From chemistry-request@ccl.net Wed May 12 07:03:44 2004 Received: from grassmarket.ucs.ed.ac.uk (grassmarket.ucs.ed.ac.uk [129.215.166.64]) by server.ccl.net (8.12.8/8.12.8) with ESMTP id i4CC3gkh023531 for ; Wed, 12 May 2004 07:03:42 -0500 Received: from grassmarket.ucs.ed.ac.uk (localhost [127.0.0.1]) by grassmarket.ucs.ed.ac.uk (8.12.10/8.12.10) with ESMTP id i4CC5v1L019638 for ; Wed, 12 May 2004 13:05:57 +0100 (BST) Received: (from nobody@localhost) by grassmarket.ucs.ed.ac.uk (8.12.10/8.12.9/Submit) id i4CC5uRZ019625 for chemistry.-at-.ccl.net; Wed, 12 May 2004 13:05:56 +0100 (BST) X-Authentication-Warning: grassmarket.ucs.ed.ac.uk: nobody set sender to P.Lozano-Casal.-at-.ed.ac.uk using -f Received: from parsons-res11.chem.ed.ac.uk (parsons-res11.chem.ed.ac.uk [129.215.220.249]) by sms.ed.ac.uk (IMP) with HTTP for ; Wed, 12 May 2004 13:05:56 +0100 Message-ID: <1084363556.40a213249c898.-at-.sms.ed.ac.uk> Date: Wed, 12 May 2004 13:05:56 +0100 From: Patricia Lozano-Casal To: chemistry.-at-.ccl.net Subject: Coulombic Potential Summary MIME-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-15 Content-Transfer-Encoding: 8bit User-Agent: Internet Messaging Program (IMP) 3.2.3 X-Spam-Status: No, hits=0.9 required=7.5 tests=LINES_OF_YELLING autolearn=no version=2.61 X-Spam-Checker-Version: SpamAssassin 2.61 (1.212.2.1-2003-12-09-exp) on servernd.ccl.net Dear all, this are the answers that I got to my problem with the Coulombic Potential. Thank you very much for your help. Patricia SUBJECT: QUESTION ABOUT COULOMBIC POTENTIAL Dear all, I am trying to write a program to calculate the electrostatic contribution of the lattice energy for molecular crystals. I have run plane-wave DFT calculations (using the CASTEP package) to calculate the total lattice energy and using Mulliken charges I hoped to estimate the electrostatic contribution by coding a simple Coulombic potential. But, I get really weird results!! I am overestimating my lattice energy by many orders of magnitude. I know that Mulliken charges are not the ideal point charge to use for this but I had not expected my numbers to be that wrong. Could anybody help or if not, could anybody recommend and existing program to use instead (preferably free)? Many thanks, Patricia Lozano-Casal (p.lozano-casal.-at-.ed.ac.uk) -------------------------------------------------------------------- Calculating electrostatic potentials of crystals can be a very daunting task considering that the sum of Coulombic interactions in an infinite crystal is only conditionally convergent. That is it only converges (when it actually does converge) to a finite number because of the alternating signs of the terms in the series. As a result the summation can have serious problems with numerical error. So, I hope you are using an Ewald method to calculate the electrostatic contribution and that your Muliken charges in your unit cell sum to exactly zero. To the first point, the Ewald sum is THE way to calculate the electrostatic energy of an infinite lattice, but even with it care must be taken to reduce round-off error in the numerical calculation. To the second point, if your unit cell is not neutral, then the strict sum includes the summation of an infinite number of positive interactions, which of course is an infinite value. In a truncated (i.e. any real numerical) Ewald summation, this could just lead to a large finite value such as you are getting. Just a thought. Malcolm ------------------------------------------------------------- Dear Patricia, GULP (General Utility Lattice Code), written by Julian Gale, can calculate very easily the electrostatic contribution to the lattice energy of a crystal. You should specify the charges in the input file, although some methods to estimate them (e.g. by electronegativity equalisation) are also implemented. The code is free for academic users. Best regards, Ricardo. -- Room 85, School of Chemistry, Joseph Black Building, Kings Buildings, West Mains Road, Edinburgh, EH9 3JJ United Kingdom Tel: +441316504804