Mass enhancement in 3d and s−p perovskites from symmetry breaking

Abstract

In some d-electron oxides the measured effective mass 𝑚 exptl * has long been known to be significantly larger than the model effective mass 𝑚 model * deduced from mean-field band theory, i.e., 𝑚 exptl * = 𝛽𝑚 model * , where 𝛽 > 1 is the 'mass enhancement', or 'mass renormalization' factor. Previous applications of density functional theory (DFT), based on the smallest number of possible magnetic, orbital, and structural degrees of freedom, missed such mass enhancement, a fact that was taken as evidence of strong electronic correlation. The current paper reports in a range of d-electron perovskites SrVO3, SrTiO3, BaTiO3 and LaMnO3 as well as p-electron perovskites CsPbI3 and SrBiO3 that the symmetry-breaking spin and structural effects included in DFT captures already the magnitudes and trends in mass enhancement for both electrons and holes, but only when enlarged unit cells, which are large enough to allow for symmetry breaking distortions and concomitant variations in spin order, are explored for their ability to lower the total energy. The paper analyzes the different symmetry breaking modalities contributing to mass enhancement, finding common effects in range of d-as well as p-electron perovskites. * , where 𝛽 is the 'mass enhancement', or 'mass renormalization' factor. Effective masses 𝑚 exptl * are generally deduced from experiment via model assumptions (such as band parabolicity or various averages over the mass tensor), leading to different effective mass definitions in different experiments, including the mass 𝑚 * ∝ 1 𝑣 𝐹 ⁄ , deduced from Fermi-velocity (𝑣 𝐹 ), or from density of states (DOS) 𝑚 * ∝ (𝐷(𝐸)) 2 3 ⁄ , from specific-heat coefficient 𝑚 * ∝ 𝛾, from magnetic susceptibility 𝜒 ∝ 𝑚 * (1 − 𝑚 0 2 3𝑚 * 2 ), and from bandwidth 𝑊 ∝ 1 𝑚 * ⁄ . Values of 𝛽 > 1 were reported in the * were compared with theoretical values 𝛽(Theory: model) = 𝑚 Theory * /𝑚 model * obtained from advanced theory (such as dynamic mean-field theory, DMFT [13-23]). Because 𝑚 model * is obtained from mean-field band theory, the predicted theoretical enhancement 𝛽(Theory: model) > 1 has been interpreted to be due to strong correlation effects [13-23]. For example, in DMFT, wavefunction localizes and bandwidth narrows (thus leading to mass enhancement) due to pure electronic symmetry breaking [15,24] induced by the dynamic self-energy from the impurity atom embedded in a mean-field bath. Finding for a compound that 𝛽(DMFT: model) > 1 consistent with 𝛽(exptl: model) > 1 helped classify the pertinent compounds as being highly correlated. This line of thinking, however, does not consider the possibility that sources of mass enhancement other than symmetry-preserving correlation might exist. The model calculations used to extract 𝑚 model * have invariably been [13-23] rather naïve (N) level of density functional theory (N-DFT), based on the least number of possible magnetic, orbital and structural degrees of freedom. Indeed, they have assumed one or a few of the following approximations: A highly symmetric unit cell symmetry (e.g., 𝑃𝑚3 ̅...

Piezoelectricity in Nominally Centrosymmetric Phases

“…pair distribution function PDF, Ref [75]), whereas they often escape detection by volumeaveraging techniques such as conventional XRD. Indeed, so are local time reversal symmetry breaking predicted theoretically in paramagnets [75,77], even when the global magnetism (vector sum of local moments) vanishes. Similarly, we see in the current work that, whereas the global dipole in a paraelectric could be small or vanishing, symmetry breaking calculations [75] indicate that local dipoles need not vanish.…”

“…Internal vs external macroscopic symmetry breaking: We intentionally did not distinguish between internal [31,32,[75][76][77][78] and external symmetry breaking [1,2] on the detected piezoelectricity. Although, there is no experimental evidence of polar nanostructures in NaCl, silica glass, CaF 2 , KTaO 3 , and SrTiO 3 at room temperature, recent observations of macroscopic symmetry breaking in cubic materials due to internal effects [75][76][77][78] suggest that bulk symmetry breaking could occur even in the absence of defects. For example, birefringence observed in CaF 2 has been proposed to stem from intrinsic effects [80].…”

“…These concepts have been recently challenged [74]. Indeed, there are recent theoretical reasons [75][76][77][78][79] to believe that intrinsic mechanisms might be responsible for the formation of atomicscale symmetry breaking in para phases, manifesting short-range-ordered (SRO) local motifs ("polymorphous network") [75]. Such symmetry breaking were argued theoretically to lower the internal energy U 0 even before thermal agitation sets in, leading to the formation of a distribution of local motifs in all para phases.…”

“…This then allows the observation of physical effects that reflect noncentrosymmetric symmetry, including piezoelectricity. Recent first-principles calculations [75][76][77][78] considered a supercell of numerous Pm-3m unit cells, preserving the global cubic shape symmetry, but then minimizes the cell-internal atomic forces in Density Functional Theory (DFT). One finds in such constrained minimization of the internal energy U 0 a symmetry breaking.…”

“…Depending on the type of pertinent local degree of freedom, these local motifs can correspond to local octahedra rotations in cubic halide perovskites (CsPbI 3 ) [75,76], or to local magnetic moments in paramagnetic oxides (YTiO 3 , YNiO 3 ) [78], or to local displacements creating local dipoles in a paraelectric (BaTiO 3 ) [75]. All such symmetry breakings represent short range order, SRO, and lead to macroscopic effects on the band structure, effective masses, optical properties [75][76][77][78].…”