guanine was attributed to an enhanced nonadiabatic coupling promoted by out-ofplane vibrations. Later, SH/TD-DFTB [133] predicted an excited-state lifetime of
11 ps for adenine gas, ten times longer than the experimental result. The deviation
was then attributed to the distribution of initial energies. More recently, systematic
investigations of adenine gas with SH/TDDFT with several functionals once more
led to very long excited-state lifetimes [35, 148]. These results are conveniently
plotted in Fig. 6 in terms of the S 0 population 1 ps after photoexcitation. Even the
most optimistic SH/TDDFT simulation at the limit of the error bar is inferior to
40%, while the experimental result reaches 68%. The root of the problem seems to
be connected to an overstabilization of the ground state energy along planar
distortions in comparison to nonplanar distortions. This unbalanced ground state
profile leads to a wrong description of the excited-state minimum [148].
0 fs
4 fs
15 fs
8 fs
Fig. 5 Evolution of the
(S 1 ÀS 0 ) electron density
difference during a single
surface-hopping trajectory
of N-methylformamide
dimer computed with
TDDFT with LC-BLYP
(μ ¼ 0.2 a 0
À1
). Green
(orange) indicates electron
acceptor (donor) regions
(Based on data from [127])
E x p e r i m
e n t a l
T D D F T - P B E
T D D F T - w B 9 7 X D
T D D F T - B 3 L Y P
T D D F T B - B 3 L Y P
T D D F T - P B E 0
T D D F T - C A M
- B 3 L Y P
T D D F T - B H L Y P
T D D F T - M
0 6 - H F
0
20
40
60
80
Ground State Population at 1 ps (%)
Fig. 6 Ground-state
population of adenine gas
1 ps after photoexcitation
according to the experiment
[182] and surface hopping
based on diverse
LR-TDDFT simulations.
TDDFT data from [35,
148]. TDDFTB from [133]
(supposing simple
exponential decay).
Sampling error of the
simulations for 90%
confidence level
Surface Hopping Dynamics with DFT Excited States
437
11 ps for adenine gas, ten times longer than the experimental result. The deviation
was then attributed to the distribution of initial energies. More recently, systematic
investigations of adenine gas with SH/TDDFT with several functionals once more
led to very long excited-state lifetimes [35, 148]. These results are conveniently
plotted in Fig. 6 in terms of the S 0 population 1 ps after photoexcitation. Even the
most optimistic SH/TDDFT simulation at the limit of the error bar is inferior to
40%, while the experimental result reaches 68%. The root of the problem seems to
be connected to an overstabilization of the ground state energy along planar
distortions in comparison to nonplanar distortions. This unbalanced ground state
profile leads to a wrong description of the excited-state minimum [148].
0 fs
4 fs
15 fs
8 fs
Fig. 5 Evolution of the
(S 1 ÀS 0 ) electron density
difference during a single
surface-hopping trajectory
of N-methylformamide
dimer computed with
TDDFT with LC-BLYP
(μ ¼ 0.2 a 0
À1
). Green
(orange) indicates electron
acceptor (donor) regions
(Based on data from [127])
E x p e r i m
e n t a l
T D D F T - P B E
T D D F T - w B 9 7 X D
T D D F T - B 3 L Y P
T D D F T B - B 3 L Y P
T D D F T - P B E 0
T D D F T - C A M
- B 3 L Y P
T D D F T - B H L Y P
T D D F T - M
0 6 - H F
0
20
40
60
80
Ground State Population at 1 ps (%)
Fig. 6 Ground-state
population of adenine gas
1 ps after photoexcitation
according to the experiment
[182] and surface hopping
based on diverse
LR-TDDFT simulations.
TDDFT data from [35,
148]. TDDFTB from [133]
(supposing simple
exponential decay).
Sampling error of the
simulations for 90%
confidence level
Surface Hopping Dynamics with DFT Excited States
437