Firefly and PC GAMESS-related discussion club

 
Learn how to ask questions correctly  
 
 

 

Yura Vishnevskiy
yuri.vishnevskiy@gmail.com

>Well, I tried to play with your file (with 6-31G(d,p) basis set, for speed, but it does not matter).

>First of all, your input geometry is far from being optimal. After optimization, both 1-2-3-4 and 2-3-4-5 torsions become substantially nonzero.

Hmm, actually I've optimized this geometry (with two fixed torsions and applying exactly the same DFT functional and basis) before using it in Firefly as a starting point.

>Second, in spite of warnings, PES scan runs successfully. If you remove (rather than set to 0) ORIG1 and ORIG2, scan will start with the values defined by $DATA. If ORIG1 or ORIG2 are set to 0, the scan will start with zero torsions 1-2-3-4 and 2-3-4-5.

I thought I know how to do relaxed scans in Firefly but now I am not sure anymore. :)
Let me explain. I need a 2D relaxed PES scan along torsional coordinates 1-2-3-4 and 2-3-4-5. They both should start from 0.0. Obviously, it is irrelevant whether this starting point corresponds to equilibrium. As I understand ORIG1 and ORIG2 are starting values for my PES coordinates. And ORIG1=0 and ORIG2=0 in my case exactly correspond to what is defined in $DATA, that is both torsions are exactly zero. Therefore it is strange to observe that removing these keywords somehow influences the calculation.

>>Black box methods are everywhere. You do know what I an talking about. And please, in this particular question, let me decide which approximation I need.

>Well, don't be rude. It was a joke, I thought you understand :)

Sorry. ;)

>But seriously, I just wanted to say that you may try any other program. You don't actually have to do any calculations. If this alternative program manages to generate internal coordinates without warnings, you may use them in FireFly. If it fails, the molecule is really problematic and the only solution is to generate the internal coordinates manually. But after playing with your molecule I see that there is no problem at all.

Do we use the same version of Firefly? Mine is 7.1.G.

We use up to 10 QC program packages for different purposes. Among different advantages Firefly can also do quite fast DFT calculations (if we are not talking about RI approximation) and the DLC allows to speedup the convergence of geom. optimizations. All this made me to hope that Firefly could be used as a working horse for such calculations.

P.S.: I've checked again examples from Alex. He used additionally IRZMAT to remove coordinates, which are coupled with scan coordinate. The calculation runs fine if I remove two torsions 2-3-4-6 and 2-3-4-7.

[ This message was edited on Wed Dec 26 '12 at 1:33pm by the author ]