sci.math
[Top] [All Lists]

Help with dihedral derivatives

Subject: Help with dihedral derivatives
From: Jellby
Date: Sat, 30 Jul 2005 18:52:41 +0200
Newsgroups: sci.math, sci.math.symbolic
Hi all,

I'm seeking some help for obtaining the analytical derivatives of the
dihedral angle defined by 4 points with respect to the cartesian
coordinates of these points.

Instead of trying to explain the problem with poor ascii, I think it's
better to give a pointer. It's explained clearly in the paper J. Chem.
Phys. 117, 9160, the relevant part of it I have uploaded to:
<http://djelibeibi.unex.es/files/diedro.pdf>. What I'm looking for is the
right form for equations 34 and 35.

I have some reasons to believe the minus sign in eq. 34 should actually be a
plus. Comparing with similar expressions in other books and computer
programs, sometimes there's a minus and sometimes there's a plus, but
things work better with a plus and at least in one case the minus was later
corrected, as it was a typo.

Eq. 35 is a bigger monster. I'd say the "u_j" in the fourth term should be
"v_j" (nothing else in that term is related to u, but to v), and the very
last "zeta" factor should be zeta_(bon) instead of zeta_(bom) for similar
reasons. However, even with those changes, the second derivatives are not
symmetrical due the the last two terms, which are symmetrical with respect
to the a<->b interchange but antisymmetrical with respect to i<->j; and the
sum of all second derivatives does not vanish, which I believe it should,
due again to these last two terms.

So, I'm almost sure these equations are not fully correct, probably just
because of a couple of typos, but I can't get the right expressions. I have
written to the paper's author but have got no reply yet. Could anyone help
me pinpoint the mistakes in eqs. 34 and 35?

Thanks in advance

-- 
     Ignacio __ Fernández Galván
            / /\
Linux user / /  \  PGP Pub Key
#289967   / / /\ \  0x01A95F99
         / / /\ \ \
                                   djelibeibi.unex.es">http://djelibeibi.unex.es
       /________\ \ \
jellby \___________\/ yahoo.com

<Prev in Thread] Current Thread [Next in Thread>
  • Help with dihedral derivatives, Jellby <=
Privacy Policy