3
$\begingroup$

My input file is:

%CPU=8
#P SCF=TIGHT CCSD(T)/aug-cc-pVTZ Opt(ReadOptimize,tight,z-matrix)

h2o test 3

0 1
O
H                  1    B1
H                  1    B1     2  A1

B1 0.96000000
A1 109.50000032

atoms=3 notatoms=1,2

And I am getting this error:

 ************************************************
 ** ERROR IN INITNF. NUMBER OF VARIABLES (  0) **
 **   INCORRECT (SHOULD BE BETWEEN 1 AND 50)   **
 ************************************************

I followed these instructions. But, they did not work for me.

**What is wrong with my input? **

$\endgroup$
2
  • $\begingroup$ Hi, do you want a full optimization (i.e. no any coordinate constraint), or just want to optimize some geometry variables? Answers depend on your optimization target. $\endgroup$
    – jxzou
    Dec 6, 2023 at 2:59
  • $\begingroup$ Hi @jxzou , I want to optimize only the second hydrogen. $\endgroup$ Dec 6, 2023 at 9:01

3 Answers 3

2
$\begingroup$

Non-gradient optimizations only work in internal coordinates, and if you only want the last H to optimize, you cannot have the B1 variable to also define atom 2.

$\endgroup$
2
$\begingroup$

TL/DR You cannot easily do optimisations with CCSD(T) in Gaussian. Try a different method instead (CCSD perhaps).

Gaussian only supports numerical gradients for CCSD(T), and the optimisation procedure defaults to the 'EigenvalueFollow' (Opt=EF) procedure when analytic gradients are not available. This method is limited to solving max 50 variables, which is likely too be too few for a 'real' complex molecule. See http://www.ccl.net/chemistry/resources/messages/2012/03/13.016-dir/index.html. Non-analytic optimisations are also famously, painfully slow. Unless you are really interested in solving the structure of water, I would recommend using a method with analytic gradients instead (pretty much anything other than CCSD(T), MP5 and MP4 I think).

Also

%CPU=8 requests the calculation run only on CPU 8 (ie, one CPU max). If you instead wanted to run a parallel calculation using 8 CPUs, you want:

%NProcShared=8

$\endgroup$
1
$\begingroup$

Are there any extra empty lines after the final line you've posted? In my experience, you need to add a few empty lines at the end of your input file if you want the calculation to work.

In other words, the last part of your input should look like this:

0 1
O
H  1  B1
H  1  B1  2  A1

B1 0.96000000
A1 109.50000032

atoms=3 notatoms=1,2


 

(Note the extra lines below the end of the input.)

That being said, if your plan is to optimise one O–H bond and not the other, you should probably add a B2 variable to your input, otherwise I'm not sure anything other than the angle will change.

$\endgroup$

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .