Help for a Complete Beginner

Hello All. Background Chemical Physics, molecular vibrational spectroscopy, nuclear industry, criticality safety. I am retired and have decided to return to my first love - orbitals, electrons and their behaviour - as an intellectual exercise for an old brain. I am running a Psi4 executable under Ubuntu from command line. I have run some test cases. Some very simple questions if I may…

  • how do I get psi4 to output energies in eV?
  • what are the definitions of “one electron energy” and “two electron energy”?
  • can alpha, beta, virtual orbital energies be easily translated to s,p,d, sigma, pi terminology?
  • I have run a single H atom (wow!). How do i get the excited state energies?
  • Multiplicity - why must the spin state be specified up front? I was hoping that ab-initio would converge to the lowest energy at whatever calculated electron structure?
    Any help appreciated.
    Best regards, Donald.

Hi. Welcome to quantum chemistry :slight_smile:

how do I get psi4 to output energies in eV?

I am afraid you would need to convert the energies yourself. For some properties the output will be in eV already

  • what are the definitions of “one electron energy” and “two electron energy”?

This depends on the Hamiltonian you are using. Are you familiar with Hartree-Fock theory?

  • can alpha, beta, virtual orbital energies be easily translated to s,p,d, sigma, pi terminology?

Yes and No. You can break down the molecular orbitals (MOs) down to the atomic basis functions (s,p,d,f…) using population analysis. The chemical sigma and pi orbital do not come straightforwardly out of MO theory. You can localize orbitals and sometimes they look like chemical valence orbitals. You want to check out intrinsic bond orbitals (https://pubs.acs.org/doi/10.1021/ct400687b).

There are also tons of examples in the /samples of the psi4 installation directory of various calculations.

As to some of the other questions:

  • Excited states are still difficult for quantum chemistry, and I’m not as familiar with excited state methods as I’d like to be, but I can offer something. Could you talk a bit more about what you’d like to do? What system sizes do you want to look at? Are you working on a personal computer? How much do you care about quantitative precision? Depending on the answer, I’m either going to recommend EOM-CC or TD-DFT.
  • Specifying the spin isn’t always needed. Psi will “guess” the spin. I believe the guess is singlet for even numbers of electrons and doublets for odd numbers. As to why Psi doesn’t converge to the lowest spin state, you can make Psi converge to the lowest spin state with something called instability following. That part of Psi, however, needs a bit of work before I can recommend it.

Many thanks to @hokru and @jmisiewicz for your replies. I will certainly follow up on the suggested study material. I am generally interested in the electron energies / structures and orbital shapes (visualisation) of molecules consisting of quite small numbers of atoms eg H2O, O2, O3, N2O4, H2O2 and many others. Then moving on to the electron structures and vibration frequencies of the molecules I studied in the infrared (vibration normal modes / interatomic potential enegies) as a student (SiH3NCO, SiH3NCS, GeH3NCO, GeH3NCS). Also interested in the electronic structures of transition metal atoms eg 3d/4s relative energies.

When you say “electron structures,” do you mean ground state information or also low-lying excited states? The electronic states of transition metal atoms are a minefield.

I can offer a lot more advice on ground state problems than I can excited state.