# Fileset

[91_Iimura_AENM_MoO3 for Mg and Ca storage.pdf](https://mdr.nims.go.jp/filesets/8eb62d6a-4629-4a1c-b470-c0b5350006fa/download)

## Creator

[Reona Iimura](https://orcid.org/0009-0006-4791-918X), [M. D. Hashan C. Peiris](https://orcid.org/0000-0002-8706-452X), [Takashi Yabu](https://orcid.org/0009-0000-8127-6110), [Toshihiko Mandai](https://orcid.org/0000-0002-2403-7794), [Ruijie Zhu](https://orcid.org/0000-0001-5317-1151), [Akira Nasu](https://orcid.org/0009-0002-7779-5721), [Saneyuki Ohno](https://orcid.org/0000-0001-8192-996X), [Masaki Matsui](https://orcid.org/0000-0003-1499-7457), [Itaru Honma](https://orcid.org/0000-0002-6536-576X), [Manuel Smeu](https://orcid.org/0000-0001-9548-4623), [Hiroaki Kobayashi](https://orcid.org/0000-0001-6705-9515)

## Rights

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

## Other metadata

[Ultra‐Low‐Strain Calcium and Magnesium Ion Storage Enabled by Tunnel‐Structured MoO                    <sub>3</sub>                    Positive Electrode](https://mdr.nims.go.jp/datasets/8a290029-696d-4535-a90d-31aaedffa1ca)

## Fulltext

Ultra‐Low‐Strain Calcium and Magnesium Ion Storage Enabled by Tunnel‐Structured MoO3 Positive ElectrodeAdvanced Energy Materialswww.advenergymat.deRESEARCH ARTICLEUltra-Low-Strain Calcium andMagnesium Ion StorageEnabled by Tunnel-Structured MoO3 Positive ElectrodeReona Iimura1,2 M. D. Hashan C. Peiris3 Takashi Yabu2 Toshihiko Mandai4 Ruijie Zhu2Akira Nasu2 Saneyuki Ohno1 Masaki Matsui2 Itaru Honma1 Manuel Smeu3,5Hiroaki Kobayashi21Institute of Multidisciplinary Research for Advanced Materials, Tohoku University, Sendai, Japan 2Department of Chemistry, Faculty of Science, HokkaidoUniversity, Sapporo, Japan 3Materials Science and Engineering, Binghamton University, Binghamton, New York, USA 4Research Center For Energy andEnvironmental Materials, National Institute for Materials Science (NIMS), Tsukuba, Japan 5Department of Physics, Binghamton University, Binghamton, NewYork, USACorrespondence: Reona Iimura (reona.iimura.r5@dc.tohoku.ac.jp) Hiroaki Kobayashi (kobahi@g.ecc.u-tokyo.ac.jp)Received: 6 January 2026 Revised: 24 March 2026 Accepted: 18 April 2026Keywords: calcium batteries | divalent batteries | magnesium batteries | molybdenum oxide | positive electrodeABSTRACTRechargeable divalent batteries employing Ca orMgmetal negative electrodes have attracted considerable interest due to their lowcost and potentially high energy density. However, the development of high-energy Ca and Mg batteries remains limited by thelack of oxide positive electrodes capable of reversibly accommodating divalent ions at room temperature. Here, we demonstratea new positive electrode material, a nano-sized hexagonal tunnel-structured MoO3 (nano-h-MoO3), as a structurally robust hostfor both Ca2+ and Mg2+ storage, exhibiting markedly improved reversibility and capacities. Comprehensive structural analyses,supported by computational modeling, reveal a unique charge–discharge mechanism in which divalent-ion (de)insertion occursthrough reversible modulation of host metal–oxygen bond lengths while retaining an intact host framework, resulting in minimallattice expansion (<2%). This structurally resilient tunnel-oxide design provides a promising pathway for developing high-energy,practical divalent metal battery systems.1RptforTossTc©AhIntroductionechargeable batteries represented by lithium-ion batteries sup-ort today’s portable electronics and electric vehicles owing toheir high energy densities, long cycle life, and mature manu-acturing ecosystem [1]. However, the accelerating electrificationf transport and stationary storage is straining supplies ofare metals due to cost volatility and sustainability concerns.hese pressures have intensified interest in “earth-abundant”r “cost-effective” elements that can detach future energytorage from supply-chain risk while maintaining strict safetytandards.his is an open access article under the terms of the Creative Commons Attribution License, which permited.2026 The Author(s). Advanced Energy Materials published by Wiley-VCH GmbHdvanced Energy Materials, 2026; 16:e71006ttps://doi.org/10.1002/aenm.71006Nonaqueous divalent metal batteries that employ Ca or Mgmetal negative electrode are promising candidates owing to theabundance of these elements and their high volumetric and gravi-metric capacity [2–5]. In addition, under appropriate interfacedesign between the negative electrode and electrolyte, Ca andMg metal exhibit reduced dendrite formation compared withlithiummetal, pointing to a compelling route to safer, high-energystorage [6, 7]. Yet the realization of practical rechargeable Ca orMg batteries remains fundamentally limited. This is not only bythe scarcity of electrolytes that enable reversible metal deposi-tion/stripping at room temperature, but also the lack of promisingpositive electrodematerials. In particular, nomaterials have beenits use, distribution and reproduction in any medium, provided the original work is properly1 of 12http://www.advenergymat.dehttps://doi.org/10.1002/aenm.71006https://orcid.org/0009-0006-4791-918Xhttps://orcid.org/0000-0002-8706-452Xhttps://orcid.org/0009-0000-8127-6110https://orcid.org/0000-0002-2403-7794https://orcid.org/0000-0001-5317-1151https://orcid.org/0009-0002-7779-5721https://orcid.org/0000-0001-8192-996Xhttps://orcid.org/0000-0003-1499-7457https://orcid.org/0000-0002-6536-576Xhttps://orcid.org/0000-0001-9548-4623https://orcid.org/0000-0001-6705-9515mailto:reona.iimura.r5@dc.tohoku.ac.jpmailto:kobahi@g.ecc.u-tokyo.ac.jphttp://creativecommons.org/licenses/by/4.0/https://doi.org/10.1002/aenm.71006http://crossmark.crossref.org/dialog/?doi=10.1002%2Faenm.71006&domain=pdf&date_stamp=2026-04-28awvtntk∼ecioDpTilcsrmtobctiIMtAicaIatdo[hw(ttHspotrdaTihrI2 16146840, 2026, 25, Downloaded from https://advanced.onlinelibrary.wiley.com/doi/10.1002/aenm.71006 by Toshihiko Mandai - National Institute For , Wiley Online Library on [07/07/2026]. 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 Creable to reversibly host divalent cations at room temperaturehile simultaneously achieving high capacity and high operatingoltage. To date, the most reliable host structures for room-emperature operation have been a small subset of chalcogenides,otably chain-type VS4 and chevrel-phase Mo6S8 [8, 9]. Whilehese phases accommodate Ca2+/Mg2+ with comparatively fastinetics, their average discharge voltages are low (e.g., Ca‖VS41.7 V, Mg‖Mo6S8 ∼1.0 V), thereby constraining the achievablenergy density. Pushing the potential upward toward valuesompetitive with state-of-the-art positive electrodes of lithium-on batteries requires oxide materials that enable higher redoxperation [10, 11].esigning oxide hosts for divalent-ion intercalation, however,resents significant challenges for room-temperature operation.he high charge density of divalent ions such as Mg2+ resultsn strong electrostatic interactions with the anionic framework,eading to sluggish solid-state diffusion and reduced accessibleapacity [12, 13]. Additionally, accumulated divalent ions aturface or defect sites induce phase transformations or conversioneactions, and ultimately accelerate capacity fade [14, 15]. Toitigate these issues, many studies have explored elevated-emperature operation (up to ∼150◦C). According to a surveyf more than 220 reports on oxide positive electrodes for Mgatteries [11], most studies that provide evidence of Mg2+ inter-alation reactions operated above room temperature to enhancehe divalent-ion mobility in the solid state. The median capacitymproved from 59 to 106 mAh g−1 under these conditions.n particular, α-MnO2 (150◦C) [16], ZnMnO3 (150◦C) [17], andgCrMnO4 (95◦C) [18] electrodes exhibited reversible capaci-ies of approximately 100, 120, and 110 mAh g−1, respectively.lthough these elevated-temperature approaches achieved somemprovement in capacity and reversibility, they remain impracti-al for most applications due to the lack of suitable electrolytesnd cell components for high-temperature operation.n recent research on oxide materials for room-temperature oper-tion, particle downsizing has emerged as an effective strategyo enhance electrochemical performance by shortening cationiffusion pathways [12, 19]. In particular, our studies have focusednα-MnO2, a promising host identified throughDFT calculations20] and 150◦C operation test [16] as both a Ca and Mg batteryost, respectively. By employing an alcohol-solution process,e successfully synthesized ultrasmall α-MnO2 nanoparticles<10 nm) with a low aspect ratio (c/a ≈ 2) [21]. This nanos-ructured material exhibited markedly improved utilization ofhe positive electrode compared with conventional counterparts.owever, despite improved utilization in both Ca andMg batteryystems, its reversibility remained limited due to irreversiblehase transitions, either into a highly distorted tunneled structurer into a stable rock salt phase, indicating that divalent ions tendo be trapped within the tunnel framework and progressivelyeduce accessible capacity. These findings highlight that theesign of a structurally robust host framework is essential tochieving long-life oxide-type electrodes.o demonstrate the necessity of a robust framework for divalent-on insertion, we suggest a new positive-electrode material, aexagonal tunnel-structured MoO3 (h-MoO3), which is geomet-ically predisposed to accommodate structural relaxation [22].n both Ca and Mg battery systems, the nanosized h-MoO3of 12(nano-h-MoO3) positive electrode exhibited superior capacityand remarkable structural reversibility, enabling reversible diva-lent cation insertion with no phase transitions and minimallattice distortion, supported by computational approaches. Inparticular, a Mg battery employing a fluorinated-alkoxyborateelectrolyte [23], which offers superior compatibility with the Mgnegative electrode, delivered excellent cyclability, retaining stableperformance for 100 cycles.2 Results and Discussion2.1 Synthesis and Characterization of h-MoO3Molybdenum trioxide (MoO3) exists in three main structuralpolymorphs: the thermodynamically stable orthorhombic phase(α-MoO3), which has a layered structure; the metastable mon-oclinic phase (β-MoO3), composed of MoO6 octahedral unitsforming a 1 × 1 tunnel structure; and the metastable hexagonalphase (h-MoO3), which features a 1D tunnel framework. In h-MoO3, the tunnels run along the c-axis and are surrounded by12 MoO6 octahedra (Figure 1a). The octahedra are corner-sharedalong the a direction, while edge-sharing occurs along the cdirection, establishing a robust 1D tunnel network. Additionally,the tunnel size of h-MoO3 is sufficiently large to accommodatecations of various ionic radii, and the nano h-MoO3 synthesizedin this study is therefore expected to facilitate reversible insertionand de-insertion of bothmonovalent and divalent cations at roomtemperature.Figure S1 shows the schematic illustration of the synthesis processfor h-MoO3 and nano-h-MoO3. Commercial α-MoO3 and NH4Clwere first dissolved in anH2O2 aqueous solution, transferred to anautoclave, and heated at 150◦C to obtain h-MoO3. The productwasthen ball-milled using ZrO2 balls at 600 rpm, yielding the blueishnano-h-MoO3.Figure 1b shows the X-ray diffraction (XRD) patterns andRietveldrefinement results of h-MoO3 and nano-h-MoO3, the latterobtained by ball-milling of the as-synthesized h-MoO3. The as-prepared h-MoO3 exhibited high crystallinity and a single phase,which was indexed to the monoclinic structure with the P63/mspace group. CHN elemental analysis and thermogravimetric(TG) analyses (Figure S2), corrected for 2 wt.% absorbed water,confirmed the presence of nitrogen with an N/Mo ratio of 0.140.This nitrogen is reasonably attributed to NH4+ ions accommo-dated within the hexagonal tunnels, and the measured contentagrees well with previously reported values [22]. In contrast, ball-milling yields broader diffraction peaks with reduced intensitiesin the XRD patterns of nano-h-MoO3, while no additional phaseswere observed, indicating that particle downsizing occurredwithout structural decomposition. The reduced crystallite sizeestimated using the Halder–Wagner equation based on Rietveldrefinement further supports the successful reduction in grain size(Table S1). Notably, the N/Mo ratio decreased to 0.118 after ball-milling. According to a previous study [24], the decrease in NH4+content is attributable to partial decomposition of tunnel NH4+species during ball-milling, whereNH3 is released by the heat andmechanical impact generated duringmilling, and residual protonspecies are likely left in the tunnel framework. In this study, theseproton specieswere not directly quantified; however, the decreaseAdvanced Energy Materials, 2026tive Commons LicenseFIGURE 1 (a) The crystal structure of h-MoO3 and ionic radius of each guest cation, (b) XRD patterns with fitting curves by Rietveld refinement,(c) EPR spectra, (d) Raman spectra, (e) SEM images of (i) h-MoO3 and (iii) nano-h-MoO3, HR-TEM images of (ii) h-MoO3 and (iv) nano-h-MoO3.iapihMwfwpaiA 16146840, 2026, 25, Downloaded from https://advanced.onlinelibrary.wiley.com/doi/10.1002/aenm.71006 by Toshihiko Mandai - National Institute For , Wiley Online Library on [07/07/2026]. 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 Crean the N/Mo ratio from 0.140 to 0.118 suggests an upper-limitmount of approximately 0.022 per Mo, assuming one residualroton remains for each released NH3 molecule. As shownn Figure S2, the lowered phase-transition temperature from-MoO3 to layered α-MoO3 [25] was also observed in the nano-h-oO3. This decrease is attributed to a reduction in NH4+ content,hich serves as a structural pillar stabilizing thehexagonal tunnelramework. Therefore, the reduction in NH4+ concentrationeakens the structural integrity of the tunnel, facilitating easierhase transformation of nano-h-MoO3 compared to h-MoO3. Inddition, the lattice volume of nano-h-MoO3 increased by approx-mately 1% compared with h-MoO3, while the average Mo─Odvanced Energy Materials, 2026bond length became longer. This bond-length variation was alsoconfirmed by pair distribution function (PDF) analysis. As shownin Figure S3, the Mo─O(1) distance, which is assigned to theMo–isolated O, was longer in nano-h-MoO3 than in h-MoO3 [26].This structural change can be mainly attributed to the differencein the NH4+ content in the tunnels. The smaller amount ofNH4+ in nano-h-MoO3 likely induces shrinkage of the isolated-Ohexagonal area, resulting in elongation of the Mo─O(1) distance.To investigate the Mo valence state before and after ball-milling,electron spin resonance (EPR) measurements were performed(Figure 1c).h-MoO3 exhibited no detectable signal, whereas nano-h-MoO3 displayed two distinct peaks at g = 1.927 and 1.886,3 of 12tive Commons Licensewsgrt(aMiMacqrttv3ooaomniaeeeoHmTtpccTfen7ns2PFnwacoMFcbrc4 16146840, 2026, 25, Downloaded from https://advanced.onlinelibrary.wiley.com/doi/10.1002/aenm.71006 by Toshihiko Mandai - National Institute For , Wiley Online Library on [07/07/2026]. 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 Creahich are unique peaks to Mo5+ species [27]. Importantly, noignal corresponding to oxygen vacancies (typically observed near≈ 2.0) was detected, indicating that the oxygen frameworkemained unchanged upon nanoparticulation. To further inves-igate the Mo state, both Mo 3d X-ray photoelectron spectroscopyXPS) and Mo K-edge X-ray absorption fine structure (XAFS)nalyses, as shown in Figure S4, were conducted. In nano-h-oO3, new spectral features corresponding to Mo5+ appearedn the XPS spectra (Mo6+: 3d5/2 = 232.8 eV, 3d3/2 = 236.02 eV,o5+: 3d5/2 = 231.7 eV, 3d3/2 = 234.5 eV) [28, 29] and the XAFSbsorption edge exhibited a slightly lower-energy shift. Thesehanges were relatively minor, as XPS analysis with the semi-uantitative fitting revealed that only about 3% of Mo6+ waseduced to Mo5+. Raman spectroscopy (Figure 1d) was employedo probe the vibrational modes of the tunnel framework. Most ofhe observed peaks, assigned to Mo─O─Mo and terminal Mo═Oibrations, were consistent with previously reported results [30,1] and remained unchanged before and after nanoparticulationf h-MoO3. However, additional peaks appeared in the spectrumf nano-h-MoO3 at around 800 cm−1. These new features arettributed to O─Mo─O vibrational modes that are characteristicf the layered α-MoO3 structure [32]. The emergence of theseodes is likely related to the reduced amount of NH4+ ions inano-h-MoO3; with fewer NH4+ ions occupying the tunnels, then-plane O─Mo─O vibrations toward the tunnel might becomectivated [33]. Figure 1e(i, iii) and (ii, iv) present the scanninglectron microscopy (SEM) and high-resolution transmissionlectron microscopy (HR-TEM) images, respectively. h-MoO3xhibited well-defined hexagonal prismatic crystals with lengthsf up to ∼30 µm, consisting of single primary particles from theR-TEM image (Figure 1e(ii)). In contrast, nano-h-MoO3 showedarkedly reduced particle sizes in the range of 0.5–2 µm. HR-EMobservation of nano-h-MoO3 (Figure 1e(iv)) further revealedhat these particleswere composed of aggregated, nearly sphericalrimary crystallites with an average size of ∼30 nm. In addition,lear lattice fringes with a spacing of 9.12 Å were observed,orresponding to the (100) plane identified in the XRD patterns.able S2 summarizes the specific surface area values obtainedrom Brunauer–Emmett–Teller (BET) measurements. h-MoO3xhibited a very low specific surface area (<1 m2 g−1), whereasano-h-MoO3 showed a markedly higher specific surface area of5m2 g−1. This substantial increase is primarily attributable to theanoparticulation of h-MoO3, which effectively reduces particleize and increases the exposed active surface area..2 Electrochemical Performance of h-MoO3ositive Electrode in Ca andMg Battery Cellsor the electrochemical evaluation ofCa andMgbattery cellswithano-h-MoO3 positive electrode, discharge-charge cycling testsere conducted, highlighting the markedly enhanced capacitynd reversibility. The tests were performed using coin-type cellsomposed of a Ca orMg foil negative electrode and the electrolytef 0.3mol dm−3 Ca[B(hfip)4]2 inmonoglyme (G1) or 0.3mol dm−3g[B(hfip)4]2 in diglyme (G2) (hfip: hexafluoroisopropoxyl).igure S5 presents the voltage profiles of Ca and Mg batteryells employing h-MoO3 positive electrode. The Ca and Mgattery cells delivered discharge capacities of 40 and 52 mAh g−1,espectively, which are significantly lower than the theoreticalapacity of 372 mAh g−1 (MoO3 + M2+ + 2e− → MMoO3, M: Caof 12or Mg). This limited capacity can be attributed to the restrictedelectrochemically active surface area, as h-MoO3 consists ofmicron-sized, well-ordered particles. This morphology is evidentin the SEM image of the electrode (Figure S6a), where the whiteregions correspond to h-MoO3 particles and the surroundingblack regions to acetylene black. In contrast, Ca and Mg bat-tery cells employing nano-h-MoO3 positive electrode exhibitedmarkedly enhanced capacities, as shown in Figure 2a,b. Thisimprovement is attributed not only to the significantly increasedelectrochemically active surface area but also to the shortenedsolid-state diffusion length, both of which result from particlesize reduction at the secondary particle level (Figure 1e(iii) andFigure S6b) and, more importantly, at the primary crystallite level(Figure 1e(iv)).In the Ca battery cell with the nano-h-MoO3 positive electrode(Figure 2a), the initial discharge capacity reached ∼195 mAh g−1.However, significant capacity fade occurred, with the dischargecapacity decreasing to ∼70 mAh g−1 by the fifth cycle. This rapiddegradation is primarily attributed to the formation of passivationlayers (e.g., CaF2 and CaO) on the negative electrode duringcycling, as widely reported in previous studies [34–36]. Theseinsulating layers strongly hinder Ca stripping/plating, leadingto large overpotential and severe capacity loss in the full cell,particularly during charging. The corresponding dQ/dV plots upto the second cycle revealed discharge peaks at ∼1.8 V and chargepeaks at ∼3.5 V.On the other hand, the Mg battery cell delivered an initial dis-charge capacity of ∼130 mAh g−1, which is lower than Ca batterycell (Figure 2b). Although the first charge capacity was lowerthan the discharge capacity, both discharge and charge capacitiesgradually increased upon cycling, with the coulombic efficiencyapproaching nearly 100% after several cycles. Notably, the voltageprofiles retained their overall shape even after repeated cycling.The correspondingdQ/dVplots revealed discharge peaks at∼1.1 Vand charge peaks at ∼3.0 V, with negligible peak shifts duringcycling, indicating that the electrochemical reactions were highlyreversible.To summarize the electrochemical performance of both batteries,the Mg battery cell exhibited markedly higher reversibility,primarily owing to the excellent interfacial compatibility betweenthe electrolyte and the Mg metal electrode. In contrast, the Cabattery cell showed higher capacity (195 mAh g−1, compared with130 mAh g−1 for the Mg battery cell), implying that Ca2+ exhibitshigher mobility than Mg2+ inside nano-h-MoO3 structure.To quantify the amount of divalent-ion insertion and de-insertionand to clarify the reactionmechanism of nano-h-MoO3 electrodesin Ca and Mg battery cells, inductively coupled plasma opticalemission spectrometry (ICP-OES), SEM-EDX, and TEM-EDXanalyses (Table S3 and Figure S7) were conducted. These elemen-tal analyses confirmed that the large amounts of Ca2+ and Mg2+insertion and de-insertion. However, the quantified values weresmaller than those estimated from the electrochemical charge–discharge profiles, suggesting the presence of side reactions thatcontribute to the measured capacities. According to previousstudies, side reactions frequently occur and can substantiallyaffect the measured capacity, particularly prominent in Mgbattery systems paired with oxide-type positive electrodes. AAdvanced Energy Materials, 2026tive Commons LicenseFIGURE 2 Voltage profiles and corresponding dQ/dV plots of nano-h-MoO3 positive electrode in (a) Ca battery cell (electrolyte water content is200 ppm) and (b)Mg battery cell (electrolytewater content is 20 ppm) at 0.1 C rate (37mAg−1), (c) Lattice parameter and volume changes of nano-h-MoO3during cycling in Ca andMg cells calculated by Rietveld refinement, (d) Volume expansion rate of nano-h-MoO3 in Ca andMg battery cells (e) HR-TEMimages of nano-h-MoO3 after discharge/charge in Ca and Mg battery cells. The colored arrows indicate the electrochemical states (red: discharge, blue:charge), and the numbers denote the corresponding amounts of inserted/de-inserted ions in each state. (f) TEM-EDX analysis of discharged nanoh-MoO3 from Ca and Mg battery cells.p[[fg[tHMaciA 16146840, 2026, 25, Downloaded from https://advanced.onlinelibrary.wiley.com/doi/10.1002/aenm.71006 by Toshihiko Mandai - National Institute For , Wiley Online Library on [07/07/2026]. 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 Crealausible explanation involves contributions from H+ insertion37] and the formation of a cathode–electrolyte interface (CEI)38, 39]. In the case of H+ insertion, protons may originaterom trace H2O inherently present in the electrolyte or may beenerated via ether decomposition on the Mg negative electrode40]. Note that the residual H+ species originally present inhe tunnels may, in principle, be electrochemically de-inserted.owever, their amount is limited to approximately 0.022 pero, corresponding to only ∼4 mAh g−1, and therefore theyre unlikely to be the main source of the side-reaction-relatedapacity. Additional factors may also contribute to side reactions,ncluding corrosion of the current collector [37] and dissolutiondvanced Energy Materials, 2026of active materials into the electrolyte [41], both of which arestrongly influenced by the water content in the electrolyte.Figure S8 presents the cycling performance of the cell. The nano-h-MoO3 electrode exhibited excellent cyclability, maintaininga stable capacity of 73 mAh g−1 for up to 100 cycles. Thiscapacity retention is remarkable when compared with other Mgbattery cells employing the same non-corrosive [B(hfip)4]−-basedelectrolyte but different oxide-type positive electrodes, whichgenerally show much lower cycle life—for example, ∼20 cyclesfor nano tunnel-type MnO2 (hollandite or romanechite) [42],∼10 cycles for nano Cu─Mn spinel oxide [19], and ∼10 cycles5 of 12tive Commons LicenseftcTceitethwacaSw2mlaSMwfecoSMfcdoMadatvcc(FMbwrTmabibawbCl6 16146840, 2026, 25, Downloaded from https://advanced.onlinelibrary.wiley.com/doi/10.1002/aenm.71006 by Toshihiko Mandai - National Institute For , Wiley Online Library on [07/07/2026]. 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 Creaor amorphous Mg─Li─Ti–Mo oxide [43]. However, similar tohe preceding works, capacity fade and a gradual decrease inoulombic efficiency were also observed here.o explore the origin of this capacity fade in the Mg batteryell with a nano-h-MoO3 positive electrode and support thestimated side reaction, the water content in the electrolyte wasntentionally increased from 20 to 200 ppm. Figure S9a showshe voltage profile of the Mg battery cell with the nano-h-MoO3lectrode using an electrolyte containing 200 ppm water. Inhe first cycle, both the discharge and charge capacities wereigher than those obtainedwith an electrolyte containing 20 ppmater. However, upon further cycling, the operating dischargend charge voltages gradually shifted to a lower position, and theoulombic efficiency dramatically decreased significantly due ton excess of charge capacity relative to discharge capacity. Figure9b compares theMg/Mo ratios determined by elemental analysisith those calculated from electrochemical capacities. With the00 ppm H2O-electrolyte, the fraction of Mg contributing to theeasured capacity was dramatically reduced, indicating that aarge portion of the capacity arose from side reactions. XRDnalysis of the nano-h-MoO3 electrodes after 10 cycles (Figure9c) revealed that the diffraction peaks corresponding to the h-oO3 phase had almost completely disappeared when cycledith the 200 ppm water-containing electrolyte. In contrast,or the cell with 20 ppm water, the h-MoO3 phase was stillvident after 10 cycles, although with weaker peak intensitiesompared to the pristine state. Furthermore, SEM–EDX analysisf the GF/C separator on the positive electrode side (Figure9d) revealed the homogeneous distributions of Mo along withg and F elements, where Mg and F are supposed to berom Mg[B(hfip)4]2 salt, after 10 cycles in the 200 ppm water-ontaining electrolyte. These analyses strongly confirmed theissolution of h-MoO3 into the electrolyte. This dissolutionf MoO3 leads to the loss of the higher-potential, reversibleg2+ insertion reaction. The dissolved Mo species subsequentlycts as a redox shuttle, promoting irreversible decomposition,issolution, and precipitation reactions that occur predominantlyt low potentials and induce significant polarization. As a result,he charge–discharge profiles progressively shift toward loweroltages, as observed in Figure S9a. This side-reactionmechanisman also account for the capacity fade observed in Mg batteryells even when the electrolyte contains as little as 20 ppm waterFigure S8).igure S10(i, ii) shows the ex situ XRD patterns of nano-h-oO3 electrodes after discharge and charge in Ca and Mgattery cells. No apparent peak shifts or phase transformationsere observed in either system, indicating that nano-h-MoO3emains electrochemically stable over a wide potential window.o further investigate subtle structural changes, Rietveld refine-ent was performed on the nano-h-MoO3 electrodes before andfter cycling (Figure 2c). During the first discharge in bothattery systems, the lattice constants a and c exhibited slightncreases, resulting in only minor lattice volume expansions (Caattery: 1.79%; Mg battery: 1.09%). These small expansions can bettributed to the reduction of Mo6+ to Mo5+, which is consistentith the Mo K-edge XANES negative shifts shown in Figure S11a,. Notably, the lattice expansion was more pronounced in thea battery cells than in the Mg battery cells, primarily becausearger Ca insertion induced a greater extent of Mo6+ reduction toof 12Mo5+. In addition, ion-specific local structural changes associatedwith Ca2+ and Mg2+ insertion, which will be discussed in detaillater, also contribute to this difference. Upon the first charge,the lattice parameters nearly returned to their pristine values,demonstrating the high structural robustness of nano-h-MoO3.These ultra-low-strain Ca2+ and Mg2+ insertion properties ofnano-h-MoO3 are particularly remarkable, especially consideringits role as a host material for high charge-density carriers suchas Mg2+. Compared with other reported divalent-ion insertionmaterials summarized in Figure 2d (experimental data) and TableS4 (experimental and computational data), nano-h-MoO3 exhibitsa favorable combination of relatively high reversible capacityand exceptionally small lattice volume change. Although highercapacities in many reported materials are often accompaniedby larger structural expansion, nano-h-MoO3 deviates from thisgeneral trend, highlighting the robust nature of the h-MoO3tunnel framework.For instance, the volume change of Ca insertion is even smallerthan that of typical polyanion-type positive electrodes, which arewell known for their structural robustness. Furthermore, nano-h-MoO3 demonstrates only ∼1–2% volume expansion even after10 cycles in Mg battery cells, highlighting its superior structuralstability compared to conventional divalent-ion electrode mate-rials. Consistent with this behavior, no apparent phase changesof nano-h-MoO3 were observed in either system from HR-TEMobservation (Figure 2e). The lattice fringes corresponding to the(100) plane remained almost unchanged after discharge andcharge, with only a slight expansion detected upon discharge.Furthermore, TEM-EDX mapping (Figure 2f) revealed that Mgand Ca were distributed almost homogeneously within theprimary particles of the discharged electrodes. Note that Mg wasnot uniformly distributed to the particle center because its smalleroverall insertion amount compared to Ca, which resulted inpartially unreacted regions. Furthermore, Figure S12 presents thecyclic voltammetry (CV) curves of the Mg battery two-electrodecell, providing additional evidence for the Mg intercalationreaction. As shown in Figure S12b, the well-defined cathodic andanodic peaks originate predominantly from diffusion-controlledprocesses, with minimal contribution from (pseudo)capacitivebehavior.With increasing sweep rate, the (pseudo)capacitive con-tribution becomes more pronounced (Figure S12c). In contrast,at lower sweep rates, the electrochemical response is dominatedby diffusion-controlled reactions, indicating that Mg2+ storagein nano-h-MoO3 primarily proceeds via solid-state intercalationrather than surface-controlled processes. Taken together, theabsence of significant lattice disruption and the uniform distribu-tion of inserted ions demonstrate that nano-h-MoO3 undergoesa stable intercalation-type reaction, rather than a conversion orsurface-captive mechanism.To clarify the origin of the ultra-low-strain intercalation behaviorof Ca2+ and Mg2+, it is necessary to identify their insertion sitesand diffusion pathways within the h-MoO3 framework. However,the large hexagonal tunnels and the complexity of divalent cationcoordination environments make purely experimental deter-mination impractical. Therefore, in the following section, weemploy density functional theory (DFT) combined with nudgedelastic band (NEB) calculations to predict the stable insertion sitesand diffusion pathways of Ca2+ and Mg2+, providing mechanisticinsight into their distinct intercalation behaviors.Advanced Energy Materials, 2026tive Commons LicenseFIGURE 3 (a) Stable Ca (left) and Mg (right) insertion sites in h-MoO3 crystal, (b) Calculated voltage profile with the Ca/Mg insertion with theshaded region indicating the experimental range for cation insertion, (c) Diffusion pathway for the intercalant, Ca(left) andMg(right) with the diffusingintercalant atoms superimposed into one structure to illustrate the diffusion path, (d) Energetics of intercalant diffusion calculated using NEB.2DToeStd(tmtsiwcilMCTt(TA 16146840, 2026, 25, Downloaded from https://advanced.onlinelibrary.wiley.com/doi/10.1002/aenm.71006 by Toshihiko Mandai - National Institute For , Wiley Online Library on [07/07/2026]. 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 Creat.3 Computational Prediction of Ca2+ andMg2+iffusion in h-MoO3o predict the Mg and Ca diffusion pathways inside the tunnelf the h-MoO3 structure, various potential insertion sites werexamined to evaluate the stability of Ca/Mg insertion (Figures13–S15). The inserted ions (M), Ca and Mg, were both foundo stabilize by anchoring to four framework oxygen atoms atistances of 2.25–2.47 Å for Ca─O and 1.95–2.10 Å for Mg─OFigure 3a). The inserted ion resides off-axis relative to theunnel center, giving an asymmetric four-coordinated environ-ent. Additionally, three energetically equivalent sites exist inhe hexagonal tunnel framework, and another three equivalentites on the next layer. Bond-valence analysis of the relaxedntercalated cell identifiesM asM2+, yielding amixedMo valence,ith twoMo5+ sites and the remaining Mo as Mo6+. This satisfiesharge neutrality per M2+ and indicates that the two electronsntroduced byCa localize on nearbyMo centers. The shorter bondength of Mg─O compared to Ca─O results in the ability of theg2+ to embed deeper into the tunnel framework compared toa2+.he theoretical voltage was calculated by progressively addinghe intercalant atoms into the structurally relaxed h-MoO3 systemFigure 3b).he voltage can be obtained as,V = −𝐸discharged − 𝐸charged𝜂dvanced Energy Materials, 2026whereEcharged,Edischarged, and η refer total energy of the constituentbattery components in the charged/discharged state and thenumber of electrons transferred, respectively [44]. Here, wheninserted cations (M) are inserted.V = −𝐸𝑀𝑥+𝑛𝑀𝑜𝑂3− 𝐸𝑀𝑥𝑀𝑜𝑂3− 𝑛𝐸𝑀nZwhere 𝐸𝑀𝑥MoO3and 𝐸𝑀𝑥+𝑛MoO3 are the energies of the charged anddischarged positive electrodes, EM is the energy per atom of thepure metal anode (Ca/Mg), and Z is the valency of the insertedcation (+2 for Ca and Mg), and n number of ions inserted [44].To ensure conformity to thermodynamic principles, this equationcan only be applied to those cation insertion concentrationsresiding on a convex hull of formation energies. Ca insertion intoh-MoO3 results in an initial voltage of 2.43 V, whereas it provides amuch lower 1.76V forMg insertion.Asmore intercalant atoms areadded, the voltage in CaxMoO3 drops to 1.63 V, whereasMgxMoO3is driven to 1.25 V.Then, the migration energy barriers for the inserted atoms werecalculated tomove through the tunnel structure of the h-MoO3, asshown in Figure 3c. The Ca2+ diffuses as an off-axis ‘double hop’.In the initial lattice, the adjacent site Ca2+ is fourfold coordinatedwith O atoms. During migration, it stays close to the tunnelframework and moves along a shallow arc, successively breakingand forming Ca─O bonds with adjacent oxygens in two well-defined hops (Figure S16a). At the saddle point, the coordinationdecreases transiently to 3 weak O contacts, and Ca passes a localbottleneck defined by opposing O rows. Ca then re-establishes7 of 12ive Commons Licensefbce0tTwirMtCtadt2NTbitclwEcFioisMtatcpse(nierittMaaidb8 16146840, 2026, 25, Downloaded from https://advanced.onlinelibrary.wiley.com/doi/10.1002/aenm.71006 by Toshihiko Mandai - National Institute For , Wiley Online Library on [07/07/2026]. 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 Creaour short Ca─O bonds at the neighboring wall adjacent siteefore repeating the secondhop. The trajectory traced in Figure 3corresponds to this wall-hugging path, and the calculated nudgedlastic band (NEB) profile in Figure 3d gives a low barrier of about.270 eV for this hop, consistent with facile diffusion of Ca alonghe tunnels.he Mg ion diffuses in a similar manner to Ca ion diffusion,ith a migration barrier for a single hop of 0.745 eV, as shownn Figure 3d. Comparatively, the significant energy differenceelative to Ca is attributed to the more inward embedding of theg ion into the contact oxygens in the intermediate layer due tohe shorter Mg─O bond lengths (2.86 Å). The longer length ofa─O seems to enable the Ca ions to glide along the wall of theunnelmore easily compared toMg,which gets trapped (Figure 3cnd Figure S16b). Therefore, Ca ion shows considerably easieriffusion enabled by the ability to glide more easily along theunnel framework..4 Local and Tunnel Structural Changes inano-h-MoO3 Induced by Ca2+ andMg2+ Insertiono elucidate the origin of the ultra-low-strain discharge/chargeehavior of nano-h-MoO3, it is crucial to identify the preferrednsertion sites and diffusion pathways of Mg2+ and Ca2+ withinhe hexagonal tunnels, which were clarified by DFT and NEBalculations, and to correlate them with experimentally observedocal structural changes. Here, we first employed Mo K-edgeavelet transform-extended X-ray absorption fine structure (WT-XAFS), which is highly sensitive to local coordination and canapture subtle bond-length variations.igure 4a shows the WT-EXAFS maps of the pristine state, Ca-nserted state, and Mg-inserted state of nano-h-MoO3. Basedn the previous report [45, 46], the dotted lines were set tondicate the Mo─O correlation (from 1.1 to 1.7 Å) in the pristinetate, and spectral peaks near ∼3.0 Å correspond to closedo─Mo interactions (Figure 4a(i)). Upon Ca and Mg insertion,he strong spectral intensity of the Mo─O bond length aroundll k values (2–11 Å−1) and 1.2 Å in the pristine state shiftedoward the lower radial distance region, outside of the Mo─Oorrelation window in the pristine state (Figure 4a(ii,iii)). Inarticular, this change is clearer with Ca insertion, indicating ahortening of the Mo─O bond. In contrast, the Mg-inserted statexhibits noticeable intensity around ∼1.8 Å at higher k values∼8 Å−1), indicative of elongated Mo─O components. Moreover,ew spectral peaks emerge at higher k values near ∼2.0 Å,mplying the formation of local Mo─Mg-associated coordinationnvironments. RegardingMo─Mo interactions, only Ca insertionesults in a clear shift toward shorter radial distances, as shownn Figure 4a(ii), whereas no appreciable change is observed forhe Mg-inserted state. Overall, the WT-EXAFS results indicatehat Ca insertion predominantly induces contraction of botho─O and Mo─Mo distances, while Mg insertion leads tomixed response characterized by partial Mo─O shorteningccompanied by more prominent bond elongation, with minimalmpact on Mo─Mo distances. Importantly, all observed radial-istance variations are highly reversible upon charging in bothattery systems, as shown in Figure S17.of 12Raman spectroscopy further supports the distinct local responsesto Ca and Mg insertion (Figure S18). No measurable shift in theMo═O vibrational mode, corresponding to the Mo─O bond tothe isolated oxygen atom located inside the hexagonal tunnel,was observed for Ca-inserted nano-h-MoO3, whereas a distinctred shift appeared in the discharged state of Mg-inserted nano-h-MoO3, indicating bond softening and elongation of the Mo─Obond within the tunnel, which is consistent with the presence ofelongatedMo─Ocomponents suggested byWT-EXAFS in theMgbattery system. After charging, the peak position returns close tothat of the pristine electrode, supporting the reversibility of theMg-induced local distortion.To quantitatively assign the bond-length changes suggested by thelocal-probe analyses and to connect them to specific structuralmotifs, Rietveld refinement was performed for nano-h-MoO3electrodes in the pristine, discharged, and charged states for bothCa and Mg battery systems. All XRD patterns of nano-h-MoO3in all the states were well-fitted (Table S5, S6 and Figure S19).These results support that both Ca2+ and Mg2+ are inserted atthe predicted sites and that their insertion induces systematicchanges in the Mo─O and Mo─Mo distances, consistent withthe ion-dependent local structural variations observed in theWT-EXAFS spectra. Using the refined structural parameters,we investigated six distinct Mo─O bond lengths (bond 1–6)within the Mo─O polyhedral unit. In the nano-h-MoO3 frame-work, these bonds are equivalent within the structural unit;however, they respond differently upon divalent-ion insertion.As shown in Figure 4b, bonds 1 and 2 (along the c axis)remain nearly unchanged after either Ca2+ or Mg2+ insertion.In contrast, bond 6, which is the Mo-isolated O bond (i.e., theoxygen atom located inside the tunnel), exhibits pronouncedion-dependent variation: it elongates upon Mg2+ insertion butshortens upon Ca2+ insertion. Given that the uncertainty inbond lengths obtained from Rietveld refinement is less than0.02 Å, these changes are sufficiently large to be consideredsignificant. These results indicate that divalent cations residinginside the tunnel strongly modulate the position of the isolatedoxygen, thereby governing the local distortion mode of the MoO6polyhedral unit. To further confirm these trends, Mo─O bondlengths shorter than the shortest bond in the pristine statewere classified as “short bonds,” whereas those longer than thelongest pristine bond were classified as “long bonds,” followingthe same criteria used to define the dotted lines indicatingthe Mo─O correlation window in the WT-EXAFS spectra. Ca2+insertion increased the number of bonds in the short-bond region,reducing the average Mo─O distance from 2.08 to 1.96 Å. Incontrast, the Mo─O bond lengths became widely distributedacross the short–to–long range, including 1/3 of the numbersin the “long bond” region, resulting in an increased averageMo─O distance from 2.08 to 2.21 Å. These Mo─O bond-lengthtrends are in good agreement with those observed in the WT-EXAFS spectra and are found to be predominantly governedby changes in the Mo-isolated O bond. Figure 4c,d showsthe Mo─M (M═Ca or Mg) and Mo─Mo bond lengths. Theshortest Mo─Mg distances were smaller than the correspondingMo─Ca distances. Additionally, only Ca2+ insertion caused acontraction of the nearest Mo─Mo distance, from 3.31 to 3.15Å. This shorter length change was clearly observed in WT-EXAFS, while the Mo─Mo length was not changed with Mginsertion.Advanced Energy Materials, 2026tive Commons LicenseFIGURE 4 (a) Mo K-edge k3-weighted WT-EXAFS of nano h-MoO3 (i) pristine state, (ii) Mg inserted state, (iii) Ca inserted state. Local structuralchange of nano-h-MoO3 by Rietveld refinement for Ca andMg insertion (b) Mo─O bond length, (c) Mo─Ca, Mg bond length, (d) Shortest Mo─Mo bondlength.TpcirsattmoMcrbbuTrpclA 16146840, 2026, 25, Downloaded from https://advanced.onlinelibrary.wiley.com/doi/10.1002/aenm.71006 by Toshihiko Mandai - National Institute For , Wiley Online Library on [07/07/2026]. 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 Creo capture the overall impact of local distortions in the MoO6olyhedral units on the tunnel framework, the O- and Mo-onstituted hexagonal units and the corresponding Mo─O (Mo-solated O) bond lengths were investigated through Rietveldefinement as illustrated in Figure 5a. As shown in Figure 5b,hortening of the Mo-isolated O bond upon Ca2+ occupationt the three equivalent insertion sites leads to expansion ofhe O-hexagonal area, which increases by 53.1%. In addition,he Mo-hexagonal framework expands slightly by 5.9%. Thisodest lattice expansion is attributed to the larger extentf Mo6+ reduction induced by Ca2+ insertion compared withg2+, as supported by the XANES results (Figure S11). Inontrast, Mg2+ insertion results in a distinctly different structuralesponse. As shown in Figure 5c, the O-hexagonal area contractsy 46.5%, accompanied by elongation of the Mo-isolated Oond, while the Mo-hexagonal framework remains essentiallynchanged.his opposite behavior is possibly attributed to the larger ionicadius of Ca2+ (1.00Å) compared toMg2+ (0.72Å), and to theDFT-redicted diffusionmode inwhichCa2+ diffuseswithin the tunnelavity rather than embedding into the framework as Mg2+ does,eading to expansion of the O-hexagonal tunnel (Figure S20).dvanced Energy Materials, 2026After charging, the structural parameters largely returned to theirpristine values, indicating reversible Ca2+ and Mg2+ insertion.The Mg2+ and Ca2+ insertion sites predicted by DFT and NEBcalculations show excellent agreement with the experimentalobservations, and multiple crystallographic analyses consistentlyvalidate the local structural evolution of nano-h-MoO3. To thebest of our knowledge, no other divalent-ion insertion system hasbeen reported in which reversible ion insertion and de-insertionproceed exclusively through modulation of host metal–oxygenbond lengths while preserving an intact Mo framework, assummarized in Figure 5d. In conventional zero-strain lithium-and sodium-ion insertion materials [47, 48], zero-strain is oftenachieved by incorporating rigid pillars or redox-inactive metalspecies into the framework to suppress lattice distortion inducedby guest-ion insertion. In contrast, the divalent-ion insertionsystem in this study represents a fundamentally different strategy:lattice distortion is mitigated through dynamic and reversibleadjustment of host metal–oxygen bond lengths rather than bystatic structural reinforcement. This highly unusual charge–discharge mechanism enables ultralow lattice distortion duringcycling and represents an unprecedented approach to achievingstructurally resilient oxide hosts for divalent metal batteries.9 of 12ative Commons LicenseFIGURE 5 (a) The crystal structure of nano-h-MoO3 illustrating the Mo and O hexagonal tunnel frameworks and Mo-isolated O bond. Theirchanges of nano-h-MoO3 by Rietveld refinement in (b) Ca and (c) Mg battery cell, (d) Summary of tunnel framework change of h-MoO3 for both batterysystems.3IotigreectcdbbrlorACT1 16146840, 2026, 25, Downloaded from https://advanced.onlinelibrary.wiley.com/doi/10.1002/aenm.71006 by Toshihiko Mandai - National Institute For , Wiley Online Library on [07/07/2026]. 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 CreaConclusionn this study, we approached material design from the viewpointf structural robustness, in addition to nanoparticulation ofhe oxide positive electrode, for room-temperature operationn divalent metal batteries. By focusing on h-MoO3 with itseometrically stable hexagonal tunnel structure, we achievedeversible Ca and Mg insertion in full cells using nano-h-MoO3lectrodes. In Mg battery systems, where the Mg metal electrodexhibits excellent stability, the nano-electrode delivered superioryclability up to 100 cycles. Moreover, comprehensive struc-ural analyses revealed minimal lattice distortion (<2% volumehange) and a unique mode of local structural evolution duringivalent-ion (de)insertion, both of which were further supportedy computationalmodeling. This reversiblemodulation ofMo─Oond lengths while preserving an intact Mo-based frameworkepresents a previously unreported insertion mechanism. Col-ectively, our findings highlight a new design principle forxide-type positive electrodes, paving the way toward practicaloom-temperature divalent batteries.uthor Contributionsonceptualization: R.I.; Methodology: R.I.; Formal Analysis: R.I. (lead),.Y, H.K; Investigation: R.I. (synthesis, Mg battery, Ca battery), H.P. (cal-0 of 12culation), M.S. (calculation); Resources: T.M. (Mg electrolyte), H.K. (Caelectrolyte); Writing – Original Draft: R.I.; Writing – Review & Editing:H.K. (lead), T.M.,R.Z., S.O., M.P., T.Y., A.N., A.M., M.M., I.H., H.P., M.S.;Supervision: H.K., T.M., M.S., S.O., I.H; Project Administration: R.I., H.K;Funding Acquisition: R.I., H.K.AcknowledgementsParts of this work were supported by JSPS KAKENHI (23KJ0214,23K13816), the Light Metal Educational Foundation. Computationalwork was performed using the Spiedie HPC at Binghamton Universityand Expanse, a part of the Advanced Cyberinfrastructure CoordinationEcosystem: Services & Support (ACCESS) program, which is supportedby the National Science Foundation under Grants No. 2138259, No.2138286, No. 2138307, No. 2137603, and No. 2138296 under allocationTGDMR180009 [49]. We gratefully acknowledge Prof. Akira Miura atHokkaido University for his support of the synchrotron PDF measure-ments. The synchrotron radiation experiments were performed at theBL13XU of SPring-8 with the approval of the Japan Synchrotron RadiationResearch Institute (JASRI) (proposal no. 2024B1777). Saneyuki Ohnogratefully acknowledges the financial support from Toyota Physical andChemical Research Institute through the Rising Fellow Program.Conflicts of InterestThe authors declare no conflicts of interest.Data Availability StatementThe data that support the findings of this study are available from thecorresponding author upon reasonable request.Advanced Energy Materials, 2026tive Commons LicenseRRo2A&c3Ma(4Pic5a16o27tL8Rd9ER(1Os01OE11sR11Sf81cA11a(1Rth1LBaA 16146840, 2026, 25, Downloaded from https://advanced.onlinelibrary.wiley.com/doi/10.1002/aenm.71006 by Toshihiko Mandai - National Institute For , Wiley Online Library on [07/07/2026]. 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 Creferences1. J. M. Tarascon and M. Armand, “Issues and Challenges Facingechargeable Lithium Batteries,” Nature 414 (2001): 359–367, https://doi.rg/10.1038/35104644.. H. D. Yoo, I. Shterenberg, Y. Gofer, G. Gershinsky, N. Pour, and D.urbach, “Mg Rechargeable Batteries: An On-Going Challenge,” EnergyEnvironmental Science 6 (2013): 2265–2279, https://doi.org/10.1039/3ee40871j.. J. Muldoon, C. B. Bucur, and T. Gregory, “Fervent Hype Behindagnesium Batteries: An Open Call to Synthetic Chemists—Electrolytesnd Cathodes Needed,” Angewandte Chemie International Edition 562017): 12064–12084, https://doi.org/10.1002/anie.201700673.. M. E. Arroyo-de Dompablo, A. Ponrouch, P. Johansson, and M. R.alacín, “Achievements, Challenges, and Prospects of Calcium Batter-es,” Chemical Reviews 120 (2020): 6331–6357, https://doi.org/10.1021/acs.hemrev.9b00339.. A. Ponrouch, C. Frontera, F. Bardé, and M. R. Palacín, “TowardsCalcium-Based Rechargeable Battery,” Nature Materials 15 (2016):69–172, https://doi.org/10.1038/nmat4462.. M. Matsui, “Study on Electrochemically Deposited MgMetal,” Journalf Power Sources 196 (2011): 7048–7055, https://doi.org/10.1016/j.jpowsour.010.11.141.. S. D. Pu, C. Gong, X. Gao, et al., “Current-Density-Dependent Elec-roplating in Ca Electrolytes: From Globules to Dendrites,” ACS Energyetters 5 (2020): 2283–2290, https://doi.org/10.1021/acsenergylett.0c01153.. D. Aurbach, Z. Lu, A. Schechter, et al., “Prototype Systems forechargeable Magnesium Batteries,”Nature 407 (2000): 724–727, https://oi.org/10.1038/35037553.. Z. Li, B. P. Vinayan, P. Jankowski, et al., “Multi-Electron Reactionsnabled by Anion-Based Redox Chemistry for High-Energy Multivalentechargeable Batteries,” Angewandte Chemie International Edition 592020): 11483–11490, https://doi.org/10.1002/anie.202002560.0. M. Rashad,M. Asif, I. Ahmed, et al., “Quest for Carbon and Vanadiumxide Based Rechargeable Magnesium-Ion Batteries,” Journal of Magne-ium and Alloys 8 (2020): 364–373, https://doi.org/10.1016/j.jma.2019.09.10.1. I. D. Johnson, B. J. Ingram, and J. Cabana, “The Quest for Functionalxide Cathodes for Magnesium Batteries: A Critical Perspective,” ACSnergy Letters 6 (2021): 1892–1900, https://doi.org/10.1021/acsenergylett.c00416.2. H. Kobayashi, Y. Fukumi, H. Watanabe, et al., “Ultraporous, Ultra-mall MgMn2O4 Spinel Cathode for a Room-Temperature Magnesiumechargeable Battery,” ACS Nano 17 (2023): 3135–3142, https://doi.org/10.021/acsnano.2c12392.3. R. Yokozaki, H. Kobayashi, T. Mandai, and I. Honma, “Effect of Alubstitution on Structure and Cathode Performance of MgMn2O4 Spinelor Magnesium Rechargeable Battery,” Journal of Alloys and Compounds72 (2021): 159723, https://doi.org/10.1016/j.jallcom.2021.159723.4. T. S. Arthur, R. Zhang, C. Ling, et al., “Understanding the Electro-hemical Mechanism of K-αMnO2 for Magnesium Battery Cathodes,”CS Applied Materials & Interfaces 6 (2014): 7004–7008, https://doi.org/0.1021/am5015327.5. R. Zhang and C. Ling, “Unveil the Chemistry of Olivine FePO4s Magnesium Battery Cathode,” ACS Applied Materials & Interfaces 82016): 18018–18026, https://doi.org/10.1021/acsami.6b03297.6. T. Hatakeyama, H. Li, N. L. Okamoto, et al., “Accelerated Kineticsevealing Metastable Pathways of Magnesiation-Induced Transforma-ions in MnO2 Polymorphs,” Chemistry of Materials 33 (2021): 6983–6996,ttps://doi.org/10.1021/acs.chemmater.1c02011.7. K. Shimokawa, T. Atsumi, N. L. Okamoto, et al., “Structure Design ofong-Life Spinel-Oxide Cathode Materials for Magnesium Rechargeableatteries,”AdvancedMaterials 33 (2021): 2007539, https://doi.org/10.1002/dma.202007539.dvanced Energy Materials, 2026ea18. B. J. Kwon, L. Yin, H. Park, et al., “High Voltage Mg-Ion Bat-tery Cathode via a Solid Solution Cr–Mn Spinel Oxide,” Chemistry ofMaterials 32 (2020): 6577–6587, https://doi.org/10.1021/acs.chemmater.0c01988.19. R. Iimura, H. Watanabe, T. Mandai, I. Honma, H. Imai, and H.Kobayashi, “AnElectrically Conductive CuMn2O4 Ultrananospinel Cath-ode for Room-Temperature Magnesium Rechargeable Batteries,” ACSApplied Energy Materials 7 (2024): 5308–5314, https://doi.org/10.1021/acsaem.4c01211.20. T. R. Juran, J. Young, andM. Smeu, “Density Functional TheoryMod-eling ofMnO2 Polymorphs asCathodes forMultivalent IonBatteries,”TheJournal of Physical Chemistry C 122 (2018): 8788–8795, https://doi.org/10.1021/acs.jpcc.8b00918.21. R. Iimura, S. Kawasaki, T. Yabu, et al., “Ultrasmall α -MnO2 With LowAspect Ratio: Applications to Electrochemical Multivalent-Ion Interca-lation Hosts and Aerobic Oxidation Catalysts,” Small 21 (2025): 2411493,https://doi.org/10.1002/smll.202411493.22. H. Guo, S. Wu, W. Chen, et al., “Hydronium Intercalation EnablesHigh Rate in Hexagonal Molybdate Single Crystals,” Advanced Materials36 (2024): 2307118, https://doi.org/10.1002/adma.202307118.23. T. Mandai, “Critical Issues of Fluorinated Alkoxyborate-Based Elec-trolytes in Magnesium Battery Applications,” ACS Applied Materials &Interfaces 12 (2020): 39135–39144, https://doi.org/10.1021/acsami.0c09948.24. V. Kumar, X. Wang, and P. S. Lee, “Formation of Hexagonal-Molybdenum Trioxide (h-MoO3 ) Nanostructures and Their Pseudoca-pacitive Behavior,” Nanoscale 7 (2015): 11777–11786, https://doi.org/10.1039/C5NR01505G.25. A. Chithambararaj, N. Rajeswari Yogamalar, and A. C. Bose,“Hydrothermally Synthesized h-MoO3 and α-MoO3 Nanocrystals: NewFindings on Crystal-Structure-Dependent Charge Transport,” CrystalGrowth & Design 16 (2016): 1984–1995, https://doi.org/10.1021/acs.cgd.5b01571.26. T. Lindahl Christiansen, E. T. S. Kjær, A. Kovyakh, et al., “StructureAnalysis of Supported Disordered Molybdenum Oxides Using Pair Dis-tribution Function Analysis and Automated Cluster Modelling,” Journalof Applied Crystallography 53 (2020): 148–158, https://doi.org/10.1107/S1600576719016832.27. W. Kong, W. Liu, X. Zheng, and Q. Xu, “Sunlight Driven ReversibleandTunable PlasmonResonance in 2DAmorphousMolybdenumOxide,”Advanced Optical Materials 12 (2024): 2301821, https://doi.org/10.1002/adom.202301821.28. S. Y. Chiam, B. Dasgupta, D. Soler, et al., “Investigating the Stabilityof Defects in MoO3 and its Role in Organic Solar Cells,” Solar EnergyMaterials and Solar Cells 99 (2012): 197.29. L. Barkat, M. Hssein, Z. El Jouad, et al., “Efficient Hole-TransportingLayer MoO3:CuI Deposited by Co-Evaporation in Organic PhotovoltaicCells,” Physica Status Solidi 214 (2017): 1600433.30. C. I. Vargas-Consuelos, M. A. Camacho-López, V. H. Ramos-Sanchez,and O. A. Graeve, “Phase and Morphology Control of Hexagonal MoO3Crystals via Na+ Interactions: A Raman Spectroscopy Study,” The Journalof Physical Chemistry C 127 (2023): 13136–13148, https://doi.org/10.1021/acs.jpcc.3c02821.31. J. V. B. Moura, J. V. Silveira, J. G. da Silva Filho, A. G. SouzaFilho, C. Luz-Lima, and P. T. C. Freire, “Temperature-Induced PhaseTransition in h-MoO3: Stability Loss Mechanism Uncovered by RamanSpectroscopy and DFT Calculations,” Vibrational Spectroscopy 98 (2018):98–104, https://doi.org/10.1016/j.vibspec.2018.07.008.32. X. Guan, Y. Ren, S. Chen, et al., “Charge Separation and StrongAdsorption-Enhanced MoO3 Visible Light Photocatalytic Performance,”Journal of Materials Science 55 (2020): 5808–5822, https://doi.org/10.1007/s10853-020-04418-8.33. R. K. Nitharwal, A. Sahoo, V. Kumar, M. S. R. Rao, T. Dixit,and S. Krishnan, “Spectroscopic Visualization of Polarons and Interva-lence Charge Transfer in MoO3– x Nanostructures Via Defect Engineer-11 of 12tive Commons Licensehttps://doi.org/10.1038/35104644https://doi.org/10.1039/c3ee40871jhttps://doi.org/10.1002/anie.201700673https://doi.org/10.1021/acs.chemrev.9b00339https://doi.org/10.1038/nmat4462https://doi.org/10.1016/j.jpowsour.2010.11.141https://doi.org/10.1021/acsenergylett.0c01153https://doi.org/10.1038/35037553https://doi.org/10.1002/anie.202002560https://doi.org/10.1016/j.jma.2019.09.010https://doi.org/10.1021/acsenergylett.1c00416https://doi.org/10.1021/acsnano.2c12392https://doi.org/10.1016/j.jallcom.2021.159723https://doi.org/10.1021/am5015327https://doi.org/10.1021/acsami.6b03297https://doi.org/10.1021/acs.chemmater.1c02011https://doi.org/10.1002/adma.202007539https://doi.org/10.1021/acs.chemmater.0c01988https://doi.org/10.1021/acsaem.4c01211https://doi.org/10.1021/acs.jpcc.8b00918https://doi.org/10.1002/smll.202411493https://doi.org/10.1002/adma.202307118https://doi.org/10.1021/acsami.0c09948https://doi.org/10.1039/C5NR01505Ghttps://doi.org/10.1021/acs.cgd.5b01571https://doi.org/10.1107/S1600576719016832https://doi.org/10.1002/adom.202301821https://doi.org/10.1021/acs.jpcc.3c02821https://doi.org/10.1016/j.vibspec.2018.07.008https://doi.org/10.1007/s10853-020-04418-8ia3aBd3“I33wC3IRh3DCOe3“Ba4oL4iM54“Mh4Em04IfJ14“ES14AIE4tId4oEh1 16146840, 2026, 25, Downloaded from https://advanced.onlinelibrary.wiley.com/doi/10.1002/aenm.71006 by Toshihiko Mandai - National Institute For , Wiley Online Library on [07/07/2026]. 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 applicableng,” ACS Materials Letters 7 (2025): 1195–1202, https://doi.org/10.1021/csmaterialslett.4c01429.4. Z. Li, O. Fuhr, M. Fichtner, and Z. Zhao-Karger, “Towards Stablend Efficient Electrolytes for Room-Temperature Rechargeable Calciumatteries,” Energy & Environmental Science 12 (2019): 3496–3501, https://oi.org/10.1039/C9EE01699F.5. Z. Zhao-Karger, Y. Xiu, Z. Li, A. Reupert, T. Smok, and M. Fichtner,Calcium-Tin Alloys as Anodes for Rechargeable Non-Aqueous Calcium-on Batteries at Room Temperature,” Nature Communications 13 (2022):849, https://doi.org/10.1038/s41467-022-31261-z.6. D. Bier, Z. Li, S. Klyatskaya, et al., “Long Cycle-Life Ca Batteriesith Poly(anthraquinonylsulfide) Cathodes and Ca−Sn Alloy Anodes,”hemsuschem 16 (2023): 202300932.7. R. Zhu, T. Yabu, C. Yang, et al., “Decoding Cathode-Electrolytenterface Issues in Conventional Ethers Electrolytes-Based Magnesiumechargeable Batteries,” Advanced Energy Materials 15 (2025): 2502050,ttps://doi.org/10.1002/aenm.202502050.8. R. Iimura, H. Kobayashi, and I. Honma, “Suppressing Electrolyteecomposition at Cathode/Electrolyte Interface by Mg-Fe Binary Oxideoating towards Room-Temperature Magnesium Rechargeable Batteryperation,” Electrochemistry 90 (2022): 067002, https://doi.org/10.5796/lectrochemistry.22-00045.9. W. Zhou, C. Xu, B. Gao, M. Nakayama, S. Yagi, and Y. Tateyama,Glyme Solvent Decomposition on Spinel Cathode Surface inMagnesiumattery,” ACS Energy Letters 8 (2023): 4113–4118, https://doi.org/10.1021/csenergylett.3c01084.0. Y. Gao, H. Noguchi, and K. Uosaki, “Online Real-Time Detectionf the Degradation Products of Lithium Oxygen Batteries,” ACS Energyetters 8 (2023): 1811–1817, https://doi.org/10.1021/acsenergylett.3c00132.1. X. Ye, H. Li, T. Hatakeyama, et al., “Examining Electrolyte Compatibil-ty on PolymorphicMnO2 Cathodes for Room-Temperature Rechargeableagnesium Batteries,” ACS Applied Materials & Interfaces 14 (2022):6685–56696, https://doi.org/10.1021/acsami.2c14193.2. T. Yabu, R. Iimura, A. Nasu, T. Mandai, M. Matsui, and H. Kobayashi,Romanechite, an Asymmetric Tunnel-Type MnO2, for Rechargeableagnesium Battey Cathodes,” Batteries & Supercaps 8 (2025): 2500118,ttps://doi.org/10.1002/batt.202500118.3. T. Kawaguchi, H. Sakurai, S. Fukui, et al., “AmorphousOxide Cathodenabling Room-Temperature Rechargeable Magnesium Batteries,” Com-unications Materials 6 (2025): 203, https://doi.org/10.1038/s43246-025-0921-0.4. R. Shepard, S. Brennan, T. R. Juran, J. Young, and M. Smeu, “Abnitio Determination of a Simultaneous Dual-Ion Charging Mechanismor Ni0.25Mn0.75O2 Through Redox Reactions of Ni2+/Ni4+ and O2− /O−,”ournal of Materials Chemistry A 10 (2022): 18916–18927, https://doi.org/0.1039/D2TA03938A.5. F. H. Hsu, S. Y. Hsu, B. H. Chen, J. L. Chen, J. M. Chen, and K. T. Lu,Correlation of the Crystal Structure and Ion Storage Behavior of MoO 3lectrode Materials for Aluminum-Ion Energy Storage Studied Using Initu X-Ray Spectroscopy,”Nanoscale 14 (2022): 7502–7515, https://doi.org/0.1039/D2NR00182A.6. A. V. Avani, Z. Li, C. R. Babu, A. V. Soldatov, A. Hossain, and E. I.nila, “Lattice Distortion Suppressed in MoO3 by Incorporating Minormpurities of rGO: Strategy for Enhanced Electrocatalytic Hydrogenvolution,” ACS Applied Energy Materials 8 (2025): 4456–4465.7. L. Yang, G. Liang, H. Cao, et al., “A New Sodium Calcium Cyclote-ravanadate Framework: “Zero-Strain” During Large-Capacity Lithiumntercalation,”Advanced Functional Materials 32 (2022): 2105026, https://oi.org/10.1002/adfm.202105026.8. M.-Y. Shen, J.-S.Wang, Z. Ren, et al., “Quasi-Zero Volume Strain Cath-de Materials for Sodium Ion Battery Through Synergetic Substitutionffect of Li and Mg,” Advanced Functional Materials 33 (2023): 2303812,ttps://doi.org/10.1002/adfm.202303812.2 of 1249. S. Strande, H. Cai, M. Tatineni, et al., “Expanse: Computing WithoutBoundaries: Architecture, Deployment, and Early Operations Experi-ences of a Supercomputer Designed for the Rapid Evolution in Scienceand Engineering,” PEARC ’21: Practice and Experience in AdvancedResearch Computing 2021: Evolution Across All Dimensions 47 (2021):1–4, https://doi.org/10.1145/3437359.3465588.Supporting InformationAdditional supporting information can be found online in the SupportingInformation section.SupportingFile: aenm71006-sup-0001-SuppMat.pdf.Advanced Energy Materials, 2026 Creative Commons Licensehttps://doi.org/10.1021/acsmaterialslett.4c01429https://doi.org/10.1039/C9EE01699Fhttps://doi.org/10.1038/s41467-022-31261-zhttps://doi.org/10.1002/aenm.202502050https://doi.org/10.5796/electrochemistry.22-00045https://doi.org/10.1021/acsenergylett.3c01084https://doi.org/10.1021/acsenergylett.3c00132https://doi.org/10.1021/acsami.2c14193https://doi.org/10.1002/batt.202500118https://doi.org/10.1038/s43246-025-00921-0https://doi.org/10.1039/D2TA03938Ahttps://doi.org/10.1039/D2NR00182Ahttps://doi.org/10.1002/adfm.202105026https://doi.org/10.1002/adfm.202303812https://doi.org/10.1145/3437359.3465588 Ultra-Low-Strain Calcium and Magnesium Ion Storage Enabled by Tunnel-Structured MoO3 Positive Electrode 1 | Introduction 2 | Results and Discussion 2.1 | Synthesis and Characterization of h-MoO3 2.2 | Electrochemical Performance of h-MoO3 Positive Electrode in Ca and Mg Battery Cells 2.3 | Computational Prediction of Ca2+ and Mg2+ Diffusion in h-MoO3 2.4 | Local and Tunnel Structural Changes in Nano-h-MoO3 Induced by Ca2+ and Mg2+ Insertion 3 | Conclusion Author Contributions Acknowledgements Conflicts of Interest Data Availability Statement References Supporting Information