Hello everyone,
I am trying to perform IRCs for calculations that involve ECP. As per usual, I performed my transition state optimization and frequency calculations, and the structures had 1 imaginary frequency.
When I tried to perform IRCs, I was not sure how the input file should look.
First, I plugged in the TS coordinates, with irc(CalcFC,...). This method didn't work.
Second, I performed a frequency calculation, linked it to an IRC calculation with irc(rcfc) by directing the checkpoint file to the frequency calculation. Then, I specified the ECP after the charge and multiplicity line. This method didn't work either.
If anyone has an experience with IRC calculations involving ECP, I would greatly appreciate any instructions or some template on how to perform this.
Dear Mikey,
I suggest you to perform the calculation in different files (this do not change the command lines).
On the other hand, in the input you have too much spaces in the %chk to %mem commands. Also, in order to call to a irc the input command is # irc followed by the method and basis set. Also, if the error is "Maximum number of corrector steps exceeded." I recommend you to disable correct steps (sometime they gives numerical problems), the command that you must add is:
# irc=( maxpoints=40,recalc=1,recorrect=never,rcfc)
PD: In the part where you introduce the basis set, the part os pseudo-potential, you must include 2 spaces better.
I hope it helps you,
Joaquim Rius
Dear Mikey,
In order to perform this calculation, you only must take care that the basis set is added and the command to read it. First is the command lines, then the geometry of TS and finally the basis set. Check the example of input. This example is not a TS but the calculation run correctly (of course, it will not find nothing). Even so, follow the same scheme for your TS (remember the option pseudo=read or the key genecp).
On the other hand, the problem could come from other factors. It is typical that IRC gives numerical problems. Sometimes a solution to calculation errors, is disable "recorrect steps to never " or add less points.
I hope it helps you,
Joaquim
http://www.gaussian.com/g_tech/g_ur/k_irc.htm
Hello Joaquim,
I would like to thank you for your response. Here is what I've done, and here is what I am planning to do.
%chk=frequency
%nprocshared=16
%mem=3800mw
#p B3LYP/GENECP freq nosymm iop(1/8=5) scf=novaracc SCF=Tight int(grid=ultrafine)
Title
0 1
coord
C H P 0
6-31G(d)
****
Rh 0
SDD
****
Rh 0
SDD
--Link1--
%chk=frequency
%nprocshared=16
%mem=3800mw
#p B3LYP/SDD irc(rcfc,maxpoints=40,recalc=1,stepsize=12,hpc) geom(allcheck) nosymm iop(1/8=5) guess=mix scf=novaracc SCF=Tight int=grid=ultrafine
IRC B3LYP/GENECP
0 1
(no coordinates herein since i specified the frequency checkpoint file and demanded the IRC to read the frequency checkpoint file for force constants)
Should I perhaps change the "#p B3LYP/SDD irc(..)" line to #p B3LYP/GENECP pseudo=read irc(...)" or is there anything in the IRC that could contribute to the calculation not running?
I've also increased the irc maxpoints from 20 to 40, due to the error "Maximum number of corrector steps exceeded." Will that be advisable as well?
Thank you so much,
Mikey
Addendum: I have already performed a separate frequency calculation for the TS structure in question, and it contained only one negative frequency.
Dear Mikey,
I suggest you to perform the calculation in different files (this do not change the command lines).
On the other hand, in the input you have too much spaces in the %chk to %mem commands. Also, in order to call to a irc the input command is # irc followed by the method and basis set. Also, if the error is "Maximum number of corrector steps exceeded." I recommend you to disable correct steps (sometime they gives numerical problems), the command that you must add is:
# irc=( maxpoints=40,recalc=1,recorrect=never,rcfc)
PD: In the part where you introduce the basis set, the part os pseudo-potential, you must include 2 spaces better.
I hope it helps you,
Joaquim Rius
- Badges
- Science topic