From chemistry-request@www.ccl.net Wed Jan 6 01:47:50 1999 Received: from exstudent3.levels.unisa.edu.au (EXSTUDENT3.Levels.UniSA.Edu.Au [130.220.22.184]) by www.ccl.net (8.8.3/8.8.6/OSC/CCL 1.0) with ESMTP id BAA15684 Wed, 6 Jan 1999 01:47:48 -0500 (EST) Received: by EXSTUDENT3.Levels.UniSA.Edu.Au with Internet Mail Service (5.5.2232.9) id ; Wed, 6 Jan 1999 17:17:44 +1030 Message-ID: From: "Soar, Terri Jacqueline - SOATJ001" To: chemistry@www.ccl.net Subject: error/accuracy determinations for S.E. application Date: Wed, 6 Jan 1999 17:17:35 +1030 MIME-Version: 1.0 X-Mailer: Internet Mail Service (5.5.2232.9) Content-Type: text/plain; charset="iso-8859-1" Hi CCLers. I'm after some ideas, pointers or even just people's thoughts as to the topic of calculating the accuracy of calculations. This is the situation...For my PhD in physical and inorganic chemistry about one quarter of my work has involved applying computational chemistry. The computational chemistry work has focused on a range of aluminium-oxyhydroxide species (monomer, dimer, trimer, tetramer and some polymers) that may exist in aging caustic aluminate solutions where crystalline Al(OH)3, i.e. gibbsite, finally precipitates out. We are using MOPAC93 with AM1 (i.e. semi-empirical quantum mechanical methodology) and also with and without COSMO (to include solvent effects) to perform geometry optimisations and SADDLE calculations on these species where our main interest is the stability and also the relative energetics of the species (wrt to heats of formation and reaction and also Gibbs free energy of formation and reaction) to indicate what species are more likely to exist in these solutions and participate in reactions possibly leading to gibbsite formation. So, how would I go about estimating errors and determining accuracy for this work (keeping in mind that many of the solution species we've modelled cannot be measured experimentally (i.e. heats of formation))? As I've said above, I'm after any answers to this question in addition to people's thoughts as to what may be expected in the way of accuracy determinations for this type of work. Thanks in advance and, of course, a summary will be produced. Terri ------------------------------------------------------------------------------- -- Ms. Terri Soar, PhD student, Ian Wark Research Institute, tel: +61 8 8302 3706 University of South Australia, fax: +61 8 8302 3683 The Levels Campus, Mawson Lakes, S.A. 5095, AUSTRALIA email: soatj001@students.unisa.edu.au From chemistry-request@www.ccl.net Wed Jan 6 05:55:54 1999 Received: from alpha.luc.ac.be (alpha.luc.ac.be [193.190.2.30]) by www.ccl.net (8.8.3/8.8.6/OSC/CCL 1.0) with ESMTP id FAA16550 Wed, 6 Jan 1999 05:55:53 -0500 (EST) Received: from alpha (alpha [193.190.2.30]) by alpha.luc.ac.be (8.9.1/8.9.1) with ESMTP id LAA16141 for ; Wed, 6 Jan 1999 11:55:54 +0100 (MET) Date: Wed, 6 Jan 1999 11:55:54 +0100 (MET) From: Sergiusz Kwasniewski To: CCL Subject: Summary CCL: Vibrational spectra. Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Thank you for your responses. Or should I say "dziekuje bardzo". As it is custom to make a summary of all the responses, I will serve you the following emails: S. Kwasniewski ++++++++++++++++ (original email) ++++++++++++++++ Does anyone know how to calculate excitation spectra ? And what kind of program would you prefer ? We've been using MOPAC here (semi-empirical techniques) to run MECI calculations. In the end we want to be able to calculate UV/Vis absorption spectra. We were also thinking about buying Hyperchem (if there are versions for workstations available that is). I think I read somewhere that the calculations of certain spectra is included in this. Does anyone has another suggestion for programs/methods to use, or some pros/contras on MOPAC and Hyperchem ? Thank you. S. Kwasniewski +++++++++++++++ (first respons) +++++++++++++++ Sergiusz, ZINDO and Argus have been parameterized explicitly to do UV/visible spectra. ZINDO may be available from the Quantum Theory Project at the University of Florida, the home of Michael Zerner, its author. It is available commercially from Molecular Simulations (www.msi.com) Argus is free and is available at http://www.seanet.com/~mthompson/ArgusLab/index.htm Best wishes, Paul Soper DuPont Central Research & Development Wilmington, DE 19880-0328 USA DuPont Central Research & Development Wilmington, DE 19880-0328 USA ++++++++++++++++ (second respons) ++++++++++++++++ Dear Sergiusz, In my experience calculations of electronic excitacitions using AM1/PM3 produce results that are in disagreement with experimental results. The positions of the transitions are OK, but the orbitals involved in that transitions are assigned in a very strange manner. For a reference look: Spectrochimica Acta A, 54, 339-348 (1998). In this work we conclude that the best method for calculations of electronic spectra of organic compounds, with small computational resources, is INDO1/S. A version of this method is available free in the net, the ArgusLab program at: http://www.seanet.com/~mthompson/ArgusLab/index.htm Best regards, Sergio (segalemb@usp.br) +++++++++++++++ (third respons) +++++++++++++++ I would advise against mopac. Traditional semiempirical methods are not advisable for excited state calculations. I would advise using ZINDO which is a CIS method derived from MOPAC. Deepak Singh Email: desingh@syr.edu Graduate Fellow URL: http://web.syr.edu/~desingh Department of Chemistry and Biochemistry, Tel: (315) 443 1739 -- work Syracuse University (315) 472 9659 -- home 1-014 Scitech Syracuse NY 13244 Fax: (315) 443 4070 "Violence is the last resort of the incompetent". -- Salvor Hardin ++++++++++++++++ (fourth respons) ++++++++++++++++ My advice is to use HyperChem. The version 4.5 for SGI (and probably for other machines) can be downloaded for testing and a trial licence is valid for a month. Sam skorzystalem z tej mozliwosci. Uzyskalem dosc dobra zgodnosc obliczonych i wyznaczonych spektrow. Ale optymalizacje struktur prosze zrobic programem z opcja ab initio. Opcja ZINDO nie nadaje sie do tego. If you understand the above, Good luck! S.Kucharski ********************************************************************* * Prof. Dr. Stanislaw Kucharski * * Institute of Organic and Polymer Technology * * Wroclaw Technical University * * ul. Wybrzeze Wyspianskiego 27 * * 50-370 Wroclaw, Poland * ********************************************************************* * E-mail: kucharski@itots.ch.pwr.wroc.pl * ********************************************************************* * Tel. +48 71 34202862; +48 71 34202426 * * Fax: +48 71 34203678 * ********************************************************************* +++++++++++++++ (fifth respons) +++++++++++++++ Hi, ZINDO is a semi-empirical package specially designed to calculate spectra. This program is from Dr. Zerner. Yongxing Liu ****************************** Sergiusz P. Kwasniewski Limburgs Universitair Centrum Universitaire Campus, Gebouw D 3590 Diepenbeek Belgium Tel. (+32) (0)11/26.83.15 Fax. (+32) (0)11/26.83.01 e-mail: skwasnie@luc.ac.be ****************************** From chemistry-request@www.ccl.net Wed Jan 6 08:16:06 1999 Received: from nitrogen.cem.uct.ac.za (nitrogen.cem.uct.ac.za [137.158.165.214]) by www.ccl.net (8.8.3/8.8.6/OSC/CCL 1.0) with SMTP id IAA17002 Wed, 6 Jan 1999 08:15:46 -0500 (EST) Received: from nitrogen (localhost [127.0.0.1]) by nitrogen.cem.uct.ac.za (950413.SGI.8.6.12/950213.SGI.AUTOCF) via SMTP id NAA19935 for ; Wed, 6 Jan 1999 13:18:49 -0800 Sender: achim@psipsy.uct.ac.za Message-ID: <3693D339.167E@psipsy.uct.ac.za> Date: Wed, 06 Jan 1999 13:18:49 -0800 From: Achim Lienke X-Mailer: Mozilla 3.01SGoldC-SGI (X11; I; IRIX 6.3 IP32) MIME-Version: 1.0 To: chemistry@www.ccl.net Subject: Ab initio on Ir(IV), Os(IV) and Ru(III, IV) Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit I am calculating a series of chloro and aqua-chloro complexes of Platin-Group-Metals (PGMs) (e.g. PtCl6-2, RhCl6-3, IrCl6-2) with ab initio methods in order to obtain geometries, IR frequencies and point charges. The results for all closed shell systems are good and reflect experimental data and trends, but for the open-shell systems like IrCl6-2 and OsCl6-2 this is probably not the case. There is a slight Jahn-Teller distorsion (which is expected), but the avaraged bond lengths are smaller than expected and - more important - the point charges on the axial and equatorial chlorines differ substantially and are larger than one would expect from experimental data (solvation). My question: Can it be that those results are just artefacts because some important effects and couplings are neglected? I am using Gaussian 94, Methods are SVWN (produces good results), B3LYP and MP2 (to check, methods are in for this application inferior), basisset is LANL2DZ for the metal and 6-31G for Cl, H and O Achim Lienke Department of Chemistry University of Cape Town achim@psipsy.uct.ac.za From chemistry-request@www.ccl.net Wed Jan 6 11:12:39 1999 Received: from hartree.chem.kuleuven.ac.be (hartree.quantchem.kuleuven.ac.be [134.58.49.1]) by www.ccl.net (8.8.3/8.8.6/OSC/CCL 1.0) with ESMTP id LAA18130 Wed, 6 Jan 1999 11:12:37 -0500 (EST) Received: from localhost (steven@localhost) by hartree.chem.kuleuven.ac.be (8.9.0/8.9.0) with ESMTP id RAA29524 for ; Wed, 6 Jan 1999 17:12:05 +0100 Date: Wed, 6 Jan 1999 17:12:05 +0100 (NFT) From: Steven Creve To: Computational Chemistry List Subject: CCL:QSAR:New question Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Hi all, Next, I'm looking for some good review papers or books about the current status of QSAR. Thanks, steven -------------------------------------------------------------------------- Steven Creve steven.creve@chem.kuleuven.ac.be Labo Quantumchemie Celestijnenlaan 200F 3001-HEVERLEE tel: (32) (16) 32 73 93 BELGIUM fax: (32) (16) 32 79 92 From chemistry-request@www.ccl.net Wed Jan 6 11:26:19 1999 Received: from mailhost.lanl.gov (mailhost.lanl.gov [128.165.3.12]) by www.ccl.net (8.8.3/8.8.6/OSC/CCL 1.0) with ESMTP id LAA18323 Wed, 6 Jan 1999 11:26:18 -0500 (EST) Received: from [128.165.22.209] (machgs.lanl.gov [128.165.22.209]) by mailhost.lanl.gov (8.9.1a/8.9.1/(cic-5, 7/9/98)) with SMTP id JAA04566 for ; Wed, 6 Jan 1999 09:26:15 -0700 X-Sender: schrecke@t12mail.lanl.gov Message-Id: Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Date: Wed, 6 Jan 1999 09:29:15 -0600 To: CHEMISTRY@www.ccl.net (CCL) From: schrecke@t12.lanl.gov (Georg Schreckenbach) Subject: G94/98: How to skip SCF completely? Dear CCL, here is a GAUSSIAN related question: I would like to calculate some properties (like forces or NMR) based solely on the "input guess". How is this best done? (Or is it possible at all in GAUSSIAN?) More specifically, I imagine to have some input MOs using the "Guess=Read" command. I would specify a certain DFT method, say, and then go on to calculate the desired properties. What I have tried so far is the following. A) specify a "non-standard route" that resembles a standard route but contains no 500 links. Didn't work because the subsequent links requested some files that had not been written. Perhaps one can do this in a more clever way? B) Specify "SCF(MaxCycle=1)" as well as the appropriate flags to avoid the error termintation due to incomletely converged SCF. These flags are IOp(8/7=1,5/13=1). This procedure comes close to what I want, in particular if I use SCF=QC. However, the one SCF cycle is still too much because it does modify the input MOs as is clearly visible from the output. Hence, I would want to avoid it! Any comments or suggestions? -- I will summarize as usual. Thank you very much! Georg -- Dr. Georg Schreckenbach Tel: (USA)-505-667 7605 Theoretical Chemistry T-12 FAX: (USA)-505-665 3909 M.S. B268, Los Alamos National E-mail: schrecke@t12.lanl.gov Laboratory, Los Alamos, New Mexico, 87545, USA Internet: http://www.t12.lanl.gov/~schrecke/ -- ============================================================== Dr. Georg Schreckenbach Tel: (USA)-505-667 7605 Theoretical Chemistry T-12 FAX: (USA)-505-665 3909 M.S. B268, Los Alamos National E-mail: schrecke@t12.lanl.gov Laboratory, Los Alamos, New Mexico, 87545, USA Internet: http://www.t12.lanl.gov/~schrecke/ ============================================================== From chemistry-request@www.ccl.net Wed Jan 6 11:58:01 1999 Received: from naomi.eerie.fr ([146.19.248.22]) by www.ccl.net (8.8.3/8.8.6/OSC/CCL 1.0) with ESMTP id LAA18576 Wed, 6 Jan 1999 11:58:00 -0500 (EST) Received: from syntem.eerie.fr (localhost [127.0.0.1]) by naomi.eerie.fr (980427.SGI.8.8.8/970903.SGI.AUTOCF) via ESMTP id RAA14053 for ; Wed, 6 Jan 1999 17:01:21 +0100 (MET) Sender: domi@syntem.eerie.fr Message-ID: <369388D1.45F2AFEF@syntem.eerie.fr> Date: Wed, 06 Jan 1999 17:01:21 +0100 From: Dominique Gorse Reply-To: dgorse@syntem.eerie.fr Organization: Synt:em X-Mailer: Mozilla 4.05C-SGI [en] (X11; I; IRIX 6.5 IP32) MIME-Version: 1.0 To: chemistry@www.ccl.net Subject: Algorithms for variable orthogonalization and PLS Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit Dear CCLers, I am working on variable orthogonalization to be used in QSAR Analysis. I found at http://lib.stat.cmu.edu/apstat/ the Gram-Schmidt orthogonalization algorithm. I would like to know more about this technique but I can't find the paper "Applied Statistics (J. R. Statist. Soc. C) 1971 vol20, N3". Can someone provide me with some information? I am also looking for a detailed algorithm of Partial Least Squares. Of course a source code will be welcome! Best Regards **************************************************************** Dominique Gorse Synt:em Parc Scientifique G.Besse Computational Drug Discovery 30000 Nimes email: dgorse@syntem.eerie.fr France Tel: +33 (0)4 66 04 86 65 Fax: +33 (0)4 66 04 86 67 **************************************************************** Discover New Drugs, Discover Synt:em http://www.syntem.com ****************************************************************