From ragno@serifos.caspur.it Tue Nov 5 03:19:28 1996 Received: from serifos.caspur.it for ragno@serifos.caspur.it by www.ccl.net (8.8.2/950822.1) id DAA28540; Tue, 5 Nov 1996 03:13:36 -0500 (EST) Received: by serifos.caspur.it; id AA04319; Tue, 5 Nov 1996 09:14:01 +0100 Date: Tue, 5 Nov 1996 09:14:01 +0100 (MET) From: Gianluca Sbardella To: Eric VANGREVELINGHE Cc: CHEMISTRY@www.ccl.net Subject: Re: CCL:Sybyl - adding files in database In-Reply-To: <327E2845.1AEE@univ-orleans.fr> Message-Id: Mime-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII On Mon, 4 Nov 1996, Eric VANGREVELINGHE wrote: > Dear Netter > > I failed to contact a sybyl mail list. > Does someone know a good address? > If I'm not wrong, ther must be a Sybyl users group. Try and send a request to join to: sybylreq@quant.chem.rpi.edu. Hope this helps, Gianluca Sbardella *************************************************************** * * * Dr. Gianluca Sbardella E-mail: ragno@kea.caspur.it * * Dip. Studi Farmaceutici ragno@serifos.caspur.it * * Universita' "La Sapienza" Phone: 39-6-49913814 * * P.le A. Moro, 5 Fax: 39-6-491491 * * 00185 Roma * * ITALY * * * * !!The electronic address is going to be changed into!! * * * * -->r.ragno@caspur.it<-- * * * *************************************************************** From ragno@serifos.caspur.it Tue Nov 5 03:19:38 1996 Received: from serifos.caspur.it for ragno@serifos.caspur.it by www.ccl.net (8.8.2/950822.1) id DAA28512; Tue, 5 Nov 1996 03:12:12 -0500 (EST) Received: by serifos.caspur.it; id AA04783; Tue, 5 Nov 1996 09:12:52 +0100 Date: Tue, 5 Nov 1996 09:12:52 +0100 (MET) From: Gianluca Sbardella To: Eric VANGREVELINGHE Cc: CHEMISTRY@www.ccl.net Subject: Re: CCL:Sybyl - adding files in database In-Reply-To: <327E2845.1AEE@univ-orleans.fr> Message-Id: Mime-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII On Mon, 4 Nov 1996, Eric VANGREVELINGHE wrote: > Dear Netter > > I failed to contact a sybyl mail list. > Does someone know a good address? > If I'm not wrong, ther must be a Sybyl users group. Try and send a request to join to: sybylreq@quant.chem.rpi.edu. Hope this helps, Gianluca Sbardella *************************************************************** * * * Dr. Gianluca Sbardella E-mail: ragno@kea.caspur.it * * Dip. Studi Farmaceutici ragno@serifos.caspur.it * * Universita' "La Sapienza" Phone: 39-6-49913814 * * P.le A. Moro, 5 Fax: 39-6-491491 * * 00185 Roma * * ITALY * * * * !!The electronic address is going to be changed into!! * * * * -->r.ragno@caspur.it<-- * * * *************************************************************** From cburkhart@goodyear.com Tue Nov 5 13:11:26 1996 Received: from goodyear.com for cburkhart@goodyear.com by www.ccl.net (8.8.2/950822.1) id MAA03205; Tue, 5 Nov 1996 12:57:53 -0500 (EST) Received: from rdsrv2 (rdsrv2.goodyear.com) by goodyear.com (4.1/SMI-4.1) id AA12876; Tue, 5 Nov 96 12:57:48 EST Received: from rds2721 by rdsrv2 (4.1/SMI-4.1) id AA00545; Tue, 5 Nov 96 12:57:48 EST Sender: burkhart@goodyear.com Message-Id: <327F8004.41C6@goodyear.com> Date: Tue, 05 Nov 1996 12:57:24 -0500 From: Craig Burkhart Organization: Goodyear Research X-Mailer: Mozilla 2.02S (X11; I; IRIX64 6.2 IP28) Mime-Version: 1.0 To: Computational Chemistry List Subject: SUMMARY: Hydrocarbon Forcefields Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Dear Netters, After much clacking of virtual tongues and over 25 responses from many parts of cyberspace, I am finally delivering on the promised summary of responses concerning liquid state simulations of alkanes/alkenes. Before proceding, much thanks is in order to the following people-- On forcefields: Bill Jorgensen (Yale), Ilja Siepmann (UMinn), David Maxwell (DuPont-Merck), Erin Duffy (Yale) and Andreas Bick (MSI) On methodology: Mike Kotelyanskii (UDelaware), Ilja Siepmann, Eric Wallis (Phillips Petroleum) and Carol Parrish (Columbia) ...And now on to the summary :) My Queries: 1) Do you have experience with liquid density NpT MD simulations of branched hydrocarbons? 2) If so, can you point me in the direction of a forcefield which you have found appropriate for the calculation of liquid densities and delta-Hvap? 1) Concerning experience/methods: Mike Kotelyanskii and Ilja Siepmann reminded me of an old rule of thumb concerning cutoffs--that you should put them at a place where g(r) is structureless; that is, where g(r) is close to unity. Good point. Most of us learn this along the way and then promptly forget about it. This brought up an important back-and-forth about the nature of how one deals with long-ranged nonbonded interactions (van der Waals and coulombic). Most of my simulations were underestimating the densities of my alkanes by 5-7% or so. I was basically using nonbonded cutoffs with smooth splines in about the 8-8.5 Angstrom range. This turns out to be way too short, and if you look at the g(r) you realize that there is significant liquid structure in this range. I redid a few of my alkane simulations using 12-13 Angstrom cutoffs. The net result is that I gained about 2% in density so that my errors were now in the 4% range using Jorgensen's 1994 OPLS-AA. For this I thank Mike Kotelyanskii for being persistent enough in prodding me to look at the effect of cutoffs. We now turn our attention to the question of "tail corrections". Tail corrections are done to handle the effect of truncating the nonbonded potential. That is, it corrects the tail that exists past the cutoff. For strictly van der Waals interactions, this correction is always attractive (this may not be true with electrostatics). The standard method is discussed in detail in Allen & Tildesley's book "Computer Simulation of Liquids". Note however that this correction is most applicable in hydrocarbon liquids for NVT ensembles. This is no problem for software using Monte Carlo (MC) dynamics, which recalculates the "tail term" only when a volume move, or some suitable interval, occurs. My thanks to Eric Wallis in discussing his experience with MC tail corrections. It is however a problem in standard Newtonian (MD) dynamics using the NpT (Gibbs) ensemble, where both molecules and system volume are constantly changing. In this instance, using the Ewald summation may be a more viable approach. A useful implementation of this method for both coulombic and Lennard- Jones potentials is given by Karasawa & Goddard, J. Phys. Chem. 93, 7320(1989). It is basically a generalization of Donald Williams' "accelerated convergence" technique (see his series of Acta Cryst. articles in the late 60's-early 70's) which was based on Ewald's seminal work. This newest rendition of the Ewald sum has its advantages over systematically increasing your nonbonded cutoff. It computes an estimate of the full long-ranged interaction which is dependent only on your convergence criteria. I've have used a 0.01 kcal/mole/atom cutoff with good success. The bottom line is that the use of the Ewald method reduces my density errors to about 3% using the 1994 OPLS-AA. 2) Concerning forcefields: There are 3 groups currently working on improving liquid alkane/alkene forcefields, each with a slightly different emphasis. Bill Jorgensen's group (which includes past members David Maxwell and Erin Duffy) has published a great deal on their Optimized Potentials for Liquid Simulations (OPLS) approach. Professor Jorgensen has informed me that a new OPLS-AA forcefield will be published in an upcoming J. Amer. Chem. Soc. November issue. He maintains that OPLS-AA(1996) will give densities generally within 2%. With his help, I'm currently running a test on a branched alkane liquid cell. My thanks to Bill Jorgensen for helping me with the details of implementing OPLS-AA. The second group involves Ilja Siepmann's recent collaborations. He has published work with B Smit, S Balasubramanian, C Mundy and ML Klein on united-atom approaches to hydrocarbon liquid critical parameters (critical density, temperature, ...) and solute partitioning. His approach may be appealing for polymer based problems as well. His most recently published work [S Balasubramanian, ML Klein and JI Siepmann, J. Phys. Chem. 1996, 100, 11960] use an updated forcefield. He informs me that they, too, are putting the finishing touches on this forcefield, which is due to be published next year; he has upcoming articles submitted to Mol. Phys. and Fluid Phase Eq. which may be of use to this audience. The third group involves Andreas Bick and coworkers at MSI. They are currently revamping the Biosym PCFF parameterization for liquids. However, it is only available at this time to their consortium members, and may be published sometime in the next year (we hope). Summary: The Jorgensen OPLS-AA forcefield, coupled with the appropriate long-range corrections may be the most comprehensive method of currently dealing with liquid hydrocarbons of small-to-intermediate size. However, Siepmann's forcefield may be also useful depending on the properties you wish to simulate, and may be more appropriate for long-chain systems. As always, I thank the members of the CCL for their help. Even if Jan doesn't want to burn bandwidth on debating the CCL's survival, I must add my $0.02 worth :). IMHO it is a unique GLOBAL resource and that we should do what is necessary to guarantee its survival. Time to put your money where your mouth is... Sincerely yours, Craig. -- ---------------------------------------------------------------------------- ! ONLY 102 DAYS 'TIL SPRING TRAINING ! ONLY 102 DAYS 'TIL SPRING TRAINING ! ----------------------------------+----------------------------------------- Papernet: Craig W. Burkhart |"I believe in the Church of Baseball. I've Goodyear Research |tried all the major religions and most of 142 Goodyear Blvd. |the minor ones...but there's too much Akron, OH 44305 |guilt in religion--you see there's no Mouthnet: 330.796.3163 |guilt in baseball...The only church that Faxnet: .3304 |really feeds the soul, day in and day out, Internet: cburkhart@goodyear.com |is the Church of Baseball". |--Annie Savoy, "Bull Durham" ----------------------------------+----------------------------------------- !!! WAIT 'TIL NEXT YEAR !!! WAIT 'TIL NEXT YEAR !!! WAIT 'TIL NEXT YEAR !!! ----------------------------------------------------------------------------