Mössbauer parameters from DFTbased WIEN2k calculations for extended systems Peter Blaha Institute of Materials Chemistry TU Vienna Main Mössbauer parameters: The main (conventional) Mössbauer spectroscopy parameters which we want to calculate by theory are: Isomer shift: d = a (r0Sample – r0Reference); Magnetic Hyperfine fields: Btot = Bcontact + Borb + Bdip it is proportional to the electron density r0 at the nucleus the constant a is proportional to the change of nuclear radii during the transition (we use a=-.291 au3mm s-1) these fields are proportional to the spin-density at the nucleus and the orbital moment of the probed atom as well as the spin moment distribution in the crystal Quadrupole splitting: D e Q Vzz given by the product of the nuclear quadrupole moment Q times the electric field gradient Vzz. The EFG is proportional to an integral over the non-spherical charge density (weighted by 1/r3) Schrödinger equation From the previous slide it is obvious, that we need an accurate knowledge of the electron (and magnetization) density, which in principle can be obtained from the solution of the manybody Schrödinger equation in the corresponding solid. HY=EY However, a many-body problem with ~1023 particles is not solvable at all and we must create models for the real material and rely on an approximate solution of the Schrödinger equation. (This will be briefly discussed in the next slides and my preferred options are marked in red.) Concepts when solving Schrödingers-equation Treatment of spin Non-spinpolarized Spin polarized (with certain magnetic order) Form of potential “Muffin-tin” MT atomic sphere approximation (ASA) pseudopotential (PP) Full potential : FP Relativistic treatment of the electrons exchange and correlation potential non relativistic semi-relativistic fully-relativistic 1 2 k k k V ( r ) i i 2 i Representation of solid non periodic (cluster, individual MOs) periodic (unit cell, Blochfunctions, “bandstructure”) Hartree-Fock (+correlations) Density functional theory (DFT) Local density approximation (LDA) Generalized gradient approximation (GGA) Beyond LDA: e.g. LDA+U Schrödinger - equation Basis functions plane waves : PW augmented plane waves : APW atomic oribtals. e.g. Slater (STO), Gaussians (GTO), LMTO, numerical basis Representation of the solid • cluster model: • approximate the solid by a finite (small) number of atoms. This can be a good approximation for “molecular crystals”. • periodic model: • use a “unit cell”, which is repeated infinitely. This approximates the “real” solid (finite with surfaces, imperfect) by an infinite “ideal” solid. • with “supercells” also surfaces, vacancies or impurities can be modelled. BaSnO3 SnS Exchange and correlation The electron-electron interaction can be approximated by: • Hartree-Fock • exact exchange, but no correlation at all (in solids often very important !) • correlation can approximately be added at various levels: MP2, CC, CI, ... • HF (and even more adding correlation) is very expensive in solids. • Density functional theory: • the exact functional is unknown, thus one must approximate exchange + correlation. Fast method, fairly accurate in solids. • LDA: local density approximation, “free electron gas” • GGA: generalized gradient approximation, various functionals • hybrid-DFT: mixing of HF + GGA, various functionals • LDA+U, DMFT: explicit (heuristic) inclusion of correlations • a brief summary of DFT follows DFT Density Functional Theory Hohenberg-Kohn theorem: (exact) The total energy of an interacting inhomogeneous electron gas in the presence of an external potential Vext(r ) is a functional of the density r E Vext ( r ) r ( r )dr F [ r ] • This (exact) statement states, that we need to know only the electron density r, but we do NOT need to know the wavefunctions (solution of Schrödinger equation not necessary). • However, the exact functional form of F[r] is unknown and any approximation so far is too crude. DFT - Kohn Sham (still exact) Lets decompose the total functional F[r] into parts which can be calculated „exactly“ and some „unknown, but small“ rest: ) 1 r ( r ) r ( r E To [ r ] Vext r ( r )d r dr dr E xc [ r ] 2 | r r | Ekinetic Ene Ecoulomb Eee Exc exchange-correlation non interacting • Ekinetic is the kinetic energy of non-interacting particles • Ecoulomb is the “classical” electrostatic interaction between electrons (including the self-interaction energy) • Exc is the exchange-correlation energy and should correct for the self-interaction and approx. kinetic energy • The exact Exc is again unknown, but now an approximation for the “small correction” may be accurate enough. Kohn-Sham equations 1 r ( r ) r ( r ) E To [ r ] Vext r ( r )d r dr dr E xc [ r ] 2 |r r | vary r 1-electron equations (Kohn Sham) 1 2 { Vext ( r ) VC ( r ( r )) Vxc ( r ( r ))} i ( r ) i i ( r ) 2 E xc ( r ) r (r ) -Z/r 2 d r r ( r ) | | i | r r | r i EF Exc and Vxc are unknown and must be approximated LDA or GGA treat both, exchange and correlation effects approximately Walter Kohn, Nobel Prize 1998 Chemistry Approximations to EXC Local density approximation (LDA): E LDA xc r (r) hom . xc [ r ( r )] dr r (r) 4 3 dr xc is the exchange-correlation energy density of the homogeneous electron gas at density r. Second order gradient expansion (GEA): E GEA xc E LDA xc C ( r ) r (r) 2 3 r ( r ) dr , The GEA XC-hole nXC(r,r’) is not a hole of any physical system and violates nX(r,r’) ≤ 0 ∫ nX(r,r’) dx = -1 ∫ nC(r,r’) dx = 0 exchange hole must be negative must contain charge -1 e- at r nXC(r,r’) Generalized gradient approximations (GGA) GGA E xc r ( r ) F [ r ( r ), r ( r )] dr “construct” GGAs by obeying as many known constraints as possible (Perdew) recover LDA for slowly varying densities obey sum rules and properties of XC-holes long range limits: lim(r -> ∞): exc=-1/2r ; vxc=C-1/r scaling relations Lieb-Oxford bound fitting some parameters to recover “exact” energies of small systems (set of small molecules) or lattice parameters in solids (Becke, Handy, Hammer, ..) Perdew-Burke-Enzerhof – GGA (PRL 1996): well balanced GGA; equally “bad” for “all” systems better approximations are constantly developed meta-GGAs: Perdew,Kurth,Zupan,Blaha (PRL 1999): m GGA Exc r ( r ) F [ r ( r ), r ( r ), 2 r ( r ), ( r )] dr 1 2 i ( r ) 2 i use laplacian of r, or the kinetic energy density ( r ) analytic form for Vxc not possible (Vxc = dExc/dr) , SCF very difficult better meta-GGAs under constant development … more “non-local” functionals (“beyond LDA”) Self-Interaction correction (Perdew,Zunger 1981; Svane+ Temmermann) vanishes for Bloch-states, select “localized states” by hand LDA+U, DMFT (dynamical mean field theory) approximate HF for selected “highly-correlated” electrons (3d,4f,5f) empirical parameter U Exact exchange (similar to HF but DFT based, misses correlation) Hybrid functionals (mixing of LDA (GGA) + HF) DFT ground state of iron LSDA GGA LSDA GGA GGA LSDA NM fcc in contrast to experiment FM bcc Correct lattice constant Experiment FM bcc LDA: Fe is nonmagnetic and in fcc structure GGA correctly predicts Fe to be ferromagnetic and in bcc structure “Everybody in Austria knows about the importance of DFT” “The 75th GGA-version follows the 52nd LDA-version” (thanks to Claudia Ambrosch (TU Leoben)) basis set for the wave functions Even with an approximate e--e- interaction the Schrödinger equation cannot be solved exactly, but we must expand the wave function into a basis set and rely on the variational principle. • “quantum chemistry”: LCAO methods • Gauss functions (large “experience” for many atoms, wrong asymptotic, basis set for heavier atoms very large and problematic, .. ) • Slater orbitals (correct r~0 and r~ asymptotic, expensive) • numerical atomic orbitals • “physics”: plane wave (PW) based methods • plane waves + pseudo-potential (PP) approximation • PP allow fast solutions for total energies, but not for Hyperfine parameters • augmented plane wave methods (APW) • spatial decomposition of space with two different basis sets: • combination of PW (unbiased+flexible in interstitial regions) • + numerical basis functions (accurate in the atomic regions, correct cusp) Computational approximations • relativistic treatment: • non- or scalar-relativistic approximation (neglects spin-orbit, but includes Darvin s-shift and mass-velocity terms) • adding spin-orbit in “second variation” (good enough) • fully-relativistic treatment (Dirac-equation, very expensive) • point- or finite-nucleus • restricted/unrestricted treatment of spin • use correct long range magnetic order (FM, AFM) • approximations to the form of the potential • shape approximations (ASA) • pseudopotential (smooth, nodeless valence orbitals) • “full potential” (no approximation) Concepts when solving Schrödingers-equation • in many cases, the experimental knowledge about a certain system is very limited and also the exact atomic positions may not be known accurately (powder samples, impurities, surfaces, ...) • Thus we need a theoretical method which can not only calculate HFFparameters, but can also model the sample: • total energies + forces on the atoms: • perform structure optimization for “real” systems • calculate phonons • investigate various magnetic structures, exchange interactions • electronic structure: • bandstructure + DOS • compare with ARPES, XANES, XES, EELS, ... • hyperfine parameters • isomer shifts, hyperfine fields, electric field gradients WIEN2k software package An Augmented Plane Wave Plus Local Orbital Program for Calculating Crystal Properties Peter Blaha Karlheinz Schwarz Georg Madsen Dieter Kvasnicka Joachim Luitz November 2001 Vienna, AUSTRIA Vienna University of Technology WIEN2k: ~1530 groups mailinglist: 1800 users http://www.wien2k.at APW based schemes APW (J.C.Slater 1937) Non-linear eigenvalue problem Computationally very demanding LAPW (O.K.Andersen 1975) Generalized eigenvalue problem Full-potential (A. Freeman et al.) Local orbitals (D.J.Singh 1991) treatment of semi-core states (avoids ghostbands) APW+lo (E.Sjöstedt, L.Nordstörm, D.J.Singh 2000) Efficience of APW + convenience of LAPW Basis for K.Schwarz, P.Blaha, G.K.H.Madsen, Comp.Phys.Commun.147, 71-76 (2002) APW Augmented Plane Wave method The unit cell is partitioned into: atomic spheres Interstitial region unit cell Rmt Basisset: i ( k K ).r PW: ul(r,) are the numerical solutions e Atomic partial waves K A mu (r, )Ym (rˆ) m rI join of the radial Schrödinger equation in a given spherical potential for a particular energy AlmK coefficients for matching the PW Slater‘s APW (1937) Atomic partial waves K A mu (r, )Ym (rˆ) m Energy dependent basis functions lead to Non-linear eigenvalue problem H Hamiltonian S overlap matrix One had to numerically search for the energy, for which the det|H-ES| vanishes. Computationally very demanding. “Exact” solution for given (spherical) potential! Linearization of energy dependence LAPW suggested by center O.K.Andersen, Phys.Rev. B 12, 3060 (1975) kn [ A m antibonding bonding ( E , r )]Ym ( rˆ) (kn )u ( E , r ) Bm (kn )u m expand ul at fixed energy El and add ul ul / Almk, Blmk: join PWs in value and slope General eigenvalue problem (diagonalization) additional constraint requires more PWs than APW Atomic sphere LAPW APW PW Full-potential in LAPW (A.Freeman etal.) SrTiO3 The potential (and charge density) can be of general form (no shape approximation) VLM ( r )YLM ( rˆ) { V (r ) Full potential LM VK e iK . r rI K Inside each atomic sphere a local coordinate system is used (defining LM) Muffin tin approximation TiO2 rutile r Ra Ti O Problems of the LAPW method LAPW can only treat ONE principle quantum number per l. Problems with high-lying “semi-core” states Extending the basis: Local orbitals (LO) LO [ AmuE1 BmuE1 CmuE2 ]Ym (rˆ) LO: contains a second ul(E2) Ti atomic sphere is confined to an atomic sphere has zero value and slope at R can treat two principal QN n for each azimuthal QN (3p and 4p) corresponding states are strictly orthogonal (no “ghostbands”) tail of semi-core states can be represented by plane waves only slight increase of basis set (matrix size) D.J.Singh, Phys.Rev. B 43 6388 (1991) New ideas from Uppsala and Washington E.Sjöstedt, L.Nordström, D.J.Singh, An alternative way of linearizing the augmented plane wave method, Solid State Commun. 114, 15 (2000) • Use APW, but at fixed El (superior PW convergence) • Linearize with additional lo (add a few basis functions) kn Am (kn )u ( E , r )Ym (rˆ) lo [ AmuE1 BmuE1 ]Ym (rˆ) m optimal solution: mixed basis • use APW+lo for states which are difficult to converge: (f or d- states, atoms with small spheres) • use LAPW+LO for all other atoms and ℓ basis for Relativistic treatment For example: Ti Valence states Scalar relativistic Spin orbit coupling on demand by second variational treatment Semi-core states Scalar relativistic on demand mass-velocity Darwin s-shift spin orbit coupling by second variational treatment Additional local orbital (see Th-6p1/2) Core states Fully relativistic Dirac equation Relativistic semi-core states in fcc Th additional local orbitals for 6p1/2 orbital in Th Spin-orbit (2nd variational method) J.Kuneš, P.Novak, R.Schmid, P.Blaha, K.Schwarz, Phys.Rev.B. 64, 153102 (2001) Quantum mechanics at work w2web GUI (graphical user interface) Structure generator step by step initialization symmetry detection automatic input generation SCF calculations spacegroup selection import cif file Magnetism (spin-polarization) Spin-orbit coupling Forces (automatic geometry optimization) Guided Tasks Energy band structure DOS Electron density X-ray spectra Optics An example: In the following I will demonstrate on one example, which kind of problems you can solve using a DFT simulation with WIEN2k. Of course, also Mössbauer parameters will be calculated and interpreted. Verwey Transition and Mössbauer Parameters in YBaFe2O5 by DFT calculations Peter Blaha, C. Spiel, K.Schwarz Institute of Materials Chemistry TU Wien Thanks to: P.Karen (Univ. Oslo, Norway) C.Spiel, P.B., K.Schwarz, Phys.Rev.B.79, 085104 (2009) “Technical details”: WIEN2k (APW+lo) calculations Rkmax=7, 100 k-points spin-polarized, various spin-structures + spin-orbit coupling based on density functional theory: LSDA or GGA (PBE) Exc ≡ Exc(ρ, ∇ρ) description of “highly correlated electrons” using “non-local” (orbital dep.) functionals LDA+U, GGA+U hybrid-DFT (only for correlated electrons) mixing exact exchange (HF) + GGA WIEN2K An Augmented Plane Wave Plus Local Orbital Program for Calculating Crystal Properties Peter Blaha Karlheinz Schwarz Georg Madsen Dieter Kvasnicka Joachim Luitz http://www.wien2k.at Verwey-transition: Fe3O4, magnetite E.Verwey, Nature 144, 327 (1939) phase transition between a mixed-valence and a charge-ordered configuration with temp. 2 Fe2.5+ Fe2+ + Fe3+ cubic inverse spinel structure AB2O4 Fe2+A (Fe3+,Fe3+)B O4 Fe3+A (Fe2+,Fe3+)B O4 B A small, but complicated coupling between lattice and charge order Double-cell perovskites: RBaFe2O5 ABO3 O-deficient double-perovskite Ba Y (R) square pyramidal coordination Antiferromagnet with a 2 step Verwey transition around 300 K Woodward&Karen, Inorganic Chemistry 42, 1121 (2003) experimental facts: structural changes in YBaFe2O5 • above TN (~430 K): tetragonal (P4/mmm) • 430K: slight orthorhombic distortion (Pmmm) due to AFM all Fe in class-III mixed valence state +2.5; • ~334K: dynamic charge order transition into class-II MV state, visible in calorimetry and Mössbauer, but not with X-rays • 308K: complete charge order into class-I MV state (Fe2+ + Fe3+) large structural changes (Pmma) due to Jahn-Teller distortion; change of magnetic ordering: direct AFM Fe-Fe coupling vs. FM Fe-Fe exchange above TV structural changes charge ordered (CO) phase: Pmma a:b:c=2.09:1:1.96 (20K) valence mixed (VM) phase: Pmmm a:b:c=1.003:1:1.93 (340K) c b a Fe2+ and Fe3+ form chains along b contradicts Anderson charge-ordering conditions with minimal electrostatic repulsion (checkerboard like pattern) has to be compensated by orbital ordering and e--lattice coupling antiferromagnetic structure CO phase: G-type AFM AFM arrangement in all directions, also across Y-layer Fe moments in b-direction VM phase: AFM for all Fe-O-Fe superexchange paths FM across Y-layer (direct Fe-Fe exchange) 4 8 independent Fe atoms results of GGA-calculations: Metallic behaviour/No bandgap Fe-dn t2g states not splitted at EF overestimated covalency between O-p and Fe-eg Magnetic moments too small Experiment: Calculation: CO: 4.15/3.65 (for Tb), 3.82 (av. for Y) VM: ~3.90 CO: 3.37/3.02 VM: 3.34 no significant charge order charges of Fe2+ and Fe3+ sites nearly identical CO phase less stable than VM LDA/GGA NOT suited for this compound! Fe-eg t2g eg* t2g eg “Localized electrons”: GGA+U Hybrid-DFT ExcPBE0 [r] = ExcPBE [r] + a (ExHF[sel] – ExPBE[rsel]) LDA+U, GGA+U ELDA+U(r,n) = ELDA(r) + Eorb(n) – EDCC(r) separate electrons into “itinerant” (LDA) and localized e- (TM-3d, RE 4f e-) treat them with “approximate screened Hartree-Fock” correct for “double counting” Hubbard-U describes coulomb energy for 2e- at the same site orbital dependent potential Vm,m', (U J )( 1 nm,m', ) 2 Determination of U Take Ueff as “empirical” parameter (fit to experiment) or estimate Ueff from constraint LDA calculations constrain the occupation of certain states (add/subtract e-) switch off any hybridization of these states (“core”-states) calculate the resulting Etot we used Ueff=7eV for all calculations DOS: GGA+U vs. GGA GGA+U GGA singleinsulator, lower Hubbard-band t2g band splits in VM splits in CO with Fe3+ states metallic lower than Fe2+ insulator GGA+U metal magnetic moments and band gap magnetic moments in very good agreement with exp. LDA/GGA: CO: 3.37/3.02 VM: 3.34 mB orbital moments small (but significant for Fe2+) band gap: smaller for VM than for CO phase exp: semiconductor (like Ge); VM phase has increased conductivity LDA/GGA: metallic Charge transfer (in GGA+U) Charges according to Baders “Atoms in Molecules”theory Define an “atom” as region within a zero flux surface r n 0 Integrate charge inside this region Structure optimization (GGA+U) O2a CO phase: Fe2+: shortest bond in y (O2b) 3+: shortest bond in z (O1) Fe O2b VM phase: all Fe-O distances similar theory deviates along z !! Fe-Fe interaction different U ?? finite temp. ?? O3 O1 O1 strong coupling between lattice and electrons ! Fe2+ (3d6) strong covalency effects in eg and d-xz orbitals CO Fe3+ (3d5) VM Fe2.5+ (3d5.5) majority-spin fully occupied very localized states at lower energy than Fe2+ minority-spin states d-xz fully occupied (localized) empty short bond in y short bond in z (missing O) d-z2 partly occupied FM Fe-Fe; distances in z ?? Difference densities Dr=rcryst-ratsup CO phase VM phase Fe2+: d-xz Fe3+: d-x2 O1 and O3: polarized toward Fe3+ Fe: d-z2 Fe-Fe interaction O: symmetric dxz spin density (rup-rdn) of CO phase Fe3+: no contribution Fe2+: dxz weak p-bond with O tilting of O3 p-orbital Mössbauer spectroscopy: Isomer shift: d = a (r0Sample – r0Reference); a=-0.291 au3mm s-1 proportional to the electron density r at the nucleus Magnetic Hyperfine fields: Btot=Bcontact + Borb + Bdip Bcontact = 8p/3 mB [rup(0) – rdn(0)] … spin-density at the nucleus … orbital-moment … spin-moment S(r) is reciprocal of the relativistic mass enhancement nuclear quadrupole interaction 1 2V (0) e ( x) xi x j dx ij Q 2 ij xi x j 2h Nuclei with a nuclear quantum number I≥1 have an electric quadrupole moment Q, which describes the asphrericity of the nucleus Nuclear quadrupole interaction (NQI) can aid to determine the distribution of the electronic charge surrounding such a nuclear site Experiments NMR NQR Mössbauer PAC eQ / h electric field gradient (EFG) 1 ij Vij d ij 2V 3 2V (0) Vij xi x j Electric field gradients (EFG) EFG traceless tensor 1 ij Vij d ij 2V 3 2V (0) Vij xi x j with Vxx V yy Vzz 0 Vaa Vab Vac Vba Vbb Vbc Vca Vcb Vcc V xx 0 0 0 V yy 0 0 0 Vzz similarity transformation traceless |Vzz| |Vyy| |Vxx| „EFG“: Vzz principal component / Vxx / / V yy / / Vzz / asymmetry parameter …0-1 directions of Vxx, Vyy, Vzz cubic: no EFG; hex, tetragonal (axial symm.): only Vzz theoretical EFG calculations EFG is tensor of second derivatives of VC at the nucleus: 2V (0) Vij xi x j Vzz Vc (r ) r ( r )Y20 1 V 3 r r3 p zz 1 V 3 r p d xy r r dr VLM (r )YLM (rˆ) LM x p y ) pz Vzz V20 ( r 0) Vxx 1 V20 V22 2 d x 2 y 2 1 (d xz d yz ) d z 2 2 Cartesian LM-repr. V yy 1 V20 V22 2 dr Vzzp Vzzd 12 ( p d d zz r (r) EFG is proportial to differences of orbital occupations Mössbauer spectroscopy CO VM Isomer shift: charge transfer too small in LDA/GGA CO VM Isomer shifts d = a (r0Sample – r0Ref.) The observed IS is proportional to the charge transfer between the different sites (source and absorber). The 4s (valence) and 3s (semicore) contributions provide 110-120% of the effect, and are reduced slightly by opposite 1s,2s contributions. a is proportional to the change of nuclear radii during the transition, (in principle a “known nuclear constant” ). However, a meaningful a depends on details of the calculations (we use a=-0.291 au3mm s-1): radial mesh for numerical basis functions (first radial mesh point, RMT) relativistic treatment (NREL, scalar/fully relativistic, point/finite nucleus) Statements for YBa2Fe2O5 standard LDA/GGA lead to much too small charge transfer and thus the IS are too similar for all 3 Fe sites. need better XC-treatment with increased on-site correlation U , the “localization” and thus the charge transfer is larger leading to more different IS in agreement with experiment. Fe site charge IS Fe2+ +1.36 0.96 Fe3+ Fe2.5+ +1.84 +1.62 0.28 0.51 Hyperfine fields: Fe2+ has large Borb and Bdip CO VM contact hyperfine fields The contact HFF dominates in many cases. It is proportional to the “spin-density” at the nucleus. Bcontact = 8p/3 mB [rup(0) – rdn(0)] It is also proportional to the spin-magnetic moment (but with opposite sign) because the (3d) magnetic moment polarizes the core s-electrons (see below), provided the valence (4s) contribution is small In many cases the “core-polarization” dominates (the core electrons are polarized due to the 3d magnetic moment): 1s 2s,p 3s,p 4s,p total - 1.4 T - 113.7 T opposite sign; exchange effect with 3s wf. + 60.7 T same sign as 3d moment, strong overlap of 3d and 3s,p wf. + 16.3 T often even smaller, but can be large due to ligand - 38.1 T effects (transferred hyperfine fields) Typically, in DFT the contact HFF is about 10-20% too small Orbital and dipolar - hyperfine fields Borb is proportional to the “orbital moment” (an effect due to spin-orbit coupling) In metals and for high-spin Fe3+ (3d5-up) compounds (closed shell ions) it is usually small. In insulators/semiconductors with partially filled 3d-shells it can be very large (high-spin Fe2+: 3d5-up, 3d1-dn) LDA/GGA usually underestimates this contribution Bdip is proportional to the anisotropy of the spin-moments around the nucleus In many cases it is fairly small EFG: Fe2+ has too small anisotropy in LDA/GGA CO VM theoretical EFG calculations • The coulomb potential Vc is a central quantity in any theoretical calculation (part of the Hamiltonian) and is obtained from all charges r (electronic + nuclear) in the system. Vc (r ) r (r ) r r dr VLM (r )YLM (rˆ) LM • The EFG is a tensor of second derivatives of VC at the nucleus: 2V (0) Vij ; xi x j Vzz r (r )Y20 r 3 dr • Since we use an “all-electron” method, we have the full charge distribution of all electrons+nuclei and can obtain the EFG without further approximations. • The spherical harmonics Y20 projects out the non-spherical (and non-cubic) part of r. The EFG is proportional to the differences in orbital occupations (eg. pz vs. px,py) • We do not need any “Sternheimer factors” (these shielding effects are included in the self-consistent charge density) theoretical EFG calculations r (r )Y20 The charge density r in the integral Vzz r3 be decomposed in various ways for analysis: dr can according to energy (into various valence or semi-core contributions) according to angular momentum l and m (orbitals) spatial decomposition into “atomic spheres” and the “rest” (interstital) Due to the 1/r3 factor, contributions near the nucleus dominate. EF theoretical EFG calculations We write the charge density and the potential inside the atomic spheres in an lattice-harmonics expansion r (r ) r LM (r )YLM (rˆ) V (r ) vLM (r )YLM (rˆ) LM LM spatialdecomposition : Vzz r (r )Y20 r3 d r 3 Vzz r LM r3 r 20 (r ) sphere r 3 d r 3 LM sphere (r )YLM Y20 int erstital r (r )Y20 r3 d 3r dr interstitial orbitaldecomposition : r 20 (r ) Y drˆ nk * nk lm l m 20 k , n.l ,l , m , m Vzz Vzzpp Vzzdd ..... interstitial p p; d d ; ( s d ) contr. theoretical EFG calculations Vzz Vzzpp Vzzdd ..... interstitial V V pp zz dd zz 1 3 r 1 3 r p 1 2 ( p d d x p y ) pz 1 d (d xz d yz ) d z 2 2 2 xy x y 2 • EFG is proportial to differences of orbital occupations , e.g. between px,py and pz. • if these occupancies are the same by symmetry (cubic): Vzz=0 • with “axial” (hexagonal, tetragonal) symmetry (px=py): =0 In the following various examples will be presented. Nuclear Quadrupole moment Q of 56Fe Compare theoretical and experimental EFGs 1 D Q eQV zz 2 ( previous value) FeSi Fe2O3 FeS2 FeCl2 YFe2 Fe4N FeNi FeZr3 Fe2P FeBr2 (exp.) Q FeF2 testing various DFT approximations: GGA works surprisingly well in FeF2: LSDA • There is ONE e- in the three Fe d-t2g states. • LDA: wrong metallic state, e- is distributed in all 3 orbitals wrong charge density and EFG • GGA splits Fe d-t2g states into a1g and eg’ correct charge density and EFG Fe-EFG LSDA: GGA: exp: in FeF2: 6.2 16.8 16.5 GGA EFGs in fluoroaluminates 10 different phases of known structures from CaF2-AlF3, BaF2-AlF3 binary systems and CaF2-BaF2-AlF3 ternary system Isolated octahedra Isolated chains of octahedra linked by corners a-BaAlF5 a-CaAlF5, b-CaAlF5, b-BaAlF5, g-BaAlF5 Rings formed by four octahedra sharing corners a-BaCaAlF7 Ca2AlF7, Ba3AlF9-Ib, b-Ba3AlF9 Ba3Al2F12 Q and Q calculations using XRD data AlF3 Q,exp = 0,803 Q,cal R2 = 0,38 a-CaAlF -16 |V | with R2 = 0,77 Q = 4,712.10performed Attributions with respect to the proportionality between |Vzz| zz b-CaAlF 1,0 AlF and Q for the multi-site compounds Ca a-BaAlF 5 5 1,8e+6 2 AlF3 5 a-CaAlF5 1,6e+6 b-BaAlF5 b-CaAlF5 g-BaAlF5 Ca2AlF7 Ba3AlF9-Ib b-BaAlF5 b-Ba3AlF9 g-BaAlF5 a-BaCaAlF7 Ba3Al2F12 Experimental Q Experimental Q (Hz) 1,2e+6 Ba3Al2F12 0,8 a-BaAlF5 1,4e+6 Ba3AlF9-Ib b-Ba3AlF9 1,0e+6 7 a-BaCaAlF7 Régression 8,0e+5 Regression Q,mes. =Q, cal. 0,6 0,4 6,0e+5 4,0e+5 0,2 2,0e+5 0,0 0,0 0,0 1,0e+21 2,0e+21 3,0e+21 -2 Calculated Vzz (V.m ) 0,0 0,2 0,4 0,6 0,8 Calculated Q Important discrepancies when structures are used which were determined from X-ray powder diffraction data 1,0 Q and Q after structure optimization 1,8e+6 1,0 Q = Vzz R2 = 0,993 5,85.10-16 1,6e+6 0,8 1,4e+6 1,2e+6 Experimental Q Experimental Q (Hz) Q, exp = 0,972 Q,cal R2 = 0,983 1,0e+6 AlF3 a-CaAlF5 8,0e+5 b-CaAlF5 Ca2AlF7 6,0e+5 0,6 AlF3 a-CaAlF5 b-CaAlF5 Ca2AlF7 a-BaAlF5 0,4 a-BaAlF5 b-BaAlF5 b-BaAlF5 g-BaAlF5 g-BaAlF5 4,0e+5 Ba3Al2F12 Ba3Al2F12 Ba3AlF9-Ib Ba3AlF9-Ib 0,2 b-Ba3AlF9 b-Ba3AlF9 2,0e+5 a-BaCaAlF7 a-BaCaAlF7 Regression Q,exp. =Q, cal. Regression 0,0 0,0 0,0 5,0e+20 1,0e+21 1,5e+21 2,0e+21 2,5e+21 3,0e+21 0,0 0,2 -2 Calculated Vzz (V.m ) 0,4 0,6 0,8 Calculated Q Very fine agreement between experimental and calculated values M.Body, et al., J.Phys.Chem. A 2007, 111, 11873 (Univ. LeMans) 1,0 EFG (1021 V/m2) in YBa2Cu3O7 Site Y theory exp. Ba theory exp. Cu(1) theory exp. Cu(2) theory exp. O(1) theory exp. O(2) theory exp. O(3) theory exp. O(4) theory exp. Vxx -0.9 -8.7 8.4 -5.2 7.4 2.6 6.2 -5.7 6.1 12.3 10.5 -7.5 6.3 -4.7 4.0 Vyy 2.9 -1.0 0.3 6.6 7.5 2.4 6.2 17.9 17.3 -7.5 6.3 12.5 10.2 -7.1 7.6 Vzz -2.0 9.7 8.7 -1.5 0.1 -5.0 12.3 -12.2 12.1 -4.8 4.1 -5.0 3.9 11.8 11.6 0.4 0.8 0.9 0.6 1.0 0.0 0.0 0.4 0.3 0.2 0.2 0.2 0.2 0.2 0.3 standard LDA calculations give good EFGs for all sites except Cu(2) K.Schwarz, C.Ambrosch-Draxl, P.Blaha, Phys.Rev. B42, 2051 (1990) D.J.Singh, K.Schwarz, K.Schwarz, Phys.Rev. B46, 5849 (1992) Interpretation of O-EFGs in YBa2Cu3O7 px py pz Vaa Vbb Vcc O(1) 1.18 0.91 1.25 -6.1 18.3 -12.2 O(2) 1.01 1.21 1.18 11.8 -7.0 -4.8 O(3) 1.21 1.00 1.18 -7.0 11.9 -4.9 O(4) 1.18 1.19 0.99 -4.7 -7.0 11.7 O(1) O(4) z O(2),O(3) y Asymmetry count x 1 Dn p p z ( p x p y ) 2 EFG (p-contribution) EF Cu1-d 1 p Vzz Dn p 3 p r difference density Dr EFG is proportional to the asymmetric charge distribution around a given nucleus O1-py partly occupied Cu(2) and O(4) EFG as function of r EFG is determined by the non-spherical charge density inside sphere r (r )Y20 r (r ) r LM (r )YLM Vzz dr r 20 (r ) / r dr 3 r LM Cu(2) r O(4) r final EFG semicore and valence Cu-EFG contributions semicore Cu 3p-states have very little importance valence 3d-states: large contribution due to smaller d-x2-y2 occupation valence 4p-states: large contribution of opposite sign. Originates from the tails of the O-2p orbitals (“re-expanded” as Cu 4p in the Cu atomic sphere, “off-site” contribution) usually only contributions within the first node or within 1 bohr are important. general statements for EFG contributions Depending on the atom, the main EFG-contributions come from anisotropies in occupations of different orbitals or the radial wave functions of different orbitals) semicore p-states: they are of course always fully occupied, but due to an anisotropic neighborhood they may slightly contract/expand in different directions. The effect depends on: the energy and spatial localization: Ti 3p more important than Cu 3p (almost inert) the distance, type (atom) and geometry of the neighbors (a small octahedral distortion will produce a much smaller effect than a square-planar coordination) valence p-states : always important (“on-site” O 2p or “off-site” Cu 4p) valence d-states : in ionic or covalent TM compounds. In metals usually “small”. valence f-states : very large contributions for “localized” 4f,5f systems unless the f-shell is “half-filled” (or full) LDA/GGA problems in correlated TM oxides As shown before, the approximations for exchange-correlation (LDA, GGA) can have significant influence on the quality of the results for the class of compounds with so called “correlated electrons”. Such electrons are in particular TM-3d electrons in TM-oxides (or, more general, in most ionic TM-compounds), but also the 4f and 5f electrons of lanthanides or actinides. For these systems, “beyond-GGA” schemes like “LDA+U” or “hybridfunctionals” (mixing of Hartree-Fock+GGA) are necessary, as was demonstrated before for YBa2Fe2O5. Another example of very bad EFGs within GGA would be the class of the undoped Cuprates (La2CuO4, YBa2Cu3O6), which are nonmagnetic metals instead of antiferromagnetic insulators in GGA. Both, doped and undoped cuprates have a planar Cu – EFG in GGAcalculations, which is by a factor of 2-3 too small compared to experiments. Cuprates: La2CuO4 LDA: nonmagnetic metal LDA+U: AFM insulator (in agreement with experiment) lower Hubbard-band upper HB Magn. moments and EFG in La2CuO4 AMF exp. FLL LDA+U gives AF insulator with reasonable moment U of 5-6 eV gives exp. EFG GGAs “mimic” a U of 1-2 eV (EV-GGA more “effective” than PBE !!) Cu-EFG Vzz (1021 V/m2) in YBa2Cu3O6 for LDA, AMF, FLL and DFT-double counting corrections. NM and AF refers to non-magnetic and antiferromagnetic solutions. U and J of 8 and 1 eV is applied to both Cu sites, except for LDA+U* where U is applied only to the Cu(2) site. Type Vzz -Cu(1) LDA (NM) -8.1 AMF-LDA+U (NM) -4.6 AMF-LDA+U*(NM) -8.0 AMF-LDA+U (AF) -4.5 AMF-LDA+U*(AF) -8.0 DFT-LDA+U(AF) -4.8 FLL-LDA+U (AF) -8.3 FLL-LDA+U*(AF) -8.0 Experiment 11.8 Vzz -Cu(2) -3.7 -7.3 -7.2 -13.3 -13.3 -12.0 -12.3 -13.3 9.0 Cu(2) too small, Cu(1) ok Cu(2) still too small, Cu(1) wrong Cu(2) ok, Cu(1) wrong DFT similar to AMF Cu(2) ok,Cu(1) remains ok Antiferromagn. FLL-calc. with U=6eV give again best results EFG analysis in YBa2Cu3O6: Vzzp 1 r3 12 ( p p ) p x p y z EFG contributions (for both spins and p-p and d-d contributions) in LSDA and LDA+U(DFT) Cu(1) Cu(2) NM-LDA AF-LDA+U(DFT) NM-LDA AF-LDA+U(DFT) p-p (up) -12.2 -12.2 5.7 5.8 large charge in z p-p (dn) -12.2 -12.2 5.7 6.3 large negative EFG d-d (up) 7.1 8.8 -7.4 5.5 d-d (dn) 7.1 8.8 -7.4 -31.1 Semicore 2.1 2.0 -0.1 1.7 large charge in xy Total -8.1 -4.8 -3.7 -12.0 large positive EFG Exp. 11.8 9.0 Partial charges and anisotropy counts Dn in LDA and antiferromagnetic LDA+U(DFT) Cu(1) Cu(2) NM-LDA AF-LDA+U(DFT) NM-LDA AF-LDA+U(DFT) 4pz 0.103 0.104 0.027 0.028 4px+py 0.041 0.042 0.131 0.140 Dnp -0.083 -0.083 0.039 0.042 2 2 dx -y 1.776 1.782 1.433 1.228 Cu(1) anisotropy increased in 2 dz 1.474 1.398 1.757 1.784 DFT and AMF schems: wrong! Cu(2) has O in xy plane daver 1.802 1.818 1.815 1.841 Dnd 0.257 0.324 -0.294 -0.529 Cu(2) dx2-y2 depopulated: ok Cu(1) has O only in z Summary EFGs can routinely be calculated for all kinds of solids. “semi-core” contribution large for “left”-atoms of the periodic table p-p contribution always large (on-site (eg. O-2p) vs. off-site (Fe-4p)) d-d (f-f) contributions for TM (lanthanide/actinide) compounds EFG stems from different orbital occupations due to covalency or crystal field effects forget “point-charge models” and “Sternheimer anti-shielding factors” EFG is very sensitive to correct structural data (internal atomic positions) correct theoretical description of the electronic structure “highly correlated” transition metal compounds (oxides, halides) 4f and 5f compounds “beyond” LDA (LDA+U, Hybrid-DFT, …) Literature WIEN2k and APW-based methods: P.Blaha, K.Schwarz, P.Sorantin and S.B.Trickey: Full-potential, linearized augmented plane wave programs for crystalline systems , Comp.Phys.Commun. 59, 399 (1990) P.Blaha and K.Schwarz: WIEN93: An energy band-structure program for ab initio calculations of electric field gradients in solids, NQI Newsletter Vol 1 (3), 32 (1994) K.Schwarz and P.Blaha: Description of an LAPW DF Program (WIEN95), in: Lecture notes in chemistry, Vol. 67, p.139, Ed. C.Pisani, Springer (Berlin 1996) H.Petrilli, P.E.Blöchl, P.Blaha, and K.Schwarz: Electric-field-gradient calculations using the projector augmented wave method, Phys.Rev. B57, 14690 (1998) G.Madsen, P.Blaha, K.Schwarz, E.Sjöstedt and L.Nordström: Efficient linearization of the augmented plane-wave method, Phys.Rev. B64, 195134 (2001) P.Blaha, K.Schwarz, G.Madsen, D.Kvasnicka and J.Luitz: WIEN2k: An augmented plane wave plus local orbitals program for calculating crystal properties. K.Schwarz, TU Wien, 2001 (ISBN 3-9501031-1-2) (http://www.wien2k.at) K.Schwarz, P.Blaha and G.K.H.Madsen: Electronic structure calculations of solids using the WIEN2k package for material sciences, Comp.Phys.Commun. 147, 71 (2002) D. J. Singh and L. Nordstrom: Planewaves, Pseudopotentials and the LAPW Method, Springer, Berlin 2006. Literature DFT and beyond: P. Hohenberg and W. Kohn: Inhomogeneous electron gas, Phys. Rev. 136, B864 1964. W. Kohn and L. J. Sham: Selfconistent equations including exchange and correlation effects, Phys. Rev. 140, A1133 1965. V. I. Anisimov, J. Zaanen, and O. K. Andersen: Band theory and Mott insulators: Hubbard U instead of Stoner I, Phys. Rev.B 44, 943 (1991). J. P. Perdew, K. Burke, and M. Ernzerhof: Generalized gradient approximation made simple, Phys. Rev. Lett. 77, 3865 (1996); 78, 1396 (1997). A Primer in Density Functional Theory, edited by C. Fiolhais, F. Nogueira, and M. Marques, Springer, Berlin, 2003. F. Tran, P. Blaha, K. Schwarz, P. Novak: Hybrid exchange-correlation energy functionals for strongly correlated electrons: Applications to transition-metal monoxides; Phys.Rev B, 74 (2006), 155108. P. Haas, F. Tran, P. Blaha: Calculation of the lattice constant of solids with semilocal functionals; Physical Review B, 79 (2009), 085104, Physical Review B, 79 (2009), 209902(E) F.Tran and P.Blaha: Accurate band gaps of semiconductors and insulators with a semilocal exchange-correlation potential, Physical Review Letters, 102, (2009) 226401. Literature Hyperfine parameters using WIEN2k: P.Blaha, K.Schwarz and P.Herzig: First-principles calculation of the electric field gradient of Li3N , Phys.Rev.Lett. 54, 1192 (1985) P.Blaha, K.Schwarz and P.H.Dederichs: First-principles calculation of the electric field gradient in hcp metals , Phys.Rev. B37, 2792 (1988) P.Blaha and K.Schwarz: Theoretical investigation of isomer shifts in Fe, FeAl, FeTi and FeCo , J.Phys.France C8, 101 (1988) P.Blaha, P.Sorantin, C.Ambrosch and K.Schwarz: Calculation of the electric field gradient tensor from energy band structures , Hyperfine Interactions 51, 917 (1989) P.Blaha and K.Schwarz: Electric field gradient in Cu2O from band structure calculations, Hyperfine Interact. 52, 153 (1989) P.Blaha: Calculation of the pressure dependence of the EFG in bct In, hcp Ti and Zn from energy band structures , Hyperfine Interact. 60, 773 (1990) C.Ambrosch-Draxl, P.Blaha and K.Schwarz: Calculation of EFGs in high Tc superconductors , Hyperfine Interactions 61, 1117 (1990) K.Schwarz, C.Ambrosch-Draxl, P.Blaha, Charge distribution and electric-field gradients in YBa2Cu3O7-x , Phys.Rev. B42, 2051 (1990) P.Blaha, K.Schwarz and A.K.Ray: Isomer shifts and electric field gradients in Y(Fe1-xAlx)2, J.Magn.Mag.Mat. 104-107, 683 (1992) D.J.Singh, K.Schwarz, K.Schwarz: Electric-field gradients in YBa2Cu3O7: Discrepancy between experimental and local-density-approximation charge distributions, Phys.Rev. B46, 5849 (1992) Literature P.Blaha, D.J.Singh, P.I.Sorantin and K.Schwarz: Electric field gradient calculations for systems with giant extended core state contributions, Phys.Rev. B46, 1321 (1992) W.Tröger, T.Butz, P.Blaha and K.Schwarz: Nuclear quadrupole interaction of 199mHg in mercury(I) and mercury(II) halides, Hyperfine Interactions 80, 1109 (1993) P.Blaha, P.Dufek, and K.Schwarz: Electric field gradients, isomer shifts and hyperfine fields from band structure calculations in NiI2, Hyperfine Inter. 95, 257 (1995) P.Dufek, P.Blaha and K.Schwarz: Determination of the nuclear quadrupole moment of 57Fe, Phys.Rev.Lett. 75, 3545 (1995) P.Blaha, P.Dufek, K.Schwarz and H.Haas: Calculations of electric hyperfine interaction parameters in solids, Hyperfine Int. 97/98, 3 (1996) K.Schwarz, H.Ripplinger and P.Blaha: Electric field gradient calculations of various borides; Z.Naturforsch. 51a, 527 (1996) B.Winkler, P.Blaha and K.Schwarz: Ab initio calculation of electric-field-gradient tensors of forsterite, American Mineralogist, 81, 545 (1996) H.Petrilli, P.E.Blöchl, P.Blaha, and K.Schwarz: Electric-field-gradient calculations using the projector augmented wave method, Phys.Rev. B57, 14690 (1998) P.Blaha, K.Schwarz, W.Faber and J.Luitz: Calculations of electric field gradients in solids: How theory can complement experiment, Hyperfine Int. 126, 389 (2000) M.Divis, K.Schwarz, P.Blaha, G.Hilscher, H.Michor , S.Khmelevskyi: Rare earth borocarbides: Electronic structure calculations and electric field gradients, Phys.Rev. B62, 6774 (2000) Literature G.Principi, T.Spataru, A.Maddalena, A.Palenzona, P.Manfrinetti, P.Blaha, K.Schwarz, V.Kuncser and G.Filotti: A Mössbauer study of the new phases Th4Fe13Sn5 and ThFe0.22Sn2, J.Alloys and Compounds 317-318, 567 (2001) R. Laskowski, G.K.H. Madsen, P. Blaha, K. Schwarz: Magnetic structure and electric-field gradients of uranium dioxide: An ab initio study; Physical Review B, 69 (2004), S. 140408(R). P. Palade, G. Principi, T. Spataru, P. Blaha, K. Schwarz, V. Kuncser, S. Lo Russo, S. Dal Toe, V. Yartys: Mössbauer study of LaNiSn and NdNiSn compounds and their deuterides; Journal of Radioanalytical and Nuclear Chemistry, 266 (2005), 553 - 556. P. Blaha, K. Schwarz, P. Novak: Electric Field Gradients in Cuprates: Does LDA+U give the Correct Charge Distribution ?; International Journal of Quantum Chemistry, 101 (2005), 550 M. Body, C. Legein, J. Buzare, G. Silly, P. Blaha, C. Martineau, F. Calvayrac: Advances in Structural Analysis of Fluoroaluminates Using DFT Calculations of 27Al Electric Field Gradients; Journal of Physical Chemistry A, 111 (2007), S. 11873 - 11884. Seung-baek Ryu, Satyendra K. Das, Tilman Butz, Werner Schmitz, Christian Spiel, Peter Blaha, and Karlheinz Schwarz: Nuclear quadrupole interaction at 44Sc in the anatase and rutile modifications of TiO2 : Time-differential perturbed-angular-correlation measurements and ab initio calculations, Physical Review B, 77 (2008), S. 094124 C. Spiel, P. Blaha, K. Schwarz: Density functional calculations on the chargeordered and valence-mixed modification of YBaFe2O5; Phys. Rev. B 79 (2009), 115123.