From trunec@elanor.sci.muni.cz Wed Jan 17 06:19:33 1996 Received: from aragorn.ics.muni.cz for trunec@elanor.sci.muni.cz by www.ccl.net (8.6.10/950822.1) id GAA01058; Wed, 17 Jan 1996 06:17:34 -0500 Received: from elanor.sci.muni.cz by aragorn.ics.muni.cz with SMTP id AA10360 (5.67a8/IDA-1.4.4 for ); Wed, 17 Jan 1996 12:17:08 +0100 Received: from [147.251.76.16] (casiopeia.physics.muni.cz [147.251.76.16]) by elanor.sci.muni.cz (8.7.1/8.7.1) with SMTP id MAA02822 for ; Wed, 17 Jan 1996 12:17:23 +0100 (MET) From: "David Trunec" Date: Wed, 17 Jan 96 12:23:36 CST Message-Id: <290.trunec@elanor.sci.muni.cz_POPMail/PC_3.2.2> X-Popmail-Charset: English To: chemistry@www.ccl.net Subject: frequency calculation with G94 Dear netters, I have obtained the following error at the calculation of frequencies with G94: ------------------------------------- JobTyp=0 Pass 1: I= 1 to 35. Spin components of T(2) and E(2): alpha-alpha T2 = 0.4198965008D-01 E2= -0.1736529860D+00 alpha-beta T2 = 0.2085778943D+00 E2= -0.8670563493D+00 beta-beta T2 = 0.4198965008D-01 E2= -0.1736529860D+00 ANorm= 0.1136906854D+01 E2= -0.1214362321D+01 EUMP2= -0.99519596711999D+03 PickT4: no shell combinations can fit! NKLS2p= 63 NKLS2= 63 MaxCom= 26 Error termination via Lnk1e in /usr/packages/g94/l1111.exe. Job cpu time: 0 days 0 hours 13 minutes 44.2 seconds. File lengths (MBytes): RWF= 842 Int= 0 D2E= 0 Chk= 3 Scr= 1 ----------------------------------- The job was ----------------------------------- %Chk=SF6 # MP2/6-31G* SCF=DIRECT FOPT TEST SF6 calculation 0 1 S1 F1 S1 sf F2 S1 sf F1 90.0 F3 S1 sf F1 90.0 F2 90.0 F4 S1 sf F1 90.0 F2 180.0 F5 S1 sf F1 90.0 F2 -90.0 F6 S1 sf F2 90.0 F1 180.0 sf=1.58 --Link1-- %Chk=SF6 # MP2/6-31G* SCF=DIRECT FREQ geom=checkpoint guess=check SF6 frequency 0 1 ---------------------------------------- Could anybody tell me, what is wrong ? Thanks, David Trunec From xiaopeng@astro.ocis.temple.edu Wed Jan 17 07:49:34 1996 Received: from astro.ocis.temple.edu for xiaopeng@astro.ocis.temple.edu by www.ccl.net (8.6.10/950822.1) id HAA01496; Wed, 17 Jan 1996 07:41:45 -0500 Received: (from xiaopeng@localhost) by astro.ocis.temple.edu (8.7.1/8.7.1) id HAA21090; Wed, 17 Jan 1996 07:41:44 -0500 (EST) Date: Wed, 17 Jan 1996 07:41:43 -0500 (EST) From: Peking To: ccl Subject: Gaussian 94 IRC calculation bugs? Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Dear netters, I post a question yesterday, from the respond, I found that I may not express myself clearly. Here I add some info about what I did. Sorry for this Anybody can kindly point out if it is a bug or I did something wrong? I used LST and got a saddle point at rhf/6-31g* level, When I finish the transition state optimization and frequency calculation so that I am sure the stationary point I got is a saddle point. then I try to use IRC calculation to confirm that the saddle point connects my starting materia and the product. however, when I submitted that job, the computer told me that "operation on file out of range" "ER terminatnion in NtrErr". "NtrErr called from FileIO" But I did similar job before using Gaussian92. The command file is : --link1-- %chk=ts #T rhf/6-31g(d) IRC=(rcfc, stepsize=5, maxpoint=20) guess=read geom=checkpoint test run IRC to see if the LST generated and TS_optimized transition state really connects the starting materia and the product. 0 1 So is this a bug of G94 but not of G92? Thanks in advance!!! xiaopeng ======================================================= Chemistry Graduate student 215-2047149(office) 215-5357188(home) xiaopeng@astro.ocis.temple.edu ====================================================== From Jeffrey.Gosper@brunel.ac.uk Wed Jan 17 08:34:34 1996 Received: from monge.brunel.ac.uk for Jeffrey.Gosper@brunel.ac.uk by www.ccl.net (8.6.10/950822.1) id IAA01757; Wed, 17 Jan 1996 08:24:11 -0500 Received: from chem-pc-18.brunel.ac.uk by monge.brunel.ac.uk with SMTP (PP) id <16577-0@monge.brunel.ac.uk>; Wed, 17 Jan 1996 13:12:40 +0000 Date: Wed, 17 Jan 1996 13:12:34 GMT From: Jeffrey J Gosper Reply-To: Jeffrey.Gosper@brunel.ac.uk Subject: Error in MOPAC7 To: chemistry@www.ccl.net Message-ID: Priority: Normal MIME-Version: 1.0 Content-Type: TEXT/PLAIN; CHARSET=US-ASCII Dear colleagues, we have been setting up MOPAC7 on our Sun UNIX system and discovered a bug in the code which performs DRC/IRC calculations. The offending line is in the force.f subroutine. The line originally was IF(INDEX(KEYWRD,'IRC')+INDEX(KEYWRD,'DRC').eq.677)THEN which prevents activation of the drc subroutine. This should be replaced by IF(INDEX(KEYWRD,'IRC')+INDEX(KEYWRD,'DRC').ne.0)THEN as it appears in MOPAC6. Does the same bug appear in MOPAC93? /\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\ Dr. Jeff Gosper Dept. of Chemistry BRUNEL University Uxbridge Middx UB8 3PH, UK voice: 01895 274000 x2187 facsim: 01895 256844 internet/email/work: Jeffrey.Gosper@brunel.ac.uk internet/WWW: http://http2.brunel.ac.uk:8080/~castjjg \/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/\/ From owner-chemistry@ccl.net Wed Jan 17 11:04:37 1996 Received: from bedrock.ccl.net for owner-chemistry@ccl.net by www.ccl.net (8.6.10/950822.1) id KAA04148; Wed, 17 Jan 1996 10:59:13 -0500 Received: from relay5.UU.NET for frisch@lorentzian.com by bedrock.ccl.net (8.7.1/950822.1) id KAA10135; Wed, 17 Jan 1996 10:59:11 -0500 (EST) Received: from uucp6.UU.NET by relay5.UU.NET with SMTP id QQzyxr28676; Wed, 17 Jan 1996 10:59:08 -0500 (EST) Received: from m10.UUCP by uucp6.UU.NET with UUCP/RMAIL ; Wed, 17 Jan 1996 10:59:10 -0500 Received: from mjf by m10.lorentzian.com; Wed, 17 Jan 1996 09:31:08 -0500 Received: by mjf (AIX 4.1/UCB 5.64/4.03) id AA09650; Wed, 17 Jan 1996 09:31:08 -0500 From: frisch@lorentzian.com (Mike Frisch) Message-Id: <9601171431.AA09650@mjf> Subject: Re: CCL:G:Gaussian 94 IRC calculation bugs? To: chemistry@ccl.net Date: Wed, 17 Jan 1996 09:31:07 -0500 (EST) In-Reply-To: from "" at Jan 16, 96 06:18:00 pm X-Mailer: ELM [version 2.4 PL24] Content-Type: text Peking writes: > > Dear netters, > Anybody can kindly point out if it is a bug or I did something wrong? > > I used LST and got a saddle point at rhf/6-31g* level, then I try > to use IRC calculation to confirm that the saddle point connects my > starting materia and the product. however, when I submitted that job, the > computer told me that "operation on file out of range" "ER terminatnion > in NtrErr". "NtrErr called from FileIO" But I did similar job before using > Gaussian92. > > So is this a bug of G94 but not of G92? > > Thanks in advance!!! > > xiaopeng > This topic is a common source of confusion, which I'd like to clarify. The IRC path leads down from a transition state to the minima it connects. It starts at a true transition structure, where the forces are zero and the Hessian has one negative eigenvalue. The algorithm in Gaussian also expects to start with the Hessian at the TS. The LST procedure does NOT locate a transition structure -- it locates an initial guess for a transition state optimization. In particular, it finds a point at which the Hessian has at least one negative eigenvalue (but perhaps more) and at which the forces are far from zero. The results of an LST calculation are not chemically meaningful, but are only useful to start an optimization to locate the actual transition structure. In Gaussian 94, LST is really obsolete, since the Opt=QST2 algorithm both uses synchronous transit methods to locate a guess for the transition structure and then goes on to complete the optimization. This replaces the combination of LST and Opt=TS in G92 with a single (and much more reliable) algorithm. The combination of LST and IRC without an intervening Opt=TS in G92 would give a path, but not one satisfying the definition of the IRC and not one containing the transition structure, so it's probably better than G94 won't let you do it. The recommended procedure is Opt=QST2, Freq, IRC=RCFC. The Freq step both confirms that the structure found by Opt=QST2 is a proper first order saddle point and provides the Hessian read by the RCFC option to the IRC. Mike Frisch frisch@lorentzian.com From jtgolab@amoco.com Wed Jan 17 11:12:36 1996 Received: from interlock.amoco.com for jtgolab@amoco.com by www.ccl.net (8.6.10/950822.1) id KAA04030; Wed, 17 Jan 1996 10:53:10 -0500 Received: by interlock.amoco.com id AA03027 (InterLock SMTP Gateway 3.0 for CHEMISTRY@www.ccl.net); Wed, 17 Jan 1996 09:52:38 -0600 Message-Id: <199601171552.AA03027@interlock.amoco.com> Received: by interlock.amoco.com (Protected-side Proxy Mail Agent-3); Wed, 17 Jan 1996 09:52:38 -0600 Received: by interlock.amoco.com (Protected-side Proxy Mail Agent-2); Wed, 17 Jan 1996 09:52:38 -0600 Received: by interlock.amoco.com (Protected-side Proxy Mail Agent-1); Wed, 17 Jan 1996 09:52:38 -0600 From: jtgolab@amoco.com (Joe Golab) Date: Wed, 17 Jan 1996 09:53:34 -0600 X-Mailer: Z-Mail (3.2.2 10apr95 MediaMail) To: CHEMISTRY@www.ccl.net Subject: SUMMARY - Flouresence Spectra Question Cc: jtgolab@amoco.com Mime-Version: 1.0 Content-Type: text/plain; charset=us-ascii Hi CCLers: Here are the responses I received concerning computational methods for determining flouresence spectra. I did not receive many responses probably because this is not a straightforward spectra to obtain (like an IR for example). It seems as though a semiempirical mixture method using MOPAC and ZINDO is the "bleeding-edge" solution for now (although I expected at least one Multi-configurational guru to have something - no matter how crude - in development). If I receive other responses after this summary is posted, I will send a second summary to the list. Thanks for the leads. >-- Attachment: Text File >! Date: Fri, 12 Jan 1996 11:01:18 -0600 >! From: jtgolab@nap.amoco.com (Joe Golab) >! Subject: CCL:Computational Flouresence Spectra? >! >! Hi CCLers: >! >! Is anyone out here aware of a method to calculate flouresence spectra? >! (already programmed)? >! >! Is there a semiempirical way to do it? >! >! Please reply to me and I will summarize for the list. >! >! Thanks. >! >! :Joe >! jtgolab@amoco.com >! (708) 961-7878 >! >! ----- Reply Separator >! >! Date: Fri, 12 Jan 1996 13:27:15 -0600 (CST) >! From: "Jeffry D. Madura" >! Subject: Re: CCL:Computational Flouresence Spectra? >! >! Dear Joe, >! >! Greetings from sunny Mobile. I hope that this note finds you doing >! well. As for your question the answer is yes and no. Fluoresence is a >! decay from an excited state to the ground state and is not readily >! calculated using semiempirical methods or even ab initio methods. Right >! now the only way I know how (which is limited to my feable mind) to do >! this is through molecular dynamics and time correlation functions. Let >! me know if you find an ab initio or semiempirical answer. This would be >! of great interest to biophysicists... >! >! I hope that your New Year is off and going... >! >! Best Regards, >! >! >! Jeffry D. Madura >! >! >! Address: Department of Chemistry >! University of South Alabama >! Mobile, AL 36688 >! >! Phone: (334) 460-7430 >! FAX: (334) 460-7359 >! >! e-mail: jmadura@jaguar1.usouthal.edu >! >! ----- Reply Separator >! >! Date: Sun, 14 Jan 96 17:04:23 CST >! From: "Ernest Chamot" >! Subject: Re: CCL:Computational Flouresence Spectra? >! >! Hi Joe, >! >! Funny you should ask about flouresence spectra. >! >! I was just looking into that for someone last month. Basically, the answer >! is YES! Fluorescence excitation and emission wavelengths can be calculated >! on reasonable sized molecules with the two semiempirical programs: MOPAC and >! ZINDO. >! >! Flourescence is the emission of light from the excited singlet state, after >! decay to the lowest excited singlet level, but before intersystem crossing >! to a triplet. One assumes the electronic reorganization (orbitals and >! orbital occupancy) is much faster than any geometric reorganization, so what >! you need is the energy difference between the lowest excited singlet state >! and the ground state singlet, both at the excited singlet geometry. >! >! MOPAC gives better geometries than ZINDO and has a CI option, so use that to >! optimize the geometry of the first excited singlet. MOPAC doesn't calculate >! transition moments, however, but ZINDO does and is specifically >! parameterized for UV spectral simulation: the excitation spectrum (UV) is >! readily calculated via a CI calculation on the ground state geometry (and >! displayed graphically in the CAChe system, including intensities based on >! calculated transition moments). The same CI calculation on an excited state >! singlet geometry gives the flourescence emission wavelength instead of the >! highest UV absorbance. >! >! There are actually two options for calculating the fluorescence wavelength: >! instead of the ZINDO CI calculation of the transition between the lowest >! singlet and the first excited singlet at the excited singlet geometry, you >! could do a pair of MOPAC calculations, and look at the difference in energy >! between the excited singlet geometry modeled as the lowest singlet and as >! the first excited singlet electronic configuration. Using ZINDO with the >! MOPAC geometries worked better for me, though. >! >! I took a quick look at a dozen compounds that I could find fluorescence data >! on readily (ranging from 250-500 nm, non-polar solvent). Two were off >! substantially: benzoin borate and tetracyanohydroquinone. ZINDO is reported >! to be less accurate for modeling electronic transitions of compounds with >! unshared electron pairs, but these two problematic data points may also >! reflect complexation, hydration, and/or solvation. Otherwise the Zindo CI >! calculations in my quick test were closer than the MOPAC energy differences: >! consistently about 20 nm or so low. >! >! Both MOPAC and ZINDO have features for including solvation (the COSMO >! solvation model in MOPAC and an SCRF model in ZINDO), so you can also >! correct for solvatochromic shift. In fact, you can even model ionic systems >! (although there will be big differences if you don't find the right >! geometry!) Fluorescein is a flourescent dye in water as the disodium salt. >! Using MOPAC and ZINDO with aqueous solvation, you can predict an emission >! maximum of 541 nm. This compares favorably with the reported flourescence >! at 520 nm. For reference, I've been told that "at the DYECon95 meeting, >! many of the predictions on dye absorbance maxima were off by 100 nm in this >! wavelength range." >! >! If you are interested in predicting flourescence on compounds with nonbonded >! electron pairs, I would certainly recommend running some similar, known >! examples and using that data to calibrate your calculations. Otherwise, you >! should be able to do very well calculating flourescence with MOPAC and >! ZINDO. If you would like, I would be happy to run a sample compound for you. >! >! EC >! --- >! Ernest Chamot >! Consultant in Computational Chemistry Applications >! Chamot Laboratories, Inc. >! 530 E. Hillside Rd. >! Naperville, Illinois 60540 >! echamot@xnet.com >! >! ----- Reply Separator >! >! Date: 14 Jan 1996 13:07:41 -0700 >! From: "Paul A Cahill" >! Subject: Re: CCL:Computational Flouresence Spectra? >! >! Joe, >! I've had mixed success by using MOPAC geometries for the first excited >! singlet state (slow process) followed by a ZINDO calculation of the molecule >! in that minimized geometry. One can argue that the lowest electronic >! transition of the resulting ZINDO calculation will be similar to the >! fluorescence spectrum of the system. >! >! Paul Cahill >! Sandia National Laboratories >! pacahil@sandia.gov >! 505-844-5754 >! >! ----- Reply Separator >! >! Date: Tue, 16 Jan 96 11:27:04 -0500 >! From: mitchell@bdrc.bd.com (Mike Mitchell) >! Subject: Re: Fluorescence Spectra >! >! Joe- >! >! I'm currently looking for a way to calculate absorption and emission >! transition dipole moments for organic dyes. I've also had some modeling >! requests for calculated emission (fluorescence) specra. Unfortunately, my >! expertise doesn't run very far in that direction. I've discovered >! that MOPAC 6 won't do it, but that MOPAC 93 might be able to do so (to >! calculate transition dipoles) I don't have MOPAC 93, though. I'm going to >! try to contact Zerner to see if Zindo will help out any. You might call >! Zerner as well to see if his program will calculate entire spectra. There >! is a version of Zindo in Hyperchem that will calculate absorption maxima >! fairly well, but that doesn't appear to include transition dipole moments >! (at least from looking at the Hyperchem manual). >! >! About 10 years ago, I wrote a program that used experimental and tabulated >! values for Einstein coefficients, etc. and a calculated slit width to create >! absorption and fluorescence spectra. As I recall, this program worked pretty >! well. I would have no idea where this code is located nowdays, though. >! Sorry. >! >! Please let me know if you find out anything useful. I'll e-mail you >! if I learn anything. >! >! Mike Mitchell >! Becton Dickinson Research Center >! mitchell@bdrc.bd.com >-- End of Attachment -- :Joe jtgolab@amoco.com (708) 961-7878 +--------------------------------------------------------+ | Many of life's failures are people who did not realize | | how close they were to success when they gave up. | | - T. Edison | +--------------------------------------------------------+ From qtmtulio@cc.fc.ul.pt Wed Jan 17 13:34:38 1996 Received: from skull.cc.fc.ul.pt for qtmtulio@cc.fc.ul.pt by www.ccl.net (8.6.10/950822.1) id NAA07003; Wed, 17 Jan 1996 13:31:32 -0500 Received: by skull.cc.fc.ul.pt id aa26558; 17 Jan 96 17:01 LISBOA From: Mario Tulio Rosado To: chemistry@www.ccl.net Subject: Different mopac and gaussian AM1 results Cc: rfausto@gemini.ci.uc.pt X-Mailer: ScoMail 1.0 Date: Wed, 17 Jan 1996 16:53:33 +0100 (LISBOA) Message-ID: <9601171653.aa26205@skull.cc.fc.ul.pt> Dear CCL, I started to make some semiempirical calculations (AM1 and PM3) under MOPAC 6.0 (DOS version) some time ago. Right now i had to repeat some of them and i did it under gaussian. This is what happened: The ENERGY VALUES after optimization that led exactly to the same structure (Glycine SCF-RHF/AM1) didn't match the earlier MOPAC results. I used the "HF" value given by gaussian in hartrees and compared to the "Total Energy" given by MOPAC in eV. The "entalpy of formation" value didn't match also (this one as expected). The CHARGES of the atoms of the same structure calculated with the same method (AM1) in both programs were also significantly different (for example: the charge in cabonyl O is ca 0.36 with one program and ca.0.40 with the other). If requested i can post the actual outputs. I checked the AM1 parameterization references for every atom in both programs and it is the same. Can anybody elucidate this "strange" happenings??? Thanks in advance. Mario Tulio Rosado Dept of Chemistry Faculty of Science University of Lisbon PORTUGAL From LLOPES@iqsc.sc.usp.br Wed Jan 17 14:19:39 1996 Received: from uspqsc.iqsc.sc.usp.br for LLOPES@iqsc.sc.usp.br by www.ccl.net (8.6.10/950822.1) id OAA07755; Wed, 17 Jan 1996 14:17:47 -0500 Date: Wed, 17 Jan 1996 17:18:14 -0500 (EST) From: "Luiz Gonzaga de F. Lopes" To: chemistry@www.ccl.net Message-Id: <960117171814.4d82@iqsc.sc.usp.br> Subject: NO solution stability Dear netters, I would like to know about the NO solution stability. If somenoe can help me with any information or reference. Thanks Luiz Lopes ph.D chemistry student E-mail "LLOPES@IQSC.SC.USP.BR" PS - I would like to know about how many hours we can keep NO solution (free of oxygen). From boyd@chem.iupui.edu Wed Jan 17 10:28:05 1996 Received: from indyvax.iupui.edu for boyd@chem.iupui.edu by www.ccl.net (8.6.10/950822.1) id KAA03212; Wed, 17 Jan 1996 10:28:03 -0500 Return-receipt-to: Boyd Received: from macgw.chem.iupui.edu by INDYVAX.IUPUI.EDU (PMDF V5.0-5 #5862) id <01I043L6Y7SG0079ZU@INDYVAX.IUPUI.EDU> for chemistry@www.ccl.net; Wed, 17 Jan 1996 10:24:26 -0500 Date: Wed, 17 Jan 1996 10:21:58 -0500 From: Boyd Subject: FREE BOOKS To: OSC CCL Message-id: <01I043L6ZADE0079ZU@INDYVAX.IUPUI.EDU> Content-transfer-encoding: 7BIT Would you like your library to receive a set of "Reviews in Computational Chemistry"? With tight library budgets these days, many libraries have an urgent need for current books. To show our appreciation to our readers, we will donate a set (Volumes 1-7, 1990-1995) with a retail value of $724 to a departmental, college, or university library. The selection will be at random based on your letting us know the need. To enter on behalf of your library, just send the name of your librarian and their postal address to: boyd@chem.iupui.edu. In the Subject line put the words FREE BOOKS. In the body of the e-mail include one sentence about why your library could use these books, and include your own email address and affiliation. The winner will be picked by random and announced over the OSC CCL. The deadline for receipt of entries is February 1, 1996. Other information about "Reviews in Computational Chemistry" is in the Home Page at http://chem.iupui.edu/~boyd/rcc.html Best wishes for 1996, Dr. Kenny B. Lipkowitz Dr. Donald B. Boyd Indiana University-Purdue University at Indianapolis Co-Editors, "Reviews in Computational Chemistry" From zuilhof@chem.chem.rochester.edu Wed Jan 17 15:28:25 1996 Received: from chem.chem.rochester.edu for zuilhof@chem.chem.rochester.edu by www.ccl.net (8.6.10/950822.1) id PAA08948; Wed, 17 Jan 1996 15:28:19 -0500 Received: (from zuilhof@localhost) by chem.chem.rochester.edu (8.7.1/8.7.1) id PAA01578; Wed, 17 Jan 1996 15:23:17 -0500 (EST) Date: Wed, 17 Jan 1996 15:23:16 -0500 (EST) From: Han Zuilhof To: chemistry@www.ccl.net cc: frisch@lorentzian.com, Han Zuilhof Subject: IRC problems in Gaussian 94 occur also with OPT=QST2 option Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Status: RO Dear CLL netters, Recently a problem regarding IRC calculations in Gaussian 94 was brought up. Mike Frisch clarified the discussion by stating that LST does in itself not yield a real transition state, and that the new OPT=QST2 routine should be used. However, this does not solve the 'out of range' error! I've bumped into it several times myself, using the OPT=QST2 method (and confirming frequencies) in G94 (versions B3 and C3). A condensed file example of input and output is given at the end of this letter. I hope that further discussion of this problem on the CCL net will clarify this situation. IRC is conceptually a very nice tool, but I've found it in several cases to be troublesome or even impossible to get it to work within G94. The 'out of range' problem is just one of them. It would be helpful to find out precisely a) what this error means, and -more importantly- b) how one can get around it. Another problem I bumped into is that the IRC on a TS (as optimized with OPT=QST2 and checked with FREQ) simply doesn't "leave the area" on the potential energy surface around the TS, and yields after 2-5 steps a geometry nearly identical to the TS as one of its minima. Here a 'way around' is sometimes to start the IRC from a geometry close to the TS along a 'self-chosen' reaction coordinate (choice base on e.g the character of the imaginary frequency), but this trick is both somewhat arbitrary and also not-infallable. If someone knows how to deal with this problem, I would be very glad! Thanks in advance to anyone who can help, Han Zuilhof -------------------------------------------------------------------------- Gaussian 94 .com input file and excerpt of .log output file: %chk=c110507for #p irc=(maxpoints=100,rcfc,forward) geom=check guess=read uhf/6-31G(d) c110507for;forward IRC TS H2O + 2-Me-butane catrad at 2-pos; 1,2 ----------------------------------------------------------------------- *************************************** Gaussian 94: SGI-G94RevC.3 26-Sep-1995 13-Jan-1996 *************************************** %chk=c110507for ---------------------------------------------------------------------- #p irc=(maxpoints=100,rcfc,forward) geom=check guess=read uhf/6-31G(d) ---------------------------------------------------------------------- c110507for;forward IRC TS H2O + 2-Me-butane catrad at 2-pos; ------------------------------------------------------------ Redundant internal coordinates taken from checkpointfile: c110507for.chk (stuff deleted) Route: #P FREQ=NORAMAN GUESS=READ GEOM=CHECK UHF/6-31G* Cartesian force constants read from checkpoint file: (stuff deleted) (Enter /usr/local/g94/l716.exe) OPERATION ON FILE OUT OF RANGE. FILEIO: IOPER= 2 IFILNO(1)=-10511 LEN= 4559 IPOS= 2 Q= 27356885 6 dumping /fiocom/, unit = 1 nfiles = 69 sizext = 524288 winblk = 64 defal = T lstwrd = 4099136 number 0 0 0 0 0 0 0 base 151424 116480 46656 1128448 1187456 1093376 122880 end 152704 118720 46720 1138304 3574784 1107328 122944 end1 152704 118720 46720 1138304 3574784 1107328 122944 wr pntr 151424 116480 46656 1128448 1187456 1093376 122880 rd pntr 151424 116480 46656 1128448 1187456 1093376 122880 (stuff deleted) Error termination in NtrErr: NtrErr called from FileIO. ****************************************************************************** ** Dr. Han Zuilhof ** e-mail: ZUILHOF@chem.chem.rochester.edu ** ** Department of Chemistry ** ** ** University of Rochester ** ** ** Rochester, NY, 14627 ** Fax: (716) 473-6889 ** ** USA ** Voice: (716) 275-2219 ** ****************************************************************************** ** ** ** "Excite a photochemist!" ** ** ** ****************************************************************************** From qtmtulio@cc.fc.ul.pt Wed Jan 17 15:49:40 1996 Received: from skull.cc.fc.ul.pt for qtmtulio@cc.fc.ul.pt by www.ccl.net (8.6.10/950822.1) id PAA09234; Wed, 17 Jan 1996 15:38:49 -0500 Received: by skull.cc.fc.ul.pt id aa26558; 17 Jan 96 17:01 LISBOA From: Mario Tulio Rosado To: chemistry@www.ccl.net Subject: Different mopac and gaussian AM1 results Cc: rfausto@gemini.ci.uc.pt X-Mailer: ScoMail 1.0 Date: Wed, 17 Jan 1996 16:53:33 +0100 (LISBOA) Message-ID: <9601171653.aa26205@skull.cc.fc.ul.pt> Dear CCL, I started to make some semiempirical calculations (AM1 and PM3) under MOPAC 6.0 (DOS version) some time ago. Right now i had to repeat some of them and i did it under gaussian. This is what happened: The ENERGY VALUES after optimization that led exactly to the same structure (Glycine SCF-RHF/AM1) didn't match the earlier MOPAC results. I used the "HF" value given by gaussian in hartrees and compared to the "Total Energy" given by MOPAC in eV. The "entalpy of formation" value didn't match also (this one as expected). The CHARGES of the atoms of the same structure calculated with the same method (AM1) in both programs were also significantly different (for example: the charge in cabonyl O is ca 0.36 with one program and ca.0.40 with the other). If requested i can post the actual outputs. I checked the AM1 parameterization references for every atom in both programs and it is the same. Can anybody elucidate this "strange" happenings??? Thanks in advance. Mario Tulio Rosado Dept of Chemistry Faculty of Science University of Lisbon PORTUGAL From owner-chemistry@ccl.net Wed Jan 17 17:49:43 1996 Received: from bedrock.ccl.net for owner-chemistry@ccl.net by www.ccl.net (8.6.10/950822.1) id RAA11466; Wed, 17 Jan 1996 17:44:51 -0500 Received: from osuunx.ucc.okstate.edu for pran@osuunx.ucc.okstate.edu by bedrock.ccl.net (8.7.1/950822.1) id RAA16532; Wed, 17 Jan 1996 17:44:49 -0500 (EST) Date: Wed, 17 Jan 1996 11:09:57 -0600 (CST) From: Ran Pan To: chemistry@ccl.net Subject: help with mopac Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: pran@osuunx.ucc.okstate.edu Dear CCL-users, Recently, I moved the MOPAC from RS/6000 to DEC/3000 workstation, the mopac pacage worked well in RS/6000, however, I cannot compile the package in the DEC/6000 workstation. I use the "cmpall" file which was ever used in RS/6000 and replaced all "forg" with "f77 -c" in DEC workstation, it didn't work. Would some experts in the net like to help us to solve this problem? I appreciate it very much. Your sincerely, Ran - * - * - * - * - * - * - * - * - * - * - * - * - * - * - * - * - * - * - - Ran Pan pran@osuunx.ucc.okstate.edu - - http://www.okstate.edu/welcome.html - - Oklahoma State University - - * - * - * - * - * - * - * - * - * - * - * - * - * - * - * - * - * - * - From pran@osuunx.ucc.okstate.edu Wed Jan 17 19:21:23 1996 Received: from osuunx.ucc.okstate.edu for pran@osuunx.ucc.okstate.edu by www.ccl.net (8.6.10/950822.1) id TAA14151; Wed, 17 Jan 1996 19:18:10 -0500 Date: Wed, 17 Jan 1996 15:41:48 -0600 (CST) From: Ran Pan To: chemistry Subject: mopac6 Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Sender: pran@osuunx.ucc.okstate.edu Dear netter, Recently, I try to use mopac to excute some calculations, the mopac package was ever used in RS/6000 very well. I moved the souce code to DEC/6000 workstation, however, I cannot compile it. There is a compiler file named "cmpall" in which the fortran compile command is "forg" for RS/6000. For the fortran compiler in DEC/6000, the compile command is "f77 -c". I change the "cmpall" file by replacing all "forg" with "f77 -c" and then excute "cmpall", it give the error message. I don't know why? whould some experts like to give me some suggestions for the use of mopac6? Thank you very much. Ran Pan - * - * - * - * - * - * - * - * - * - * - * - * - * - * - * - * - * - * - - Ran Pan pran@osuunx.ucc.okstate.edu - - - - Oklahoma State University - - * - * - * - * - * - * - * - * - * - * - * - * - * - * - * - * - * - * -