From noertema@theochem.tu-muenchen.de Sat Jul 18 09:06:07 1998 Received: from theo1.theochem.tu-muenchen.de (theo2.theochem.tu-muenchen.de [129.187.191.1]) by www.ccl.net (8.8.3/8.8.6/OSC/CCL 1.0) with SMTP id JAA22344 Sat, 18 Jul 1998 09:06:06 -0400 (EDT) Received: from theo3.theochem.tu-muenchen.de (theo3.theochem.tu-muenchen.de [129.187.157.6]) by theo1.theochem.tu-muenchen.de (950413.SGI.8.6.12/950213.SGI.AUTOCF) via ESMTP id PAA11843 for ; Sat, 18 Jul 1998 15:06:06 +0200 Received: by theo3.theochem.tu-muenchen.de (SMI-8.6/SMI-SVR4) id PAA02479; Sat, 18 Jul 1998 15:06:13 +0200 Date: Sat, 18 Jul 1998 15:06:13 +0200 From: noertema@theochem.tu-muenchen.de (Folke Noertemann) Message-Id: <199807181306.PAA02479@theo3.theochem.tu-muenchen.de> To: chemistry@www.ccl.net Subject: x,y,z vs. internal coordinates X-Sun-Charset: US-ASCII Cartesian <-> internal coordinates: This question has to be answered with is clear "It depends". Usually cartesian coordinates introduces a great deal of undesired couplings. One may get rid of them by choosing the right internal coordinates. My experience is that nothing can beat a well-chosen Z-matrix. Other internal coordinate systems, e.g. Pulays natural internal coordinates compare also very well in the literature. However, the choice of the coordinate system depends -at least in my view- entirely on two things: 1. The nature of the molecule one wishes to optimize. Establishing a Z-matrix for polycyclic systems is a nightmare that nobody really wants to go through. Some people claim that it is impossible to compose a good Z-matrix for such a system. So if the only two options available are cartesians and Z-matrix AND the molecule is unpleasantly complex, cartesians should be preferred. In agreement with this, many papers mention that cartesians are superior to anything else for polycyclic molecules. However, these statements are NOT independent of 2.) 2. The quality of the initial Hessian that is available. Because of the couplings they introduce, cartesians do in general not have the best reputation. However, Baker and Hehre showed that if a good approximation to the Hessian is available, cartesians perform comparably to other sets of internal coordinates. They also stress, that for FLEXIBLE polycyclic systems (whatever that may be) they found the performance of cartesian coordinates to be below internal coordinates. A few of the relevant references are: J.Baker and W.J. Hehre, J.Comp.Chem. 12, 606 (1991) J.Baker, J.Comp.Chem. 14, 1085 (1993) J.Baker and D.Bergeron, J.Comp.Chem. 14, 1339 (1993) P.Pulay and G.Fogarasi, J.Chem.Phys. 96, 2856 (1991) Hope this helps, Folke Folke Noertemann Lehrstuhl f. Theoret. Chemie TU Muenchen noertema@theochem.tu-muenchen.de From idimitr@cc.uoi.gr Sat Jul 18 10:33:41 1998 Received: from cc.uoi.gr (solon.cc.uoi.gr [193.92.4.5]) by www.ccl.net (8.8.3/8.8.6/OSC/CCL 1.0) with ESMTP id KAA22615 Sat, 18 Jul 1998 10:33:39 -0400 (EDT) Received: (from idimitr@localhost) by cc.uoi.gr (8.8.4/8.8.3) id TAA01009; Sat, 18 Jul 1998 19:35:01 +0200 (GMT+0200) Date: Sat, 18 Jul 1998 19:35:01 +0200 (GMT+0200) From: Ioannis Dimitropoulos Message-Id: <199807181735.TAA01009@cc.uoi.gr> To: MARYJO@neu.edu, chemistry@www.ccl.net Subject: Re: CCL:gradients Dear Mary, On Fri, 17 Jul 1998 you wrote asking about programs to calculate the gradient of a function. I would suggest to look at the UR http://nrt.cs.uoi.gr/merlin/ The source code of the program can be downloaded - though the program is an environment for multidimensional minimization it provides for analytical and numerical gradients (hessians as well) - the user has to provide the function of cource, regards I.N. Demetropoulos Chemistry Dept Ioannina,Greece From mckelvey@ncsa.uiuc.edu Sat Jul 18 11:23:50 1998 Received: from ex1.ncsa.uiuc.edu (ex1.ncsa.uiuc.edu [141.142.3.16]) by www.ccl.net (8.8.3/8.8.6/OSC/CCL 1.0) with ESMTP id LAA22816 Sat, 18 Jul 1998 11:23:49 -0400 (EDT) Received: from mckelvey (stolen40-240.ncsa.uiuc.edu [141.142.40.240]) by ex1.ncsa.uiuc.edu (8.9.1/8.8.8) with SMTP id KAA23956; Sat, 18 Jul 1998 10:23:47 -0500 (CDT) Message-Id: <3.0.5.32.19980718102517.009a6540@pop.ncsa.uiuc.edu> X-Sender: mckelvey@pop.ncsa.uiuc.edu X-Mailer: QUALCOMM Windows Eudora Pro Version 3.0.5 (32) Date: Sat, 18 Jul 1998 10:25:17 -0700 To: E.Tajkhorshid@DKFZ-Heidelberg.de From: John McKelvey Subject: Re: CCL:G:Dipole moment of a charged species (Summary) Cc: chemistry@www.ccl.net In-Reply-To: <9807161610.ZM9140@mbp-sgi7.inet.dkfz-heidelberg.de> Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii" X-NCSA-md5: d22ad4762c950c59f3b402cd8f78c092 Content-MD5: 0irUdiyVDFnztALNj3jAkgAA Content-MD5-Origin: ex1.ncsa.uiuc.edu An origin independent dipole moment of a charged molecule _is not defined_. The leading term of the multipole moment expansion is Q*Z where Q is the molecular charge and Z represents a position in space. If a molecule is charged then some _arbitrary_ choice has to be made for Z; thus you can have a dipole moment for a molecule if you are convinced about the location of Z. John McKelvey NCSA At 04:10 PM 7/16/98 -0700, you wrote: >Thank you very much for your replies. > >However, it is still not clear to me whether I can report the dipole moments of >two charged species in a scientific paper or not. I mean, can I compare dipole >moments of molecule_1 and molecule_2 (both negative ions, and the dipole >moments are calculated after transforming the coordinate to a mass-weighted or >charge-wieghted system) and can I say that the dipole moment of 1 is larger >than 2, or if we align two molecules, the direction of the dipole vector with >respect to the carbon chain is the opposite for molecule 1 as compared to the >one in molecule 2. Since I would like to report these numbers in a scientific >papers, I am a bit conservative. Perhaps I may pose the question in another >way, asking: >Do you know any paper where the dipole moments of the ionic species are >compared and/or discussed? Or the people prefer not to discuss these numbers in >their paper .... > >The Q and summary of replies: > >*************************************************************************** **** >Dear Colleagues > >I have calculated two molecules using gas phase calculations and a number of >different solvent models. They are both negatively charged species. Now the >question is whether and how we can compare the dipole moments between these >molecules and among different solvent models. I mean, the values, and also the >directions, calculated for the dipole moment of a charged species are >completely dependent on the origin of the coordinates. With this regard, how >are the dipole moments calculated in quantom chemical calculations >(particularly, I am interested in the reported values in the output of the >Gaussian94 or DMOL programs)? And (how) can we use them for further analysis of >the results or it is better to look at other properties like the atomic >charges? > >*************************************************************************** *** >From: Wayne Steinmetz >Date: Tue Jul 14, 10:17am -0700 >To: E.Tajkhorshid@DKFZ-Heidelberg.de >Cc: >Subject: Re: CCL:G:G:Dipole moment of a charged species > >Be very careful since the species to be compared possess net charge. >It is a consequence of the principles of electricity and magnetism that >the dipole moment is dependent on the choice of the origin when the >molecule has net charge. In fact, the dipole moment can be >made zero by choosing the origin of the charge distribution at the center >of the charge. The value of the dipole moment is independent of the origin >when the molecule has no net charge. > >***************************************************************************** >From: liang@wavefun.com >Date: Tue Jul 14, 10:35am -0700 >To: E.Tajkhorshid@dkfz-heidelberg.de >Cc: liang@wavefun.com >Subject: Re: CCL:G:G:Dipole moment of a charged species > >Dear Emad, > >For an N point charge system of total charge Q, the dipole moment is > >D = sum_i [ x_i * q_i ] > >where x and q are position vector and charge on point i. By translating >to the center of geometry, > >X = (1/N) sum_i [ x_i ] > >one obtains > >D = sum_i [ (x_i - X) * q_i ] + X * sum_i [ q_i ] > = D0 + X*Q > >D0 is a value that is independent of the positions of the point charge and >therefore may be used for comparison purposes. For a continuous charge >distribution, substitute an integral for the sum. > >Regards, >Liang > >*************************************************************************** >From: Doug Fox >Date: Tue Jul 14, 12:19pm -0400 >To: uunet.uu.net!uunet!DKFZ-Heidelberg.de!E.Tajkhorshid%lorentzian.com >Cc: >Subject: Re: CCL:G:G:Dipole moment of a charged species > > Emad, > > The dipole moment of a charge species includes a contribution from >the separation of charge and the center of nuclear charge. Thus one >approach is to compare dipoles with the origin at the center of charge >to zero this contribution. In Gaussian this is the origin for the >standard orientation and the orientation used in multipole report in the >log file, not in the archive entry or if you use NoSymm. I don't know >what DMOL reports but it should be fairly simple to correct one to the >other. > >**************************************************************************** >From: John McKelvey >Date: Tue Jul 14, 2:06pm -0700 >To: E.Tajkhorshid@dkfz-heidelberg.de >Cc: >Subject: Re: CCL:G:G:Dipole moment of a charged species > >Emad, > >I like using the charge-reference-coordinate as being the neutral molecule >valence-electron-weighted (x,y,z) point in space. It puts the emphasis on >where the meaningful chemistry is. > >Mass-weighted systems might also be useful, but there could be a >significant shift, for example, if a H or F atom were replaced by, say, >Iodine. > >Regards, > >John McKelvey >NCSA > >*********************************************************************** >From: Preston J. MacDougall >Date: Tue Jul 14, 1:42pm -0700 >To: E.Tajkhorshid@DKFZ-Heidelberg.de >Cc: >Subject: atomic first moments > >Dr. Tajkhorshid, > You are correct that the dipole moment of a charged species is >origin dependent. Packages must select origins anyway, so the computed >dipole moment must be taken for what it is worth. You can assign >first moments to the consituent atoms in a molecule or ion, whether >there are partial charges on them or not (see J. Chem. Phys., vol. 87, >p. 1142 (1987)). The first moment of an atom is also origin dependent, >but is naturally chosen as the nucleus. I am not sure, but I believe >that these first moments can be calculated from a G94 wavefunction using >the AIMPAC option. If not, the AIMPAC software can be downloaded off of >the internet at http://www.chemistry.mcmaster.ca/aimpac/ > >Hope this helps in your research, > >Dr. Preston J. MacDougall >NASA-ASEE-Stanford Summer Faculty Fellow >MS T27-A, NASA Ames Research Center >Moffett Field, CA 94035 > >************************************************************************** > > >-- >Emad >********************************************************************* >E. Tajkhorshid >German Cancer Research Center; DKFZ Tel: +49 6221 42 2340 >Dept. Molecular Biophysics (H0200) FAX: +49 6221 42 2333 >P.O.Box 101949 >69009 Heidelberg, Germany Email: E.Tajkhorshid@DKFZ-Heidelberg.de >********************************************************************* >* "Never express yourself more clearly than you think." -Niels Bohr * >********************************************************************* > >-------This is added Automatically by the Software-------- >-- Original Sender Envelope Address: tajkhors@mbp-sgi7.inet.dkfz-heidelberg.de >-- Original Sender From: Address: tajkhors@mbp-sgi7.inet.dkfz-heidelberg.de >CHEMISTRY@www.ccl.net: Everybody | CHEMISTRY-REQUEST@www.ccl.net: Coordinator >MAILSERV@www.ccl.net: HELP CHEMISTRY or HELP SEARCH | Gopher: www.ccl.net 73 >Anon. ftp: www.ccl.net | CHEMISTRY-SEARCH@www.ccl.net -- archive search > Web: http://www.ccl.net/chemistry.html > >