Abstract
The theoretical understanding of emergent phenomena in quantum materials is one of the greatest challenges in condensed matter physics. In contrast to simple materials such as noble metals and semiconductors, macroscopic properties of quantum materials cannot be predicted by the properties of individual electrons. One of the examples of scientific importance is strongly correlated electron system. Neither localized nor itinerant behaviors of electrons in partially filled 3d, 4f, and 5f orbitals give rise to rich physics such as Mott insulators, high-temperature superconductors, and superior thermoelectricity, but hinder quantitative understanding of low-lying excitation spectrum. Here we present a new first-principles approach to strongly correlated solids. It is based on a combination of the quasiparticle self-consistent GW approximation and the dynamical mean-field theory. The sole input in this method is the projector to the set of correlated orbitals for which all local Feynman graphs are being evaluated. For that purpose, we choose very localized quasiatomic orbitals spanning large energy window, which contains most strongly hybridized bands, as well as upper and lower Hubbard bands. The self-consistency is carried out on the Matsubara axis. This method enables the first-principles study of Mott insulators in both their paramagnetic and antiferromagnetic phases. We illustrate the method on the archetypical charge transfer correlated insulators La2CuO4 and NiO, and obtain spectral properties and magnetic moments in good agreement with experiments.
Similar content being viewed by others
Introduction
The first-principles description of strongly correlated materials is currently regarded as one of the greatest challenges in condensed matter physics. The interplay between correlated electrons in open d- or f- shell and itinerant band states gives rise to rich physics that makes these materials attractive for a wide range of applications such as oxide electronics, high-temperature superconductors, and spintronic devices. Various theoretical approaches are currently being pursued.1 One of the most successful approaches is the dynamical mean-field theory (DMFT).2 In combination with density functional theory,3,4 it has described many features of strongly correlated materials successfully and highlighted the surprising accuracy of treating correlations local to a small subset of orbitals exactly, while treating the remainder of the problem in a static mean-field manner.5
The numerous successes of DMFT in different classes of correlated materials revived the interest in the long-sought goal of achieving a diagrammatically controlled approach to the quantum many-body problem of solids, starting from the Green’s function G and the screened Coulomb interactions W.6,7 The lowest order diagram in perturbation theory in this functional gives rise to the GW approximation,8 whereas the local approximation applied to the most correlated orbitals gives rise to an extended DMFT approach to the electronic structure problem.7 The addition of the GW and DMFT graphs was proposed and implemented in model Hamiltonian studies9 and in realistic electronic structure.10,11 There is now intense activity in this area with many recent publications12–15 triggered by advances in the quality of the impurity solvers,16,17 insights into the analytic form of the high-frequency behavior of the self-energy,18 and improved electronic structure codes.
Several conceptual issues remain to be clarified before the long-sought goal of a robust electronic structure method for solids is attained. The first issue is the choice of local orbitals on which to perform the DMFT method (summation of all local Feynman graphs). The second issue is the level of self-consistency that should be used in the calculation of various parts of the diagrams included in the treatment (free or bare Green’s function G0 versus self-consistent interacting Green’s functions G). These central issues are addressed in this letter.
The self-consistency issue appears already at the lowest order, namely the GW level and it has been debated over time. The corresponding issue in GW+DMFT is expected to be at least as important, but has not been explored, except for model Hamiltonians.19,20 At the GW level, it is now well established that Hedin’s fully self-consistent formulation,8 while producing good total energies in solids,21 atoms, and molecules,22,23 does not produce a good approximation to the spectra of even three-dimensional electron gas and aluminum in comparison with non-self-consistent GW results.21,24 Instead, using a free (quasiparticle (QP)) Green’s function in the evaluation of the polarization graph of the GW method gives much better results for spectral functions. This is the basis of the one-shot QP GW, starting from local density approximation (LDA)25 or from others.26 Unfortunately, the answer depends on the starting point. A solution for this problem is to impose a self-consistency equation to determine G0. This method, called the QP self-consistent GW (QSGW),27 is very successful reproducing the spectra of many systems.27 How to combine it with DMFT is an important open challenge.28,29
Previous GW+DMFT studies typically used a G0, which depends on the LDA starting point, and projectors spanning a relatively small energy window.12–15 In this work, we propose a different approach to the level of self-consistency and the choice of the DMFT orbital. We do a self-consistent QSGW calculation and then calculate local self-energy using DMFT with static Ud and JH without feedback to non-local self-energy within GW. For the DMFT step, we choose a very localized orbital spanning large energy window, which contains most strongly hybridized bands, as well as upper and lower Hubbard bands.
In the LDA+DMFT context, the choice of very localized orbitals has provided a great deal of universality, as the interactions do not vary much among compounds of the same family. This has been demonstrated in the studies of iron pnictides30 and transition metal oxides.31 This choice results in a second advantage as we will show below, namely the frequency dependence of the interaction matrix can be safely ignored. Having chosen the correlated orbitals, all the other parameters are self-consistently determined. This is the first ab initio QP self-consistent GW+DMFT implementation and the first study on a paramagnetic (PM) Mott insulator within the GW+DMFT method.
Results
Figure 2a shows the frequency dependence of real and imaginary parts of Ud of La2CuO4 shown in Figure 1. It is calculated on an imaginary frequency axis and analytically continued by a maximum entropy method.32 We also plot the fully screened Coulomb interaction Wd for comparison. Static Ud is 12.0 eV and Ud remains almost constant up to 10 eV. In contrast, in Wd, there are several peaks due to low-energy collective excitations below 10 eV. At very high energy, Ud approaches the bare coulomb interaction of 28 eV. Static value of Upd is 2.0 eV, much smaller than Ud; hence, we do not discuss its treatment further (The renormalization of Ud by the nonlocal Coulomb interaction can be understood by using the definition of Hubbard U by C Herring. According to Herring,33 the Hubbard U can be obtained from the reaction of between Cu-d and O-p orbitals, where nd and np are the number of electrons in Cu-d and O-p orbitals, respectively. The energy cost for the reaction is approximately Ud−2Upd+Up and this should be the sum of the renormalized Ud and renormalized Up. In the simplest case, where we have one electron in p and d orbitals, renormalized U of Cu-d orbitals is Ud−Upd from the reaction of d1p1→d2p0. The renormalized Ud is derived more rigorously by Schüler et al.34). Calculated JH is 1.4 eV and has negligible frequency dependence. By contrast, conventional constrained random phase approximation, in which 10 bands of mostly Cu-3d character are excluded from screening, results in static Ud=7.6 eV, which is too small to open the Mott gap, and which is also inconsistent with photoemission experiments on CuO charge transfer insulators.35
We introduced a complementary method to compute the static Ud. The key idea is to first calculate the excitation spectra of La2CuO4 within Matsubara QSGW (MQSGW)+DMFT using local GW (with a static Ud) as the impurity solver and then determine Ud, by finding the value that best matches the full spin-polarized MQSGW spectra. The procedure starts from the non-spin-polarized MQSGW band structure without magnetic long-range order. We then allow spontaneous magnetic long-range order by embedding a polarized impurity self-energy for the Cu-3d electrons computed in a local GW approximation. We find that indeed magnetic ordering associated with Cu-3d is captured by spin-polarized local MQSGW using a static value of Ud and JH, and spectral properties such as energy gap are very similar in value to the full spin-polarized MQSGW spectra. In Figure 2b, we allowed Ud to vary between 8 and 13 eV (at fixed JH=1.4 eV) and we plot the size of the indirect gap. The gap size of this method matches the gap of spin-polarized MQSGW when Ud≈12 eV. If this choice of Ud and JH is correct, the resulting spectra must be similar to the prediction of spin-polarized MQSGW method. We show this comparison in Figure 2c to confirm a good match. In addition, the relative position of the Cu-d band (the lowest energy conduction band at S) to the La-d band (the lowest energy conduction band at Y) is also well matched, justifying the approximation of . ΣDC(iωn=0) for Cu- orbital differs from nominal double counting energy36 by only 1%, highlighting again the advantages of using a broad window and narrow orbitals.
We now discuss the magnetic moment associated with Cu and the electronic excitation spectra of La2CuO4 by using MQSGW+DMFT (with Ud=12.0 eV, JH=1.4 eV) in which the impurity is solved by the numerically exact continuous-time quantum Monte-Carlo16,17 and compare them with other methods. Local spin density approximation does not have a magnetic solution. In contrast, spin-polarized MQSGW, QSGW,27 and MQSGW+DMFT predict 0.7 μB, 0.7 μB, and 0.8 μB, respectively. This is consistent with experimental measurements, although the later span quite large range 0.4~0.8 μB.37–39
In the low-energy spectrum of La2CuO4, local spin density approximation does not have an insulating solution; there is a single non-magnetic solution with zero energy gap as shown in the bandstructure(Figure 3a) and total density of states (Figure 4a). The non-spin-polarized MQSGW also predicts metal as shown in Figure 4a, but the two bands of primarily Cu- character near the Fermi level are well-separated from the rest of the bands (dashed lines in Figure 3b). Spin-polarized MQSGW calculation (dashed lines in Figure 3c) yields qualitative different results from local spin density approximation and non-spin-polarized MQSGW calculation. The two Cu- bands are now well separated from each other with a bandgap of 3.4 eV. Spin-polarized QSGW27 also yields insulating phase with a gap of 4.0 eV. In the experiment, the larger direct gap, as measured by optics, is ~2 eV.40,41
We show that these deficiencies of LDA, QSGW and MQSGW in the low-energy spectra can be remedied by adding all local Feynman diagrams for the Cu-d orbitals using the DMFT. The LDA+DMFT calculation in Figure 4a, carried out by the all-electron LDA+DMFT method,31,36 predicts reasonable gap of 1.5 and 1.8 eV in PM and antiferromagnetic (AFM) phases, in good agreement with experiment and previous LDA+DMFT studies.31,42–45 Within MQSGW+DMFT, we find gaps of 1.5 and 1.6 eV in PM and AFM phases, respectively, as shown in Figure 4b. The excitation spectra of MQSGW+DMFT in PM and AFM phase as shown in Figure 3b,c are very similar, as both are insulating with well-separated Cu- bands, which is now also substantially broadened due to large scattering rate in Hubbard-like bands. In addition, MQSGW+DMFT improves the line-shape of LDA+DMFT. Near the top of the valence bands with oxygen p character, the lineshape within LDA+DMFT is too sharp in comparison with the experiments as shown in Figure 4c. By treating oxygen p levels within GW, the lineshape becomes smoother and in a better agreement with experiments.
In the high-energy region of La2CuO4, the most distinctive difference is the position of La-f peak. It appears at ~3 eV within LDA and LDA+DMFT, but at around ~9 eV in the inverse-photoemission spectra (cyan dotted line in Figure 4a).46 By treating La-f within GW approximation, it appears at ~10 eV within MQSGW and MQSGW+DMFT. The underestimation of unoccupied La-f excitation energy is attributed to the local approximation to the electron self-energy within LDA. Within LDA, Hartree and exchange-correlation potential applied to La-f orbitals are orbital independent, as charge density is averaged over 14 different m channels.47 In contrast, these potentials within MQSGW are orbital dependent and non-local. The effect of orbital-dependent potential can be tested within LDA+U approaches, as LDA+U adds orbital-dependent potential and subtracts orbital-independent potential explicitly3. From LDA+U approaches, we can also understand MQSGW better, as LDA+U can be regarded as a local and static approximation to GW approximation3. According to Czyzyk and Sawatzky48, La-f peaks shift from EF+3 eV to EF+3 eV+U/2 with U=11 eV for La-f.
We also tested our proposed scheme with one more charge transfer insulator, NiO. Figure 5a shows the frequency dependence of Ud and Wd for the Ni-3d orbitals in the low-energy region. In contrast to Wd, Ud is almost constant up to 5 eV. Static Ud is 9.6 eV. In the high-energy limit, Ud and Wd approach the bare value of 26.0 eV. Calculated JH for the Ni-3d orbitals has negligible frequency dependence and static JH is 1.4 eV. Figure 5b shows the total density of states of NiO within LDA+DMFT and MQSGW+DMFT in its PM phase. Photoemission/inverse photoemission data are also plotted for comparison.49 The LDA+DMFT calculation is being carried out by the all-electron LDA+DMFT method36 with Ud=10 eV, JH=0.9 eV and nominal double-counting energy. In the PM phase, LDA+DMFT and MQSGW+DMFT predict insulator in an agreement with previous LDA+DMFT studies,50,51 but MQSGW+DMFT improves the lineshape of LDA+DMFT. Near the top of the valence bands, the lineshape within LDA+DMFT is too sharp in comparison with the experiments. By treating oxygen p levels within GW, the lineshape becomes smoother and in a better agreement with experiments. In the AFM phase, magnetic moment associated with Ni-d orbitals is 1.6 μB within MQSGW+DMFT, in agreement with experimental value of 1.6–1.9 μB.50,52,53
In summary, we introduced a new methodology within MQSGW+DMFT and tested it in the classic charge transfer insulator La2CuO4 and NiO. Our methodology predicts a Mott-insulating gap in the PM phase, thus overcoming the limitation of LDA and QSGW. It yields more precise peak positions of the La-f states in La2CuO4 and valence band lineshape, thus improving the results of LDA+DMFT. The method should be useful in understanding electronic excitation spectrum of other strongly correlated materials, in particular, those where precise position of both the itinerant and correlated states is important.
Methods
Our approach is carried out entirely on the Matsubara axis, which requires a different approach to the QP self-consistency in GW,54 called MQSGW, where the QP Hamiltonian is constructed by linearizing the self-energy and renormalization factor (MQSGW is a form of QP self-consistency that replaces with on the Matsubara axis. It is similar to, but not identical with the form in reference 55 on the real frequency axis, which replaces with , derived from a norm minimization principle. Here, and are QP energy and wavefunction, respectively.). Working on the Matsubara axis is numerically very stable, provide a natural interface with advanced DMFT solvers such as continuous-time quantum Monte-Carlo,16,17 and has very good scaling in system size as in the space-time method (see Supplementary Note on MQSGW calculations).
For DMFT, it is essential to obtain bandstructures in a fine enough crystal momentum (k) mesh to attain desired frequency resolution of physical quantities. To achieve such momentum resolution, we use a Wannier-interpolated MQSGW bandstructure in a large energy window using maximally localized Wannier function56 and then constructed local projector in a fine momentum mesh. In contrast to SrVO3,12–15 where a set of t2g states is reasonably well separated from the other bands, correlated 3d orbitals in La2CuO4 and NiO, and are strongly hybridized with other itinerant bands. In this case, it is necessary to construct local projectors from states in a wide-enough energy windows to make projectors localized near the correlated atoms. We constructed local projectors in the energy window EF±10 eV in which there are ~82 bands at each k point, where EF is the Fermi level for La2CuO4. For NiO, we constructed local projectors in the energy window of EF−11 eV to EF+10 eV. Next, we confirmed that absolute value of its overlap to the muffin-tin orbital (of which radial function is determined to maximize electron occupation in it) is >95%. Our choice of energy window is justified by the Cu-3d spectra being entirely contained in this window. Using constructed maximally localized Wannier function s, we defined our local projector , where WR i(r) is the maximally localized Wannier function with an index i, is the QP wavefunction with an index n, and Nk is the number of k points in the first Brillouin zone.
Static Ud and JH are evaluated by a modification of the constrained random phase approximation method,57 which avoids screening by the strongly hybridized bands. This screening by hybridization is included in our large-energy window DMFT. For details, see Supplementary Note on Ud and JH. We divide dynamic polarizability within MQSGW approximation χQP into two parts, . Here, is defined by all transitions between the states in the energy window accounted for by the DMFT method (EF±10 eV for La2CuO4 and EF−11 eV to EF+10 eV for NiO). Using , we evaluate partially screened Coulomb interaction and parametrize static Ud and JH by Slater’s integrals,58,59 where V is bare Coulomb interaction.
The Feynman graphs included in both MQSGW and DMFT (double counting) are the local Hartree and the local GW diagram. They are computed using the local projection of the MQSGW Green’s function and the local Coulomb matrix constructed from Slater’s integrals. For the details, see Supplementary Note on double counting energy.
References
Anisimov, V. I . Strong Coulomb Correlations in Electronic Structure Calculations (CRC Press, 2000).
Georges, A., Kotliar, G., Krauth, W. & Rozenberg, M. J. Dynamical mean-field theory of strongly correlated fermion systems and the limit of infinite dimensions. Rev. Mod. Phys. 68, 13–125 (1996).
Anisimov, V. I., Poteryaev, A. I., Korotin, M. A., Anokhin, A. O. & Kotliar, G. First-principles calculations of the electronic structure and spectra of strongly correlated systems: dynamical mean-field theory. J. Phys. Condens. Matter 9, 7359 (1997).
Lichtenstein, A. I. & Katsnelson, M. I. Ab initio calculations of quasiparticle band structure in correlated systems: LDA++ approach. Phys. Rev. B 57, 6884–6895 (1998).
Kotliar, G. et al. Electronic structure calculations with dynamical mean-field theory. Rev. Mod. Phys. 78, 865–951 (2006).
Almbladh, C.-O., Barth, U. V. & Leeuwen, R. V. Variational total energies from-and-derivable theories. Int. J. Mod. Phys. B 13, 535–541 (1999).
Chitra, R. & Kotliar, G. Effective-action approach to strongly correlated fermion systems. Phys. Rev. B 63, 115110 (2001).
Hedin, L. New method for calculating the one-particle Green’s function with application to the electron-gas problem. Phys. Rev. 139, A796–A823 (1965).
Sun, P. & Kotliar, G. Extended dynamical mean-field theory and GW method. Phys. Rev. B 66, 085120 (2002).
Biermann, S., Aryasetiawan, F. & Georges, A. First-principles approach to the electronic structure of strongly correlated systems: combining the GW approximation and dynamical mean-field theory. Phys. Rev. Lett. 90, 086402 (2003).
Kotliar, G. & Savrasov, S. Y. in New Theoretical Approaches to Strongly Correlated Systems, No. 23 in NATO Science Series (ed. Tsvelik, A. M. ) 259–301 (Springer, 2001).
Tomczak, J. M., Casula, M., Miyake, T., Aryasetiawan, F. & Biermann, S. Combined GW and dynamical mean-field theory: dynamical screening effects in transition metal oxides. EPL 100, 67001 (2012).
Sakuma, R., Werner, P. & Aryasetiawan, F. Electronic structure of SrVO3 within GW+DMFT. Phys. Rev. B 88, 235110 (2013).
Taranto, C. et al. Comparing quasiparticle GW+DMFT and LDA+DMFT for the test bed material SrVO3 . Phys. Rev. B 88, 165119 (2013).
Tomczak, J. M., Casula, M., Miyake, T. & Biermann, S. Asymmetry in band widening and quasiparticle lifetimes in SrVO3: competition between screened exchange and local correlations from combined gw and dynamical mean-field theory gw+dmft. Phys. Rev. B 90, 165138 (2014).
Werner, P., Comanac, A., de Medici, L., Troyer, M. & Millis, A. J. Continuous-time solver for quantum impurity models. Phys. Rev. Lett. 97, 076405 (2006).
Haule, K. Quantum Monte Carlo impurity solver for cluster dynamical mean-field theory and electronic structure calculations with adjustable cluster base. Phys. Rev. B 75, 155113 (2007).
Casula, M., Rubtsov, A. & Biermann, S. Dynamical screening effects in correlated materials: plasmon satellites and spectral weight transfers from a Green’s function ansatz to extended dynamical mean field theory. Phys. Rev. B 85, 035115 (2012).
Sun, P. & Kotliar, G. Many-body approximation scheme beyond GW. Phys. Rev. Lett. 92, 196402 (2004).
Hansmann, P., Ayral, T., Vaugier, L., Werner, P. & Biermann, S. Long-range coulomb interactions in surface systems: a first-principles description within self-consistently combined GW and dynamical mean-field theory. Phys. Rev. Lett. 110, 166401 (2013).
Kutepov, A., Savrasov, S. Y. & Kotliar, G. Ground-state properties of simple elements from GW calculations. Phys. Rev. B 80, 041103 (2009).
Stan, A., Dahlen, N. E. & Leeuwen, R. V. Fully self-consistent GW calculations for atoms and molecules. EPL 76, 298 (2006).
Stan, A., Dahlen, N. E. & Leeuwen, R. V. Levels of self-consistency in the GW approximation. J. Chem. Phys. 130, 114105 (2009).
Holm, B. & von Barth, U. Fully self-consistent gw self-energy of the electron gas. Phys. Rev. B 57, 2108–2117 (1998).
Hybertsen, M. S. & Louie, S. G. Electron correlation in semiconductors and insulators: band gaps and quasiparticle energies. Phys. Rev. B 34, 5390–5413 (1986).
Rinke, P., Qteish, A., Neugebauer, J., Freysoldt, C. & Scheffler, M. Combining GW calculations with exact-exchange density-functional theory: an analysis of valence-band photoemission for compound semiconductors. New J. Phys. 7, 126 (2005).
Kotani, T., van Schilfgaarde, M. & Faleev, S. V. Quasiparticle self-consistent GW method: a basis for the independent-particle approximation. Phys. Rev. B 76, 165106 (2007).
Tomczak, J. M., van Schilfgaarde, M. & Kotliar, G. Many-body effects in iron pnictides and chalcogenides: nonlocal versus dynamic origin of effective masses. Phys. Rev. Lett. 109, 237010 (2012).
Tomczak, J. M. QSGW+DMFT: an electronic structure scheme for iron pnictides and beyond. J. Phys.: Conf. Ser. 592, 012055 (2015).
Yin, Z. P., Haule, K. & Kotliar, G. Kinetic frustration and the nature of the magnetic and paramagnetic states in iron pnictides and iron chalcogenides. Nat. Mater. 10, 932–935 (2011).
Haule, K., Birol, T. & Kotliar, G. Covalency in transition-metal oxides within all-electron dynamical mean-field theory. Phys. Rev. B 90, 075136 (2014).
Jarrell, M. & Gubernatis, J. E. Bayesian inference and the analytic continuation of imaginary-time quantum Monte Carlo data. Phys. Rep. 269, 133–195 (1996).
Rado, G. & Suhl, H . Magnetism: Exchange Interactions Among Itinerant Electrons, by C. Herring, Magnetism (Academic Press, 1966).
Schüler, M., Rösner, M., Wehling, T. O., Lichtenstein, A. I. & Katsnelson, M. I. Optimal Hubbard models for materials with nonlocal coulomb interactions: graphene, silicene, and benzene. Phys. Rev. Lett. 111, 036601 (2013).
Ghijsen, J., Tjeng, L. H., Eskes, H., Sawatzky, G. A. & Johnson, R. L. Resonant photoemission study of the electronic structure of CuO and Cu2O. Phys. Rev. B 42, 2268–2274 (1990).
Haule, K., Yee, C.-H. & Kim, K. Dynamical mean-field theory within the full-potential methods: electronic structure of CeIrIn5, CeCoIn5, and CeRhIn5 . Phys. Rev. B 81, 195107 (2010).
Borsa, F. et al. Staggered magnetization in La2−xSrxCuO4 from 139La nqr and μsr: effects of sr doping in the range 0<x<0.02. Phys. Rev. B 52, 7334–7345 (1995).
Reehuis, M. et al. Crystal structure and high-field magnetism of La2CuO4 . Phys. Rev. B 73, 144513 (2006).
Vaknin, D. et al. Antiferromagnetism in La2CuO4−y . Phys. Rev. Lett. 58, 2802–2805 (1987).
Ginder, J. M. et al. Photoexcitations in La2CuO4: 2-eV energy gap and long-lived defect states. Phys. Rev. B 37, 7506–7509 (1988).
Cooper, S. L. et al. Optical studies of gap, exchange, and hopping energies in the insulating cuprates. Phys. Rev. B 42, 10785–10788 (1990).
Weber, C., Haule, K. & Kotliar, G. Optical weights and waterfalls in doped charge-transfer insulators: a local density approximation and dynamical mean-field theory study of La2−xSrxCuO4 . Phys. Rev. B 78, 134519 (2008).
Weber, C., Haule, K. & Kotliar, G. Strength of correlations in electron- and hole-doped cuprates. Nat. Phys. 6, 574–578 (2010).
Wang, X. et al. Covalency, double-counting, and the metal-insulator phase diagram in transition metal oxides. Phys. Rev. B 86, 195136 (2012).
Werner, P., Sakuma, R., Nilsson, F. & Aryasetiawan, F. Dynamical screening in La2CuO4 . Phys. Rev. B 91, 125142 (2015).
Nücker, N. et al. Experimental electronic structure studies of La2xSrxCuO4 . Z. Phys. B Condens. Matter 67, 9–14 (1987).
Anisimov, V. I., Zaanen, J. & Andersen, O. K. Band theory and mott insulators: Hubbard U instead of stoner I. Phys. Rev. B 44, 943–954 (1991).
Czyzyk, M. T. & Sawatzky, G. A. Local-density functional and on-site correlations: the electronic structure of La2CuO4 and LaCuO3 . Phys. Rev. B 49, 14211–14228 (1994).
Zimmermann, R. et al. Electronic structure of 3d-transition-metal oxides: on-site Coulomb repulsion versus covalency. J. Phys. Condens. Matter 11, 1657 (1999).
Ren, X. et al. LDA+DMFT computation of the electronic spectrum of NiO. Phys. Rev. B 74, 195114 (2006).
Yin, Q., Gordienko, A., Wan, X. & Savrasov, S. Y. Calculated momentum dependence of Zhang-rice states in transition metal oxides. Phys. Rev. Lett. 100, 066406 (2008).
Cheetham, A. K. & Hope, D. A. O. Magnetic ordering and exchange effects in the antiferromagnetic solid solutions MnxNi1−xO. Phys. Rev. B 27, 6964–6967 (1983).
Fender, B. E. F., Jacobson, A. J. & Wedgwood, F. A. Covalency parameters in MnO, MnS, and NiO. J. Chem. Phys. 48, 990–994 (1968).
Kutepov, A., Haule, K., Savrasov, S. Y. & Kotliar, G. Electronic structure of Pu and Am metals by self-consistent relativistic GW method. Phys. Rev. B 85, 155129 (2012).
vanSchilfgaarde, M., Kotani, T. & Faleev, S. Quasiparticle self-consistent GW theory. Phys. Rev. Lett. 96, 226402 (2006).
Marzari, N., Mostofi, A. A., Yates, J. R., Souza, I. & Vanderbilt, D. Maximally localized Wannier functions: theory and applications. Rev. Mod. Phys. 84, 1419–1475 (2012).
Aryasetiawan, F. et al. Frequency-dependent local interactions and low-energy effective models from electronic structure calculations. Phys. Rev. B 70, 195104 (2004).
van der Marel, D. & Sawatzky, G. A. Electron-electron interaction and localization in d and f transition metals. Phys. Rev. B 37, 10674–10684 (1988).
Kutepov, A., Haule, K., Savrasov, S. Y. & Kotliar, G. Self-consistent GW determination of the interaction strength: application to the iron arsenide superconductors. Phys. Rev. B 82, 045105 (2010).
Acknowledgements
This work was supported by the U.S Department of energy, Office of Science, Basic Energy Sicences as a part of the Computational Materials Science Program and by the Simons Foundation under the Many Electron Problem collaboration. An award of computer time was provided by the INCITE program. This research used resources of the Oak Ridge Leadership Computing Facility, which is a DOE Office of Science User Facility supported under Contract DE-AC05-00OR22725.
Author information
Authors and Affiliations
Contributions
G.K. designed the framework of the code. S.C. developed the code, building on earlier developments of A.K and K. H. and performed the calculations. G.K., K.H. and S.C. analyzed the data with the help of A.K. and M.v.S. All authors provided comments on the paper.
Corresponding author
Ethics declarations
Competing interests
The authors declare no conflict of interest.
Additional information
Supplemental Information accompanies the paper on the npj Quantum Materials website (http://www.nature.com/npjquantmats)
Supplementary information
Rights and permissions
This work is licensed under a Creative Commons Attribution 4.0 International License. The images or other third party material in this article are included in the article’s Creative Commons license, unless indicated otherwise in the credit line; if the material is not included under the Creative Commons license, users will need to obtain permission from the license holder to reproduce the material. To view a copy of this license, visit http://creativecommons.org/licenses/by/4.0/
About this article
Cite this article
Choi, S., Kutepov, A., Haule, K. et al. First-principles treatment of Mott insulators: linearized QSGW+DMFT approach. npj Quant Mater 1, 16001 (2016). https://doi.org/10.1038/npjquantmats.2016.1
Received:
Revised:
Accepted:
Published:
DOI: https://doi.org/10.1038/npjquantmats.2016.1
This article is cited by
-
Infinite-layer nickelates as Ni-eg Hund’s metals
npj Quantum Materials (2023)
-
Breakdown of the scaling relation of anomalous Hall effect in Kondo lattice ferromagnet USbTe
Nature Communications (2023)
-
Quantum embedding theories to simulate condensed systems on quantum computers
Nature Computational Science (2022)
-
Orbital selective Kondo effect in heavy fermion superconductor UTe2
npj Quantum Materials (2022)
-
Unusual electronic and vibrational properties in the colossal thermopower material FeSb2
Scientific Reports (2018)