# Fileset

[Advanced Science - 2025 - Dutta - Interconnected Hierarchically Porous Graphene‐Based Membrane Electrode for High‐Power and.pdf](https://mdr.nims.go.jp/filesets/df36433a-e235-434a-820e-d8afb23d4540/download)

## Creator

[Arghya Dutta](https://orcid.org/0000-0002-3769-7820), [Takashi Kameda](https://orcid.org/0000-0003-2080-3540), [Taiga Ozawa](https://orcid.org/0009-0004-5608-8520), Anna Myojin, Minako Nishioka, Wei Yu, Hirotomo Nishihara, [Shoichi Matsuda](https://orcid.org/0000-0002-0640-3404)

## Rights

[Creative Commons BY Attribution 4.0 International](https://creativecommons.org/licenses/by/4.0/)

## Other metadata

[Interconnected Hierarchically Porous Graphene‐Based Membrane Electrode for High‐Power and Long‐Cycle Lithium–Oxygen Battery](https://mdr.nims.go.jp/datasets/6eaf1008-8f40-4f63-9a6e-4bf69ced4dcb)

## Fulltext

Interconnected Hierarchically Porous Graphene‐Based Membrane Electrode for High‐Power and Long‐Cycle Lithium–Oxygen BatteryRESEARCH ARTICLEwww.advancedscience.comInterconnected Hierarchically Porous Graphene-BasedMembrane Electrode for High-Power and Long-CycleLithium–Oxygen BatteryArghya Dutta,* Takashi Kameda, Taiga Ozawa, Anna Myojin, Minako Nishioka, Wei Yu,Hirotomo Nishihara, and Shoichi Matsuda*The energy–power trade-off in lithium–oxygen batteries (LOBs) arises fromsluggish oxygen (O2) transport in the porous positive electrode and poreclogging by lithium peroxide (Li2O2). While increasing porosity enhanceselectrolyte accessibility and Li2O2 storage, it also increases electrolytedemand, compromising the overall energy density of the cell andnecessitating alternative strategies to boost power capabilities withoutsacrificing energy density. In this study, theoretical simulations of O2transport reveal that reducing tortuosity by improving pore interconnectivityhas a more significant impact on O2 transport than porosity itself. Based onthis insight, a freestanding graphene-based electrode with a highlyinterconnected macroporous network is fabricated via a non-solvent-inducedphase separation approach using polyacrylonitrile (PAN) as a carbon scaffoldand polyethylene oxide (PEO) as a sacrificial porogen. The selectivedecomposition of PEO creates spatially interconnected macropores,effectively reducing tortuosity. The resulting electrode enables LOB cells toachieve >2500 mAh g−1 at 1.0 mA cm−2 under lean-electrolyte conditions.Stable cycling at 4 mAh cm−2 is maintained with only 3.25 g Ah−1 electrolyte,and high-rate performance persists over 90 cycles at 1.5 mA cm−2. This workdemonstrates a robust strategy to simultaneously improve energy and powerperformance in practical LOBs through rational electrode architecture.A. Dutta, T. Kameda, T. Ozawa, A. Myojin, M. Nishioka, S. MatsudaCenter for Green Research on Energy and Environmental MaterialsNational Institute for Materials Science1-1 Namiki, Tsukuba, Ibaraki 305-0044, JapanE-mail: DUTTA.Arghya@nims.go.jp;MATSUDA.Shoichi@nims.go.jpW. Yu,H.NishiharaAdvanced Institute forMaterials Research (WPI-AIMR)TohokuUniversitySendai,Miyagi 980-8577, JapanS.MatsudaCenter for AdvancedBattery CollaborationNational Institute forMaterials Science1-1Namiki, Tsukuba, Ibaraki 305-0044, JapanThe ORCID identification number(s) for the author(s) of this articlecan be found under https://doi.org/10.1002/advs.202519091© 2025 The Author(s). Advanced Science published by Wiley-VCHGmbH. This is an open access article under the terms of the CreativeCommons Attribution License, which permits use, distribution andreproduction in any medium, provided the original work is properly cited.DOI: 10.1002/advs.2025190911. IntroductionLithium–oxygen batteries (LOBs) havegarnered significant interest as a next-generation energy storage system due totheir exceptionally high theoretical gravi-metric energy density (≈3500 Wh kg−1),far surpassing conventional lithium-ionbatteries (LIBs).[1,2] However, despite theirpromise, the practical realization of LOBsis hindered by several critical challenges,including lower-than-expected specificcapacity, poor power performance, andlimited cycle life.[3–7] These challengesprimarily stem from the fundamentalelectrochemical processes and the struc-tural limitations of LOB components. Atypical LOB consists of a lithium (Li) metalnegative electrode, a non-aqueous elec-trolyte, and a porous carbon-based positiveelectrode. During discharge, lithium ions(Li+) react with oxygen (O2) to form solidlithium peroxide (Li2O2), which is subse-quently decomposed during the chargecycle (2Li+ + O2 + 2e− ↔ Li2O2). However,the accumulation of insoluble Li2O2 withinthe porous carbon obstructs ion and O2transport, leading to electrode passivation, lower-than-expectedcapacity, large overpotential, poor rate capability, and short cy-cle life.[8–10] Therefore, designing an optimized carbon struc-ture that facilitates effective Li2O2 deposition and decomposi-tion while maintaining open pathways for ion and gas diffu-sion is crucial for enhancing LOB performance. Extensive in-vestigations have highlighted the critical role of carbon elec-trode architecture in governing the discharge behavior and over-all performance of LOBs.[11–19] A consensus has emerged thatan ideal air electrode must consist of three structural features:i) a large surface area to maximize electrochemical reactionsites, ii) sufficiently wide pores to facilitate efficient O2 andLi+-ion transport while preventing pore clogging, and iii) ahigh pore volume to accommodate the substantial formation ofsolid Li2O2 discharge products without severely compromisingperformance.While considerable progress has been made in tuning theporosity of carbon electrodes for LOBs, several critical aspects,such as the spatial interconnectivity of pores, the packing of car-Adv. Sci. 2026, 13, e19091 e19091 (1 of 10) © 2025 The Author(s). Advanced Science published by Wiley-VCH GmbHhttp://www.advancedscience.commailto:DUTTA.Arghya@nims.go.jpmailto:MATSUDA.Shoichi@nims.go.jphttps://doi.org/10.1002/advs.202519091http://creativecommons.org/licenses/by/4.0/http://creativecommons.org/licenses/by/4.0/http://crossmark.crossref.org/dialog/?doi=10.1002%2Fadvs.202519091&domain=pdf&date_stamp=2025-12-01www.advancedsciencenews.com www.advancedscience.combon particles, and their implications for energy and power den-sities, have remained largely underexplored. Traditional designstrategies focus primarily on increasing electrode porosity to en-hance O2 transport and boost discharge capacity.[14–16] However,highly porous electrodes require substantial volumes of elec-trolyte to ensure adequate wetting.[20,21] This increased electrolytedemand significantly compromises the cell-level energy density,posing a major challenge to the practical deployment of LOBs.Despite its importance, the role of electrolyte volume has re-ceived limited attention in the field. Most reported systems stilloperate with excess electrolyte (>50 μL cm−2) and low areal ca-pacities (<1 mAh cm−2), resulting in cell-level energy densitiesthat fall short of those achieved in commercial LIBs.[22,23] Fur-thermore, conventional porous electrodes are typically fabricatedusing polymeric binders, which promote dense particle aggre-gation. This compact microstructure limits O2 diffusivity and re-duces the accessible pore volume for Li2O2 growth, leading to lowactive material utilization, limited capacity, and poor rate perfor-mance. In addition, the presence of binders can introduce par-asitic reactions that further deteriorate battery stability.[24,25] Toovercome these limitations, the development of binder-free, self-supporting carbon architectures with engineered porosity andimproved pore interconnectivity (low tortuosity) is highly desir-able. Such designs aim to simultaneously enhance O2 transportpathways and facilitate more efficient Li2O2 formation, therebyimproving both energy and power performance in practical LOBsystems.Beyond structural optimization, the surface chemistry of car-bon electrodes plays a crucial role in determining LOB stabil-ity. Highly graphitized carbons with minimal surface functionalgroups exhibit superior long-term cycling performance by re-ducing side reactions with reactive oxygen species.[23,26] Conse-quently, significant research efforts have been directed towardenhancing the degree of graphitization and crystallinity of carbonmaterials. Our recent works introduced edge-site-free grapheneelectrodes with hierarchical porosity, achieving high capacity andextended cycle life.[27,28] The mesoporous framework was createdusing aluminum oxide (Al2O3) nanoparticles as hard templates,while macropores formed between spherical carbon aggregates.However, the absence of a well-interconnected long-rangemacro-porous network and unoptimized porosity resulted in limitedrate capability and excessive electrolyte dependence to maintainlong cycling stability.Based on these considerations, at first, we theoretically sim-ulate the effects of porosity and the pore-connectivity (tortuos-ity) on O2 diffusion through the porous positive electrode. Then,we focus on developing spatially interconnected hierarchicallyporous self-standing graphene mesosponge (GMS) membranesoptimized for high capacity, superior rate performance, and sta-ble cycling under lean electrolyte conditions. The highly intercon-nected meso-macroporous network provides efficient O2 and iondiffusion pathways while offering sufficient space for Li2O2 de-position. The electrode fabrication process involved three steps:First, a porous GMS carbon was synthesized, enabling modu-lation of structural properties for improved electrochemical sta-bility and porosity in different dimensions, from micropore tomesopore. Second, self-standing membranes were fabricated us-ing the doctor blade technique with a carbon-based slurry con-taining polyacrylonitrile (PAN) and polyethylene oxide (PEO).The non-solvent-induced phase separation (NIPS) method intro-duced amacroporous network into the polymermatrix. The thirdstep employed the carbonization of the hierarchically porousmembrane to convert the polymer matrix into macroporous car-bon. The differences in the thermal stabilities of PAN and PEOwere leveraged to regulate themacroporous network. During car-bonization, the PAN matrix transformed into an isolated macro-porous carbon scaffold, whereas PEO decomposed into low-molecular-weight non-solid products, opening up the porous net-work and producing an interconnectedmacropore structure. Theformation of this interconnected macroporous network is crucialto achieving high-power operation of an LOB. This hierarchicaldesign balances meso- and macroporosity, optimizing electrodecapacity and electrolyte requirements for sufficient wetting underlean electrolyte conditions.2. Results and Discussions2.1. Simulation of Quantitative Impacts of Porosity and PoreConnectivity (Tortuosity) on the Oxygen DiffusionDuring the discharge process of LOBs, molecular O2 dissolvedin the liquid electrolyte diffuses into the porous positive elec-trode, where it undergoes electrochemical reduction. This reduc-tion process is oftenmodeled as a first-order reactionwith respectto the local O2 concentration, and it is strongly influenced by thebalance between O2 supply via diffusion and its electrochemi-cal consumption.[29] Under constant current (galvanostatic) op-eration, the applied current density dictates the rate at which O2must be reduced, which in turn requires a steady diffusive fluxof O2 across the porous structure of the electrode. If this flux isinsufficient to meet the electrochemical demand, O2 concentra-tion gradients develop, ultimately leading to mass transport lim-itations. Therefore, the rate capability or the power density of theLOB cell directly depends on the efficiency of O2 diffusion in theporous positive electrode. In this regard, the architecture of theporous positive electrode plays a crucial role in determining theefficiency of O2 transport to reaction sites. Two microstructuralparameters, porosity (𝜖) and tortuosity (𝜏), are especially impor-tant. Porosity is defined as the fraction of void volume withinthe total electrode volume and relates to the space available forelectrolyte infiltration and gas transport. A higher porosity gen-erally provides more open channels for O2 diffusion. Tortuosity,on the other hand, characterizes the geometric complexity andconnectivity of the pore network. It quantifies how much longeror more convoluted the actual diffusion path is compared to astraight-line path across the electrode. Increased tortuosity re-sults in greater resistance tomass transport due to effects such asdead-end pores, narrow necks, and poorly connected pathways.The effective diffusion coefficient (Deff) captures how these twoparameters modulate O2 transport in the electrode and is typi-cally expressed as:[30]Deff = D0𝜀𝜏(1)where D0 is the bulk diffusion coefficient of O2 in the electrolyte.This relationship makes clear that while increasing porosity fa-cilitates diffusion, increasing tortuosity significantly impedes it.Adv. Sci. 2026, 13, e19091 e19091 (2 of 10) © 2025 The Author(s). Advanced Science published by Wiley-VCH GmbH 21983844, 2026, 9, Downloaded from https://advanced.onlinelibrary.wiley.com/doi/10.1002/advs.202519091 by National Institute For, Wiley Online Library on [16/02/2026]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons Licensehttp://www.advancedsciencenews.comhttp://www.advancedscience.comwww.advancedsciencenews.com www.advancedscience.comFigure 1. Variation of the effective diffusion coefficient (Deff) of oxygen (O2) with a) porosity (𝜖) and b) tortuosity (𝜏). c–h) Contour plots of normalizedO2 concentration across the electrode thickness under different current densities.Figure 1a,b illustrates the dependence ofDeff on porosity and tor-tuosity, respectively. As shown in Figure 1a, higher porosity leadsto a linear increase in Deff, as more free volume becomes avail-able forO2 molecules. In contrast, Figure 1b highlights a strongernon-linear inverse dependence of Deff on tortuosity, where smallincreases in 𝜏 can lead to a substantial reduction in effective dif-fusivity.To quantitatively evaluate how the microstructure of the posi-tive electrode, specifically 𝜖 and 𝜏, influences O2 availability, wemodeled steady-state 1D diffusion across the electrode thickness,incorporating a homogeneous first-order reaction term, by thefollowing equation:[21]C (x) = C0exp(−xjnFC0Deff)(2)where C(x), C0, j, n, F, and Deff are the concentration of O2 in theelectrode at distance x from electrode-O2 interface, O2 solubilityin the bulk electrolyte, current density normalized to the elec-trochemically active surface area (ECSA) of the carbon, numberAdv. Sci. 2026, 13, e19091 e19091 (3 of 10) © 2025 The Author(s). Advanced Science published by Wiley-VCH GmbH 21983844, 2026, 9, Downloaded from https://advanced.onlinelibrary.wiley.com/doi/10.1002/advs.202519091 by National Institute For, Wiley Online Library on [16/02/2026]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons Licensehttp://www.advancedsciencenews.comhttp://www.advancedscience.comwww.advancedsciencenews.com www.advancedscience.comof electrons in the rate limiting step, Faraday constant, and effec-tive diffusion coefficient of O2 in the electrolyte inside the porouselectrode, respectively. The value of n is taken to be 1, based onprevious reports, considering one electron reduction step (O2 +e− ↔ O2−) to be kinetically rate-limiting during the discharge ofLOBs.[29,31,32] This framework captures the critical coupling be-tween structural parameters and electrochemical kinetics undervarious discharge conditions. Figure 1c–h displays normalizedO2 concentration (C(x)/C0) profiles as contour plots across theelectrode thickness, resolved for a matrix of porosity values (𝜖 =0.9 and 0.7) and tortuosity factors (𝜏 = 1, 2.5, and 5). These pa-rameter sets represent idealized open structures to more restric-tive, highly convoluted diffusion pathways, respectively. At highporosity (𝜖 = 0.9), the diffusion pathways offer substantial voidspace for gas-phase transport. In the ideal case with minimal tor-tuosity (𝜏 = 1, Figure 1c),> 25% of the initially available O2 is effi-ciently delivered throughout the electrode depth of 100 μm, evenat high current densities approaching 1.0 mA cm−2. However,as 𝜏 increases to 2.5 and 5 (Figure 1d,e), the effective diffusivitydecreases due to more tortuous transport paths. Consequently,under 𝜏 = 5, O2 becomes significantly depleted before reachingthe separator side, with < 2.5% of initial O2 concentrations ob-served at 1.0 mA cm−2, marking a clear transition to diffusion-limited behavior. The effect of reduced porosity (𝜖 = 0.7) furtherexacerbates the limitations onO2 transport. Even under favorabletortuosity (𝜏 = 1, Figure 1f), the O2 profile shows markedly de-creased concentration at current density 1.0 mA cm−2. Only 15%of the initial O2 concentration can be transported through thethickness of 100 μm at 1.0 mA cm−2. When 𝜏 is increased to 2.5and 5 (Figure 1g,h), these O2 concentration values drop sharply.In particular, at 𝜖 = 0.7 and 𝜏 = 5, O2 is almost entirely depletednear the separator at discharge rates above 0.5 mA cm−2, leavingmuch of the electrode underutilized due to insufficient reactantsupply. The solid white lines in the contour plots indicate regionswhere the O2 concentration drops to 2% of its initial value. Asporosity decreases and tortuosity increases, thesewhite lines shifttoward lower electrode thicknesses and current densities, under-scoring the increasing limitation on O2 transport under less fa-vorable structural conditions. These simulations clearly demon-strate that achieving high porosity alone is insufficient to ensureeffective O2 transport at practical discharge rates in LOBs. Oursimulations suggest that lowering tortuosity could be a more ef-fective strategy to improve O2 transport and enable higher powerperformance in LOBs, provided that such control can be realizedexperimentally.2.2. Design Strategy of Spatially Interconnected Graphene-BasedPorous ElectrodeDespite significant efforts to tailor carbon porosity, the spatialorganization of carbon within the electrode and the tortuosityfactor remain largely overlooked. Conventional electrodes, com-posed of carbon particles bound by polymers, often form denselypacked structures that restrict O2 diffusion and limit active mate-rial utilization.[33] Figure 2 schematically illustrates, through ex-treme hypothetical examples, the contrast between ideal and non-ideal electrode architectures. The top panel shows conventionalelectrodes with poor pore connectivity, absence of macropores,Figure 2. Schematic illustration of the structural differences among var-ious carbon electrodes depending on the fabrication approach. The toppanel depicts a particle-based porous carbon electrode fabricated using abinder, where pores are primarily confined within individual particles. Themiddle panel shows a nonideal meso-macroporous carbon membrane inwhich macropores exist but remain spatially isolated, limiting mass trans-port. The bottompanel illustrates the ideal architecture, a freestanding car-bon membrane featuring a continuous network of interconnected macro-pores that facilitate efficient electrolyte infiltration and gas diffusion. Smallwhite circles denote mesopores, while large white circles indicate macrop-ore. The micropores and interconnectivity through microporous channelsare not shown.and blocked transport channels. The middle panel representsa nonideal structure with macropores lacking interconnectivity,thus consuming a high volume of electrolyte without improvingmass transport. The bottom panel illustrates the optimally idealcase: a network of interconnected macropores that supports effi-cient gas diffusion and electrolyte transport.Motivated by this understanding, we developed a freestandinggraphene-based electrode with a hierarchically interconnectedporous framework. A GMS carbon was synthesized based ona previously reported method.[34] The scanning electron micro-graph (SEM) in Figure S1 (Supporting Information) reveals aspherical morphology of the GMS particles. The particle-sizedistribution curve of the GMS, derived from SEM analysis andshown in Figure S2 (Supporting Information), indicates a uni-form size distribution centered around 50 μm. The transmissionelectron micrograph (TEM) in Figure 3a shows that the GMSparticles are actually aggregates of graphene flakes. The struc-tural and surface chemical properties of the GMS carbon werethoroughly characterized using X-ray diffraction (XRD), Ramanspectroscopy, and X-ray photoelectron spectroscopy (XPS), andthe results were compared with those of the commercial KB car-bon. The results from XRD patterns for GMS and KB powdersamples are shown in Figure S3 (Supporting Information). Acomparison of the Raman spectra, shown in Figure S4 (Sup-porting Information), also reveals that GMS exhibits compara-tively sharper peaks, and a lower ID/IG ratio (1.62 vs 1.82 forKB) in Figure S5 (Supporting Information) confirms a lowerstructural defect in GMS. Surface chemical analysis via XPS re-veals slightly higher carbon content in GMS compared to KB,Adv. Sci. 2026, 13, e19091 e19091 (4 of 10) © 2025 The Author(s). Advanced Science published by Wiley-VCH GmbH 21983844, 2026, 9, Downloaded from https://advanced.onlinelibrary.wiley.com/doi/10.1002/advs.202519091 by National Institute For, Wiley Online Library on [16/02/2026]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons Licensehttp://www.advancedsciencenews.comhttp://www.advancedscience.comwww.advancedsciencenews.com www.advancedscience.comFigure 3. a) Transmission electron micrograph (TEM) of GMS powder sample. Cross-section scanning electron micrograph (SEM) of b) GMS and c)iGMS membranes. d) Schematic representation of the electrode fabrication processes following NIPS method with and without using PEO. e) Thermo-gravimetric analysis (TGA) data of PAN and PEO under a helium atmosphere. Pore size distribution curves of GMS and iGMS membranes measuredby f) N2 adsorption/desorption and g) Hg porosimetry.as shown in Figure S6 (Supporting Information). Nevertheless,both carbons exhibit similar types of oxygen-containing func-tional groups. Deconvoluted XPS spectra (Figures S7–S10, Sup-porting Information) identify C1s peaks at 285.2–285.6, 286.8–287.0, and 289.8–290.2 eV, which correspond to C−O, C═O,and COO− functional groups, respectively, with correspondingO1s peaks at 533.2–533.6 eV (C−O), 532.0–532.5 eV (C═O), and535.5–535.6 eV (COO−).[23] These results indicate that the GMSelectrode possesses enhanced graphitization and a higher carboncontent, which are expected to enhance the electrochemical sta-bility of GMS-based electrodes relative to KB electrodes.As mentioned previously, the porous carbon positive elec-trode in LOBs should have high O2 diffusion ability and ef-ficient electrochemical properties. From a practical applicabil-ity and scalability point of view, a self-standing hierarchicallyporous electrode is a desirable choice. Therefore, self-standingcarbon membranes were fabricated using a slurry casting tech-nique combined with NIPS process, developed within our re-search group.[35] The slurry was prepared by mixing carbon pow-ders, carbon nanotubes (CNTs), PAN, and/or PEO inN-methyl-2-pyrrolidone (NMP) solvent. CNTs were incorporated to enhancethe mechanical integrity of the membranes. A uniform carbonfilm was prepared using a doctor blade technique and subse-quently immersed in methanol, a poor solvent, to initiate theNIPS process, leading to the development of macroporous voidswithin the film. The membranes were dried and then underwentheat treatment followed by carbonization at 1050 °C under aninert atmosphere. During the carbonization process, PAN trans-formed into a robust carbon scaffold with isolated macropores,while PEO decomposed into volatile compounds, opening up themacroporous channels and forming an interconnected hierarchi-cal macro-mesoporous network.[36,37] The sacrificial decomposi-tion of PEO played a pivotal role in connecting the macropores,enabling efficient transport of O2 dissolved in the electrolyte. Thecross-sectional focused ion beam (FIB) SEM image in Figure 3breveals that the GMS/PAN electrode without PEO (denoted asAdv. Sci. 2026, 13, e19091 e19091 (5 of 10) © 2025 The Author(s). Advanced Science published by Wiley-VCH GmbH 21983844, 2026, 9, Downloaded from https://advanced.onlinelibrary.wiley.com/doi/10.1002/advs.202519091 by National Institute For, Wiley Online Library on [16/02/2026]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons Licensehttp://www.advancedsciencenews.comhttp://www.advancedscience.comwww.advancedsciencenews.com www.advancedscience.comTable 1. Surface area and pore volume of different membrane electrodesmeasured by N2 adsorption/desorption and Hg porosimetry.MembraneelectrodeBETsur-facearea[m2g−1]Pore volume [cm3 g−1]<2 nm 2–20nm>20nm0.2–10μmGMS 1667 0.59 2.56 0.17 3.28iGMS 1711 0.60 2.45 0.35 1.18KB 802 0.35 2.15 2.45 2.33GMS) displays the formation of robust macropores, which ap-pear to be isolated without extended connection via macropores.These macropores can be linked through microporous walls thatpermit electrolyte infiltration, but such connections via microp-ores provide limited pathways for efficient mass transport dur-ing electrochemical reactions. However, in contrast, when PEOwas incorporated, as shown in the FIB-SEM image in Figure 3c,the non-oxidative thermal decomposition of PEO created openchannels through the pore walls and established an extended net-work of interconnected macropores in the electrode (denoted asiGMS). A proposed mechanism of these two types of electrodepreparation is shown in Figure 3d. The thermogravimetric anal-ysis (TGA) data for PAN and PEO in Figure 3e under an inertatmosphere show that upon heating to 1100 °C, PAN exhibited a69.2%mass loss, while PEO lost 97.2% of its mass. These resultsconfirm the carbonization of PAN and the non-oxidative decom-position of PEO.We performed a series of structural characterizations forthe three kinds of carbon electrodes: KB, GMS, and iGMSmembranes. Figures S11 and S12 (Supporting Information) il-lustrate the N2 adsorption/desorption isotherms for the GMSand iGMS membranes, while a reference KB membrane is in-cluded for comparison in Figure S13 (Supporting Information).Brunauer−Emmett−Teller (BET) surface area analysis revealsthat the addition of PEO has a negligible impact on the BET sur-face area of the GMS electrodes, which measured 1711 m2 g−1with PEO and 1667 m2 g−1 without PEO. Similarly, the Barrett-Joyner-Halenda (BJH) pore size distribution curves (Figure 3f)indicate no difference in mesopore diameters between GMSelectrodes fabricated with and without PEO. Detailed analysis(Table 1) shows that the iGMS electrode also exhibits nearly sim-ilar mesopore volumes (2.8 cm3 g−1) compared to the GMS elec-trode (2.73 cm3 g−1). Macropore volume analysis using mercury(Hg) porosimetry, as shown in Figure 3g, highlights significantdifferences in the macroporous structures of the electrodes. Themacropores with diameters of approximately 2 μm in the GMSmembrane were collapsed due to the sacrificial decompositionof PEO during carbonization, resulting in a reduction inmacrop-ore volume from 3.28 cm3 g−1 in the GMSmembrane to 1.18 cm3g−1 in the iGMS membrane. Interestingly, the decomposition ofPEO facilitated the opening of larger macropores with diametersexceeding 10 μm. These large macroporous channels are criticalfor enabling efficient electrolyte diffusion within the electrode,particularly for practical applications requiring thicker electrodesto achieve high specific energy. Overall, these results underscorethe advantages of the NIPS process combined with the strategicselection of polymer materials, allowing precise control over theporosity of the carbon membrane electrode.2.3. Application of GMS Electrodes in LOBsTo evaluate the impact of pore optimization on electrodeperformance, the discharge capacities of the electrodeswere assessed in a stack-type cell (Figure S14, Support-ing Information) using a controlled amount of 1 m lithiumbis(trifluoromethanesulfonyl)imide (LiTFSI) in tetraethyleneglycol dimethyl ether (TEGDME or G4) as the electrolyte. Theelectrolyte loading amount was set to the electrolyte to carbonmass ratio (EL/C) of 5. Discharge tests were performed at currentdensities of 0.4 and 1.0 mA cm−2 to investigate the effect of poreoptimization on rate-dependent capacities. The actual mem-brane electrode masses, along with the mass-normalized currentdensity and specific capacity values, are listed in Table S1 of theSupporting Information. At a low discharge rate of 0.4 mA cm−2,the GMS and iGMS electrodes exhibited higher capacities of2350 and 2850mAh g−1, respectively (Figure 4a), whereas the KBelectrode achieved a slightly lower capacity of approximately 2100mAh g−1 (Figure S15, Supporting Information). In contrast, ata higher discharge rate of 1.0 mA cm−2, the advantages of GMSelectrodes and pore optimization become strikingly evident. Asshown in Figure 4b and Figure S16 (Supporting Information),the GMS electrode achieved a significantly high capacity of1800 mAh g−1, and the iGMS electrode further improved toan impressive 2510 mAh g−1. The concave feature observed inthe initial stage of the discharge curve of GMS in Figure 4barises from transient limitations in electrolyte transport withinthe porous GMS electrode under lean-electrolyte and high-rateconditions. As discharge proceeds, Li2O2 gradually depositswithin the pores, partially blocking the pores and mitigatingelectrolyte scarcity, which allows the voltage to recover. The KBelectrode, in comparison, exhibited a negligible capacity of just12 mAh g−1 (Figure S17, Supporting Information). Moreover,Figure 4c shows that the average discharge potential of iGMSelectrode (2.46 V vs Li/Li+) is higher than that of GMS electrode(2.38 V vs Li/Li+) at a high rate of 1 mA cm−2. It is importantto mention that the capacity trends observed here inverselycorrelate with the total pore volume (combining meso- andmacropores) of the electrodes: KB (6.93 cm3 g−1) > GMS (6.01cm3 g−1) > iGMS (3.98 cm3 g−1). These results emphasize that,under lean electrolyte and high current density conditions,discharge capacity is not directly correlated with the total porevolume of the electrode. Instead, an optimized pore structureconcerning efficient electrolyte diffusion and proper electrodefilling is critical for achieving high capacity, particularly underfast discharge conditions. A highly interconnected porous net-work that promotes efficient diffusion, together with a balancedpore volume that minimizes excessive electrolyte demand whilemaintaining sufficient capacity, provides a viable approach toaddress the energy–power trade-off in LOBs. While large porevolumes in carbon electrodes intuitively provide sufficient spacefor Li2O2 deposition and facilitate O2 and Li-ion transport, theyalso require substantial electrolyte volumes for effective pore-filling. Under lean electrolyte conditions (EL/C = 5), the largepore volume of the KB-based electrode resulted in incompleteAdv. Sci. 2026, 13, e19091 e19091 (6 of 10) © 2025 The Author(s). Advanced Science published by Wiley-VCH GmbH 21983844, 2026, 9, Downloaded from https://advanced.onlinelibrary.wiley.com/doi/10.1002/advs.202519091 by National Institute For, Wiley Online Library on [16/02/2026]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons Licensehttp://www.advancedsciencenews.comhttp://www.advancedscience.comwww.advancedsciencenews.com www.advancedscience.comFigure 4. Galvanostatic discharge profiles of LOB cells with GMS and iGMS electrodes at a) 0.4 mA cm−2 and b) 1.0 mA cm−2 up to 2.0 V versus Li/Li+.c) Average discharge voltages at 1.0 mA cm−2. Cycling profiles of d) GMS and e) iGMS at 0.4 mA cm−2 with 4 mAh cm−2 capacity under EL/C = 6, andf) corresponding cycle lives. Cycling profiles of g) GMS and h) iGMS under EL/C = 4, and i) corresponding cycle lives. High-rate cycling of j) GMS andk) iGMS electrodes. l) Comparison of high-rate cycling stabilities of iGMS electrode with previously reported cells cycled at current densities ≥ 0.8 mAcm−2 and limited capacities.pore filling, leading to significant voltage polarization and lowcapacity at high rates. Conversely, the reduced pore volumesof GMS-based electrodes minimized electrolyte requirements,enabling higher capacities under the same conditions. Betweenthe GMS and iGMS electrodes, the connection of macroporesand reduction of macropore volume in the iGMS electrodefurther improved capacity and rate capability. The enhancedelectrolyte diffusion facilitated by interconnected pores (lowertortuosity) and the reduction of inactive porosity were key tothis performance improvement. The impedance spectra ofcarbon|separator|carbon symmetric cells using GMS and iGMSelectrodes with three electrolyte loadings (EL/C = 4, 5, and 6)Adv. Sci. 2026, 13, e19091 e19091 (7 of 10) © 2025 The Author(s). Advanced Science published by Wiley-VCH GmbH 21983844, 2026, 9, Downloaded from https://advanced.onlinelibrary.wiley.com/doi/10.1002/advs.202519091 by National Institute For, Wiley Online Library on [16/02/2026]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons Licensehttp://www.advancedsciencenews.comhttp://www.advancedscience.comwww.advancedsciencenews.com www.advancedscience.comare presented in Figures S18–S20 (Supporting Information).When the electrodes are adequately filled with electrolyte atEL/C = 5 and 6, both cells exhibit comparable impedancevalues. However, under lower electrolyte loading conditions(EL/C = 4), the iGMS electrode with its interconnected porousarchitecture displays significantly lower impedance. Because thesymmetric cell configuration eliminates faradaic contributions,the measured impedance primarily reflects electrolyte transportwithin the porous membrane. Under lean electrolyte conditions,limitations in electrolyte diffusion and insufficient electrodefilling play a more dominant role than total pore volume indetermining discharge capacity, particularly during high-poweroperations. The cell level energy and power density values ofKB, GMS, and iGMS electrodes along with the high performingelectrodes reported in the literature are compared in Figure S30and Table S2 (Supporting Information). Themasses of all the cellcomponents including the electrolyte mass are considered forthe calculation of gravimetric energy and power densities. Theresults reveal that for GMS-based electrodes, particularly iGMS,increasing the power density has minimal impact on energydensity. In contrast, the KB electrode exhibits a pronounced dropin energy density under high-power operation. Furthermore, theiGMS electrode exhibits a cell level energy density of ≈990 Whkg−1, which is comparable to that of several recently reportedhigh-energy-density pouch-type LOB cells at similar powerdensities (≈30 W kg−1), while maintaining excellent retention(> 800 Wh kg−1) under high-power operation (≈80 W kg−1).These findings underscore the importance of optimizing porestructures to balance efficient electrolyte utilization with highcapacity, especially under practical conditions requiring leanelectrolyte and high-power performance.We next examined the impact of optimized porosity on the cy-cling performance of electrodes under lean electrolyte and highareal capacity conditions. For the cycling experiments, we set a ca-pacity of 4 mAh cm−2 and used two levels of very lean electrolyteloading, corresponding to EL/C of 6 and 4. The electrolyte usedfor the cycling experiments was 0.5 m LiTFSI + 0.5 m lithiumnitrate (LiNO3) + 0.2 m lithium bromide (LiBr) in TEGDME.Figure 4d,e depicts the galvanostatic discharge/charge profilesof the GMS and iGMS electrodes, respectively, for selected cy-cles with an EL/C ratio of 6. The XRD patterns of the dischargedand charged electrodes in Figures S21–S24 (Supporting Informa-tion) confirm the deposition and decomposition of Li2O2 as thedischarge product. As shown in Figure 4f, the GMS and iGMSelectrodes demonstrated 24 and 23 stable cycles, respectively atan EL/C ratio of 6. Similarly, Figure S25 (Supporting Informa-tion) shows that the KB electrode achieved 19 cycles under simi-lar conditions. XRD patterns in Figures S26 and S27 (SupportingInformation) show the evidence of Li2O2 deposition and decom-position during discharge and recharge of the KB electrode. Withthe progress of cycling, all cells exhibited a gradual increase involtage polarization and capacity fade, which are typical degrada-tion phenomena in LOBs caused by the accumulation of insu-lating side products that increase impedance and clog the elec-trode pores. The slightly higher cycling stability of GMS-basedelectrodes compared to KB can be attributed to the better crys-tallinity and higher degrees of graphitization of GMS. A signif-icant difference in stability emerged when the electrolyte load-ing was reduced to an EL/C of 4. Under this condition, the LOBcell with GMS electrode could not even charge during the firstcycle, and KB electrode failed within three cycles, as shown inFigure 4g, and Figure S28 (Supporting Information), respectively.These failures were marked by a sudden voltage spike duringcharging, indicating a sharp rise in cell impedance likely dueto electrolyte diffusion limitation. In contrast, the LOB cell withiGMS electrode, in Figure 4h, exhibited stable cycling for 23 cy-cles, even with the low EL/C of 4 (which is equal to the capac-ity normalized value of 3.25 g Ah−1). The cycling performancecomparison of GMS and iGMS electrodes at an EL/C ratio of 4(Figure 4i) clearly highlights that pore interconnectivity plays acritical role, beyond the contribution of pore volume alone, in de-termining long-term cell stability, particularly with lean amountof electrolyte. The interconnected porous network and reducedpore volume of the iGMS electrode likely enhanced the diffusionand retention of the electrolyte within the electrode pores, par-ticularly under extremely lean electrolyte conditions. As a result,the iGMS electrode demonstrated the most stable cycling withthe lowest electrolyte loading.2.4. High-Rate Cycling of LOBsAs mentioned earlier, the rate performance of LOBs is governedby the efficiency of O2 diffusion within the positive electrode andthe ionic conductivity of the electrolyte. Given the highly inter-connected pore structure of the iGMS electrode, designed forhigher electrolyte and O2 diffusion and minimal electrolyte re-quirement, a high-rate capability is expected. To confirm this,we compared the discharge/charge performances of cells us-ing GMS, iGMS, and KB electrodes at a high current densityof 1.5 mA cm−2. The tests employed a lean electrolyte with alow EL/C ratio of 5. As shown in Figure 4j, the LOB cell withGMS electrode failed to discharge even in the first cycle underthese demanding conditions. Similarly, the cell with KB elec-trode (Figure S29, Supporting Information) failed to completethe charge process during the first cycle, prematurely reachingthe cut-off voltage of 4.5 V vs Li/Li+. These sudden voltage polar-izations in the cases of GMS and KB electrodes indicate high cellimpedance, which makes high-rate cycling impossible. In con-trast, the cell with iGMS electrode successfully cycled for 90 cy-cles at the same high current density and low electrolyte load-ing. The discharge/charge voltage profiles for selected cycles, pre-sented in Figure 4k, illustrate the stable performance of iGMSelectrode. This finding of better rate-capability of iGMS electrodeconfirms the benefits of interconnected pores and proper porefilling under lean electrolyte conditions. Figure 4l compares thehigh-power cycling stability of the iGMS electrode with previ-ously reported electrodes cycled with aminimum current densityof 0.8 mA cm−2.[38,39] First of all, the number of reports on high-rate cycling of LOB is very limited. Moreover, none of the priordesigns could achieve any significant cycling stability at high rate.In contrast, the iGMS electrode achieved an impressive 90 cyclesat a high current density of 1.5 mA cm−2 under lean electrolyteconditions, setting a new benchmark for high-power LOBs.2.5. Analysis of Cell FailureAlongside electrolyte depletion, the degradation of both the elec-trode and the electrolyte plays a critical role in determining theAdv. Sci. 2026, 13, e19091 e19091 (8 of 10) © 2025 The Author(s). Advanced Science published by Wiley-VCH GmbH 21983844, 2026, 9, Downloaded from https://advanced.onlinelibrary.wiley.com/doi/10.1002/advs.202519091 by National Institute For, Wiley Online Library on [16/02/2026]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons Licensehttp://www.advancedsciencenews.comhttp://www.advancedscience.comwww.advancedsciencenews.com www.advancedscience.comFigure 5. Galvanostatic discharge/charge profiles and O2 evolution ratesduring charging for the a) 1st and b) 5th cycles using KB, GMS, and iGMSelectrodes. Comparison of O2 yield in different electrodes for the c) 1stand d) 5th cycles.cycle life of LOBs.[40,41] To investigate this, we conducted onlineelectrochemical mass spectrometry (online MS) to quantify gasevolution during cell charging, providing insights into the rel-ative stabilities of the electrodes. The fundamental operation ofLOBs relies on the reversible electrochemical reduction ofO2 dur-ing discharge and its evolution during charging. As a result,mon-itoring O2 evolution is a keymetric for evaluating the reversibilityof LOB processes. However, alongside the expected O2 evolution,undesirable parasitic reactions involving both the electrode andthe electrolyte can lead to the release of CO2. Consequently, an-alyzing the gases generated during cycling provides valuable in-sights into the reactions occurring at the positive electrode andtheir influence on the stability of the cells. We employed a spe-cialized two-compartment cell design for the online MS exper-iments. This setup isolates the Li negative electrode from thepositive electrode using a glass ceramic separator, allowing forthe exclusive detection of gases evolved from the positive elec-trode. We compared the gas evolution trends of KB, GMS, andiGMS electrodes to evaluate the stability of these carbon mate-rials. Figure 5a,b presents the voltage profiles for the 1st and5th cycles of LOBs using these electrodes with a limited capac-ity of 4 mAh cm−2, along with the corresponding gas evolutionrates. The quantitative results for O2 evolution are summarizedin Figure 5c,d. During the 1st charge cycle, there was no signifi-cant difference in the evolution rates of O2 between the LOB cellswith KB and two GMS-based electrodes (Figure 5c). However, bythe 5th charge cycle (Figure 5d), a noticeable decrease in O2 evo-lution was observed in the cell with KB electrode compared to thecell with GMS electrodes. Specifically, the KB electrode exhibited63%O2 evolution, while both the GMS electrodes achieved about69%O2 yield. The O2 yield is calculated based on the oxygen con-sumption rate of 2e− per O2 molecule during discharge. Sincethe same electrolyte was used and a consistent cut-off voltage wasemployed, the difference in the O2 evolution can be attributed tothe differences in the stabilities of the electrodes. As shown bythe physicochemical characterizations in Figures S3–S10 (Sup-porting Information), theGMS carbon exhibits a higher degree ofgraphitization, fewer defects, and a greater carbon content thanthe KB carbon. The more ordered carbon framework, with re-duced surface functional groups and defect sites, is less proneto parasitic reactions with reactive oxygen species such as sin-glet oxygen and superoxide intermediates. This structural stabil-ity improves the chemical robustness of the carbon matrix dur-ing repeated discharge/charge cycles, mitigating the formationof irreversible carbonates and other degradation products. Theseintrinsic advantages of the GMS carbon account for its enhancedO2 yield. Notably, optimization of the electrode pore architecturedoes not appear to significantly influence the O2 reversibility.3. ConclusionThis study highlights the pivotal role of an interconnected porousnetwork and optimized porosity in carbon electrodes for achiev-ing high capacity, long cycling life, and stable performance un-der high-power operations in LOBs operating with lean elec-trolyte conditions. While increasing electrode porosity is com-monly associated with higher capacity, proper connection of theporous network is crucial to ensure efficient diffusion of elec-trolyte and O2, which are essential for high-power performance.Additionally, electrodes with unoptimized large pore volumes of-ten require higher electrolyte volumes to maintain stable cycling,offsetting the specific capacity benefits. Conversely, upon lower-ing the electrolyte volume, highly porous electrodes face severeelectrolyte depletion, leading to pronounced voltage polarizationand premature cell failure. The development of electrodes witha highly interconnected pore network and optimized pore vol-ume, minimizing inactive porosity, offers a promising pathwayto overcoming the energy-power trade-off in LOBs. For example,a GMS electrode with an optimized pore structure demonstrateda high discharge capacity of 2520 mAh g−1 at a discharge rateof 1.0 mA cm−2 under a low electrolyte loading ratio (EL/C) of 5.Furthermore, this pore-optimized GMS electrode achieved stablecycling at a high areal capacity of 4mAh cm−2 under an EL/C ratioof 4 and delivered exceptional high-power cycling performance,sustaining 90 cycles at a high rate of 1.5 mA cm−2. In contrast,electrodes with greater pore volumes but lacking proper poreconnectivity exhibited inferior capacity, reduced cycling stability,and lower rate-capability under comparable lean electrolyte con-ditions. These findings underscore the necessity of finely tunedelectrode pore structures to balance capacity, cycling stability, rateperformance, and electrolyte utilization in LOBs. They also offervaluable insights into the design of advanced carbon materials,paving the way for the development of next-generation practicallithium-oxygen batteries.Supporting InformationSupporting Information is available from the Wiley Online Library or fromthe author.AcknowledgementsThe authors gratefully acknowledge financial support from ALCA-SPRING(Advanced Low Carbon Technology Research and Development Pro-gram─Specially Promoted Research for Innovative Next Generation Bat-teries) Project of the Japan Science and Technology Agency (JST: grantAdv. Sci. 2026, 13, e19091 e19091 (9 of 10) © 2025 The Author(s). Advanced Science published by Wiley-VCH GmbH 21983844, 2026, 9, Downloaded from https://advanced.onlinelibrary.wiley.com/doi/10.1002/advs.202519091 by National Institute For, Wiley Online Library on [16/02/2026]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons Licensehttp://www.advancedsciencenews.comhttp://www.advancedscience.comwww.advancedsciencenews.com www.advancedscience.comnumber JPMJAL1301) and also from Green Technologies of Excellence(GteX), Grant Number JPMJGX24S0. This work was also supported byJSPS KAKENHI grant number JP24K08590. Technical assistance was pro-vided by the Battery Research Platform at the National Institute for Mate-rials Science (NIMS).Conflict of InterestThe authors declare no conflict of interest.Data Availability StatementThe data that support the findings of this study are available from the cor-responding author upon reasonable request.Keywordscarbon electrode, high power battery, lean electrolyte, lithium–oxygen bat-tery, porosity optimization, rechargeable batteryReceived: September 28, 2025Revised: November 13, 2025Published online: December 2, 2025[1] K. Song, D. A. Agyeman, M. Park, J. Yang, Y.-M. Kang, Adv. Mater.2017, 29, 1606572. https://doi.org/10.1002/adma.201606572.[2] W.-J. Kwak, Rosy, D. S, C. Xia, H. Kim, L. R. Johnson, P. G. Bruce,L. F. Nazar, Y.-K. Sun, A. A. Frimer, M. Noked, S. A. Freunberger, D.Aurbach, Chem. Rev. 2020, 120, 6626.[3] O. L. Li, T. Ishizaki, in Emerging Materials for Energy Conversion andStorage, (Eds: K. Y. Cheong, G. Impellizzeri, M. A. Fraga), Elsevier,Amsterdam 2018, pp. 115–152.[4] M.-K. Song, S. Park, F. M. Alamgir, J. Cho, M. Liu,Mater. Sci. Eng., R2011, 72, 203. https://doi.org/10.1016/j.mser.2011.06.001.[5] A. A. Franco, K.-H. Xue, ECS J. Solid State Sci. Technol. 2013, 2, M3084.https://doi.org/10.1149/2.012310jss.[6] L.Ma, T. Yu, E. Tzoganakis, K. Amine, T.Wu, Z. Chen, J. Lu,Adv. EnergyMater. 2018, 8, 1800348. https://doi.org/10.1002/aenm.201800348.[7] D. Geng, N. Ding, T. S. A. Hor, S. W. Chien, Z. Liu, D. Wuu, X. Sun, Y.Zong, Adv. Energy Mater. 2016, 6, 1502164. https://doi.org/10.1002/aenm.201502164.[8] J. Liu, S. Khaleghi Rahimian, C. W. Monroe, Phys. Chem. Chem. Phys.2016, 18, 22840. https://doi.org/10.1039/C6CP04055A.[9] J. Wang, Y. Zhang, L. Guo, E. Wang, Z. Peng, Angew. Chem., Int. Ed.2016, 55, 5201. https://doi.org/10.1002/anie.201600793.[10] P. Albertus, G. Girishkumar, B. McCloskey, R. S. Sánchez-Carrera, B.Kozinsky, J. Christensen, A. C. Luntz, J. Electrochem. Soc. 2011, 158,A343. https://doi.org/10.1149/1.3527055.[11] S. Meini, M. Piana, H. Beyer, J. Schwämmlein, H. A. Gasteiger, J. Elec-trochem. Soc. 2012, 159, A2135.[12] N. Ding, S. W. Chien, T. S. A. Hor, R. Lum, Y. Zong, Z. Liu, J. Mater.Chem. A 2014, 2, 12433. https://doi.org/10.1039/C4TA01745E.[13] T. Kuboki, T. Okuyama, T. Ohsaki, N. Takami, J. Power Sources 2005,146, 766. https://doi.org/10.1016/j.jpowsour.2005.03.082.[14] S. B. Ma, D. J. Lee, V. Roev, D. Im, S.-G. Doo, J. Power Sources 2013,244, 494. https://doi.org/10.1016/j.jpowsour.2013.03.150.[15] S. R. Younesi, S. Urbonaite, F. Björefors, K. Edström, J. Power Sources2011, 196, 9835. https://doi.org/10.1016/j.jpowsour.2011.07.062.[16] M. Kim, E. Yoo, W.-S. Ahn, S. E. Shim, J. Power Sources 2018, 389, 20.https://doi.org/10.1016/j.jpowsour.2018.03.080.[17] K. Sakai, S. Iwamura, S. R. Mukai, J. Electrochem. Soc. 2017, 164,A3075. https://doi.org/10.1149/2.0791713jes.[18] T. Huang, F. Wu, S. Liu, G. Liu, R. Ran, W. Zhou, K. Liao, Small 2025,21, 2412208. https://doi.org/10.1002/smll.202412208.[19] J. Ye, C. Wang, Z. Shao, K. Liao, Energy Fuels 2024, 38, 15001. https://doi.org/10.1021/acs.energyfuels.4c02606.[20] S. Matsuda, E. Yasukawa, T. Kameda, S. Kimura, S. Yamaguchi, Y.Kubo, K. Uosaki, Cell. Rep. Phys. Sci. 2021, 2, 100506. https://doi.org/10.1016/j.xcrp.2021.100506.[21] A. Dutta, T. Kameda, J. Takada, Y. Nakajima, T. Morishita, S. Matsuda,Adv. Sci. 2025, 14406. https://doi.org/10.1002/advs.202514406.[22] A. Dutta, K. Ito, Y. Kubo, J. Mater. Chem. A 2019, 7, 23199. https://doi.org/10.1039/C9TA07427A.[23] A. Dutta, K. Ito, A. Nomura, Y. Kubo, Adv. Sci. 2020, 7, 2001660. https://doi.org/10.1002/advs.202001660.[24] J. K. Papp, J. D. Forster, C. M. Burke, H. W. Kim, A. C. Luntz, R. M.Shelby, J. J. Urban, B. D.McCloskey, J. Phys. Chem. Lett. 2017, 8, 1169.https://doi.org/10.1021/acs.jpclett.7b00040.[25] R. Black, S. H. Oh, J.-H. Lee, T. Yim, B. Adams, L. F. Nazar, J. Am.Chem. Soc. 2012, 134, 2902. https://doi.org/10.1021/ja2111543.[26] R. A. Wong, A. Dutta, C. Yang, K. Yamanaka, T. Ohta, A. Nakao, K.Waki, H. R. Byon, Chem. Mater. 2016, 28, 8006. https://doi.org/10.1021/acs.chemmater.6b03751.[27] W. Yu, Z. Shen, T. Yoshii, S. Iwamura, M. Ono, S. Matsuda, M. Aoki,T. Kondo, S. R. Mukai, S. Nakanishi, H. Nishihara, Adv. Energy Mater.2024, 14, 2303055. https://doi.org/10.1002/aenm.202303055.[28] W. Yu, T. Yoshii, A. Aziz, R. Tang, Z.-Z. Pan, K. Inoue, M. Kotani,H. Tanaka, E. Scholtzová, D. Tunega, Y. Nishina, K. Nishioka, S.Nakanishi, Y. Zhou, O. Terasaki, H. Nishihara, Adv. Sci. 2023, 10,2300268. https://doi.org/10.1002/advs.202300268.[29] J. Read, K. Mutolo, M. Ervin, W. Behl, J. Wolfenstine, A. Driedger,D. Foster, J. Electrochem. Soc. 2003, 150, A1351. https://doi.org/10.1149/1.1606454.[30] A. Torayev, S. Engelke, Z. Su, L. E. Marbella, V. De Andrade, A.Demortière, P. C. M. M. Magusin, C. Merlet, A. A. Franco, C. P. Grey,J. Phys. Chem. C 2021, 125, 4955. https://doi.org/10.1021/acs.jpcc.0c10417.[31] C. O. Laoire, S. Mukerjee, K. M. Abraham, E. J. Plichta, M. A.Hendrickson, J. Phys. Chem. C 2009, 113, 20127. https://doi.org/10.1021/jp908090s.[32] C. O. Laoire, S. Mukerjee, K. M. Abraham, E. J. Plichta, M. A.Hendrickson, J. Phys. Chem. C 2010, 114, 9178. https://doi.org/10.1021/jp102019y.[33] Y. Zhang, H. Zhang, J. Li, M. Wang, H. Nie, F. Zhang, J. Power Sources2013, 240, 390. https://doi.org/10.1016/j.jpowsour.2013.04.018.[34] H. Nishihara, T. Simura, S. Kobayashi, K. Nomura, R. Berenguer,M. Ito, M. Uchimura, H. Iden, K. Arihara, A. Ohma, Y. Hayasaka, T.Kyotani, Adv. Funct. Mater. 2016, 26, 6418. https://doi.org/10.1002/adfm.201602459.[35] J. Saengkaew, T. Kameda, M. Ono, S. Matsuda, Mater. Adv. 2022, 3,3536. https://doi.org/10.1039/D1MA01001H.[36] L. Costa, A.M.Gad, G. Camino, G. G. Cameron,M. Y.Qureshi,Macro-molecules 1992, 25, 5512. https://doi.org/10.1021/ma00046a059.[37] N. Grassie, R. McGuchan, Eur. Polym. J. 1970, 6, 1277. https://doi.org/10.1016/0014-3057(70)90046-7.[38] Z. Liang, Y.-C. Lu, J. Am. Chem. Soc. 2016, 138, 7574. https://doi.org/10.1021/jacs.6b01821.[39] X. Gao, Y. Chen, L. R. Johnson, Z. P. Jovanov, P. G. Bruce,Nat. Energy2017, 2, 17118. https://doi.org/10.1038/nenergy.2017.118.[40] M. Ono, J. Saengkaew, S.Matsuda, Adv. Sci. 2023, 10, 2300896. https://doi.org/10.1002/advs.202300896.[41] B. D. McCloskey, D. S. Bethune, R. M. Shelby, G. Girishkumar, A.C. Luntz, J. Phys. Chem. Lett. 2011, 2, 1161. https://doi.org/10.1021/jz200352v.Adv. Sci. 2026, 13, e19091 e19091 (10 of 10) © 2025 The Author(s). Advanced Science published by Wiley-VCH GmbH 21983844, 2026, 9, Downloaded from https://advanced.onlinelibrary.wiley.com/doi/10.1002/advs.202519091 by National Institute For, Wiley Online Library on [16/02/2026]. See the Terms and Conditions (https://onlinelibrary.wiley.com/terms-and-conditions) on Wiley Online Library for rules of use; OA articles are governed by the applicable Creative Commons Licensehttp://www.advancedsciencenews.comhttp://www.advancedscience.comhttps://doi.org/10.1002/adma.201606572https://doi.org/10.1016/j.mser.2011.06.001https://doi.org/10.1149/2.012310jsshttps://doi.org/10.1002/aenm.201800348https://doi.org/10.1002/aenm.201502164https://doi.org/10.1002/aenm.201502164https://doi.org/10.1039/C6CP04055Ahttps://doi.org/10.1002/anie.201600793https://doi.org/10.1149/1.3527055https://doi.org/10.1039/C4TA01745Ehttps://doi.org/10.1016/j.jpowsour.2005.03.082https://doi.org/10.1016/j.jpowsour.2013.03.150https://doi.org/10.1016/j.jpowsour.2011.07.062https://doi.org/10.1016/j.jpowsour.2018.03.080https://doi.org/10.1149/2.0791713jeshttps://doi.org/10.1002/smll.202412208https://doi.org/10.1021/acs.energyfuels.4c02606https://doi.org/10.1021/acs.energyfuels.4c02606https://doi.org/10.1016/j.xcrp.2021.100506https://doi.org/10.1016/j.xcrp.2021.100506https://doi.org/10.1002/advs.202514406https://doi.org/10.1039/C9TA07427Ahttps://doi.org/10.1039/C9TA07427Ahttps://doi.org/10.1002/advs.202001660https://doi.org/10.1002/advs.202001660https://doi.org/10.1021/acs.jpclett.7b00040https://doi.org/10.1021/ja2111543https://doi.org/10.1021/acs.chemmater.6b03751https://doi.org/10.1021/acs.chemmater.6b03751https://doi.org/10.1002/aenm.202303055https://doi.org/10.1002/advs.202300268https://doi.org/10.1149/1.1606454https://doi.org/10.1149/1.1606454https://doi.org/10.1021/acs.jpcc.0c10417https://doi.org/10.1021/acs.jpcc.0c10417https://doi.org/10.1021/jp908090shttps://doi.org/10.1021/jp908090shttps://doi.org/10.1021/jp102019yhttps://doi.org/10.1021/jp102019yhttps://doi.org/10.1016/j.jpowsour.2013.04.018https://doi.org/10.1002/adfm.201602459https://doi.org/10.1002/adfm.201602459https://doi.org/10.1039/D1MA01001Hhttps://doi.org/10.1021/ma00046a059https://doi.org/10.1016/0014-3057(70)90046-7https://doi.org/10.1016/0014-3057(70)90046-7https://doi.org/10.1021/jacs.6b01821https://doi.org/10.1021/jacs.6b01821https://doi.org/10.1038/nenergy.2017.118https://doi.org/10.1002/advs.202300896https://doi.org/10.1002/advs.202300896https://doi.org/10.1021/jz200352vhttps://doi.org/10.1021/jz200352v Interconnected Hierarchically Porous Graphene-Based Membrane Electrode for High-Power and Long-Cycle Lithium9040�Oxygen Battery 1. Introduction 2. Results and Discussions 2.1. Simulation of Quantitative Impacts of Porosity and Pore Connectivity (Tortuosity) on the Oxygen Diffusion 2.2. Design Strategy of Spatially Interconnected Graphene-Based Porous Electrode 2.3. Application of GMS Electrodes in LOBs 2.4. High-Rate Cycling of LOBs 2.5. Analysis of Cell Failure 3. Conclusion Supporting Information Acknowledgements Conflict of Interest Data Availability Statement Keywords