[gmx-users] Normal modes

[Andrew Horsfield]

Hi,

I am looking at the normal modes of a single water molecule, and am
hitting a problem. First, I relax the structure to the ground state using
mdrun_d:

Step 1, E-Pot = 1.0346800043e-02, F-max =  5.45631e+01
Step 2, E-Pot = 6.0199001961e-04, F-max =  1.99443e+01
Step 3, E-Pot = 1.3253953913e-08, F-max =  6.71523e-02
Step 4, E-Pot = 6.4280446913e-09, F-max =  4.07163e-02
Step 5, E-Pot = 1.5057999323e-10, F-max =  1.02353e-02
Step 6, E-Pot = 5.1519237405e-14, F-max =  1.01403e-04
Step 7, E-Pot = 2.4508589290e-14, F-max =  9.99452e-05
Step 8, E-Pot = 5.9851088013e-16, F-max =  2.09201e-05
Step 9, E-Pot = 3.2854699832e-19, F-max =  2.14438e-07

This is clearly well converged. Then I build the hessian, and finally I
diagonalise it. The program for building the Hessian (nmrun_d) complains
that:

  Maximum force: 8.59693e+01
  Maximum force probably not small enough to ensure that you are in a 
  energy well. Be aware that negative eigenvalues may occur when the
  resulting matrix is diagonalized.

Why?

The eigenvalues I get from g_nmeig_d are:

     1     6.31598e-07
     2     1.63845e-06
     3    -2.55816e-06
     4         84.4515
     5         519.739
     6        -726.238
     7          107624
     8          432807
     9          450119

Values 4, 5 and 6 should be zero, but clearly are not.

Are the units ps^{-2}?

Cheers,

Andrew

  +----------------------------------------------------+
   Andrew Horsfield       e-mail: horsfield at fecit.co.uk 
    FECIT, Hayes Park Central, Hayes End Road, Hayes,
           Middlesex UB4 8FE, United Kingdom.           
   phone: +44-(0)20-8606-4653  FAX: +44-(0)20-8606-4422 
  +----------------------------------------------------+