A superlattice approach to doping infinite-layer nickelates
R. A. Ortiz, H. Menke, F. Misják, D. T. Mantadakis, K. Fürsich, E. Schierle, G. Logvenov, U. Kaiser, B. Keimer, P. Hansmann, E. Benckiser
AA superlattice approach to doping infinite-layer nickelates
R. A. Ortiz, D. T. Mantadakis, F. Misj´ak, K. F¨ursich, E. Schierle, G. Christiani, G. Logvenov, U. Kaiser, B. Keimer, P. Hansmann,
4, 5 and E. Benckiser ∗ Max Planck Institute for Solid State Research, Heisenbergstraße 1, 70569 Stuttgart, Germany University of Ulm, Albert-Einstein-Allee 11, 89081 Ulm, Germany Helmholtz-Zentrum Berlin f¨ur Materialien und Energie,Albert-Einstein-Straße 15, 12489 Berlin, Germany Max Planck Institute for Chemical Physics of Solids, N¨othnitzerstraße 40, 01187 Dresden, Germany Department of Physics, University of Erlangen-N¨urnberg, 91058 Erlangen, Germany (Dated: February 16, 2021)The recent observation of superconductivity in Sr-doped infinite-layer NdNiO thin films hasattracted a lot of attention, since this compound is electronically and structurally analogous tothe high- T c superconducting cuprates. Since phase stabilization upon chemical doping with Sr ischallenging, we synthesized artificial superlattices of LaNiO embedded in insulating LaGaO , andused layer-selective topotactic reactions to reduce the nickelate layers to LaNiO . Hole dopingis achieved via interfacial oxygen atoms and the layer thickness. We used electrical transportmeasurements and x-ray spectroscopy together with ab initio calculations to track changes in thelocal nickel electronic configuration upon reduction and found that these changes are reversible.Experimental and theoretical data indicate that the doped holes are trapped at the interfacialquadratic pyramidal Ni sites, which then form the stable Ni valence state. Calculations forelectron-doped cases predict a different behavior, with evenly distributed electrons among the layers,thus opening up interesting perspectives for interfacial doping of transition metal oxides. Ever since the discovery of unconventional supercon-ductivity in high- T c cuprates, the search is ongoing forother 3 d transition-metal oxides that exhibit this intrigu-ing and not yet understood quantum state of matter.This is especially true for nickelates, because nickel isright next to copper in the periodic table. Recently thefirst observation of superconductivity in heteroepitaxiallygrown thin films of Sr-doped NdNiO has raised a lot ofinterest [1]. Some time passed before these results couldbe reproduced experimentally, which is largely related tothe delicate growth conditions [2]. Similarly to the high- T c cuprates the phase diagram of the infinite-layer nicke-lates shows a dome-like doping dependence for the super-conducting transition temperature T c [2–4]. The optimaldoping with maximum T c ≈
15 K has been reported forNd . Sr . NiO . The synthesis route of this compoundrequires the growth of a precursor Nd . Sr . NiO thinfilm, which is subsequently treated by soft-chemistry re-duction with CaH . The preparation of the precursor filmconstitutes the first challenge, since the Sr-doped per-ovskite phase competes with Ruddlesden-Popper phases,leaving a very narrow growth window [5]. The homogene-ity of the Sr-dopant distribution and oxygen reduction,and the role of the heteroepitaxy with SrTiO imposefurther challenges on the synthesis, with many detailsthat remain unexplored.The soft-chemistry reduction of nickelate powders andthin films was studied for the first time several yearsago [6–12]. Starting from LaNiO , the reduction toLaNiO takes place with the formation of relatively sta-ble phases with intermediate oxygen contents, LaNiO . and LaNiO . [13]. Structurally, the reduction goes alongwith successive removal of apical oxygen ions, such that for the so-called brownmillerite phase LaNiO . , columnsof Ni ions with square-planar coordination are placed be-tween the octahedrally coordinated nickel ions in the per-ovskite phase. The lattice parameters change from nearlycubic 3 .
84 ˚A for LaNiO to enlarged in-plane a = 3 .
96 ˚A(perpendicular to the NiO columns) and reduced out- Reduction with CaH at 280 °C4 unit cells of LaGaO LaGaNiOAnnealing in O at FIG. 1. Left: Sketch of the pristine (re-oxidized) LaNiO (blue)- LaGaO (grey) superlattice with m = 8, n = 4 stack-ing sequences grown on a (001) SrTiO substrate. Right:Structure stabilized by layer-selective reduction of the nick-elate slab with CaH . We label individual layers within thenickelate slab by 1 to 4. The dashed horizontal line indicatesthe mirror plane. a r X i v : . [ c ond - m a t . s up r- c on ] F e b -2 I n t en s i t y ( a . u . ) d (Å -1 ) pristine reduced re-oxidized S T O ( ) S T O ( )( ) ( ) ( ) ( )( ) ( ) ( )( )( ) ( )( ) ( ) (a) R e s i s t i v i t y ( µ W c m ) Temperature (K) pristine reduced re-oxidized pristine re-oxidized reduced (b) * ** FIG. 2. (a) XRD patterns of a set of pristine, reduced and re-oxidized LaNiO x /LaGaO superlattices including the (001)and the (002) reflections of the SrTiO substrate. The aster-isks mark peaks arising from the sample holder. (b) Resistiv-ity of the same set of samples. of-plane c = 3 .
38 ˚A values in LaNiO [6, 10]. The largecollapse of the c -axis parameter makes the reduction pro-cess easily trackable by x-ray diffraction, even for verythin films and multilayers.Theoretical interest in this new, supposedly unconven-tional superconductor naturally arises from the possi-ble similarity to, or insightful differences from high- T c cuprates [14–18]. Some theories predict much weakermagnetic correlations for the infinite-layer nickelates, andthus question their relevance for the superconductingpairing mechanism [19]. Other theoretical studies, how-ever, come to different conclusions [18]. Furthermore,Ni with a formal 3 d valence electron configuration israrely found in bulk compounds, and the hybridizationstrength with the oxygen ligands appears to be weakerthan the one of Cu in cuprates [20]. Last but not least,the role of the rare-earth ions was put into question, sincethe closely analogous (La,Sr)NiO heterostructure wasfound to be not superconducting [1].Here we report on an alternative approach to holedoping infinite-layer nickelates that considers the layer-selective topotactical reduction of LaNiO -LaGaO su-perlattices [21]. The essential idea is that in a layer-selective reduction process the interfacial, apical oxygen I n t en s i t y ( a . u . ) d (Å -1 ) exp. pristine octahedral (0024) (0025) (a) I n t en s i t y ( a . u . ) d (Å -1 ) exp. reduced square planar pyramidal (0024) (b) FIG. 3. QUAD simulations of the XRD data around the (002)SrTiO substrate reflections for different crystal structures ofthe (a) pristine and (b) reduced LaNiO superlattices. Inpanel (b), the simulation, of the proposed quadratic pyra-midal (black box inset) interfacial nickel coordination yieldsa better agreement with the experimental results than thesquare planar (grey box inset). ions remain, forming a structure depicted in Fig. 1. Asimilar interface structure has also been suggested inrecent density functional theory studies [15, 22]. As-suming that the additional charge originating from theinterfacial oxygen ions is homogeneously distributed inthe LaNiO layer stack, this opens the possibility of atunable doping level, where the average nickel valencestate Ni m is controlled by the number of consecutivelayers m (see Fig. 1). Further, the growth of the pre-cursor superlattices is stable and the optimal conditionshave been well established [23]. In the following we fo-cus on (LaNiO ) m =8 /(LaGaO ) n =4 superlattices grownon (001) SrTiO substrates (see Fig. 1). With respectto the bulk perovskite nickelate, the substrate induces amoderate tensile strain that translates into an enlarged c axis parameter for the epitaxially-strained infinite-layerphase. LaGaO was chosen to facilitate layer-selective re-duction, since it is expected to be chemically more stableagainst oxygen removal (the Ga ions have a closed-shell configuration). Further, there is no change of theLa cation sublattice across the interfaces (except from theout-of-plane La atomic distances), which removes a pos-sible source of interfacial disorder. We emphasize thatsuperconductivity was found in analogous infinite-layercuprate superlattices [24].The superlattices were grown by pulsed-laser deposi- (b)(a)(c)
840 860 88001020 1120 1140 1160 1180012525 530 535 540 545 5500123 pristine reduced I n t en s i t y ( a . u . ) Energy (eV)La- M Ni- L Ni- L La- M Ga- L O -K I n t en s i t y ( a . u . ) Energy (eV) I n t en s i t y ( a . u . ) Energy (eV)
FIG. 4. Polarization-averaged XAS spectra ( I average =(2 I E ⊥ c + I E (cid:107) c ) /
3; see Fig. 5) of the energy range cover-ing (a) La- M , and Ni- L , edges, (b) of the Ga- L , -edges(smoothed because of the lower flux at these high energies),and (c) around the O- K edge. In each panel data for pristineand reduced samples are offset for clarity. All spectra weremeasured in total-electron yield detection mode. tion using high-density stoichiometric targets of LaNiO and LaGaO under the conditions reported in Ref. 23.High-resolution transmission electron microscopy im-ages of as-grown samples reveal well-ordered superlat-tice structures with sharp interfaces (see SupplementaryMaterial). In order to reduce the influence of small dif-ferences in superlattice structures, samples cut out of asingle growth were either kept pristine, or reduced. Someof the reduced samples were subsequently re-oxidized (in-set in Fig. 2 (b)). The as-grown superlattices were re-duced using CaH powder as a reduction agent. Spec-imens with dimensions 2 . × . were placed un-der Ar-atmosphere inside an open aluminum foil box ontop of approximately 50 mg CaH powder. The quartztubes were then sealed in vacuum (10 − − − mbar) andsubsequently annealed at 280 ◦ C for 72 h. X-ray diffrac-tion and electrical transport measurements (performed invan-der-Pauw geometry) show that the reduction processis reversible (Fig. 2 (a,b)). Such reversibility of a topotac-tic reduction has been previously reported in LaNiO δ thin films [11] and CaFeO based superlattices [21].While the electrical transport of pristine and re-oxidized samples exhibits metallic behavior closely analo-
855 860 865 870 8750.00.51.0 I n t. d i ff e r en c e Energy (eV) E ^ c E || c E ^ c - E || c I n t en s i t y ( a . u . ) I n t en s i t y ( a . u . ) pristine reduced 𝐸 ⊥ 𝑐 𝐸 ∥ 𝑐
FIG. 5. Polarization-dependent XAS, measured in total-fluorescence yield for pristine and reduced samples across theNi- L , absorption edges, where the La- M lines were previ-ously subtracted. The top and middle panel show the spectrataken with x-ray polarization E (cid:107) c and E ⊥ c , respectively.A sketch of the geometry is shown in the inset of the toppanel. The bottom panel shows the linear dichroism definedby the difference of intensities I E ⊥ c − I E (cid:107) c . gous to bulk LaNiO , the resistivity of the reduced super-lattices shows a semiconducting temperature dependence(Fig. 2 (b)). In several samples we found that the conduc-tivity of the re-oxidized samples is higher than that of thepristine ones. We attribute this finding to the presence ofoxygen vacancies in the as-grown samples. Although inthe last step of the PLD growth samples were annealedin 1-bar oxygen atmosphere at 690 ◦ C was applied, thisannealing of the pristine samples is less efficient than re-duction and subsequent annealing. Whereas the fillingof individual, isolated oxygen vacancies in the annealingprocess of pristine samples is limited by oxygen diffusion,this restriction is lifted when the apical oxygen sites arefirst emptied and then refilled.Detailed analysis of the (00 L ) x-ray diffraction scansand comparison with patterns calculated by the simula-tion software QUAD [25] confirmed the interfacial struc-ture depicted in Fig. 1. In particular, the relative in-tensities of the L = 23 , ,
25 superlattice reflections(high-low-high) are in good agreement with the simu-lation of quadratic pyramidal Ni-O coordination, whilethe monotonically increasing intensity pattern predictedfor square-planar coordination disagrees with the exper-imental data. This characteristic intensity pattern wasobserved in various samples that had been subjected tothe topotactic reduction treatment. electron doping hole doping
L1Ni3d O2p O2p La5d La5dO2pO2pNi3d Ni3d Ni3dL2 L3 L4 L1 L2 L3 L4 L1 L2 L3 L4 L1 L2 L3 L4 a) b)c) d)
FIG. 6. Distribution of doped holes (left panels) and electrons (right panels) for the superlattice structure depicted in the rightpanel of Fig. 1. Panels a) and b) resolve the charge-difference in Ni 3 d orbitals per layer (L1-L4) of the heterostructure. Panelsc) and d) show the distribution of the doped charge carriers within the nickel 3 d , oxygen 2 p and lanthanum 5 d states. To study modifications of the electronic configura-tion of the Ni ions upon reduction we examined soft x-ray absorption spectra across several relevant absorptionedges at room temperature. The measurements were car-ried out at the UE46 PGM-1 beamline of the BESSY-IIsynchrotron at Helmholtz-Zentrum Berlin. The spectraaround the La- M , (3 d → f ) and Ni- L , (2 p → d )absorption edges are shown in Fig. 4 (a). Compared tothe pristine piece of our sample, the center of mass ofthe Ni- L spectrum of the reduced sample is shifted tolower energies by about 1.5 eV, while the La- M and Ga- L lines remain unchanged within in our energy resolution.The absence of a shift and nearly identical line shapesof the Ga- L edge spectra before and after reduction arestrong indications of the layer-selectivity of the reduction(Fig. 4 (b)). The comparison of spectra at the O- K edge,on the other hand, indicates a strongly modified Ni-Ohybridization (Fig. 4 (c)). Perovskite LaNiO is a nega-tive charge-transfer material, where the largest contribu-tion to the Ni ground state configuration is 3 d L (where L denotes a ligand hole) [26]. This strong hybridiza-tion with nearly one hole in the oxygen 2 p states resultsin a pronounced O- K edge pre-peak around 528 eV. Inthe spectrum of the reduced superlattice, the spectralweight shifts towards higher energies and decreases con-siderably, indicating a loss of hybridization. A relatedeffect has been observed in resonant inelastic x-ray scat-tering (RIXS) experiments on NdNiO and in model cal-culations predicting that the infinite-layer nickelates fallinto the Mott-Hubbard regime, where the lowest-energyelectron-addition states are of Ni 3 d character [27, 28].Next we turn to the polarization dependence of theXAS spectra. The spectra of pristine LaNiO are nearlyisotropic, because the two d -shell holes on the octahe- drally coordinated Ni ions are approximately evenlydistributed over the d x − y - and d z − r -based orbitals[23]. In reduced samples, on the other hand, a strong dif-ference between XAS spectra with x-rays polarized along(I E (cid:107) c ) and perpendicular to the c -axis (I E ⊥ c ) is expectedto arise from the single hole on the Ni ions in planarcoordination, which favors the d x − y orbital (inset inFig. 5). In line with this expectation, we only observeda weak negative dichroism (defined as I E ⊥ c -I E (cid:107) c ) in thepristine samples (blue lines in Fig. 5), which can be at-tributed to tensile epitaxial strain from the SrTiO sub-strate [23]. As expected from the change in electron fill-ing and crystal field, the dichroism of the reduced sampleis greatly enhanced and shows a change in sign. Interest-ingly, its line shape is remarkably similar to the dichroismof Cu in high- T c cuprates [29].In order to further analyze the interfacial self-doping ofthe heterostructure we performed ab initio density func-tional plus dynamical mean-field calculations. We re-laxed the ionic positions of the reduced unit-cell shownin Fig. 1 on the right hand side with the Vienna ab ini-tio simulation package (VASP) code [30] using the gen-eralized gradient approximation (GGA) functional [31].The subsequent Wannier-projection on a large Ni 3 d , O2 p , and La 5 d basis was performed with the Wannier90package [32]. For the dynamical mean-field calculationswe used the TRIQS library [33] with a continuous timequantum Monte-Carlo solver in the hybridization expan-sion version [34]. In the chosen dp -basis, which includesexplicitly oxygen 2 p and La 5 d states we define the lo-cal interaction operator in the Kanamori parametriza-tion with Hubbard U = 8 . J H = 0 . U (cid:48) = U − J H ). As usual the total charge of the lat-tice model remains fixed within the DMFT self-consistentcalculation and allows for a straight forward procedureto include doping (approximating that the dopand hasa negligible effect on the structure). Further calculationdetails can be found in [35].Our DFT+DMFT study concludes two main messages:i) Doped holes will get trapped at the interface due to aHund’s rule mechanism which is well known from nomi-nal Ni compounds while doped electrons will be evenlydistributed among the layers despite a small leakage toLa 5 d states. ii) Despite of the Hund’s trapping in theouter layer, the inner layers are practically well describedby a single-band Hubbard model also in hole and electrondoped cases.To be more specific, we now discuss the distribution ofadditionally doped electrons (holes) resolved in layer andorbital degrees of freedom. In Fig. 6 we show the distri-bution of the doped charge ∆ ρ = | ρ − ρ | as histogramsfor two values (∆ ρ tot. ∈ { , } per unit cell) of absoluteelectron and hole doping, respectively. The upper panelsa) and b) show a layer resolved ∆ ρ Ni , the lower panelsc) and d) the layer-integrated orbital distribution of thedoped charge.Without any extra charges the “inner” three lay-ers with planar quadratic nearest-neighbor coordinationhave an almost ionic-limit filling of oxygen and nickelstates with a single half-filled x − y band at the Fermilevel. The nickel ion at the interface which resides in thebasal plane of a quadratic pyramid of oxygen has a simi-lar filling. However, due to its coordination the 3 z − r orbital can hybridize strongly with the p z orbital of theapical oxygen forming an anti-bonding linear combina-tion, which is pushed just below the Fermi level in starkcontrast to the other nickel sites. Indeed, this higher ly-ing “axial” orbital of the interface nickel is so close tothe Fermi level that - together with the Hund’s energy- it will trap doped holes to form a local high-spin con-figuration. This can be seen quantitatively in Fig. 6 (a)which shows that there are more doped holes in the in-terface layer (L1) than in all other layers combined. Onthe electron doped side, panel (b), we observe a quasi ho-mogeneous distribution of the extra charge in all layers.We can draw another important conclusion from the dis-tribution of doped charge when we resolve by element.In panels (c) and (d) of Fig. 6 we see that the majorpart of doped charges resides in nickel 3 d orbitals. Theamount of extra oxygen charge on the hole-doped side isexpected due to a remaining finite hybridization of Ni3 d and O2 p orbitals, but is quantitatively small compared tocharge-transfer systems like high- T c cuprates or nominalNi systems. On the electron-doped side in panel (d) wesee some “leakage” to O2 p which are part of anti-bondinghybrids and La5 d states. At this point we have, however,no indication of major effects from the La5 d states. Ofcourse we cannot exclude their importance at very lowenergies as they clearly contribute to the Fermi surface. Despite having a very similar doping level as the su-perconducting Nd . Sr . NiO films reported in Ref. [1],the superlattices we have investigated experimentally donot exhibit superconductivity (Fig. 2). Our electronic-structure calculations yield an explanation for this strik-ing difference and provide a very interesting outlook forthe superlattice approach with electron-doping.In summary, we have shown that reversible, layer-selective removal of oxygen ions yields superlattices withwell-ordered infinite-layer nickelate stacks, and we usedx-ray diffraction to characterize the interfacial structure.Our combined results from x-ray spectroscopy and ab ini-tio calculations based on the experimentally determinedstructure show that holes donated by interfacial oxy-gen atoms remain trapped near the interfaces with theband insulator. However, the same calculations predictthat additional electrons spread out uniformly across theLaNiO layer stack, opening up interesting perspectivesfor interfacial doping of nickelates and other transition-metal oxide superlattices.We acknowledge financial support by the Center forIntegrated Quantum Science and Technology (IQ ST )and the Deutsche Forschungsgemeinschaft (DFG, Ger-man Research Foundation): Projektnummer 107745057- TRR 80 and Projektnummer KA 1295/26-2. We thankHZB for the allocation of synchrotron radiation beam-time. We thank R. Merkle and M. Klett for insightfuldiscussions. ∗ [email protected][1] D. Li, K. Lee, B. Y. Wang, M. Osada, S. Crossley, H. R.Lee, Y. Cui, Y. Hikita, and H. Y. Hwang, Nature ,624 (2019).[2] S. Zeng, C. S. Tang, X. Yin, C. Li, M. Li, Z. Huang,J. Hu, W. Liu, G. J. Omar, H. Jani, Z. S. Lim, K. Han,D. Wan, P. Yang, S. J. Pennycook, A. T. S. Wee, andA. Ariando, Phys. Rev. Lett. , 147003 (2020).[3] D. Li, B. Y. Wang, K. Lee, S. P. Harvey, M. Osada, B. H.Goodge, L. F. Kourkoutis, and H. Y. Hwang, Phys. Rev.Lett. , 027001 (2020).[4] M. Osada, B. Y. Wang, K. Lee, D. Li, and H. Y. Hwang,Phys. Rev. Materials , 121801 (2020).[5] K. Lee, B. H. Goodge, D. Li, M. Osada, B. Y. Wang,Y. Cui, L. F. Kourkoutis, and H. Y. Hwang, APL Ma-terials , 041107 (2020).[6] M. Crespin, P. Levitz, and L. Gatineau, J. Chem. Soc.,Faraday Trans. 2 , 1181 (1983).[7] P. Levitz, M. Crespin, and L. Gatineau, J. Chem. Soc.,Faraday Trans. 2 , 1195 (1983).[8] M. A. Hayward, M. A. Green, M. J. Rosseinsky, andJ. Sloan, Journal of the American Chemical Society ,8843 (1999).[9] M. Crespin, O. Isnard, F. Dubois, J. Choisnet, andP. Odier, Journal of Solid State Chemistry , 1326(2005).[10] Y. Shimakawa, S. Inoue, M. Kawai, N. Ichikawa, M. Mizumaki, and N. Kawamura, MRS Proceedings , 1148-PP07 (2008).[11] M. Kawai, S. Inoue, M. Mizumaki, N. Kawamura,N. Ichikawa, and Y. Shimakawa, Applied Physics Letters , 082102 (2009).[12] D. Kaneko, K. Yamagishi, A. Tsukada, T. Manabe, andM. Naito, Physica C: Superconductivity , 936 (2009),proceedings of the 21st International Symposium on Su-perconductivity (ISS 2008).[13] T. Moriga, O. Usaka, I. Nakabayashi, T. Kinouchi,S. Kikkawa, and F. Kanamaru, Solid State Ionics , 252(1995), proceedings of the 20th Commemorative Sympo-sium on Solid State Ionics in Japan.[14] F. Lechermann, Phys. Rev. B , 081110 (2020).[15] F. Bernardini and A. Cano, Journal of Physics: Materials , 03LT01 (2020).[16] A. S. Botana and M. R. Norman, Phys. Rev. X ,011024 (2020).[17] Y. Zhang, L.-F. Lin, W. Hu, A. Moreo, S. Dong, andE. Dagotto, Phys. Rev. B , 195117 (2020).[18] V. M. Katukuri, N. A. Bogdanov, O. Weser, J. van denBrink, and A. Alavi, Phys. Rev. B , 241112 (2020).[19] G. Sawatzky, Nature , 592 (2019).[20] K.-W. Lee and W. E. Pickett, Phys. Rev. B , 165109(2004).[21] K. Matsumoto, M. Haruta, M. Kawai, A. Sakaiguchi,N. Ichikawa, H. Kurata, and Y. Shimakawa, ScientificReports , 27 (2011).[22] B. Geisler and R. Pentcheva, Phys. Rev. B , 020502(2020).[23] M. Wu, E. Benckiser, M. W. Haverkort, A. Frano,Y. Lu, U. Nwankwo, S. Br ˜A ¼ ck, P. Audehm, E. Go-ering, S. Macke, V. Hinkov, P. Wochner, G. Chris-tiani, S. Heinze, G. Logvenov, H. U. Habermeier, andB. Keimer, Phys. Rev. B , 1 (2013).[24] D. Di Castro, M. Salvato, A. Tebano, D. Innocenti, C. Aruta, W. Prellier, O. I. Lebedev, I. Ottaviani, N. B.Brookes, M. Minola, M. Moretti Sala, C. Mazzoli, P. G.Medaglia, G. Ghiringhelli, L. Braicovich, M. Cirillo, andG. Balestrino, Phys. Rev. B , 134524 (2012).[25] S. Macke, J. E. Hamann-Borrero, R. J. Green, B. Keimer,G. A. Sawatzky, and M. W. Haverkort, Phys. Rev. Lett. , 115501 (2016).[26] R. J. Green, M. W. Haverkort, and G. A. Sawatzky,Phys. Rev. B , 195127 (2016).[27] M. Hepting, D. Li, C. J. Jia, H. Lu, E. Paris, Y. Tseng,X. Feng, M. Osada, E. Been, Y. Hikita, Y.-D. Chuang,Z. Hussain, K. J. Zhou, A. Nag, M. Garcia-Fernandez,M. Rossi, H. Y. Huang, D. J. Huang, Z. X. Shen,T. Schmitt, H. Y. Hwang, B. Moritz, J. Zaanen, T. P.Devereaux, and W. S. Lee, Nature Materials , 381(2020).[28] M. Jiang, M. Berciu, and G. A. Sawatzky, Phys. Rev.Lett. , 207004 (2020).[29] C. T. Chen, L. H. Tjeng, J. Kwo, H. L. Kao, P. Rudolf,F. Sette, and R. M. Fleming, Phys. Rev. Lett. , 2543(1992).[30] G. Kresse and J. Hafner, Phys. Rev. B , 558 (1993).[31] J. P. Perdew, K. Burke, and M. Ernzerhof, Phys. Rev.Lett. , 3865 (1996).[32] A. A. Mostofi, J. R. Yates, G. Pizzi, Y.-S. Lee, I. Souza,D. Vanderbilt, and N. Marzari, Computer Physics Com-munications , 2309 (2014).[33] O. Parcollet, M. Ferrero, T. Ayral, H. Hafermann,I. Krivenko, L. Messio, and P. Seth, Computer PhysicsCommunications , 398 (2015).[34] P. Seth, I. Krivenko, M. Ferrero, and O. Parcollet, Com-puter Physics Communications200