# Fileset

[milan-et-al-2025-filling-the-gaps-in-the-libr-lioh-phase-diagram-a-study-on-the-high-temperature-li3(oh)2br-phase-1.pdf](https://mdr.nims.go.jp/filesets/96f436c6-d14d-4f70-8d48-ab421d8d583d/download)

## Creator

Emily Milan, James A. Quirk, [Kenjiro Hashi](https://orcid.org/0000-0002-0320-4768), [John Cattermull](https://orcid.org/0009-0006-5209-3132), Andrew L. Goodwin, [James A. Dawson](https://orcid.org/0000-0002-3946-5337), [Mauro Pasta](https://orcid.org/0000-0002-2613-4555)

## Rights

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

## Other metadata

[Filling the Gaps in the LiBr-LiOH Phase Diagram: A Study on the High-Temperature Li<sub>3</sub>(OH)<sub>2</sub>Br Phase](https://mdr.nims.go.jp/datasets/1ee412cb-8d04-49ee-94c5-b04bd06e2eff)

## Fulltext

Filling the Gaps in the LiBr-LiOH Phase Diagram: A Study on the High-Temperature Li3(OH)2Br PhaseFilling the Gaps in the LiBr-LiOH Phase Diagram: A Study on theHigh-Temperature Li3(OH)2Br PhaseEmily Milan, James A. Quirk, Kenjiro Hashi, John Cattermull, Andrew L. Goodwin, James A. Dawson,and Mauro Pasta*Cite This: Chem. Mater. 2025, 37, 2899−2906 Read OnlineACCESS Metrics & More Article Recommendations *sı Supporting InformationABSTRACT: In this paper, we build on previous work tocharacterize a phase with stoichiometry Li3(OH)2Br existingbetween ∼225 and ∼275 °C in the LiBr-LiOH phase diagram.Diffraction studies indicate that the phase takes a hexagonal unitcell, and theoretical modeling is used to suggest a possible crystalstructure. Nuclear magnetic resonance spectroscopy and electro-chemical impedance spectroscopy measurements demonstrateexcellent lithium-ion dynamics in this phase, with an ionicconductivity of 0.12 S cm−1 at 250 °C. Initial attempts to stabilizethis phase at room temperature through quenching were notsuccessful. Instead, a metastable state demonstrating poor ionicconductivity is found to form. This is an important considerationfor the synthesis of Li2OHBr solid-state electrolytes (also found inthe LiBr-LiOH phase diagram) which are synthesized by cooling through phase fields containing Li3(OH)2Br, and are hencesusceptible to these impurities.■ INTRODUCTIONRecently, materials in the LiBr-LiOH system have receivedsignificant attention in energy storage research: Li2OHBr as apromising solid-state electrolyte for lithium−metal batteries,and Li4(OH)3Br as a phase change material for thermal energystorage.1−10 Initial work by Scarpa and Hartwig believed theseto be the only compounds existing within the LiBr-LiOHphase diagram;11,12 however, a high-temperature phase withstoichiometry Li3(OH)2Br was mentioned by Reshetnikov in1953.13 No information about the phase is available, and in2000, Sangster chose to omit it from their proposed phasediagram.14In 2022, Mahroug et al. investigated to substantiate claims ofa Li3(OH)2Br phase.9 Crucially, differential scanning calorim-etry (DSC) measurements taken across a range ofcompositions indicated a phase change at ∼230 °C whichhad not previously been identified, with a maximum transitionenthalpy for the composition 67 mol % LiOH, indicating thatthe stoichiometry of the unknown phase lies at this LiBr-LiOHratio. They additionally conducted in situ and ex situ X-raydiffraction (XRD) measurements which showed the formationof new diffraction reflections upon heating samples from theroom-temperature “Li2OHBr + Li4(OH)3Br” phase field,although the validity of these observations is unknown sincetheir starting Li4(OH)3Br has since been shown to be ametastable hydrated phase by Milan et al.15 Nevertheless, thesefindings suggest that there may be a compound at thiscomposition. Based on these findings, Mahroug et al.developed the existing phase diagram to include Li3(OH)2Br.As shown in Figure 1a, these alterations would introduceLi3(OH)2Br as a peritectic phase (liquid + Li4(OH)3Br →Li3(OH)2Br).9Understanding the LiBr-LiOH phase diagram may be criticalto understanding impurity formation in Li2OHBr andsolidification phenomena in Li4(OH)3Br, as well as thepossibility of exciting properties offered by the discovery ofnovel materials. In this paper, the existence of a Li3(OH)2Brphase is confirmed, and the phase is characterized for the firsttime. We use a combination of diffraction studies andmodeling to determine a possible crystal structure. Itslithium-ion dynamics are evaluated through the use of nuclearmagnetic resonance (NMR) spectroscopy and electrochemicalimpedance spectroscopy (EIS), and supported further bymolecular dynamics (MD) simulations. We find that this phasecannot easily be retained at room temperature, instead forminga metastable state. The properties of this material are alsoconsidered, and its implications are addressed.Received: January 27, 2025Revised: March 25, 2025Accepted: March 26, 2025Published: April 4, 2025Articlepubs.acs.org/cm© 2025 The Authors. Published byAmerican Chemical Society2899https://doi.org/10.1021/acs.chemmater.5c00206Chem. Mater. 2025, 37, 2899−2906This article is licensed under CC-BY 4.0Downloaded via NATL INST FOR MATLS SCIENCE (NIMS) on May 9, 2025 at 02:38:05 (UTC).See https://pubs.acs.org/sharingguidelines for options on how to legitimately share published articles.https://pubs.acs.org/action/doSearch?field1=Contrib&text1="Emily+Milan"&field2=AllField&text2=&publication=&accessType=allContent&Earliest=&ref=pdfhttps://pubs.acs.org/action/doSearch?field1=Contrib&text1="James+A.+Quirk"&field2=AllField&text2=&publication=&accessType=allContent&Earliest=&ref=pdfhttps://pubs.acs.org/action/doSearch?field1=Contrib&text1="Kenjiro+Hashi"&field2=AllField&text2=&publication=&accessType=allContent&Earliest=&ref=pdfhttps://pubs.acs.org/action/doSearch?field1=Contrib&text1="John+Cattermull"&field2=AllField&text2=&publication=&accessType=allContent&Earliest=&ref=pdfhttps://pubs.acs.org/action/doSearch?field1=Contrib&text1="Andrew+L.+Goodwin"&field2=AllField&text2=&publication=&accessType=allContent&Earliest=&ref=pdfhttps://pubs.acs.org/action/doSearch?field1=Contrib&text1="James+A.+Dawson"&field2=AllField&text2=&publication=&accessType=allContent&Earliest=&ref=pdfhttps://pubs.acs.org/action/doSearch?field1=Contrib&text1="Mauro+Pasta"&field2=AllField&text2=&publication=&accessType=allContent&Earliest=&ref=pdfhttps://pubs.acs.org/action/doSearch?field1=Contrib&text1="Mauro+Pasta"&field2=AllField&text2=&publication=&accessType=allContent&Earliest=&ref=pdfhttps://pubs.acs.org/action/showCitFormats?doi=10.1021/acs.chemmater.5c00206&ref=pdfhttps://pubs.acs.org/doi/10.1021/acs.chemmater.5c00206?ref=pdfhttps://pubs.acs.org/doi/10.1021/acs.chemmater.5c00206?goto=articleMetrics&ref=pdfhttps://pubs.acs.org/doi/10.1021/acs.chemmater.5c00206?goto=recommendations&?ref=pdfhttps://pubs.acs.org/doi/10.1021/acs.chemmater.5c00206?goto=supporting-info&ref=pdfhttps://pubs.acs.org/doi/10.1021/acs.chemmater.5c00206?fig=tgr1&ref=pdfhttps://pubs.acs.org/toc/cmatex/37/8?ref=pdfhttps://pubs.acs.org/toc/cmatex/37/8?ref=pdfhttps://pubs.acs.org/toc/cmatex/37/8?ref=pdfhttps://pubs.acs.org/toc/cmatex/37/8?ref=pdfpubs.acs.org/cm?ref=pdfhttps://pubs.acs.org?ref=pdfhttps://pubs.acs.org?ref=pdfhttps://doi.org/10.1021/acs.chemmater.5c00206?urlappend=%3Fref%3DPDF&jav=VoR&rel=cite-ashttps://pubs.acs.org/cm?ref=pdfhttps://pubs.acs.org/cm?ref=pdfhttps://acsopenscience.org/researchers/open-access/https://creativecommons.org/licenses/by/4.0/https://creativecommons.org/licenses/by/4.0/https://creativecommons.org/licenses/by/4.0/■ RESULTS AND DISCUSSIONExistence of Li3(OH)2Br. Anhydrous LiBr-LiOH wasground together in the correct stoichiometric ratio, 66.7 mol% LiOH to 33.3 mol % LiBr, and heated at 5 °C/min to itsmolten state at 350 °C for 30 min. Upon furnace cooling at 3°C/min to room temperature, a material containing Li2OHBrand the anhydrous Li4(OH)3Br phase reported by Milan et al.is formed (Figures 1b and S1).15 This material was ground intoa powder using a mortar and pestle and investigated by DSCand variable-temperature (VT) XRD from room temperatureto 300 °C. Figure 1c,d shows these DSC and VT-XRDmeasurements, taken at 5 and 6 °C/min, respectively. In bothinstances, a phase transition to a new crystalline phase isobserved at ∼230 °C, as expected from Mahroug’s phasediagram.This phase persists until melting. The first indications ofmelting onset between 260 and 270 °C, where a uniformreduction in XRD peak intensity and rapid narrowing of the7Li NMR line shape occurs (Figure S2), with the Li3(OH)2Brphase fully disappearing by ∼290 °C. Unlike indicated inFigure 1. Existence of Li3(OH)2Br. (a) LiBr-LiOH phase diagram proposed by Mahroug et al.,9 drawn here with phase-field labels. Stoichiometriccompounds LiBr, Li2OHBr, Li3(OH)2Br, Li4(OH)3Br, and LiOH are indicated with colored lines. Li3(OH)2Br is included as a peritectic compoundat 67 mol % LiOH, between ∼230 and ∼280 °C. The widths of lines corresponding to stoichiometric compounds are not representative of theirstoichiometric ranges, which is unknown. (b) XRD pattern from room-temperature 67 mol % LiOH sample cooled from 350 °C at 3 °C/min.Bragg peak positions for Li4(OH)3Br and Li2OHBr phases are indicated by ticks beneath the data. (c) Heating and cooling DSC of 67 mol % LiOHsample at a rate of 5 °C/min. Two peaks are seen in each instance, expected to correspond to the formation and decomposition of Li3(OH)2Br. (d)Film plot showing synchrotron variable-temperature XRD measurements (λ = 0.82311 Å) of the 67 mol % LiOH sample shown in part b, heated at6 °C/min. A phase transition occurs at ∼225 °C, followed by melting from ∼270 °C.Chemistry of Materials pubs.acs.org/cm Articlehttps://doi.org/10.1021/acs.chemmater.5c00206Chem. Mater. 2025, 37, 2899−29062900https://pubs.acs.org/doi/suppl/10.1021/acs.chemmater.5c00206/suppl_file/cm5c00206_si_001.pdfhttps://pubs.acs.org/doi/suppl/10.1021/acs.chemmater.5c00206/suppl_file/cm5c00206_si_001.pdfhttps://pubs.acs.org/doi/10.1021/acs.chemmater.5c00206?fig=fig1&ref=pdfhttps://pubs.acs.org/doi/10.1021/acs.chemmater.5c00206?fig=fig1&ref=pdfhttps://pubs.acs.org/doi/10.1021/acs.chemmater.5c00206?fig=fig1&ref=pdfhttps://pubs.acs.org/doi/10.1021/acs.chemmater.5c00206?fig=fig1&ref=pdfpubs.acs.org/cm?ref=pdfhttps://doi.org/10.1021/acs.chemmater.5c00206?urlappend=%3Fref%3DPDF&jav=VoR&rel=cite-asMahroug’s phase diagram, no Li4(OH)3Br is observed prior tomelting. Additionally, on further heating to 300 °C, no LiOHforms, as expected from the phase diagram, but instead anotherpeak appears, as of yet unidentified. It is difficult to ascertainwhether these observations correspond to equilibriumconditions and hence indicate a discrepancy with the phasediagram, or can be attributed to nonequilibrium phenomena.Nevertheless, our findings support the existence of a phasewith composition Li3(OH)2Br which forms from 225 °C uponheating and melts around 270 °C.Structural Characterization. A cold-pressed pellet of the67 mol % LiOH sample was heated to 250 °C at 1 °C/min. Toensure equilibrium conditions, the sample was held at 250 °Cfor 1 h under in situ XRD. The diffraction pattern was found tobe stable in this time, indicating that equilibrium conditionswere being obtained and what is believed to be a single phase.The observed peak positions for the Li3(OH)2Br phase wereindexed and space group searching was conducted usingTOPAS-Academic software16 (Table S1). The peaks werefound to be fit well by several space groups with hexagonal unitcells. A Pawley refinement for the P63 space group, shown inFigure S3 (Table S2), fits the diffraction pattern in excellentagreement, giving lattice parameters of a = b = 6.57192(6) Å, c= 10.74643(17) Å.A low-energy structure for Li3(OH)2Br was determinedcomputationally through ab initio random structure searching.Figure 2. Crystal structure. Refined P63/mmc structure for Li3(OH)2Br viewed along (a) [100] and (b) [001]. Lithium occupancies arerepresented by pie charts and are assumed to be equal (0.24) across all sites. Blue and green lithium atoms are used to distinguish between cage-like(green) and network-like (blue) sites in the structure. The green lithium atoms are extended beyond a single unit cell in the a−b plane toemphasize the cage-like structures they form. (c) XRD pattern of Li3(OH)2Br after 1 h annealing at 250 °C, with matching Rietveld refinementusing the proposed P63/mmc structural model of Li3(OH)2Br (Rwp = 11.9%). The corresponding difference curve is offset below the data, and ticksindicate the positions of the Bragg reflections. (d) 7Li NMR spectra of Li3(OH)2Br at 230 °C. The line shape can be deconvoluted into twoGaussian peaks, indicating 2 lithium environments in the structure.Chemistry of Materials pubs.acs.org/cm Articlehttps://doi.org/10.1021/acs.chemmater.5c00206Chem. Mater. 2025, 37, 2899−29062901https://pubs.acs.org/doi/suppl/10.1021/acs.chemmater.5c00206/suppl_file/cm5c00206_si_001.pdfhttps://pubs.acs.org/doi/suppl/10.1021/acs.chemmater.5c00206/suppl_file/cm5c00206_si_001.pdfhttps://pubs.acs.org/doi/10.1021/acs.chemmater.5c00206?fig=fig2&ref=pdfhttps://pubs.acs.org/doi/10.1021/acs.chemmater.5c00206?fig=fig2&ref=pdfhttps://pubs.acs.org/doi/10.1021/acs.chemmater.5c00206?fig=fig2&ref=pdfhttps://pubs.acs.org/doi/10.1021/acs.chemmater.5c00206?fig=fig2&ref=pdfpubs.acs.org/cm?ref=pdfhttps://doi.org/10.1021/acs.chemmater.5c00206?urlappend=%3Fref%3DPDF&jav=VoR&rel=cite-asThousands of structures were generated in which the Li−O−Br sublattice was arranged in the P63 space group, with Hignored in symmetry determination as it could not be refinedin XRD. Each candidate structure was optimized using aCHGNet machine-learned force field fine-tuned for the Li−O−H−Br phase space.17 A possible structure for Li3(OH)2BrTable 1. Crystallographic Parameters from the Rietveld Refinement of Li3(OH)2Br In Situ at 250 °C Shown in Figure 2caspace group P63/mmca = b (Å) 6.57137(8)c (Å) 10.7454(2)V (Å3) 401.851(13)atom Wyckoff position x y z occupancy Uiso (Å2)Br1 4f 1/3 2/3 0.9373(2) 1 0.0344(11)O1 6h 0.1631(7) 0.8369 0.25 1 0.061(3)O2 2a 0 0 0 1 0.061(3)Li1 4e 0 0 0.3185 0.24 0.25(4)Li2 6h 0.1775 0.8225 0.75 0.24 0.25(4)Li3 12k 0.0938 0.9062 0.6454 0.24 0.25(4)Li4 12k 0.1456 0.8544 0.0879 0.24 0.25(4)Li5 4f 1/3 2/3 0.1905 0.24 0.25(4)Li6 12k 0.0790 0.5395 0.1804 0.24 0.25(4)aErrors on refined parameters are indicated in parentheses.Figure 3. Li-ion dynamics. (a) NMR relaxometry and diffusivity measurements of 67 mol % LiOH stoichiometry sample as a function oftemperature. Li3(OH)2Br forms from 220 °C and starts to melt from 260 °C. A linear fit to the SLR data is included for what is believed to be thepure Li3(OH)2Br phase region, used to calculate the activation energy for Li-ion hopping. (b) Nyquist plots from EIS of 67 mol % LiOHstoichiometry sample as a function of temperature. (c) Temperature dependence of ionic conductivity of 67 mol % LiOH stoichiometry sample.The ionic conductivity increases dramatically when Li3(OH)2Br forms above 225 °C. (d, e) Predicted lithium trajectories in the Li3(OH)2Br phasefrom molecular dynamics simulations, overlaid on the P63/mmc unit cell and viewed along (d) [100] and (e) [001]. Lithium density is seen to beconfined to cages with facile intra-cage hopping.Chemistry of Materials pubs.acs.org/cm Articlehttps://doi.org/10.1021/acs.chemmater.5c00206Chem. Mater. 2025, 37, 2899−29062902https://pubs.acs.org/doi/10.1021/acs.chemmater.5c00206?fig=fig3&ref=pdfhttps://pubs.acs.org/doi/10.1021/acs.chemmater.5c00206?fig=fig3&ref=pdfhttps://pubs.acs.org/doi/10.1021/acs.chemmater.5c00206?fig=fig3&ref=pdfhttps://pubs.acs.org/doi/10.1021/acs.chemmater.5c00206?fig=fig3&ref=pdfpubs.acs.org/cm?ref=pdfhttps://doi.org/10.1021/acs.chemmater.5c00206?urlappend=%3Fref%3DPDF&jav=VoR&rel=cite-aswas found and equilibrated at 250 °C so that fractionaloccupations at a finite temperature could be determined. Thelithium atoms were found to be highly mobile, meaning thatprecise occupancies could not be determined from MD. Assuch, the occupancies have been assumed to be equal across allsites. The resultant structure is shown in Figure S4, and it isdetailed in Table S3.The pseudosymmetry observed in the structure and theabsence of a clear mechanism for inversion-symmetry breakingindicate that the phase may belong to a higher symmetry spacegroup than P63. The atomic positions were adjusted to beconsistent with those of the P63/mmc space group, whichrefined well with the observed diffraction data. The refinedstructural model for the proposed P63/mmc structure is shownin Figure 2a,b, and compared with the computational startingmodel in Figure S4. A combination of network-like lithiumsites and distinct “cage” geometries is predicted to exist in thestructure, shown with blue and green lithium atoms,respectively. An XRD pattern taken at 250 °C, along with acorresponding Rietveld refinement, is shown in Figure 2c(Table 1). Information from the lithium in the structure wasnot refined due to its poor X-ray scattering cross section toavoid unphysical and inaccurate results. In order to establishinformation about lithium in the structure accurately,complementary techniques, such as neutron diffraction, willbe necessary.In addition to diffraction studies, 7Li NMR lineshapes weremeasured. Between 220 and 270 °C, in the temperature rangeof the Li3(OH)2Br phase, a line shape consisting of a narrowand a broad component is observed, which does not changesignificantly over the phase’s temperature range. An examplespectrum taken at 230 °C is shown in Figure 2d. The narrowcomponent suggests a highly mobile site, whereas the broadcomponent indicates another less mobile environment. It ispossible that the narrower element arises from the “cage”lithiums in the structure, in which rapid intra-cage hoppingmay occur.Lithium-Ion Dynamics. Another LiBr-LiOH compound,Li2OHBr, has high lithium-ion conductivity (∼10−6 S cm−1 atroom temperature), making it of interest for solid-electrolyteapplications.1−5,18−20 It is interesting to assess the mobility oflithium ions in Li3(OH)2Br to see whether it also exhibitssuperionic conductivity.To evaluate the lithium-ion dynamics in the Li3(OH)2Brphase, we took NMR 7Li spin−lattice relaxation (SLR)measurements and pulsed-field gradient (PFG) NMR meas-urements across a range of temperatures, as shown in Figure3a. The SLR measurements display three distinct regions. Atlow temperatures, the low-temperature flank of the rate peakcorresponding to the regions of mixed Li2OHBr andLi4(OH)3Br can be seen. Around 220 °C, Li3(OH)2Br beginsto form, and data points on the high-temperature flank of thecorresponding rate peak can be seen. From 260 °C, anotherchange corresponding to the sample melting is observed. Thenarrow temperature range in which the Li3(OH)2Br phaseexists means that it is challenging to fit a Bloembergen−Purcell−Pound (BPP) model to the SLR data. To obtain anestimate of the activation energy for lithium-ion hopping, anapproximation can instead be made using an Arrheniusrelationship.21,22 A linear fit between 225 and 250 °C suggestsa low activation energy of 0.22 eV. PFG-NMR measurementscould be obtained from 220 °C (coinciding with the onset ofLi3(OH)2Br formation). A lithium diffusivity of 1.53 × 10−10m2 s−1 was found for the Li3(OH)2Br phase at 250 °C.Activation energy cannot be obtained from the diffusivitymeasurements due to microscopic changes occurring uponheating. The high temperatures involved result in sintering andgrain growth in the powder sample, which impacts PFGdiffusivity measurements probing a similar length scale.To study ionic conductivity in Li3(OH)2Br over macro-scopic length scales, EIS was conducted on cold-pressed pelletswith stainless steel blocking electrodes and a high stackpressure of 70 MPa applied to minimize interfacial resistance.The Nyquist plots obtained are shown in Figure 3b.Characteristic semicircles are observed at lower temperatures,which could be fitted using an equivalent circuit containing abulk and grain boundary component. As the temperatureincreases, low resistances result in the semicircles disappearing,and so the intersection of the low-frequency tail with the realaxis was instead used to calculate the total resistance. Thetemperature dependence of the ionic conductivity is shown inFigure 3c. A step change corresponding to the phase transitionforming Li3(OH)2Br can be seen at 230 °C, with an ionicconductivity of 0.12 S cm−1 at 250 °C and an activation energyof 0.32 ± 0.04 eV found between 240 and 260 °C. Uponextrapolating to room temperature, an ionic conductivity of 5.2± 1.6 × 10−4 S cm−1 is found. It would be desirable to retainthis phase and hence the excellent ionic conductivity to lowertemperatures. Li3(OH)2Br could offer several significantadvantages over many solid electrolytes currently underconsideration. Li3(OH)2Br contains inexpensive and abundantprecursors, unlike popular options such as Li7La3Zr2O12(LLZO) and lithium argyrodite sulfides.23 Li2OHBr exhibitsgood stability with lithium metal at room temperature, as is thecase with other oxide electrolytes, which is anticipated totranslate to Li3(OH)2Br.3,5 Furthermore, the density ofLi3(OH)2Br (2.23 g cm−3) is significantly reduced comparedto other oxide electrolytes, such as LLZO (5.07 g cm−3),Li0.34La0.56TiO3 (5.01 g cm−3), and Li1.5Al0.5Ge1.5(PO4)3 (3.56g cm−3).24To identify and eliminate the impacts of sintering occurringat high temperatures, EIS measurements were also taken usinganother heating protocol (see Supporting Note 1). Similarbehavior was observed for the Li3(OH)2Br phase in bothsamples (Figure S5), indicating that the obtained conductivityis representative of that of bulk Li3(OH)2Br.Molecular dynamics simulations were carried out on asupercell of the computational Li3(OH)2Br structure contain-ing 576 Li ions across a temperature range of 220 to 240 °C.Trajectories, shown in Figure 3d,e, reveal that the Li density islargely confined to “cages” in which Li can easily movebetween sites within the cage (intra-cage), but where jumpsbetween the cages (inter-cage) are less frequent. Thecorresponding activation energies for intracage and intercagejumps were calculated to be 0.18 ± 0.02 and 0.35 ± 0.02 eVrespectively. Accordingly, the macroscopic activation energy iscalculated to be 0.27 ± 0.03 eV, which is in reasonableagreement with the EIS findings. The underestimate inactivation energy between MD and EIS is reasonable giventhe differences between the computational and experimentallyrefined structure and the absence of extended defects such asgrain boundaries in MD, which assumes a pristine crystal.Attempts to engineer Li3(OH)2Br to improve ionic con-ductivity may focus on lowering the barriers for intercagediffusion to enable long-range transport. Discrepanciesbetween the intra- and intercage activation energies are wellChemistry of Materials pubs.acs.org/cm Articlehttps://doi.org/10.1021/acs.chemmater.5c00206Chem. Mater. 2025, 37, 2899−29062903https://pubs.acs.org/doi/suppl/10.1021/acs.chemmater.5c00206/suppl_file/cm5c00206_si_001.pdfhttps://pubs.acs.org/doi/suppl/10.1021/acs.chemmater.5c00206/suppl_file/cm5c00206_si_001.pdfhttps://pubs.acs.org/doi/suppl/10.1021/acs.chemmater.5c00206/suppl_file/cm5c00206_si_001.pdfhttps://pubs.acs.org/doi/suppl/10.1021/acs.chemmater.5c00206/suppl_file/cm5c00206_si_001.pdfhttps://pubs.acs.org/doi/suppl/10.1021/acs.chemmater.5c00206/suppl_file/cm5c00206_si_001.pdfpubs.acs.org/cm?ref=pdfhttps://doi.org/10.1021/acs.chemmater.5c00206?urlappend=%3Fref%3DPDF&jav=VoR&rel=cite-asstudied in the familiar argyrodite family of solid electrolytesLi6PS5Cl,25 where increased disorder is a possible avenue toencouraging macroscopic diffusion.Metastable Retention of Li3(OH)2Br. To see whether theLi3(OH)2Br phase could be retained at room temperaturemetastably, samples were annealed at 250 °C for 4 h such thatLi3(OH)2Br fully forms, followed by quenching to roomtemperature to try and “freeze-in” the phase. Both samplesquenched from the liquid state at 400 °C and from theLi3(OH)2Br phase field at 250 °C exhibited a similar XRDpattern with a different set of reflections to Li3(OH)2Br,Li2(OH)Br, or Li4(OH)3Br (Figure S6a). The thermodynami-cally unstable nature of the P63/mmc phase at low temper-atures, coupled with the fast cooling rate inhibiting thenecessary atomic rearrangements for attaining equilibrium,results in the metastable state observed here. The ionicconductivity of this metastable state was assessed using EIS(Figure S6b,c), but did not demonstrate the same promisingbehavior as the high-temperature phase, yielding an ionicconductivity of 3.6 × 10−8 S cm−1 at 30 °C and an activationenergy of 0.75 eV in the temperature range 30 to 50 °C. DCchronoamperometry measurements indicated an electronicconductivity of 2.0 × 10−11 S cm−1 at 25 °C (Figure S6d). Thisionic conductivity is lower than that of the related compoundand solid-electrolyte candidate, Li2OHBr, which typicallyexhibits ionic conductivities of ∼10−6 S cm−1 for cold-pressedpellets at room temperature.1−5,18−20 Synthesis of Li2OHBrfrom the melt, as is typical in the literature, requires coolingthrough phase fields containing Li3(OH)2Br (Figure 1a).Consequently, it is important to consider undesirableLi3(OH)2Br impurity formation in the synthesis of Li2OHBr.The small crystallite sizes formed during cooling, combinedwith potentially high levels of microstrain from thediffusionless metastable transformation, may mean that it isnot obvious in XRD, typically used to screen for impurities.■ CONCLUSIONSIn this work, we investigated the high-temperature Li3(OH)2Brphase which is thermodynamically stable between ∼225 and275 °C. Diffraction studies suggest the phase takes a hexagonalunit cell with lattice parameters of a = b = 6.572 Å and c =10.746 Å at 250 °C. Through a combination of XRDrefinements and theoretical prediction, a structural modelwith the P63/mmc space group is suggested. MD indicates thatthe structure contains cage-like structures of lithium, in whichintracage diffusion is facile. A macroscopic ionic conductivityof 0.12 S cm−1 is measured at 250 °C, which would extrapolateto 5 × 10−4 S cm−1 upon retention at room temperature.However, attempts to stabilize this phase were not successful,resulting in a different metastable state with a worse ionicconductivity.■ EXPERIMENTAL METHODSSynthesis. For the synthesis described in the main text, anhydrousLiOH (98%, Sigma-Aldrich) and LiBr (≥99%, Sigma-Aldrich) wereused. Synthesis was carried out in an MTI compact muffle furnaceinside an argon-filled glovebox (MBraun, H2O < 0.5 ppm, O2 < 0.5ppm). All utensils and consumables were dried in a vacuum oven (∼1mbar, 70 °C) for at least 4 h prior to use.XRD. Room-temperature XRD measurements were taken using aRigaku Miniflex diffractometer (Cu Kα) inside a nitrogen-filledglovebox (MBraun, H2O < 0.5 ppm, O2 < 0.5 ppm). Powder sampleswere loaded onto a single-crystal silicon holder to minimizebackground contributions. VT XRD measurements in Figure 1dand Figure S2a were carried out on the I11 beamline at DiamondLight Source (λ = 0.82311 Å) on powders sealed in borosilicate glasscapillaries under argon and heated continuously at 6 °C/min using anFMB Oxford cyberstar hot air blower. Diffraction patterns weremeasured in capillary transmission geometry using a Mythen2Position Sensitive Detector. Two data collections, each 5 s, weretaken at angles 0.25° apart and summed to account for gaps indetector coverage. The in situ 250 °C measurements used forrefinements were taken on a Rigaku Smartlab diffractometer (Cu Kα),following a heating step at 1 °C/min and a holding step for 1 h at 250°C. For this, the powder was pressed into a 10 mm diameter pelletand loaded onto a heating stage under argon flow.Pawley and Rietveld refinements were conducted using TOPAS-Academic software.16 In both, the unit cell, background, and peakshape parameters were allowed to refine freely. In Rietveldrefinements, the O and Br fractional positions were allowed to refine,in addition to 3 thermal displacement parameters for the Br, O, and Liatoms.NMR. 7Li NMR measurements were performed using an ECZ-500(JEOL, Japan) spectrometer and a homemade high-temperature PFG-NMR probe.26 The sample was packed into a quartz NMR tube SP-405 (SHIGEMI, Japan) and sealed in an Ar-filled glovebox. Theresonance frequency of 7Li was 194.4 MHz. NMR spectra wererecorded by using a single-pulse sequence. The chemical shift wasreferenced to a 1.0 M LiCl solution at 0 ppm. Spin−lattice relaxationtimes were measured by using a saturation recovery method. The datawas fitted to f(t) = f∞(1 − H0 exp(−t/T1ρ)). The diffusion coefficientswere measured by using the STE-PFG sequence.DSC. DSC measurements were taken under argon on a NetzschTGA-MS using ∼5 mg of sample in an alumina pan covered with analumina lid. An argon shower was used to protect the sample from airexposure during loading into the instrument. A heating and coolingrate of 5 °C/min was used. Background subtraction was carried outusing OriginLab software.Electrochemical Measurements. For electrochemical measure-ments, the powder samples were pressed into 5 mm diameter pelletsfor 3 min at 370 MPa. For measurements of the quenched metastablestate, nickel foil (Advent Materials, 99.95%, 0.0125 mm) blockingelectrodes were placed on either side of the electrolyte pellet. Forhigh-temperature measurements of the Li3(OH)2Br phase, stainlesssteel pistons were used as blocking electrodes due to the reactivity ofnickel. A custom-built cell applied a uniaxial pressure of 70 MPa toensure good contact between the electrolyte and blocking electrodes.Assembly and measurements were carried out in an argon-filledglovebox.For EIS measurements, the cells were heated in an MTI compactmuffle furnace. The samples were held at the desired temperature for30 min prior to measurement, followed by a 15 min ramp period tothe next temperature. Measurements were taken using a BioLogicMTZ35 frequency response analyzer in a two-point probeconfiguration in the frequency range 35 MHz to 0.1 Hz with avoltage amplitude of 10 mV. Conductivities were obtained fromNyquist plots. For measurements of the quenched metastable state,equivalent circuits consisting of a resistor and contributionscorresponding to bulk and grain boundary resistances, R1 + Q2/R2+ Q3/R3, were used to model the data using BioLogic EC lab software.DC chronoamperometry measurements were taken for 10 h at 0.3,0.6, and 1 V.■ COMPUTATIONAL METHODSA preliminary ab initio random structure searching (AIRSS)27 run wasperformed on several hundred geometries for Li3(OH)2Br andLi4(OH)3Br using the pretrained CHGNet foundation model. Then, aCHGNet model was fine-tuned for Li−O−H−Br systems by trainingthem against a set of molecular dynamics trajectories. The systemswere: solid Li2OHBr, Li3(OH)2Br, and Li4(OH)3Br at 900 K withNPT ensemble; and molten Li2OHBr, Li3(OH)2Br, and Li4(OH)3Brat 2000 K with NVT ensemble. The fine-tuned model was then usedfor the final AIRSS to determine the stable structure in this work andChemistry of Materials pubs.acs.org/cm Articlehttps://doi.org/10.1021/acs.chemmater.5c00206Chem. Mater. 2025, 37, 2899−29062904https://pubs.acs.org/doi/suppl/10.1021/acs.chemmater.5c00206/suppl_file/cm5c00206_si_001.pdfhttps://pubs.acs.org/doi/suppl/10.1021/acs.chemmater.5c00206/suppl_file/cm5c00206_si_001.pdfhttps://pubs.acs.org/doi/suppl/10.1021/acs.chemmater.5c00206/suppl_file/cm5c00206_si_001.pdfhttps://pubs.acs.org/doi/suppl/10.1021/acs.chemmater.5c00206/suppl_file/cm5c00206_si_001.pdfpubs.acs.org/cm?ref=pdfhttps://doi.org/10.1021/acs.chemmater.5c00206?urlappend=%3Fref%3DPDF&jav=VoR&rel=cite-asfor all further MD runs. Due to the importance of H-bonding inmaterials containing OH species, dispersion was treated using theDFT-D3 method28 using the implementation in torch-dftd.29Analysis of the MD trajectories was performed with Pymatgen30and Gemdat,31 which allows for decomposition of activationenergies into distinct hops between atomic sites.Fully occupied unit cells were produced by structure searching.Then, a supercell was constructed and equilibrated at 250 °C. Partiallyoccupied sites were determined by wrapping the trajectory back intothe unit cell to determine the stable sites. Because Li diffuses soreadily in the structure, the convergence of the site occupancies ispoor. For simplicity, the occupation of the determined stable sites wasnormalized to give the correct number of Li atoms per unit cell.All geometry optimizations were carried out until the forces on theions were less than 0.05 eV/Å. The time step for molecular dynamicswas 0.5 fs to ensure numerical stability when equations of motionwere integrated due to the small mass and rapid acceleration ofprotons. The computational activation energy was fitted to temper-atures of 220, 230, 230, 235, 240, 245, 250, 260, and 270 °C. Alltrajectories are at least 250 ps long. For fitting to the Einstein relation,the trajectories were split into 5 parts to calculate mean-squareddisplacements in order to reduce errors arising from deviations fromlinearity in long trajectories.32 Simulated ab initio MD for fine-tuningwas performed in VASP. Exchange correlation was treated with thePBE functional.33 Projector-augmented wave (PAW) pseudopoten-tials34,35 were employed in which the following electrons were treatedas valence: 1s1 for H, 1s22s1 for Li; 2s22p4 for O; and 4s24p5 for Br.The plane-wave basis used a cutoff energy of 500 eV.■ ASSOCIATED CONTENT*sı Supporting InformationThe Supporting Information is available free of charge athttps://pubs.acs.org/doi/10.1021/acs.chemmater.5c00206.Additional computational and experimental resultsincluding XRD, Raman spectroscopy, NMR, and EIS(PDF)■ AUTHOR INFORMATIONCorresponding AuthorMauro Pasta − Department of Materials, University of Oxford,Oxford OX1 3PH, U.K.; orcid.org/0000-0002-2613-4555; Email: mauro.pasta@materials.ox.ac.ukAuthorsEmily Milan − Department of Materials, University of Oxford,Oxford OX1 3PH, U.K.James A. Quirk − Chemistry�School of Natural andEnvironmental Sciences, Newcastle University, Newcastleupon Tyne NE1 7RU, U.K.Kenjiro Hashi − National Institute for Materials Science,Tsukuba 305-0044, JapanJohn Cattermull − Department of Chemistry, University ofOxford, Oxford OX1 3QR, U.K.; Department of Materials,University of Oxford, Oxford OX1 3PH, U.K.; orcid.org/0009-0006-5209-3132Andrew L. Goodwin − Department of Chemistry, Universityof Oxford, Oxford OX1 3QR, U.K.James A. Dawson − Chemistry�School of Natural andEnvironmental Sciences, Newcastle University, Newcastleupon Tyne NE1 7RU, U.K.; orcid.org/0000-0002-3946-5337Complete contact information is available at:https://pubs.acs.org/10.1021/acs.chemmater.5c00206NotesThe authors declare no competing financial interest.■ ACKNOWLEDGMENTSThis work was supported by funding from the Engineering andPhysical Sciences Research Council, partially sponsored byMorgan Advanced Materials (grant number EP/T517811/1),and the Henry Royce Institute for capital equipment (throughthe U.K. Engineering and Physical Sciences Research Councilgrant EP/R010145/1). J.A.D is grateful for funding from UKResearch and Innovation (UKRI) under the U.K. govern-ment’s Horizon Europe funding Guarantee (EP/Z000254/1).Via membership of the UK’s HEC Materials ChemistryConsortium, which was funded by the EPSRC (EP/X035859), this work used the ARCHER2 U.K. NationalSupercomputing Service. J.C. and A.L.G. gratefully acknowl-edge the E.R.C. for funding (Advanced Grant 788144) and theprovision of a BAG allocation (CY25166) on the I11 beamlineat the Diamond Light Source, U.K.■ REFERENCES(1) Schwering, G.; Honnerscheid, A.; Wüllen, L. V.; Jansen, M. HighLithium Ionic Conductivity in the Lithium Halide Hydrates Li3-n(OHn)Cl (0.83 ≤ n ≤ 2) and Li3-n(OHn)Br (1 ≤ n ≤ 2) atAmbient Temperatures. ChemPhysChem 2003, 4, 343−348.(2) Zhao, Y.; Daemen, L. L. Superionic conductivity in lithium-richanti-perovskites. J. Am. Chem. Soc. 2012, 134, 15042−15047.(3) Hood, Z. D.; Wang, H.; Pandian, A. S.; Keum, J. K.; Liang, C.Li2OHCl Crystalline Electrolyte for Stable Metallic Lithium Anodes.J. Am. Chem. Soc. 2016, 138, 1768−1771.(4) Sacci, R. L.; Bennett, T. H.; Fang, H.; Han, K. S.; Lames, M.;Murugesan, V.; Jena, P.; Nanda, J. Halide sublattice dynamics driveLi-ion transport in antiperovskites. J. Mater. Chem. A 2022, 10,15731−15742.(5) Lee, H. J.; Darminto, B.; Narayanan, S.; Diaz-Lopez, M.; Xiao, A.W.; Chart, Y.; Lee, J. H.; Dawson, J. A.; Pasta, M. Li-ion conductivityin Li 2 OHCl 1 x Br x solid electrolytes: grains, grain boundaries andinterfaces. J. Mater. Chem. A 2022, 10, 11574−11586.(6) Achchaq, F.; Barrio, E. P. D.; Lebraud, E.; Péchev, S.; Toutain, J.Development of a new LiBr/LiOH-based alloy for thermal energystorage. J. Phys. Chem. Solids 2019, 131, 173−179.(7) Legros, P.; Lebraud, E.; Duquesne, M.; Achchaq, F. Li4Br(OH)3microstructure monitoring over its synthesis to tackle the lithium-based salts exploitation challenges as advanced phase change materialsfor storage technologies. Mater. Des. 2020, 196, 109160−109170.(8) Mahroug, I.; Doppiu, S.; Dauvergne, J. L.; Echeverria, M.;Toutain, J.; del Barrio, E. P. Study of peritectic compoundLi4(OH)3Br for high temperature thermal energy storage in solarpower applications. Sol. Energy Mater. Sol. Cells 2021, 230, 111259−111261.(9) Mahroug, I.; Doppiu, S.; Dauvergne, J. L.; Toutain, J.; del Barrio,E. P. Extended investigation of LiOH−LiBr binary system for high-temperature thermal energy storage applications. J. Therm. Anal.Calorim. 2022, 147, 12455−12465.(10) Mahroug, I.; González-Fernández, L.; Grosu, Y.; Doppiu, S.;Palomo, E. Compatibility of container materials with peritectic phasechange material for high-temperature thermal energy storage. J. EnergyStorage 2023, 73, No. 108892.(11) Scarpa, G. Atti Accad. Naz. Lincei Rend. Classe Sci. Fis. Mater.Nat. Set. V 1915, 24, No. 476.(12) Hartwig, P.; Rabenau, A.; Weppner, W. Lithium HydroxideHalides: Phase Equilibria and Ionic Conductivities. J. Less CommonMet. 1981, 78, 227−233.(13) Reshetnikov, N. A.; Unzhakov, G. M.Izv Fiz-Khim Nauch-Issled Inst Irkutsk Gosud Univ. 1953; Vol. 2, p 23.Chemistry of Materials pubs.acs.org/cm Articlehttps://doi.org/10.1021/acs.chemmater.5c00206Chem. Mater. 2025, 37, 2899−29062905https://pubs.acs.org/doi/10.1021/acs.chemmater.5c00206?goto=supporting-infohttps://pubs.acs.org/doi/suppl/10.1021/acs.chemmater.5c00206/suppl_file/cm5c00206_si_001.pdfhttps://pubs.acs.org/action/doSearch?field1=Contrib&text1="Mauro+Pasta"&field2=AllField&text2=&publication=&accessType=allContent&Earliest=&ref=pdfhttps://orcid.org/0000-0002-2613-4555https://orcid.org/0000-0002-2613-4555mailto:mauro.pasta@materials.ox.ac.ukhttps://pubs.acs.org/action/doSearch?field1=Contrib&text1="Emily+Milan"&field2=AllField&text2=&publication=&accessType=allContent&Earliest=&ref=pdfhttps://pubs.acs.org/action/doSearch?field1=Contrib&text1="James+A.+Quirk"&field2=AllField&text2=&publication=&accessType=allContent&Earliest=&ref=pdfhttps://pubs.acs.org/action/doSearch?field1=Contrib&text1="Kenjiro+Hashi"&field2=AllField&text2=&publication=&accessType=allContent&Earliest=&ref=pdfhttps://pubs.acs.org/action/doSearch?field1=Contrib&text1="John+Cattermull"&field2=AllField&text2=&publication=&accessType=allContent&Earliest=&ref=pdfhttps://orcid.org/0009-0006-5209-3132https://orcid.org/0009-0006-5209-3132https://pubs.acs.org/action/doSearch?field1=Contrib&text1="Andrew+L.+Goodwin"&field2=AllField&text2=&publication=&accessType=allContent&Earliest=&ref=pdfhttps://pubs.acs.org/action/doSearch?field1=Contrib&text1="James+A.+Dawson"&field2=AllField&text2=&publication=&accessType=allContent&Earliest=&ref=pdfhttps://orcid.org/0000-0002-3946-5337https://orcid.org/0000-0002-3946-5337https://pubs.acs.org/doi/10.1021/acs.chemmater.5c00206?ref=pdfhttps://doi.org/10.1002/cphc.200390060https://doi.org/10.1002/cphc.200390060https://doi.org/10.1002/cphc.200390060https://doi.org/10.1002/cphc.200390060https://doi.org/10.1021/ja305709z?urlappend=%3Fref%3DPDF&jav=VoR&rel=cite-ashttps://doi.org/10.1021/ja305709z?urlappend=%3Fref%3DPDF&jav=VoR&rel=cite-ashttps://doi.org/10.1021/jacs.5b11851?urlappend=%3Fref%3DPDF&jav=VoR&rel=cite-ashttps://doi.org/10.1039/D2TA02598Ahttps://doi.org/10.1039/D2TA02598Ahttps://doi.org/10.1039/D2TA01462Ahttps://doi.org/10.1039/D2TA01462Ahttps://doi.org/10.1039/D2TA01462Ahttps://doi.org/10.1016/j.jpcs.2019.04.001https://doi.org/10.1016/j.jpcs.2019.04.001https://doi.org/10.1016/j.matdes.2020.109160https://doi.org/10.1016/j.matdes.2020.109160https://doi.org/10.1016/j.matdes.2020.109160https://doi.org/10.1016/j.matdes.2020.109160https://doi.org/10.1016/j.solmat.2021.111259https://doi.org/10.1016/j.solmat.2021.111259https://doi.org/10.1016/j.solmat.2021.111259https://doi.org/10.1007/s10973-022-11468-4https://doi.org/10.1007/s10973-022-11468-4https://doi.org/10.1016/j.est.2023.108892https://doi.org/10.1016/j.est.2023.108892https://doi.org/10.1016/0022-5088(81)90132-6https://doi.org/10.1016/0022-5088(81)90132-6pubs.acs.org/cm?ref=pdfhttps://doi.org/10.1021/acs.chemmater.5c00206?urlappend=%3Fref%3DPDF&jav=VoR&rel=cite-as(14) Sangster, J. Thermodynamics and phase diagrams of 32 binarycommon-ion systems of the group Li,Na,K,Rb,Cs//F,Cl,Br,I,OH,-NO3. J. Phase Equilib. 2000, 21, 241−268.(15) Milan, E.; Quirk, J. A.; Cattermull, J.; Goodwin, A. L.; Dawson,J. A.; Pasta, M. Synthesis and characterization of Li4(OH)3Br forthermal energy storage ChemRxiv 2025 DOI: 10.26434/chemrxiv-2025-r1zlr. (accessed March 25, 2025).(16) Coelho, A. A. TOPAS and TOPAS-Academic: an optimizationprogram integrating computer algebra and crystallographic objectswritten in C++. J. Appl. Crystallogr. 2018, 51, 210−218.(17) Deng, B.; Zhong, P.; Jun, K.; Riebesell, J.; Han, K.; Bartel, C. J.;Ceder, G. CHGNet as a pretrained universal neural network potentialfor charge-informed atomistic modelling. Nat. Mach. Intell. 2023, 5,1031−1041.(18) Deng, Z.; Ou, M.; Wan, J.; Li, S.; Li, Y.; Zhang, Y.; Deng, Z.;Xu, J.; Qiu, Y.; Liu, Y.; Fang, C.; Li, Q.; Huang, L.; Zhu, J.; Han, S.;Han, J.; Zhao, Y. Local structural changes and inductive effects on ionconduction in antiperovskite solid electrolytes. Chem. Mater. 2020, 32,8827−8835.(19) Yoshikawa, K.; Yamamoto, T.; Sugumar, M. K.; Motoyama, M.;Iriyama, Y. Room Temperature Operation and High Cycle Stability ofan All-Solid-State Lithium Battery Fabricated by Cold Pressing UsingSoft Li2OHBr Solid Electrolyte. Energy Fuels 2021, 35, 12581−12587.(20) Sugumar, M. K.; Yamamoto, T.; Motoyama, M.; Iriyama, Y.Room temperature synthesis of anti-perovskite structured Li2OHBr.Solid State Ionics 2020, 349, No. 115298.(21) Kuhn, A.; Narayanan, S.; Spencer, L.; Goward, G.;Thangadurai, V.; Wilkening, M. Li self-diffusion in garnet-typeLi7La3Zr 2O12 as probed directly by diffusion-induced Li7 spin-lattice relaxation NMR spectroscopy. Phys. Rev. B 2011, 83, 94302−94312.(22) Bottke, P.; Rettenwander, D.; Schmidt, W.; Amthauer, G.;Wilkening, M. Ion Dynamics in Solid Electrolytes: NMR Reveals theElementary Steps of Li+ Hopping in the Garnet Li6.5La3Zr1.75-Mo0.25O12. Chem. Mater. 2015, 27, 6571−6582.(23) Blengini, G. A.; Latunussa, C. E. L.; Eynard, U.; de Matos, C.T.; Wittmer, D.; Georgitzikis, K.; Pavel, C.; Carrara, S.; Mancini, L.;Unguru, M.; Blagoeva, D.; Mathieux, F.; Pennington, D. Study on theEU’s List of Critical Raw Materials, Final Report; EuropeanCommission: Directorate-General for Internal Market, Industry,Entrepreneurship and SMEs. Publications Office of the EuropeanUnion, 2020.(24) Jiang, Z.; Wang, S.; Chen, X.; Yang, W.; Yao, X.; Hu, X.; Han,Q.; Wang, H. Tape-Casting Li0.34La0.56TiO3 Ceramic ElectrolyteFilms Permit High Energy Density of Lithium-Metal Batteries. Adv.Mater. 2020, 32, No. 1906221.(25) Gautam, A.; Sadowski, M.; Ghidiu, M.; Minafra, N.; Senyshyn,A.; Albe, K.; Zeier, W. G. Engineering the Site-Disorder and LithiumDistribution in the Lithium Superionic Argyrodite Li6PS5Br. Adv.Energy Mater. 2021, 11, No. 2003369.(26) Hashi, K.; Ohki, S.; Mogami, Y.; Goto, A.; Shimizu, T. High-Temperature Pulsed-Field-Gradient 7Li-NMR Measurements ofLi2CO3 over 700 K. Anal. Sci. 2021, 37, 1477−1479.(27) Pickard, C. J.; Needs, R. J. Ab initio random structuresearching. J. Phys.: Condens. Matter 2011, 23, No. 053201.(28) Grimme, S.; Antony, J.; Ehrlich, S.; Krieg, H. A consistent andaccurate ab initio parametrization of density functional dispersioncorrection (DFT-D) for the 94 elements H-Pu. J. Chem. Phys. 2010,132, No. 154104.(29) Takamoto, S.; Shinagawa, C.; Motoki, D.; Nakago, K.; Li, W.;Kurata, I.; Watanabe, T.; Yayama, Y.; Iriguchi, H.; Asano, Y.;Onodera, T.; Ishii, T.; Kudo, T.; Ono, H.; Sawada, R.; Ishitani, R.;Ong, M.; Yamaguchi, T.; Kataoka, T.; Hayashi, A.; Charoenphakdee,N.; Ibuka, T. Towards universal neural network potential for materialdiscovery applicable to arbitrary combination of 45 elements. Nat.Commun. 2022, 13, No. 2991.(30) Ong, S. P.; Richards, W. D.; Jain, A.; Hautier, G.; Kocher, M.;Cholia, S.; Gunter, D.; Chevrier, V. L.; Persson, K. A.; Ceder, G.Python Materials Genomics (pymatgen): A robust, open-sourcepython library for materials analysis. Comput. Mater. Sci. 2013, 68,314−319.(31) Azizi, V.; Smeets, S.; Lavrinenko, A. K.; Ciarella, S.; Famprikis,T. GEMDAT, https://gemdat.readthedocs.io. (accessed March 25,2025).(32) He, X.; Zhu, Y.; Epstein, A.; Mo, Y. Statistical variances ofdiffusional properties from ab initio molecular dynamics simulations.npj Comput. Mater. 2018, 4, 1−9.(33) Perdew, J. P.; Burke, K.; Ernzerhof, M. Generalized GradientApproximation Made Simple. Phys. Rev. Lett. 1996, 77, 3865−3868.(34) Kresse, G.; Joubert, D. From ultrasoft pseudopotentials to theprojector augmented-wave method. Phys. Rev. B 1999, 59, 1758−1775.(35) Blöchl, P. E. Projector augmented-wave method. Phys. Rev. B1994, 50, 17953−17979.Chemistry of Materials pubs.acs.org/cm Articlehttps://doi.org/10.1021/acs.chemmater.5c00206Chem. Mater. 2025, 37, 2899−29062906https://doi.org/10.1361/105497100770340048https://doi.org/10.1361/105497100770340048https://doi.org/10.1361/105497100770340048https://doi.org/10.26434/chemrxiv-2025-r1zlrhttps://doi.org/10.26434/chemrxiv-2025-r1zlrhttps://doi.org/10.26434/chemrxiv-2025-r1zlr?urlappend=%3Fref%3DPDF&jav=VoR&rel=cite-ashttps://doi.org/10.26434/chemrxiv-2025-r1zlr?urlappend=%3Fref%3DPDF&jav=VoR&rel=cite-ashttps://doi.org/10.1107/S1600576718000183https://doi.org/10.1107/S1600576718000183https://doi.org/10.1107/S1600576718000183https://doi.org/10.1038/s42256-023-00716-3https://doi.org/10.1038/s42256-023-00716-3https://doi.org/10.1021/acs.chemmater.0c02173?urlappend=%3Fref%3DPDF&jav=VoR&rel=cite-ashttps://doi.org/10.1021/acs.chemmater.0c02173?urlappend=%3Fref%3DPDF&jav=VoR&rel=cite-ashttps://doi.org/10.1021/acs.energyfuels.1c01190?urlappend=%3Fref%3DPDF&jav=VoR&rel=cite-ashttps://doi.org/10.1021/acs.energyfuels.1c01190?urlappend=%3Fref%3DPDF&jav=VoR&rel=cite-ashttps://doi.org/10.1021/acs.energyfuels.1c01190?urlappend=%3Fref%3DPDF&jav=VoR&rel=cite-ashttps://doi.org/10.1016/j.ssi.2020.115298https://doi.org/10.1103/PhysRevB.83.094302https://doi.org/10.1103/PhysRevB.83.094302https://doi.org/10.1103/PhysRevB.83.094302https://doi.org/10.1021/acs.chemmater.5b02231?urlappend=%3Fref%3DPDF&jav=VoR&rel=cite-ashttps://doi.org/10.1021/acs.chemmater.5b02231?urlappend=%3Fref%3DPDF&jav=VoR&rel=cite-ashttps://doi.org/10.1021/acs.chemmater.5b02231?urlappend=%3Fref%3DPDF&jav=VoR&rel=cite-ashttps://doi.org/10.1002/adma.201906221https://doi.org/10.1002/adma.201906221https://doi.org/10.1002/aenm.202003369https://doi.org/10.1002/aenm.202003369https://doi.org/10.2116/analsci.21A001https://doi.org/10.2116/analsci.21A001https://doi.org/10.2116/analsci.21A001https://doi.org/10.1088/0953-8984/23/5/053201https://doi.org/10.1088/0953-8984/23/5/053201https://doi.org/10.1063/1.3382344https://doi.org/10.1063/1.3382344https://doi.org/10.1063/1.3382344https://doi.org/10.1038/s41467-022-30687-9https://doi.org/10.1038/s41467-022-30687-9https://doi.org/10.1016/j.commatsci.2012.10.028https://doi.org/10.1016/j.commatsci.2012.10.028https://gemdat.readthedocs.iohttps://doi.org/10.1038/s41524-018-0074-yhttps://doi.org/10.1038/s41524-018-0074-yhttps://doi.org/10.1103/PhysRevLett.77.3865https://doi.org/10.1103/PhysRevLett.77.3865https://doi.org/10.1103/PhysRevB.59.1758https://doi.org/10.1103/PhysRevB.59.1758https://doi.org/10.1103/PhysRevB.50.17953pubs.acs.org/cm?ref=pdfhttps://doi.org/10.1021/acs.chemmater.5c00206?urlappend=%3Fref%3DPDF&jav=VoR&rel=cite-as