[molpro-user] Numerical Freq calculations using F12-CCSD(T)

[John Travers]

Hi
I was trying to do a numerical frequency calculation using f12-ccsd(t) in c1 symmetry for a small molecule. The program always produces some warning messages like : "Poor overlap with frozen basis set. May have deleted the wrong vectors! ? The problem occurs in SvdConstructOrthoBasis"
When checked the detailed output,  I also found some possible numerical errors, something looks like :
 Construction of CABS: Pseudo-inverse stability               NaN Smallest eigenvalue of S         -3.71E-13  (threshold= 1.00E-08) Ratio eigmin/eigmax              -3.71E-13  (threshold= 1.00E-09) Smallest eigenvalue of S kept    -3.71E-13  (threshold=-3.71E-13, 0 functions deleted, 658 kept) Maximum overlap with frozen deleted vectors: (nDel=  44)    Vector:         1       2       3       4       5       6       7       8       9      10      11      12      13      14      15      16      17      18      51      25      67      44      36      35      43      32      40      38      33      39      55      52      34      22      37      21      42      50      20      41      74      19      53      54      26      23     231      28      24      95      93      27     158      75      61     123      68     108      31      63     163     121     203     152      62      29     157      45      94      72      97      77      65     204     103      30     115     126      66     250      73     131     101     162     116      91     200     122    Overlap:       NaN     NaN     NaN     NaN     NaN     NaN     NaN     NaN     NaN     NaN     NaN     NaN     NaN     NaN     NaN     NaN     NaN     NaN   1.436   1.248   1.102   1.000   0.999   0.997   0.996   0.994   0.985   0.983   0.973   0.964   0.962   0.912   0.896   0.871   0.858   0.853   0.827   0.825   0.742   0.596   0.594   0.537   0.523   0.485   0.465   0.386   0.375   0.373   0.346   0.327   0.321   0.310   0.307   0.268   0.259   0.257   0.249   0.247   0.216   0.204   0.194   0.191   0.183   0.166   0.164   0.159   0.135   0.134   0.130   0.126   0.118   0.114   0.110   0.100   0.095   0.093   0.092   0.092   0.091   0.089   0.080   0.079   0.079   0.078   0.076   0.075   0.072   0.070

My questions are related to these "NaN" and warnings. Do these matter and how to avoid this? Thanks!
Best wishes

John


-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://www.molpro.net/pipermail/molpro-user/attachments/20170116/854fb7fc/attachment.html>