Talk:Mod:Hunt Research Group/Chemshell Step By Step

Step 1 - Molcluster

The first thing that needs to be done to run any chemshell job is generate the input files. To cut the clusters a successful MD simulation is required and MolCluster needs to be installed.

For information on how to obtain, install and run MolCluster, read here.

Step 2 - Editing input files

The next stage of the process is editing the input files to do what you want them to do. From the MolCluster directory on your local machine you need to copy the opt.chm, ff.dat and cluster_1.chm files to the hpc.

the cluster_1.chm file might be named something else! the file will always end *_1.chm, *_2.chm *_3.chm etc.

The files need to be edited before you can run a ChemShell job.

ff.dat needs to contain the parameters from the MD simulation. This is not a simple copy and paste, the numbers needs by ChemShell are in kCal while the numbers often used in DL_poly are in kJ/mol

HOW DO YOU CHANGE BETWEEN THE TWO

the cluster_1.chm (created by MolCluster) file needs to become the cluster.pun (input file for ChemShell) file required by ChemShell. There are 2 methods to do this depending on the ChemShell code you are using. Both methods are explained on the Running ChemShell page here

opt.chm is essentially the control file. It contains all of the keywords and tells ChemShell what you want it to do.

Each of these files are explained here

Errors that have been encountered

Point charge position with Pseudopotentials

As I have been trying to model an La ion in molten NaCl I need to employ a pseudo potential for my La ion. When trying to use an in-built PP, such as LanL2DZ the gaussian.com file has point charges between the basis set and ecp sections. This resulted in ECP's not being included on the ions and the job would fail. The gaussian.log file included a large section with all ions having "No pseudo potential on this centre."

To overcome this problem it was necessary to edit the gaussian.tcl file in the ChemShell tcl directory. When editing tcl files it is important to first make a copy of the file calling it XXXX_original this way no matter what happens you have a copy to start from again! The second important thing is to count the brackets, you have not got to the end of a section if more brackets have been opened than have been closed. I searched for sections on ecp's and moved them (by copying and pasting) to just below sections on basis sets.

LINK TO MY GAUSSIAN.TCL FILE

gaussian.update

In the opt.out file a section was included showing the following

******************** DL_POLY Pairlist calculation requested ********************

hybrid: dl_poly updated OK
hybrid: Warning gaussian update failed
        invalid command name "gaussian.update"

This particular problem is that the gaussian.update section had not been specified in the gaussian.tcl file or the tclIndex file.

In the tclIndex file the following line needs to be on line number 371

set auto_index(gaussian.update) [list source [file join $dir gaussian.tcl]]

In the gaussian.tcl file the following need to the final section

proc gaussian.update { args } {
    return 0
}