On Wed, 2014-05-21 at 10:32 +0200, giulia palermo wrote:
> Dear James,
>
> thank you for your reply. However, I still have some doubts.
>
> In frequencies calculations (with g09) the force constant is expressed in
> mDyne/Ang.
> However, when asking g09 to print the force constant for each degree of
> freedom, one gets the following output:
>
> ! Name Definition Value Derivative
> Info. !
> --------------------------------------------------------------------------------
> ! R1 R(1,2) 1.5017 D2E/DX2 =
> 0.3353 !
> ! R2 R(1,4) 1.0794 D2E/DX2 =
> 0.3863 !
> ! R3 R(1,5) 1.0808 D2E/DX2 =
> 0.3807 !
>
> where R(1,2) is the bond and D2E/DX2 is the force constant. *It seems to me
> that these values are expressed in Hartree*, even if it is not clearly
> specified in the output file.
This Gaussian manual page suggests that force constants are printed in
mDyne/Angstrom, as you stated:
http://www.gaussian.com/g_whitepap/vib.htm
Why do these values seem like they are Hartrees to you? 0.3353
mDyne/Angstrom is equal to roughly 48 kcal/mol/A^2. While this value
seems a bit small -- between a factor of 2 and 5 smaller than other bond
force constants -- assuming these units are Hartrees gives you force
constants on the order of 35,000 kcal/mol/A^2, which is orders of
magnitude too large.
These are useful exercises, I think. Rather than rely on suspicion in
cases like this, take out a pencil, paper, and calculator and do some
quick estimates to see what seems reasonable. This should satisfy you
that the units Gaussian prints are what they claim.
> Do you ever had any experience with this problem??
> The conversion of either mDyne/Ang or Hartree to kcal/mol/(Ang**2) will
> give different values of the Force Constants that will be unreliavle if the
> startung units are wrong.
Certainly if the units are wrong then the force constants will be
unreliable. But a Hartree is over 600 times larger than a kcal/mol (and
over 2500 times larger than a kJ/mol), so it should be easy to determine
which units were used just using basic order-of-magnitude estimates in
this case.
HTH,
Jason
--
Jason M. Swails
BioMaPS,
Rutgers University
Postdoctoral Researcher
_______________________________________________
AMBER mailing list
AMBER.ambermd.org
http://lists.ambermd.org/mailman/listinfo/amber
Received on Wed May 21 2014 - 05:30:02 PDT
