From - Thu Apr 10 12:09:01 1997 From: barclay@hyper.com Message-Id: <199704101202.IAA06004@relay.hyper.com> Comments: Authenticated sender is To: hyperchem@hyper.com Date: Thu, 10 Apr 1997 07:59:15 +0000 MIME-Version: 1.0 Content-type: text/plain; charset=US-ASCII Content-transfer-encoding: 7BIT Subject: new location for Hypercube Reply-to: support@hyper.com Priority: normal X-mailer: Pegasus Mail for Win32 (v2.52) X-Mozilla-Status: 0001 Content-Length: 5073 PRESS RELEASE April 14, 1997 FOR IMMEDIATE RELEASE Contact: LouAnn Bergeron (352) 371-7744 Hypercube, Inc. Now in the USA. NEW LOCATION FOR HYPERCUBE, INC. SAN FRANCISCO, CALIFORNIA - At the National Meeting of the American Chemical Society (ACS) in San Francisco, Hypercube, Inc., a Canadian Corporation and the developers of the popular molecular modeling package, HyperChem, announced that it had made a major move into the US by opening a new office in Gainesville, Florida. The company has purchased a 2.5-acre, 12,000 square foot site in downtown Gainesville as the location of its new office. The head office for Hypercube, Inc. remains in Waterloo, Ontario, Canada but the Gainesville office is expected to grow rapidly and eventually become the principal business location for the company. "Our principal market is, of course, in the US and this new location will allow us to serve our US customers better," according to Mr. Millard Pate, Hypercube's Director of Sales. "The US-Canadian border is not truly a barrier but this new location will make many operations flow smoother when we are dealing with our US customers, distributors, dealers, and all the other American entities that we do business with constantly." Initially, sales, marketing, and product support will all be transferred to the US office, as well as a number of senior management positions. "We have been working on our new US operation for over a year," commented Ms. LouAnn Bergeron, the Office Manager for the new facility. "We are finally open for business in a very comfortable new location. We should be able to serve our customers and friends much better from this new office." Gainesville, Florida is home to the 40,000+ students of the University of Florida as well as the Quantum Theory Project (QTP), the largest group of quantum chemists in the world. In 1996, Money Magazine rated Gainesville as the "Most Livable City in the USA". "We have been looking for a US site for some time," said Dr Neil S. Ostlund, President and CEO of Hypercube. "Gainesville turns out to be an ideal location for the company. Personally, I just love living here - most Canadians love Florida, of course - particularly in January." The new site for Hypercube, Inc. will allow the company to interact with QTP faculty, hire their students, and expand its scientific base. The company expects to have close interactions with many members of the University of Florida community. A new development group will be initiated at the Florida site in addition to the business operations that are now in place there. "Our 12,000 square foot building gives us plenty of space for now and we have a couple of acres to expand into if we need more space", explained Dr. Ostlund. "We are very proud of the new home for our US operation." Shipping and manufacturing will continue to operate out of the Canadian location for the foreseeable future. "Shipping to US sites from Gainesville makes some sense," said Irvine Nichols, Hypercube's Director of Manufacturing, "but the Toronto airport allows us to serve our overseas customers better for the time being." Other facets of Hypercube's business will continue in Canada but the US operation is expected to eventually become the major site for the company. Effective immediately, all sales, marketing, and support questions are to be directed to the US site. The new US address for Hypercube, Inc. is: Hypercube, Inc. Florida Science and Technology Park 1115 NW 4th Street Gainesville, Florida 32601 Tel: (352) 371-7744 or (800)-960-1871 Fax: (352) 371-3662 E-Mail: info@hyper.com Web-Site: http://www.hyper.com Background Hypercube is a privately held scientific software company incorporated in 1985 and headquartered in Waterloo, Ontario, Canada. Its principal product is HyperChem for Intel-based PCs running Microsoft Windows 95 or NT and machines from Silicon Graphics Inc. Other products include HyperChem Lite, an introductory version of HyperChem, ChemPlus, a set of extensions to HyperChem, and HyperNMR, a package that performs a priori quantum mechanical simulation of NMR spectra. A Windows NT version of HyperChem is also available for the Digital Equipment Corporation Alpha PC, as are client-server versions of HyperChem that use UNIX computational servers from Silicon Graphics, Digital Equipment Corporation, and IBM. HyperChem is a molecular modeling software product noted for its ease of use. It is a comprehensive desktop productivity tool for visualizing, analyzing and communicating information about molecular structures as well as for performing the calculations of computational chemistry. It is the leading molecular modeling software for the Microsoft Windows environment of IBM-compatible PCs. -30 - HyperChem is a registered trademark of Hypercube, Inc. All other trade and product names mentioned are the service marks, trademarks or registered trademarks of their respective holders. From - Thu Apr 10 09:50:40 1997 Received: from MAILQUEUE by TRURL (Mercury 1.21); 10 Apr 97 11:07:35 GMT+2 From: "Maria Rutkowska" Organization: Faculty of Chemistry, UJ To: hyperchem@hyper.com Date: Thu, 10 Apr 1997 11:07:32 MET+2 Subject: literture Priority: normal X-mailer: Pegasus Mail v3.22 Message-ID: <74678B65DA@Trurl.ch.uj.edu.pl> X-Mozilla-Status: 0001 Content-Length: 370 I would be grateful for the references to the papers which describe the details of using Conformational Search and PhiPsi Scans routins of ChemPlus - extension for Hyperchem. If you have used it yourself just let me know as well. I would like to share my experience in using this routines and clear some points. Maryla Rutkowska Jagiellonian University, Poland From - Fri Mar 21 23:03:49 1997 From: ross@cgl.ucsf.EDU Date: Fri, 21 Mar 1997 11:03:04 -0800 (PST) Message-Id: <199703211903.LAA01433@socrates.ucsf.EDU> To: barclay@hyper.hyper.com, hyperchem@hyper.com Subject: Re: AMBER 4.1 X-Mozilla-Status: 0011 Content-Length: 564 If you want to implement a new set of parameters, use a text editor to modify the parameter files (as per Appendix B of the Reference Manual, "Force Field Files"). You will then choose the parameter set using Setup/Select Parameter Set and compile it prior to using it, with Setup/Compile Parameter Set. I don't know how Hyperchem works, but in the new ff, the atom types are different, so in the amber progs themselves, we have to load different residue templates so that the atoms are typed properly. Bill Ross, former amber support person From - Thu Apr 10 08:18:15 1997 Date: Wed, 9 Apr 1997 12:43:19 -0800 (AKDT) From: KELLER JOHN W To: hyperchem@hyper.com Subject: Organic Computational Review Message-Id: Mime-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII X-Mozilla-Status: 0001 Content-Length: 1024 Hi folks, Does anyone know of a book, monograph or text, that goes through the different types of organic reactions and/or intermediates (SN2, SN1, Diels-Alder, enolates, free radicals, carbocations, etc etc), and for each describes the current state of computational assessment? This would be a great help to those of us working on integration of computational methods in undergraduate chemistry. Of course the literature references in any such compendium would be quite helpful too. Thanks, JOhn Keller From - Thu Apr 10 12:09:01 1997 Comments: Authenticated sender is From: "Millard E. Pate" To: support@hyper.com Date: Thu, 10 Apr 1997 11:45:12 +0000 MIME-Version: 1.0 Content-type: text/plain; charset=US-ASCII Content-transfer-encoding: 7BIT Subject: (Fwd) Problems with Windows NT 4.0, HyperChem 4(.5) and ChemPl X-Confirm-Reading-To: "Millard E. Pate" X-pmrqc: 1 Return-receipt-to: "Millard E. Pate" Priority: normal X-mailer: Pegasus Mail for Win32 (v2.52) X-Mozilla-Status: 0001 Content-Length: 1183 Hi there! I was using HyperChem 4.5 and ChemPlus 1.5 with Windows 3.1. For some reasons I had to switch over to Windows NT 4.0. After I installed the rainbow driver (from your FTP site) HyperChem worked correctly. I also installed ChemPlus 1.5 and changed the NT system settings to 'English (USA)'. But whenever I want to do a conformational search (by clicking onto the icon or from within HyperChem), the following messages are prompted on the screen: Notify on update: Unknown variable 'atomic number' and Error beginning Conformational search: Foreign application won't perform DDE method or operation and that's it. Conformational search won't start! Related problems occured by calling Molecule presentations. What is wrong? Horst Sommer e-mail debay258@ibmmail.com Millard E. Pate pate@hyper.com Director of Dealer Relations Hypercube, Inc. Florida Science and Technology Park 1115 N.W. 4th Street Gainesville, Florida 32601 Voice: 1-352-371-7744/1-800-960-1871 Fax: 1-352-371-3662 Email: info@hyper.com Web Site: http://www.hyper.com From - Thu Apr 10 12:09:01 1997 From: debay258@ibmmail.com Message-Id: <199704101251.IAA06145@relay.hyper.com> Date: Thu, 10 Apr 1997 08:48:54 EDT To: hyperchem@hyper.com MIME-Version: 1.0 Content-Type: text/plain; charset=US-ASCII Content-Transfer-Encoding: 7bit Subject: Problems with Windows NT 4.0, HyperChem 4(.5) and ChemPlus 1 X-Mozilla-Status: 0001 Content-Length: 89 Hi there! I was using HyperChem 4.5 and ChemPlus 1.5 with Windows 3.1. For some reasons I had to switch over to Windows NT 4.0. After I installed the rainbow driver (from your FTP site) HyperChem worked correctly. I also installed ChemPlus 1.5 and changed the NT system settings to 'English (USA)'. But whenever I want to do a conformational search (by clicking onto the icon or from within HyperChem), the following messages are prompted on the screen: Notify on update: Unknown variable 'atomic number' and Error beginning Conformational search: Foreign application won't perform DDE method or operation and that's it. Conformational search won't start! Related problems occured by calling Molecule presentations. What is wrong? Horst Sommer e-mail debay258@ibmmail.com From - Wed Apr 09 15:38:05 1997 Date: Wed, 9 Apr 1997 13:20:34 -0300 Message-Id: <9704091620.AA18142@abello.dic.uchile.cl> X-Sender: facien03@abello.dic.uchile.cl X-Mailer: Windows Eudora Version 1.4.4 Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii" To: hyperchem@hyper.com From: facien03@abello.dic.uchile.cl (Juan S. Gomez-Jeria.) Subject: Values of the isolines. X-Mozilla-Status: 0001 Content-Length: 390 Dear Hyperchemists: It would be a VERY GOOD idea to implement one option making appear the values of the isolines in the 2d and 3d drawings (for our sake and for the papers!!). Thanks, Juan S. Gomez-Jeria Associate Professor Universidad de Chile Facultad de Ciencias Casilla 653 Santiago CHILE. e-mail: facien03@abello.dic.uchile.cl Phone: (562)-672-7261. Fax: (562)-271-3888 From - Wed Apr 09 10:04:52 1997 Date: Wed, 9 Apr 1997 09:21:44 -0400 Message-Id: <199704091321.JAA00271@relay.hyper.com> To: support@hyper.com From: hyperchem-request@relay.hyper.com Subject: hyperchem-request results Reply-To: hyperchem-request@relay.hyper.com Errors-To: listadmin@relay.hyper.com X-Mozilla-Status: 0001 Content-Length: 1041 >>>> subscribe **** subscribe: must specify list **** Help for hyperchem-request@relay.hyper.com. hyperchem-request understands the following commands: subscribe [
] Subscribe yourself (or
if specified) to the named . unsubscribe [
] Unsubscribe yourself (or
if specified) from the named . which [
] Find out which lists you (or
if specified) are on. info Retrieve the general introductory information for the named . lists Show the lists served by this list server. help Retrieve this message. end Stop processing commands (useful if your mailer adds a signature). Commands should be sent in the body of an email message to "hyperchem-request@relay.hyper.com". Commands in the "Subject:" line NOT processed. If you have any questions or problems, please contact "listadmin@relay.hyper.com". This is Brent Chapman's "Majordomo" mailing list manager, Revision 1.32. From - Wed Apr 09 14:41:45 1997 From: barclay@hyper.com Message-Id: <199704091449.KAA00514@relay.hyper.com> Comments: Authenticated sender is To: hyperchem@hyper.com Date: Wed, 9 Apr 1997 10:46:28 +0000 MIME-Version: 1.0 Content-type: text/plain; charset=US-ASCII Content-transfer-encoding: 7BIT Subject: patch for Molecular Dynamics bug Reply-to: support@hyper.com Priority: normal X-mailer: Pegasus Mail for Win32 (v2.52) X-Mozilla-Status: 0001 Content-Length: 792 Greetings HyperChemists, Several users have reported a Molecular Dynamics error with HyperChem 5. When a Molecular Mechanics force-field is used, and a non-zero cool-time is specified, the calculation gives an error at the point when cooling begins. MM+: "Floating point error encountered. May have been caused by an ill-conditioned molecular system." AMBER, BIO+, OPLS: "Two or more atoms are in the same position." The patches are now available through our web site. Go to http://www.hyper.com/Support/MDpatch.html and follow the directions. Regards, Victoria ----------------------------- Victoria Barclay, PhD. Manager, Scientific Support Hypercube, Inc. 1115 4th St. NW, Gainesville, FL 32601 (352)-371-7744 voice (352)-371-3662 fax http://www.hyper.com From - Thu Apr 10 08:18:15 1997 Date: Thu, 10 Apr 1997 12:08:24 +1000 (EST) From: Adnan Hazar To: hyperchem Subject: Re: Organic Computaional Review Message-Id: Mime-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII X-Mozilla-Status: 0011 Content-Length: 1036 John, I think publications of Wavefunction - (founder Prof Warren J Hehre) should be a good start for what you want. http://www.wavefun.com/ I have a copy of " Experiments in Computational Organic Chemistry" written by Prof Hehre and his collegues. It is consist of many small experiments you'd find very useful for undergraduate teaching. They also have a software called SPARTAN and the books are written to help SPARTAN users but as I found that books are useful tools that can be used with any software. Regards ****************************************************************************** ADNAN HAZAR hazad@cougar.vut.edu.au Dep. Chemical Sciences Victoria University of Technology Phone: +61 3 9365 2557 or 2711 PO Box 14428 Fax : +61 3 9365 2465 MCMC, Melbourne, Victoria 8001 Australia ****************************************************************************** keep smiling!!! ****************************************************************************** From - Fri Apr 11 09:37:49 1997 Message-Id: <01BC4674.17A57000@pcmb01.phys.chemie.tu-muenchen.de> From: "Dr. Hans Heitele" To: "'HyperChem users' group'" Subject: Spin 0.76/0.46 for benzene with ZINDO/S Date: Fri, 11 Apr 1997 12:30:07 +-200 Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: quoted-printable X-Mozilla-Status: 0001 Content-Length: 2291 Hi HyperChemists, a ZINDO/S calculation with full CI for benzene (charge 0, spin 0, AM1 = optimized geometry, overlap weighting factors: sigma-sigma 1, pi-pi = 0.585) yielded the following log file (only the interesting section of = the UV spectrum is shown): 4 (Transition) Excitation Energy 262.7 nm 38072.3 1/cm 1 -> 5 Spin S 0.76 State Dipole 0.0000 Oscillator Strength 0.0000 State Dipole Components 0.0000 -0.0000 = 0.0000 Transition Dipole Components 0.0000 -0.0000 = 0.0000 Spin Up : Occ. MO --> Unocc. MO Coefficients ------------------------------------------------- Spin Down: Occ. MO --> Unocc. MO Coefficients ------------------------------------------------- 15 --> 16 0.621382 15 --> 17 0.316696 14 --> 16 0.316696 14 --> 17 -0.621382 5 (Transition) Excitation Energy 262.7 nm 38072.3 1/cm 1 -> 6 Spin S 0.46 State Dipole 0.0000 Oscillator Strength 0.0000 State Dipole Components 0.0000 0.0000 = 0.0000 Transition Dipole Components -0.0000 0.0000 = -0.0000 Spin Up : Occ. MO --> Unocc. MO Coefficients ------------------------------------------------- 15 --> 16 0.621382 15 --> 17 0.316696 14 --> 16 0.316696 14 --> 17 -0.621382 Spin Down: Occ. MO --> Unocc. MO Coefficients ------------------------------------------------- Spin 0.76 and 0.46?? The 'Electronic Spectrum' box showed: =20 Transition 4, Degeneracy 2, Spin Multiplicity 2.510131 Transition 5, Degeneracy 2, Spin Multiplicity 1.923341 ????? Thanks in advance for all constructive comments. Hans From - Fri Apr 11 10:25:23 1997 Message-ID: <334E49D3.406A@hyper.com> Date: Fri, 11 Apr 1997 10:25:23 -0400 From: Krassimir Stavrev Organization: Hypercube, Inc. X-Mailer: Mozilla 3.0 (Win95; I) MIME-Version: 1.0 To: "Dr. Hans Heitele" Subject: Re: Spin 0.76/0.46 for benzene with ZINDO/S References: <01BC4674.17A57000@pcmb01.phys.chemie.tu-muenchen.de> Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Mozilla-Status: 0011 Content-Length: 3242 Dr. Hans Heitele wrote: > > Hi HyperChemists, > > a ZINDO/S calculation with full CI for benzene (charge 0, spin 0, AM1 optimized geometry, overlap weighting factors: sigma-sigma 1, pi-pi 0.585) yielded the following log file (only the interesting section of the UV spectrum is shown): > > 4 (Transition) Excitation Energy 262.7 nm > 38072.3 1/cm > > 1 -> 5 Spin S 0.76 > State Dipole 0.0000 > Oscillator Strength 0.0000 > > State Dipole Components 0.0000 -0.0000 0.0000 > Transition Dipole Components 0.0000 -0.0000 0.0000 > > Spin Up : Occ. MO --> Unocc. MO Coefficients > ------------------------------------------------- > > Spin Down: Occ. MO --> Unocc. MO Coefficients > ------------------------------------------------- > 15 --> 16 0.621382 > 15 --> 17 0.316696 > 14 --> 16 0.316696 > 14 --> 17 -0.621382 > > 5 (Transition) Excitation Energy 262.7 nm > 38072.3 1/cm > > 1 -> 6 Spin S 0.46 > State Dipole 0.0000 > Oscillator Strength 0.0000 > > State Dipole Components 0.0000 0.0000 0.0000 > Transition Dipole Components -0.0000 0.0000 -0.0000 > > Spin Up : Occ. MO --> Unocc. MO Coefficients > ------------------------------------------------- > 15 --> 16 0.621382 > 15 --> 17 0.316696 > 14 --> 16 0.316696 > 14 --> 17 -0.621382 > > Spin Down: Occ. MO --> Unocc. MO Coefficients > ------------------------------------------------- > > ============================================================== > > Spin 0.76 and 0.46?? The 'Electronic Spectrum' box showed: > > Transition 4, Degeneracy 2, Spin Multiplicity 2.510131 > Transition 5, Degeneracy 2, Spin Multiplicity 1.923341 > ????? > > Thanks in advance for all constructive comments. > Hans Hi Hans, You are probably mixing the spin-states somehow. What are your CI specifications exactly? I used the AM1 optimized benzene with energy cut-off in the CI window 10 eV (big enough), and did not have any funny looking spin-states, just singlets and triplets. Have you tried both the original sigma-sigma interaction factors suggested by Mike Zerner: 1.267 and pi-pi 0.64; as opposed to Jaffe's pi-pi 0.585. Your dipole component lines (all zeros) look also suspicious. Try it the above way and let me know what happened. Regards, Krassimir --- Krassimir Stavrev, PhD Hypercube, Inc. Florida Science and Technology Park 1115 N.W. 4th Street Gainesville, Florida 32601 Voice: 1-352-371-7744/1-800-960-1871 Fax: 1-352-371-3662 From - Fri Apr 11 12:56:18 1997 Message-Id: <199704111346.JAA12069@relay.hyper.com> Date: Fri, 11 Apr 1997 09:45:28 -0400 X-Sender: underhil@137.94.5.141 X-Mailer: Windows Eudora Version 1.4.3 Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii" To: hyperchem@hyper.com From: underhill-r@rmc.ca (Ross Underhill) Subject: Van der Waals parameters for Ions X-Mozilla-Status: 0001 Content-Length: 1049 Does anyone have some words of wisdom about determining van der Waals radii for metal ions (I'm specifically interested in Al with a formal charge between +2 and +3). There's a discussion of this in the Computational Chemistry Manual under the default force field but it seems to me that this is not going to work for ions. One approach I thought of was to use crystal data to obtain the equilibrium distance between oppositely charged ions and then calculate the pair function using the van der Waals and electrostatic interactions so that I get the correct equilibrium distance. The problem with this approach is that it would be very sensitive to the exact choice of formal charge on the ions. For ions like Na+ this isn't really a problem, but there's a lot of debate as to what the formal charge on Al in Al2O3 really is. If anyone has any experience with this or knows of a good reference, I would be very keen to hear. Dr. Ross Underhill Royal Military College of Canada Kingston, Ontario K7K 5L0 (613) 541-6000 X6175 From - Sat Apr 12 09:17:42 1997 Message-Id: From: "Bill Ross" To: Subject: Fw: Van der Waals parameters for Ions Date: Sat, 12 Apr 1997 01:39:05 +0400 X-MSMail-Priority: Normal X-Priority: 3 X-Mailer: Microsoft Internet Mail 4.70.1157 MIME-Version: 1.0 Content-Type: text/plain; charset=ISO-8859-1 Content-Transfer-Encoding: 7bit X-Mozilla-Status: 0001 Content-Length: 1285 > Does anyone have some words of wisdom about determining van der Waals radii > for metal ions (I'm specifically interested in Al with a formal charge > between +2 and +3). .... One approach I thought of was to use crystal > data to obtain the equilibrium distance between oppositely charged ions and > then calculate the pair function using the van der Waals and electrostatic > interactions so that I get the correct equilibrium distance. The problem > with this approach is that it would be very sensitive to the exact choice of > formal charge on the ions. For ions like Na+ this isn't really a problem, > but there's a lot of debate as to what the formal charge on Al in Al2O3 > really is. If anyone has any experience with this or knows of a good > reference, I would be very keen to hear. > I don't remember if Al is covered, but Vedani & collaborators developed an ion ff based on xtal data; it is worth looking at even if hyperchem may not support it. There may be refs to it on the amber web http://www.amber.ucsf.edu/amber under 'developing new parameters'; if not (and in any case for issues in developing parameters for monovalent ions in complex), see the Ross & Hardin paper referred to there, which has the Vedani refs. Bill Ross From - Sat Apr 12 10:31:14 1997 Message-ID: <334F9CB2.296D@hyper.com> Date: Sat, 12 Apr 1997 10:31:14 -0400 From: Krassimir Stavrev Organization: Hypercube, Inc. X-Mailer: Mozilla 3.0 (Win95; I) MIME-Version: 1.0 To: Hans Heitele Subject: Re: AW: Spin 0.76/0.46 for benzene with ZINDO/S References: <01BC46E1.0286B5E0@dial0094.ppp.lrz-muenchen.de> Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Mozilla-Status: 0011 Content-Length: 1632 Hans Heitele wrote: > > Dear Krassimir, > > thanks for your prompt response. Actually, I didn't mix anything, > I just clicked icons. I did the CI calculation with different > specifications (all single excitations, 9 occupied/9 unoccupied > orbitals, 12/12 orbitals etc., using both sets of interaction factors > each). The results was always the same: noninteger values for spin and > multiplicity for transitions 4 and 5. The vanishing dipole components > are not really surprising. Benzene is highly symmetric, and what is the > transition dipole moment of a transition with a spin change change > 0.xxx anyway?!? > > For no other molecule have a encountered such odd > results. > > Regards, > Hans Hi Hans, I was able to reproduce your experiment, and indeed the CI-singles mixes the singlet and triplet for these 2 CI states. The most probable reason for that seems to be the high symmetry and the fact that the excitations are from degenerate pi-to-pi* orbitals. If you break the symmetry -- by making, for example, one of the of the H-C-C angles 122 deg.-- these two states go back to a singlet and a triplet at 240nm. Also, I am not sure you don't need do go beyond the CI-singles level and do at least doubles to correctly handle the excitations from these degenerate orbitals. CISD, however, will make the calculations prohibitively long and expensive. Regards, Krassimir --- Krassimir Stavrev, PhD Hypercube, Inc. Florida Science and Technology Park 1115 N.W. 4th Street Gainesville, Florida 32601 Voice: 1-352-371-7744/1-800-960-1871 Fax: 1-352-371-3662 From - Fri Apr 11 08:02:50 1997 From: barclay@hyper.com Message-Id: <199704110301.XAA08576@relay.hyper.com> Comments: Authenticated sender is To: hyperchem@hyper.com Date: Thu, 10 Apr 1997 23:00:42 +0000 MIME-Version: 1.0 Content-type: text/plain; charset=US-ASCII Content-transfer-encoding: 7BIT Subject: Hello and good-bye Reply-to: support@hyper.com Priority: normal X-mailer: Pegasus Mail for Win32 (v2.52) X-Mozilla-Status: 0001 Content-Length: 1172 Welcome back to the newsgroup! We had greater down-time for our newsgroup than we planned, due to unanticipated complicating factors during the move from Waterloo to Gainesville, Florida. Our apologies if you have been trying to post a message to hyperchem@hyper.com or down-load items from our web-site (which you should now be able to access) or our ftp-site (which we are working on). My position as Support Manager ends this week. I would like to take this opportunity to say farewell to the many Hyperchem folk who have made my tenure here so enjoyable. It has been a real pleasure getting to know you... and the myriad ways that HyperChem and related products are being used in the community. Please welcome my replacement, Krassimir Stavrev. Fresh from the world-reknowned QTP right here in Gainesville, Dr. Stavrev brings considerable insight and experience to this position. I know you will be in good hands! Regards, Victoria ----------------------------- Victoria Barclay, PhD. Manager, Scientific Support Hypercube, Inc. 1115 4th St. NW, Gainesville, FL 32601 (352)-371-7744 voice (352)-371-3662 fax http://www.hyper.com From - Wed Apr 16 16:56:43 1997 Message-ID: <01BC4A5D.D0078920@burton1.qo.fcen.uba.ar> From: Gerardo Burton To: "'hyperchem@hyper.com'" Subject: new bug in MM+? Date: Wed, 16 Apr 1997 12:00:42 -0300 MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: quoted-printable X-Mozilla-Status: 0001 Content-Length: 689 Hello I was preparing an exercise for my students and tried to minimize ciclooctatetraene with MM+ in HyperChem 5.0 to obtain the typical "bathtub" conformation, however MM+ gave a flat conformer, no matter what the starting geometry was. Of course AM1 gave the correct conformer and so does an old version of PCModel that I still keep, which uses the standard MM2 from Allinger. Could this be a similar bug as that found some time ago, that gave an incorrect conformer with cyclohexadiene? Prof. Gerardo Burton e-mail: burton@qo.fcen.uba.ar Departamento de Quimica Organica Facultad de Ciencias Exactas y Naturales Universidad de Buenos Aires Tel/FAX 54 1 788-6915 From - Wed Apr 16 16:56:43 1997 Received: from MAILQUEUE by VIRGIL (Mercury 1.21); 14 Apr 97 16:17:38 +0100 From: "MORTEN LANGGaaRD - BIOKEM" Organization: Roskilde Universitetscenter To: hyperchem@hyper.com Date: Mon, 14 Apr 1997 16:17:36 +0100 Subject: Single-point and Orbitals? Priority: normal X-mailer: Pegasus Mail for Windows (v2.23) Message-ID: <67F5174BDA@virgil.ruc.dk> X-Mozilla-Status: 0001 Content-Length: 741 Just a short question. When I for a MO job, request a single point calculation from a script by sending "do-single-point" then HyperChem finish by displaying an orbital. This is somewhat annoying and waste of time. This does not occur when a single point calculation is requested from the menu. Can this misbehavior be avoided? I am running HyperChem 5.01 and Windows 95. Best regards Morten =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=- Morten Langgaard Department of Chemistry Phone: +45 46757781 * 2471 Roskilde University (RUC) Fax: +45 46757721 POB 260, DK-4000 Roskilde, Denmark E-Mail: ml@virgil.ruc.dk =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=- From - Wed Apr 16 16:56:43 1997 Message-Id: <01BC4874.13D40700@dial0108.ppp.lrz-muenchen.de> From: Hans Heitele To: "'HyperChem Internet email users' group'" Subject: incomplete log-files with ZINDO/S Date: Sun, 13 Apr 1997 22:37:02 +-200 Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: quoted-printable X-Mozilla-Status: 0001 Content-Length: 5675 Hi HyperChem users, I am interested in calculations of UV spectra of rather simple organic molecules using ZINDO/S. For testing I did CI calculations for benzene and p-xylene using all singly excited configurations. Benzene was ok (with some qualifications) but the log-files produced for xylene (QuantumPrintLevel=3D1) - and for larger molecules even more so - were incomplete. The unabridged beginning of the section on the UV spectrum of p-xylene was: ********************************* ********** UV Spectrum ********** ********************************* ---- Absolute Energy in eV. ---- Dipole Moments in Debye. Transition Dipole Components 0.0000 -0.0000 = -0.0000 21 --> 23 0.455103 Transition Dipole Components -0.0000 0.0000 -0.0000 1 -> 4 Spin S 1.00 State Dipole 0.0000 Oscillator Strength 0.0000 20 --> 22 0.483807 Transition Dipole Components 0.1291 -0.0192 -1.2418 State Dipole Components 0.0000 -0.0000 0.0000 State Dipole Components -0.0000 -0.0000 -0.0000 Transition Dipole Components -0.0000 0.0000 -0.0000 Transition Dipole Components 2.3891 -0.1159 0.3910 Transition Dipole Components 0.0000 -0.0000 0.0000 9 (Transition) Excitation Energy 218.1 nm 45852.6 1/cm 21 --> 24 -0.672837 Spin Down: Occ. MO --> Unocc. MO Coefficients ------------------------------------------------- Spin Up : Occ. MO --> Unocc. MO Coefficients ------------------------------------------------- Spin Up : Occ. MO --> Unocc. MO Coefficients ------------------------------------------------- 21 --> 26 0.517043 Transition Dipole Components 0.0049 0.0982 -0.0010 1 -> 16 Spin S 1.00 State Dipole 0.0000 Oscillator Strength 0.0000 19 --> 22 0.502569 18 --> 22 -0.365389 Spin Up : Occ. MO --> Unocc. MO Coefficients ------------------------------------------------- Spin Up : Occ. MO --> Unocc. MO Coefficients ------------------------------------------------- Transition Dipole Components -0.0000 -0.0000 0.0000 18 --> 23 0.488354 Spin Down: Occ. MO --> Unocc. MO Coefficients ------------------------------------------------- 21 (Transition) Excitation Energy 183.4 nm 54539.9 1/cm 18 --> 22 0.568425 Spin Down: Occ. MO --> Unocc. MO Coefficients ------------------------------------------------- Transition Dipole Components -0.0000 -0.0000 0.0000 Transition Dipole Components 0.0000 0.0000 0.0000 Transition Dipole Components 0.0021 0.0423 -0.0004 State Dipole Components -0.0000 -0.0000 -0.0000 27 (Transition) Excitation Energy 169.7 nm 58918.6 1/cm 28 (Transition) Excitation Energy 166.9 nm 59932.7 1/cm 20 --> 25 0.676005 Spin Down: Occ. MO --> Unocc. MO Coefficients ------------------------------------------------- 30 (Transition) Excitation Energy 161.3 nm 61989.7 1/cm State Dipole Components -0.0000 -0.0000 -0.0000 19 --> 25 0.474389 Spin Down: Occ. MO --> Unocc. MO Coefficients ------------------------------------------------- 19 --> 25 0.474389 16 --> 22 -0.363111 1 -> 32 Spin S 1.00 State Dipole 0.0000 Oscillator Strength 0.0000 State Dipole Components -0.0000 -0.0000 -0.0000 Transition Dipole Components 0.0000 -0.0000 0.000 Spin Up : Occ. MO --> Unocc. MO Coefficients ------------------------------------------------- 18 --> 25 0.378950 Spin Down: Occ. MO --> Unocc. MO Coefficients ------------------------------------------------- 18 --> 25 0.378950 32 (Transition) Excitation Energy 157.9 nm etc. Thus it completely dropped or messed up the first 31 excitations which is almost the entire spectroscopically relevant part of the UV spectrum, although these transitions could be inspected in the 'Electronic Spectrum' box of the 'Compute' menu. Is there a size limit for molecules, log files etc. that I might be unaware of? Is there a way to retrieve the lost information? Thanks in advance for any help. Hans From - Wed Apr 16 16:56:43 1997 From: "Christian Pilger" Message-Id: <9704141740.ZM22370@OC30.uni-paderborn.de> Date: Mon, 14 Apr 1997 17:40:59 -0600 X-Mailer: Z-Mail (3.2.2 10apr95 MediaMail) To: hyperchem@hyper.com Subject: HyperChem and DDE Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii X-Mozilla-Status: 0001 Content-Length: 1892 Hi everybody, I've got some questions concerning the control of HyperChem 4.0 by Excel 5 and Visual Basic 4 (16-bit) via DDE. 1.) When I tried to query the variable "is-ring-atom" from an Excel macro, HyperChem sent the message: "ignoring query for unregistered variable: is-ring-atom". A couple of weeks ago I had a discussion about this with Victoria and we came up with the theory that the DDE channels are somehow "confused" with the German versions of Windows and Excel. For that reason I tested an English version of Windows 3.11 with a Visual Basic program - but the result was the same. I'd like to know if somebody else has experienced this problem. 2.) Is it possible for a Visual Basic program to find out, wether HyperChem is allready running, and, if not, to start it ? 3.) When I first installed Visual Basic, I used the 32-bit version under Windows 95. I sent a command-sequence to perform a MD-calculation followed by an optimization. All calculations went very slow, stopped sometimes, went on again, and the calculation-speed changed upon moving the mouse. When I used the 16-bit version of Visual Basic, these effects disappeared and the calculation-speed was much higher. Strange, isn't it ? I'd really appreciate, if somebody could help me. I'm a novice to Visual Basic, but I think it is easy to write very helpful applications using HperChem as a "plattform". I'd like to get in touch with people, who are beginners in this area (like me), and people, who are more experienced. Bye Christian -- Christian Pilger Uni-GH Paderborn FB 13 - Organische Chemie J4.308 Warburger Str. 100 33098 Paderborn Telefon : 05251/60-2183 e-mail : cpilger@oc30.uni-paderborn.de From - Wed Apr 16 17:31:53 1997 Date: Wed, 16 Apr 1997 22:22:47 +0100 (BST) From: patrick kane <94970459@tolka.dcu.ie> To: HChem User , HChem Supp , CCL Every Subject: HyperChem and Conjugate Gradient Search Methods Message-Id: Mime-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII X-Mozilla-Status: 0001 Content-Length: 826 Hi, Page 213 of the HyperChem, version 4 manual states that with conjugate gradient search methods, a descent direction (to lower energy) is chosen by consdering [among other things - my brackets] the conjugate of the gradient which is initially the negative of the current gradient. Could someone please tell me what is meant by the CONJUGATE of a gradient. THanks in response for your replies. Rgds, Paddy. ************************************************************************* * * * Paddy Kane email: 94970459@tolka.dcu.ie * * School of Chemical Sciences * * Dublin City University Tel: 00-353-1-7045641 * * Dublin 9 * * Ireland. Fax: 00-353-1-7045503 * * * ************************************************************************* From - Thu Apr 17 07:56:48 1997 From: THORSTEN.LOEHL@DENOTES.henkel.de To: internet@henkel.de, hyperchemhypercom Subject: Request for help //17.04.97 11:30:37 Message-ID: <0006700000265599000002L092*@MHS> Date: Thu, 17 Apr 1997 11:39:31 +0100 X-Mozilla-Status: 0001 Content-Length: 555 Dear Hyperchemers, for demonstration purposes I would like to show a cubic block of Al-Atoms (similar to a crystal) with Hyperchem 5. While I am waiting to receive ChemPlus 1.6 and to evaluate its use I am running out of time for this very demonstration. Question: Is anyone able to provide a file with this information or can someone give me a hint on how to construct it manually and the use of Hyperchem 5. Thanks for any reply. Dr. Thorsten Loehl, Henkel KGaA Corp. Germany, EMail: Thorsten.Loehl@henkel.de (Please NO attached files) From - Thu Apr 17 07:56:48 1997 Received: from ITS-PHARM/SpoolDir by its-pharm.vcp.monash.edu.au (Mercury 1.21) ; Thu, 17 Apr 1997 10:56:09 +1000 Received: from SpoolDir by ITS-PHARM (Mercury 1.30); Thu, 17 Apr 1997 10:56:08 +1000 Date: Thu, 17 Apr 1997 10:55:56 +0000 From: Elizabeth Yuriev Subject: Re: Publications ... To: hyperchem@hyper.com, tbender@ccs.carleton.ca (timothy bender) Reply-to: elizabeth.yuriev@vcp.monash.edu.au Message-id: <5428D3B6F5B@its-pharm.vcp.monash.edu.au> Organization: Victorian College of Pharmacy X-Mailer: Pegasus Mail for Windows (v2.33) Priority: normal X-Mozilla-Status: 0011 Content-Length: 1146 Tim Bender asked: > I am in the process of putting together a publications which >includes molecular modelling results obtained in hyperchem using the MM+ >force field and the AM1 semi-empirical method, > what I am looking for are examples in the scientific literature >where the results of models produced in hyperchem are used, in order to >get a better idea how to present my results, Hi Tim, you may want to have a look at our recent papers using HyperChem: 1. Modelling steric effects in DNA-binding platinum(II)-am(m)ine complexes. E. Yuriev and J.D. Orbell, J. Comput.-Aided Mol. Design, 10 (1996), 589 - 606 (MM+ and AMBER force fields, AM1 and ZINDO/1 semiempirical methods were used) 2. Steric Parameters for Metal Binding Sites on Nucleobases. E. Yuriev and J.D. Orbell, 35 (1996), 7914-7915 (MM+ was used) Hope that helps, Elizabeth ********************* Elizabeth Yuriev Department of Medicinal Chemistry Victorian College of Pharmacy Monash University 381 Royal Parade, Parkville 3052 Victoria, Australia Ph: 61-3-9903-9636 Fax: 61-3-9903-9582 e-mail: elizabeth.yuriev@vcp.monash.edu.au From - Thu Apr 17 07:56:47 1997 Message-Id: <3.0.32.19970416130547.006ce748@mbi.ucla.edu> X-Sender: lavelle@mbi.ucla.edu X-Mailer: Windows Eudora Pro Version 3.0 (32) Date: Wed, 16 Apr 1997 13:05:49 -0700 To: hyperchem@hyper.com From: Laurence Lavelle Subject: ZINDO/1 calculations and the MMX force field. Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii" X-Mozilla-Status: 0001 Content-Length: 1308 Hi, ZINDO/1 single point calculations give too high a conformational energy difference (765 kcal/mol) for two conformations of the octahedral coordination compound [Ni(NCS)2(Py)4. If anyone is aware of this "ZINDO/1 problem" I would appreciate your comments. Are there other packages (semi-empirical) that I can use to calculate the single point energy of different conformations and compare with ZINDO/1 results? ............................................................................ ......................................... The MMX force field appears to model coordination compounds with some sophistication (trans effect, high spin/low spin in different coordination geometries etc.). What plans are there for HyperChem to include some of the MMX features for organometallic and coordination compounds? Thanks Laurence Lavelle """"""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""" Laurence Lavelle, Ph.D. University of California Los Angeles Laboratory of Structural Biology & Molecular Medicine Molecular Biology Institute Los Angeles, CA90095-1570, USA Email:LAVELLE@MBI.UCLA.EDU Phone:(310)206-8270 Fax:(310)825-0982 http://www.mbi.ucla.edu/people/lavelle/lavelle.html """"""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""" From - Thu Apr 17 07:56:47 1997 From: tbender@ccs.carleton.ca (timothy bender) Message-Id: <199704162147.RAA08573@rideau> Subject: Publications ... To: hyperchem@hyper.com Date: Wed, 16 Apr 97 17:47:56 EDT X-Mailer: ELM [version 2.3 PL11] X-Mozilla-Status: 0001 Content-Length: 1091 ---------------------------------------------------------------------- Tim Bender, B.Sc | To know chemistry, Student, Department of Chemistry | is to know a little Carleton University, Ottawa, Ont, Canada | piece of nature ... Email address: tbender@ccs.carleton.ca | ---------------------------------------------------------------------- Hello ... I am in the process of putting together a publications which includes molecular modelling results obtained in hyperchem using the MM+ force field and the AM1 semi-empirical method, what I am looking for are examples in the scientific literature where the results of models produced in hyperchem are used, in order to get a better idea how to present my results, if anyone has a reference involving hyperchem models could you please forward me the citation, thank you in advance ... Tim :-) ------------------------------------------------------------------------- "... can money pay for all the days, I lived awake but half asleep? ..." - Primitive Radio Gods From - Thu Apr 17 07:56:47 1997 From: tbender@ccs.carleton.ca (timothy bender) Message-Id: <199704162145.RAA07934@rideau> Subject: Annealing .... To: hyperchem@hyper.com Date: Wed, 16 Apr 97 17:45:02 EDT X-Mailer: ELM [version 2.3 PL11] X-Mozilla-Status: 0001 Content-Length: 984 ---------------------------------------------------------------------- Tim Bender, B.Sc | To know chemistry, Student, Department of Chemistry | is to know a little Carleton University, Ottawa, Ont, Canada | piece of nature ... Email address: tbender@ccs.carleton.ca | ---------------------------------------------------------------------- Hello ... I would appreciate any information, helpful hints or specific details on performing simulated anneallings on protiens and/or large macromolecules (more than 200 atoms), I am trying to find the global minium of a large macromolecule with only limited success to this point, any help that my fellow hyperchem users can offer would be greatly appreciated, thank-you in advance Tim :-) ------------------------------------------------------------------------- "... can money pay for all the days, I lived awake but half asleep? ..." - Primitive Radio Gods From - Thu Apr 17 08:33:14 1997 Message-ID: <3356188A.71FF@hyper.com> Date: Thu, 17 Apr 1997 08:33:14 -0400 From: Krassimir Stavrev Organization: Hypercube, Inc. X-Mailer: Mozilla 3.0 (Win95; I) MIME-Version: 1.0 To: patrick kane <94970459@tolka.dcu.ie> Subject: Re: HyperChem and Conjugate Gradient Search Methods References: Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Mozilla-Status: 0011 Content-Length: 1261 > > Could someone please tell me what is meant by the CONJUGATE of a gradient. Hi Patrick, In the Fletcher-Reeves scheme, the search for a minimum explores not only the current gradient to find a descent direction (as, for example, in the Steepest Descent method), but also its "conjugate" which in the beginning is set to the negative of the current gradient, as well as the gradient from the step before, thus utilizing the 'gradient history' of the search. A refinement to this scheme was done in the Polak-Ribiere method which takes into account also the previous conjugate direction, thus avoiding the need of resetting it to the negative of the gradient, as implemented originally in the Fletcher-Reeves scheme. This P-R method usually converges faster, and is therefore chosen as the default of HyperChem. You can learn more about these methods in P. Gill et.al's book, Practical Optimization, Academic Press, NY (1981. A good outline is also given in the computational chemistry manual of HyperChem. Regards, Krassimir --- Krassimir Stavrev, PhD Hypercube, Inc. Florida Science and Technology Park 1115 N.W. 4th Street Gainesville, Florida 32601 Voice: 1-352-371-7744/1-800-960-1871 Fax: 1-352-371-3662 From - Thu Apr 17 12:51:45 1997 From: Joel Polowin To: CCL Every , HChem User Cc: "'94970459@tolka.dcu.ie'" <94970459@tolka.dcu.ie> Subject: RE: HyperChem and Conjugate Gradient Search Methods Date: Thu, 17 Apr 1997 12:18:00 -0400 Message-Id: <33564F92@office.agissa> X-Mailer: Microsoft Mail V3.0 X-Mozilla-Status: 0011 Content-Length: 657 Paddy Kane <94970459@tolka.dcu.ie> writes: > Page 213 of the HyperChem, version 4 manual states that with conjugate > gradient search methods, a descent direction (to lower energy) is chosen > by consdering [among other things - my brackets] the conjugate of the > gradient which is initially the negative of the current gradient. > Could someone please tell me what is meant by the CONJUGATE of a gradient. Take a look in the HyperChem _Computational Chemistry_ manual, in the "Conjugate Gradient Methods" section in the "Computational Options" chapter. All of that stuff is defined/explained there. Joel joelp@agiss.com From - Thu Apr 17 13:08:27 1997 From: Joel Polowin To: "'hyperchem@hyper.com'" Cc: "'LAVELLE@MBI.UCLA.EDU'" Subject: RE: ZINDO/1 calculations and the MMX force field. Date: Thu, 17 Apr 1997 12:06:00 -0400 Message-Id: <33564BD4@office.agissa> X-Mailer: Microsoft Mail V3.0 X-Mozilla-Status: 0011 Content-Length: 1244 > The MMX force field appears to model coordination compounds with some > sophistication (trans effect, high spin/low spin in different coordination > geometries etc.). The key word is "appears"; if I recall correctly, the only REAL effect of specifying the spin state with MMX (the MM2 variant in PCMODEL) is to adjust the metal-ligand r_0 value by about 0.2 Angstroms in the case of a "low-spin" system. To get a "square planar" system, it's necessary to put atoms into the axial coordination sites as place-holders/space-fillers; all X-M-X interactions are strictly steric. I don't think there's anything that can be done with MMX/PCMODEL that can't also be done with MM+ in HyperChem by adjusting parameters, using structural restraints, writing scripts for driving a dihedral angle, etc. (I did my Ph.D. project with PCMODEL; the above comments are with respect to the DOS/SUN release 4. I spent the last two years or so of my research up to the elbows in PCMODEL's source code.) (Hi, everyone, I'm back on the list, at least once in a while. I'm with the Agiss Software Corp.; my current project is the development of a package which uses HyperChem as its main numerical engine.) Joel joelp@agiss.com From - Thu Apr 17 13:11:43 1997 Date: Thu, 17 Apr 1997 10:06:08 -0700 (MST) From: Mark D Shenderovich To: patrick kane <94970459@tolka.dcu.ie> cc: HChem User , HChem Supp , CCL Every Subject: Re: CCL:HyperChem and Conjugate Gradient Search Methods In-Reply-To: Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII X-Mozilla-Status: 0011 Content-Length: 1043 On Wed, 16 Apr 1997, patrick kane wrote: > > Hi, > > Page 213 of the HyperChem, version 4 manual states that with conjugate > gradient search methods, a descent direction (to lower energy) is chosen > by consdering [among other things - my brackets] the conjugate of the > gradient which is initially the negative of the current gradient. > > Could someone please tell me what is meant by the CONJUGATE of a gradient. > > THanks in response for your replies. > > Rgds, > Paddy. Dear Paddy, CONJUGATE means that the direction of each iterative step is defined as a linear combination of gradient vectirs in at least two consequtive points of the path, i.e. r(i) = a*g(i) + b*g(i-1) + ... This makes conjugated gradient methods very efficient when a starting point is far from an energy mimimum. But, if you have a choice, use second order algorithms (Fletcher-Powel, Newton-Raphson, etc.) when you are close to a minimum and conjugated gradient becomes too slow. Yours, Mark Shenderovich From - Fri Apr 18 09:01:49 1997 Date: Fri, 18 Apr 1997 07:42:16 -0300 (GRNLNDST) From: Sergio Emanuel Gelembeck X-Sender: segalemb@swan.uspnet.usp.br To: Laurence Lavelle Cc: hyperchem@hyper.com Subject: Re: ZINDO/1 calculations and the MMX force field. In-Reply-To: <3.0.32.19970416130547.006ce748@mbi.ucla.edu> Message-Id: Mime-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII X-Mozilla-Status: 0011 Content-Length: 831 Dear Laurence, On Wed, 16 Apr 1997, Laurence Lavelle wrote: > Hi, > > ZINDO/1 single point calculations give too high a conformational energy > difference (765 kcal/mol) for two conformations of the octahedral > coordination compound [Ni(NCS)2(Py)4. > > If anyone is aware of this "ZINDO/1 problem" I would appreciate your comments. > ZINDO/1 works only for geometry calculations. Electronic structure may be made with ZINDO/S. But, in my experience with ruthenium complexes, ZINDO/S furnishes to large differences for calculations of conformers. > Are there other packages (semi-empirical) that I can use to calculate the > single point energy of different conformations and compare with ZINDO/1 > results? The original ZINDO program is sold by BIOSYM. Bye, Sergio Emanuel Galembeck From - Fri Apr 18 11:46:19 1997 From: cyyi@musica.McGill.CA X-SMTP-Posting-Origin: 198.168.189.115 (I-15.DAS.McGill.CA [198.168.189.115]) Message-Id: <199704181308.JAA11708@sirocco.CC.McGill.CA> MIME-Version: 1.0 Content-Type: text/plain Content-Transfer-Encoding: 7bit Date: Fri, 18 Apr 97 09:08:52 -0500 Subject: Sugar Builder problem To: Hyperchem@hyper.com X-Mailer: SPRY Mail Version: 04.00.06.17 X-Mozilla-Status: 0001 Content-Length: 965 I want to do calculations on polysaccharides and have encountered the following problems with "Sugar Builder" of ChemPlus 1.6 : Attempt to add blocking group O-methyl, using menu Add/End group, to positions 2, 3, 6 of glucose residue causes the following message (from Hyperchem) to appear: "Poorly formed or out of range array/vector indexing for atom-name" Sometimes a second message from Sugar Builder also appears: DDE error: Foreign application won't perform DDE operation ------. The program failed to add O-methyl group to any position. Can any one explain why I am getting these messages? Any suggestions to solve this problem would be much appreciated. Other relevant (?) information: a) Operating system: Windows 95; b) HyperChem 5.0 & ChemPlus 1.6 installed with Chem.scr modified as suggested in Readme.txt; c) Tutorial on Sugar Builder (ChemPlus menu pg. 81) working correctly. C.T. Yim Dawson College, Montreal From - Fri Apr 18 12:29:18 1997 From: Joel Polowin To: "'hyperchem@hyper.com'" Cc: "'burton@qo.fcen.uba.ar'" Subject: RE: new bug in MM+? Date: Fri, 18 Apr 1997 12:12:00 -0400 Message-Id: <33579F02@office.agissa> X-Mailer: Microsoft Mail V3.0 X-Mozilla-Status: 0011 Content-Length: 1626 Prof. Gerardo Burton (burton@qo.fcen.uba.ar) writes: > I was preparing an exercise for my students and tried to minimize = > ciclooctatetraene with MM+ in HyperChem 5.0 to obtain the typical = > "bathtub" conformation, however MM+ gave a flat conformer, no matter = > what the starting geometry was. Of course AM1 gave the correct = > conformer and so does an old version of PCModel that I still keep, = > which uses the standard MM2 from Allinger. > Could this be a similar bug as that found some time ago, that gave an = > incorrect conformer with cyclohexadiene? No, this is the behaviour that one would expect from a molecular mechanics calculation. All of the carbon atoms are assigned type C3 (sp2 carbon); each torsion angle in the system has two of these C3-type atoms in the middle, and so "tries" to come out planar during optimization. As a result, the entire system goes planar. If, instead, you set the carbon atoms to be of "unknown" type ("**"), optimization gives torsions around the single bonds of about 64', and 0' around the double bonds. (AM1 gives torsions of 56.9' and 0', respectively.) It's essentially the same problem as trying to model butadienes or biphenyls; the modelling scheme, which works on the basis of atom types, assigns incorrect bond lengths and torsions to single bonds between sp2 atoms. If the atom types are unknown, MM+ uses the default parameter scheme, which considers the bond orders. What surprises me a bit is that PCMODEL didn't have the same problem, assuming that you specified sp2 carbons to it. Joel joelp@agiss.com From - Fri Apr 18 15:12:01 1997 From: "Jens Spanget-Larsen" Organization: Roskilde Universitetscenter To: hyperchem@hyper.com, JoelP@agiss.com, burton@qo.fcen.uba.ar Date: Fri, 18 Apr 1997 18:20:29 +0100 Subject: RE: new bug in MM+? Priority: normal X-mailer: Pegasus Mail for Windows (v2.23) Message-ID: X-Mozilla-Status: 0011 Content-Length: 1309 Joel, on the MM+ prediction of planar cyclooctatetraene: > What surprises me a bit is that PCMODEL didn't have the same problem, > assuming that you specified sp2 carbons to it. I just checked that PCMODEL does indeed predict a tubshaped cyclooctatetraene, provided the initial configuration is not completely planar (in this case, the molecule does not allways escape from the hilltop, but remains planar). To me, this is not surprising, since in PCMODEL, conjugated pi-systems are treated by an SCF quantum chemical procedure, predicting anti-aromatic destabilization of the 4N-perimeter pi-system in planar cyclooctadiene (according to Huckel's rule). In the tub-shaped configuration the destabilization is minimized, but HyperChem MM+ does not include this contribution, since there is no SCF treatment of pi-systems. Why is the SCF part omitted in MM+? Jens >--< =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= JENS SPANGET-LARSEN Department of Chemistry Phone: +45 46757781 + 2710 Roskilde University (RUC) Fax: +45 46757721 P.O.Box 260 E-Mail: JSL@virgil.ruc.dk DK-4000 Roskilde, Denmark http://frederik.ruc.dk/dis/chem/psos =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= From - Fri Apr 18 16:17:29 1997 From: Joel Polowin To: burton , hyperchem , jsl Subject: RE: new bug in MM+? Date: Fri, 18 Apr 1997 16:06:00 -0400 Message-Id: <3357D5CD@office.agissa> X-Mailer: Microsoft Mail V3.0 X-Mozilla-Status: 0011 Content-Length: 2777 JENS SPANGET-LARSEN (JSL@virgil.ruc.dk) writes: > I just checked that PCMODEL does indeed predict a tubshaped > cyclooctatetraene, provided the initial configuration is not > completely planar (in this case, the molecule does not allways escape > from the hilltop, but remains planar). > To me, this is not surprising, since in PCMODEL, conjugated > pi-systems are treated by an SCF quantum chemical procedure, > predicting anti-aromatic destabilization of the 4N-perimeter > pi-system in planar cyclooctadiene (according to Huckel's rule). In > the tub-shaped configuration the destabilization is minimized, but > HyperChem MM+ does not include this contribution, since there is no > SCF treatment of pi-systems. Why is the SCF part omitted in MM+? The SCF stuff in PCMODEL's MMX isn't part of MM2; it was written by John McKelvey and added to the PCMODEL code. (In fact, when I discovered errors in the data storage procedure for pi systems during file-saving in PCMODEL, and requested clarification from Kevin Gilbert of Serena Software, he told me that he couldn't help -- McKelvey was the only one who knew how that code was supposed to work!) MMX running with the SCF calculations is a peculiar hybrid between molecular mechanics and semi-empirical calculations. But that explains why PCMODEL is giving the different results; it's not running as a simple molecular mechanics program. The interaction between the SCF stuff and the molecular mechanics calculations is a bit strange; it involves automatic revision of the force constants for planar systems' torsion angles with each iteration between SCF calculations and optimizations. (It also had a nasty bug which I spent several frustrating months trying, unsuccessfully, to track down, involving an interaction between the dihedral driver and the SCF calculations. After the first dihedral value was done, with each remaining dihedral angle, the force constants for planar systems would change sign; each "planar" torsional angle would suddenly try to go anti-planar, and the structure would do a somersault or try to tie itself into knots. And then with the next SCF iteration, everything would try to go planar again. This didn't happen with all structures, but I did get it to happen consistently with -- if I recall correctly -- using the dihedral driver on the CH3 group of propene, where the other two C atoms were marked for pi calculations. I don't know if that bug has been fixed in later versions of PCMODEL.) If you specify that the carbon atom types for cyclooctatetraene in PCMODEL are "CA" (aromatic carbon) or other SP2 type, and omit the pi calculations, you will probably find that the structure will go planar. Joel joelp@agiss.com From - Mon Apr 21 08:27:41 1997 X-Sender: lavelle@mbi.ucla.edu X-Mailer: Windows Eudora Pro Version 3.0 (32) Date: Fri, 18 Apr 1997 13:40:01 -0700 To: hyperchem@hyper.com From: Laurence Lavelle Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii" X-Mozilla-Status: 0001 Content-Length: 1736 Hi, Thanks for all the responses. I would like to let others know what software recommendations were made and clarify some points. **************************************************************************** ****************** Dear Laurence, I suspect the semi-empirical parameters for Nickel may be causing the problem. Try Walter Thiel's MNDO/d, in the new version of AMPAC, marketted by SemiChem. **************************************************************************** ***************** Points to clarify. 1) The ground state is a triplet state with 2 unpaired e-. 2) It is very clear to me that one uses ZINDO/1 to obtain single point energies. I use the description "different conformations" for the possible different positions (conformations) of the pyridine ligands which are free to rotate around the Ni-N bond. This means the Oh geometry is essentially maintained. All calculations were done in the electronic ground state (triplet) using ZINDO/1, otherwise I would have stated "different electronic" states. Dr. Krassimir Stavrev (HyperChem) has made several interesting suggestions which I will follow up on in the near future. Welcome Krassimir, and I'm pleased to see Joel still has a finger in the pie. Thanks Laurence """"""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""" Laurence Lavelle, Ph.D. University of California Los Angeles Laboratory of Structural Biology & Molecular Medicine Molecular Biology Institute Los Angeles, CA90095-1570, USA Email:LAVELLE@MBI.UCLA.EDU Phone:(310)206-8270 Fax:(310)825-0982 http://www.mbi.ucla.edu/people/lavelle/lavelle.html """"""""""""""""""""""""""""""""""""""""""""""""""""""""""""""""" From - Mon Apr 21 08:27:41 1997 From: E.W.Thomas@biosci.salford.ac.uk Message-Id: <199704190931.FAA22543@relay.hyper.com> Date: 19 Apr 97 10:29 To: hyperchem@hyper.com Subject: MM+ parameters X-Mailer: University of Salford cc:Mail/SMTP gateway 1.75 X-Mozilla-Status: 0001 Content-Length: 409 During Hyperchem MM+ modelling of chlorinated aliphatics I notice that default parameters are being used for torsions(notably cl c4 co o1 and cl c4 s4 o1)with warnings about bend defaults in some instances.How should I assess the validity of such calculations?Is there a likely source of transferable parameters for such structures?I'm using HyperChem 4.0. Emrys. From - Mon Apr 21 08:27:42 1997 with TCP; Sat, 19 Apr 97 19:33:00 CDT Message-ID: <33597ED6.7FEA@truman.edu> Date: Sat, 19 Apr 1997 19:26:30 -0700 From: sc18 Reply-To: sc18@truman.edu Organization: Truman St. Un. X-Mailer: Mozilla 3.01 (Win95; I; 16bit) MIME-Version: 1.0 To: hyperchem@hyper.com Subject: Re: Building cyclooctatetraene References: <33579F02@office.agissa> Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Mozilla-Status: 0011 Content-Length: 832 Hi All, I've found that HyperChem indeed, does not build cyclooctatraene in any way. There is a build for it in HyperChem though. Start with cubane and optimize. Then delect alternate single bonds and add appropriate double bonds to get the gross structure. Do a partial miminization with MM+ ( all of the other MM modules seem to give flat COT also) until you get something resembling COT. Then finish with AM1 or PM3, both of which give acceptable COT geometry. NB: If you optimize even these QCHM structure in any MM feature of HyperChem, you again get flat COT. Apparently it really is a function of the MM code, because the same optimizing schemes that give good COT geometry in the QCHM give flat COT in the MM programs. Hope this helps if you really need a COT structure. Sincerely, Ken Fountain From - Mon Apr 21 12:24:14 1997 X-Mozilla-Status: 0001 Content-Length: 2193 Message-ID: <01BC4E36.83D97B80@burton1.qo.fcen.uba.ar> From: Gerardo Burton To: "'sc18'" Cc: "'hyperchem@hyper.com'" Subject: RE: Building cyclooctatetraene Date: Mon, 21 Apr 1997 09:29:29 -0300 MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: quoted-printable X-Mozilla-Status: 0010 Content-Length: 1775 Hello I didn't have any trouble getting a good cyclooctatetraene geometry with a semiempirical method. If you build a perfectly flat COT like the one that results from MM+, optimization with AM1 or PM3 will keep it flat and you get the transition state (one negative normal mode, sorry in my previous message I had not checked the vibrational spectra and said that this was a conformer). If you twist it, e.g. by moving two carbons from a double bond slightly out of the plane, and then reminimize with AM1 it will rapidly go to the tub-shaped COT. Gerardo Prof. Gerardo Burton e-mail: burton@qo.fcen.uba.ar Departamento de Quimica Organica Facultad de Ciencias Exactas y Naturales Universidad de Buenos Aires Tel/FAX 54 1 788-6915 ---------- From: sc18[SMTP:sc18@truman.edu] Sent: S=E1bado 19 de Abril de 1997 23:27 To: hyperchem@hyper.com Subject: Re: Building cyclooctatetraene Hi All, I've found that HyperChem indeed, does not build cyclooctatraene in any way. There is a build for it in HyperChem though. Start with cubane and optimize. Then delect alternate single bonds and add appropriate double bonds to get the gross structure. Do a partial miminization = with MM+ ( all of the other MM modules seem to give flat COT also) until you get something resembling COT. Then finish with AM1 or PM3, both of which give acceptable COT geometry. NB: If you optimize even these = QCHM structure in any MM feature of HyperChem, you again get flat COT. =20 Apparently it really is a function of the MM code, because the same optimizing schemes that give good COT geometry in the QCHM give flat = COT in the MM programs. Hope this helps if you really need a COT structure. Sincerely, Ken Fountain From - Mon Apr 21 13:54:07 1997 From: Joel Polowin To: hyperchem , jsl Subject: RE: new bug in MM+? Date: Mon, 21 Apr 1997 14:23:00 -0400 Message-Id: <335BB2DE@office.agissa> X-Mailer: Microsoft Mail V3.0 X-Mozilla-Status: 0011 Content-Length: 1217 JENS SPANGET-LARSEN writes: > No! If you adopt the default MMX atom type assignment (sp2 alkene, > atom type 2), you get a tub-shaped structure with alternating > bondlengths, regardles of whether you select the pi-scf option or not! Okay -- it's been a few years since I worked with MMX; I thought that it based its bonding patterns on atom types alone, instead of considering bond orders. > But if > you assign all carbons to be "aromatic" (atom type 40), the system is > forced to behave like an aromatic pi-system and you get of course a > planar structure with no bond-alternation; the scf pi-procedure is > automatically deactivated when atom type 40 is selected. But why > choose "aromatic" carbons, COT is not aromatic! For the work I was doing -- modelling pi complexes of transition metals -- I didn't have a lot of choice. Trying to include the SCF calculations with the dihedral driver gave the disastrous results I described earlier; the only way I could get my hydrocarbon structures to have approximately the correct properties was to specify "aromatic" carbons. It was the least- inappropriate work-around that I could find. Joel joelp@agiss.com From - Mon Apr 21 14:22:46 1997 Received: from MAILQUEUE by VIRGIL (Mercury 1.21); 21 Apr 97 16:41:44 +0100 From: "Jens Spanget-Larsen" Organization: Roskilde Universitetscenter To: JoelP@agiss.com, hyperchem@hyper.com Date: Mon, 21 Apr 1997 16:41:37 +0100 Subject: RE: new bug in MM+? Priority: normal X-mailer: Pegasus Mail for Windows (v2.23) Message-ID: <11060BD0FC6@virgil.ruc.dk> X-Mozilla-Status: 0011 Content-Length: 4151 Joel Polowin: > The SCF stuff in PCMODEL's MMX isn't part of MM2; it was > written by John McKelvey and added to the PCMODEL code. (In > fact, when I discovered errors in the data storage procedure > for pi systems during file-saving in PCMODEL, and requested > clarification from Kevin Gilbert of Serena Software, he told > me that he couldn't help -- McKelvey was the only one who > knew how that code was supposed to work!) MMX running with > the SCF calculations is a peculiar hybrid between molecular > mechanics and semi-empirical calculations. Yes, MMX is a hybrid, just like modern procedures like MM3 and later versions of MM2 that all combine the MM approach with an SCF procedure for pi-systems. This is necessary in order to describe the properties of conjugated systems with non-localizable bonds; these systems can only be properly represented within a quantum chemical context. There is a long tradition for this kind of modelling. M.J.S Dewar worked with conceptually related models in the 60'es, before he turned to MINDO, MNDO, etc., and already in 1972 Warshel and Karplus (JACS 94:5612) launched their QCFF/PI hybrid. The latter was applicable to ground as well as excited pi-pi* states, and molecular vibrations and Condon-factors could be computed. > But that explains why PCMODEL is giving the different results; > it's not running as a simple molecular mechanics program. The > interaction between the SCF stuff and the molecular mechanics > calculations is a bit strange; it involves automatic revision > of the force constants for planar systems' torsion angles with > each iteration between SCF calculations and optimizations. (It > also had a nasty bug which I spent several frustrating months > trying, unsuccessfully, to track down, involving an interaction > between the dihedral driver and the SCF calculations. After > the first dihedral value was done, with each remaining dihedral > angle, the force constants for planar systems would change sign; > each "planar" torsional angle would suddenly try to go anti-planar, > and the structure would do a somersault or try to tie itself into > knots. And then with the next SCF iteration, everything would > try to go planar again. This didn't happen with all structures, > but I did get it to happen consistently with -- if I recall > correctly -- using the dihedral driver on the CH3 group of propene, > where the other two C atoms were marked for pi calculations. I don't > know if that bug has been fixed in later versions of PCMODEL.) I understand your frustration. But program bugs and technical problems are not necessarily an indication that the underlying theoretical concepts are invalid. With PCMODEL Version 5.1 I have not experienced problems of the kind you describe. > If you specify that the carbon atom types for cyclooctatetraene in > PCMODEL are "CA" (aromatic carbon) or other SP2 type, and omit the > pi calculations, you will probably find that the structure will go > planar. No! If you adopt the default MMX atom type assignment (sp2 alkene, atom type 2), you get a tub-shaped structure with alternating bondlengths, regardles of whether you select the pi-scf option or not! Their are only minor differences in calculated geometry but significant differences in MMX energy and heat of formation. But if you assign all carbons to be "aromatic" (atom type 40), the system is forced to behave like an aromatic pi-system and you get of course a planar structure with no bond-alternation; the scf pi-procedure is automatically deactivated when atom type 40 is selected. But why choose "aromatic" carbons, COT is not aromatic! Jens >--< =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= JENS SPANGET-LARSEN Department of Chemistry Phone: +45 46757781 + 2710 Roskilde University (RUC) Fax: +45 46757721 P.O.Box 260 E-Mail: JSL@virgil.ruc.dk DK-4000 Roskilde, Denmark http://frederik.ruc.dk/dis/chem/psos =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= From - Mon Apr 21 16:33:33 1997 From: Gerardo Burton To: "'hyperchem@hyper.com'" Subject: RE: new bug in MM+? Date: Mon, 21 Apr 1997 15:28:44 -0300 MIME-Version: 1.0 Content-Type: text/plain; charset="us-ascii" Content-Transfer-Encoding: quoted-printable X-Mozilla-Status: 0011 Content-Length: 5708 Hello Joel and Jens Actually I was trying to reproduce an exercise from a paper in J. Chem. = Ed. with the calculations done with MM2 in Macromodel. In PCModel I = used the standard MM2 which is supposed to be equivalent to MM+, I DID = NOT use the extension for conjugated pi systems in MMX (i.e. mark pi = atoms and do SCF calculation to recalculate force constants for the = specific pi system), this is an option that you have to turn on, its = not the default. All carbon atoms were type 2 (i.e. alkene sp2) and = hydrogens were type 5. I checked the output file from PCModel and the = torsion energy expressions used in the force field (as described in the = manual) and it seems that PCModel doing JUST PLAIN MOLECULAR MECHANICS, = (i.e NO correction for conjugation with SCF calculations) takes into = account that the bond between 2 sp2 carbons may be single or double = (the parameters read are different for torsions involving single or = double bonds and also for bond lengths in the stretch term). So Joel = don't be so hard on PCModel, it does have its good thing. Actually PCModel predicts both conformers the flat and the tub shaped = with a 12.5 kcal/mol energy difference in favor of the tub shaped as = Jens mentioned. AM1 also predicts both conformers with practically the = same energy difference (ca 12.6 kcal/mol). AGAIN these calculations DO = NOT use the SCF option in PCModel, this is just plain MM. I guess that the MM2 implementation in Macromodel (from which PCModel = derived) probably also differentiates single and double bonds between = sp2 carbons and that is why the author of the J. Chem. Ed. paper didn't = have any trouble. By the way putting all carbons of cyclooctatetraene = as CA in PCModel would of course give the planar structure, but that's = cheating... Finally a word about Joel's comments on the corrections for pi systems = in PCModel. You should be aware that these corrections, i.e VESCF = calculation of the pi system to obtain corrected force constants, were = developed in part by Prof. Allinger (MMP2), incorporated in the latest = versions of MM2 and are an integral part of MM3, being very well = documented in the literature and in the MM3 manual. So I think that = this corrections could very well be implemented in Hyperchem MM+ = program. When I optimized ciclooctatetraene with MM3 (Allinger's original code) = it gave the flat conformer and a warning that pi atoms and VESCF = calculation should be included. Doing this, the tub-shaped conformer = resulted! Regards Gerardo Prof. Gerardo Burton e-mail: burton@qo.fcen.uba.ar Departamento de Quimica Organica Facultad de Ciencias Exactas y Naturales Universidad de Buenos Aires Tel/FAX 54 1 788-6915 ---------- From: Joel Polowin[SMTP:JoelP@agiss.com] Sent: Viernes 18 de Abril de 1997 17:06 To: burton; hyperchem; jsl Subject: RE: new bug in MM+? JENS SPANGET-LARSEN (JSL@virgil.ruc.dk) writes: > I just checked that PCMODEL does indeed predict a tubshaped > cyclooctatetraene, provided the initial configuration is not > completely planar (in this case, the molecule does not allways escape > from the hilltop, but remains planar). > To me, this is not surprising, since in PCMODEL, conjugated > pi-systems are treated by an SCF quantum chemical procedure, > predicting anti-aromatic destabilization of the 4N-perimeter > pi-system in planar cyclooctadiene (according to Huckel's rule). In > the tub-shaped configuration the destabilization is minimized, but > HyperChem MM+ does not include this contribution, since there is no > SCF treatment of pi-systems. Why is the SCF part omitted in MM+? The SCF stuff in PCMODEL's MMX isn't part of MM2; it was written by John McKelvey and added to the PCMODEL code. (In fact, when I discovered errors in the data storage procedure for pi systems during file-saving in PCMODEL, and requested clarification from Kevin Gilbert of Serena Software, he told me that he couldn't help -- McKelvey was the only one who knew how that code was supposed to work!) MMX running with the SCF calculations is a peculiar hybrid between molecular mechanics and semi-empirical calculations. But that explains why PCMODEL is giving the different results; it's not running as a simple molecular mechanics program. The interaction between the SCF stuff and the molecular mechanics calculations is a bit strange; it involves automatic revision of the force constants for planar systems' torsion angles with each iteration between SCF calculations and optimizations. (It also had a nasty bug which I spent several frustrating months trying, unsuccessfully, to track down, involving an interaction between the dihedral driver and the SCF calculations. After the first dihedral value was done, with each remaining dihedral angle, the force constants for planar systems would change sign; each "planar" torsional angle would suddenly try to go anti-planar, and the structure would do a somersault or try to tie itself into knots. And then with the next SCF iteration, everything would try to go planar again. This didn't happen with all structures, but I did get it to happen consistently with -- if I recall correctly -- using the dihedral driver on the CH3 group of propene, where the other two C atoms were marked for pi calculations. I don't know if that bug has been fixed in later versions of PCMODEL.) If you specify that the carbon atom types for cyclooctatetraene in PCMODEL are "CA" (aromatic carbon) or other SP2 type, and omit the pi calculations, you will probably find that the structure will go planar. Joel joelp@agiss.com From - Wed Apr 23 08:00:22 1997 Message-Id: <335D5E92.7291@cec.uchile.cl> Date: Wed, 23 Apr 1997 01:57:59 +0100 From: Alexandre Hocquet Reply-To: ahocquet@tamarugo.cec.uchile.cl Organization: ucsf, department of anesthesia X-Mailer: Mozilla 3.01 (Macintosh; I; PPC) Mime-Version: 1.0 To: hyperchem@hyper.com Subject: MM+ conjugation... References: <199704190931.FAA22543@relay.hyper.com> Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit content-length: 2230 X-Mozilla-Status: 0001 Content-Length: 2278 ... and other interesting features like treatment of missing parameters (isn't this title too long ?) Dear Hyperchemists, There seems to begin a discussion about MM+ features like assesment of default parameters or treatment of conjugation (with the example of cyclooctatetraene), especially compared to MMX. May i give my two cents contribution ? First of all, i do not think that MMX implemented a "true" MM2 code. A quick look at the Gundertofte test (J. Comput Chem march or april 1996) shows some differences INCLUDING some examples where comjugation is the key. I do not know if the MMX VESCF is exactly the same as in the MM2 (there are actually several of them : see Liljefors JcomputChem 1987 and Sprague JComput Chem 1987) (By the way sorry for those inaccurate quotings, if you have the patience to wait until i get back to chile, i will be happy to provide the exact ones on email request) but i doubt it. It's funny that my questions about MMX to McKelvey were answered that Gilbert knows it whereas Joel Polowin experienced the opposite. Anyway, i never could see the book reference where MMX is described so the mystery seems unresolved. What about MM+ anyway ? Of course, MM+ does not provide a SCF treatment of conjugation like MMX or the MM2 version of Chem3D. BUT, it does provide a simple bond type based procedure to deal with it (which is not the case of Spartan's MM2, for example, that will optimise butadiene with 3 equal bonds). "Why?" said Jens Spanget Larsen. I guess the reason is that the MM+ force field was designed to be suitable for molecular dynamics and for that purpose, as simple as possible. After this long introduction, i am pleased to announce that i am currently performing an evaluation of the MM+ forcefield on that matter (a poster about it was presented in the last ACS meeting in San Francisco) It compares the results of Hyperchem on a variety of conjugated molecules compared to MM2. Of course the results are worse, but in some cases resulsts are acceptable and in others case, not. This evaluation will deal with default parameters in a second step, by comparing with default parameters in DREIDING and in MM3. If anyone has any comment on this, please mail on ... Date: Thu, 24 Apr 1997 16:38:36 -0400 (EDT) From: Anna Gorokhov To: hyperchem@hyper.com Subject: hyperchem log files Message-Id: Mime-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII X-Mozilla-Status: 0001 Content-Length: 470 I am trying to read a log file of a semi-empirical ZINDO/S calculation with singly excited CI on a butadiene molecule. The file indicates an occ.mo to unocc.mo coefficient for every transition. I was wondering what that refers to and how I can use it if I wanted to generate a wavefunction of butadiene by starting with a ground state and adding the various CI states into it. Thank you. Anna Gorokhov Washington and Lee University Lexington, VA. From - Fri Apr 25 07:39:19 1997 Date: Thu, 24 Apr 1997 15:49:28 +0100 (BST) From: Kat Street X-Sender: pcxks@granby To: Hyperchem Mailing List Subject: ECP Basis sets Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII X-Mozilla-Status: 0001 Content-Length: 2507 Hi all, We have version 4.5 of Hyperchem. I want to create a basis set for Platinum, but the only practical way to do this would seem to be using Effective Core Potentials. Can Hyperchem cope with ECPs? If so, please can someone explain the following notation to me and give me an idea of how to use it?! Thanks very much, Kat Street Basis Query for Pt ECP1 (taken from http://www.tc.bham.ac.uk/molpro/basisform.html) ECP Pt ECP1 Core charge=68 ECP1, Er core, P.J. Hay, J. Chem. Phys. 82,270 m Exponent Contraction Coefficient V(4) 0 415.29905 -0.16193 1 117.99010 -44.10566 2 41.27165 -225.82083 2 11.85728 -99.34621 2 3.54521 -20.92722 2 1.20315 -2.75094 V(0)-V(4) 0 118.31236 2.73342 1 40.22698 76.87883 2 13.64206 208.32206 2 3.98844 60.03122 2 0.92467 25.98122 2 0.72601 -8.64182 V(1)-V(4) 0 111.79319 1.88786 1 36.45269 57.10330 2 11.92715 150.67809 2 3.25539 37.29851 2 0.77838 12.30816 V(2)-V(4) 0 92.31127 2.93437 1 39.37729 52.86911 2 14.69183 170.64111 2 3.02341 99.83539 2 2.56670 -54.31707 2 0.40191 -0.28729 V(3)-V(4) 0 1.65246 3.95343 1 23.21552 44.98893 2 6.73369 85.31119 2 1.49854 28.76190 2 0.35673 6.88557 2 0.26996 -2.67428 ________________________________________________________________________ Okay, boring .sig type-thing: Kat Street - http://www.nottingham.ac.uk/~pcxks/kat.htm Physical Chemistry, University of Nottingham. Tel: 0115 951 3455 ________________________________________________________________________ From - Fri Apr 25 07:39:19 1997 Received: from 149.174.217.136 by pp200.hyper.com with SMTP (Microsoft Exchange Internet Mail Service Version 5.0.1457.7) id JS2L4WQP; Thu, 24 Apr 1997 15:08:22 -0400 Received: by arl-img-6.compuserve.com (8.6.10/5.950515) id LAA08470; Thu, 24 Apr 1997 11:09:14 -0400 Date: Thu, 24 Apr 1997 11:08:26 -0400 From: "CYCLOLAB R&D Lab." Subject: Re: Re: calculation of spectra To: Krassimir Stavrev Message-ID: <199704241107_MC2-14DE-B4B@compuserve.com> X-Mozilla-Status: 0010 Content-Length: 7898 Hi Krassimir, thank you for the answers. Here are the short readme files (I hope there are not too much grammatical errors): ******1. NewOpen The HyperChem's manual contains a sample Excel macro for opening HyperChem from an external program. That macro is written in Excel 4 macro language. The Visual Basic for Applications has became "official macro language" at MicroSoft what has much more flexibility than the original Excel macro language. Before the use, you must enter your path for HyperChem in HCPath$ = "d:\chem\hc\exe45\" line. If you had problems with HyperChemMode=6, you could substitute the value from 6 to 4 (not selected, non-iconized run of an application). This is necessary in case of using Windows' (and HyperChem's) Copy/Cut-Paste procedures, otherwise you will get an HyperChem internal erro with unrealistic paste of molecule(s). The reason is unknown. The NewOpen macro replaces that old version and has some other new features: - you can select a (molecule) file for HyperChem using an Excel file open facility; - you can call this routine from other modules/application; - you wil get the channel number; - you will get not only the filename but its path too. How to call this macro? syntax1: CALL GetFileToHyperChem(Channel, FilePath$, Filename$) or syntax2: GetFileToHyperChem Channel, FilePath$, Filename$ In case of syntax 1, please check for parentheses around the parameters! Other tips&triks are documented in the macro's remark lines. ******2. NewCage It is a tool for generating more molecules from one molecule with random locations of a certain, (user) predefined space. You can use it to generate molecules for simulating solvents. This macro DOES NOT optimize the created "solvent". It is advisable to make a geometry optimization by a semiempirical method on the molecule what is intended to be solvent. It puts molecules randomly into a certain box of space and creates a solvent periodic box around them. The dimension of the resulted periodic box is about 1.5 times larger (in one dimension) than the given box for generating. This macro generates the location of molecules similar way as Timothy Bender's FillCage macro but works completely different way. It creates an HIN file without opening HyperChem and therefore in case of small molecules (<50 atoms) works at least 5 times faster than the other macro. Due to the method of creation some special cases (like hexachloroethane, tetrachloroethylene, etc.) you have to check that all molecules are within the solvent box or not and the "excluded" molecules must be translated "manually" into the solvent box. In some certain cases of cyclic molecules (benzene, cyclohexane, etc.) you have to check the creation for "magic rings". The formation of "magic rings" was found to be incidental and it seems to depend on the generated random numbers. If you want to create more molecules from a file which contains more than 1 (one) molecule, use Timothy's macro! This macro were checked under the following systems (by L. J.): Hyperchem 3 and 4.5 Windows 3.11 and Win'95 with the following molecules: H2O NH3 CH3NH2 CH4 CHCl3 MeOH C2H4 C2Cl2H2 C2Cl6 C6H6 toluene phenol phenolate cyclohexane - First you have to select a molecule from your deposit; In case of 16 bit version of HyperChem you must use not only maximum 5 character for the filename but your path must not contain more than 8 character fragments! This macro does not check any of above parameters! - then you can determine the number of molecules; - you can set the approximate dimensions of the solvent cage; - the macro generates your "solvent" file on the name of "MYMOL_sb.hin" where the name of your molecule is MYMOL. - opens the HyperChem and arranges the molecules into the periodic box; - saves the file under the same name; - leaves you "solvent" on the screen for further studies. ******************************* > >**your cage micro sets nicely the solvent box. Can you > suggest how to put the solute inside? Manual docking > would be the simplest (and perhaps most primitive) way > of doing so, followed by a minimization. The other, more > scientific way, would be to have a packing algorithm that > fits the two (solvent cage, and solute) the way it is > done now in HyperChem. I've spent plenty of time to set all molecules into the box, and I found that "manual" action is necessary. Why the quoted manual action. Because this step requires the opening of HyperChem. Now, I'me telling to you step-by-step the whole process (more detailed as in the proposed readme): (0. gets the location of HyperChem) 1. selection of the file The use of excel's built-in function to select your molecule has a disadvantage (what is in our case an advantage!). Namely, this built-in function does not open the selected file! So, you can import your file into an Excel worksheet. 2. there are constant elements (0) in the transformation matrices, what are generated. I know, that in Basic all variables have the zero starting values but I don't like miracles in programs. 3. creates a worksheet and imports the selected molecule file 4. calculates the number of atoms in the molecule (for dimensioning). Unfortunately, it calculates the number of atoms from the to lines, i.e. for multiple molecules in a file this macro is not suitable. 5. gets the dimensions for building 6. creates a file for the "solvent" 7. creates a copy of the comment section of the basic molecule. In this step the macro writes the dimensions, view, location, etc. of the solvent box. 8. puts (writes) the molecules 9. finishes the creating of HIN file and closes the (temporary) worksheet 10. starts or activates the HyperChem program 11. select the all molecules in the system and translates them into the origo (I supposed this means a geometric centre of all molecules). This step is "manual" action but the macro drives it. I found that in all cases the molecules are in a certain part of the space what is different from the created solvent box. I have not wanted to spend more time to create all molecules within the box, and I suspect this would be longer than it is worth, so I made a little cheat and I select all molecules and translates them into the origo. By this way all the molecules are taken within the solvent box. 12. saves the molecule 13. closes the macro I do hope that there are not too much redundacies or inconsistencies in the macro. Why have not I used my new open macro? The answer is simple: I've started to prepare the transciption of Timothy's macro before Victoria asked me about the generally usable Open macro. I have finished the transcription but I was not so satisfied with the results. Otherwise, Victoria mentioned a lot of inconsistencies and contradictions in the transcripted macro (origined from a simple translation and I did not intend to re-write anything), so I was full with the corrections. Finally, I got an almost completely different program. During the test period I was boring due to its slowness, and I've decided to write a new macro (despite my original plan), independently from Timothy's applet. In that step I changed step-by-step the modules and I did only the necessary changes. Meanwhile, Victoria asked to wrirte a general purpose open macro to substitute the previous sample module. I found that it would be too much work to change the idea in the "solvent cage project", and I kept my original version for opening HyperChem. Regards, Laszlo From - Fri Apr 25 10:09:06 1997 Message-ID: <3360BB02.20E1@hyper.com> Date: Fri, 25 Apr 1997 10:09:06 -0400 From: Krassimir Stavrev Organization: Hypercube, Inc. X-Mailer: Mozilla 3.0 (Win95; I) MIME-Version: 1.0 To: Anna Gorokhov Subject: Re: hyperchem log files References: Content-Type: text/plain; charset=us-ascii Content-Transfer-Encoding: 7bit X-Mozilla-Status: 0011 Content-Length: 1157 > I am trying to read a log file of a semi-empirical ZINDO/S calculation > with singly excited CI on a butadiene molecule. The file indicates an > occ.mo to unocc.mo > coefficient for every transition. I was wondering what that refers to and > how I can use it if I wanted to generate a wavefunction of butadiene by > starting with a ground state and adding the various CI states into it. Hi Anna, The coefficient you see is the unsquared eigenvector for the corresponding CI state /eigenvalue/. Hyper- chem prints it, at lower print levels, only if it is larger than 0.3.. in absolute value, but to form the entire CI eigenfunction, you would need to go higher quantum print levels, 6 and larger to get the entire information from the diagonalization of the CI matrix. The occ->unocc is the MO excitation which is involved in the transition from the ground CI state to this particular excited CI state. Regards, Krassimir --- Krassimir Stavrev, PhD Hypercube, Inc. Florida Science and Technology Park 1115 N.W. 4th Street Gainesville, Florida 32601 Voice: 1-352-371-7744/1-800-960-1871 Fax: 1-352-371-3662