Asymmetric orbital-lattice interactions in ultra-thin correlated oxide films J. Chakhalian, 1, ∗ J.M. Rondinelli, 2, 3 Jian Liu, 1, 4 B.A.Gray, 1 M. Kareev, 1 E.J.Moon, 1 N. Prasai, 5 J.L. Cohn, 5 M.Varela, 6 I.C. Tung, 7 M.J. Bedzyk, 7 S.G. Altendorf, 8 F. Strigari, 8 B. Dabrowski, 9 L.H. Tjeng, 10 P.J. Ryan, 3 and J.W. Freeland 3 1 Department of Physics, University of Arkansas, Fayetteville, Arkansas 70701, USA 2 Department of Materials Science & Engineering, Drexel University, Philadelphia, Pennsylvania 19104, USA 3 Advanced Photon Source, Argonne National Laboratory, Argonne, Illinois 60439, USA 4 Advanced Light Source, Lawrence Berkeley National Laboratory, Berkeley, California 94720, USA 5 Department of Physics, University of Miami, Coral Gables, Florida, 33124, USA 6 Oak Ridge National Laboratory, Oak Ridge, Tennessee 37831, USA 7 Materials Science and Engineering, Northwestern University, Evanston, Illinois, 60208, USA 8 II. Physikalisches Institut, Universit¨at zu K¨oln, Z¨ ulpicher Str. 77, 50937 K¨oln, Germany 9 Department of Physics, Northern Illinois University, Dekalb, Illinois 60115, USA 10 Max Planck Institute for Chemical Physics of Solids, N¨othnitzerstr. 40, 01187 Dresden, Germany (Dated: November 1, 2018) Using resonant X-ray spectroscopies combined with density functional calculations, we find an asymmetric bi-axial strain-induced d-orbital response in ultra-thin films of the correlated metal LaNiO3 which are not accessible in the bulk. The sign of the misfit strain governs the stability of an octahedral “breathing” distortion, which, in turn, produces an emergent charge-ordered ground state with an altered ligand-hole density and bond covalency. Control of this new mechanism opens a pathway to rational orbital engineering, providing a platform for artificially designed Mott materials. PACS numbers: 73.20.-r, 73.50.-h, 68.55.-a, 71.20.Be Heteroepitaxial synthesis is a powerful avenue to mod- ify orbital–lattice interactions in correlated materials with strong electron–electron interactions derived from transi- tion metals with open d-shell configurations [1–3]. Epitax- ial strain allows access to latent electronic functionalities and phases that do not exist in bulk equilibrium phase diagrams [4–8]. However, efforts to rationally control properties that are exceedingly sensitive to small per- turbations [9] through the orbital–lattice interaction are impeded by the poor understanding of how heteroepi- taxy imposes constraints on the orbital response [10–12]. Despite the recent progress in strain-induced orbital engi- neering, a crucial fundamental question remains: when a single electron occupies a doubly degenerate d-orbital in a cubic crystal field as in perovskites with Cu 2+ , Mn 3+ or low spin Ni 3+ cations, how does the substrate imposed epi- taxial constraints dictate the correlated orbital responses of ultrathin films? The exceptional strain control of the frontier atomic or- bitals relies on the susceptibility of the orbital occupations and their energy level splittings to biaxial strain-induced lattice deformations, i.e. orbital–lattice interactions. The conventional orbital engineering approach in perovskite- structured oxides is often rationalized as follows: coherent heteroepitaxy imposes a tetragonal distortion on the film’s primitive unit cell, which then modifies the chemical bond lengths of the functional octahedral building blocks. The bond distortions in turn alter the crystal field symme- try and remove the cubic twofold e g electron degeneracy. Since each e g orbital state is of the same symmetry, it is generally anticipated that both tensile and compressive strains should symmetrically alter the d x 2 −y 2 and d 3z 2 −r 2 orbital states. They are either lowered or raised relative to the strain-free band center of mass, which for finite filling, leads to an orbital polarization. This symmetric strain- induced orbital polarization (SIOP) concept is routinely used to rationalize the orbital responses of many complex oxide systems [12, 13] and even theoretically suggested to be a possible route to replicate high-T c cuprate physics in nickelates [14]. However, the strategies for tuning orbital ground states in ultrathin films devised from the suppos- edly symmetric strain response of the orbital occupations alone are violated more often than they apply: Doped manganite thin films exhibit d 3z 2 −r 2 orbital polarization regardless of the bi-axial strain sign [15, 16]. Ultrathin cuprate bilayers also show variable critical temperatures [17] correlated with the sign of the interface lattice misfit [18], and cobaltite films are either ferro- or diamagnetic depending on the strain state [19]. In this Letter, we provide insight into why many ultra- thin correlated oxides violate the SIOP model by investi- gating the orbital–lattice interactions in 10 unit cell thick films of the correlated and orbitally degenerate (t 6 2g e 1 g ) metal LaNiO 3 (LNO). This representative spin- 1 2 system has no magnetic or structural transitions below 500 K, which otherwise might obscure the investigation [20–22]. We identify the microscopic mechanism responsible for the asymmetric strain-induced orbital response and show that it originates from latent instabilities of the bulk ma- terial, which are strongly enhanced due to the epitaxial constraints imposed by the heterointerface. We further demonstrate deterministic control of the ligand-hole den- sity and covalency through the asymmetric orbital–lattice response, enabling access to an emergent charge-ordered phase not attainable in the bulk. For this purpose, we synthesized high quality epitax- arXiv:1008.1373v2 [cond-mat.mtrl-sci] 25 Aug 2011