Notably, the initial open-circuit voltage (OCV) of the film was ~3 V, closer to the expected OCV of Li2Mn2O4 rather than LiMn2O4 (~3.5 V). Additionally, the first charge (shown in purple) displayed evidence of a lithium extraction process near 3 V, consistent with the presence of the Li2Mn2O4-like component. Importantly, RF-sputtered LiPON can partially lithiate the underlying oxide cathode, slightly increasing the Li extracted on the first charge in the solid-state cells relative to the liquid-electrolyte one. Unless stated otherwise, capacities are reported as true areal values; derived gravimetric estimations are based on the measured loading, given in the Methods. In organic electrolyte at low current densities within the narrower window (3.5-4.5 V), the films achieved gravimetric capacities close to the practical LiMn2O4 values reported, reaching device areal capacities of ~27 and ~14 µAh·cm-2 (107 and 59 mAh·g-1) at 6.8 and 160 µA·cm-2, respectively (Fig. 3d). Cycling was stable under these conditions. However, extending cycling to the wider voltage window introduces the well-known Mn dissolution limitations (see Supporting Information, section III), typically only mitigated in very thin films (<200 nm thick).28 Electrochemical impedance spectroscopy (EIS) shows an initial total impedance of ~11 kΩ·cm-2, decreasing to ~600 Ω·cm-2 after the cycling protocol between 3.5-4.5 V (Fig. 3g). This drop is consistent with an interface improvement, such as removal of resistive surface species). One plausible contributor to this behavior is the dissolution of the surface carbonate layer, whose presence is suggested by the transient carbon signal observed in GDOES. Alternatively, recent work by Ou et al.34 suggests that LiMn2O4 cathodes in carbonate-based electrolytes form a dynamic cathode–electrolyte interphase, whose continuous restructuring can improve charge transfer. While this improvement suggests a reduction in film resistance during cycling, the issues at lower voltages exemplify the limitations of widening the useful cycling range of LiMn2O4 films in conventional liquid cells. Figure 3: a, b, c) Galvanostatic cycling of LiMn2O4 in liquid organic electrolyte (narrower window), and in solid state cells of LiMn2O4/LiPON/Li (narrower and wider windows), respectively. d, e, f) Extracted discharge capacities for the corresponding galvanostatic cycling. g, h, i) EIS before and after cycling for each of the cells. Insets show the full EIS spectra for the liquid electrolyte cell and DRT analysis of the solid-state cells, respectively. Capacities are true areal values. EIS spectra were acquired at ~4.0 V after the indicated protocols to enable comparison. The electrochemical performance of the LiMn2O4 films in solid-state configuration was evaluated within both the narrower and wider voltage ranges as shown (Fig. 3b-c). In both cases, the precycle (in purple) exhibits a ~3 V plateau, with the initial OCV also close to ~3V. Although this behavior is typical of cathodes coated with a LiPON layer, which has been reported to lithiate the cathode underneath during deposition,35 in our case, the structural and electrochemical evidence from the liquid electrolyte cells support prelithiation during the PLD process itself. This prelithiation likely contributes to the Li-rich spinel footprints observed and to subsequent performance upon cycling. As seen in Fig. 3e, LiMn2O4 films show stable cycling in the narrower range (3.5 -4.5 V). At low current densities, areal capacities are comparable to those measured in liquid electrolyte of ~27 µAh·cm-2. However, the C-rate capability is worse for the solid-state configuration, highlighting diusion-limited behavior for our ~850 nm planar films. In the liquid case, there is likely electrolyte infiltration into porosity within the film, which aids in the ionic transport, accounting for the dierences between the configurations. This emphasizes the challenges associated with Li+ diusion in thicker cathode films (as opposed to LiCoO2), where long diusion paths can limit the cycling kinetics for LiMn2O4 without any compositional tuning that could increase its ionic diusion in the bulk. When the voltage range is widened to lower cut-o voltage of 2.5 V (Fig. 3c, f), several eects appear. At very low current densities, the films achieved areal capacities of ~30-50 µAh·cm-2, markedly higher than for the narrower voltage window. Interestingly, we observe an increase in capacity over the initial cycles, suggesting a progressive stabilization of the cathode–solid electrolyte interphase during early cycling or beneficial restructuring of the cathode layer. Similar behavior has been reported in micron-sized Aldoped LiMn₂O₄ samples cycled at low voltages, although those exhibited lower capacities overall.9 This interface improvement is corroborated by EIS: in the narrower window the total resistance shows no substantial change (Fig. 3h), whereas in the wider window it decreases from ~1200 Ω·cm-2 to ~250 Ω·cm2 after cycling (Fig. 3i), consistent with an improved transport within the electrode and at the interface. To further examine these changes, Distribution of Relaxation Times (DRT) analysis was performed in the solid-state configurations (insets, Fig. 3h–i). Four processes were tentatively identified:36,37 P1, likely associated with bulk LiPON resistance; P2, related to the cathode–electrolyte interface; P3, linked to the LiPON/Li metal interface; and P4, corresponding to charge transfer and ionic diusion within the LiMn2O4 layer. In the narrower window, P2 showed a moderate decrease, while P4 increased, suggesting modest improvement at the CEI but some degradation in the cathode transport. In contrast, in the wider window P2 was strongly reduced and P4 also decreased, consistent with a more conductive interface and enhanced transport within the cathode. While these assignments are consistent with prior studies,36,37 the physical origin of the relaxation processes arising from the DRT analysis in thin-film solid-state batteries remains under discussion, as charge-transfer phenomena may involve both interfacial and bulk contributions in such compact geometries. Long-term cycling and detailed interfacial characterization will be reported separately. These findings highlight the importance of interfacial stability in solid-state battery operation. Xia et al. reported LiMn2O4/LiPON interfaces that evolve into an inactive Mn3O4/Li2O phase during cycling, accompanied by Mn diusion into the LiPON.24 In contrast, our films exhibit a substantial impedance decrease and improved capacity in the wider window. Overlithiation during PLD, leading to a partial Li₂Mn₂O₄-like component and the presence of cationic defects in our films may help mitigate such degradation. By accommodating volume changes and reducing interfacial stress, this phase could help preserve contact and transport, and suggests that cathode pre-engineering could be a suitable approach toward robust, high-performance solid-state LiMn2O4 systems. As summarized in Fig. 4a, at very low current densities (<10 µA/cm2) and within the narrower window, our films approach in both liquid and solid-state configurations equivalent gravimetric capacities close to the practical values of LiMn2O4 (~120 mAh·g-1). The rate dependence diers, with solid-state capacities declining more rapidly due to transport limitations in ~850 nm planar films. Liquid cells can partly oset this via electrolyte access to porosity. Extending cycling to include the ~3 V plateau in the wider window compensates for these limitations, yielding significantly higher areal (and derived gravimetric) capacities. In this wider window, films reach >50 µAh·cm-2 at low current densities. Even at ~(~50–100 µA·cm-2, capacities approach the practical limits of the narrower-window counterparts. While some earlier reports show ≥80 µAh·cm-2, reproducibility has been challenging, likely reflecting stoichiometry control issues in RF sputtering.23,38 Our results highlight a key advantage of the solid-state configuration in the wider window: the absence of Mn dissolution that limits liquid cells. While the exact mechanisms remain under study, we suggest that the partial Li2Mn2O4 formation during fabrication may help accommodate volume changes during deep (de)lithiation, contributing to mechanical stability and preserving interfacial contact. Additionally, cation vacancies, as proposed in our earlier works28,29, could aid strain accommodation. Together, these features may enable a more robust and conductive interface for solid-state LiMn2O4-based thin film cathodes operating in the wider voltage window. To contextualize these results, Fig. 4b compares our data with representative reports over the past two decades of Li-Mn-O thin film device integration, based on the literature summarized in Table 1. The plot shows that our films are competitive within device-class thickness even against thicker films or 3D architectures. These observations suggest that incorporating a Li-rich spinel component during fabrication can be a practical route to stabilize polycrystalline LiMn2O4 films for wider window operation. This approach can be further optimized to enable higher energy and power density thin-film LiMn2O4 for more sustainable microbatteries. Figure 4: a) Summary of capacities as a function of current density for the three configurations tested. b) Comparison of the results of this work with the literature on Li-Mn-O systems integrated into solid-state thin film batteries. Conclusions This work demonstrates the successful integration of polycrystalline LiMn2O4 thin films with LiPON solid electrolytes using PLD. Through a multilayer deposition approach, partial incorporation of a Li-rich spinel (Li2Mn2O4-like) component was achieved during film growth. Structural, electrochemical data and comparison towards literature suggest that this phase might contribute to stable cycling in the wider voltage window. As a result, areal capacities of ~30-50 µAh·cm-2 were obtained. Wider-window cycling also coincided with a progressive decrease in interfacial and bulk resistances (EIS/DRT), in contrast to liquid organic electrolytes wherein Mn dissolution limits extension to lower voltages. The interface appears to improve during early cycling in the wider window, potentially benefiting from the Li-rich spinel and the films’ cation-vacancy landscape, which may help accommodate volume change. Reduced impedance accompanies a gradual capacity increase over initial cycles. Despite these advances, diusion-limited kinetics remain a challenge for thick-film architectures, especially at higher current densities. Future work should aim to unravel the mechanisms underlying interface rearrangement during long-term cycling and explore ways to further optimize ion transport within device-class thickness. Conflicts of interest The authors declare no conflict of interest. Data availability Data for this article, including TERS maps and simulated profiles will be available at IREC’s institutional repository, hosted in Zenodo at https://zenodo.org/communities/irec/ Acknowledgements The SPINELKING project (2025 prod 00005) has the support of the Department of Research and Universities of the Generalitat de Catalunya. This project (PCI2022-132960) has received funding from MCIN/AEI/10.13039/501100011033 and the European Union “NextGenerationEU”/PRTR” (AfreeSSB project) and the “Generalitat de Catalunya” (2021 SGR 00750, NANOEN, and 2021 SGR 01286). J.C.G.-