Angewandte
Chemie
data, see Table S1 in the Supporting Information). Firstly,
a comparison between 1 and 2 shows the importance of the
electron-donating properties of the diphenylamino group in 1.
While the absorption spectra of the anisyl derivative 2 showed
subtle solvatochromism (labs = 359–367 nm), the emission
spectra showed a bathochromic shift with increasing solvent
for the radiative (k ) and nonradiative (k ) decay were
r nr
determined based on fluorescence quantum yield and lifetime
values (t). While the k values in DMSO are comparable for 1,
r
8
À1
4, and 5 (0.89–1.1 ꢁ 10 s ), the k values of 4 and 5 (52 ꢁ
nr
8
À1
8
À1
10 s and 15 ꢁ 10 s , respectively) are much higher than
8
À1
that of 1 (0.60 ꢁ 10 s ). These results indicate that even
though the presence of a strong electron-accepting ring
skeleton is crucial for the bathochromically shifted emission,
excessive electron-accepting properties may accelerate the
nonradiative decay process from the excited singlet state.
Although the photophysical properties of 1 and 3 are
almost comparable, their solubility in polar solvents differs
greatly. The solubility of phosphole oxide 1 in DMSO
polarity. However, the shift for 2 from toluene to ethanol
À1
(
1310 cm ) was measured to be significantly smaller than
À1
that in 1 (2080 cm ). Based on TD-DFT calculations at the
PBE0/6-31 + G(d) level of theory, the HOMO of 1 is mostly
localized on the electron-donating triphenylamine moiety,
whereas that of 2 is delocalized over the 2-anisylbenzophosp-
hole moiety. Accordingly, the character of S –S transitions in
0
1
À1
1
exhibits the intramolecular charge transfer (ICT) character
(17.4 gL ) is significantly higher than that of phosphole
sulfide 3 (0.8 gL ). This difference most probably results
À1
from the triphenylamine moiety to the electron-accepting
benzophosphole oxide moiety, whereas that of 2 is attribut-
able to p–p* transition. The calculated values for the
oscillator strengths are 0.396 and 0.265 in 1 and 2, respectively.
The higher value for 1 is consistent with a larger molar
from the more polarized P=O moiety relative to the P=S
moiety. Electron densities on the P, O, and S atoms in 1 and 3
[16]
were calculated by natural population analysis, using the
B3LYP hybrid functional in combination with the 6-311G(d)
basis set. For the P and O atoms in 1, electron density values
of 1.88 and À1.06, respectively, were calculated, whereas in 3,
values of 1.32 and À0.58 were obtained for the P and S atoms,
respectively.
4
À1
À1
extinction coefficient for 1 (e = 1.87 ꢁ 10 m cm ) compared
4
À1
À1
to 2 (e = 1.03 ꢁ 10 m cm ), which should, at least in part,
account for the higher quantum yield of 1 relative to 2.
Secondly, a comparison of 1 with the phosphole sulfide 3
and phosphonium salt 4 illustrates the impact of the electron-
accepting character of the benzophosphole moiety. In terms
of the LUMO levels, the parent phosphole rings with P=O or
P=S moieties should possess a comparable electron-accepting
An advantageous feature of 1 for the application as
a fluorescent bioimaging probe is the drastic color change of
the fluorescence emission as a function of the solvent polarity,
that is, the color change from a bluish green in toluene to
a reddish orange in DMSO (Figure 2a). Notably, the Lippert–
character (see Figure S10 in the Supporting Information).
Commensurate with this similarity, 1 and 3 exhibit compara-
ble photophysical properties. For example, 3 shows high
[6]
Mataga plot for 1 in various solvents, including the protic
2
solvents ethanol and methanol, showed high linearity (R =
fluorescence quantum yields both in polar (DMSO: F =
0.93; Figure 2b). This linearity suggests that the specific
F
0
.61) and in protic (EtOH: F = 0.76) solvents, and the
F
maximum fluorescence emission wavelengths are comparable
to those of 1. In contrast, 4 exhibits, even in toluene, a 33 nm
bathochromically shifted absorption maximum, as well as
a 144 nm red-shifted fluorescence maximum relative to that
of 1 (see Figure S4 in the Supporting Information and
Table 1). These large differences are obviously due to the
much stronger electron-accepting character of the phospho-
nium moiety. Irrespective of the solvent polarity, 4 exhibits
substantially lower F values associated with these red-
F
shifted emissions.
Thirdly, a comparison with the thiophene dioxide con-
gener 5 provides insight into the crucial role of the P=O
moiety with respect to the strong fluorescence. The parent
thiophene dioxide has a lower LUMO level compared to that
of phosphole oxide (Figure S10). Whereas 5 exhibits, in
nonpolar solvents, a maximum emission wavelength at
slightly shorter wavelengths compared to 1, its maximum
emission wavelength in DMSO is bathochromically shifted by
3
5 nm (see Figure S5 in the Supporting Information). An even
more significant difference is observed for the fluorescence
quantum yield of 5 in polar or protic solvents, which is much
lower than that of 1 (Table 1).
These comparisons clearly demonstrate that the use of
a sufficiently strong electron-accepting ring is crucial to
preserving intense fluorescence in polar or protic solvents.
The fluorescence intensity in such solvents should largely rely
on the dynamics of the excited state. Therefore, rate constants
Figure 2. Solvent dependence of the fluorescence of 1: a) Photos of
the color change of the fluorescence emission of 1 in various solvents.
b) A Lippert–Mataga plot for 1 in various solvents (R =0.93), whereby
2
Dn and Df represent the Stokes shift (nabsÀn ) and the orientation
em
2
2
polarizability of solvents (eÀ1)/(2e+1)À(n À1)/(2n +1), respectively.
e=dielectric constant, n=refractive index.
Angew. Chem. Int. Ed. 2015, 54, 1 – 6
ꢀ 2015 Wiley-VCH Verlag GmbH & Co. KGaA, Weinheim
3
These are not the final page numbers!