TDDFT Excited state geometry relaxation goes up in energy.


Click here for full thread
Clicked A Few Times
Dear developers,

I am having difficulty understanding a calculation I am running in order to get zero phonon line of MgP (Mg-Porphyrin)

MgP has two degenerate roots, which split if you break the symmetry of the ring. There next root is well seperated by around 1.2 eV

The vertical excitation (Ground state energy + Root1 = -1187.669221804102 +0.089440466) is at -1187.5797813 Ha

When I relax the system in excited state, the energy starts at -1187.5797813 Ha, as expected, but goes UP to -1187.45264016 Ha. The forces in the meantime go down.

As far as I know, it is impossible to have ZPL higher in energy than a vertical excitation, at most they can be at the same energy.

Am I interpreting the values printed in output incorrectly?

How does a geometry relaxation can go up in energy?

Are the values given in the output somehow contain other terms, i.e. the constraints etc. ?

Please find the input file attached.
All the best,
Baris


START MgP-medium-lrtddft-hse06-rlxroot1
TITLE "MgP-medium-lrtddft HSE06"
memory total 3000 mb
GEOMETRY "HSE06" noautoz noautoz nocenter noautosym
C         7.589163       12.423772        5.000000
C         8.911513       12.862607        5.000000
C         7.142217       11.104027        5.000000
C        11.104127       12.856943        5.000000
C         7.136107        8.911806        5.000000
C        12.424227       12.410662        5.000000
C         7.576135        7.589836        5.000000
C         9.332355       14.237513        5.000000
C         5.764970       10.690630        5.000000
C        12.864239       11.088390        5.000000
C         8.896154        7.143870        5.000000
C        10.690051       14.234182        5.000000
C         5.761182        9.332844        5.000000
C        12.411200        7.575905        5.000000
C        12.858283        8.895815        5.000000
C        11.088760        7.136906        5.000000
C        14.239210       10.666937        5.000000
C         9.308957        5.766676        5.000000
C        14.235534        9.309599        5.000000
C        10.666699        5.762197        5.000000
N        10.005638       12.049970        5.000000
N         7.948937       10.005645        5.000000
N        12.051413        9.994363        5.000000
N         9.994979        7.950062        5.000000
H         6.823282       13.194398        5.000000
H        13.194434       13.177073        5.000000
H         6.806329        6.823143        5.000000
H        13.176954        6.805048        5.000000
H         8.671200       15.095293        5.000000
H         4.910102       11.355453        5.000000
H        11.354188       15.089661        5.000000
H         4.902821        8.672514        5.000000
H        15.098017       11.327250        5.000000
H         8.642878        4.912827        5.000000
H        15.090827        8.644769        5.000000
H        11.326312        4.903302        5.000000
Mg        9.999605        9.999319        5.000000
symmetry C1
END
BASIS "very large" SPHERICAL
* library aug-cc-pvtz
END
BASIS "large" SPHERICAL
* library maug-cc-pvdz
END
BASIS "large2" SPHERICAL
* library aug-cc-pvdz
END
BASIS "medium" SPHERICAL
* library 6-31+g*
END
BASIS "small" SPHERICAL
* library sto-3g
END

echo



###################Production run
set "ao basis" "medium"
set geometry "HSE06"
SCF
maxiter 200
semidirect filesize 0
vectors output MgP-medium-lrtddft-hse06-hf.movecs
END

DFT
direct
iterations 100
decomp
XC xpbe96 1.0 xcampbe96 -0.25 cpbe96 1.0 srhfexch 0.25
cam 0.11 cam_alpha 0.0 cam_beta 1.0
vectors input  MgP-medium-lrtddft-hse06-hf.movecs
#vectors input MgP-medium-lrtddft-hse06.movecs
vectors output MgP-medium-lrtddft-hse06.movecs
END
TDDFT
RPA
SINGLET
NOTRIPLET
NROOTS 10
TARGET 1
CIVECS
GRAD
  ROOT 1
END
FREEZE ATOMIC
END
driver
 maxiter 200
 XYZ MgP-medium-lrtddft-hse06-rlxroot1
 TRUST 0.1
 SADSTP 0.05
end
TASK SCF ENERGY
TASK DFT ENERGY
task tddft optimize
-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=-=
Dr. O. Baris Malcioglu,
University of Liege,
Bât. B5 Physique de la matière condensée
allée du 6 Août 17
4000 Liège 1
Belgique