Ceilidh ... Re^2: ***** FAILURE TO LOCATE STATIONARY POINT, SCF HAS NOT CONVERGED

Firefly and PC GAMESS-related discussion club

Re^2: ***** FAILURE TO LOCATE STATIONARY POINT, SCF HAS NOT CONVERGED

On Sun Jan 10 '10 6:23pm, sanya wrote
-------------------------------------
>One thing every user should remember forever: poor SCF convergence is the first warning that there's something wrong with your problem. It may be just a misprint in the $DATA group, but most frequently poor convergence means that something's wrong with the problem statement. Either the molecule cannot be described within a single-reference approach (for example, an open-shell system) or the molecule is not a molecule at all.

>Your molecule is in fact a fragment of a crystal structure. Are you sure the multiplicity of your cluster is 2? Are you sure that clusters of this type really exist in the gas phase? In crystals, such structures are surrounded by other atoms and, therefore, have no free valences, in contrast to your cluster.

>Again and again: If you're trying to simulate a fragment of a crystal within some cluster approach, take care of capping your cluster properly. Otherwise, even in you force SCF to converge, the result will be far from reality.

>An alternative way is to use plane-wave packages for periodic calculations. But, please, don't ask me about these packages, I have no experience of using them!

>On Sun Jan 10 '10 11:22am, luxiaohua wrote
>------------------------------------------
>>I optimize some molecule structure, but it don't converge even though I add some keyword. such as diis=.t. damp=.t. shift=.t. and so on. it still don't change.

first of all. thank sanya reply to me. 狢 optimize the same structure with other DFT(such as B3LYP) or with GAUSSIAN. it is showed that these structure are converged. How possible it is to converge these structure
what some good ways there are other!
thank you!

Mon Jan 11 '10 7:13am

This message read 1146 times