From amasunov@shiva.Hunter.CUNY.EDU Tue Mar 5 15:36:06 1996 Received: from hcrelay.hunter.cuny.edu for amasunov@shiva.Hunter.CUNY.EDU by www.ccl.net (8.7.1/950822.1) id PAA07013; Tue, 5 Mar 1996 15:09:27 -0500 (EST) Received: from shiva.hunter.cuny.edu (amasunov@shiva.hunter.cuny.edu [146.95.128.96]) by hcrelay.hunter.cuny.edu (8.6.12/george0995) with SMTP id PAA01490; Tue, 5 Mar 1996 15:08:49 -0500 Date: Tue, 5 Mar 1996 15:10:29 -0500 (EST) From: Artem Masunov To: Computational Chemistry List cc: Magda Wajrak , Artem Masunov Subject: Summary: Anharmonic Vibrations. Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII Content-Transfer-Encoding: QUOTED-PRINTABLE Dear Netters,=20 Thank you all who replied.=20 I was pleasently surprized by many responces to my question. Here they are. I edited them a bit to make the text more readable and added some relevant information.=20 ____________________Original Question________________________ I am looking for software to solve (multidimensional) vibrational Schroedinger equation. Gaussian94 apply harmonic model and just warns you it is incorrect for soft modes, but what is correct? I can calculate accurate PES, but how to get frequencies and zero-point vibrational levels for these selected vibrations (in my case they have double-well shape)? Another problem - how to get vibrational modes expressed in internal (not cartesian) coordinates? ______________________Abstract____________________________ Besides direct solution of the Schroedinger equation, it is possible to= =20 "correct" experimental frequencies for anharmonicity or use isotopic=20 substitution. The molecular vibrations software could be devided into several categories= : 1. Normal coordinate analizes of experimental spectra:=20 SprioVib, Prometheus, BGF/NCA, ASYM20, Fitvib, MolVib, Gamma,=20 Macrosearch, Specedit, SPSIM, Morass;=20 2. Normal modes (incl. normal modes in internal coordinates) and spectra= =20 from theoretical calculations (using Hessian from DGauss, Gaussian, etc.): Vibra, Noko, Cospeco, Gar2ped; 3. Visualization of normal modes from theoretical calculations: Molden, Moviemol, XMol, KGNGraph(MOTECC91)=20 4. Anharmonic correction factors from ab initio PPE: Spectro, SurVibTM; 5. Solution of the vibrational Schroedinger equation (one or=20 mutidimensional) Level(=3DDVibro), QMC, Nivelon. _________________________ URL Addresses ____________________________ SPECTRO http://gmos.ch.cam.ac.uk/spectro.html PROMETHEUS http://www.ruhr-uni-bochum.de/www-public/goetzhbv/prom.html SPIROVIB http://spiro.princeton.edu/spirolabs/svib.html SURVIBTM http://www.cs.qub.ac.uk/cpc/summaries/ABDZ and other vibrational programs from CPC (mostly, for triatomics) http://www.cs.qub.ac.uk/cpc/mol.html#molvib From: Tuomas J Lukka =20 I have done some research on rovibrational properties of small (XY_2 and XY_3) molecules and have some software.=20 16 degrees of freedom is a LOT when talking about anharmonic molecular vibrations. I don't think that there's any software available - most work has so far been done on a per-molecule basis, because simplifications and models have to be generated.=20 From: Martin Jursch The force constances of an ab initio calculation are related to carthesian coordinates and not to internal coordinates as in Wilsons theory. Beside that IR-calculation in molecular modeling are a sideeffect of geometrie optimization and therefore not made to calculate good wavenumbers and meaningful force constants.=20 Normal modes are given in displacements of carthesian coordinates. I have never seen normal modes given in internal coordinates. All models for NCA using classic force fields are harmonic but the experimental wave numbers can be "corrected" to harmonic conditions. PROMETHEUS offers the posibility to use "correction factors" for wave numbers beeing strongly anharmonic. For installation download either prom.zip or disc1.zip / disc2.zip from our WWW-page and unpack them in a temporary directory. Then run nkasetup.exe for installation. From: Dr. Philippe Youkharibache If you already have code that calculates normal modes in cartesian X =3D Lx Q and to determine the "B matrix" R =3D B X ... (a program that does the former does the latter if the force constants=20 are expresse d in internal space) then the Lr matrix you are looking for is Lr =3D B * Lx where R =3D int coords X cartesian coords Q normal coords this is much simpler. programs at QCPE exist for these things. see the book of Wilson on that topic From: MR M Govender There is a solution to your problem, i.e you have to use isotopic substitution, in order to predict the anharmonic vibrations. This is however a bit difficult if you are going to go via intensities. Vibra can handle such calcs:)) The program Vibra (by Dr Steele, for a full normal mode treatment and ped, and dipole moment derivat. to predict the IR intensities after correcting for rotation etc..), uses the cartesian force field from gaussian...I have never tried using the program, with input containing imaginary freq... But however the soft modes are well predicted by the program:)... The address of Dr STeele is aslc801@vmsfe.ulcc.ac.uk. We also have ASYM20, by Ian Mills , obtained from QCPE. I also have SPectro, author Handy, and Willets, and lastly there is a program obtainable via ftp iqm.unicamp.br, in the chem or pub dir, which is called BGF and NCA, similar to ASyM20.=20 From: Derek Steele =09I am not clear as to what it is that you need. My program is very comprehensive for harmonic vibrational problems. Among the many things it will do is to take the cartesian force field (in atomic units) and the atomic coordinates and recompute the vibrational frequencies. However this is done in internal valence coordinates which must be defined (stretches, torsions, angle bends etc.). The resulting frequencies should be exactly equal to the values from Gaussian. The potential energy distribution is given; It is possible to scale different types of internal coordinates separately; Various intensity models for both i.r. and Raman bands can be tested and various parameter modelling, such as valence force fields , can be optimised. I will send by e-mail a document on the programme. The programme is in fortran. If it is of interest I will send the listing, again via e-mail. Should you find that you wish to continue using it then I request a nominal payment of $100 to help maintain the program. =09Reverting to your problem; Clearly it is not possible for any=20 classical program to improve the ab initio, except by scaling. It follows that anything you do has to begin with the ab initio force field. If you get a negative frequency with the Gaussian , then you will with the=20 classical programs. Of course you can derive (from my program for instance) a symmetrised valence force field and modify the quadratic term for the torsion to generate a suitable positive frequency to match that derived by your proposed PES. A comparison of your potential energy barrier with the ZPE is of course trivial, and the low torsional frequencies are not going to contribute significantly. From: Thomas Nowak If you have the B matrix which transform the cartesian force constant matrix to the internal force constant matrix you can build the G matrix by G=3DB M**-1 B#. Where B# is the transpose of the B matrix and M**-1 is the inverse of the M matrix. M is a diagonal matrix of tripels of the mass. The product FG is a nonsymmetric matrix. The simplest way to get the eigenvalue and eigenvectors is to bring the matrix to a hessenberg matrix then you can get the eigenvalue and eigenvectors by the qr methode.=20 I wrote a FORTRAN program which can build the internal force constants matrix from the cartesian force constant matrix. Futhermore it will calculate the G matrix and frequences. There is also a procedure which bild the B matrix. My program can handle your problem. It was designed to work with DGAUSS, therefore it needs the cartesian Hessian in a sepearate file and in an other file the definition of your internal coordinates. The programm used the GF-methode from Wilson and is an FORTRAN written program. It is tesed on ibm and SGI workstation but should be run also on other platforms. Most parts of the program are written in German. So if you have problems with the data, let me know and I will make an english version. If you are loocking for some literature for this topic:=20 Califano Vibrational States, JOHN WILLEY & SONS 1976 T.Miyazawa J. Chem. Phys. 1958 29 P.246 R.J.Malriot J. Chem. Phys. 1955 23 P.30 Prof. Diem at CUNY, Hunter College has a program. The reference is J. Chem. Educ. 1991, vol. 68, p.35-39.=20 Hedberg and Mills have a program called ASYM20, which is available from QCPE. The most recent version is ASYM40. The reference is J. Mol.=20 Spectroscopy, 1993, vol. 160, p.117-142.=20 Prof. Thomas Bally of the University of Fribourg, Switzerland, has a very nice program which is able to extract force constants from ab initio calculations. The program is going to be distributed by Gaussian. I do not know the date. Vijay, A. and Sathyanarayana, D. N. J. Mol. Struct. 1994, 328, 269-276. Durig et al. J. Mol. Struct. 1994, 327, 55-69. From: "Robert J. Le Roy" The program I believe you are referring to is the one I call LEVEL. For any radial or effective one-dimensional potential (including double/multiple minimum cases) it can calculate bound state levels, and if appropriate, automatically add a centrifugal potential to account for the effect of rotation. The potential may be in the form of an array of points, to be interpolated over in a manner specified by the user in the input data file, or one of a range of typical diatomic analytic finctions. However, a user can readily replace the internal potential generating routine with their own function, if so desired. If this code seems useful to you, let me know and I'll e-mail you the source and post you the user's manual.=09 From: Alfonso Nino , The program used for the anharmonic treatment of acetaldehyde is available from QCPE. The program is designed as a general tool for an arbitrary number of vibrations. Thus, I think you can use it for your problem. My only worry is that at present the program uses a Givens-Houselholder routine for diagonalizing the hamiltonian. The routine has shown to work well with matrix sizes of several thousands (what we found with 3 vibrations). In your case, 4 vibrations, the hamiltonian size can be very large and may be the Givens-Houselholder is not the ideal routine.=20 In its present implementation the program can use arbitrary potential functions expressed as Fourier series and Taylor expansions on the vibrational coordinates. Even gaussian perturbations can be introduced in the program. Thus, inversion (double-well potentials) can be introduced as a) Polynomial forms:=20 something like (a*X^4-b*X^2) b) Gaussian perturbed potentials: something like (a*X^2+b*exp[-c*X^2]) We have sent you two references describing the capabilities of the program= :=20 a) C. Munoz-Caro & A. Nino. Computers. Chem, 18(4), 413-417 (1994) b) A. Nino & C. Munoz-Caro. Computers. Chem, 19(4), 371-378 (1995) The first one describes the general features of the first implementation. The second one describes the use of hybrid free rotor+ harmonic oscillator basis functions. In addition, this second paper analyzes the ability of different polynomial and gaussian perturbed functions to describe double-well potentials. In particular, the inversion of ammonia is used as a test case.=20 The program is available as: NIVELON. QCPE PROGRAM #665 (workstation version) NIVELON. QCMP 142 (PC version) From: Brian Hammond Bill Lester's group at Berkeley has a QMC program that will give a number of vibrational states to very high accuracy. The person to contact is Will Brown at UC Berkeley, wbrown@garlic.CChem.Berkeley.EDU From: "Eric R. Bittner" Your problem is a matter of how large of a system are you talking about and how accurate you want the vibrational energy levels. If you want very high accuracy, you're limited to (at most) a few internal degrees of freedom + rotation. John Zhang (at NYU) and John Light (at U. Chicago) are the world experts in this domain. As far as I know (being one of Light's former students), neither group has made their codes avalable to the public, but all the details are in the literature. This also means you're going to have to do a lot of work to really nail down the potential energy surface for a large number of accessible classical configurations and then fit to some functional form.=20 If the dimensionality of the problem is relatively small (2-3 degree of vibrational freedom) then I'd recommend using DVR methods (ala Light). For big molecules, ala proteins,there was a paper in Science a few months ago by Roitberg, Gerber, Elber, and Ratner (Science 268, 1319 (1995)) where they computed the vibrational spectra of BPTI using a SCF treatment of the vibrational modes. You may want to check out their results and methodology.=20 The matrix which diagonalizes the Hessian (2nd deriv. of potential) matrix is the transformation between the cartesian (lab) frame and the internal normal vibrational modes. These are OK for the normal modes, but are not so hot for the really floppy modes.=20 From: William Mccarthy The short answer to the question of "how to get frequencies ... for these selected vibrations" is: it takes a lot of work. Here are the steps: 1) get an analytical form for your PES and for the dependence of mass on your chosen internal coordinates. 2) expand your nuclear wavefunction into a suitable basis. Now your 90% of the way there. There is a freeware program out there called dvibro. I haven't used it though. (I wrote my own, and feel more comfortable with it since it allows me complete control over the subroutines that determine contributions to the Hamiltonian and Overlap matrices.)=20 If you want to get the secular equation in internal coordinates, check out J Chem Phys 103(2)(1995) 656-662.=20 With regard to getting the Hessian in internal coordinates, then getting its eigenvalues and eigenvectors, you could just read the Hessian in internal coords from the g94 output and diagonalize it. Otherwise, you'r stuck with determining the B matrix and the "L" matrix which diagonalizes the Hessian in cartesian coords. Several people have offered their programs over CCL which will probable do what you want in this regard, but it isn't hard to code it yourself (see Califano's "Vibrational States", pages 81-88) From: "A. Willetts" =09In response to your request for some information about the SPECTRO program, I have included below a short description of some of the major features. SPECTRO v3.0 is a modular FORTRAN77 computer program which has been developed at the University of Cambridge. It now consists of approximately 48000 lines of code. As it was originally developed to verify the accuracy of ab initio calculated potentials (and dipole fields), it is typically used with the output of electronic structure theory codes such as CADPAC. However, it is flexible enough to accept, for example, an experimentally=20 derived internal coordinate potential.=20 The program itself uses the formulae derived from perturbation theory= =20 to calculate a wide variety of spectroscopic properties. For example, the following list includes a selection of these properties: Equilibrium and vibrationally excited rotational constants; Quartic and sextic centrifugal distortion constants; Harmonic and fundamental frequencies; Anharmonic constants; Fermi and coriolis resonance; Absolute intensities of fundamentals and first overtones and=20 combination bands. Originally the program was used to examine the detailed effects of anharmonicity on small gas phase molecules. Recently, however, it has been used in the calculation of such properties as reaction rates (using semiclassical transition state theory) and the vibrational correction to=20 (hyper)polarisabilities.=20 Ongoing work with SPECTRO includes the integration of a sophisticated least squares fitting program (developed at the University of Bologna) whic= h allows us to refine an entire internal coordinate quartic potential to all= =20 of the experimental data which SPECTRO is able to calculate.=20 If you would like a copy of the program (along with a sample=20 input/output and documentation in LaTeX) please let me know and I will be happy to send it to you. SPECTRO has been distributed to a number of groups throughout the World, both theoretical and experimental. It is currently running on a number of platforms ranging from an Apple Macintosh (where some code amendment was required) to a Cray YMP. The program is free= . From: Kris Van Alsenoy Jan Martin and myself have develloped the utility gar2ped (a fortran program) which reads a Gaussian94 archive record containing the cartesian second derivatives and calculates PED's and a few other things. pullarc.f : extracts the archive record from a log file and writes this to a file with extension .arch gar2ped.f : reads the archive record (extension .arch) and calculates PED's. The program gar2ped produces : an output file, tetrazine.out, and several .xyz files which can be viewed using XMOL. The file tetrazine.nomos.xyz contains all normal modes (one after another) with arrows indicating in which direction atoms will move. The files tetrazine.'number'.xyz contain for normal mode 'number' the input for the normal mode 'to be played' using XMOL, showing the molecule as it vibrates. A script 'go' contains the input as it can be given interactively, for triazine it shows how internal coordinates can de defined. From: "Andrey V. Khavryutchenko" ,=20 andrey@compchem.kiev.ua We have program set to calculate vibrational spectra and it has the feature you need - the vibrational modes are treated in internal coordinates. The cartesian force field from external program is converted to the force field in internal coord's. Then you can calculate frequencies, the distribution of the potential energy between internal vibration coordinates, the amplitudes of vibrations and directions of atom movement, etc. IR and inelatric neutron scattering spectra intensities could be calculated to.=20 Unfortunaly it is the comercial product, which may be unsuitable for you. We are using home-made semiempirical package to produce the force field. = =20 But since the input format is simply ASCII, I beleive our spectroscopy=20 programs can read almost every output after little reformating (done by=20 computer). C O S P E C O A Computational Spectroscopy Program. --- Functional Description=20 Geometry structure: - Input of structure from different molecular files formats - Matrix conversion from Cartesian to internal dependent coordinates.=20 - Calculation of a matrix of kinematic coefficients. - Calculation of a set of equivalent vibrational coordinates. - Calculation of a set of vibrational coordinates that have no=20 common atoms. Force field: - Convertion of a force field matrix from Cartesian coordinates=20 to dependent internal coordinates. - Scaling of force constants obtained in the course of quantum chemistry calculations. - Creation of a force field matrix from a force constant data bank. - Creation of force constant and scaling factor banks. Vibrational problem: - Solving of a direct harmonic vibrational problem. - Solving of an inverted frequency problem. - Incorporation of different models for force field. Intensity: - Calculation of IR-spectrum intensities. - Calculation of inelastic neutron scattering spectrum intensities. - Solving of an inverted intensity problem and seeking of scaling factors. - Creation of a scaling intensity factor bank. --- Application Field=20 - COSPECO is used in many areas of material science, =09just some of them: =09 high temperature superconductors,=20 =09 vitamines,=20 =09 proteins and DNA=20 surface and inteface studies=20 =09 catalysis,=20 =09 adsorption, =09 surface zone vibrations study --- Requirements=20 Hardware: - IBM PC-compatible computers: =09Intel 486 with 8 Mb of RAM or more and 10 Mb of free disk space - SUN SPARCStation ( 2 or better ) - other ports available on request Operating system: - IBM PC-compatible computers: =09DOS =09Windows 3.11 =09Windows'95 =09Windows NT 3.5 - Unix SYSV User interface: - Windows GUI (Graphical User Interface) under Windows on PC's - X Windows on Unix'es ____________________ Vibration Programs in QCPE ___________________ 665. NIVELON: Calculation of Anharmonic Vibrational Energy Levels by Camelia Munoz-Caro and Alfonso Nino, E. U. Inform=A0tica de Ciudad Real= , Universidad de Castilla-La Mancha, Ronda de Calatrava 5, 13071 Ciudad Real, Espa*a NIVELON is a program for the computation of energy levels and populations for any kind of anharmonic vibration. Several coupled vibrations can be handled simultaneously with the kinetic and potential terms expanded in Fourier or Taylor series. Within a variational framework [1], free rotor and harmonic oscillator eigenfunctions can be used in the basis set. The program permits one to introduce Gaussian perturbations in the potential function and to compute the quantum mechanical average of a magnitude in different vibrational states. The non-rigid group theory can be used for the factorization of the Hamiltonian and the classification of vibrational states. NIVELON is written in standard FORTRAN 77 and [1] C. Mu*oz Caro, A. Ni*o and D. C. Moule, Chem. Phys., l86, 221-231 (1994).=20 Lines of Code: 1826 FORTRAN 77 656. spiroVib: Graphics-Based Normal Mode Analysis by Arka Mukherjee and Thomas G. Spiro, Department of Chemistry, Princeton University, Princeton, New Jersey 08544 The spiroVib program is a graphics-based, menu-driven interface to Normal Mode Analysis. It allows the user to interactively perform vibrational analysis of small to medium-size molecules (<150 atoms) and to visualize the frequencies, eigenvectors and potential energy distributions of the normal modes almost instantaneously. The program enables users to perform calculations with relative ease and to output the frequency and mode compositions in clear formats.=20 The important features of the program are as follows:=20 Graphics-based, menu-driven Normal Coordinate Analysis program. Real-time, mouse-driven rotation (in the x, y and z directions) and scaling of molecule Compatible with Schachtschneider original programs^1. The ability to run CART, GMAT, ZSYM, UBZM, GFROOT and VSEC using the menu-driven user interface.=20 Visual description of normal modes through animation The ability to import other formats like BIOGRAF (bgf) and BROOKHAVEN (pdb= ) Limit the input to Cartesians, masses, connectivity and force=20 constants: The program will be capable of automatically generating the=20 complete z-matrix, which the user can edit. Symmetry coordinates can be=20 generated automatically. The ability to measure geometric parameters (distances, angles, dihedrals and wags) Capable of producing plots of eigenvectors Can be easily redimensioned to perform calculations on larger molecules Hypertext help available Can generate macros that record menu-selection and data input-output Reference: 1. Schachtschneider, J. M., Shell Development Company, Technical Reports No. 57-65 and 231-264 (1962). Lines of Code: 75,387 FORTRAN 77 (Silicon Graphics) 658. FITVIB: Refinement of Kinetic and Potential Energy Functions for Several Large Amplitude Vibrations Using Experimental Data by Alfonso Nino and Camelia Munoz-Caro, E. U. Informatica de Ciudad Real, Universidad de Castilla-La Mancha, Ronda de Calatrava 5, 13071 Ciudad Real, Spain FITVIB is a program for refining kinetic and potential energy functions for several large amplitude vibrations. These functions are expressed as Fourier expansions on the anharmonic large amplitude coordinates^1,2. The program minimizes the sum of squares of the differences between calculated and experimental energy levels using a multidimensional Newton-like algorithm. For large amplitude vibrations the position of the energy levels depends on the kinetic and potential terms. Thus, all these terms are included in the minimization procedure. The program is especially useful for deriving accurate potential functions for torsional (internal rotation) motions. Thus, reliable values for the barriers to rotation and the equilibrium conformations can be obtained. The program is written in standard FORTRAN 77 and runs under the UNIX operating system.=20 References: 1. A. Ni*no, C. Munoz-Caro and D. C. Moule, J. Phys. Chem., 98, 1519 (1994)= . 2. C. Mu*oz-Caro, A. Ni*o and D. C. Moule, Chem. Phys., in press. Lines of Code: 2731 FORTRAN 77 (UNIX) QCMP142. NIVELON: Calculation of Anharmonic Vibrational Energy Levels by Camelia Mu*oz-Caro and Alfonso Ni*o This is the PC version of QCPE 665, announced in this issue of the QCPE Bulletin (see page 48).=20 QCMP103. MOLVIB: Calculation of Harmonic Force Fields and Vibrational Modes of Molecules by T. Sundius, Department of Physics, University of Helsinki, SF-00170 Helsinki, Finland MOLVIB is a program for the calculation of harmonic force fields and vibrational modes of molecules with up to 30 atoms. In the calculation of crystal vibrations, a totality of 50 atoms divided among 11 sub-units can be treated. However, the main sub-unit should not contain more than 30 atoms.=20 Lines of Code: 5200 FORTRAN (Microsoft v. 4.01) 628. REDONG and VISUVIB by A. Allouche, CNRS URA 773, Campus de St. Jer*me, Box 541, 13397 Marseille Cedex 13, France This package consists of two separate programs, REDONG and VISUVIB. REDONG is designed to carry out vibrational analysis for internal coordinates starting from a GAUSSIAN 88 ab initio calculation of vibrational frequencies. The full set of internal coordinates can be transformed into any set of symmetry coordinates after scaling of the vibrational force field.=20 The atomic Cartesian coordinates and the Cartesian force-constant matrix (f) are read from the GAUSSIAN output file (file 22).=20 f is transformed into the mass-weighted matrix f=AB.=20 f=AB is diagonalized in order to obtain the vibrational frequencies.= =20 The full set of internal coordinates is read from file 5 and the B matrix is built (BMAT--QCPE 576--is included in REDONG).=20 Vibrational frequencies can be fitted to experiment by direct scaling.=20 The program VISUVIB is a fully interactive program designed to produce a graphical representation of the normal vibrational coordinates. It uses the IBM GDDM graphics library and an IBM color display station.=20 The only input is the GAUSSIAN 88 or 86 OUTPUT file.=20 VISUVIB displays the atomic coordinates and then the symmetry class, vibrational frequency and eigenfunction for each normal mode as it is read from G88 output file. One can then select one of the normal modes for display.=20 Lines of Code: 3183 VS FORTRAN (IBM) operating under VM/CMS on IBM 30XX systems. 629. KICO: Kinetic Constants Calculation Program by Camelia Mu*oz-Caro and Alfonso Ni*o, E. U. Inform=A0tica de Ciudad Real= , Universidad de Castilla-La Mancha, Ronda de Calatrava s/n, 13071 Ciudad Real, Spain KICO (KInetic COnstants calculation program) is a software tool designed to obtain the kinetic part of the molecular Hamiltonian for internal motions. The program evaluates the vibrational-rotational G matrix1 (which represents the kinetic energy corresponding to the overall rotation and vibrations of the molecule). the internal coordinates are defined as bond lengths, bond angles and dihedral angles through the Z matrix. Dummy atoms are identified with chemical symbol X, and deuterium is identified with chemical symbol D. The program calculates the Cartesian coordinates, the inertial tensor and the principal inertial moments and axis. It obtains the atomic coordinates referred to the center of mass coordinates using the principal axis system. The program is particularly useful for obtaining the kinetic constants of large amplitude vibrations such as internal rotation, inversion or ring puckering. Reference: 1. M. A. Harthcock and J. Laane, J. Phys. Chem., 89, 4231-4240 (1985).=20 Lines of Code: 1609 FORTRAN 77 (SUN, IBM RS/6000, etc.) 631. FCARTP: A Series of Programs Used to Generate a Predicted Set of Fundamental Vibrational Frequencies by William B. Collier, Department of Chemistry, Oral Roberts University, Tulsa, Oklahoma 74171 The method can be summarized as follows. First, a force constant matrix in Cartesian coordinates is generated using a suitable molecular orbital program or other method. Second, from the structure of the molecule a B matrix is calculated and used as input to the FCARTP program. Lastly, FCARTP reads the Cartesian F matrix file and its main input file to obtain all the input data necessary to complete the calculations and produce a set of predicted fundamental vibrational frequencies. Version 1.10 of FCARTP has the capability of:=20 1. Calculating the fundamental vibrational frequencies of a molecule or series of molecules or molecular symmetry blocks after scaling (or not scaling) the input Cartesian F matrix according to the method of Pulay, et al., J.A.C.S., 101, 2550 (1979); J. Chem. Phys., 72, 3999 (1980).=20 2. Adjusting the user-specified scale factors according to the method of nonlinear least squares so that the predicted frequencies of the scaled molecular orbital calculated Cartesian F matrices are minimized with respect to user-entered and weighted experimental vibrational frequencies.= =20 3. Calculating the Cartesian components of and the total intensity of each vibration from user-inputted Cartesian dipole derivatives calculated from the AMPAC molecular orbital program or others.=20 4. Calculating and printing the scaled and adjusted molecular F matrix in internal coordinates for each molecule entered, the potential energy distribution of each vibrational frequency calculated, the Cartesian vibrational displacements of each vibration calculated, the initial and final adjusted scale factors, and various other spectroscopic quantities and matrices of interest.=20 5. Performing overlay isotopic calculations and refining scale factors for an isotopic set of vibrational frequencies.=20 6. Providing instructions for modifying matrices for use in FCARTP.=20 7. Giving specific line by line instructions for modifying QCPE 506 (AMPAC) and BMAT. FOR of QCPE 342 to provide the user with the necessary Cartesian F matrices and B matrices needed for input into the program.=20 Lines of Code: 5235 FORTRAN 77 (VAX) QCMP128. =20 ASYM20: A Program for Force Constant and Normal Coordinate Analysis by Lise Hedberg, Department of Chemistry, Oregon State University, Corvallis, Oregon 97331 and Ian M. Mills, Department of Chemistry, University of Reading, Reading RG6 2AD, England Program ASYM20 is a normal coordinate program which may be used to refine force constants in the harmonic approximation using the following experimental data: wave numbers, isotopic wavenumber shifts, Coriolis coupling constants, root-mean-square amplitudes of vibration, and centrifugal distortion constants. The name ASYM20 is intended to indicate that the calculations may be carried out for molecules of any symmetry with up to 20 atoms. The program has special provisions for handling molecules of higher symmetry, specifically molecules with doubly and triply degenerate normal modes.=20 The theory behind the calculations, and the way it is implemented in the program, is discussed in the paper by Lise Hedberg and Ian Mills, "ASYM20:= =20 A Program for Force Constant and Normal Coordinate Calculations, with a Critical Review of the Theory Involved", which is scheduled for publication in the Journal of Molecular Spectroscopy in the spring of 1993.=20 The refinement is usually carried out using symmetrized force constants.= =20 For small molecules where there are no redundancies, it is possible to carry out a refinement based on valence force constants. If redundancies are present, these must be eliminated before the calculations can be carried out. The structure of the molecule is entered in the form of Cartesian coordinates, and the symmetry coordinates are defined by use of the U-matrix. When symmetry coordinates are used, the G-matrix will be blocked according to symmetry species, and the program will calculate normal coordinates for one symmetry species at a time. These will be kept separate throughout the calculations. Consequently, the observed wavenumbers must be entered according to the symmetry species to which they belong.=20 There is no provision in the program for calculation of Urey-Bradley coordinates, but if these are known, they may be entered in terms of the Z-matrix, and the refinement may be carried out based on Urey-Bradley force constants.=20 The FORTRAN code has been written in very simple form to make it easy to transfer the program to different computers. But there are two statements, CALL DATE and CALL TIME, which are dependent on the compiler and the operating system one is using. The information obtained from these calls is listed at the top of the output file, and we find it useful. It may be necessary to write a small subroutine to make these calls compatible with the computer one is using.=20 Lines of Code: 5647 FORTRAN (Microsoft 5.0) 639. ASYM40: A Program for Force Constant and Normal Coordinate Analysis by Lise Hedberg, Department of Chemistry, Oregon State University, Corvallis, Oregon 97331 and Ian M. Mills, Department of Chemistry, University of Reading, Reading RG6 2AD, England Program ASYM40 is a normal coordinate program which may be used to refine force constants in the harmonic approximation using the following experimental data: wave numbers, isotopic wavenumber shifts, Coriolis coupling constants, root-mean-square amplitudes of vibration, and centrifugal distortion constants. The name ASYM40 is intended to indicate that the calculations may be carried out for molecules of any symmetry with up to 40 atoms. The program has special provisions for handling molecules of higher symmetry, specifcally molecules with doubly and triply degenerate normal modes.=20 The theory behind the calculations, and the way it is implemented in the program, is discussed in the paper by Lise Hedberg and Ian Mills, "ASYM20: A Program for Force Constant and Normal Coordinate Calculations, with a Critical Review of the Theory Involved", which is scheduled for publication in the Journal of Molecular Spectroscopy.=20 The refinement is usually carried out using symmetrized force constants.= =20 For small molecules where there are no redundancies, it is possible to carry out a refinement based on valence force constants. If redundancies are present, these must be eliminated before the calculations can be carried out. In an earlier version of the program the van der Waals radii of atoms consist of internal data. In this version these values are contained in a text file named "VDW.DAT". Thus the user can edit this file in order to change the radius of an atom or to add non-existing values.=20 Lines of Code: 6791 FORTRAN 77 (VAX) 642. RKR (Rydberg, Klein, Rees) Method Program by Emilio Gallicchio, Department of Chemistry, Columbia University, New York, New York 10027 and Franco Battaglia, Dipartimento di Chimica, Universita degli Studi della Basilicata, 85100 Potenza, Italy The RKR method is a computational tool in the spectroscopy of diatomic molecules since it provides the Born-Oppenheimer nuclear potential energy curve for a particular electronic state given the experimental vibrational levels of the electronic state.=20 This program differs from older implementations in that it does not require as input the coefficients of a Taylor Series expansion in (v+1/2) of the standard spectroscopy terms G(v). This approach can achieve good accuracy over the entire range of v.=20 Lines of Code: 621 FORTRAN 77 (IBM RS/6000) 648. ROVI: RO-VIbrational Properties Calculation Program by Alfonso Nino and Camelia Munoz-Caro, E. U. Informatica de Ciudad Real, Universidad de Castilla-La Manca, Ronda de Calatrava s/n, 13071 Ciudad Real, Spain ROVI is a software tool designed to obtain the fine vibrational structure of electronic spectra for large amplitude vibrations. The program permits the user to obtain the energy levels for the transitions and the intensities. ROVI generates two files for plotting the spectrum where the transitions are represented by vertical lines or Lorentzian shapes.=20 Potential energy functions, kinetic constants and transitions dipole moments can be input as a function of the vibrational coordinates. The non-rigid group theory can be used for factorizing the Hamiltonian matrix and classifying the energy levels. The user must supply the basis functions for the calculations. ROVI has been written in standard FORTRAN 77 and has been tested under the UNIX operating system.=20 Lines of Code: 1748 QCMP012. General Vibrational Analysis Programs Utilizing the Wilson GF Matrix Method for a General Unsymmetrized Molecule by Douglas F. McIntosh and Michael R. Peterson, Department of Chemistry, University of Toronto, Toronto, Ontario Canada M5S 1A1 Converted by Timothy J. O'Leary, Department of Health and Human Services, Bethesda, Maryland 20205 This set of three programs allows the user to analyze a general vibrational problem in terms of the method of Wilson, Decius and Cross1.=20 The programs allow for a general solution with a complete set of internal coordinates or, if desired, a less complete and more restricted basis set.= =20 Previous methods of solving the general vibrational equation GFL =3D Ll where G and F are the familiar matrices of the Wilson method, L is the eigenvector matrix and l is the diagonal eigenvalue matrix, have been hampered by the requirement that the basis vectors (internal coordinates) be orthonormalized. This is generally accomplished by applying both symmetry projection operators and a Schmidt orthogonalization process.=20 In the methodology of the present programs, the basis vectors need not be orthogonal. This allows for the direct solution of the vibrational problem without recourse to the use of symmetry coordinates, although provision has been made for their use if desired. The programs will generate the required 3N-6 (or 3N-5, for linear molecules) eigenvalues expected for a complete analysis from 3N-6 (3N-5) to 3N basis vectors. The normal difficulties inherent in redundant coordinates or modes present no problems.=20 Program 1 computes the Wilson B matrix which will be used in all the remaining programs. The BMAT program is a modification of the original GMAT program of J. H. Schactschneider2 and allows for the calculation of six different types of internal coordinates, namely: (1) Bond Stretch, (2) Valence Angle Bend, (3) Out-of- Plane Wag, (4) Torsion, (5) Linear Bend (defining 2 internal coordinates), and (6) Linear Bend (defining 1 internal coordinate). Two important differences between Schactschneider's original GMAT program and BMAT is the inclusion of R. L. Hilderbrandt's method of normalization of the torsional coordinate3 and the use of new formulae for the Out-of-Plane Wag.=20 Program 2 is a dual-purpose program allowing the user to compute either the complete set of eigenvalues for a series of isotopically related molecules or the series of matrices important to the interpretation of a vibrational problem. These matrices include (1) the L matrix, (2) the Potential Energy Distribution matrix, (3) the Atom Displacement matrix, and (4) the Root Mean Square Amplitudes of Vibration between Atom Pairs (bonded and nonbonded), Cartesian Displacements for the equilibrium geometry and Internal Coordinates. The first part of program 2 also allows the user to manipulate the F matrix with individual calculations to obtain an approximate fit of experimental data.=20 Program 3 utilizes the SIMPLEX optimization algorithm of Nelder and Mead4 to refine the guessed force constants via a non-linear least-squares analysis between calculated and observed frequencies (eigenvalues). References: 1. E. Bright Wilson, Jr., J. C. Decius and Paul Cross, Molecular Vibrations: The Theory of Infrared and Raman Vibrational Spectra (New York: McGraw-Hill Book Co.), 1955.=20 2. J. H. Schactschneider, Reports 231/64 and 57/65, Shell Development Co., West Hollow Research Center, P.O. Box 1380, Houston, Texas 77001.=20 3. R. L. Hilderbrandt, J. Mol. Spectroscopy, 44, 599 (1972).=20 4. J. A. Nelder and R. Mead, Computer Journal, 7, 1809 (1965).=20 FORTRAN 77 (Microsoft FORTRAN) Lines of Code: 2650 QCMP038. =20 NCRDWC: A Program to Determine Vibration Frequencies and Normal Modes of Vibration by H. L. Sellers, L. B. Sims, Lothar Schafer and D. E. Lewis, Department of Chemistry, University of Arkansas, Fayetteville, Arkansas 72701 Converted by K. J. Tupper, QCPE The potential energy for vibration of a molecule is expressed in terms of internal valence displacement coordinates. The internal coordinates which are used are bond stretching, valence-angle bending, out-of- plane bending and torsion about chemical bonds. For details, see the article by J. C. Decius in J. Chem. Phys., 17, 1315 (1949).=20 As small displacements are involved, the potential energy is expanded in the displacements from equilibrium, yielding: ... The program accepts a general or simple valence force field with force constants corresponding to the previously mentioned internal coordinates as described in Molecular Vibrations (New York: McGraw-Hill Book Co., 1955), pp. 54-61, by Wilson, Decius and Cross.=20 The program converts the potential energy matrix to one which is appropriate to mass- weighted Cartesian displacement coordinates, since this affords an efficient solution for the frequencies.=20 This software differs from QCPE 339 in that no output is written to the screen as in the original program, and two files, INPUT and OUTPUT, are opened for I/O of data. Thus, a data file must be named INPUT in order to be read. This may be changed by the user to suit his individual needs by changing the statement:=20 OPEN(5,FILE=3D'INPUT',STATUS=3D'OLD') Also note that the output file = is given a status of NEW, so all existing files of this name will be deleted by executing this program. The name of the output file may also be changed by the user by changing the "OPEN(6, " statement. Data which the user does not want written to the OUTPUT file may be written to the screen by substituting a ' * ' for the number '6' in the appropriate WRITE statements.=20 The software package includes a complete version of QCPE 339 as well as the individual subroutines. Sample input and the resulting output are provided, as is the executable program QCP339.EXE which was used to generate the output. The software was compiled using Microsoft FORTRAN-77, version 3.2, using an 8087 coprocessor. The data for cis and trans (NO)2 are not included in the INPUT data file but have been included in the file CISTRANS.DAT.=20 FORTRAN 77 (Microsoft FORTRAN) Lines of Code: 624 QCMP067. UMAT: General Vibrational Analysis Program by Douglas F. McIntosh and Michael R. Peterson, Depatment of Chemistry, Lash Miller Chemical Laboratories, University of Toronto, 80 St. George Street, Toronto, Ontario, Canada M5S 1A1 This set of four programs is version 2.0 of the system BMAT (QCMP012), which is designed to allow the user to analyze the general vibrational problem in terms of the method of Wilson, Decius, and Cross.1 The program allows for a general solution with a complete set of internal coordinates or, if desired, a less complete and more restrictive basis set.=20 The four programs included in this package are:=20 1) UMAT - A multifunctional program which is intended to "set-up" the entire analysis, 2) BMAT - A modification of the GMAT program of J. H. Schachtschneider.2 BMAT generates an unsymmetrical B matrix in a form suitable for use by programs 3 and 4. BMAT is included as a subroutine in UMAT.=20 3) FTRY-ATOM-RMSA-INTY - May be used to obtain the complete set of frequencies for a set of isotopically related molecules, the normal coordinate analysis of the first atom of this series, the R.M.S.=20 amplitude of vibration, and the infrared intensities of the frequencies of a series of isotopically related molecules.=20 4) FFIT - Uses the Simplex method of Nelder and Mead3 to refine the best-guess force constants via a non-linear least-squares analysis between the calculated and observed frequencies. References: 1. E. Bright Wilson, Jr., J. C. Decius, and Paul Cross, Molecular Vibrations: The Theory of Infrared and Raman Vibrational Spectra (New York: McGraw-Hill Co.), 1955.=20 2. J. H. Schachtschneider, Report 231/64 and 57/65, Shell Development Co., West Hollow Research Center, P.O. Box 1380, Houston, Texas 77001.=20 3. J. A. Nelder and R. Mead, Computer Journal, 7, 1809 (1965).=20 FORTRAN 77 (Microsoft, v. 4.1)=20 Lines of Code: 6309 __ _________=20 / \ / _ _ \ Artem Masunov - amasunov@shiva.hunter.cuny.edu / \\ \\ \\ \ Chemistry Department, Hunter College / /\ \\ \\ \\ \ City University of New York / ____ \\ \\ \\ \ 695 Park Avenue, New York, NY 10021 /__/\__/\__\\__\\__\\__\ Tel: (212) 725-0317, Fax: (212) 772-5332 \__\/ \/__//__//__//__/ From yuan@nka1.med.uc.edu Tue Mar 5 16:32:33 1996 Received: from nka1.med.uc.edu for yuan@nka1.med.uc.edu by www.ccl.net (8.7.1/950822.1) id PAA07502; Tue, 5 Mar 1996 15:41:45 -0500 (EST) Received: by nka1.med.uc.edu (951211.SGI.8.6.12.PATCH1042/951211.SGI.AUTO) id PAA02434; Tue, 5 Mar 1996 15:42:34 -0500 Date: Tue, 5 Mar 1996 15:42:27 -0500 (EST) From: Jie Yuan To: Computational Chemistry List Subject: LaTex on Macintosh Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII I think the most important one is missed in the summary (sorry I have deleted the email). One Mac implimentation of LaTeX is called OzteX, available from ftp site "midway.uchicago.edu", in the pub/OzTeX/oztex directory. There are 3 versions, with the oldest version being freeware and newer versions shareware. I have not used OzTeX but many people used and praised it very much in the c.s.mac.apps group. If you search the hyperarchive for latex, you'll find a readme file that can lead you to it. Cheers! Jie -- Jie Yuan, PhD - U. Cincinnati - Pharmacology & C.B. -- == POBox 670575, Cin., OH 45267-0575 = 513-558-2352 == == Jie.Yuan@UC.edu = www.uc.edu/~yuanj = using Pine == From ysandler@post.smu.edu Tue Mar 5 17:32:33 1996 Received: from post.cis.smu.edu for ysandler@post.smu.edu by www.ccl.net (8.7.1/950822.1) id QAA07835; Tue, 5 Mar 1996 16:33:19 -0500 (EST) Received: by post.cis.smu.edu (/\==/\ Smail3.1.28.1 #28.1) id ; Tue, 5 Mar 96 15:33 CST Date: Tue, 5 Mar 1996 15:33:15 -0600 (CST) From: Yakov Sandler Subject: unsubscribe To: chemistry@www.ccl.net Message-ID: MIME-Version: 1.0 Content-Type: TEXT/PLAIN; charset=US-ASCII unsubscribe Yakov Sandler