Original Paper
The electron-gas pair density and its geminal representation II. The long-range asymptotics of the Kimball–Overhauser geminals
P. Ziesche,
P. Ziesche
MPI für Physik komplexer Systeme, Nöthnitzer Str. 38, 01187 Dresden, Germany
Search for more papers by this authorP. Ziesche,
P. Ziesche
MPI für Physik komplexer Systeme, Nöthnitzer Str. 38, 01187 Dresden, Germany
Search for more papers by this authorAbstract
In part I [phys. stat. sol. (b) 241, 3544 (2004)] it has been shown, for the homogeneous electron gas, how the momentum distribution n(k) determines the geminal occupancy μ(k), which appears in the Kimball–Overhauser partial-wave analysis of the pair density g(r). The Fermi surface singularities of n(k) cause singularities in μ(k). These singularities determine the large-r asymptotics of the 1-matrix and of the pair density, respectively. Because the pair-density geminals are scattering states, their asymptotic behavior is characterized by phase shifts. Whereas the normalization sum rule of g(r) leads to a Friedel-like phase-shift sum rule [Phys. Rev. B 67, 233102 (2003)], the plasmon sum rule cannot be satisfied by Kimball–Overhauser geminals. Possibilities to cure this failure are discussed. (© 2005 WILEY-VCH Verlag GmbH & Co. KGaA, Weinheim)
References
- [1] The 1-matrix γ1(1|1′) [“1” stands for position and spin] can be diagonalized in terms of natural orbitals and corresponding natural 1-body occupancies. The cumulant partitioning of the 2-matrix γ1(1|1′, 2|2′) = γ1(1|1′)γ1(2|2′) – γ1(1|2′)γ1(2|1′) – χ2(1|1′, 2|2′) defines the cumulant 2-matrix χ2(1|1′, 2|2′). The first and second term may be referred to as Hartree and Fock term, respectively. γ1(1|1′) and χ2(1|1′, 2|2′) are size-extensively normalized, what means ∫ d1 γ1(1|1) = N, ∫ d1d2 χ2(1|1, 2|2) = cN ∝ N, to be compared with ∫ d1d2 γ2(1|1, 2|2) = N(N – 1) ∝ N2. The quantity c= ∫ d1 [γ(1|1) – ∫ d2 γ(1|2, γ2|1)], measuring the non-idempotency of the 1-marix, is called Löwdin parameter. Both γ2(1|1′, 2|2′) and χ2(1|1′, 2|2′) can be diagonalized. γ(1|1′, 2|2′) yields natural geminals (2-body wave functions) and cor-responding natural 2-body occupancies, χ2(1|1′, 2|2′) yields cumulant geminals and cumulant 2-body occu- pancies. The contraction SR ∫ d2 χ2(1|1′, 2|2) = γ(1|1′) – ∫ d2 γ(1|2′)γ(2|1′) is a quadratic equation for the 1-matrix, supposed χ(1|1′, 2|2) is known. For reduced density matrices in general cf. [2–4], for the 2-matrix and its cumulant partitioning in particular cf. e.g. [5, 6], for inequalities cf. [7–9].
- [2] E. R. Davidson, Reduced Density Matrices in Quantum Chemistry (Academic, New York, 1976).
- [3] A. J. Coleman and V. I. Yukalov, Reduced Density Matrices: Coulson's Challenge (Springer, New York, 2000).
- [4] J. Cioslowski, Many-Electron Densities and Reduced Density Matrices (Kluwer/Plenum, New York, 2000).
- [5] P. Ziesche, Int. J. Quantum Chem. 90, 342 (2002). The last line of the Abstract has to be deleted.
- [6] P. Ziesche and F. Tasnádi, Ann. Physik 13, 124 (2004).
- [7] G. Garrod and J. M. Hannon, Int. J. Quantum Chem. 13, 125 (1978).
- [8] P. Ziesche, J. Tao, M. Seidl, and J. P. Perdew, Int. J. Quantum Chem. 77, 819 (2000).
- [9] E. R. Davidson, Int. J. Quantum Chem. 91, 1 (2003).
- [10] F. Furche, Phys. Rev. A 70, 022514 (2004).
- [11] J. M. Soler, Phys. Rev. B 69, 195101 (2004).
- [12] P. Gori-Giorgi and P. Ziesche, Phys. Rev. B 66, 235116 (2002).
- [13] M. Vogt, R. Zimmermann, and R. J. Needs, Phys. Rev. B 69, 045113 (2004).
- [14] P. Gori-Giorgi, F. Sacchetti, and G. B. Bachelet, Physica A 280, 199 (2000); Phys. Rev. B 61, 7353 (2000); Phys. Rev. B 66, 159901(E) (2002).
- [15] P. Gori-Giorgi and J. P. Perdew, Phys. Rev. B 66, 165118 (2002).
- [16] J. C. Kimball, Phys. Rev. A 7, 1648 (1973). Therein the cusp theorem for antiparallel-spin pairs of the Coulomb hole is derived, whereas A. K. Rajagopal, J. C. Kimball, and M. Banerjee, Phys. Rev. B 18, 2339 (1978) present the curvature theorem for parallel-spin pairs of the Fermi hole.
- [17] A. W. Overhauser, Can. J. Phys. 73, 683 (1995).
- [18] P. Gori-Giorgi and J. P. Perdew, Phys. Rev. B 64, 155102 (2001).
- [19] P. Gori-Giorgi and A. Gonis et al., in: Electron Correlations and Materials Properties 2 (Kluwer/Plenum, New York, 2002), p. 379.
- [20] M. Corona, P. Gori-Giorgi, and J. P. Perdew, Phys. Rev. B 69, 045108 (2004).
- [21] P. Gori-Giorgi and A. Savin, Phys. Rev. A 71, 032513 (2005).
- [22] B. Davoudi, M. Polini, R. Asgari, and M. P. Tosi, Phys. Rev. B 66, 075110 (2002).
- [23] B. Davoudi, M. Polini, R. Asgari, and M. P. Tosi, cond-mat/0206456.
- [24] B. Davoudi, R. Asgari, M. Polini, and M. P. Tosi, Phys. Rev. B 68, 155112 (2003).
- [25] R. Asgari, M. Cardenas, M. Polini, B. Davoudi, and M. P. Tosi, cond-mat/0408293.
- [26] I. Nagy, J. I. Juaristi, R. Díez Muiño, and P. M. Echenique, Phys. Rev. B 67, 073102 (2003).
- [27] I. Nagy, R. Díez Muiño, J. I. Juaristi, and P. M. Echenique, Phys. Rev. B 69, 233105 (2004).
- [28] R. Díez Muiño, I. Nagy, and P. M. Echenique, Phys. Rev. B, in press (2005).
- [29] P. Ziesche, Phys. Rev. B 67, 233102 (2003). Therein the singlet/triplet PDs are erroneously normalized as g±(∞) = 1/2 instead of the correct normalization g±(∞) = 1.
- [30] I. Mertig, E. Mrosan, and P. Ziesche, Multiple Scattering Theory of Point Defects in Metals: Electronic Properties (Teubner, Leipzig, 1987), pp. 44, 46.
- [31] P. Ziesche, phys. stat. sol. (b) 241, 3544 (2004). In the line before Eq. (2.5) read (κ, k). After Eq. (5.1), read (the term ∝z2F holds for k < 1, the next two terms ∝A1 hold for k ≪ ), instead of (the first line …). In Eq. (B.1) read sin (kr) k · n(k) instead of krk · n(k); similar corrections are in Eqs. (C.1), (C.5), (D.1) and Table D.1.
- [32] P. Ziesche and J. Cioslowski, Physica A, in press (2005).
- [33] For rs = 0 it is μ0(k) = 8(1 – 3k/2 + k3/2) Θ(1 –k), cf. Refs. [8], [34–36].
- [34] K. A. Brueckner, Phys. Rev. 96, 508 (1954).
- [35] N. H. March, W. H. Young, and S. Sampanthar, The Many-Body Problem in Quantum Mechanics (Cambridge University Press, Cambridge, 1967), p. 210.
- [36] P. M. W. Gill et al., J. Mol. Struct. (Theochem) 506, 303 (2002).
- [37] P. Ziesche and G. Lehmann (eds. ), Ergebnisse in der Elektronentheorie der Metalle (Springer, Berlin, 1983), p. 31.
- [38] J. C. Kimball, Phys. Rev. 14, 2371 (1976).
- [39] D. Pines and P. Nozières, The Theory of Electron Liquids Vol. I (Benjamin, New York, 1966).
- [40] From the f-SR or Thomas–Reich–Kuhn SR of the dynamic structure factor S(q,ω) follows the plasmon SR S(q → 0) ∼ q2/(2ωpl) of the static structure factor S(q).
- [41] Eqs. (3.3–3.5) agree with Eq. (J6) of Ref. [42], this because of dk sin (kr)[k n(k)]″ ∼ πA1 for r ≫ rc.
- [42] B. Farid, Philos. Mag. B 82, 1413 (2002).
- [43] This divergence compensation resembles to some extent to the well-known error compensation of the exchange and correlation energy in DFT, cf. e.g. R. M. Dreizler and E. K. U. Gross, Density Functional Theory. An Approach to the Quantum Many-Body Problem (Springer, Berlin, 1990), p. 187, (ii).
- [44] N. Iwamoto, Phys. Rev. A 30, 3289 (1984).
- [45] N. Iwamoto, Phys. Rev. A 33, 1940 (1986).
- [46] L. Zecca, P. Gori-Giorgi, S. Moroni, and G. B. Bachelet, Phys. Rev. B 70, 205127 (2004).
- [47] Eq. (2.12) excludes non-oscillatory asymptotics as g(r ≫ rc – 1 ∼ 1/rn, n ≤ 7, but oscillatory terms e.g. as ∼ (cos 2r)/r7 and ∼(sin 2r)/r7 do not conflict with Eq. (2.12). This is because of dr (sin 2r)/r = π/2 and dr (cos 2r)/r = –Ci(2rc). Such Friedel oscillations have their origin in the singular behavior of S(q) at qF = 2, which arise from the singular behavior of n(k) at the Fermi surface k = 1. These oscillations tend to decrease with increasing coupling strength rs. But in addition to this there are other oscillations (with another wave length), which emerge and increase with rs as a precursor of the phase transition from a spin-unpolarized uniform liquid to a ferromagnetic bcc crystal at a critical value rs*. These Wigner-like oscillations with decaying terms ∼cos qwr and ∼sin qwr describe the short-range order of the liquid phase and show up in g(r) through the formation of a first-neighbor shell and further oscillations [23]. They originate from a peak-like structure of S(q) at qw, which is not the above qF, but roughly the reciprocal of the first-neighbor distance in such a Wigner lattice.
