From chemistry-request@server.ccl.net Thu Nov 30 02:36:36 2000 Received: from qclgw.qcl.t.u-tokyo.ac.jp (qclgw.qcl.t.u-tokyo.ac.jp [133.11.70.190]) by server.ccl.net (8.8.7/8.8.7) with ESMTP id CAA20829 for ; Thu, 30 Nov 2000 02:36:35 -0500 From: ajitha@qcl.t.u-tokyo.ac.jp Received: from ashura.qcl01.qcl.t.u-tokyo.ac.jp (ashura.qcl01.qcl.t.u-tokyo.ac.jp [192.168.1.1]) by qclgw.qcl.t.u-tokyo.ac.jp (8.8.7/3.7W/MGATE-20000124) with ESMTP id QAA19347; Thu, 30 Nov 2000 16:35:30 +0900 Received: (from ajitha@localhost) by ashura.qcl01.qcl.t.u-tokyo.ac.jp (8.9.3/3.7W/MHOST-19990324) id QAA81456; Thu, 30 Nov 2000 16:35:30 +0900 Date: Thu, 30 Nov 2000 16:35:30 +0900 Message-Id: <200011300735.QAA81456@ashura.qcl01.qcl.t.u-tokyo.ac.jp> To: CHEMISTRY@server.ccl.net Subject: Hirshfeld charges Cc: ramesh@qcl.t.u-tokyo.ac.jp, roy@qcl.t.u-tokyo.ac.jp Dear Colleagues : There is a famous paper by Hirshfeld (Theoret. Chim. Acta(Berl) 44, 129(1977))which defines the atomic charges by using the so-called stockholder partitioning technique. These charges are derived in terms of a weight function w(r) which he has defined on p 130 of the paper. However, the analysis has normally been applied to neutral molecules. A specific question I have is : How to get the weight function for a cationic or anionic species? The main problem there is the correct definition of premolecule. Has anyone seen such an analysis for molecular cation or anion? How is the w(r) chosen in such situations? Thank you in anticipation!...........Ram Kinkar Roy From chemistry-request@server.ccl.net Thu Nov 30 10:24:15 2000 Received: from silver.ch.unito.it (silver.ch.unito.it [130.192.118.1]) by server.ccl.net (8.8.7/8.8.7) with ESMTP id KAA23701 for ; Thu, 30 Nov 2000 10:24:14 -0500 Received: from chpc47.ch.unito.it (chpc47.ch.unito.it [130.192.118.47]) by silver.ch.unito.it (8.9.3/8.9.3) with SMTP id QAA04891; Thu, 30 Nov 2000 16:39:53 +0100 (MET) Message-Id: <3.0.3.32.20001130162159.00796850@silver.ch.unito.it> X-Sender: teorica@silver.ch.unito.it X-Mailer: QUALCOMM Windows Eudora Light Version 3.0.3 (32) Date: Thu, 30 Nov 2000 16:21:59 +0100 To: chemistry@ccl.net From: Theoretical Chemistry Group Subject: MSSC2001 - Modeling in Solid State Chemistry Cc: mssc@ch.unito.it Mime-Version: 1.0 Content-Type: text/plain; charset="us-ascii" European Summerschool Ab initio Modeling in Solid State Chemistry MSSC2001 September 11-15, 2001 -Torino, Italy Director R. Dovesi, University of Torino The Theoretical Chemistry Group of the Torino University is organizing a Summer School on the ab initio simulation of crystalline and defective solids. The school is addressed to PhD students with interests in solid state chemistry, physics, materials science, surface science, catalysis. Students will be expected to be familiar with the basic theoretical apparatus of quantum chemistry and solid state physics. The school will provide an overview of the possibilities offered by ab initio quantum mechanical techniques when applied to characterize solid state materials. The capabilities of the new release of the CRYSTAL code, CRYSTAL2001, will be illustrated during afternoon sessions in our computer room. For further information send a mail to mssc@ch.unito.it, and see: http://www.ch.unito.it/ifm/teorica/mssc2001/annuncio.html Theoretical Chemistry Group University of Torino Via Giuria 5 - I 10125 Torino Italy Phone: +39 011 670 7564 Fax: +39 011 670 7855 From chemistry-request@server.ccl.net Thu Nov 30 10:18:57 2000 Received: from ccl.net (atlantis.ccl.net [192.148.249.4]) by server.ccl.net (8.8.7/8.8.7) with ESMTP id KAA23678 for ; Thu, 30 Nov 2000 10:18:57 -0500 Received: from oscdesk53.ccl.net (oscdesk53.ccl.net [192.148.249.53]) by ccl.net (8.9.3/8.9.3/OSC 2.0) with ESMTP id KAA17178; Thu, 30 Nov 2000 10:18:42 -0500 (EST) Date: Thu, 30 Nov 2000 10:18:42 -0500 From: Gerald Lushington To: chemistry@ccl.net cc: gerald@ccl.net Subject: G98 / displacement issue with frozen coordinates Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Dear all, In working with a colleague on large cluster optimizations of enzyme active site (via G98), I've encountered a problem that seems to be: - part philosophical - part methodological - part bug (?) Anyway, the issue has to do with doing geometry optimizations with some degrees of freedom contrained. In general it's proving necessary to freeze the coordinates for at least a couple of atoms in order to keep the cluster in some sort of shape comparable to the real biosystem. The tradeoff, of course, comes in the fact that forces and displacements will never achieve global minima. Standard protocol seems to be to run the job until these quantities level off to some reasonably converged values. When we apply this strategy we find that force and net displacement eventually level off, but the net displacement remains very high. This brings me to question number 1: 1) If we stop the geometry optimization at a structure whose force and displacement values have leveled, but whose displacement is still high, would any subsequent vibrational analysis be at all physically meaningful? Are there any effective guidelines for this? My first thought on this was that Gaussian would likely decompose both force and displacement as a sum over coordinates. Since the only freezing we do is way out on the perifery of the cluster, while the structural and vibrational information of interest is in the center of the system, maybe we can track the forces and displacements in the central region of interest and feel comfortable about the results when those central contributions are low. So, to try to follow that strategy, we've been refering to a table generated as standard G98 output during the geometry optimization that looks like: Variable Old X -DE/DX Delta X Delta X Delta X New X (Linear) (Quad) (Total) X1 -2.50100 -0.00465 0.00000 0.00000 2.50100 0.00000 Y1 8.90656 0.01749 0.00000 0.00000 -8.90656 0.00000 Z1 -8.10547 -0.01647 0.00000 0.00000 8.10547 0.00000 X11 -3.32922 0.00421 0.00000 0.00000 2.50100 -0.82822 Y11 -9.47808 -0.00542 0.00000 0.00000 -8.90656 -18.38465 Z11 -1.01768 -0.00913 0.00000 0.00000 8.10547 7.08779 X38 3.49152 0.00310 0.00000 0.00000 2.50100 5.99252 Y38 14.83066 0.00184 0.00000 0.00000 -8.90656 5.92410 Z38 -5.24649 -0.00415 0.00000 0.00000 8.10547 2.85898 X48 -1.96597 0.00535 0.00000 0.00000 2.50100 0.53502 Y48 -4.25077 -0.01552 0.00000 0.00000 -8.90656 -13.15733 Z48 9.58721 0.02164 0.00000 0.00000 8.10547 17.69267 X50 1.54048 -0.00800 0.00000 0.00000 2.50100 4.04148 Y50 -8.97461 0.00161 0.00000 0.00000 -8.90656 -17.88117 Z50 5.17857 0.00810 0.00000 0.00000 8.10547 13.28404 R1 2.84387 0.01822 0.00000 0.00000 0.00000 2.84388 R2 2.02201 0.01493 0.00000 0.00000 0.00000 2.02201 R3 2.02201 0.01403 0.00000 0.00000 0.00000 2.02201 R4 2.02201 0.01490 0.00000 0.00000 0.00000 2.02201 R5 2.81303 -0.01054 0.00000 -0.01044 -0.00995 2.80307 R6 2.10327 -0.02056 0.00000 -0.01706 -0.01706 2.08621 R7 2.10327 -0.01967 0.00000 -0.01633 -0.01633 2.08694 R8 2.60262 -0.00578 0.00000 -0.00448 -0.00457 2.59805 R9 2.55751 -0.00018 0.00000 -0.00230 -0.00185 2.55566 .. .. .. NOTE: in the above, the first bunch of coordinates (X1-R4 inclusive) are all frozen. R5 onwards (many coordinates not shown) are freely varying. Anyway, in the variable coordinates toward the bottom, it seems that the "DE/DX" is a real force, the "Delta X" are real displacements, and the "Old X" and "New X" are exactly as they sound -- initial and final values for the coordinate. However: 2) What's going on with these the "Delta X" and "New X" values for the frozen coordinates?? They're frozen (and really do not vary during the optimization) so shouldn't "New X" = "Old X" regardless of nonzero forces and predicted displacements? 3) Furthermore, the values written in "Delta X (Total)" look extremely fishy to me? They're all either "2.50100", "-8.90656", "8.10547" or "0.00000"! Is this a bug, and if so, might it be corrupting the value given for net displacement at the end of each geometry step? Anyway, this may be a complex issue, but any insight you may have would be greatly appreciated! - Gerry P.S. I did drop Gaussian Inc. a note on this several weeks ago -- I guess it's somewhere in their queue. ---------------------------------------------------------------------- Gerald H. Lushington Ph: 614-292-6036 Research Specialist Fax: 614-292-7168 OSC / PET-CCM e-mail: gerald@ccl.net 1224 Kinnear Road http://www.arl.hpc.mil/PET/cta/ccm Columbus OH 43212-1163 http://www.asc.hpc.mil/PET/CCM From chemistry-request@server.ccl.net Thu Nov 30 10:41:20 2000 Received: from mel.ruc.dk (mel.ruc.dk [130.225.220.27]) by server.ccl.net (8.8.7/8.8.7) with ESMTP id KAA23865 for ; Thu, 30 Nov 2000 10:41:19 -0500 Received: from smtprelay.ruc.dk (smtprelay.ruc.dk [130.225.220.22]) by mel.ruc.dk (8.9.3/8.9.1) with ESMTP id QAA17908; Thu, 30 Nov 2000 16:41:02 +0100 (MET) Received: from virgil.ruc.dk (virgil.ruc.dk [130.225.220.110]) by smtprelay.ruc.dk (8.9.3+Sun/8.9.3) with ESMTP id QAA07883; Thu, 30 Nov 2000 16:28:25 +0100 (MET) Received: from VIRGIL/SpoolDir by virgil.ruc.dk (Mercury 1.44); 30 Nov 00 16:41:02 +0100 Received: from SpoolDir by VIRGIL (Mercury 1.44); 30 Nov 00 16:40:58 +0100 From: "Jens Spanget-Larsen" Organization: Roskilde Universitetscenter To: Ali N Rashid Date: Thu, 30 Nov 2000 16:40:45 +0100 Subject: CCL:Energy of Transition State is higher that ground state. CC: chemistry@ccl.net Priority: normal X-mailer: Pegasus Mail for Windows (v2.23) Message-ID: Ali N Raschid: > Hi, I user HF6-31+G* and 6-311+G+ to optimize the ground state of > p-nitroamine. The amine come out pyramidal. When I hold the amine in a > planar conformation I get one negative frequency. The energy of the > pyramidal conformer is higher than that of the pyramidal one, but when I > take into consideration the zero point energy correction the planar > conformer (with the negative frequency) has lower energy than the > pyramidal one (all positive frequencies). > Does anyone have a clue as to why? Dear Ali, the ZPE corrections predicted for the equilibrium geometry and the transition state geometry are not strictly comparable. The transition state has one vibrational mode with a negative force constant and an imaginary frequency. This mode is not considered in the computation of the ZPE. - Anyway, the harmonic approximation is not reliable for this mode, that corresponds to the amine "inversion vibration". The molecule is situated in a double minimum potential well with a relatively small barrier between the minima. This situation is associated with large anharmonic effects, leading to tunnelling splitting (inversion doubling). Yours, Jens >--< =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= JENS SPANGET-LARSEN Phone: +45 4674 2000 (RUC) Department of Chemistry +45 4674 2710 (direct) Roskilde University (RUC) Fax: +45 4674 3011 P.O.Box 260 E-Mail: JSL@virgil.ruc.dk DK-4000 Roskilde, Denmark http://www.rub.ruc.dk/dis/chem/psos/ =-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-= From chemistry-request@server.ccl.net Thu Nov 30 18:12:38 2000 Received: from stillwater.cs.umt.edu (stillwater.cs.umt.edu [150.131.15.39]) by server.ccl.net (8.8.7/8.8.7) with ESMTP id SAA26135 for ; Thu, 30 Nov 2000 18:12:37 -0500 Received: from reimel.cs.umt.edu (reimel [150.131.15.44]) by stillwater.cs.umt.edu (8.9.3+Sun/8.9.3) with ESMTP id QAA10487 for ; Thu, 30 Nov 2000 16:12:36 -0700 (MST) Received: from localhost (cabal@localhost) by reimel.cs.umt.edu (8.9.3+Sun/8.9.1) with ESMTP id QAA05048 for ; Thu, 30 Nov 2000 16:12:35 -0700 (MST) X-Authentication-Warning: reimel.cs.umt.edu: cabal owned process doing -bs Date: Thu, 30 Nov 2000 16:12:35 -0700 (MST) From: Clair W Cabal X-Sender: cabal@reimel To: chemistry@ccl.net Subject: Looking for BABEL and DRAGON users Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII All, I am using Babel to convert a chemical data set from SMILES notation to MOL file notation so that I can use Prof. Todeschini's DRAGON program to generate theoretical descriptors. I think the conversion is OK (based on comparing my generated files to the sample files in DRAGON. However, I always get an error in trying to calculate the descriptors with Dragon. I was in hopes that someone with a real understanding of these two formats might point out my error. Here is the output from Babel for one molecule. This SMILES data does not even list the molecule name, so on line two Babel puts the SMILES representation of the molecule. I have added the .mol2 which is necessary for DRAGON to recognize this file. @MOLECULE O=C(Cn1ccnc1)c2ccc3ccccc3c2.mol2 18 20 0 0 0 SMALL NO_CHARGES @ATOM 1 O1 0.0000 0.0000 0.0000 O.2 1 <1> 0.0000 2 C1 0.0000 0.0000 0.0000 C.2 1 <1> 0.0000 3 C2 0.0000 0.0000 0.0000 C.3 1 <1> 0.0000 4 N1 0.0000 0.0000 0.0000 N.pl3 1 <1> 0.0000 5 C3 0.0000 0.0000 0.0000 C.2 1 <1> 0.0000 6 C4 0.0000 0.0000 0.0000 C.2 1 <1> 0.0000 7 N2 0.0000 0.0000 0.0000 N.pl3 1 <1> 0.0000 8 C5 0.0000 0.0000 0.0000 C.2 1 <1> 0.0000 9 C6 0.0000 0.0000 0.0000 C.2 1 <1> 0.0000 10 C7 0.0000 0.0000 0.0000 C.2 1 <1> 0.0000 11 C8 0.0000 0.0000 0.0000 C.2 1 <1> 0.0000 12 C9 0.0000 0.0000 0.0000 C.2 1 <1> 0.0000 13 C10 0.0000 0.0000 0.0000 C.2 1 <1> 0.0000 14 C11 0.0000 0.0000 0.0000 C.2 1 <1> 0.0000 15 C12 0.0000 0.0000 0.0000 C.2 1 <1> 0.0000 16 C13 0.0000 0.0000 0.0000 C.2 1 <1> 0.0000 17 C14 0.0000 0.0000 0.0000 C.2 1 <1> 0.0000 18 C15 0.0000 0.0000 0.0000 C.2 1 <1> 0.0000 @BOND 1 1 2 2 2 2 3 1 3 2 9 1 4 3 4 1 5 4 5 ar 6 4 8 ar 7 5 6 ar 8 6 7 ar 9 7 8 ar 10 9 10 ar 11 9 18 ar 12 10 11 ar 13 11 12 ar 14 12 13 ar 15 12 17 ar 16 13 14 ar 17 14 15 ar 18 15 16 ar 19 16 17 ar 20 17 18 ar With this exact input, the error I get in DRAGON when I attempt to calculate descriptors is: Errors reading molecules ..... 1 Mol. 1 d:\user\cabal\chemdata\out15-~2.mol Input past end of file. (62) Here is a sample mol2 file from DRAGON which runs in that program fine: # Name: 2-Iodophenol.mol2 # Creating user name: alex # Creation time: Tue Oct 10 10:25:54 2000 # Modifying user name: alex # Modification time: Tue Oct 10 10:29:16 2000 @MOLECULE 2-Iodophenol.mol2 13 13 1 0 0 SMALL NO_CHARGES @ATOM 1 I1 -0.0098 -0.1299 -0.6467 I 1 <1> 0.0000 2 C2 0.0262 -0.0379 1.4009 C.ar 1 <1> 0.0000 3 C3 0.0048 0.0872 4.2021 C.ar 1 <1> 0.0000 4 C4 1.2354 0.0420 2.1198 C.ar 1 <1> 0.0000 5 C5 -1.1835 -0.0552 2.1025 C.ar 1 <1> 0.0000 6 C6 -1.1943 0.0067 3.4943 C.ar 1 <1> 0.0000 7 C7 1.2186 0.1049 3.5225 C.ar 1 <1> 0.0000 8 O8 2.4126 0.0558 1.4091 O.3 1 <1> 0.0000 9 H9 -2.1324 -0.1185 1.5467 H 1 <1> 0.0000 10 H10 -2.1530 -0.0085 4.0328 H 1 <1> 0.0000 11 H11 2.1654 0.1677 4.0777 H 1 <1> 0.0000 12 H12 -0.0064 0.1370 5.3013 H 1 <1> 0.0000 13 H13 3.1420 0.1084 2.0448 H 1 <1> 0.0000 @BOND 1 1 2 1 2 2 5 ar 3 5 6 ar 4 6 3 ar 5 3 7 ar 6 7 4 ar 7 2 4 ar 8 6 10 1 9 7 11 1 10 3 12 1 11 4 8 1 12 5 9 1 13 8 13 1 @SUBSTRUCTURE 1 **** 1 TEMP 0 **** **** 0 ROOT Babel has not output any substructure information, so I have tried typing some into my file and then I get an overflow error in DRAGON. Any suggestions or feedback about what I might be doing wrong would be greatly appreciated - I'm really out of my depth in this chemistry stuff :-) Thank you. Clair Cabal Research Assistant Machine Learning Research Lab Department of Computer Science University of Montana Missoula, Montana email: cabal@cs.umt.edu fax: (406) 243-5139 From chemistry-request@server.ccl.net Thu Nov 30 15:28:08 2000 Received: from qtp.ufl.edu (zwart.qtp.ufl.edu [128.227.89.3]) by server.ccl.net (8.8.7/8.8.7) with ESMTP id PAA25360 for ; Thu, 30 Nov 2000 15:28:08 -0500 Received: from pc1 (pc1 [128.227.192.187]) by qtp.ufl.edu (8.9.1a/8.9.1) with SMTP id PAA08934; Thu, 30 Nov 2000 15:28:27 -0500 (EST) Message-ID: <06ae01c05b0c$7fd032e0$bbc0e380@qtp.ufl.edu> From: "Stefan Fau" To: "CCL - all" Cc: "Gerald Lushington" References: Subject: Re: CCL:G98 / displacement issue with frozen coordinates Date: Thu, 30 Nov 2000 15:31:20 -0500 MIME-Version: 1.0 Content-Type: text/plain; charset="iso-8859-1" Content-Transfer-Encoding: 7bit X-Priority: 3 X-MSMail-Priority: Normal X-Mailer: Microsoft Outlook Express 5.00.2314.1300 X-MIMEOLE: Produced By Microsoft MimeOLE V5.00.2314.1300 Hi, an interesting practical problem... Here are my 2 cents, comments invited. to 2) and 3) Delta X and New X are the differences and new values of the geometry descriptors. G98 allows mixed geometry specifications using internal coordinates together with cartesian coordinates. Freezing cartesian coordinates is equivalent to freezing all internal coordinates between this group of atoms (provided you have only one spatial group of frozen atoms). If you have two clusters of frozen atoms linked by a unconstrained part of the molecule, one way of optimizing the molecule is to allow movement of the two clusters with respect to each other. I don't think G98 does it this way, but you can check by tracking the distances between atoms of different frozen clusters over the optimization. Another reason for blockwise movement of the frozen atom clusters is that sometimes the standard orientation of the molecule flips from one optimization step to the next. (I think that the standard orientation is determined by the moments of inertia. Sometimes two of them are very similar and change sequence from one geometry to the next.) I think the list of geometry variables contains both the original set of geometry descriptors and the redundant internal coordinates. I am not sure if original descriptors that are NOT redundant internal coordinates enter in the computation of the displacement used as a convergence criterion. You may want to browse the CCL archives: I faintly remember a mail from Gaussian (maybe Doug Fox) on that topic a year or two ago. to 1) Vibrational frequencies are molecular properties. They all depend on all of the molecule. At the same time, some are quite specific for functional groups. Therefore the influence of other parts of the molecule must be small, probably rapidly declining with distance. The common implementation of harmonic frequencies requires the forces (=first derivatives of the energy w.r.t. nuclear positions) to be (nearly) zero. I don't know how the second derivatives change with displacements that cause nearly no forces, but the displacements must be along floppy modes with very small force constants and frequencies. I suppose that big residual displacements will mainly influence small frequencies. In the harmonic approximation you may diagonalize the matrix of mass weighted second derivatives of the energy with respect to nuclear coordinates and get the vibrational modes and vibrational energies. Like with the solution of the HF equation, you can apply unitary transformations to get a localized description (ideally only one internal coordinate varying in a mode) that is not diagonal anymore. The off-diagonal matrix elements describe how the localized modes influence each other. More to your problem: I think you can use frequencies that have NO movement of frozen atoms in the corresponding vibrations. You may try to block-diagonalize the matrix with a unitary transformation that constricts vibrations with motion of frozen atoms in one block. Since the coupling elements with the other block would be zero, the frequencies >from that block would be valid. One caveat: When you truncate a molecule, you omit interactions between the cut off parts as well as interactions between the cut off and the retained part of the molecule. If you have to freeze clusters of atoms to force the truncated molecule into the original configuration, some of the omitted interactions were important. I would like to make sure that none of the cut off parts had a significant interaction with the interesting part of the molecule. I hope it answers some of your questions, Stefan ______________________________________________________________________ Dr. Stefan Fau | fau@qtp.ufl.edu Quantum Theory Project | (352) 392-6714 University of Florida Gainesville, FL 32611-8435