TDDFT state energies incorrect in the compact table?

[Sergi Ruiz-Barragan]

Dear users,

I'm doing several NAMD and I have noticed that the Pot. energy that appears in every QMD step information does not coincide with the energy value of the state in the table of TDDFT.

When your are doing a TDDFT calculation, the state information is printed in a compact table with, between other information the E(Diff), energy difference between ground state and the state. However, if you recalculate the the energy, it does not coincide with the information.

Using the option "print medium", a part of the compact table, there is specific information of every state. In this case, the energy difference is correct respect the QMD information.

Do I misunderstood anything?

[Sergi Ruiz-Barragan]

In particular:

Input for CO TDDFT:

start CO-EX-01
MEMORY total 24 gb
print medium
geometry units angstrom nocenter noautoz noautosym
 O 0       0        0
 C 0       0.00     1.15
end

basis  
 * library 6-31G**
end

dft
   xc B3LYP
  grid nodisk
end

tddft
  nroots 3
  notriplet
  cis
  civecs
end


task tddft

Output:

...

    3 smallest eigenvalue differences (eV)
--------------------------------------------------------
  No. Spin  Occ  Vir  Irrep   E(Occ)    E(Vir)   E(Diff)
--------------------------------------------------------
    1    1    7    8 a        -0.373    -0.027     9.418
    2    1    7    9 a        -0.373    -0.027     9.418
    3    1    6    8 a        -0.463    -0.027    11.888
--------------------------------------------------------

...

  ----------------------------------------------------------------------------
  Root   1 singlet a              0.315768069 a.u.                8.5925 eV
  ----------------------------------------------------------------------------
     Transition Moments    X -0.62284   Y -0.00282   Z -0.00000
     Transition Moments   XX  0.00000  XY  0.00000  XZ -2.30441
     Transition Moments   YY  0.00000  YZ -0.01044  ZZ -0.00000
     Dipole Oscillator Strength                    0.0816651393
     Electric Quadrupole                           0.0000008904
     Magnetic Dipole                               0.0000063814
     Total Oscillator Strength                     0.0816724111

     Occ.    4  a   ---  Virt.    8  a       -0.19342
     Occ.    7  a   ---  Virt.    8  a        0.97778

--------------------------

root1 have two different values for ediff: 9.418 and 8.5925eV

El dia dilluns, 27 de maig del 2024 a les 22:25:52 UTC+2, Sergi Ruiz-Barragan va escriure:

[Edoardo Aprà]

9.418 eV is the energy difference from eigenvalues (resulting from the ground state calculation)

8.5925 eV is the result of the Time-Dependent DFT calculation, a more accurate way of computing the excitation.

[Sergi Ruiz-Barragan]

Dear Edoardo,

Thanks for the information.

That my fault, I supposed that a list that appears in the TDDFT module must be the solution, but it is the "initial guess".

Thank you.

El dia dissabte, 1 de juny del 2024 a les 0:44:49 UTC+2, Edoardo Aprà va escriure: