# Dipole moment

6 years 9 months ago #262
by tamas821

Replied by

*tamas821*on topic*Dipole moment*
Dear Professor Kállay,

I recently returned to the matter of computing the dipole moment values of H2He+ in its ground and excited electronic states, and would like to ask for your help in something I don't understand.

In my FCI/aug-cc-pVTZ calculations I found that the computed dipole moments for the first and second excited states behaved strange, i.e. they changed "discontinously" with respect to small changes in the nuclear distances.

Therefore, I made some test calculations to see what the issue might be. My question is about these test calculations. In my understanding, for H2He+ with three electrons, the FCI and CCSDT methods should be the same in principle. But when comparing FCI/cc-pVDZ and CCSDT/cc-pVDZ computations, I find a huge difference in the dipole moments of the excited electronic states:

FCI/cc-pVDZ

Energy -3.049118919 dipole 0 0 0.01723041

Energy -2.893181294 dipole 0 0 0.00205867

Energy -2.890695799 dipole 0 0 0.0234418

CCSDT/cc-pVDZ

Energy -2.890695799 dipole 0 0.00000498 -5.83185399

Energy -2.713822633 dipole 0 -0.00001074 -5.58909303

Energy -2.713822633 dipole 0 -0.00001074 -5.58909304

For the ground state the two calculations are in perfect agreement, but for the excited states (with energies -3.126094303 and -2.890695799) the dipole values are very different. Can you please explain the origin of the differences?

Attached you can find the input and output files.

Many thanks in advance,

Tamás

I recently returned to the matter of computing the dipole moment values of H2He+ in its ground and excited electronic states, and would like to ask for your help in something I don't understand.

In my FCI/aug-cc-pVTZ calculations I found that the computed dipole moments for the first and second excited states behaved strange, i.e. they changed "discontinously" with respect to small changes in the nuclear distances.

Therefore, I made some test calculations to see what the issue might be. My question is about these test calculations. In my understanding, for H2He+ with three electrons, the FCI and CCSDT methods should be the same in principle. But when comparing FCI/cc-pVDZ and CCSDT/cc-pVDZ computations, I find a huge difference in the dipole moments of the excited electronic states:

FCI/cc-pVDZ

**Energy -3.488512899 dipole 0 0 0.00560133**

Energy -3.126094303 dipole 0 0 -0.00893903Energy -3.126094303 dipole 0 0 -0.00893903

Energy -3.049118919 dipole 0 0 0.01723041

Energy -2.893181294 dipole 0 0 0.00205867

Energy -2.890695799 dipole 0 0 0.0234418

CCSDT/cc-pVDZ

**Energy -3.488512899 dipole 0 0 0.00560132**

Energy -3.126094303 dipole 0 0 -5.66094205Energy -3.126094303 dipole 0 0 -5.66094205

Energy -2.890695799 dipole 0 0.00000498 -5.83185399

Energy -2.713822633 dipole 0 -0.00001074 -5.58909303

Energy -2.713822633 dipole 0 -0.00001074 -5.58909304

For the ground state the two calculations are in perfect agreement, but for the excited states (with energies -3.126094303 and -2.890695799) the dipole values are very different. Can you please explain the origin of the differences?

Attached you can find the input and output files.

Many thanks in advance,

Tamás

Please Log in or Create an account to join the conversation.

6 years 9 months ago #263
by kallay

Best regards,

Mihaly Kallay

Replied by

*kallay*on topic*Dipole moment*
Dear Tamás,

Thank you very much for reporting this. Please edit mrcc.f, at the beginning of subroutine step insert the line

if(calc.eq.0.and..not.osc) ampfile=tfile

before line

if(osc.and.(l3route.or.d2route)) lamfile=lfile

At the middle of the routine insert the line

if(calc.eq.0.and..not.osc) file2=tfile

after line

if(qcc.eq.1) file2=ampfile

This will fix this problem.

Please also note that a CC excited state property calculation is only possible for one excited state at a time, but this is not true for CI. That is, you can set nstate=5 if a CI calculation is run, but if you use CC, you must run a separate calculation for each excited state with nstate=2.

Thank you very much for reporting this. Please edit mrcc.f, at the beginning of subroutine step insert the line

if(calc.eq.0.and..not.osc) ampfile=tfile

before line

if(osc.and.(l3route.or.d2route)) lamfile=lfile

At the middle of the routine insert the line

if(calc.eq.0.and..not.osc) file2=tfile

after line

if(qcc.eq.1) file2=ampfile

This will fix this problem.

Please also note that a CC excited state property calculation is only possible for one excited state at a time, but this is not true for CI. That is, you can set nstate=5 if a CI calculation is run, but if you use CC, you must run a separate calculation for each excited state with nstate=2.

Best regards,

Mihaly Kallay

Please Log in or Create an account to join the conversation.

6 years 9 months ago #264
by tamas821

Replied by

*tamas821*on topic*Dipole moment*
Dear Professor Kállay,

thank you again for your prompt help, it solved the problem.

With warm regards,

Tamás

thank you again for your prompt help, it solved the problem.

With warm regards,

Tamás

Please Log in or Create an account to join the conversation.

Time to create page: 0.016 seconds