# Fileset

[PhysRevLett.132.156701.pdf](https://mdr.nims.go.jp/filesets/68c2c429-a78b-49cc-93b1-75eef13079e9/download)

## Creator

Jie Chen, Hongze Li, Javier Gainza, Angel Munoz, Jose A. Alonso, Jue Liu, Yu-Sheng Chen, [ベリック アレクセイ](https://orcid.org/0000-0001-9031-2355), [山浦 一成](https://orcid.org/0000-0003-0390-8244), Jiaming He, Xinyu Li, John B. Goodenough, J.-S. Zhou

## Rights

[In Copyright](http://rightsstatements.org/vocab/InC/1.0/)

## Other metadata

[Exotic magnetism in perovskite KOsO3](https://mdr.nims.go.jp/datasets/cb0d14d8-84ae-4765-9627-6dc2e09e6bde)

## Fulltext

Exotic Magnetism in Perovskite KOsO3Exotic Magnetism in Perovskite KOsO3Jie Chen ,1 Hongze Li ,1 Javier Gainza ,2 Angel Muñoz ,3 Jose A. Alonso ,2 Jue Liu,4 Yu-Sheng Chen,5Alexei A. Belik,6 Kazunari Yamaura,6,7 Jiaming He,1 Xinyu Li,1 John B. Goodenough,1 and J.-S. Zhou 1,*1Materials Science and Engineering program, Mechanical Engineering, University of Texas at Austin, Austin, Texas 78712 USA2Instituto de Ciencia de Materiales de Madrid, CSIC, Cantoblanco, E-28049 Madrid, Spain3Universidad Carlos III, Avenida Universidad 30, E-28911, Leganés-Madrid, Spain4Neutron Scattering Division, Oak Ridge National Laboratory, Oak Ridge, Tennessee 37830, USA5NSF’s ChemMatCARS, The University of Chicago, Chicago, Illinois 60437, USA6Research Center for Materials Nanoarchitectonics (MANA), National Institute for Materials Science,Namiki 1-1, Tsukuba, Ibaraki 305-0044, Japan7Graduate School of Chemical Sciences and Engineering, Hokkaido University,North 10 West 8, Kita-ku, Sapporo, Hokkaido 060-0810, Japan(Received 28 January 2023; accepted 12 March 2024; published 9 April 2024)A new perovskite KOsO3 has been stabilized under high-pressure and high-temperature conditions. It iscubic at 500 K (Pm − 3m) and undergoes subsequent phase transitions to tetragonal at 320 K (P4=mmm)and rhombohedral (R − 3m) at 230 K as shown from refining synchrotron x-ray powder diffraction(SXRD) data. The larger orbital overlap integral and the extended wave function of 5d electrons in theperovskite KOsO3 allow to explore physics from the regime where Mott and Hund’s rule couplingsdominate to the state where the multiple interactions are on equal footing. We demonstrate an exoticmagnetic ordering phase found by neutron powder diffraction along with physical properties via a suite ofmeasurements including magnetic and transport properties, differential scanning calorimetry, and specificheat, which provide comprehensive information for a system at the crossover from localized to itinerantelectronic behavior.DOI: 10.1103/PhysRevLett.132.156701The Bloch wave function consists of atomic orbitals andplane waves for describing electrons in a crystal. TheCoulomb interaction U between electrons increases theprobability for electrons to stay in atomic orbitals, i.e., morelocalized in real space. A sufficiently large U opens a gapnear the Fermi energy, which leads to a Mott insulator [1].Transition-metal perovskite oxides provide a good play-ground for studying the Mott physics. In addition to theCoulomb interaction, the Hund’s coupling plays an impor-tant role in determining physical properties in the stronglycorrelated systems, especially those with electrons in the t2gorbitals in the cubic crystal field [2]. The threshold Uc forthe Mott transition becomes the lowest for systems withthree electrons in the t2g orbital. In other words, thesesystems are most likely to be Mott insulators. The cubicperovskite SrMnO3 with the electron configuration t32ge0g isa G-type antimagnetic insulator with a Néel temperatureTN ≈ 260 K [3]. Another family of 3d perovskites RCrO3having t32ge0g is also a G-type antiferromagnetic (AFM)insulator with TN in a range from 120 to 298 K [4]. Asorbitals become more extended for 4d electrons, a weakerU in SrTcO3 with t32ge0g has been invoked in an argumentthat the oxide approaches the crossover from localized toitinerant electronic behavior so as to exhibit a higherTN ∼ 1000 K; it also exhibits the G-type AFM order [5,6]and presumably an insulator. The even more extended 5delectrons with a much weakerUmakes AOs5þO3 with t32ge0ga good candidate to test whether the G-type AFM remainsstable at the crossover. The orthorhombic NaOsO3 exhibitstheG-typeAFMorderingwith TN ¼ 410 K [7,8]. However,a reduced TN in NaOsO3 relative to that in SrTcO3 may beattributed to a highly distorted structure that reduces thebandwidth like that in the orthorhombic RCrO3.The cubic perovskite KOsO3 can be stabilized byquenching the high-pressure phase to ambient condition.The detailed information about the sample preparation canbe found in Supplemental Material (SM) [9]. The avail-ability of KOsO3 completes the family of A1þOs5þO3 in abroader range of geometric tolerance factor t¼ðA−OÞ=ðB−OÞ ffiffiffi2p. LiOsO3 synthesized under high pressure has at ¼ 0.86 which is too small for the perovskite structure; itcrystalizes in the LiNbO3 structure [24]. NaOsO3ðt¼ 0.98Þsynthesized under high pressure crystalizes in the ortho-rhombic perovskite structure. NaOsO3 undergoes a metal-insulator transition at TN ¼ 410 K [7] and the G-type AFMhas been determined by neutron diffraction. The metal-insulator transition has been argued as a good example ofSlater transition since there is no obvious change in the cellvolume on crossing the transition [8]. However, a kink inboth a and b axes of the Pbnm orthorhombic cell found inPHYSICAL REVIEW LETTERS 132, 156701 (2024)0031-9007=24=132(15)=156701(7) 156701-1 © 2024 American Physical Societyhttps://orcid.org/0000-0001-9609-669Xhttps://orcid.org/0000-0002-3017-383Xhttps://orcid.org/0000-0002-1999-3116https://orcid.org/0000-0002-1586-8385https://orcid.org/0000-0001-5329-1225https://orcid.org/0000-0002-7667-5640https://crossmark.crossref.org/dialog/?doi=10.1103/PhysRevLett.132.156701&domain=pdf&date_stamp=2024-04-09https://doi.org/10.1103/PhysRevLett.132.156701https://doi.org/10.1103/PhysRevLett.132.156701https://doi.org/10.1103/PhysRevLett.132.156701https://doi.org/10.1103/PhysRevLett.132.156701neutron diffraction on crossing the metal-insulator transi-tion has been well mimicked by a density-functional theorycalculation along increasing of U, which is associated witha change of the local structural distortion. This workindicates that the Mott physics plays an important rolebehind the metal-insulator transition [25]. The Néel temper-ature in a localized electron system increases with increas-ing the orbital overlap integral (OOI). The bandwidth canbe tuned by altering the structural distortion throughchemical substitution. In perovskite AMO3, a t < 1 isaccommodated by the cooperative octahedral-site rotationsthat lower the structural symmetry and bend the bond angleM─O─M from 180º found in the cubic phase for t ¼ 1.The structural change from the orthorhombic to cubic phaseenhances OOI, therefore the bandwidth. The structuralchange from the orthorhombic NaOsO3 to the cubic KOsO3broadens up the electron bandwidth. Therefore, studyingthe evolution of magnetism from NaOsO3 to KOsO3 willilluminate the enigma of magnetism at the crossover.The structural study by SXRD indicates that KOsO3undergoes subsequently phase transitions from the cubic C(Pm − 3m) to the tetragonal at 320 K T (P4=mmm) and tothe rhombohedral phase R (R − 3m) at 230 K as shown bylattice parameters versus temperature in Fig. 1(a). There isno thermal hysteresis loop in the structure on crossing thesephase transitions. The detailed information of the structuralrefinement can be found in Figs. S1–S4 and Table S1–S6in SM [9]. There are extremely small local structuraldistortions associated with the two phase transitions. TheOs─O bond length along the c axis is slightly smaller thanthat in the a − b plane in the T phase and the O─Os─Obond angle deviates slightly from 90º in the R phase.Neutron powder diffraction (ND) has been performed in thesame temperature range. Although the very tiny peaksplitting associated with the phase transitions detectedby SXRD cannot be distinguished in the ND, the com-parison of patterns between SXRD and ND at 500 K inFig. 1(b) unambiguously illustrate the magnetic diffrac-tions, which indicates a magnetically ordered phase inKOsO3 at least at 500 K. Further justifications that the extrapeaks from ND are from magnetic ordering are given inFigs. S5–S9 of SM [9].We start with the magnetic structure in the cubic phase at500 K. The initial fitting with magnetic structures havingk ¼ 0 fails. The possible magnetic structures are given bythe basis vectors of the irreducible representations of thelittle group Gk, which is formed by the symmetry elementsof the space group Pm − 3m that leave k invariant. For thecubic KOsO3, there is only one Os atom per unit cell, so thebasis vectors can only establish some relationships betweenthe mx, my, and mz components of the magnetic moment.For k1 ¼ ðkx; ky; kzÞ, the little group Gk1 is only formed bythe identity, in which there is only one irreducible repre-sentation and there is no relationship between the mx, my,and mz components [15]. The basis vectors of three phasesare presented in Table S7 [9]. A good agreement betweenthe observed and calculated neutron diffraction patterns inbank1 at 500 K based on a model of spin structure isobtained and shown in Figs. 1(b) and 1(c) for a zoomed-inplot at magnetic peaks, and Fig. S10 for the fitting results inbank2. The magnetic structure with a sinusoidally modu-lated moment and the maximum moment M ∼ 1.5 μB isincommensurate to the lattice; a schematic plot of the spinstructure projected in the a − c plane of the cubic cell isgiven in an inset of Fig. 1 and the magnetic momentdirection is shown in Fig. S11. [9] The important param-eters of fitting the neutron diffraction patterns at 500 K areshown in Table S8 [9]. The model of magnetic structure inFIG. 1. (a) Lattice parameters versus temperature for KOsO3obtained by Rietveld refinement of the synchrotron XRD profile(λ ¼ 0.61928 Å) collected on warming and cooling processes.The data acquired from warming up are denoted by red markers,while data from cooling down are represented by blue markers.(b) Neutron powder diffraction (ND), synchrotron x-ray diffrac-tion at 500 K, and the fitting results of ND (the reflections fromtop to bottom represent nuclear structure of KOsO3, nuclearstructure of Os, and magnetic structure of KOsO3); the inset is thespin structure from the fitting. (c) The zoomed-in plot of ND andthe simulation.PHYSICAL REVIEW LETTERS 132, 156701 (2024)156701-2T and R phase is similar to that in C phase, but has slightlydifferent k. The fitting results with the model are shown inFig. S10 and fitting parameters are in Table S9 [9].The nonlinear M − H curves in Fig. 2(a) from 300 to680 K and the hysteresis loop in Fig. S12 of SM [9] confirmthe magnetically ordered phase. The magnetization atH ¼ 2 T and room temperature is comparable to that inNaOsO3 [7] in which the small spontaneous magnetizationwas attributed to the spin canting allowed by the G-typeAFMordering in thePbnm perovskite structure. The samplestarts to decompose at T > 600 K from the magnetizationmeasurement, which prevents us from determining TNaccurately. However, extrapolating the curve of coerciveforce versus temperature (Fig. S12 of SM [9]) to highertemperatures can give a rough estimation of TN ∼ 800�100 K. The structural phase transitions at 230 K and 320 Kcorrespond to anomalies at these temperatures in the temper-ature dependence of magnetization in Fig. 2(b), which alsohas the same magnitude as that of NaOsO3 below TN [7].Since all the crystal structural of three phases found inKOsO3 are not compatible with spin canting, the nonlinearmagnetization must be attributed to a weak ferromagnetism,which will be further elaborated below.Measurements of resistivity in Fig. 2(c) confirm thatKOsO3 is a poormetal with an overall resistivity comparablewith that of the metallic phase at T > TIM in NaOsO3.However, the subsequent phase transitions from C to T to Rphase cause jumps in the resistivity. There are correspondinganomalies in the temperature dependence of thermoelectricpower S in Fig. 2(e); S drops abruptly and changes sign oncooling from C phase to T phase whereas the magnitude ofjSj reduces through the transition from T phase to R phase.The overall magnitude and the temperature dependence of Sare consistentwith ametal with the Fermi level located in themiddle of π� symmetry band shown in Fig. S14 [9,26]. Thespecific heat result in Fig. 2(d) can fit well to the formula ofDebye-Einstein model at all temperatures and CpðTÞ ¼γTþ βT3 at low temperatures.A γ ¼ 2.54 mJ=molK2 fromthe fitting result is the smallest one for a metal; a γ ¼4.25 mJ=molK2 is obtained from the band structure inFig. S13 [9]. The λ-shape anomaly in CpðTÞ, typical for thesecond-order phase transition can be clearly seen at 230 Kand 320 K. The entropy change ΔS ¼ 1.7 Jmole−1K−1associated with the C − T phase transition and the resultsof differential scanning calorimetry (DSC) measurementfrom 100–500 K are given in Figs. S15 and S16 of SM [9].Comparing with a ΔS ¼ 0.6 Jmole−1K−1 associated withthe orthorhombic to the rhombohedral structural transitionin the perovskite LaGaO3 [27], a huge ΔS at the C − Ttransition must be due to some change of electronic degreeof freedom.The spin canting due to theDzyaloshinskii-Moriya theoryoccurs only in some structural systems with lower sym-metries, such as the orthorhombicPbnm perovskite [28,29].The abrupt cancellation of canted moment from the ortho-rhombic to cubic perovskite in G-typeAFMCa1−xSrxMnO3is a good example [3]. None of the three phases found inKOsO3 is compatible with canted spins. The coexistenceof the magnetic structure with a sinusoidally modulatedmoment and conducting electrons in the perovskite KOsO3fits the prediction of the Hubbard Hamiltonian at thecrossover [30]. In this theory, the effect of U is not to splita band into the upper Hubbard band (UHB) and the lowerFIG. 2. (a) Field dependence of the magnetization of KOsO3 at different temperatures; temperature dependences of (b) themagnetization, (c) the resistivity, (d),(f) the specific heat, and (e) the thermoelectric power. The symbols in (c) represent the resultsmeasured on a cluster of KOsO3 crystal (WU, warm up; CD, cool down); the solid line shows the result of pressed powder sample in adiamond anvil cell.PHYSICAL REVIEW LETTERS 132, 156701 (2024)156701-3Hubbard band (LHB) in one step, but to transfer the spectralweight around the Fermi energy EF of free electrons intoUHB and LHB continuously. At the crossover, there arecoherent electrons near EF and localized electrons in LHBresponsible for static magnetic moments. The nonlinearmagnetization in the KOsO3 can be well attributed to theweak itinerant electron ferromagnetism [28,31] contributedfrom the coherent electrons at EF.The evolution of magnetism in transition-metal perov-skites can be demonstrated along pathways of either thestructural distortions or the broadening of the d orbitaldistribution function from 3d to 4d to 5d orbitals in systemswith the t32ge0g configuration. In the perovskite structure, b ofOOI is related to the orthorhombic structural distortion,specifically, the M─O─M bond angle ψ ¼ 180 − ϕ by therelationship of the bandwidthW ∼ b ∼ cosϕ. The transitionfrom the orthorhombic phase (ψ ¼ 180 − ϕ) to the cubicphase (ψ ¼ 180) enhancesW (given a constant M─O bondlength). The magnetic transition temperature, therefore, canbe linked to the structural parameter through the perturba-tion expression of superexchange interaction (SEI) [32],TN ∼ b2=U ∼ cos2ϕ=U, which implies lines with a nearlyidentical slope for different perovskite systems with local-ized electrons in the plot of ln(TN) versus ln(cosϕ).The systems having t32ge0g give the highest possible spinstate available for comparing the change of magneticproperties in perovskites with 3d, 4d, and 5d transitionmetals. In the 3d oxides, there are two families of RCrO3 [4]and Ca1−xSrxMnO3 suitable [3]. The G-type AFM orderingidentified in these two families is in line with the rules forSEI. The rare-earth substitution in the orthorhombic per-ovskite RCrO3 changes theCr─O─Cr bond angle from145ºto 160º for R ¼ Lu to La [4]. Data points from all membersin this perovskite family fall almost perfectly on a line in theplot of lnðTNÞ versus lnðcosϕÞ in Fig. 3(a). For the series ofCa1−xSrxMnO3, a line with nearly the same slope as that forRCrO3 can be obtained if it is drawn between the spots forCaMnO3 and SrMnO3. A slightly lower TN relative to theline for the Ca1−xSrxMnO3 with 0 < x < 1 can be wellaccounted for by the size variance effect [33,34]. This plotconfirms that the perturbation expression [32] of SEI isapplicable to describe the magnetic interactions in these3d perovskite systems with localized electrons. Moreover,the structural change from the orthorhombic phase to thecubic phase adds a clear restriction on the spin structure.The canted spin structure is allowed in the orthorhombicstructure based on either the DMI [35] or the single ionanisotropy [15], which is represented by a spontaneousmagnetization along the canted spin direction and a divergeincrease of M(T) on cooling through TN in CaMnO3. Thesefeatures disappear in the cubic SrMnO3 [3,36].The 4d perovskites CaTcO3 [38] and SrTcO3 [5] show thehighest magnetic transition temperature in transition-metaloxides due to the much extended 4d orbitals. The colinearG-type spin structure found in CaTcO3 and SrTcO3 andthe positive slope of a line connecting CaTcO3 and SrTcO3in the plot of lnðTNÞ versus lnðcosϕÞ in Fig. 3(a) areconsistent with the scenario of localized electrons in theseoxides. However, a reducedmagneticmoment relative to the3d perovskites in Fig. 3(b) reflects the extended wavefunction for the 4d electrons.While an even higher TN is expected in AOs5þO3ðA ¼alkaline elementsÞ because of a larger OOI through evenmore extended 5d orbitals and a weaker U based on SEI,FIG. 3. (a),(b) The correlation between the Néel temperature TN and the magnetic moment at 4.2 K and the structural distortion in theperovskite structure for the 3d, 4d, and 5d perovskites with the t32ge0g electronic configuration. The angle ϕ measures the deviation ofbond angle M─O─M from 180º in the orthorhombic perovskite structure. A in the formula stands for alkaline elements, AE for alkalineearth elements, and R for rare-earth elements. The magnetic moments in RCrO3 are from Ref. [37]; the moment of SrMnO3 at 77 K isfrom Ref. [36], which is converted to the moment at 4.2 K following the Brillouin function (J ¼ 3=2); the moment of Ca1−xSrxMnO3 isfrom Ref. [3]; the moment of CaTcO3 is from Ref. [38]; the moment of SrTcO3 is from Ref. [5]; the moment of NaOsO3 at 200 K is fromRef. [8], which is converted to the moment at 4.2 K following the Brillouin function (J ¼ 3=2). (c) A schematic phase diagram shows theevolution of the magnetic transition temperature as a function of electron bandwidth from localized electrons to itinerant electrons.PHYSICAL REVIEW LETTERS 132, 156701 (2024)156701-4the strong spin-orbit coupling associated with Osðz ¼ 78Þcompared with Tcðz ¼ 43Þ competes with the spin-spininteraction [39,40]. The spin-orbit coupling effect shouldbe negligible for the L − S coupling as the orbital momen-tum is quenched for Os5þðt32ge0gÞ. However, the j − jcoupling that leads to a Jeff ¼ 1=2 state in Sr2IrO4 [41],still places the spin-orbit coupling on the same footing withthe spin-spin interaction in these 5d oxides so as to lower themagnetic transition temperature. A large change of the bondangle Os─O─Os from the orthorhombic NaOsO3 to thecubic KOsO3 separates these two perovskites in the plot oflnðTNÞ versus lnðcosϕÞ in Fig. 3(a). The nearly same slopefor the line connection NaOsO3 and KOsO3 as those forRCrO3 andAEMnO3 in Fig. 3(a)may be an accident. But thepositive slope for the TN change between NaOsO3 andKOsO3 in the plot and the G-type magnetic ordering inNaOsO3 fit the picture of localized electrons inNaOsO3. Thetransport properties and the specific heat in Figs. 2(c)–2(f)clearly place KOsO3 on the itinerant electron side of thecrossover. Two important findings from this Letter are, first,on topof themomentmodulation, the spin arrangement in theac plane creates the interfaces of AFM coupling that isincommensurate with the lattice. The incommensurate spinstructures reported in the literature refer to the spiral spinsthat travel in the lattice transversely [42–47]. The unprec-edented incommensurate moment variation found in KOsO3can be treated as an example of electronic structure at thecrossover. The average moment of the sinusoid wave of themoment at Os sites in KOsO3 is nearly the same as that inhomogeneous NaOsO3. Second, whereas there is a sharpchange of themagnetic structure from theG-typeAFMfoundin the orthorhombic NaOsO3 to the incommensurate spins inthe cubic KOsO3, both perovskites show the colinear spinarrangement. The orthorhombic distortion in a perovskitestructure allows or induces [48] the spin canting that leads toa spontaneous magnetization in the G, A, and C types ofAFM [15]. The canted spins in NaOsO3 account for thenonlinear magnetization. However, the spontaneous mag-netization found in the KOsO3 in Fig. 2(a) and the clearcoercive force plotted in Fig. S12 comes from the weakitinerant electron ferromagnetism in coherent electronicstates near EF.The M─O bonding covalency or the d − p hybridizationin the transition-metal oxide perovskites reduces themagnetic moment in spin ordered structures. Figure 3(b)displays the change of magnetic moment in four families ofperovskites with 3d, 4d, and 5d electrons. The 3d per-ovskites exhibit the saturation moments M ¼ 2.9–2.5 μBthat are close to 3μB for the spin-only value for localizedelectrons. A reduction to M ¼ 2.1 μB in SrTcO3 has beenaccounted for by the bonding covalency in a first-principlescalculation [6]. An even smaller M ¼ 1.09 μB in NaOsO3is consistent with the enhancement of the bonding cova-lency for 5d electrons. The crystal field on t2 orbitalsbecomes ψ t ¼ Ntðft − λπφπÞ after including the covalentmixing with the anion; where Nt is the normalizationconstant, ft is d orbital, φπ is a t2-symmetrized pπ orbital.The moment reduction of the d orbital is caused by thecovalent mixing λπ ¼ bca=ðEt − EπÞ, where bca is theorbital overlap integral in M─O bonds, Et is the bottomenergy in the band for the t2 electrons, and Eπ is the topenergy in the oxygen 2p band. For the orthorhombicperovskites, the overlap integral over the M─O─M bondbecomes bπ ¼ επλ2π cosϕ. Given the nearly constant M─Obond length for the perovskites discussed here, the increaseof bπ comes from a monotonic reduction of Δ ¼ Et − Eπand a significant increase of bca due to the change of radialdistribution function of d orbital from 3d to 4d to 5d. Thechange of bπ between NaOsO3 and KOsO3 is caused by thestructural factor cosϕ. This change plus a sufficiently largeλπ as illustrated by a strong Os t2g∶ O2p hybridization asindicated in the density-functional theory calculation ofFig. S14 for KOsO3 makes it the first perovskite with thet32ge0g configuration at the crossover. The schematic plot ofTN versus bπ in Fig. 3(c) builds in the effects on thebandwidth enhancement from both the structure and theradial distribution function of different orbitals.In conclusion, the electron bandwidth in the perovskitesystems can be varied by the structural distortion and thecovalent mixing between the d − p orbitals, which issensitive to orbital distribution function. The synthesis ofcubic KOsO3 completes a group of perovskites with thehighest spin state t32ge0g crossing 3d, 4d, and 5d of transitionmetals, which enables the study of evolution of magnetism.The Hund’s coupling effect makesmost of these perovskitesa Mott insulator. Whereas the antiferromagnetic phase inthe 3d perovskites can be described by the HeisenbergHamiltonian beautifully; much enhanced TN and reducedmagnetic moment in the 4d and 5d perovskites fit predic-tions of the simulation of the Hubbard Hamiltonian. KOsO3has the highest electron bandwidth, which places it at thecrossover from localized to itinerant electronic behavior.The incommensurate magnetic moment revealed by neutrondiffraction and anomalous physical properties in connectionwith the unusual structural changes advance our knowledgeof a system at the crossover.This research was primarily supported by the NationalScience Foundation through the Center for Dynamics andControl of Materials: an NSF MRSEC under CooperativeAgreement No. DMR-1720595 and DMR-2308817. J. B. G.was supported by the Welch Foundation (F-1066). J. G.thanks Ministerio de Ciencia e Innovación (MICINN) forgranting the contract PRE2018-083398. J. A. A. thanksthe Spanish Ministry of Science and Innovation forgranting the project No. PID2021-122477OB-I00. NSF’sChemMatCARS, Sector 15 at the Advanced Photon Source(APS), Argonne National Laboratory (ANL) is supported bythe Divisions of Chemistry (CHE) and Materials Research(DMR),National Science Foundation, underGrantNo.NSF/CHE-1834750. Use of APS, an Office of Science UserPHYSICAL REVIEW LETTERS 132, 156701 (2024)156701-5Facility operated for the U.S. Department of Energy (DOE)Office of Science by ANL, was supported by the U.S. DOEunderContractNo.DE-AC02-06CH11357.Neutron powderdiffraction measurements used resources at the SpallationNeutron Source (NOMAD instrument), a DOE Office ofScience User Facility operated by the Oak Ridge NationalLaboratory. MANA is funded by MEXT’s WPI, Japan.Synchrotron radiation at SPring-8 (BL02B2)was used underJapan Synchrotron Radiation Research Institute approval(2023A2361, 2023B1676). All simulations were performedon the high-performance computational clusters of TexasAdvanced Computing Center (TACC). J. S. Z. is grateful forthe enlightening discussion with Allan MacDonald.*Corresponding author: jszhou@mail.utexas.edu[1] Masatoshi Imada, Atsushi Fujimori, and Yoshinori Tokura,Metal-insulator transitions, Rev. Mod. Phys. 70, 1039(1998).[2] Luca de’ Medici, Jernej Mravlje, and Antoine Georges,Janus-faced influence of Hund’s rule coupling in stronglycorrelated materials, Phys. Rev. Lett. 107, 256401 (2011).[3] O. Chmaissem, B. Dabrowski, S. Kolesnik, J. Mais,D. E. Brown, R. Kruk, P. Prior, B. Pyles, and J. D.Jorgensen, Relationship between structural parametersand the Neel temperature in Sr1−xCaxMnO3ð0x < 1Þ andSr1−yBayMnO3ðy < 0.2Þ, Phys. Rev. B 64, 134412 (2001).[4] J. S. Zhou, J. A. Alonso, V. Pomjakushin, J. B.Goodenough, Y. Ren, J. Q. Yan, and J. G. Cheng, Intrinsicstructural distortion and superexchange interaction in theorthorhombic rare-earth perovskites RCrO3, Phys. Rev. B81, 214115 (2010).[5] Efrain E. Rodriguez, Frédéric Poineau, Anna Llobet,Brendan J. Kennedy, Maxim Avdeev, Gordon J.Thorogood, Melody L. Carter, Ram Seshadri, David J.Singh, and Anthony K. Cheetham, High temperature mag-netic ordering in the 4d perovskite SrTcO3, Phys. Rev. Lett.106, 067201 (2011).[6] Jernej Mravlje, Markus Aichhorn, and Antoine Georges,Origin of the high Néel temperature in SrTcO3, Phys. Rev.Lett. 108, 197202 (2012).[7] Y. G. Shi, Y. F. Guo, S. Yu, M. Arai, A. A. Belik, A. Sato, K.Yamaura, E. Takayama-Muromachi, H. F. Tian, H. X. Yang,J. Q. Li, T. Varga, J. F. Mitchell, and S. Okamoto, Continu-ous metal-insulator transition of the antiferromagneticperovskite NaOsO3, Phys. Rev. B 80, 161104(R) (2009).[8] S. Calder, V. O. Garlea, D. F. McMorrow, M. D. Lumsden,M. B. Stone, J. C. Lang, J. W. Kim, J. A. Schlueter, Y. G.Shi, K. Yamaura, Y. S. Sun, Y. Tsujimoto, and A. D.Christianson, Magnetically driven metal-insulator transitionin NaOsO3, Phys. Rev. Lett. 108, 257209 (2012).[9] See Supplemental Material at http://link.aps.org/supplemental/10.1103/PhysRevLett.132.156701 for the in-formation about the material preparation, patterns of x-rayand neutron diffraction, and the results of Rietveld refine-ment, physical property measurements, and first-principlescalculations, which includes Refs. [10–23].[10] Fujio Izumi, A Rietveld-analysis program RIETAN-98 andits applications to zeolites, Mater. Sci. Forum 321, 198(2000).[11] Koichi Momma and Fujio Izumi, VESTA 3 for three-dimensional visualization of crystal, volumetric, and mor-phology data, J. Appl. Crystallogr. 44, 1272 (2011).[12] Jörg Neuefeind, Mikhail Feygenson, John Carruth, RonHoffmann, and Kenneth K. Chipley, The nanoscale orderedMAterials Diffractometer NOMAD at the Spallation Neu-tron Source SNS, Nucl. Instrum. Methods Phys. Res., Sect.B 287, 68 (2012).[13] S. Calder, K. An, R. Boehler, C. R. Dela Cruz, M. D.Frontzek, M. Guthrie, B. Haberl, A. Huq, S. A. J.Kimber, J. Liu, J. J. Molaison, J. Neuefeind, K. Page,A. M. dos Santos, K. M. Taddei, C. Tulk, and M. G.Tucker, A suite-level review of the neutron powder dif-fraction instruments at Oak Ridge National Laboratory, Rev.Sci. Instrum. 89, 092701 (2018).[14] Juan Rodríguez-Carvajal, Recent advances in magneticstructure determination by neutron powder diffraction,Physica (Amsterdam) 192B, 55 (1993).[15] E. F. Bertaut, Spin configurations of ionic structures:Theory, and practice, in Magnetism, edited by G. T.Rado and H. Suhl (Acedemic Press, New York and London,1963), Vol. III, p. 149.[16] G. Kresse and J. Furthmüller, Efficient iterative schemes forab initio total-energy calculations using a plane-wave basisset, Phys. Rev. B 54, 11169 (1996).[17] G. Kresse and D. Joubert, From ultrasoft pseudopotentialsto the projector augmented-wave method, Phys. Rev. B 59,1758 (1999).[18] G. Kresse and J. Hafner, Ab initio molecular dynamics forliquid metals, Phys. Rev. B 47, 558 (1993).[19] G. Kresse and J. Hafner, Ab initio molecular-dynamicssimulation of the liquid-metal–amorphous-semiconductortransition in germanium, Phys. Rev. B 49, 14251 (1994).[20] Uthpala Herath, Pedram Tavadze, Xu He, Eric Bousquet,Sobhit Singh, Francisco Muñoz, and Aldo H. Romero,PyProcar: A Python library for electronic structure pre/post-processing, Comput. Phys. Commun. 251, 107080 (2020).[21] Anton Kokalj, Computer graphics, and graphical userinterfaces as tools in simulations of matter at the atomicscale, Comput. Mater. Sci. 28, 155 (2003).[22] N. E. Brese and M. O’Keeffe, Bond-valence parameters forsolids, Acta Crystallogr. Sect. B 47, 192 (1991).[23] Jun-Ichi Yamaura, Shigeki Yonezawa, Yuji Muraoka, andZenji Hiroi, Crystal structure of the pyrochlore oxidesuperconductor KOs2O6, J. Solid State Chem. 179, 336(2006).[24] P. W. Anderson and E. I. Blount, Symmetry considerationson martensitic transformations: “Ferroelectric” metals?,Phys. Rev. Lett. 14, 217 (1965).[25] J. S. Zhou, X. Y. Li, and M. D. Johannes, Evidence ofelectronic structural change at the metal-insulator transitionin the perovskite NaOsO3, Phys. Rev. B 108, 195115(2023).[26] D. K. C. MacDonald, Thermoelectricity: An Introduction tothe Principles (John Wiley & Sons, New York, 1962).[27] T. Shibasaki, T. Furuya, J. Kuwahara, Y. Takahashi, H.Takahashi, and T. Hashimoto, Exploration of high pressurePHYSICAL REVIEW LETTERS 132, 156701 (2024)156701-6https://doi.org/10.1103/RevModPhys.70.1039https://doi.org/10.1103/RevModPhys.70.1039https://doi.org/10.1103/PhysRevLett.107.256401https://doi.org/10.1103/PhysRevB.64.134412https://doi.org/10.1103/PhysRevB.81.214115https://doi.org/10.1103/PhysRevB.81.214115https://doi.org/10.1103/PhysRevLett.106.067201https://doi.org/10.1103/PhysRevLett.106.067201https://doi.org/10.1103/PhysRevLett.108.197202https://doi.org/10.1103/PhysRevLett.108.197202https://doi.org/10.1103/PhysRevB.80.161104https://doi.org/10.1103/PhysRevLett.108.257209http://link.aps.org/supplemental/10.1103/PhysRevLett.132.156701http://link.aps.org/supplemental/10.1103/PhysRevLett.132.156701http://link.aps.org/supplemental/10.1103/PhysRevLett.132.156701http://link.aps.org/supplemental/10.1103/PhysRevLett.132.156701http://link.aps.org/supplemental/10.1103/PhysRevLett.132.156701http://link.aps.org/supplemental/10.1103/PhysRevLett.132.156701http://link.aps.org/supplemental/10.1103/PhysRevLett.132.156701https://doi.org/10.4028/www.scientific.net/MSF.321-324.198https://doi.org/10.4028/www.scientific.net/MSF.321-324.198https://doi.org/10.1107/S0021889811038970https://doi.org/10.1016/j.nimb.2012.05.037https://doi.org/10.1016/j.nimb.2012.05.037https://doi.org/10.1063/1.5033906https://doi.org/10.1063/1.5033906https://doi.org/10.1016/0921-4526(93)90108-Ihttps://doi.org/10.1103/PhysRevB.54.11169https://doi.org/10.1103/PhysRevB.59.1758https://doi.org/10.1103/PhysRevB.59.1758https://doi.org/10.1103/PhysRevB.47.558https://doi.org/10.1103/PhysRevB.49.14251https://doi.org/10.1016/j.cpc.2019.107080https://doi.org/10.1016/S0927-0256(03)00104-6https://doi.org/10.1107/S0108768190011041https://doi.org/10.1016/j.jssc.2005.10.039https://doi.org/10.1016/j.jssc.2005.10.039https://doi.org/10.1103/PhysRevLett.14.217https://doi.org/10.1103/PhysRevB.108.195115https://doi.org/10.1103/PhysRevB.108.195115phase in LaGaO3 and LaCrO3, J. Therm. Anal. Calorim. 81,575 (2005).[28] Tôru Moriya, Theory of itinerant electron magnetism, J.Magn. Magn. Mater. 100, 261 (1991).[29] I. Dzyaloshinsky, A thermodynamic theory of “weak”ferromagnetism of antiferromagnetics, J. Phys. Chem.Solids 4, 241 (1958).[30] M. J. Rozenberg, G. Kotliar, and X. Y. Zhang, Mott-Hub-bard transition in infinite dimensions. II, Phys. Rev. B 49,10181 (1994).[31] E. P. Wohlfarth, Very weak itinerant ferromagnets; applica-tion to ZrZn2, J. Appl. Phys. 39, 1061 (2003).[32] P. W. Anderson, New approach to the theory of super-exchange interactions, Phys. Rev. 115, 2 (1959).[33] J. P. Attfield, A. L. Kharlanov, and J. A. McAllister,Cation effects in doped La2CuO4 superconductors, Nature(London) 394, 157 (1998).[34] Lide M. Rodriguez-Martinez and J. Paul Attfield, Cationdisorder and size effects in magnetoresistive manganeseoxide perovskites, Phys. Rev. B 54, R15622 (1996).[35] Tóru Moriya, Anisotropic superexchange interaction, andweak ferromagnetism, Phys. Rev. 120, 91 (1960).[36] Takayoshi Takeda and Sōji Ōhara, Magnetic structure of thecubic perovskite type SrMnO3, J. Phys. Soc. Jpn. 37, 275(1974).[37] E. F. Bertaut, G. Bassi, G. Buisson, P. Burlet, J. Chappert, A.Delapalme, J. Mareschal, G. Roult, R. Aleonard, R.Pauthenet, and J. P. Rebouillat, Some neutron-diffractioninvestigations at the nuclear center of Grenoble, J. Appl.Phys. 37, 1038 (1966).[38] Maxim Avdeev, Gordon J. Thorogood, Melody L. Carter,Brendan J. Kennedy, Jimmy Ting, David J. Singh, and KiaS. Wallwork, Antiferromagnetism in a technetium oxide.Structure of CaTcO3, J. Am. Chem. Soc. 133, 1654(2011).[39] JunjiroKanamori, Theory of themagnetic properties of ferrous,and cobaltous oxides, I, Prog. Theor. Phys. 17, 177 (1957).[40] Junjiro Kanamori, Theory of the magnetic properties offerrous, and cobaltous oxides, II, Prog. Theor. Phys. 17, 197(1957).[41] B. J. Kim, Hosub Jin, S. J. Moon, J. Y. Kim, B. G. Park,C. S. Leem, Jaejun Yu, T. W. Noh, C. Kim, S. J. Oh, J. H.Park, V. Durairaj, G. Cao, and E. Rotenberg, Novel Jeff ¼1=2 Mott state induced by relativistic spin-orbit coupling inSr2IrO4, Phys. Rev. Lett. 101, 076402 (2008).[42] Wei Bao, P. G. Pagliuso, J. L. Sarrao, J. D. Thompson, Z.Fisk, J. W. Lynn, and R.W. Erwin, Incommensurate mag-netic structure of CeRhIn5, Phys. Rev. B 62, R14621 (2000).[43] W. D. Hutchison, D. J. Goossens, K. Nishimura, K. Mori, Y.Isikawa, and A. J. Studer, Magnetic structure of TbNiAl4, J.Magn. Magn. Mater. 301, 352 (2006).[44] G. Venturini, D. Fruchart, and B. Malaman, Incommensu-rate magnetic structures of RMn6Sn6ðR ¼ Sc;Y;LuÞ com-pounds from neutron diffraction study, J. Alloys Compd.236, 102 (1996).[45] A. Muñoz, J. A. Alonso, M. T. Casais, M. J. Martínez-Lope,J. L. Martínez, and M. T. Fernández-Díaz, The magneticstructure of YMnO3 perovskite revisited, J. Phys. Condens.Matter 14, 3285 (2002).[46] S. A. Ivanov, P. Nordblad, R. Mathieu, R. Tellgren, C. Ritter,N. V. Golubko, E. D. Politova, and M. Weil, New type ofincommensurate magnetic ordering in Mn3TeO6, Mater.Res. Bull. 46, 1870 (2011).[47] Dmitry D. Khalyavin, Roger D. Johnson, Fabio Orlandi,Paolo G. Radaelli, Pascal Manuel, and Alexei A. Belik,Emergent helical texture of electric dipoles, Science 369,680 (2020).[48] J. S. Zhou, L. G. Marshall, Z. Y. Li, X. Li, and J. M. He,Weak ferromagnetism in perovskite oxides, Phys. Rev. B102, 104420 (2020).PHYSICAL REVIEW LETTERS 132, 156701 (2024)156701-7https://doi.org/10.1007/s10973-005-0826-zhttps://doi.org/10.1007/s10973-005-0826-zhttps://doi.org/10.1016/0304-8853(91)90824-Thttps://doi.org/10.1016/0304-8853(91)90824-Thttps://doi.org/10.1016/0022-3697(58)90076-3https://doi.org/10.1016/0022-3697(58)90076-3https://doi.org/10.1103/PhysRevB.49.10181https://doi.org/10.1103/PhysRevB.49.10181https://doi.org/10.1063/1.1656163https://doi.org/10.1103/PhysRev.115.2https://doi.org/10.1038/28120https://doi.org/10.1038/28120https://doi.org/10.1103/PhysRevB.54.R15622https://doi.org/10.1103/PhysRev.120.91https://doi.org/10.1143/JPSJ.37.275https://doi.org/10.1143/JPSJ.37.275https://doi.org/10.1063/1.1708325https://doi.org/10.1063/1.1708325https://doi.org/10.1021/ja109431thttps://doi.org/10.1021/ja109431thttps://doi.org/10.1143/PTP.17.177https://doi.org/10.1143/PTP.17.197https://doi.org/10.1143/PTP.17.197https://doi.org/10.1103/PhysRevLett.101.076402https://doi.org/10.1103/PhysRevB.62.R14621https://doi.org/10.1016/j.jmmm.2005.07.012https://doi.org/10.1016/j.jmmm.2005.07.012https://doi.org/10.1016/0925-8388(95)01998-7https://doi.org/10.1016/0925-8388(95)01998-7https://doi.org/10.1088/0953-8984/14/12/315https://doi.org/10.1088/0953-8984/14/12/315https://doi.org/10.1016/j.materresbull.2011.07.041https://doi.org/10.1016/j.materresbull.2011.07.041https://doi.org/10.1126/science.aay7356https://doi.org/10.1126/science.aay7356https://doi.org/10.1103/PhysRevB.102.104420https://doi.org/10.1103/PhysRevB.102.104420