From rovshan-0at0-atlas.chemistry.uakron.edu Mon Jul 5 11:19:41 1993 Date: Mon, 5 Jul 93 15:19:41 EDT From: rovshan -AatT- atlas.chemistry.uakron.edu (Rovshan G. Sadygov) Message-Id: <9307051919.AA11686 $#at#$ atlas.chemistry.uakron.edu> To: chemistry -AatT- ccl.net Subject: Oscillator Strength Dear netters, I use Argus package to do CI calculations on Ohio Supercomputer. Transition energies and dipole moments I got are in good agreement with experiment. I have had hard time to understand results for os- cillator strength. I thought that sum of oscillator strengths of all excited states should be equal to 1. But for some states oscillator strength length is greater than 1. Could anyone recommend me any cor- responding reference? Thanx. Rovshan. From mail Mon Jul 5 19:54:40 1993 Date: Mon, 5 Jul 1993 16:31:15 -0400 From: hyper!hurst (Graham Hurst) Message-Id: <9307052031.AA29310;at;hyper.hyper.com> To: Gustavo Mercier , chemistry ( ( at ) ) ccl.net Subject: Re: HONDO and 5 canonical d orbitals Gustavo Mercier wrote: > Does anybody know how to request that HONDO 7 use the > 5 canonical d orbitals (i.e. dx2-y2 rather than dxx and dyy) instead > of the default 6 cartesian d orbitals? Although I am one of the authors of HONDO 7, it's been several years since I've used it and I don't have the manual at hand. I recall that there was no "simple" option to throw away the xx+yy+zz component and use the 5 canonical d orbitals. The closest thing that I recall is a tolerance for throwing away orbitals based on near-zero eigenvalues of the AO-basis overlap matrix. I believe that you can set the tolerance and if you've got a "close" s basis function, you might be able to set it to a value that only threw away the d xx+yy+zz. You'd have to check the output coefficients (Ci) to make sure there was no MO with equal Ci's for dxx, dyy and dzz, since it might throw away the s instead! Sorry that I can't recall or think of a better solution. This test is designed to ensure that linear dependency problems don't arise when using all 6 cartesian d functions. My experience was that, with the notatable exception of Gaussian, it was common to use all 6. I think many ab initio programs did this since they'd calculated the integrals anyway, and when they were written basis sets were so small that the extra s function was not unwelcome. > Also, I have had some difficult interpreting a line from a paper > on effective core potentials for transition metals? > The paper gives the exponents and coefficients for the gaussian > orbitals that form the pseudoorbitals when used as a minimal basis set. > It then goes on to state: > "Contracted ""double zeta" [2s2p2d] basis sets can be obtained > by leaving the outermost function uncontracted and by contracting > the remaining inner functions into a single function > ACCORDING TO THE RATIOS (Ci) IN TABLES X-XII" > The tables are the listing of the exponents and coefficients, Ci's, > for the minimal basis set. > Simply, how do you go from the minimal basis set to a double zeta as > described above? What does the phrase ACCORDING TO THE RATIOS.. > mean? > Reference: P.J.Hay and W. R. Wadt > J.Chem.Phys. 82(1) 270-283, 1985 > "Ab initio effective core potentials for molecular calculations. > Potentials for the transition metal atoms Sc to Hg" > Any help would be welcomed This is easier - just use the Ci's for the contraction coefficients. I'm pretty sure that HONDO will normalize them for you (and will print both the normalized and unnormalized values in the basis set output). Of course the "split-out" basis function should have a contraction coefficient of one. > Thanks > > gus > mercie -8 at 8- cumc.cornell.edu Hope this helps! Graham ------------ Graham Hurst Hypercube Inc, 7-419 Phillip St, Waterloo, Ont, Canada N2L 3X2 (519)725-4040 internet: hurst;at;hyper.com