SCF and Total energy in Solvation - confusion

Questions
#1

I am performing an RHF def2-SVP SMD solvation calculation on a system. The SCF energy printed at the end of the final (converged) iteration of the SCF is -23199.8621042032 However, the solvation results look like this:

(3) G-ENP(liq) elect-nuc-pol free energy of system -23199.862104203 a.u.
(4) G-CDS(liq) cavity-dispersion-solvent structure
free energy 61.1846 kcal/mol
(6) G-S(liq) free energy of system -23199.764600266 a.u.
SCF energy in the final basis set = -23199.7646002661
Total energy in the final basis set = -23199.7646002661

Why is the SCF energy in the final basis set at the end of the solvation calculation different from the SCF energy at the end of the SCF?

#2

It’s adding the solvation energy. See this sample output:

 ---------------------------------------
  Cycle       Energy         DIIS error
 ---------------------------------------
    1    -233.3252758105      8.19e-02
    2    -232.3682188055      8.89e-03
    3    -232.3486983124      1.04e-02
    4    -232.4492920863      1.24e-03
    5    -232.4502550623      5.61e-04
    6    -232.4505031841      9.76e-05
    7    -232.4505113047      4.67e-06  Convergence criterion met
 ---------------------------------------
 SCF time:   CPU 7.93s  wall 3.00s
(3)  G-ENP(liq) elect-nuc-pol free energy of system     -232.450511305 a.u.
(4)  G-CDS(liq) cavity-dispersion-solvent structure
     free energy                                               -2.5885 kcal/mol
(6)  G-S(liq) free energy of system                     -232.454636275 a.u.
 SCF   energy in the final basis set =     -232.4546362747
 Total energy in the final basis set =     -232.4546362747

Note that (3) is the same as the final SCF cycle, because SMD uses PCM electrostatics (or what SMx models call ENP) and that gets added into the SCF energy during the iterations. (4) is the nonelectrostatic contribution (which SMx calls CDS). (3) + (4) = (6), which is the same as what is reported as “SCF energy in the final basis set”. That “final basis set language” has a purpose in certain jobs, here it just means SCF energy including all the solvation corrections.

#3

Thanks John. Wouldn’t it be preferred to say “SCF energy with solvation correction included”?

After 6.1 is released I’ll take a look at making that change.

#4

The notation goes back to dual-basis methods, where “energy in the final basis” makes sense and I wouldn’t want to break that functionality. I think it’s reasonably obvious by matching up the SCF energies.

#5

I will deal with this in a late check-in for v. 6.1

#6

OK John. It’s appreciated. Thank you!