|Posted on Tuesday, November 29, 2011 - 5:34 am: |
Hi, we have recently built a server hosting Webmo with GAMESS and MOPAC2009 for our educational institution. We have received reports saying that when they carry out a 'triplet' calculation it either fails or produces an error or strange looking structure. The singlet calculations are all correct. Does anybody have an idea what could be causing this please? I have attached a successful output log which was reported as being incorrect results. Thank you.
Post Number: 211
|Posted on Tuesday, November 29, 2011 - 10:17 am: |
The ground state of methane is NOT a triplet, it is a singlet. The triplet corresponds to some excited state. Thus the geometry will look "strange". There is no fundamental problem.
|Posted on Tuesday, November 29, 2011 - 10:42 am: |
Thank you once again Schmidt, it was our Chemistry department that reported the fault but I do not know enough about it to argue the findings with them. I will let them know the problem lies with them, all the best, Leigh.
|Posted on Tuesday, November 29, 2011 - 12:00 pm: |
Hi, as the person who made the inital error with regards to methane, can I please add to the information surrounding this problem, as initally described by Leigh_H ? For structures such as methane, a 'strange' geometry IS expected. However an error is occuring even when calculating other structures that can usually be described with a triplet state, such as H2(g). Within our system, a MOPAC/PM3 calculation of H2 (singlet) can be calculated, but a H2 (triplet) calculation produces the following error message: "SPECIFIED SPIN COMPONENT NOT SPANNED BY ACTIVE SPACE".
The corresponding GAMESS/PM3 (UHF) or GAMESS/HF (UHF) calculations give the following error message: "ERROR *** ATTEMPTING A BOGUS READ OF A DAF RECORD.RECORD NUMBER 256 OF LENGTH 2 WAS NEVER PREVIOUSLY WRITTEN. THIS ERROR IS LIKELY TO BE A BUG IN THE PROGRAM, BUT PLEASE CONSIDER INPUT ERRORS AS A POSSIBLE CAUSE, TOO"
Can anyone tell me if I am making another mistake within the input? Thanks.
Post Number: 212
|Posted on Tuesday, November 29, 2011 - 2:40 pm: |
The MOPAC issue appears to be a MOPAC bug, probably because H2 triplet contains zero beta spin electrons (a strange case).
GAMESS/ROHF and GAMESS/UHF calculations on H2 tripelt work perfectly in my testing (using the default input file generated by WebMO). Perhaps your local version of GAMESS contains a bug?
|Posted on Wednesday, November 30, 2011 - 6:01 am: |
Could you please tell me if there are logs we can examine to determine where these bugs could be within the system? Many thanks.
Post Number: 216
|Posted on Wednesday, November 30, 2011 - 10:20 am: |
These are almost certainly not WebMO bugs. If you send me the exact input file that exhibits the error, I can confirm that. The GAMESS output file should contain all the relevant information.