In an attempt to do calculations on anions using Gaussian, I've systematically added diffuse s- and p- functions be scaling the exponent.
This works well, as long as the exponents are above 10^-6.. when they fall below this threshold, I receive a warning stating "Missing exponents in basis set"
I've spent a lot of time looking into the iOps reference manual looking for this elusive threshold, but to no avail. Does anyone know how to accommodate such low basis set exponents (down to 10^-10) in gaussian?
For reference, my input file looks like this currently:
#P UB3LYP/6-311++G(d,p) SCF=(XQC,NoVarAcc,MaxCycle=512)
OPT=(MaxCyc=300,VeryTight,CalcAll)
GFPrint GFInput Test
ExtraBasis ExtraDensityBasis
IOp(3/75=5000434) CPHF=SuperFineGrid
System Name: CN Test Case
-1 1
C -0.35333 -0.27437 0.00198
N 0.29335 0.66987 -0.14055
You have to do the next steps:
1) Go to get the basis set 6-311++G(d,p) with format for Gaussian; you can easily get it here:
https://bse.pnl.gov/bse/portal
2) Use the GEN keyword, examples and explanation here:
http://www.gaussian.com/g_tech/g_ur/k_gen.htm
This means that you have to change UB3LYP/6-311++G(d,p)
by UB3LYP/GEN
3) At the end of the Molecule Specification section of your input, you have to add your basis set; example the 3-21G basis set for C and H (don't forget the stars):
H 0
S 2 1.00
5.4471780 0.1562850
0.8245470 0.9046910
S 1 1.00
0.1831920 1.0000000
****
C 0
S 3 1.00
172.2560000 0.0617669
25.9109000 0.3587940
5.5333500 0.7007130
SP 2 1.00
3.6649800 -0.3958970 0.2364600
0.7705450 1.2158400 0.8606190
SP 1 1.00
0.1958570 1.0000000 1.0000000
****
4) Make any change you want in the exponents and/or coefficients. The exponents are reported at the left of the Gaussian basis set format (confirm this).
I'm here if you need more info or if this is not completely clear.
Kindest, JP
Thank you Bryan M. Wong, I changed the source and now it's working as intended using the ExtraBasis keyword. I don't know about the GEN keyword Pablo, but I expect that would encounter the same limitations as the ExtraBasis keyword for very small exponents? In any case, I appreciate the responses and it seems to be working now :)
An alternative appears to be putting IOp(3/46=2). Then, exponents smaller than SmlExp will issue a warning but not abort the job.
Ah nice, that works also yes. Contrary to previous literature on known molecules, I'm finding that the extra diffuse functions to not affect my results. I'm pretty sure this is because integration points with low density are discarded - do anyone know which iOp to use for disabling this threshold screening of integral points with low density?
- Badges
- Science topic
- Similar topics
- Psychology