1742
Cao CZ, et al. Sci China Chem November (2011) Vol.54 No.11
are different in different kinds of solvents for a given solute.
The experimental test results showed that even if the same
solute molecule, the UV absorption energy also changes in
different solvents. For example, the difference of absorption
wavenumbers of ClSBNMe2 is 536 cm–1 (about 6.9 nm) in
the two solvents C6H12 and CHCl3. This means that the sol-
vent effect is very obvious, and can not be neglected. Thus,
if we investigate the changing law of the UV absorption
energy of XSBY in different kinds of solvents, the solvent
effect must be taken into account for. At present, many sol-
vent effect parameters have been proposed, in which the
solvatochromic dye ET(30) [19] attracted the most attention.
In eq. (4), a, b, c, and d are coefficients, V, EHOMO, and
ELUMO are the volume V, the HOMO energy and LUMO
energy of the solute molecule. We suspect that the XSBY
molecules in a solvent interact with the solvent molecules,
that is the HOMO of the XSBY molecule interact with the
LUMO of the solvent molecule, and the LUMO of the
XSBY molecule interact with the HOMO of the solvent
molecule. Thus eq. (5) can be employed to express this sol-
vent effect (Ssolvent),
Ssolvent a(V
Vsolute ) b（EHOMO，solvent ELUMO，solute）
solvent
c（ELUMO，solvent EHOMO，solute
）
(5)
aV bEHOMO，solvent cELUMO，solvent
solvent
In this paper, we use the ET(30) to correlate the 242 experi-
[aVsolute bELUMO，solute cEHOMO，solute
]
ex
CC
mental
in Table 1 combining with parameters
max
As respect to a given XSBY molecule, its Vsolute, EHOMO, solute
,
and XY. See from Table 3, eq. (2) has a good correlation.
What we want to know is that whether the regression re-
sults can be improved, if the ET(30) in eq. (2) was replaced
by other parameters. The authors [27] have proposed that
the aqueous solubility of non-proton compounds is related
to their molecular volume V, HOMO energy and LUMO
energy, and the HOMO energy and LUMO energy of the
water molecule. Because the HOMO energy and LUMO
energy of the water molecule are invariable, the aqueous
solubility of the solute molecule can be quantified with its
V, HOMO energy and LUMO energy. Basing on the results
of ref. [27], we speculate that the interaction between
XSBY molecule and solvent molecule is similar to the in-
teraction between solute molecule and water molecule. That
is to say, the interaction of solute molecule and solvent
molecule may be related with their HOMO and LUMO. If
this hypothesis is correct, and because the HOMO energy
and LUMO energy are constants for a given XSBY mole-
cule, the ET(30) in eq. (2) can be substituted by the aqueous
solubility parameters logP (n-octanol/water partition coeffi-
cient). Consequently, the eq. (2) is replaced by eq. (3) (see
Table 3). Compared with eq. (2), the correlation of eq. (3) is
much better, and its standard error decreases about 15 cm–1.
The average absolute error between the experimental λmax and
the calculated values of eq. (2) is 1.1 nm, and that of eq. (3) is
only 1.0 nm. Seen from Table 4, the numbers of absolute
errors of eq. (2) and eq. (3) within 2.0 nm are 199 and 212,
and account for 82.24% and 87.61% of the total samples,
respectively. The plots of the calculated wavenumbers of eq.
(3) against the experimental values are shown in Figure 2.
The variation of νmax values in Table 1 is in a wide range.
The maximum is 32592 cm–1 (the νmax of HSBH in t-BuOH),
and the minimum is 27692 cm–1 (the νmax of ClSBNMe2 in
CHCl3). Their variation is 4900 cm–1 (54.3 nm). In such
wide range of wavenumbers, eq. (3) has an excellent corre-
lation. It indicated that the solvent effect on the UV absorp-
tion energy of XSBY can be scaled by logP. This result
may be explained as follows, the authors [25] have pro-
posed the logP of the solute can be expressed as eq. (4),
and ELUMO, solute are constants. Therefore, the last item of eq.
(5) can be replaced by a coefficient d. So eq. (5) can be
modified as eq. (4). In addition, it is generally believed that
the molecular volume is related to its polarizability. There-
fore, the bigger the molecular volume is, the larger the po-
larizability effect is, and the larger the stabilization effect on
the charge is. The logP can be used to scale the solvent ef-
fect on the UV absorption energy of XSBY, because it ex-
pressed the electronic effect between XSBY molecules and
solvent molecules. This electronic effect includes the inter-
action of their frontier orbitals, and the polarizability effect
of solvent molecule on the solute molecule. If the solute
molecule is variable, the last item of eq. (5) (the contents in
the square brackets is no longer a constant. Thus, the molec-
ular properties (such as substituent effect) of XSBY must be
considered.
Eq. (3) expressed the changing law of the energy of UV
absorption max wavelengths for a series of XSBY molecule
in different kinds of solvents, and can be used to calculate
and predict the energy of UV absorption max wavelengths of
this kind of compounds. For instance, Jiang et al. [28] has
synthesized compounds of Table 5 and measured their energy
of UV absorption max wavelengths λmax in 95% ethanol. If
the logP of 95% ethanol is approximated as that of pure eth-
anol, the λmax of these compounds can be predicted with eq.
(3). The predicated λmax are in agreement with the experi-
mental values (see Table 5). Here, it should be pointed out
that the logP of pure ethanol is not equal to that of 95% eth-
anol, so the result listed in Table 5 is only an approximation.
log P = aV +bEHOMO + cELUMO + d
(4)
Figure 2 Plot of νmax, expt versus νmax, calcd for XSBY (cm–1).
Cao, Chenzhong
