From chemistry-request@server.ccl.net Wed Aug 21 04:59:00 2002 Received: from ifp.ifp.fr ([156.118.212.2]) by server.ccl.net (8.11.6/8.11.0) with ESMTP id g7L8wxQ03900 for ; Wed, 21 Aug 2002 04:58:59 -0400 Received: from irsun165.ifp.fr (irsun165.ifp.fr [156.118.22.211]) by ifp.ifp.fr (8.9.3/8.9.3) with ESMTP id KAA13398 for ; Wed, 21 Aug 2002 10:58:58 +0200 (MET DST) Received: from irnts21.ifp.fr (irnts21 [156.118.15.51]) by irsun165.ifp.fr (8.8.5/8.8.5) with ESMTP id KAA06662 for ; Wed, 21 Aug 2002 10:58:57 +0200 (MET DST) Received: by irnts21.ifp.fr with Internet Mail Service (5.5.2653.19) id ; Wed, 21 Aug 2002 10:58:57 +0200 Message-ID: <3AA99E087354D61185E4002048403CFC59DF99@irnts23.ifp.fr> From: DE BRUIN theodorus To: "'chemistry@ccl.net'" Subject: G98: opt=stable and freq Date: Wed, 21 Aug 2002 10:58:47 +0200 MIME-Version: 1.0 X-Mailer: Internet Mail Service (5.5.2653.19) Content-Type: text/plain; charset="iso-8859-1" Dear all, I am studying a system in which I am cracking a bond homolytically in a stepwise manner with B3LYP. I found that upon increase of the bond length the wavefunction becomes unstable, i.e. it converges to a singlet state. I tested the wf with G98 with stable=opt and this calculation confirmed the instabilty. However, I wish now to do a freq analysis on the new (stable) wavefunction. The keywords freq and stable=opt do not go together: Illegal IType generated by parse. Error termination via Lnk1e in /home2/debruint/bin//g98/l1.exe. On the other hand, if I perform a --link1-- job, in which I first do the stable=opt calculation and the second is the freq analysis, G98 is not able to read in the optimized wf using guess=read (failures to project some orbitals), and starts with the default guess, which of course finally converges to the wrong (instable) wavefunction. Does anybody have a suggestion how I can do a frequency analysis on the stable wf ? Thanks in advance, Theo de Bruin. PS. I found in the archives that Tim Geo (23 Feb 2001) posted I similar problem, but I could not find any replies to his questions. From chemistry-request@server.ccl.net Wed Aug 21 07:16:05 2002 Received: from pc15190.chemie.uni-marburg.de (root@[137.248.152.16]) by server.ccl.net (8.11.6/8.11.0) with ESMTP id g7LBG5Q04342 for ; Wed, 21 Aug 2002 07:16:05 -0400 Received: from sg1503.chemie.uni-marburg.de (guest@sg1503.Chemie.Uni-Marburg.DE [137.248.151.215]) by pc15190.chemie.uni-marburg.de (8.12.5/8.12.5/schn) with ESMTP id g7LBFpGe002615 for ; Wed, 21 Aug 2002 13:15:52 +0200 Received: from localhost (karin@localhost) by sg1503.chemie.uni-marburg.de (8.12.5/8.12.5/schn) with ESMTP id g7LBG4Xv029704 for ; Wed, 21 Aug 2002 13:16:04 +0200 (MEST) Date: Wed, 21 Aug 2002 13:16:03 +0200 From: Karin Wichmann To: chemistry@ccl.net Subject: basis set extrapolation Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII X-Virus-Scanned: by amavisd-milter (http://amavis.org/) Dear CCLers, I have done CCSD(T) single point calculations with cc-pVXZ (X=D,T,Q) and aug-cc-pVXZ (X=D,T) basis sets for several structures on a reaction path. Now I would like to know whether it is possible to do a basis set extrapolation from these data and if so, how is it to be done explicitly? Thank you very much in advance, Karin Wichmann e-mail: karin@chemie.uni-marburg.de From chemistry-request@server.ccl.net Wed Aug 21 12:21:10 2002 Received: from ms1.uibk.ac.at ([138.232.1.201]) by server.ccl.net (8.11.6/8.11.0) with ESMTP id g7LGLAQ12739 for ; Wed, 21 Aug 2002 12:21:10 -0400 Received: from localhost (web-mail1.uibk.ac.at [138.232.1.225] Hannes.Loeffler@uibk.ac.at) by ms1.uibk.ac.at (8.11.3/E2) with ESMTP id g7LGL7r68809482 for ; Wed, 21 Aug 2002 18:21:07 +0200 (CEST) Received: from 138.232.32.2 ( [138.232.32.2]) as user c72409@mail1.uibk.ac.at by web-mail1.uibk.ac.at with HTTP; Wed, 21 Aug 2002 18:21:06 +0200 Message-ID: <1029946866.3d63bdf30426a@web-mail1.uibk.ac.at> Date: Wed, 21 Aug 2002 18:21:07 +0200 From: Hannes Loeffler To: chemistry@ccl.net Subject: How to determine hydrogen bonds? MIME-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 8bit User-Agent: Internet Messaging Program (IMP) 3.0 X-Originating-IP: 138.232.32.2 [Sorry for posting again, I forgot the subject line] Hello all, We have the following problem: We use a three-site water model (point charges and vdW are atom centered) and would like to study hydrogen bonding. Is it possible to unambiguously determine if a given conformation (assuming threshold values for the O...H distance and the O-H...O angle) is hydrogen bondend? How could a possible algorithm look like? Hannes. -- Hannes Loeffler | tel: +43-(0)512-507-5166 | fax: +43-(0)512-507-2714 http://www-c724.uibk.ac.at/staff/loeffler/ | mailto:Hannes.Loeffler@uibk.ac.at Institute for General, Inorganic and Theoretical Chemistry University of Innsbruck, Austria From chemistry-request@server.ccl.net Wed Aug 21 11:01:53 2002 Received: from mailhub.health.nb.ca ([207.179.181.34]) by server.ccl.net (8.11.6/8.11.0) with ESMTP id g7LF1rQ10695 for ; Wed, 21 Aug 2002 11:01:53 -0400 Received: from onco_irmb.health.nb.ca ([10.10.100.12]) by mailhub.health.nb.ca (Post.Office MTA v3.5.3 release 223 ID# 607-57574U3000L300S0V35) with ESMTP id ca for ; Wed, 21 Aug 2002 12:01:47 -0300 Message-Id: <5.1.1.6.0.20020821115632.00a0e7b0@mailhub.health.nb.ca> X-Sender: miroslavac@mailhub.health.nb.ca X-Mailer: QUALCOMM Windows Eudora Version 5.1.1 Date: Wed, 21 Aug 2002 12:01:44 -0300 To: CCL From: miroslavac@health.nb.ca (Miroslava Cuperlovic-Culf) Subject: protein-protein interaction from sequence Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii"; format=flowed Dear All, This question might look kind of naive but what I am interested in is a software or a database search tool that will give me any kind of estimate of protein-protein interaction from primary sequence only (ideally) or from terciary structure of one protein and only primary sequence for the other. Is there such a tool available? Thanks a lot for ongoing help, Mira Miroslava Cuperlovic-Culf, Ph.D. Research Scientist Beausejour Medical Research Institute (IRMB) Dr. Leon Richard Oncology Center 37 Providence Street Moncton, NB E1C 8X3 Canada fax: 506-862-4222 tel: 506-862-4848 e-mail: miroslavac@health.nb.ca From chemistry-request@server.ccl.net Wed Aug 21 04:02:27 2002 Received: from mail1.rrz.Uni-Koeln.DE ([134.95.100.208]) by server.ccl.net (8.11.6/8.11.0) with ESMTP id g7L82PQ18249 for ; Wed, 21 Aug 2002 04:02:25 -0400 Received: from campfire.rrz.Uni-Koeln.DE (IDENT:2897@campfire.rrz.Uni-Koeln.DE [134.95.19.27]) by mail1.rrz.Uni-Koeln.DE (8.12.3/8.12.2) with ESMTP id g7L82OPw005991 for ; Wed, 21 Aug 2002 10:02:24 +0200 (MEST) Received: from localhost (acp64@localhost) by campfire.rrz.Uni-Koeln.DE (8.9.1a/8.9.1) with SMTP id KAA10462 for ; Wed, 21 Aug 2002 10:02:22 +0200 (MET DST) X-Authentication-Warning: campfire.rrz.Uni-Koeln.DE: acp64 owned process doing -bs Date: Wed, 21 Aug 2002 10:02:22 +0200 (MET DST) From: Johannes Weber Reply-To: Johannes Weber To: CHEMISTRY@ccl.net Subject: SUMMARY: g98: excited state densities via TDDFT ? Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII X-Virus-Scanned: by amavisd-milter (http://amavis.org/) X-Spam-Report: X_AUTH_WARNING,SUBJ_ENDS_IN_Q_MARK,FROM_ENDS_IN_NUMS,DEAR_SOMEBODY,DOUBLE_CAPSWORD Dear CCL members, a week ago I posted a question concerning excited state densities via the TDDFT method in g98. Thanks to Doug Fox, Peter W. Thulstrup, Kiniu Wong and Jens Spanget-Larsen for answers! Here is the summary (detailed answers are collected below): The output after "Total CI Rho(1) Density" gives the density of the excited state X which was specified in the route section with TD(Root=X). However, this is the "unrelaxed" density, i.e. a more or less crude approximation, see the paper of Wiberg et al, JPC 96, 671 (1992) and Dougs mail. Therefore, the numbers cannot be considered as very reliable. They should (hopefully) show the correct trend in most cases, however. Despite of the fact that it is technically possible to calculate the densities in the KS-formalism, it still remains unclear to me, if this can be justified from a theoretical point of view. The transition density is a quantity between two different electronic states. Two wavefunctions are needed for its calculation. DFT is, in general, restricted to a single state (ground state) and a wavefunction does formally not exist. with kind regards, :-)ohannes. ------------------------------------------------------------------- The original question was : -------------------------------------------------------------------- Dear CCL members, I have a general question: Is it possible to calculate electron densities of excited states and/or excited state spin densities and/or transition densities with the TDDFT method ? With respect to the TDDFT implementation in g98 this general question splits up into several concrete questions: Which density (ground state or excited state) will be written to the checkpoint file if I specify the DENSITY keyword in the route section together with TD(Root=1) ? -- The g98 manual is not very explicit about this point. If I look at the formatted checkpoint file I find two densities, named the "Total SCF Density" and "Total CI Rho(1) Density". What densities are given here? I was not able to extract the Total CI Rho(1) Density with the cubegen utility. Is this possible? Thanks in advance for all answers. If there is interest I will summarize them, of course. With kind regards, :-)ohannes ---------------------------------------------------------------------- reply of Doug Fox: ===================================================================== Dr. Weber, Gaussian 98 does not properly deliver densities for excited states using the TD-DFT methods. So if you specify DENSITY you should get an error message. The manual incorrectly suggests that this combination is valid and it is on the wish list but won't be there in G98. The issue is that to get reliable properties, density, dipole, gradients, frequencies, you need to do better than just a first order density matrix and need to consider the change in the HF density due to the excitation. This is solved in the CIS method with a small CPCI step but the corresponding code is not implemented for TDDFT or TDHF. The field labeled Total CI Rho(1) is a density formed from the weights of the excited KS determinants. The proper CI density would be labeled Total CI Density and it would correspond to the root specified on the route, i.e. CIS(Root=2) Density would compute the correction using the second root of the CIS, use that density for properties listed in the output and store that density as the Total CI Density if you make a formatted checkpoint file. There is no easy way to recover the CI Rho(1) but if you modify the formatted checkpoint file and shorten this to Total CI Density cubegen will pick it up. ===================================================================== reply of Peter Waaben Thulstrup: ===================================================================== Dear Johannes, I have asked a similar question to the help at Gaussian. I got a reply > from Doug Fox, included (in part) below: > Peter, > > The problem with assigning spin for excited states starting from > an open shell system is not a solved problem. Since the UHF or UDFT > reference will generally be spin contaminated it is not easy to sort > out the spin. From an RHF reference the sign flips between alpha and > beta components but there is no simple relation for the spin > unrestricted reference. >... > The Density options for TDDFT are not correct for Rev. A.7 and we have > actually turned them off with A.11 because at present even the one > particle RhoCI type density is not correct. It is on the list of > things to do right but we lost the person who was going to do it to an > academic position and they have not surfaced from trying to set up a > group/lab etc. So in short I would not put any credence in trying to > analyze the densities reported there. So, unfortunately bad news. ========================================================================== reply of Kiniu Wong: ========================================================================== Dear Johannes, I also have the same problem about TDDFT. For TDHF and CIS, I am able to generate the excited state density by either "Density = CI" or "Density=Current". However, for TDDFT, if I specify "Density = CI", it shows the error. But "Density = Current", the density is just SCF (ground state). Moreover, "Density = RhoCI" and "Density(CIS=1)" work properly for TDDFT, TDHF as well as CIS. Although I am not so sure what do they mean or what is the difference between them and "Density = CI" or "Density = Current", however, from Chapter 9 Exploring Chemistry with Electronic Structure Methods, they don't prefer "Density = RhoCI" and "Density(CIS=1)". Anyway, if there are any useful or interesting replies, may you forward them to me as well? Thanks! Regards, Kiniu ========================================================================== --------------------------------------------------------------------- | Johannes Weber | Email: | | Universitaet zu Koeln | Johannes.Weber@uni-koeln.de | | Institut fuer Theoretische Chemie | tel: 0049-(0)221-470-4812 | | Luxemburger Str. 116 | fax: 0049-(0)221-470-5144 | | 50939 Koeln | | --------------------------------------------------------------------- From chemistry-request@server.ccl.net Wed Aug 21 22:40:26 2002 Received: from heretic.bri.nrc.ca ([206.167.189.33]) by server.ccl.net (8.11.6/8.11.0) with ESMTP id g7M2ePQ10270 for ; Wed, 21 Aug 2002 22:40:25 -0400 Received: from dirac.bri.nrc.ca ([206.167.189.67]) by heretic.bri.nrc.ca with esmtp (Exim 4.10) id 17hiom-0002H1-00 for chemistry@ccl.net; Wed, 21 Aug 2002 22:39:32 -0500 Received: from localhost (wujih@localhost) by dirac.bri.nrc.ca (SGI-8.9.3/8.9.3) with ESMTP id WAA56537 for ; Wed, 21 Aug 2002 22:40:24 -0400 (EDT) X-Authentication-Warning: dirac.bri.nrc.ca: wujih owned process doing -bs Date: Wed, 21 Aug 2002 22:40:24 -0400 From: Jianhui Wu To: chemistry@ccl.net Subject: protein order parameter, MD Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII X-SpamVersion: 1.2 X-SpamLevel: 2 X-SpamDetails: Domain:Good(-50) Subject:KW:Free:1(25) Body:KW:Free:1(15) Body:KW:Info:1(8) Body:KW:Misc:1(4) Dear CCL, I am looking for a software to caluate the protein order parameter from molecular dynamic trajectory (generated by Amber). Any help will be greatly appreciated. Thanks a lot, JianHui Wu McGill University Department of Oncology Lady Davis Institute Montreal, QC