From chemistry-request@www.ccl.net Tue Oct 6 03:33:25 1998 Received: from anduril.Austria.EU.net (anduril.Austria.EU.net [193.154.160.104]) by www.ccl.net (8.8.3/8.8.6/OSC/CCL 1.0) with ESMTP id DAA03900 Tue, 6 Oct 1998 03:33:24 -0400 (EDT) Received: from apollon.bender.co.at (apollon.bender.co.at [193.154.107.33]) by anduril.Austria.EU.net (8.9.1/8.9.1) with SMTP id JAA18758 for ; Tue, 6 Oct 1998 09:33:19 +0200 (MET DST) Received: from vieexch1.bender.co.at by apollon.bender.co.at via smtpd (for anduril.Austria.EU.net [193.154.160.104]) with SMTP; 6 Oct 1998 07:33:16 UT Received: by vieexch1.vie.at.bic with Internet Mail Service (5.5.1960.3) id ; Tue, 6 Oct 1998 09:26:41 +0200 Message-ID: <4195A3C6D39BD11189C30020AFFC17588F3C17@vieexch1.vie.at.bic> From: "LOEFFLER,DR,GERALD FEM BENAT" To: "'ccl'" Subject: Graph Layout in 2D Date: Tue, 6 Oct 1998 09:26:32 +0200 Hi! I would be very thankful if you could take a second to enlighten me on this subject: Question: ========= What algorithms and implementations of these algorithms (in Java?) are available to "intelligently" lay out a general graph in 2D? The resulting layout should contain a minimum number of overlaps of nodes and edges. Background: =========== An algorithms of mine produces a general graph in the sense that the ouput is a set of nodes {N_i} and a set of edges that connect pairs of nodes {(N_j,N_k)}. Every node is connected by at least one edge, i.e. there are no unconnected nodes. Furthermore, there are no isolated "islands" that have no connection to the rest of the graph, i.e. there is a path from every node to every other node in the set of nodes. I would like to automatically lay out a graph like this in 2D, such that it looks beautiful: 1) Nodes should not overlap, and 2) edges should cross as seldom as possible. A description of such a layout algorithm would be fine, but an implementation in Java that I can adapt to digest my input (set of nodes and set of edges) would be heaven on earth (-: My stupid solution: =================== Having an MD background, I thought about treating the nodes as particles that repell each other and that force the edges originating at them into a regular polygon (e.g. 4 edges on a node would be forced to an angle of 90 degrees to each other). Then I could start with maybe all nodes on a circle and solve Newton's equations in 2D until the layout "looks good" (the "energy" doesn't change much)... thank you very much in advance!!, gerald |--------------------------------------------------------------------| Gerald Loeffler - Bioinformatics Scientist Boehringer Ingelheim R&D Vienna Email: Gerald.Loeffler@vienna.at Phone: +43 676 3289588 (and +43 1 80105 634) Fax: +43 1 80105 683 Smail: Bender+Co, Dr. Boehringer-Gasse 5-11, A-1121 Vienna, Austria From chemistry-request@www.ccl.net Tue Oct 6 04:48:08 1998 Received: from pobox.csc.fi (pobox.csc.fi [128.214.46.62]) by www.ccl.net (8.8.3/8.8.6/OSC/CCL 1.0) with ESMTP id EAA04107 Tue, 6 Oct 1998 04:48:05 -0400 (EDT) Received: from laaksonen.csc.fi (laaksonen.csc.fi [193.166.1.75]) by pobox.csc.fi (8.9.0.Beta5/8.9.0.Beta5/CSC/Rtrs-1.22) with ESMTP id LAA07684 for ; Tue, 6 Oct 1998 11:48:05 +0300 (EET DST) Date: Tue, 6 Oct 1998 11:47:43 +0200 (GFT Standard Time) From: Leif Laaksonen To: CHEMISTRY@www.ccl.net Subject: Version 1.2 of gOpenMol for Windows available Message-ID: X-X-Sender: laaksone@laaksonen.csc.fi Hi, Scarecrow Computing anounces version 1.2 of gOpenMol for the Windows platform. The Unix versions will be available when I have sorted out some problems with the event queue handling and the GLUT library. Tons of bugs have been corrected and some new features have been added. Please have a look at: http://laaksonen.csc.fi/gopenmol/help/version12.html The program is available through the home page at: http://laaksonen.csc.fi/gopenmol/ This is still a binary distribution. When I have the Unix version running I will start distributing also the source code. The gOpenMol Graphics Gallery is available at: http://laaksonen.csc.fi/gopenmol/Gallery/ If you have graphics made by gOpenMol on your Web pages please let me know and I will make a pointer to your pages. Thank you all who have contributed with bug reports and suggestions. Cheerio, -leif laaksonen ------------------------------------------------------------------- Center for Scientific Computing | Phone: 358 9 4572378 P.O. Box 405 | Mobile: 358 400425203 FIN-02101 Espoo | Telefax: 358 9 4572302 FINLAND | Mail: Leif.Laaksonen@csc.fi -------- URL: http://laaksonen.csc.fi/leif.laaksonen.html --------- From chemistry-request@www.ccl.net Tue Oct 6 09:39:59 1998 Received: from main.amu.edu.pl (root@main.amu.edu.pl [150.254.65.2]) by www.ccl.net (8.8.3/8.8.6/OSC/CCL 1.0) with ESMTP id JAA05394 Tue, 6 Oct 1998 09:38:43 -0400 (EDT) Received: from main.amu.edu.pl (jolanala@main.amu.edu.pl [150.254.65.2]) by main.amu.edu.pl (8.8.8/8.8.8) with SMTP id PAA10301 for ; Tue, 6 Oct 1998 15:38:15 +0200 (MET DST) Date: Tue, 6 Oct 1998 15:38:15 +0200 (MET DST) From: Jolanta Latosinska To: "Comp. Chem. List" Subject: susceptibility Message-ID: Dear Netters, Does anybody know or have any suggestions how can I calculate susceptibility by Gaussian 94? Thanks in advance, Jola From chemistry-request@www.ccl.net Tue Oct 6 13:38:32 1998 Received: from ucs.orst.edu (pughj@UCS.ORST.EDU [128.193.4.5]) by www.ccl.net (8.8.3/8.8.6/OSC/CCL 1.0) with ESMTP id NAA09564 Tue, 6 Oct 1998 13:38:31 -0400 (EDT) Received: from localhost (pughj@localhost) by ucs.orst.edu (8.9.1/8.9.1) with SMTP id KAA24701 for ; Tue, 6 Oct 1998 10:38:33 -0700 (PDT) Date: Tue, 6 Oct 1998 10:38:33 -0700 (PDT) From: James Pugh Reply-To: James Pugh To: chemistry@www.ccl.net Subject: Carbene calculations summary Message-ID: Dear CCL, Thank you for all the responses. It will certainly help our progress here. Enclosed is the original message followed by all replies. ---------------------------------------------------------------------- > Dear CCL, > > Our lab is heavily involved in the computational study of carbenes (both > singlet and triplet) and related diradicals. Currently, we are using the > b3lyp/6-311+g(3df,2p)//b3lyp/6-31g(d,p) method. Our calculations show > this to give nice agreement to experimental data for the singlet-triplet > gap of methylene carbene. In searching the literature there are several > examples of other groups also using DFT methods for carbenes. But there > are probably just as many using a multiconfigurational self consistent > field method (MCSCF or CASSCF in Gaussian). My question is a follows: > > Does anyone else have experience with similiar calculations and more > importantly - which method would be the preferred or more correct > approach. > > Any assistance would be greatly appreciated. I will summarize all > responses. > > Jim Pugh > Freeman Research Group > Department of Chemistry > Oregon State University > e-mail: pughj@ucs.orst.edu ------------------------------------------------------------------ From: "Christopher J. Cramer" To: James Pugh Subject: Re: CCL:carbene calculations We've done rather a lot in this area. Rather than recapitulate the whole ouevre, I would just point you to our latest comprehensive work in this area. See Worthington, S. E.; Cramer, C. J. "Density Functional Calculations of the Influence of Substitution on Singlet-Triplet Gaps in Carbenes and Vinylidenes" J. Phys. Org. Chem. 1997, 10, 755. which, in addition to being on topic vis a vis DFT for carbenes, contains references to several papers we (and others) have published comparing DFT to CASPT2 for carbenes and isoelectronic systems. CJC ------------------------------------------------------------------------- From: pschrei@gwdg.de To: "James Pugh " Subject: Re: CCL:carbene calculations Dear James, We have been doing a LOT of carbene studies with various DFT approaches, also comparing them to MRCI and CAS(SCF and MP2). You can find the relevant articles in: Carbenes: A Test Ground for Electronic Structure Methods. Holger F.Bettinger, Peter R. Schreiner,* Paul von Rague Schleyer, and Henry F. Schaefer III; in "The Encyclopedia of Computational Chemistry" Paul v. Ragu^N Schleyer, Norman L. Allinger, Tim Clark, Johann Gasteiger, Peter A. Kollman, Henry F. Schaefer III, and Peter R. Schreiner (Eds.), John Wiley & Sons Inc.; Chichester, 1998, in press. (the encyclopedia was published this week) And a full book chapter on the same topic: Computational Analyses of Prototype Carbene Structures and Reactions. Holger F. Bettinger, Paul von Rague Schleyer, Peter R. Schreiner,* and Henry F. Schaefer III; in "Modern Electronic Structure Theory and Applications to Organic Chemistry," Ernest L. Davidson (Ed.), World Scientific Press Inc.; Singapore 1997, pp 89-171; (ISBN: 981-02-3168-7) And several papers with benchmarks and comparisons: The Topomerization of Benzene. Holger F. Bettinger, Peter R. Schreiner, Paul von Rague Schleyer,* and Henry F. Schaefer III J. Am. Chem. Soc. 1998, 120, 5741-5750. Substituted Cyclopropylidenes and Cyclic Allenes. Holger F. Bettinger, Paul von Rague Schleyer,* Peter R. Schreiner, and Henry F. Schaefer III J. Org. Chem. 1997, 62, 9267-9275. The Naphthylcarbene Potential Energy Hypersurface. Yaoming Xie, Peter R. Schreiner,* Paul von Rague Schleyer, and Henry F. Schaefer III J. Am. Chem. Soc. 1997, 119, 1370-1377. The Ring Opening of Cycopropylidene and the Internal Rotation of Allene. Holger F. Bettinger, Peter R. Schreiner, Paul von Rague Schleyer,* and Henry F. Schaefer III* J. Phys. Chem. 1996, 100, 16147-16154. And here you find a nice table for methylene and various levels: Carbene Rearrangements Unsurpassed: The C7H6 Potential Energy Surface Revealed. Peter R. Schreiner,* William L. Karney, Paul von Rague Schleyer,* Weston T. Borden,* Tracy P. Hamilton, and Henry F. Schaefer III J. Org. Chem. 1996, 61, 7030-7039. The groups around Wentrup, McMahon, and Borden also have been doing a lot of work along these lines. All published in JACS and JOC. Hope this helps, Peter //// ___|--00___________________________________________________________________ C ^ Dr. Peter R. Schreiner http://www.gwdg.de/~pschrei \ ~/ Institut fuer Organische Chemie <><> Georg-August Universitaet Goettingen Tammannstr. 2 Phone: +49-(0)551-393287 D-37077 Goettingen, Germany FAX: +49-(0)551-399475 ___________________________________________________________________________ From: "Wang, Lifeng" To: "'pughj@ucs.orst.edu'" Cc: "'wangl@engr.sc.edu'" Subject: b3lyp method Yes. b3lyp is a very good method. I choose it for my calculation. a reference if you are interested: J.Phys.Chem.B 1998,102,5152-5157 p 5153 l 9-15 Li-Feng Wang wangl@engr.sc.edu ------------------------------------------------------------------ From: "Preston J. MacDougall" To: pughj@ucs.orst.edu Subject: correlation in carbenes James, While we did not use DFT, my colleagues and I explored the effect of other standard correlation techniques on the local properties of carbenes, especially the atomic properties (as defined by the quantum theoespecially the atomic properties (as defined by the quantum theory of atoms in molecules). The reference is J. Chem. Phys., vol. 88, p. 3792 (1988). Sincerely, Preston J. MacDougall Preston J. MacDougall Assistant Professor Department of Chemistry, Box X101 Middle Tennessee State University Murfreesboro, TN 37132 ----------------------------------------------------------------------- From: Ivan Rossi To: James Pugh Subject: Re: CCL:carbene calculations Chris Cramer of University of Minnesota published some papers on the subject in which he compares the performance of the methods too. Check his web page for a list of his publications. http://pollux.chem.umn.edu/~cramer -- Dr. Ivan Rossi - CIRB Biocomputing Unit e-mail: ivan@lipid.biocomp.unibo.it WWW: http://www.biocomp.unibo.it/ivan ------------------------------------------------------------------------- From: "Eric V. Patterson" To: James Pugh Subject: Re: CCL:carbene calculations James, In my work with Bob McMahon at Wisconsin, and Chris Cramer at Minnesota, I have used a large number of methods to study carbenes and relatehave used a large number of methods to study carbenes and related systems. These methods have included DFT, CCSD, CCSD(T), CAS and CASPT2. My conclusions might be summarized as follows: For "well-behaved" systems (i.e. mostly single-reference), all methods show good agreement. For modestly multi-reference systems, a gradient-corrected DFT functional with NO HF exchange performs well when compared to CCSD(T) and the MC methods. When things get significantly multi-reference, only CAS seems to be trustworthy. CCSD(T) does an okay job, but CASPT2 blows it, sintrustworthy. CCSD(T) does an okay job, but CASPT2 blows it, since it is just perturbation theory. Unfortunately, MRCI may the best option in those cases, but it's soooo expensive. I would be a little leary of relying on B3LYP. The HF exchange could trip you up if your system is even modestly multi-reference. But, if you have systems which are well described by a single reference, I see nothing wrong with your approach. It's certainly cheaper than the alternatives, and the evidence is that the results hold up well versus both experiment and higher-level calcs. Regards, Eric From: Eric V. Patterson voice: (515) 269-4543 Assistant Professor FAX: (515) 269-4285 Grinnell College Department of Chemistry Room 2030, Science Bldg. P.O. Box 805 patterse@ac.grin.edu Grinnell, IA 50112 http://www.grinnell.edu/individuals/patterse ------------------------------------------------------------------------ From: STEWARTK@RAND.PPRD.Abbott.Com To: pughj@ucs.orst.edu Subject: Re: CCL:carbene calculations Jim: I recommend that you send your message to a professor at Univ. Wisconsin. He will be happy to talk with you regarding high level calcs. on carbenes. He has done those kind of calcs. He reads his e-mail regularly and you can mail him at mcmahon@chem.wisc.edu. His name is Robert McMahon and is a professor of organic chemistry. I am not in the carbene field, but I happen to know Bob McMahon personally and I know if he cant answer your question, he can tell you exactly where to go to find out. Kent Stewart Abbott Laboratories --------------------------------------------------------------