arXiv subjects
Publications and source records attributed to John P. Perdew.
Successful modern generalized gradient approximations (GGA) are biased toward atomic energies. Restoration of the first-principles gradient expansion for the exchange energy over a wide range of density gradients eliminates this bias. We introduce PBEsol, a revised Perdew-Burke-Ernzerhof GGA that improves equilibrium properties for many densely-packed solids and their surfaces.
While the exact total energy of a separated open system varies linearly as a function of average electron number between adjacent integers, the energy predicted by semi-local density functional approximations curves upward and the exact-exchange-only or Hartree-Fock energy downward. As a result, semi-local density functionals fail for separated open systems of fluctuating electron number, as in stretched molecular ions A$_2^{+}$ and in solid transition metal oxides. We develop an exact-exchange theory and an exchange-hole sum rule that explain these failures and we propose a way to correct them via a local hybrid functional.
We construct a Laplacian-level meta-generalized gradient approximation (meta-GGA) for the non-interacting (Kohn-Sham orbital) positive kinetic energy density $τ$ of an electronic ground state of density $n$. This meta-GGA is designed to recover the fourth-order gradient expansion $τ^{GE4}$ in the appropiate slowly-varying limit and the von Weizsäcker expression $τ^{W}=|\nabla n|^2/(8n)$ in the rapidly-varying limit. It is constrained to satisfy the rigorous lower bound $τ^{W}(\mathbf{r})\leqτ(\mathbf{r})$. Our meta-GGA is typically a strong improvement over the gradient expansion of $τ$ for atoms, spherical jellium clusters, jellium surfaces, the Airy gas, Hooke's atom, one-electron Gaussian density, quasi-two dimensional electron gas, and nonuniformly-scaled hydrogen atom. We also construct a Laplacian-level meta-GGA for exchange and correlation by employing our approximate $τ$ in the Tao, Perdew, Staroverov and Scuseria (TPSS) meta-GGA density functional. The Laplacian-level TPSS gives almost the same exchange-correlation enhancement factors and energies as the full TPSS, suggesting that $τ$ and $\nabla^2 n$ carry about the same information beyond that carried by $n$ and $\nabla n$. Our kinetic energy density integrates to an orbital-free kinetic energy functional that is about as accurate as the fourth-order gradient expansion for many real densities (with noticeable improvement in molecular atomization energies), but considerably more accurate for rapidly-varying ones.
Under a certain scaling, the electron densities of finite systems become both large and slowly-varying, so that the gradient expansions of the density functionals for the Kohn-Sham kinetic and exchange energies become asymptotically exact to order $\nabla^2$. Neutral atoms of large $Z$ scale similarly, but a cusp correction at the nucleus requires generalizing the gradient expansion for exchange, producing the wrong gradient coefficient in the slowly-varying limit. Meta-generalized gradient approximations (meta-GGA's) recover both the slowly-varying and large-$Z$ limits. GGA correlation energies of large-Z atoms are found to be accurate.
The negative correlation energy per particle of a uniform electron gas of density parameter $r_s$ and spin polarization $ζ$ is well known, but its spin resolution into up-down, up-up, and down-down contributions is not. Widely-used estimates are incorrect, and hamper the development of reliable density functionals and pair distribution functions. For the spin resolution, we present interpolations between high- and low-density limits that agree with available Quantum Monte Carlo data. In the low-density limit for $ζ= 0$, we find that the same-spin correlation energy is unexpectedly positive, and we explain why. We also estimate the up and down contributions to the kinetic energy of correlation.
The electron density, its gradient, and the Kohn-Sham orbital kinetic energy density are the local ingredients of a meta-generalized gradient approximation (meta-GGA). We construct a meta-GGA density functional for the exchange-correlation energy that satisfies exact constraints without empirical parameters. The exchange and correlation terms respect {\it two} paradigms: one- or two-electron densities and slowly-varying densities, and so describe both molecules and solids with high accuracy, as shown by extensive numerical tests. This functional completes the third rung of ``Jacob's ladder'' of approximations, above the local spin density and GGA rungs.
As shown by Overhauser and others, the pair-distribution function $g(r)$ of a many-electron system may be found by solving a two-electron scattering problem with an effective screened electron-electron repulsion $V(r)$. We propose a simple physical picture in which this screened repulsion is the ``dressed-dressed'' interaction between two neutral objects, each an electron surrounded by its full-coupling exchange-correlation hole. For the effective interaction between two electrons of antiparallel spin in a high-density uniform electron gas of arbitrary spin polarization, we confirm that this picture is qualitatively correct. In contrast, the ``bare-dressed'' interaction is too repulsive, and does not have the expected symmetry $V_{\ud}(r) = V_{\du}(r)$. The simple original Overhauser model interaction, independent of the relative spin polarization $ζ$, does not capture the $ζ$-dependence of the correlation contribution to $g(r=0)$.
The optimized effective potential (OEP) is the exact Kohn-Sham potential for explicitly orbital-dependent energy functionals, e.g., the exact exchange energy. We give a proof for the OEP equation which does not depend on the chain rule for functional derivatives and directly yields the equation in its simplest form: a certain first-order density shift must vanish. This condition explains why the highest-occupied orbital energies of Hartree-Fock and exact-exchange OEP are so close. More importantly, we show that the exact OEP can be constructed iteratively from the first-order shifts of the Kohn-Sham orbitals, and that these can be calculated easily. The exact exchange potential $v_\mathrm{x}(\re)$ for spherical atoms and three-dimensional sodium clusters is calculated. Its long-range asymptotic behavior is investigated, including the approach of $v_\mathrm{x}(\re)$ to a non-vanishing constant in particular spatial directions. We calculate total and orbital energies and static electric dipole polarizabilities for the sodium clusters employing the exact exchange functional. Exact OEP results are compared to the Krieger-Li-Iafrate (KLI) and local density approximations.
For exchange-correlation functionals that depend explicitly on the Kohn-Sham orbitals, the potential $V_{\mathrm{xc}σ}(\re)$ must be obtained as the solution of the optimized effective potential (OEP) integral equation. This is very demanding and has limited the use of orbital functionals like exact exchange. We demonstrate that the OEP can be obtained iteratively by solving a system of partial differential equations instead of an integral equation. This amounts to calculating the orbital shifts that exactify the Krieger-Li-Iafrate (KLI) approximation. Unoccupied orbitals do not need to be calculated. Accuracy and efficiency of the method are shown for atoms and clusters using the exact exchange energy. Counter-intuitive asymptotic limits of the exact OEP, not accessible from previous constructions, are presented.
The most widely-used density functionals for the exchange-correlation energy are inexact for one-electron systems. Their self-interaction errors can be severe in some applications. The problem is not only to correct the self-interaction error, but to do so in a way that will not violate size-consistency and will not go outside the standard Kohn-Sham density functional theory. The solution via the optimized effective potential (OEP) method will be discussed, first for the Perdew-Zunger self-interaction correction (whose performance for molecules is briefly summarized) and then for the more modern self-interaction corrections based upon unitarily-invariant indicators of iso-orbital regions. For the latter approaches, the OEP construction is greatly simplified. The kinetic-energy-based iso-orbital indicator $τ^W_σ(\re)/τ_σ(\re)$ will be discussed and plotted, along with an alternative exchange-based indicator.
We construct analytic formulas that represent the coupling-constant-averaged pair distribution function $\gxcav(r_s,ζ, k_Fu)$ of a uniform electron gas with density parameter $r_s =(9π/4)^{1/3}/k_F$ and relative spin polarization $ζ$ over the whole range $0<r_s<\infty$ and $-1<ζ<1$, with energetically-unimportant long range ($u\to \infty$) oscillations averaged out. The pair distribution function $g_{xc}$ at the physical coupling constant is then given by differentiation with respect to $r_s$. Our formulas are constructed using {\em only} known theoretical constraints plus the correlation energy $\ec(r_s,ζ)$, and accurately reproduce the $g_{xc}$ of the Quantum Monte Carlo method and of the fluctuation-dissipation theorem with the Richardson-Ashcroft dynamical local-field factor. Our $g_{xc}$ seems to be correct even in the high-density ($r_s\to 0$) and low-density ($r_s \to \infty$) limits. When the spin resolution of $\ec$ into $\uu$, $\dd$, and $\ud$ contributions is known, as it is in the high- and low-density limits, our formulas also yield the spin resolution of $g_{xc}$. We also analyze the kinetic energy of correlation into contributions from density fluctuations of various wavevectors.
We use the two-electron wavefunctions (geminals) and the simple screened Coulomb potential proposed by Overhauser [Can. J. Phys. 73, 683 (1995)] to compute the pair-distribution function for a uniform electron gas. We find excellent agreement with Quantum Monte Carlo simulations in the short-range region, for a wide range of electron densities. We are thus able to estimate the value of the second-order coefficient of the small interelectronic-distance expansion of the pair-distribution function. The results are generalized to the partially polarized gas.