From owner-hyperchem@hyper.com Tue Oct 1 13:44:27 1996 From: "Udo Haberl" To: hyperchem@hyper.com Date: Tue, 1 Oct 1996 15:15:04 GMT+1 Subject: Spin densities 2 Reply-To: haberl@uni-bonn.de Cc: "Jens Spanget-Larsen" Jens Spanget-Larsen wrote: > I wonder if the information requested by Udo, the "numerical values > of the spin densities on the different atoms", is not simply available > from the log-file? The log-file of an UHF calculation contains the > atomic orbital populations for electrons with alpha and with beta > spin, and by simple subtraction of the two the corresponding spin > densities can be obtained. Udo, is this not what you need? Thank you for your answer! Is this really correct? Or do you have to regard the population analysis? Udo _______________________________________________________________________________ From owner-hyperchem@hyper.com Tue Oct 1 14:59:08 1996 From: "Jens Spanget-Larsen" Organization: Roskilde Universitetscenter To: hyperchem@hyper.com Date: Tue, 1 Oct 1996 17:37:51 +0100 Subject: Re: Spin densities 2 > Jens Spanget-Larsen wrote: > >> I wonder if the information requested by Udo, the "numerical values >> of the spin densities on the different atoms", is not simply available >> from the log-file? The log-file of an UHF calculation contains the >> atomic orbital populations for electrons with alpha and with beta >> spin, and by simple subtraction of the two the corresponding spin >> densities can be obtained. Udo, is this not what you need? > > Thank you for your answer! > > Is this really correct? Or do you have to regard the > population analysis? > > Udo > Yes, Udo, I think this is correct. The log-file data I refer to is a result of the Mulliken population analysis. Below you find part of the HyperChem log file for an AM1-UHF calculation on the allyl radical: ATOMIC ORBITAL ELECTRON POPULATIONS Alpha Orbitals: 0.654760 0.527975 0.498700 0.856273 0.591332 0.466047 0.446206 0.287454 0.654760 0.505871 0.520803 0.856273 0.420411 0.420601 0.451523 0.420411 0.420601 Beta Orbitals: 0.606713 0.484716 0.447335 0.143394 0.628592 0.510102 0.499963 0.713212 0.606713 0.456777 0.475274 0.143394 0.465393 0.464764 0.423500 0.465393 0.464764 If you subtract the two sets of numbers you get the spin populations of the atomic arbitals (AOs): C1(2s),C1(2px),C1(2py),C1(2pz); C2(2s),C2(2px), ... If you want overall atomic spin densities, you must add up the AO contributions for each atom. I think that Joel misinterpreted your question and thought you wanted the spin density on each atomic NUCLEUS (e.g., for a Fermi analysis), and that is more difficult. Aber das war doch nicht so schlimm, oder? Mit freundlichen Gru:ssen, Jens >--< =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=- JENS SPANGET-LARSEN Department of Chemistry Phone: +45 46757711 Roskilde University (RUC) Fax: +45 46757721 POB 260, DK-4000 Roskilde, Denmark E-Mail: JSL@virgil.ruc.dk =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=- _______________________________________________________________________________ From owner-hyperchem@hyper.com Sat Oct 5 14:21:44 1996 Date: Sat, 5 Oct 1996 16:30:59 +0000 (GMT) From: Peter Gedeck To: hyperchem@hyper.com Subject: IR calculations and log files Hello, For H2O I obtain the following entry in the log-file. ---------------------------------------------------------------------------- ==== Force Constant Matrix in Milli-Dynes / Angstrom ==== (I -- Atom Index Z Atomic Number) I Z I Z I Z I Z 1 8 2 1 3 1 1 8 5.48377 3.42771 3.42771 2 1 3.42771 3.55750 0.54322 3 1 3.42771 0.54322 3.55750 ---------------------------------------------------------------------------- In Mopac or Vamp the force constant matrix is a 3Nx3N matrix and not a NxN matrix. I assume that HyperChem also calculates a 3Nx3N force matrix, so how does HyperChem calculate the NxN matrix? The meaning of the off-diagonal matrix elements seem to be the force constants of the atom-atom bond. But, how should one interpret the diagonal elements of the matrix? If I print the normal coordinates as well, I get lines like this: l( 1x) 0.33451 0.00000 0.00000 ^ Here we have a small L. Why? Should this be the name given to the transformation matrix in the manual? Let me take this letter to suggest a slight improvement of the IR vector rendering used in HyperChem. Sometimes it would be really greate to be able to invert the IR vector, so that all vectors point in the opposite direction. With regards, Peter PS: If I carry out a 2x1 singly excited state CI calculation (2 occupied and 1 unoccupied) and print the CI eigenvalues and eigenvectors I get: ===== CI Eigenvalues(in (eV) ) and Eigenvectors ===== CI State 1 2 3 4 5 Eigenvalue 0.00000 5.48531 6.54158 8.35672 9.57857 -45 0.00000*************************************** 0.00000***** 1429249331 *********************************************************** 1187692171 0.00000 0.00000************************************* 1190610748 *********************************************************** -1599662333 ********************************************************** Something seems to be really wrong here. (HyperChem 4) Dr. Peter Gedeck Centrum f. Computer Chemie - Institut f. Organische Chemie I Naegelsbachstrasse 25 - D-91052 Erlangen Deutschland / Germany Tel: ++9131 - 85 6581 Fax: ++9131 - 85 6565 E-Mail: gedeck@organik.uni-erlangen.de WWW: http://organik.uni-erlangen.de/clark/gedeck _______________________________________________________________________________ From owner-hyperchem@hyper.com Mon Oct 7 19:57:53 1996 Date: Mon, 7 Oct 96 12:14:47 -0400 From: polowin (Joel Polowin) To: Peter Gedeck , hyperchem@hyper.com Subject: Re: IR calculations and log files > Date: Sat, 5 Oct 1996 16:30:59 +0000 (GMT) > From: Peter Gedeck > > In Mopac or Vamp the force constant matrix is a 3Nx3N matrix and not a > NxN matrix. I assume that HyperChem also calculates a 3Nx3N force matrix, > so how does HyperChem calculate the NxN matrix? A 3N * 3N force constant matrix can be very confusing so HyperChem, like MOPAC, gives an N * N force constant matrix describing the atom-atom interaction forces. This N * N force constant matrix is formed from the 3N * 3N force constant matrix by calculating the norm of the tensor of each atom. You may reconstruct the 3N * 3N force constant matrix from HyperChem log file output using the matrix equation: F C = C E where F is the MASS-WEIGHTED force constant matrix (3N * 3N), C is the eigenvector (normal coordinate) matrix and E is the eigenvalue diagonal matrix. You can get C and E from the log file (with QuantumPrintLevel=9) or script output, you can calculate the matrix F by multiplying on the right by the inverse of C. The relationship between eigenvalue Ei in E and the vibrational frequency f is f = SQRT(Ei) / (2*PI*c) where c is the speed of light. > If I print the normal coordinates as well, I get lines like this: > l( 1x) 0.33451 0.00000 0.00000 > ^ Here we have a small L. Why? Should this be the name given to the > transformation matrix in the manual? The 'l' values are normalized coefficients; the symbol "L" is sometimes used instead. (_Spectroscopy and Molecular Structure_ by Gerald W. King uses the "l" symbol.) > Let me take this letter to suggest a slight improvement of the IR vector > rendering used in HyperChem. Sometimes it would be really greate to be > able to invert the IR vector, so that all vectors point in the opposite > direction. There is no trivial way to do this, as far as I know. But if you really need to, you can save the IR spectrum (with the SAVEIR.SCR script, or using the File/Export function of HyperChem 5), and then manually change the sign on each component of the normal modes. Reading the spectrum back in will reverse the direction of the IR vectors. Alternately, you could use a macro to read the normal modes, change their signs, and put them back into HyperChem. Obviously, these are a bit clumsy; I will add the suggestion for an automatic reversal to our wish list. > PS: If I carry out a 2x1 singly excited state CI calculation (2 occupied > and 1 unoccupied) and print the CI eigenvalues and eigenvectors I get: > ===== CI Eigenvalues(in (eV) ) and Eigenvectors ===== > > CI State 1 2 3 4 5 > Eigenvalue 0.00000 5.48531 6.54158 8.35672 9.57857 > > -45 0.00000*************************************** 0.00000***** > 1429249331 *********************************************************** > 1187692171 0.00000 0.00000************************************* > 1190610748 *********************************************************** > -1599662333 ********************************************************** It appears to be a formatting problem; I get the same sort of thing with HyperChem 4.5 with those settings. I'll add it to the bug list. Joel ------------ Joel Polowin, Ph.D. Manager, Scientific Support Email to: polowin@hyper.com WWW: http://www.hyper.com/ Hypercube Inc, 419 Phillip St, Waterloo, Ont, Canada N2L 3X2 (519)725-4040 Info requests to: info@hyper.com Support questions to: support@hyper.com Email group: Send "subscribe hyperchem" (or "unsubscribe hyperchem") to hyperchem-request@hyper.com; please do not send such administrative messages to the group itself. _______________________________________________________________________________ From owner-hyperchem@hyper.com Tue Oct 8 13:37:06 1996 Sender: From: "Galeazzi Roberta" To: hyperchem@hyper.com Date: Tue, 8 Oct 1996 13:00:23 +0000 Subject: need help: Hyperchem script Dear collegues, does anyone know if it is possible to write a script for Hyperchem 4.5 including commands for Chemplus? (for example to do a conformational search) I have the version 1.5 of Chemplus. thanks in advance for your help Roberta ************************************************************************** Dr. Roberta Galeazzi Dipartimento di Scienze dei Materiali e della Terra Universita di Ancona via Brecce Bianche 60131 Ancona fax: 39-71-2204714 e-mail: roberta@popcsi.unian.it _______________________________________________________________________________ From owner-hyperchem@hyper.com Tue Oct 8 22:48:08 1996 Date: Tue, 8 Oct 96 16:21:00 -0400 From: polowin (Joel Polowin) To: "Galeazzi Roberta" , hyperchem@hyper.com Subject: Re: need help: Hyperchem script > From: "Galeazzi Roberta" > Date: Tue, 8 Oct 1996 13:00:23 +0000 > > does anyone know if it is possible to write a script for Hyperchem > 4.5 including commands for Chemplus? (for example to do a conformational search) It is not possible to write a HyperChem 4.5 script that will control ChemPlus; there are no script commands in that version that can do such things. With HyperChem 5, it *may* be possible, by using a Tcl/Tk script to start the QSAR Properties module and send it commands to work with HyperChem. The process would be complicated at best, and it might not work at all. And only the QSAR Properties module is able to accept commands by DDE; the other modules can only be controlled manually. Joel ------------ Joel Polowin, Ph.D. Manager, Scientific Support Email to: polowin@hyper.com WWW: http://www.hyper.com/ Hypercube Inc, 419 Phillip St, Waterloo, Ont, Canada N2L 3X2 (519)725-4040 Info requests to: info@hyper.com Support questions to: support@hyper.com Email group: Send "subscribe hyperchem" (or "unsubscribe hyperchem") to hyperchem-request@hyper.com; please do not send such administrative messages to the group itself. _______________________________________________________________________________ From owner-hyperchem@hyper.com Wed Oct 9 02:18:04 1996 From: "Steven Desjardins" To: hyperchem@hyper.com Date: Tue, 8 Oct 1996 14:57:09 EST5EDT Subject: pKa's Hi. Can anyone refer me some literature on the calculation of pKa's using quantum chemical methods? Thanks. Steven G. Desjardins Department of Chemistry Washington and Lee University Lexington VA 24450 (703) 463-8873 _______________________________________________________________________________ From owner-hyperchem@hyper.com Wed Oct 9 22:57:35 1996 From: JJAINHUKNAN@miavx1.acs.muohio.edu Date: Wed, 09 Oct 1996 20:48:13 -0500 (EST) Subject: Bond Order calculation To: hyperchem@hyper.com Hi, Does anyone know how to use hyperchem to calculate bond order? Regard, jaran _______________________________________________________________________________ From owner-hyperchem@hyper.com Thu Oct 10 13:59:43 1996 Date: Thu, 10 Oct 96 09:47:09 -0400 From: polowin (Joel Polowin) To: JJAINHUKNAN@miavx1.acs.muohio.edu, hyperchem@hyper.com Subject: Re: Bond Order calculation > From: JJAINHUKNAN@miavx1.acs.muohio.edu > Date: Wed, 09 Oct 1996 20:48:13 -0500 (EST) > > Does anyone know how to use hyperchem to calculate bond order? It is not possible to calculate bond orders automatically with HyperChem, but they can be calculated manually from the density matrices, which can be printed in a log file. The bond order between atoms A and B is the sum of the squares of the elements in the density matrix with rows for atom A's orbitals and columns for atom B's orbitals, or vice versa. If you are doing UHF calculations, you must add the elements of the alpha and beta matrices together before squaring. I tried this for C2H4 as model-built, with an AM1 single-point calculation, and got a C--C bond order of 2.0004. Joel ------------ Joel Polowin, Ph.D. Manager, Scientific Support Email to: polowin@hyper.com WWW: http://www.hyper.com/ Hypercube Inc, 419 Phillip St, Waterloo, Ont, Canada N2L 3X2 (519)725-4040 Info requests to: info@hyper.com Support questions to: support@hyper.com Email group: Send "subscribe hyperchem" (or "unsubscribe hyperchem") to hyperchem-request@hyper.com; please do not send such administrative messages to the group itself. _______________________________________________________________________________ From owner-hyperchem@hyper.com Thu Oct 10 19:46:39 1996 Date: Thu, 10 Oct 1996 13:32:01 -0800 (AKDT) From: KELLER JOHN W To: hyperchem@hyper.com Subject: Windows NT 4.0 Joel, Could you summarize the latest information concerning the compatibility of Hyperchem 4.0, 4.5, and 5.0 (both standalone and network versions) with WinNT 4.0? Thanks John KEller Univ of Alaska Chem Dept _______________________________________________________________________________ From owner-hyperchem@hyper.com Fri Oct 11 18:51:41 1996 From: "Mei, George" To: "'Hypercube'" Subject: Molecular Dynamics of Hypercube 4.5 under Windows NT 3.51 Date: Fri, 11 Oct 1996 15:46:00 -0500 I recently switched from my Dell 50 486 to Gateway 2000 p5-200 computer and am now running Hypercube 4.5 under Windows NT 3.51 with the rainport patch driver. I am running into problems with the molecular dynamics. If I use the snapshot feature of the MD, I'll get a Hyper Cube message of "Computation aborted! Communication problem." That will lock up my computer, and I have to reboot. Otherwise, I am able to run for 0.329 ps at 0.001ps/step with 1refresh/data. If I run the same with 20 refresh/data, it will stop at 0.348 ps, not 20 times as long as described in the FAQ. Can someone tell me how can I improve or resolve the problem? _______________________________________________________________________________ From owner-hyperchem@hyper.com Sat Oct 12 09:51:46 1996 Date: Sat, 12 Oct 1996 11:11:53 +0200 From: Riccardo Vivani Organization: Dipartimento di Chimica To: hyperchem@hyper.com Subject: cartesian to cell coordinates Hi all, I am working on Hyperchem 4.0/Chemplus1 and I'm trying to build/modify some crystal fragment with them. In order to convert cartesian into cell coordinates I often need to align my fragment with a given plane (xy for instance). I tried to do this with "build/align viewer" or "build/align molecule", but I wasn't able to do that. Is there some simple way? Thanks to you all -- ------------------------------------- | Riccardo Vivani | | Laboratorio di Chimica Inorganica | | Dipartimento di Chimica | | via Elce di Sotto, 8 | | 06123 PERUGIA - ITALY | | Tel. +39 75 585 5564 | | Fax. +39 75 585 5566 | | e-mail: ric@unipg.it | ------------------------------------- _______________________________________________________________________________ From owner-hyperchem@hyper.com Mon Oct 14 12:10:07 1996 From: "A.J.M.DRIESSEN" Organization: Department of Biology RuG To: hyperchem@hyper.com Date: Mon, 14 Oct 1996 13:18:59 MET Subject: 3 D viewing Dear Joel I am using the hardware setting for 3D viewing with the SGI hyperchem 4.5 and crystaleyes spectacles. However, when the screen is switched on the 3D modus and subsequently back to the normal viewing mode, the overall refresh rate of the screen (Indigo2Extreme, 20" hardware stereo ready) remains with 60 Hz instead of returning to 72 Hz. is there a way to restore the original setting automatically when returning from 3D viewing. yours sincerely arnold driessen Dr. A.J.M. Driessen Department of Microbiology Groningen Biomolecular and Biotechnology Institute Biological Center Kerklaan 30 9751 NN HAREN Netherlands Tel. 31-50-3632164 Fax. 31-50-3632154 A.J.M.DRIESSEN@BIOL.RUG.NL _______________________________________________________________________________ From owner-hyperchem@hyper.com Mon Oct 14 17:08:30 1996 X-Sender: JeffAyres@postoffice.worldnet.att.net To: hyperchem@hyper.com From: Jeffrey J Ayres Subject: Appropriate reference manual Date: Mon, 14 Oct 1996 18:49:19 +0000 Attn: Netters, I'm looking for a particular reference manual which accompanies the Hyperchem program. The manual describes the various equations used in each of the methods. Chapter three of the manual has an excellent description of Hartree Fock calculations (69-76). Could you please send the reference particulars such as exact title, year of publication. It is probably version 4.0. _______________________________________________________________________________ From owner-hyperchem@hyper.com Tue Oct 15 06:36:06 1996 Date: Tue, 15 Oct 96 12:03:35 +0300 From: "Ananikov V.P." Reply-To: To: hyperchem@hyper.com Subject: information about TNDO/1-2 methods Dear Netters, I want to obtain more information about semiempirical methods TNDO/1 and TNDO/2 which HyperNMR uses for calculations of the shielding and coupling constants. Could anyone send me the references concerning this subject? Does anyone have information about accuracy of these methods? (the references with practical examples of HyperNMR calculations are also greatly appreciated). Thank you. Sincerely, Valentin. @ Valentin P. Ananikov (Post-Graduate Student) @ @ NMR Laboratory @ @ N.D. Zelinsky Institute of Organic Chemistry @ @ Leninsky Prospect 47 @ @ Moscow 117913 @ @ Russia @ @ email: val@nmr1.ioc.ac.ru @ _______________________________________________________________________________ From owner-hyperchem@hyper.com Tue Oct 15 14:07:21 1996 Date: Tue, 15 Oct 96 11:19:17 -0400 From: polowin (Joel Polowin) To: "Mei, George" , "'Hypercube'" Subject: Re: Molecular Dynamics of Hypercube 4.5 under Windows NT 3.51 > From: "Mei, George" > Date: Fri, 11 Oct 1996 15:46:00 -0500 > > I recently switched from my Dell 50 486 to Gateway 2000 p5-200 computer and > am now running Hypercube 4.5 under Windows NT 3.51 with the rainport patch > driver. I am running into problems with the molecular dynamics. If I use the > snapshot feature of the MD, I'll get a Hyper Cube message of "Computation > aborted! Communication problem." There is a known problem with using Snapshots and Averaging to create a graph during MD calculations under NT 3.51; a crash results. There is no problem if the graph is generated during playback of a snapshot file. The problem is not seen under Windows 95 nor under NT 4; we therefore believe that it is the result of a bug in NT 3.51 . This problem is not related to the communications problems under NT 3.51, however (as far as we know). > That will lock up my computer, and I have > to reboot. Otherwise, I am able to run for 0.329 ps at 0.001ps/step with > 1refresh/data. If I run the same with 20 refresh/data, it will stop at 0.348 > ps, not 20 times as long as described in the FAQ. Can someone tell me how > can I improve or resolve the problem? One thing that we do know about this communication problem is that the effects are cumulative -- in a single HyperChem session, if you run several short calculations, and then try a long one, the short calculations will consume some of the available resources and the long calculation will not run as long as if you had started it immediately. Could this be why your calculation did not run as long as expected -- had you done other calculations first? This problem does not appear to occur under NT 4. Joel ------------ Joel Polowin, Ph.D. Manager, Scientific Support Email to: polowin@hyper.com WWW: http://www.hyper.com/ Hypercube Inc, 419 Phillip St, Waterloo, Ont, Canada N2L 3X2 (519)725-4040 Info requests to: info@hyper.com Support questions to: support@hyper.com Email group: Send "subscribe hyperchem" (or "unsubscribe hyperchem") to hyperchem-request@hyper.com; please do not send such administrative messages to the group itself. _______________________________________________________________________________ From owner-hyperchem@hyper.com Tue Oct 15 14:37:59 1996 Date: Tue, 15 Oct 96 11:26:18 -0400 From: polowin (Joel Polowin) To: Riccardo Vivani , hyperchem@hyper.com Subject: Re: cartesian to cell coordinates > Date: Sat, 12 Oct 1996 11:11:53 +0200 > From: Riccardo Vivani > > I am working on Hyperchem 4.0/Chemplus1 and I'm trying > to build/modify some crystal fragment with them. In order > to convert cartesian into cell coordinates I often need to > align my fragment with a given plane (xy for instance). > I tried to do this with "build/align viewer" or > "build/align molecule", but I wasn't able to do that. > Is there some simple way? You should be able to align a structure with the axes using Build/Align Molecule. You could, for example, select a few atoms in a line and align them with an axis, or select a few atoms in a plane and align the tertiary axis of the atoms with a coordinate axis, thereby aligning the plane of the atoms with the perpendicular coordinate plane. In these cases, if the selected atoms are not the entire set of atoms, you would want to have whole molecule translation turned on. Another procedure is to save the structure in MOPAC Z-matrix format and then read it back in. The first atom is placed at the origin (0,0,0); the second atom is placed on the positive x axis (x,0,0); the third atom in the positive xy plane (x,y,0); remaining atoms are positioned with respect to these first three. The tricky part here is making sure that the appropriate atoms are the first three; this may require breaking and reconnecting some bonds to renumber atoms and molecules. Joel ------------ Joel Polowin, Ph.D. Manager, Scientific Support Email to: polowin@hyper.com WWW: http://www.hyper.com/ Hypercube Inc, 419 Phillip St, Waterloo, Ont, Canada N2L 3X2 (519)725-4040 Info requests to: info@hyper.com Support questions to: support@hyper.com Email group: Send "subscribe hyperchem" (or "unsubscribe hyperchem") to hyperchem-request@hyper.com; please do not send such administrative messages to the group itself. _______________________________________________________________________________ From owner-hyperchem@hyper.com Tue Oct 15 15:19:00 1996 Date: Tue, 15 Oct 96 11:11:23 -0400 From: polowin (Joel Polowin) To: KELLER JOHN W , hyperchem@hyper.com Subject: Re: Windows NT 4.0 > Date: Thu, 10 Oct 1996 13:32:01 -0800 (AKDT) > From: KELLER JOHN W > > Could you summarize the latest information concerning the > compatibility of Hyperchem 4.0, 4.5, and 5.0 (both standalone and network > versions) with WinNT 4.0? All versions appear to run correctly in standalone mode under NT 4; no additional drivers are needed; it is necessary to have the "default locale" set to "US" in order to have HyperChem's script commands and DDE commands work properly. HyperChem 5 seems to run properly over a network. We're still testing the earlier releases for network operations under NT 4, and we still need to test the operations of the license servers. Joel ------------ Joel Polowin, Ph.D. Manager, Scientific Support Email to: polowin@hyper.com WWW: http://www.hyper.com/ Hypercube Inc, 419 Phillip St, Waterloo, Ont, Canada N2L 3X2 (519)725-4040 Info requests to: info@hyper.com Support questions to: support@hyper.com Email group: Send "subscribe hyperchem" (or "unsubscribe hyperchem") to hyperchem-request@hyper.com; please do not send such administrative messages to the group itself. _______________________________________________________________________________ From owner-hyperchem@hyper.com Tue Oct 15 19:35:27 1996 Date: Tue, 15 Oct 96 17:38:29 -0400 From: polowin (Joel Polowin) To: hyperchem@hyper.com, Subject: Re: information about TNDO/1-2 methods > Date: Tue, 15 Oct 96 12:03:35 +0300 > From: "Ananikov V.P." > > I want to obtain more information about semiempirical methods > TNDO/1 and TNDO/2 which HyperNMR uses for calculations of the > shielding and coupling constants. > > Could anyone send me the references concerning this subject? > > Does anyone have information about accuracy of these methods? > (the references with practical examples of HyperNMR > calculations are also greatly appreciated). The only journal article that we know about which references HyperNMR is a software review: "HyperNMR for Windows" by Ovidiu Ivanciuc, _J. Chem. Inf. Comput. Sci._, *36*, 916 (1996). The only published description of the calculation methods is in the HyperNMR manual. We're working on other publications. As usual, we would like to hear about any journal articles which describe use of our products. Joel ------------ Joel Polowin, Ph.D. Manager, Scientific Support Email to: polowin@hyper.com WWW: http://www.hyper.com/ Hypercube Inc, 419 Phillip St, Waterloo, Ont, Canada N2L 3X2 (519)725-4040 Info requests to: info@hyper.com Support questions to: support@hyper.com Email group: Send "subscribe hyperchem" (or "unsubscribe hyperchem") to hyperchem-request@hyper.com; please do not send such administrative messages to the group itself. _______________________________________________________________________________ From owner-hyperchem@hyper.com Wed Oct 16 06:09:09 1996 From: ps@ocisgi7.unizh.ch (Serge Pachkovsky) Subject: Re: information about TNDO/1-2 methods To: polowin (Joel Polowin) Date: Wed, 16 Oct 1996 08:29:38 +0100 (MET) Cc: hyperchem@hyper.com, val@nmr1.ioc.ac.ru > The only journal article that we know about which references HyperNMR > is a software review: "HyperNMR for Windows" by Ovidiu Ivanciuc, > _J. Chem. Inf. Comput. Sci._, *36*, 916 (1996). The only published > description of the calculation methods is in the HyperNMR manual. Well, there are also: R.K. Gilpin, "PREDICTING 1D NMR SPECTRA - HYPERNMR" in Analyt. Chem. 1995 Sep 1;67(17):A 541 and M. Witanowski, Z. Biedrzycka, W. Sicinska, Z. Grabowski, and G.A. Webb, "SOLVENT EFFECTS ON THE NITROGEN NMR SHIELDINGS IN OXAZOLE AND OXADIAZOLE SYSTEMS", J. Magn. Res. A 1996 Jun;120(2):148-154 The later gives (more or less) reasonable conceptual description of TNDO/2, along with some results for Nitrogen-14. I would call the lack of (reproducible) definition of TNDO/2 in open scientific literature a major obstacle for its acceptance by the community in general... Regards, /Serge.P _______________________________________________________________________________ From owner-hyperchem@hyper.com Wed Oct 16 18:16:31 1996 Date: Wed, 16 Oct 96 14:50:41 -0400 From: polowin (Joel Polowin) To: "Jose G. de la Campa" , hyperchem@hyper.com Subject: Re: about SGI version > Date: Wed, 16 Oct 1996 03:53:37 +0100 > From: "Jose G. de la Campa" > > We are using Hyperchem version 4.5 for Silicon Graphics Workstations, mainly > for ab initio calculations. When we try to calculate molecules the size of > aniline using the 6-31G** basis set the size of the temporal integrals file > is so big that the disk becomes full and the calculation stops. To avoid > this problem we have installed a new, bigger disk and have tried to use the > instruction IntegralPath (pag 550 of the Reference manual) in the chem.ini > file, to tell the program to save the two-electron integrals in the new > disk. However, all the attempts have been completely unsuccesfull and the > program always saves the integrals in the directory /var/tmp, that > unfortunately is in the first and smaller disk. This forces us to make the > calculations in direct SCF, what is a much slower system. > The instruction we are giving is IntegralPath=/name of the disk/directory. > Even if we do not change the disk and only say IntegralPath=/directory the > instruction does not work. The procedure for specifying the integral storage location changed between the Windows and SGI versions, but that change was not properly described in the manuals. For the SGI version, you need to specify the storage directory in the TMPDIR environment variable -- for example, setenv TMPDIR /disk2/tmp This is described on page 19 of the _Getting Started_ manual. To have this set up automatically, you would put the command in your .login or .cshrc file so that it will be done automatically when you log in or start a new shell. Joel ------------ Joel Polowin, Ph.D. Manager, Scientific Support Email to: polowin@hyper.com WWW: http://www.hyper.com/ Hypercube Inc, 419 Phillip St, Waterloo, Ont, Canada N2L 3X2 (519)725-4040 Info requests to: info@hyper.com Support questions to: support@hyper.com Email group: Send "subscribe hyperchem" (or "unsubscribe hyperchem") to hyperchem-request@hyper.com; please do not send such administrative messages to the group itself. _______________________________________________________________________________ From owner-hyperchem@hyper.com Thu Oct 17 23:32:40 1996 Date: Thu, 17 Oct 96 16:00:34 -0400 From: polowin (Joel Polowin) To: hyperchem@hyper.com, poso@messi.uku.fi Subject: Molecular Dynamics > From: "Antti Poso" > Date: Thu, 17 Oct 1996 14:30:05 WET-2WET > > We are doing some quite long simulations (1 ns) with small organic > compounds (MM+). We monitor 2 torsion angle. To our suprise the value > for those 2 torsion will stabilize after 0.5 to 0.8 ns simulations > (several cases). This means that those values wary only +- 0.5 degree. > The same happens to all energetics and temperature. Allthough we want > to have constant temperature, but not that constant... Remember that the point of molecular dynamics is to allow for inter- conversion of kinetic and potential energy. Optionally, if you specify a constant-temperature bath, the kinetic energies of the atoms will be adjusted periodically. But if you start out with not very much energy in the system (potential energy from a strained geometry, plus kinetic energy from atomic velocities) and no constant temperature, or have a low constant temperature specified, the system will settle out with not much motion. To increase the amount of torsion: Include other molecules (for example, water) to collide with your system and transfer energy to it; Make sure that you have not specified "Restart", which means that the atomic velocities are not adjusted to the specified temperature; Make sure that if you have specified a constant temperature bath, the temperature is high enough. Since I do not know what kind of structures you are modelling, and I do not know what the barriers are for their torsional twisting, I cannot tell what kind of torsions should be expected. Joel ------------ Joel Polowin, Ph.D. Manager, Scientific Support Email to: polowin@hyper.com WWW: http://www.hyper.com/ Hypercube Inc, 419 Phillip St, Waterloo, Ont, Canada N2L 3X2 (519)725-4040 Info requests to: info@hyper.com Support questions to: support@hyper.com Email group: Send "subscribe hyperchem" (or "unsubscribe hyperchem") to hyperchem-request@hyper.com; please do not send such administrative messages to the group itself. _______________________________________________________________________________ From owner-hyperchem@hyper.com Sat Oct 19 14:27:03 1996 From: "williams" To: Subject: Problems running/loading HyperChem 4.5 Date: Sat, 19 Oct 1996 13:00:25 -0400 I have been having some problems running AM1 and ZINDO calculations under 4.5. The program will accept the structure I input, will run under geometry optimization through either 1 "point" up to 4 "points", then locks at "Iteration=500". This seems to occur for even the simplest structures; even an unoptimized ethane locks up. I thought this problem might be due to a corrupted file on the hard disk, so I though the next logical thing to do would be erase the existing files and re-install the program. On re-installation, the first three 4.5 floppies load ok using the install program, but the fourth will not load "NDO.exe". Upon examining this disk using ScanDisk, I get back error messages that say the original factory disk has numerous bad clusters, and that the original file is probably lost. Now I am totally dead in the water. Advice? Brian Williams Dept. Chemistry Bucknell University (717) 524-1143 fax (717) 524-1739 _______________________________________________________________________________ From owner-hyperchem@hyper.com Sat Oct 19 16:59:12 1996 Date: Sat, 19 Oct 1996 14:23:54 +0100 From: dan@tom.fos.su.se (Dan Bergman) To: hyperchem@hyper.com Subject: Drift velocity Hello I have the following problem: When I do a molecular dynamics simulation of a simple argon liquid, and add the velocities of all molecules together they dont average to zero!! There seem to be a net drift velocity in the simulation box. How can i eliminate this? With a net drift velocity, the temperature calculations can be wrong. regards Dan Bergman _______________________________________________________________________________ From owner-hyperchem@hyper.com Mon Oct 21 10:18:28 1996 From: "Udo Haberl" To: hyperchem@hyper.com Date: Mon, 21 Oct 1996 11:17:23 GMT+1 Subject: DFT calculations with HyperChem ? Reply-To: haberl@uni-bonn.de Hi, I would be very interested in DFT calculations with hyperchem. When will Density Functional Theory be implemented in Hyperchem? Udo Haberl ------------------------------------------------------------------ Dipl.-Chem. Udo Haberl Institut fuer Organische Chemie und Biochemie Gerhard-Domagk-Str. 1 D-53121 Bonn Germany e-mail: haberl@uni-bonn.de ------------------------------------------------------------------ _______________________________________________________________________________ From owner-hyperchem@hyper.com Mon Oct 21 14:57:36 1996 Date: Mon, 21 Oct 96 11:35:55 -0400 From: polowin (Joel Polowin) To: dan@tom.fos.su.se (Dan Bergman), hyperchem@hyper.com Subject: Re: Drift velocity > Date: Sat, 19 Oct 1996 14:23:54 +0100 > From: dan@tom.fos.su.se (Dan Bergman) > > When I do a molecular dynamics simulation of a simple argon > liquid, and add the velocities of all molecules together they dont > average to zero!! There seem to be a net drift velocity in the simulation > box. How can i eliminate this? With a net drift velocity, the temperature > calculations can be wrong. If you do an MD run on a set of identical atoms, either with no atoms selected or with all atoms selected, with "Restart" turned off and without constant temperature, the average velocity should be 0 (or within round-off error of 0). (The actual rule is that the average *momentum* should be 0, but if all the atoms are the same, the average velocity will also be 0.) You can check this by selecting all the atoms, and using "Setup/Set Velocity" to display the present average velocity of the selected atoms.) If you have constant temperature turned on, the periodic rescaling of velocities may result in a non-0 average velocity. And of course if you have "Restart" on, the atoms will not have their velocities adjusted to a net value of 0. If you are running MD on only part of the system in the workspace, I think that all of the velocities cannot be adjusted properly. Joel ------------ Joel Polowin, Ph.D. Manager, Scientific Support Email to: polowin@hyper.com WWW: http://www.hyper.com/ Hypercube Inc, 419 Phillip St, Waterloo, Ont, Canada N2L 3X2 (519)725-4040 Info requests to: info@hyper.com Support questions to: support@hyper.com Email group: Send "subscribe hyperchem" (or "unsubscribe hyperchem") to hyperchem-request@hyper.com; please do not send such administrative messages to the group itself. _______________________________________________________________________________ From owner-hyperchem@hyper.com Mon Oct 21 15:05:10 1996 Date: Mon, 21 Oct 96 12:20:30 -0400 From: polowin (Joel Polowin) To: "williams" , Subject: Re: Problems running/loading HyperChem 4.5 > From: "williams" > Date: Sat, 19 Oct 1996 13:00:25 -0400 > > I have been having some problems running AM1 and ZINDO calculations under > 4.5. The program will accept the structure I input, will run under > geometry optimization through either 1 "point" up to 4 "points", then locks > at "Iteration=500". This seems to occur for even the simplest structures; > even an unoptimized ethane locks up. Are you using NT 3.51 as your operating system? If so, switching to NT 4 (or to Windows 95) should fix the problem. I have had no other reports of this kind of behaviour except under NT 3.51; if you are using another operating system, I would need more details of the calculation options to try to reproduce the problem. If you are using NT 3.51, I can tell you how to try to work around the problem; the information is also on our WWW page (http://www.hyper.com) under "Product Support" in the Frequently Asked Questions. > I thought this problem might be due > to a corrupted file on the hard disk, so I though the next logical thing to > do would be erase the existing files and re-install the program. On > re-installation, the first three 4.5 floppies load ok using the install > program, but the fourth will not load "NDO.exe". Upon examining this disk > using ScanDisk, I get back error messages that say the original factory > disk has numerous bad clusters, and that the original file is probably > lost. Now I am totally dead in the water. Advice? You would need to replace the disk, or at least that file. Try copying the contents of the disk to a hard drive or to another disk; see if you can find out which file(s) are damaged. Depending on the severity of the problem, we can send you a replacement disk or make the file available to you electronically. (Things would be somewhat more complicated if the problem was with Disk #1, which has the software serial number encoded in it.) Joel ------------ Joel Polowin, Ph.D. Manager, Scientific Support Email to: polowin@hyper.com WWW: http://www.hyper.com/ Hypercube Inc, 419 Phillip St, Waterloo, Ont, Canada N2L 3X2 (519)725-4040 Info requests to: info@hyper.com Support questions to: support@hyper.com Email group: Send "subscribe hyperchem" (or "unsubscribe hyperchem") to hyperchem-request@hyper.com; please do not send such administrative messages to the group itself. _______________________________________________________________________________ From owner-hyperchem@hyper.com Mon Oct 21 19:27:06 1996 Date: Mon, 21 Oct 96 10:50:31 -0400 From: polowin (Joel Polowin) To: hyperchem@hyper.com, haberl@uni-bonn.de Subject: Re: DFT calculations with HyperChem ? > From: "Udo Haberl" > Date: Mon, 21 Oct 1996 11:17:23 GMT+1 > > I would be very interested in DFT calculations with hyperchem. > When will Density Functional Theory be implemented in Hyperchem? Well, not in Release 5. I do not know when it might be implemented. With the Chemist's Developer Kit in HyperChem 5, one could link a DFT program to HyperChem, but DFT is not an included feature. Joel ------------ Joel Polowin, Ph.D. Manager, Scientific Support Email to: polowin@hyper.com WWW: http://www.hyper.com/ Hypercube Inc, 419 Phillip St, Waterloo, Ont, Canada N2L 3X2 (519)725-4040 Info requests to: info@hyper.com Support questions to: support@hyper.com Email group: Send "subscribe hyperchem" (or "unsubscribe hyperchem") to hyperchem-request@hyper.com; please do not send such administrative messages to the group itself. _______________________________________________________________________________ From owner-hyperchem@hyper.com Tue Oct 22 14:03:27 1996 X-Sender: bolger@zygote.hsc.usc.edu Date: Tue, 22 Oct 1996 07:32:04 -0700 To: hyperchem@hyper.com From: "Michael B. Bolger" Subject: Serious problem with ISISDraw 2 and HyperChem 4.5 I think there might be a bug in one of these programs. My entire Win95 session crashes when I start geometry optimization under the following circumstances. I create a 2D structure in ISISDraw 2.0 and copy it to HyperChem 4.5. Add hydrogen and model build works fine. Start the geometry optimization and the whole system stops with an application error. There was no problem with ISIS 1.2 and HyperChem 4.0. Also, there is not problem if the molecule is created by drawing in HyperChem 4.5. Any suggestions. * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * Michael B. Bolger, Ph.D. * Phone: (213) 342-1442 * * USC School of Pharmacy * FAX: (213) 342-1390 * * Dept. of Pharm. Sci. * Internet: bolger@usc.edu * * 1985 Zonal Ave. PSC 700 * * * Los Angeles, CA 90033 * * * U.S.A. * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * _______________________________________________________________________________ From owner-hyperchem@hyper.com Tue Oct 22 16:23:50 1996 From: kfowler@icos.com Date: Tue, 22 Oct 96 09:59:12 PST To: hyperchem@hyper.com Subject: Re: Serious problem with ISISDraw 2 and HyperChem 4.5 Jeez, I thought it was only me. On my new computer I switched from Windows 3.1 to Win95 and Hyperchem 3 to 4.5. I also find that using ConSystant/Chemeleon to paste from ISISDraw 2.0 to Hyperchem causes a complete hosing of my OS upon geometry optimization. I figured there was just something odd about my whole new setup. Kerry Fowler ICOS Corp. kfowler@icos.com ______________________________ Reply Separator _________________________________ Subject: Serious problem with ISISDraw 2 and HyperChem 4.5 Author: bolger (Michael B. Bolger) { NAME: Michael B. Bolger | EMS: INTERNET | MBX: bolger@zygote.hsc.usc.edu } at MCIMAIL Date: 10/22/96 7:32 AM I think there might be a bug in one of these programs. My entire Win95 session crashes when I start geometry optimization under the following circumstances. I create a 2D structure in ISISDraw 2.0 and copy it to HyperChem 4.5. Add hydrogen and model build works fine. Start the geometry optimization and the whole system stops with an application error. There was no problem with ISIS 1.2 and HyperChem 4.0. Also, there is not problem if the molecule is created by drawing in HyperChem 4.5. Any suggestions. * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * Michael B. Bolger, Ph.D. * Phone: (213) 342-1442 * * USC School of Pharmacy * FAX: (213) 342-1390 * * Dept. of Pharm. Sci. * Internet: bolger@usc.edu * * 1985 Zonal Ave. PSC 700 * * * Los Angeles, CA 90033 * * * U.S.A. * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * * _______________________________________________________________________________ From owner-hyperchem@hyper.com Tue Oct 22 16:40:04 1996 Date: Tue, 22 Oct 96 14:22:10 -0400 From: polowin (Joel Polowin) To: "Michael B. Bolger" , hyperchem@hyper.com Subject: Re: Serious problem with ISISDraw 2 and HyperChem 4.5 Reply-To: polowin@hyper.com > Date: Tue, 22 Oct 1996 07:32:04 -0700 > From: "Michael B. Bolger" > > I think there might be a bug in one of these programs. My entire Win95 > session crashes when I start geometry optimization under the following > circumstances. > I create a 2D structure in ISISDraw 2.0 and copy it to HyperChem 4.5. Add > hydrogen and model build works fine. Start the geometry optimization and > the whole system stops with an application error. > There was no problem with ISIS 1.2 and HyperChem 4.0. Also, there is not > problem if the molecule is created by drawing in HyperChem 4.5. I have had one other report, with no follow-up information, of unspecified trouble with bringing an ISISDraw 2 structure into HyperChem. I am surprised -- if you can Add H/Model Build without trouble, you should be able to run any of the optimizers. What kind of optimization calculation were you running? Is there any chance that there were several atoms at exactly the same locations? -- that might be a problem for some of the calculation methods. Is the number of atoms in the structure (as determined, for example, by selecting all atoms) the number that you expect? If you can send me a copy of the structure by E-mail as a HIN file, after Add H/Model Build but before trying the optimizer, I'll take a look at it. Joel ------------ Joel Polowin, Ph.D. Manager, Scientific Support Email to: polowin@hyper.com WWW: http://www.hyper.com/ Hypercube Inc, 419 Phillip St, Waterloo, Ont, Canada N2L 3X2 (519)725-4040 Info requests to: info@hyper.com Support questions to: support@hyper.com Email group: Send "subscribe hyperchem" (or "unsubscribe hyperchem") to hyperchem-request@hyper.com; please do not send such administrative messages to the group itself. _______________________________________________________________________________ From owner-hyperchem@hyper.com Tue Oct 22 19:15:30 1996 Date: Tue, 22 Oct 1996 12:59:27 +0100 From: mkh100@york.ac.uk (MK Hodgson) To: hyperchem@hyper.com Subject: DNA Counter ions, I am trying to simulate a drug/DNA complex in TIP3 water using periodic boundaries. I have experimented with scaling the DNA phosphate charges to represent counter ions, but have met with little success. Using explicit counter ions leads me to the question of how they are placed. I am also surprised at the speed at which the system comes to equlibrium given the long relaxation times a solvated ion is expected to have. I would appreciate any comments,critasisms or sugestions people may have. Having minimised the system I hope to examine the binding of the drug and DNA. Does anyone have any suggestions as to the best way of doing this. Is it possible to get hyperchem to display specific contributions to the overall energy, for example the electrostatic interaction energy between drug and dna. -Michael Hodgson ,----- Michael Hodgson < mkh100@unix.york.ac.uk > ,----- | __/ | __/ l_F-< A_A_A,-mmmmm--=_ =---<==r==- l_F-< L \ _ _H_H_H_,-------|..|-,_,--| L \ r-i ) I I=L|==L<_ >L|: : : :|..|>|_<>:II r-i ) I \ " / H H H `-------|_ |-' `--' \ " / `---' U U U `---' (http://www.york.ac.uk/~mkh100) If it's green and wiggles, it's biology. If it stinks, it's chemistry. If it doesn't work, it's Physics. _______________________________________________________________________________ From owner-hyperchem@hyper.com Tue Oct 22 19:40:07 1996 From: "Mei, George" To: "'Hypercube'" Subject: Thermal Degradation by Molecular Dynamics Date: Tue, 22 Oct 1996 16:20:00 -0500 Hello: I am trying to model thermal degradation of small molecules by using molecular dynamics. The molecules of interest are energetic materials which should release energy when the proper temperatures are reached. To test the theory, I perform calculations on materials which avail us plenty of experimental data (in terms of activation energy and frequency factor) to be compared to the calculations. The advantage of modeling an energetic compound is that it is a uni-molecular reaction which, in principle, can be modeled by simply heating the molecule up until a bond (or few) is broken. The test reaction has an activation energy of 50 kcal/mole with a frequency factor of 10 ^ 24.5 per second. When observed under DSC, the material was fully decomposed at 433 K. Using Arrhenius equation, I estimated that the temperature for an event to occur within 1 ps should be at 875 K, and within 0.1 ps at 950 K, and within 0.01ps at 1041 K. But when I ran it at 2,000 K for 5 ps, nothing had happened! And followed by an additional 5 ps, still nothing. Then, I re-estimate the time for an event to occur, assuming an activation energy of 90 kcal/mole (the worst case). they turned out to be within 1 ps at 1574 K, within 0.1 ps at 1711 K, and within 0.01 ps at 1873 K. So, at 2,000 K, I would expect the molecule to break up well within the 10 ps time period. Why did it not happen? Of course, the structure had been optimized before I started. I was using PM3/UHF, 0.001 ps per step, constant temperature bath, and 0.1 ps heating period with 100 K step. I wasn't able to follow the kinetic energy and potential energy separately yet for I was running hyper under NT 3.51 with a glitch not allowing me to do just so. Nevertheless, I was able to roughly estimate the potential energy from time to time using the temperature and total energy displayed at the bottom of the screen. They were running anywhere from 70 to 120 kcal over the experimental activation energy, and not breaking the bond! The reaction did proceed, only after I raise the temperature to 5,000 K and after roughly half ps. My concern over using a high temperature, however, is that the reaction mechanism may shift on me for more bonds are made available to be broken by the higher overall energy. Was I on the right track? Any thought will be appreciated. _______________________________________________________________________________ From owner-hyperchem@hyper.com Wed Oct 23 14:44:33 1996 Date: Wed, 23 Oct 1996 10:41:43 -0400 From: Joel Polowin To: hyperchem@hyper.com Subject: Hypercube E-mail We have been having some trouble with our E-mail since our service provider changed addresses almost three weeks ago. It seems that mail that is being sent to addresses at "hyper.hyper.com" is not getting to us; mail sent to addresses at "hyper.com" is arriving safely. Unfortunately, my E-mail is being sent from an apparent address on "hyper.hyper.com", so direct replies to my messages are being lost. I am trying to work around this by adding an explicit "Reply-to" line to my mail, but when I send to this mailing list, that line is removed, leaving only the problem address. We are working on this problem, but for now, if you need to send a message to me, please make sure that it is being sent to "polowin@hyper.com". (I am sending this message from another machine, which I believe is able to get the return address right.) I would very much appreciate if anyone who has had the recently-reported problems with Isis Draw 2 could send me, by E-mail to polowin@hyper.com, a small sample HIN-format file of a structure that has been imported from Isis Draw, and which will cause HyperChem to fail on an optimization. Joel ------------ Joel Polowin, Ph.D. Manager, Scientific Support Email to: polowin@hyper.com WWW: http://www.hyper.com/ Hypercube Inc, 419 Phillip St, Waterloo, Ont, Canada N2L 3X2 (519)725-4040 Info requests to: info@hyper.com Support questions to: support@hyper.com Email group: Send "subscribe hyperchem" (or "unsubscribe hyperchem") to hyperchem-request@hyper.com; please do not send such administrative messages to the group itself. _______________________________________________________________________________ From owner-hyperchem@hyper.com Wed Oct 23 15:41:34 1996 Date: Wed, 23 Oct 96 12:08:16 -0400 From: polowin (Joel Polowin) To: "Mei, George" , "'Hypercube'" Subject: Re: Thermal Degradation by Molecular Dynamics Reply-To: polowin@hyper.com > From: "Mei, George" > Subject: Thermal Degradation by Molecular Dynamics > > The test reaction has an activation energy of 50 kcal/mole with a frequency > factor of 10 ^ 24.5 per second. When observed under DSC, the material was > fully decomposed at 433 K. Using Arrhenius equation, I estimated that the > temperature for an event to occur within 1 ps should be at 875 K, and within > 0.1 ps at 950 K, and within 0.01ps at 1041 K. But when I ran it at 2,000 K > for 5 ps, nothing had happened! And followed by an additional 5 ps, still > nothing. Then, I re-estimate the time for an event to occur, assuming an > activation energy of 90 kcal/mole (the worst case). they turned out to be > within 1 ps at 1574 K, within 0.1 ps at 1711 K, and within 0.01 ps at 1873 > K. So, at 2,000 K, I would expect the molecule to break up well within the > 10 ps time period. Why did it not happen? > > Of course, the structure had been optimized before I started. I was using > PM3/UHF, 0.001 ps per step, constant temperature bath, and 0.1 ps heating > period with 100 K step. It seems to me that there are a couple of problems here. First, you're starting with an optimized structure, which is *not* very appropriate when you're trying to model a system at an elevated temperature. If you equilibrate the system at an elevated temperature so that the structure has some internal potential energy, you might have better luck. Second, if you are trying to model a single molecule in isolation, that is a poor representation of the real system. In reality, there are many molecules, which collide and exchange energy. It seems to me that it is the statistical distribution of energies in the real world, which leads to some molecules having a lot of internal energy and some having very little at any particular moment, that allows the high-energy fraction to dissociate. If you want to get a single molecule to dissociate, you will need to get it to represent the high-energy end of that distribution, by applying a high temperature. Or so it seems to me. > I wasn't able to follow the kinetic energy and > potential energy separately yet for I was running hyper under NT 3.51 with a > glitch not allowing me to do just so. As has been noted recently, if you switch to NT 4 (or to Windows 95 or to Windows 3.1x), this problem should go away. Joel ------------ Joel Polowin, Ph.D. Manager, Scientific Support Email to: polowin@hyper.com WWW: http://www.hyper.com/ Hypercube Inc, 419 Phillip St, Waterloo, Ont, Canada N2L 3X2 (519)725-4040 Info requests to: info@hyper.com Support questions to: support@hyper.com Email group: Send "subscribe hyperchem" (or "unsubscribe hyperchem") to hyperchem-request@hyper.com; please do not send such administrative messages to the group itself. _______________________________________________________________________________ From owner-hyperchem@hyper.com Wed Oct 23 19:57:15 1996 Date: Wed, 23 Oct 96 11:59:15 -0400 From: polowin (Joel Polowin) To: mkh100@york.ac.uk (MK Hodgson) Subject: Re: DNA Counter ions, Cc: hyperchem@hyper.com Reply-To: polowin@hyper.com > Date: Tue, 22 Oct 1996 12:59:27 +0100 > From: mkh100@york.ac.uk (MK Hodgson) > > Having minimised the system I hope to examine the binding of the > drug and DNA. Does anyone have any suggestions as to the best way of > doing this. Is it possible to get hyperchem to display specific > contributions to the overall energy, for example the electrostatic > interaction energy between drug and dna. There are a couple of possible approaches. If you are using the MM+ force field, then after your system has been optimized you can run a single-point calculation, with a log file open and "Mechanics Print Level" set to 1 or higher. This will give you a detailed list of all of the electrostatic interactions in the log file. You can add together the ones that are relevant. If you are using one of the other force fields, save your optimized structure in a file. Delete everything but the drug and the DNA, and save this in a temporary file. Open a log file, and do a single-point calculation. The summary in the log file includes the total of all of the electrostatic interactions. Delete the drug, and do another single- point calculation; that will give you the electrostatic interactions within the drug. Bring back your temporary file and delete the DNA, and do another single-point calculation to get the electrostatic interactions within the DNA. The difference between the total of the electrostatic interactions of both together, and the sum of the interactions of the two separately, will give you the electrostatic interaction energy between the two. If you select part of the system in a workspace and run a single-point calculation, only the interactions that include the selected atoms will be included in the totals in the log file. Again, by selecting different parts of the system in the workspace, and comparing, adding, and subtracting their electrostatic energies, you can determine the interactions between specific parts. Joel ------------ Joel Polowin, Ph.D. Manager, Scientific Support Email to: polowin@hyper.com WWW: http://www.hyper.com/ Hypercube Inc, 419 Phillip St, Waterloo, Ont, Canada N2L 3X2 (519)725-4040 Info requests to: info@hyper.com Support questions to: support@hyper.com Email group: Send "subscribe hyperchem" (or "unsubscribe hyperchem") to hyperchem-request@hyper.com; please do not send such administrative messages to the group itself. _______________________________________________________________________________ From owner-hyperchem@hyper.com Thu Oct 24 14:41:53 1996 Date: Thu, 24 Oct 96 11:41:33 -0400 From: polowin (Joel Polowin) To: hyperchem@hyper.com Subject: Re: Isis 2.0 and HyperChem 4.5 Reply-To: polowin@hyper.com Patrick Grant informs me that one workaround for the problem is to save the imported Isis/Draw structure in the HIN format, and then reload it. The saved and loaded structure has had the problem -- whatever it is -- corrected; one can then run the geometry optimization. Michael Bolger tells me that if the molecular mechanics module is already loaded (for example, from a previous calculation), the system does not crash. The crash occurs when a new molecular mechanics module is loaded. I am still waiting to hear back from MDL, who make Isis/Draw, on this matter. Joel ------------ Joel Polowin, Ph.D. Manager, Scientific Support Email to: polowin@hyper.com WWW: http://www.hyper.com/ Hypercube Inc, 419 Phillip St, Waterloo, Ont, Canada N2L 3X2 (519)725-4040 Info requests to: info@hyper.com Support questions to: support@hyper.com Email group: Send "subscribe hyperchem" (or "unsubscribe hyperchem") to hyperchem-request@hyper.com; please do not send such administrative messages to the group itself. _______________________________________________________________________________ From owner-hyperchem@hyper.com Fri Oct 25 15:51:18 1996 To: hyperchem@hyper.com From: "Nancy Mills" Organization: Trinity University Date: Fri, 25 Oct 1996 11:24:37 -0500 (CDT) Subject: ZINDO/S parameters for chlorine Reply-To: nmills@Trinity.Edu Hi, Does anyone know of the ZINDO/S parameter for chlorine? As a newcomer to ZINDO/S calculations, is there perhaps a review article which has compiled a set of "unusual" ZINDO/S parameters? Thanks much for your help. If there is interest, I will summarize to the net. Nancy Nancy Mills Department of Chemistry Trinity University San Antonio, TX 78212-7200 210-736-7317 phone 210-736-7569 fax nmills@trinity.edu _______________________________________________________________________________ From owner-hyperchem@hyper.com Fri Oct 25 17:26:49 1996 From: "Mei, George" To: "'Hypercube'" Subject: Direct SCF vs. Indirect SCF Calculations Date: Fri, 25 Oct 1996 13:42:00 -0500 Encoding: 16 TEXT Hello: I was playing with various parameters of ab initio setup in order to optimize the speed on my system. One thing I noticed was that the direct SCF and the indirect SCF calculations sometimes gave different results. I don't know, its just my computer or something else. For example, I was following the example in the reference to calculate the energy of hydronium ion H3O+. With the direct SCF checked 'off', I obtained the same result as listed in the reference: -47270.57. However, after checking the direct SCF 'on', I obtained a value of -42920.3. with everything else kept the same. More peculiar was when I try to optimize it with the direct SCF 'on', the structure actually was going the other way - de-stabilizing rather than stabilizing, and seemed not wanting to converge. Any thoughts? My computer is a Gateway P5-200 with 64M ram. I am using version 4.5. _______________________________________________________________________________ From owner-hyperchem@hyper.com Fri Oct 25 17:42:49 1996 Date: Fri, 25 Oct 96 15:26:19 -0400 From: polowin (Joel Polowin) To: hyperchem@hyper.com Subject: More HP driver trouble For the last couple of weeks, I have had intermittent trouble in getting HyperChem on *my* machine to communicate with the hardware lock, and I have finally isolated the cause. It turns out that the printer "Status Monitor" that was automatically installed when I added the driver for the HP 1600CM to my machine sometimes blocks HyperChem's communications with the parallel port. This happened even though the printer was on another machine on our network, not on my machine, and even after I had shut down the "Status Monitor". To get rid of the stupid thing, I had to edit my WIN.INI file manually; the monitor started automatically with Windows, and this "feature" could not be disabled even with the monitor setting that was supposed to do just that. After I rebooted Windows, everything was fine. We knew that some HP (and other) drivers cause communications trouble, but we at least thought that this affected only ports that have printers attached to them. On the positive side, our E-mail problems with mail sent to addresses at 'hyper.hyper.com' have been corrected; all the mail seems to be arriving safely again. Have a nice weekend, everyone. Joel ------------ Joel Polowin, Ph.D. Manager, Scientific Support Email to: polowin@hyper.com WWW: http://www.hyper.com/ Hypercube Inc, 419 Phillip St, Waterloo, Ont, Canada N2L 3X2 (519)725-4040 Info requests to: info@hyper.com Support questions to: support@hyper.com Email group: Send "subscribe hyperchem" (or "unsubscribe hyperchem") to hyperchem-request@hyper.com; please do not send such administrative messages to the group itself. _______________________________________________________________________________ From owner-hyperchem@hyper.com Mon Oct 28 14:24:18 1996 Organization: CNR - SIAM - Milano (Italy) X-Sender: ico@siam.mi.cnr.it Date: Mon, 28 Oct 1996 15:50:04 +0100 To: hyperchem@hyper.com From: Istituto di Chimica degli Ormoni Subject: cyclodextrin Hi netters, I a begginer in this subject. I am lookin for cristallographic data of cyclodextrins. Is there anyone that can give me some sugestions on how to get it. Marcella Chiari _______________________________________________________________________________ From owner-hyperchem@hyper.com Mon Oct 28 17:53:05 1996 Date: Mon, 28 Oct 96 15:01:54 -0500 From: polowin (Joel Polowin) To: "Mei, George" , "'Hypercube'" Subject: Re: Direct SCF vs. Indirect SCF Calculations > From: "Mei, George" > Date: Fri, 25 Oct 1996 13:42:00 -0500 > > I was playing with various parameters of ab initio setup in order to > optimize the speed on my system. One thing I noticed was that the direct SCF > and the indirect SCF calculations sometimes gave different results. I don't > know, its just my computer or something else. For example, I was following > the example in the reference to calculate the energy of hydronium ion H3O+. > With the direct SCF checked 'off', I obtained the same result as listed in > the reference: -47270.57. However, after checking the direct SCF 'on', I > obtained a value of -42920.3. with everything else kept the same. More > peculiar was when I try to optimize it with the direct SCF 'on', the > structure actually was going the other way - de-stabilizing rather than > stabilizing, and seemed not wanting to converge. Any thoughts? You have just found a new bug for us. It affects results if one specifies a direct SCF calculation, which does not use the integral buffer, and a buffer size that would be larger than the total required storage if any storage were actually being used. (This has not been noticed until now because usually, if one is doing a calculation with such small storage requirements, one would not do it in direct SCF mode.) The work-around is to specify a small buffer size -- say, 100 words -- if one is doing a direct SCF calculation. We will probably do a patch for GAUSS.EXE to correct this problem after HyperChem 5 for NT has been completed. (HyperChem 5 for Windows 95 went for manufacturing last week, and it should be starting to ship in another week or so.) Joel ------------ Joel Polowin, Ph.D. Manager, Scientific Support Email to: polowin@hyper.com WWW: http://www.hyper.com/ Hypercube Inc, 419 Phillip St, Waterloo, Ont, Canada N2L 3X2 (519)725-4040 Info requests to: info@hyper.com Support questions to: support@hyper.com Email group: Send "subscribe hyperchem" (or "unsubscribe hyperchem") to hyperchem-request@hyper.com; please do not send such administrative messages to the group itself. _______________________________________________________________________________ From owner-hyperchem@hyper.com Tue Oct 29 07:37:11 1996 Sender: neuber@exp.bessy.de Date: Tue, 29 Oct 1996 10:31:21 MET From: Michael Neuber To: hyperchem@hyper.com Subject: qualified convergence criterion ? Hello folks! Hello Joel! When calculating semi-empiric calculations with HyperChem the criterion to stop the run can be either the convergence limit or an iteration limit. (Same with ab-initio I learn in these days...) Every now and then I notice that the convergence limit is reached by accident, i.e. the sequence of energy differences in a log-file (with still rather high values) is broken by one very small energy step (accidentel I suppose), and consequently the calculation is stopped. But, the calculation indeed has not reached the energy minimum which can be seen in a comparison with other data points, say in a potential energy surface. A repetition of the calculation is necessary, and meanwhile I decrease the convergence limit in order to be sure that such accidental stops don't occur too often. Sure it is annoying when a long over-night run did not produce any meaningful results because of such premature stops. My question: Is there any trick known to introduce some 'qualified' convergence limit, say, a criterion which tests whether such small energy steps occur (let's say:) three times in sequence, and if so, first then the iteration will be stopped ? I am thinking about programming a corresponding script, but, during a calculation run there won't be any chance neither to increase the iteration limit on the fly or something like that nor to go on using calculation results obtained up to that point where control is returned to a script file in order to manipulate anything from outside..... Hm. If there isn't any such trick, perhaps this is included in HC Vs 5.0 ? Greetings from Germany Michael --------------------------------------------------------------------------- Dr. Michael Neuber Universitaet Heidelberg c/o BESSY Lentzeallee 100 D-14195 Berlin e-mail neuber@exp.bessy.de --------------------------------------------------------------------------- _______________________________________________________________________________ From owner-hyperchem@hyper.com Tue Oct 29 12:31:27 1996 From: "Alison Lennon" To: Subject: Predicting uv-vis spectra Date: Tue, 29 Oct 1996 20:20:31 +1100 Hi everyone, I am trying to use HyperChem to predict the uv-vis spectra of organic (and some organometallic) chromophores. I have found a low of variation in the predicted transitions between the different semi-empirical methods. For example, I'm currently looking at the spectra of some substituted quinones (and their reduced equivalents). If I optimise the structures with ZINDO/1 and then perform a single point calculation with ZINDO/S I can get results that are reasonably close to experimental values in most cases. PM3, however, results in predicted transitions which can be up to 100 nm different from an experimental value. Which semi-empirical method(s) are best for this purpose? Does the preferred method depend on the type of chromophore being investigated? And, how accurately can I expect to predict uv-vis spectra using these methods? Can someone suggest any references that might be useful to me. Thanks for your help. Alison. -------------- Alison Lennon email: alison@silverbrook.com.au Silverbrook Research Pty Ltd, 214 Catherine St, Leichhardt, NSW, 2040, Australia. Phone: +61 2 9550 9777 Fax: +61 2 9550 9201 _______________________________________________________________________________ From owner-hyperchem@hyper.com Tue Oct 29 13:19:24 1996 From: Henryette Roth Subject: Saving IR-spectra To: hyperchem@hyper.com Date: Tue, 29 Oct 1996 15:12:55 +0100 (MET) Cc: roth@organik.uni-erlangen.de (Henryette Roth) OFFICE MEMO Date: October 29th, 1996 Saving IR-spectra Hi there, is there a possibility to save a IR-spectra under a different name than the default. The idea is to calculate a series of molecules and their spectra with the help of a script file or an Excel-script and save the spectra under the same name as the HIN files. Any idea ? Thank you very much in advance, Henryette ******************************************************************************** * Henryette Roth * * * Computer Chemie Centrum * All opinions expressed by this user * * Inst. f. Org. Chemie I * are well-reasoned and insightful. * * Universitaet Erlangen * Needless to say, they are not those * * Naegelsbachstr. 25 * of his advisor or university. * * 91052 Erlangen * * ******************************************************************************** * roth@organik.uni-erlangen.de * ******************************************************************************** _______________________________________________________________________________ From owner-hyperchem@hyper.com Tue Oct 29 15:34:27 1996 Date: Tue, 29 Oct 96 11:29:47 -0500 From: polowin (Joel Polowin) To: Michael Neuber , hyperchem@hyper.com Subject: Re: qualified convergence criterion ? > Date: Tue, 29 Oct 1996 10:31:21 MET > From: Michael Neuber > > Every now and then I notice that the convergence limit > is reached by accident, i.e. the sequence of energy differences in a > log-file (with still rather high values) is broken by one very small > energy step (accidentel I suppose), and consequently the calculation > is stopped. > But, the calculation indeed has not reached the energy minimum which can > be seen in a comparison with other data points, say in a potential > energy surface. A repetition of the calculation is necessary, and > meanwhile I decrease the convergence limit in order to be sure that such > accidental stops don't occur too often. > Sure it is annoying when a long over-night run did not produce any > meaningful results because of such premature stops. I think that there is no simple solution to this. Sometimes one will find a shallow local minimum in the potential energy surface, and depending on the "direction" of travel on the surface and the amount of "momentum" -- the step size for atomic motions in the optimization -- the optimization can get stuck in the local minimum. One way to work around this would be to specify a small convergence limit at the beginning; if you come back the following morning and find that the calculation is still running, having reached a gradient larger than the small limit but smaller than you really need, you could then just halt the calculation. > Is there any trick known to introduce some 'qualified' convergence > limit, say, a criterion which tests whether such small energy steps > occur (let's say:) three times in sequence, and if so, first then > the iteration will be stopped ? > I am thinking about programming a corresponding script, but, during a > calculation run there won't be any chance > neither to increase the iteration limit on the fly or something like > that nor to go on using calculation results obtained up to that point > where control is returned to a script file in order to manipulate anything > from outside..... > Hm. > If there isn't any such trick, perhaps this is included in HC Vs 5.0 ? I think that you would need to use a controller which is a bit "smarter" than a script file for release 4 or 4.5, with more flexibility and more control options. Something like Visual Basic or Excel would do. With HyperChem 5 you could write a script using Tcl/Tk, which offers the power that the earlier scripting language did not. Joel ------------ Joel Polowin, Ph.D. Manager, Scientific Support Email to: polowin@hyper.com WWW: http://www.hyper.com/ Hypercube Inc, 419 Phillip St, Waterloo, Ont, Canada N2L 3X2 (519)725-4040 Info requests to: info@hyper.com Support questions to: support@hyper.com Email group: Send "subscribe hyperchem" (or "unsubscribe hyperchem") to hyperchem-request@hyper.com; please do not send such administrative messages to the group itself. _______________________________________________________________________________ From owner-hyperchem@hyper.com Tue Oct 29 15:44:05 1996 Date: Tue, 29 Oct 96 13:29:38 -0500 From: polowin (Joel Polowin) To: Henryette Roth , hyperchem@hyper.com Subject: Re: Saving IR-spectra Cc: roth@organik.uni-erlangen.de (Henryette Roth) > From: Henryette Roth > Date: Tue, 29 Oct 1996 15:12:55 +0100 (MET) > > is there a possibility to save a IR-spectra under a different name than the > default. The idea is to calculate a series of molecules and their spectra > with the help of a script file or an Excel-script and save the spectra under > the same name as the HIN files. If you look at the SAVEIR.SCR file in a text editor, you will find the line that specifies the name of the file that is created: omsgs-to-file \hyper\chem_ir.scr Just change that line in the script file to specify some other path and/or file name. HyperChem 5's import/export function allows one to save and retrieve IR and UV/vis spectra using a menu function instead of having to run a script. Joel ------------ Joel Polowin, Ph.D. Manager, Scientific Support Email to: polowin@hyper.com WWW: http://www.hyper.com/ Hypercube Inc, 419 Phillip St, Waterloo, Ont, Canada N2L 3X2 (519)725-4040 Info requests to: info@hyper.com Support questions to: support@hyper.com Email group: Send "subscribe hyperchem" (or "unsubscribe hyperchem") to hyperchem-request@hyper.com; please do not send such administrative messages to the group itself. _______________________________________________________________________________ From owner-hyperchem@hyper.com Tue Oct 29 16:10:58 1996 Date: 29 Oct 1996 18:08:16 Z From: "Matthew Harbowy" To: "hyperchem@hyper.com" (Return requested) (Receipt notification requested), "alison@silverbrook.com.au" (Return requested) (Receipt notification requested) Subject: Re: Predicting uv-vis spectra --PART.BOUNDARY.x500test.0294.32764896.0001 Content-Disposition: inline RFC-822-Headers: MIME-Version: 1.0 --PART.BOUNDARY.x500test.0294.32764896.0001 Content-Disposition: inline The reason that PM3 is good for energies and not for spectra, and the opposite for ZINDO, is that one or the other is optimized for that application. Each is "fudged" to get good numbers, hence the name "semi" empirical. In either case, you should not be attempting to reproduce reality with theory. You should be looking for informative differences: shifts of absorption bands, or their appearance or disappearance. Take that 100nm difference as the error in absolute positioning to give yourself a good idea over what range you expect to find changes. You can probably do a better job of predicting spectra with a table of common absorptions, and a pencil and paper, for most molecules. The question that I am wondering about is: why do you want to predict the spectra? For different applications there are different approaches, and there are no set rules for making decisionas in a vacuum. matt ______________________________ Reply Separator _________________________________ Subject: Predicting uv-vis spectra Author: alison@silverbrook.com.au at INTERNET Date: 10/29/96 9:33 AM Hi everyone, I am trying to use HyperChem to predict the uv-vis spectra of organic (and some organometallic) chromophores. I have found a low of variation in the predicted transitions between the different semi-empirical methods. For example, I'm currently looking at the spectra of some substituted quinones (and their reduced equivalents). If I optimise the structures with ZINDO/1 and then perform a single point calculation with ZINDO/S I can get results that are reasonably close to experimental values in most cases. PM3, however, results in predicted transitions which can be up to 100 nm different from an experimental value. Which semi-empirical method(s) are best for this purpose? Does the preferred method depend on the type of chromophore being investigated? And, how accurately can I expect to predict uv-vis spectra using these methods? Can someone suggest any references that might be useful to me. Thanks for your help. Alison. -------------- Alison Lennon email: alison@silverbrook.com.au Silverbrook Research Pty Ltd, 214 Catherine St, Leichhardt, NSW, 2040, Australia. Phone: +61 2 9550 9777 Fax: +61 2 9550 9201 --PART.BOUNDARY.x500test.0294.32764896.0001-- _______________________________________________________________________________ From owner-hyperchem@hyper.com Wed Oct 30 14:07:20 1996 Date: Wed, 30 Oct 96 11:09:10 -0500 From: polowin (Joel Polowin) To: "Walter Zielinski (314) 539-2174" , matthew.harbowy@Unilever.COM Subject: Re: ir spectra query Cc: hyperchem@hyper.com, alison@silverbrook.com.au > Date: Wed, 30 Oct 1996 09:48:57 -0600 (CST) > From: "Walter Zielinski (314) 539-2174" > > Matt, > > I'd like to offer two comments to your response to Alison's > query on spectra predictions: > > (1) Why should one not attempt to reproduce reality with theory? > Don't theoreticians strive to model reality so one can move from > empiricism to prediction? If I may: one may certainly make the attempt to reproduce reality with theory. However, in this case, the currently-available theory (and computer technology) are not yet able to give highly-accurate reproductions of reality. One should not *expect* high accuracy yet from spectral calculations; one may expect reasonable accuracy and consistent trends of behaviour, which themselves give useful information. Joel ------------ Joel Polowin, Ph.D. Manager, Scientific Support Email to: polowin@hyper.com WWW: http://www.hyper.com/ Hypercube Inc, 419 Phillip St, Waterloo, Ont, Canada N2L 3X2 (519)725-4040 Info requests to: info@hyper.com Support questions to: support@hyper.com Email group: Send "subscribe hyperchem" (or "unsubscribe hyperchem") to hyperchem-request@hyper.com; please do not send such administrative messages to the group itself. _______________________________________________________________________________ From owner-hyperchem@hyper.com Wed Oct 30 15:31:54 1996 Date: Wed, 30 Oct 1996 11:20:52 -0600 (CST) From: "Walter Zielinski (314) 539-2174" Subject: Re: ir spectra query To: "polowin@hyper.hyper.COM" Cc: "Walter Zielinski (314) 539-2174" , "matthew.harbowy@Unilever.COM" , "hyperchem@hyper.COM" , "alison@silverbrook.com.AU" Posting-Date: Wed, 30 Oct 1996 11:31:00 -0600 (CST) Joel, I concur - high accuracy would (and should) not be expected at the present stage of theory and computer capabilities (i.e., the *ultimate* objective might never be attained in real practice). Nonetheless, the strides that have been made have been truly phenomenal. This however, is what drives the ship. The attainment of *reasonable* accuracies in spectral predictions however - I think - should and would be applauded! Walt My apologies to you'all - I goofed - I meant to indicate uv/vis not ir in the memo title (but the argument's the same). _______________________________________________________________________________ From owner-hyperchem@hyper.com Thu Oct 31 19:49:24 1996 From: "Alison Lennon" To: Subject: Re: Predicting UV-vis spectra of substituted quinones Date: Fri, 1 Nov 1996 08:50:10 +1100 Thanks everyone for your comments. I think it's been a useful discussion - I've learned something! Cheers, Alison. -------------- Alison Lennon email: alison@silverbrook.com.au Silverbrook Research Pty Ltd, 214 Catherine St, Leichhardt, NSW, 2040, Australia. Phone: +61 2 9550 9777 Fax: +61 2 9550 9201