3.4 eV specific to this class of molecules representative of highly luminescent
transition metal complexes.
From the results reported above for the reference molecule, namely the complex
fac-[Ir (ppy) 3 ] 1, we can conclude that the chosen computational strategy is
reasonable, leading to realistic structural and spectroscopic properties. The quality
of the PW91 based on generalized gradient approximation (GGA) functional must
be pointed out for this class of compact nearly spherical molecules with bulky
ligands. This surprisingly good agreement between the TD-DFT/PW91 absorption
spectrum in vacuum and the experimental one could also be due of compensation of
errors. Solvent corrections could induce red shift whatever the functional is, and
could improve the TD-DFT/B3LYP spectrum as well [114]. However, the presence
of LLCT states in the lowest part of the TD-DFT/B3LYP theoretical spectrum
indicates a particular problem of charge transfer description in this class of molecules with the hybrid functional. This failure of the B3LYP functional for this class
of molecules is not surprising and has already been observed for other Ir(III)
complexes with phenylisoquinoline phenylpyridine ligands [81].
When taking into account SOC, the density of states does increase drastically by
the splitting of the triplet states. The A and E “spin-orbit” states in C 3 point group
reported in Table 2 are generated by the “spin-free”
1,3 A and
1,3 E states according to
the zero-field splitting. Each
3 A state is split into A + E, each
3 E state into E + E + A.
1 A remains A and
1 E remains E in the so-called double group representation.
As already observed in other theoretical studies [75–79] the singlet/triplet
mixing induces a decrease of the intensities and a red shift (1,600 cm
À1 ) of the
visible part of the spectrum because triplet states gain weak oscillator strengths
(<10
À3 ). Whereas a few states remain nearly pure singlet or triplet, most of the
“spin-orbit” excited states reported in Table 2 present mixed character. Another
consequence of the SOC effects on the absorption spectrum of fac-[Ir (ppy) 3 ] 1a is
an increase of the MLCT/LC mixing in the lowest part of the spectrum.
Whereas the “spin-orbit” states A and E calculated between 19,120 cm
À1 and
21,760 cm
À1 remain essentially MLCT, the
1 E (S 7 , f ¼ 0.041), for instance, gains
32% of LC contribution by coupling with the
3 E (T 7 ) leading to the E6 “spin-orbit”
state calculated at 22,720 cm
À1 of decreasing intensity ( f ¼ 0.0017). The visible
band calculated between 17,000 and 25,000 cm
À1 (Fig. 1) is enlarged and decreases
in intensity compared to the one calculated in the “spin-free” spectrum. This band is
composed essentially of MLCT states (E1–E7, A1, A2). Whereas the global
energetics and the shape of the absorption spectrum of fac-[Ir (ppy) 3 ] 1 is only
slightly modified by SOC effects, the character of the transitions is affected. The
main consequence is an increase of mixed LC/MLCT character of the excited states
above 30,000 cm
À1 . The two intense LC peaks S 24 ( f ¼ 0.106) and S 44 ( f ¼ 0.109)
(Table 1) decrease in intensity, either by coupling with MLCT states or by coupling
with triplet states. The S 44 state ( f ¼ 0.109) is not affected energetically, remains
LC in character, but decreases in intensity by coupling with several triplet states,
leading to the “spin-orbit” state A28 ( f ¼ 0.080).
Absorption Spectroscopy, Emissive Properties, and Ultrafast Intersystem. . .
389
transition metal complexes.
From the results reported above for the reference molecule, namely the complex
fac-[Ir (ppy) 3 ] 1, we can conclude that the chosen computational strategy is
reasonable, leading to realistic structural and spectroscopic properties. The quality
of the PW91 based on generalized gradient approximation (GGA) functional must
be pointed out for this class of compact nearly spherical molecules with bulky
ligands. This surprisingly good agreement between the TD-DFT/PW91 absorption
spectrum in vacuum and the experimental one could also be due of compensation of
errors. Solvent corrections could induce red shift whatever the functional is, and
could improve the TD-DFT/B3LYP spectrum as well [114]. However, the presence
of LLCT states in the lowest part of the TD-DFT/B3LYP theoretical spectrum
indicates a particular problem of charge transfer description in this class of molecules with the hybrid functional. This failure of the B3LYP functional for this class
of molecules is not surprising and has already been observed for other Ir(III)
complexes with phenylisoquinoline phenylpyridine ligands [81].
When taking into account SOC, the density of states does increase drastically by
the splitting of the triplet states. The A and E “spin-orbit” states in C 3 point group
reported in Table 2 are generated by the “spin-free”
1,3 A and
1,3 E states according to
the zero-field splitting. Each
3 A state is split into A + E, each
3 E state into E + E + A.
1 A remains A and
1 E remains E in the so-called double group representation.
As already observed in other theoretical studies [75–79] the singlet/triplet
mixing induces a decrease of the intensities and a red shift (1,600 cm
À1 ) of the
visible part of the spectrum because triplet states gain weak oscillator strengths
(<10
À3 ). Whereas a few states remain nearly pure singlet or triplet, most of the
“spin-orbit” excited states reported in Table 2 present mixed character. Another
consequence of the SOC effects on the absorption spectrum of fac-[Ir (ppy) 3 ] 1a is
an increase of the MLCT/LC mixing in the lowest part of the spectrum.
Whereas the “spin-orbit” states A and E calculated between 19,120 cm
À1 and
21,760 cm
À1 remain essentially MLCT, the
1 E (S 7 , f ¼ 0.041), for instance, gains
32% of LC contribution by coupling with the
3 E (T 7 ) leading to the E6 “spin-orbit”
state calculated at 22,720 cm
À1 of decreasing intensity ( f ¼ 0.0017). The visible
band calculated between 17,000 and 25,000 cm
À1 (Fig. 1) is enlarged and decreases
in intensity compared to the one calculated in the “spin-free” spectrum. This band is
composed essentially of MLCT states (E1–E7, A1, A2). Whereas the global
energetics and the shape of the absorption spectrum of fac-[Ir (ppy) 3 ] 1 is only
slightly modified by SOC effects, the character of the transitions is affected. The
main consequence is an increase of mixed LC/MLCT character of the excited states
above 30,000 cm
À1 . The two intense LC peaks S 24 ( f ¼ 0.106) and S 44 ( f ¼ 0.109)
(Table 1) decrease in intensity, either by coupling with MLCT states or by coupling
with triplet states. The S 44 state ( f ¼ 0.109) is not affected energetically, remains
LC in character, but decreases in intensity by coupling with several triplet states,
leading to the “spin-orbit” state A28 ( f ¼ 0.080).
Absorption Spectroscopy, Emissive Properties, and Ultrafast Intersystem. . .
389