[molpro-user] MRCI: Catastrophic failure in diagonalization(hsdel)

Mario Tacconi tacconi at caspur.it
Wed Feb 15 14:51:29 GMT 2006


Dear Molpro Users,

I'm trying to describe the CsNH system at the CASSCF-MRCI level of 
theory. I used a small core ECP to describe the Cesium core electrons.
Let me give you some infos about the system I'm studying. The PES for 
the doublet spin symmetry of that system is characterized by an ionic 
ground state. Because the long range 1/r dependence of the ionic 
potential, the ionic state can cross about ten weakly bound covalent 
states when the Cs-NH distance becomes greater and greater.
I found an active orbital space that can give a smooth PES at the CASSCF 
level for relatively large values of the Cs-NH distance. The problem 
arose in the following MRCI calculation that use the CAS wavefunction as 
reference (Cs point group symmetry):

    14     1     1     1.08767736    -0.07142049   -74.82403947 
-0.00022226    -0.00006400  0.58D-04  0.12D-04   193.17
    14     2     2     1.03378445    -0.08760619   -74.81417020 
0.00000000     0.00000000  0.22D-09  0.11D-09   193.17
  Removal of internal vectors e(1)<  0.1000000E-14
  e(1).le.1.d-15  -0.1784811E-15
  Removal of internal vectors e(1)<  0.1000000E-14
  e(1).le.1.d-15  -0.5256488E-15
  Removal of internal vectors e(1)<  0.1000000E-14
  e(1).le.1.d-15  -0.1545717E-15
  Removal of internal vectors e(1)<  0.1000000E-14
  e(1).le.1.d-15   0.3437726E-16
  Removal of internal vectors e(1)<  0.1000000E-14
  e(1).le.1.d-15  -0.3852155E-15
  Removal of internal vectors e(1)<  0.1000000E-14
  e(1).le.1.d-15  -0.1596976E-15
  Removal of internal vectors e(1)<  0.1000000E-14
  e(1).le.1.d-15  -0.1243741E-15
  Removal of internal vectors e(1)<  0.1000000E-14
  e(1).le.1.d-15  -0.2740863E-15
  Removal of internal vectors e(1)<  0.1000000E-14
  e(1).le.1.d-15   0.4198031E-15
  Removal of internal vectors e(1)<  0.1000000E-14
  e(1).le.1.d-15   0.3712308E-15
  Catastrophic failure in diagonalization(hsdel)
  This difficulty can arise for many reasons
  Sometimes it helps to redefine P space
  Otherwise, try increasing or decreasing reference space or nstati

Oddly enough, The problem seems to disappear if I do the calculation 
using the C2v point group symmetry.
However, how suggested by the error message, I tried to 
increase/decrease the nstati value, but it doesn't seem to work. So I 
should try to redefine the P space. Maybe does it mean that I have to 
increase the number of active orbitals? Is there any test that I could 
do to find out the real cause of that problem?

thank you in advance for any help,

Mario Tacconi





More information about the Molpro-user mailing list