Return-Path: Received: from murder ([unix socket]) by merkur (Cyrus v2.2.13-Debian-2.2.13-10) with LMTPA; Thu, 24 Sep 2009 09:41:34 +0200 X-Sieve: CMU Sieve 2.2 Received: from localhost (localhost [127.0.0.1]) by merkur.ins.uni-bonn.de (Postfix) with ESMTP id 73CE640004B23; Thu, 24 Sep 2009 09:41:34 +0200 (CEST) X-Virus-Scanned: Debian amavisd-new at merkur.ins.uni-bonn.de Received: from merkur.ins.uni-bonn.de ([127.0.0.1]) by localhost (merkur.ins.uni-bonn.de [127.0.0.1]) (amavisd-new, port 10024) with ESMTP id FcrZPULZoC0q; Thu, 24 Sep 2009 09:41:25 +0200 (CEST) Received: from [131.220.133.122] (unknown [131.220.133.122]) by merkur.ins.uni-bonn.de (Postfix) with ESMTP id 57DFC400049F2; Thu, 24 Sep 2009 09:41:22 +0200 (CEST) Message-ID: <4ABB22A4.9050001@ins.uni-bonn.de> Date: Thu, 24 Sep 2009 09:41:24 +0200 From: Astrid Link Reply-To: astrid.link@ins.uni-bonn.de User-Agent: Thunderbird 1.5.0.2 (Windows/20060308) MIME-Version: 1.0 To: Astrid Link Subject: Sonderforschungsbereich 611 - SFB-Seminar am 1. Oktober 2009 Content-Type: multipart/alternative; boundary="------------070403080809070903090709" --------------070403080809070903090709 Content-Type: text/plain; charset=ISO-8859-15; format=flowed Content-Transfer-Encoding: 7bit Einladung zu einem Vortrag im SFB-Seminar Donnerstag, 1. Oktober 2009 Wegelerstr. 10, Kleiner Hoersaal 15.30 Uhr Prof. Dr. N. Sukumar, University of California, Davis, USA, spricht zum Thema Linear Scaling Finite Element Solution of the All-Electron Coulomb Problem in Periodic Solids Abstract: The computation of the electrostatic potential and total energy of crystalline solids has been an ongoing problem in solid state physics. In quantum mechanical calculations, the electrostatic potential is constructed as a sum of nuclear and electronic terms. In an infinite crystal, each of these terms diverges and the sum is only conditionally convergent due to the long-range 1/r nature of the Coulomb interaction. A common approach to such "all electron" Coulomb problems is to smear the nuclear point charges (distributed nucleus approximation) and solve the resulting smoothed but strongly inhomogeneous problem in a basis which can be concentrated in the vicinity of the nuclei. Accurate solutions require strongly localized nuclear charges and so highly refined basis sets in the vicinity of the nuclei. The most common approach in high-accuracy calculations employs a combined spherical-harmonic and Fourier representation and so has O (N log N) complexity. In this talk, I will present a systematically improvable, O (N) (linear scaling) formulation for the all-electron Coulomb problem in periodic solids that avoids the need for distributed nucleus approximations. Ewald sums, and operations in reciprocal space. Linear scaling is achieved by introducing smooth, strictly local neutralizing densities to render interactions strictly local and solving the remaining neutral Poisson problem for the electrons in real space. The resulting formulation is amenable to solution using basis sets that are in H1. In the numerical computations, we employ finite element and enriched finite element methods to demonstrate the accuracy and convergence of the approach by direct comparison to standard Ewald sums and application to all-electron diamond. Tee: 16.30 h -- Astrid Link Sonderforschungsbereich 611 Universitaet Bonn Poppelsdorfer Allee 82 53115 Bonn, Germany Tel.: +49 228 73-4882 (-3858) Fax: +49 228 73-7864 astrid.link@ins.uni-bonn.de www.sfb611.iam.uni-bonn.de --------------070403080809070903090709 Content-Type: text/html; charset="iso-8859-15" Content-Transfer-Encoding: quoted-printable

Einladung zu einem Vortrag im SFB-Seminar

Donnerstag, 1. Oktober 2009

Wegelerstr. 10, Kleiner Hoersaal

15.30=A0 Uhr

=A0Prof. Dr. N. Sukumar, University of California, Davis, USA,
spricht zum Thema

Linear Scaling Finite Element Solution of the All-Electron Coulomb Problem in Periodic Solids

=A0
Abstract:
The computation of the electrostatic potential and total energy of crystalline solids has been an ongoing problem in solid state physics. In quantum mechanical calculations, the electrostatic potential is constructed as a sum of nuclear and electronic terms. In an infinite crystal, each of these terms diverges and the sum is only conditionally convergent due to the long-range 1/r nature of the Coulomb interaction. A common approach to such ”all electron“ C= oulomb problems is to smear the nuclear point charges (distributed nucleus approximation) and solve the resulting smoothed but strongly inhomogeneous problem in a basis which can be concentrated in the vicinity of the nuclei. Accurate solutions require strongly localized nuclear charges and so highly refined basis sets in the vicinity of the nuclei. The most common approach in high-accuracy calculations employs a combined spherical-harmonic and Fourier representation and so has O (N log N) complexity.
In this talk, I will present a systematically improvable, O (N) (linear scaling) formulation for the all-electron Coulomb problem in periodic solids that avoids the need for distributed nucleus approximations. Ewald sums, and operations in reciprocal space. Linear scaling is achieved by introducing smooth, strictly local neutralizing densities to render interactions strictly local and solving the remaining neutral Poisson problem for the electrons in real space. The resulting formulation is amenable to solution using basis sets that are in H1. In the numerical computations, we employ finite element and enriched finite element methods to demonstrate the accuracy and convergence of the approach by direct comparison to standard Ewald sums and application to all-electron diamond.

Tee: 16.30 h

--=20

Astrid Link
Sonderforschungsbereich 611
Universitaet Bonn
Poppelsdorfer Allee 82
53115 Bonn, Germany
Tel.: +49 228 73-4882 (-3858)
Fax:  +49 228 73-7864
astrid.link@ins.uni-bonn.de=20
www.sfb611.iam.uni-bonn.de

--------------070403080809070903090709--