Journal of Molecular Modeling (2019) 25: 60

https://doi.org/10.1007/s00894-019-3948-0

ORIGINAL PAPER

State-specific electrostatic potential descriptors for estimating

solvatochromic effects
Lulu Wang 1 & Shiwei Yin 1

Received: 10 December 2018 / Accepted: 29 January 2019 / Published online: 12 February 2019
# Springer-Verlag GmbH Germany, part of Springer Nature 2019

Abstract
The Lippert-Mataga equation is used widely to describe the solvatochromic effects of fluorescent molecules through the eval-
uation of solute–solvent interactions on the basis of the point-dipole model. A large dipole deviation of molecules in the ground-
state and the lowest excited-state is a basic requirement for the design of a polarity-sensitive fluorescent probe. Some recently
synthesized probes with center-symmetry have near zero dipole deviation while undergoing notably solvatochromic behaviors.
Thus, it is necessary to find a new method beyond the Lippert-Mataga model to qualitatively estimate the molecular solvent
shifts. To this end, a state-specific descriptor (SSD) based on molecular surface electrostatic potentials (ESP) is proposed to
explain the solvatochromic behaviors of the well-studied coumarin C153 and center-symmetric DCB-1d. In contrast to the
experimental solvent shifts and state-specific TD-DFT calculations, the SSD successfully explains the solvatochromic effect of
C153 and DCB-1d molecules. In addition, the SSD was tested by using eight selected polarity-sensitive fluorescent molecules.
The SSD was found to provide a good linear relationship with solvatochromism.

Keywords Solvatochromism . State-specific descriptor . Non-equilibrium implicit solvation model . Polarity-sensitive probe

Introduction solvatochromic dyes usually contain donor and acceptor

groups. When stabilization of the excited-state probe is larger
Molecules used as fluorescent probes are usually required to than that of the ground state, the probe leads to red-shift
change their emission characters, e.g., polarity, viscosity, and solvatochromism. Vice versa, more stabilization of the ground
molecular stacking, in response to their environment. The state relative to the first excited state results in blue-shift
changed emission features not only give a wealth of informa- solvatachromism. Under a weak interaction approximation,
tion in situ about interactions between probes and their envi- the stabilization energy is also termed solvent energy, and is
ronment, but also provide insights at a molecular level with determined mainly by the electrostatic interaction between the
high spatial and temporal resolution for applications in bio- probe and its environmental solvent molecules. Thus, evalua-
logical, material and environmental areas [1, 2]. Among re- tion of the electrostatic interactions between molecules in var-
sponse types, including polarity, viscosity, ions, pH and spe- ious electronic states and solvent plays an important role in
cific reactions, there are fluorogenic and chromogenic probes, designing new solvatochromic probes. In principle, these
which change their fluorescence intensity and color, respec- solute–solvent interactions need be solved explicitly using a
tively. Among them, polarity-sensitive probes known as polarizable molecular force field [3, 4]. To deal with the
nearest solvent hydrogen bonding or other strong interactions,
explicit QM super molecule or QM/MM methods would be
Electronic supplementary material The online version of this article
(https://doi.org/10.1007/s00894-019-3948-0) contains supplementary needed [5], which is beyond the scope of this paper. These
material, which is available to authorized users. explicit methods on the basis of force field require statistical
averaging over various configurations of solute and solvent
* Shiwei Yin molecules [6, 7]. Furthermore, developing accurate state-
[email protected] specific intermolecular potential is necessary to correctly esti-
1
mate solvent shift. In practice, solvent reaction-field methods
Key Laboratory for Macromolecular Science of Shaanxi Province,
are used to implicitly evaluate solute–solvent interactions [8,
School of Chemistry and Chemical Engineering, Shaanxi Normal
University, Xi’an 710062, Shaanxi, China 9]. When the solvent–solute interactions are approximated by
60 Page 2 of 8 J Mol Model (2019) 25: 60

    2μ
point-dipole interactions, the electronic transition energy in
emi ¼ E emi − μe −μg Re ¼ E emi − μe −μg ð4Þ
e *
E sol vac vac
f
solvent (Esol) can be solved by the first-order perturbation a30
approximation showed in Fig. 1,

In practice, the polarity-sensitive behavior of solute is eval-
E sol ¼ Evac − ΔGe −ΔGg ð1Þ
uated by solvent shift (ΔEshift) when transferring a molecule
Evac is electronic transition energy in vacuum, ΔG is the from an apolar solvent into a strong polar solvent, which is
stabilization energy (also named as the solvation energy), also called the Lippert-Mataga equation.
which represents Gibbs free energy during the transfer of sol-   2μ
g=e
ute from vacuum to solvent, and the subscripts e and g mean ΔE shift
abs=emi ¼ hΔν abs=emi ¼ μe −μg Δf * ð5Þ
a30
the stabilization energies for solute in the excited and ground
state, respectively. The solvation energy is solved approxi- Here, Δf* is the deviation of polarity parameters be-
mately by the interactions between the reaction field (R) with tween nonpolar and strongly polar solvents. According to
the solute dipole moment (μ) and the magnitude of reaction Eq. 5, the large difference between μe and μg has long
field, which is determined by the Onsager model [10]: been considered a necessary requirement in order to de-
  sign polarity-sensitive fluorescent molecules. Very recent-
2μ 2μ ε−1 1 n2 −1
R ¼ 3 f* ¼ 3 − ð2Þ ly, Zhao and co-workers synthesized a series of p-bis(2,2-
a0 a0 2ε þ 1 2 2n2 þ 1
dicyanovinly)benzene (DCB)-based fluorophores, which
Here, a0 and μ are the radius of the reaction cavity and exhibit interesting red-shift solvatochromism features
solute dipole moment, respectively. f* is the solvent polarity [12]. However, the DCB shown in Fig. 2 actually has a
parameter, which can be determined by the solvent’s isotropic center-symmetric structure, which results in a zero dipole
dielectric constant (ε) and optical refractive index (n). Under moment in both the ground and excited states. According
the assumption of the Franck-Condon principle, the R of sol- to Lippert-Mataga expressions (see Eqs. 3–5), DCB-based
ute during absorption processes of electronic spectra will be a compounds should not have solvent shift fluorescence be-
fixed value, which is determined by the ground-state dipole cause of their zero-dipole moment. The point-dipole ap-
moment (μg) and the polarity of the solvent. proximation gives an incorrect picture from which to es-
timate solute–solvent interactions for zero-dipole mole-
Then, the absorption transition energy (Esolabs ) in solvent is
cules, which prevents us from finding new polarity-
calculated by using the Lippert-Mataga expression [9, 11].
sensitive fluorescent molecules. It is necessary to find a
    2μ more accurate descriptor to represent solute–solvent inter-
g *
abs ¼ E abs − μe −μg Rg ¼ E abs − μe −μg
E sol ð3Þ
vac vac
f
a30 actions at the molecular level. Besides the point-dipole
model, Politzer and coworkers also defined a set of mo-
Here Evac
abs is the transition energy of the solute absorption lecular descriptors based on electrostatic potential (ESP)
spectrum in vacuum. Similarly, the reaction field is deter- on the van der Waals (vdW) surface to describe intermo-
mined by the lowest excited-state dipole moment (μe) of the lecular interactions. They further used them to calculate
solute molecule for the emission processes, and the emission many related condensed-phase properties [13–15]. These
energy (E sol
emi ) is evaluated by Lippert-Mataga expression. ESP-based descriptors take molecular shape into account
and describe noncovalent interactions well [16, 17].
Moreover, in order to avoid underestimating intermolecu-
lar interactions for non-dipole molecules, they developed
the average deviation of the surface ESP (Π) to represent
molecular local polar or charge separation [18]. Once the
solute–solvent interaction is described by the descriptor
Π, the solvent shift can then be evaluated by the differ-
ence in average deviations for ground (Πg) and excited
(Πe) state surface ESP. The larger value of Πe–Πg leads
to more notable solvatochromism. Inspired by Politzer
ESP descriptors, here we combine elements of classical
solvent reaction field theory to develop a new state-
specific ESP-based descriptor to implicitly represent sol-
vent shifts on the absorption and emission spectra. In this
Fig. 1 Illustration of the solvent-shift from vacuum to solvent based on
paper, we take the fluorescent probe coumarin C153 [19]
ground-state and excited-state solvation energies and center-symmetric DCB-1d [12] shown in Fig. 2 as the
J Mol Model (2019) 25: 60 Page 3 of 8 60

Fig. 2 Atomic structures of

coumarin C153 (left) and center-
symmetric DCB-1d (right).
Atoms: Grey C, red O, white H,
blue N, baby blue F

prototype molecules, and study their polarity-dependent excited states during the absorption and emission process-
absorption and emission spectra. We tested the new es. As shown in Fig. 1, the large solvation energy of the
ESP-related descriptors by evaluating solvent shifts in solute is an important parameter in evaluating polarity-
electronic spectra. As a comparison, time-dependent den- sensitive fluorescent behaviors. Based on solvent reaction
sity functional theory (TD-DFT) coupled with the polar- field theory, the solute–solvent interaction is described by
izable continuum model (PCM) were applied to study the summation of the surface ESPs times their corre-
absorption and emission spectra. sponding reaction cavity response charges. The response
charge is determined using the self-consistent reaction
field (SCRF), which is positively correlated to its ESP
and solvent polarity parameter [21, 22].
Methods  
ε−1 1 n2 −1
qi ∝V i S i f * ¼ V i S i − ð6Þ
In the Lippert-Mataga model, the dipole moment of the 2ε þ 1 2 2n2 þ 1
solute is determined predominantly by intermolecular in-
teractions, where the molecule is approximately taken as a Where qi is the response charge on the ith facet of the vdW
point-dipole in a spherical cavity. If the molecular dipole surface, Vi is the ESP value on the center of the ith facet, and Si
is zero, the solute–solvent interaction is still zero even is the area of the facet. In practice, the vdW surface is defined
though there are some local dipole (DA) structures in as the isosurface of 0.001 electron density.
DCB-1d. To better describe the intermolecular interac- Similar to the aforementioned Lippert-Mataga derivation,
tions of the DCB-1d type, high-ordered multipoles, such the solvent shift of the absorption and emission processes
as quadrupoles, need to be considered [20]. Additionally, during the transfer of solute from nonpolar solvent to polar
the ESP on the molecular vdW surface, which naturally solvent is evaluated by
take the molecular shape into account is also a very good  
g=e
g=e g=e
method to describe noncovalent interactions. To overcome hΔν abs=emi ≅− ∑ V ei ⋅qi −V gi qi ∝− ∑ V ei −V gi V i S i Δ f *
i i
the failure of point-dipole approximation in the descrip-
tion of the solute–solvent interactions for the D–A type of ð7Þ
molecules with a central symmetric structure, we propose
a new state specific descriptor (SSD) to represent the dif- Here, the superscripts g and e are related to the ground-state
ferent solvation energies of the solute in its ground and and the excited-state, respectively. qg and qe represent for the
ground-state and excited-state response charges. Ve and Vg is
the ESP value produced by ground-state and excited-state
electron density, respectively. The facet on the vdW surface
is defined by using the improved matching tetrahedral algo-
rithm [23]. Therefore, the ESP SSD is defnied as

g=e
SSDg=e ¼ ∑ V ei −V gi V i S i Δf * ð8Þ
i

SSDg and SSDe represent the capability of solvatochromism

Fig. 3 Electron/hole wavefunctions for C153 (left) and DCB-1d (right)
during the electron transtion from S0 to S1. Blue Hole, green electron of the molecular absorption and emission spectra, respectively.
distribution The larger SSD value suggests a higher polarity-sensitive
60 Page 4 of 8 J Mol Model (2019) 25: 60

Table 1 Experimental and calculated solvatochromic range (Δν = νnonpolar−νpolar, in cm−1) of C153 and DCB-1d for their absorption and emission
spectra

Compound Spectra Δf* (μe−μg)μg/e ∏e−∏g SSD Δν SS-PCM Δν LR-PCM Δν Exp.

C153 Abs. 0.292 38.49 0.0139 0.0777 1589 863 1597 [19]
Emi. 0.292 72.62 0.1644 2791 540 3404 [19]
DCB-1d Abs. 0.279 0 0.0065 0.0524 1307 664 –
Emi. 0.279 0 0.0893 2823 629 1548 [12]

Theoretical values are calculated using the state-specific (SS) and linear response (LR) PCM methods coupled with TD-PBE0 approach, respectively.
Molecular dipole moments of S0 (μg) and S1 (μe) were solved using the DFT and TD-DFT methods with the optimized gas-phase S0 structure. The
Lippert-Mataga descriptors (μe−μg)μg/e is in Debye2 . The average deviation of ESP of S0(Πg) and S1(Πe) are calculated using ground and excited
surface ESP values. The unit of Π is atomic unit of ESP. The state-specific descriptors for absorption (SSDg) and emission (SSDe) are in atomic unit (a.u.)
calculated using the Multiwfn ESP analysis. Δf* represents the polarity difference of the different solvents

solvent shift. Compared with Lippert-Mataga dipole approxi- Results and discussion
mation, the SSD fully considers electronic distribution and mo-
lecular shape effects on solute–solvent interactions. To simplify Solvatochromism of C153 and DCB-1d
the evaluation of solvent–shift capability, we calculate only the
SSD based on ground-state balanced geometry. ESP values (Vg The electron/hole wavefunctions for C153 and DCB-1d dur-
and Ve) were calculated based on the ground-state and excited- ing the electronic transition from their ground state (S0) to the
state densities in vacuum. The average deviations of ESPs for lowest excited singlet state (S1) are shown in Fig. 3. It can be
ground sate (Πg) and excited state (Πe) were also calculated seen readily that the electronic transition of C153 defines the
according to Politzer’s definition [18]. As a comparison, the charge transfer character, i.e., electron transfer from the
standard linear response (LR) [24] and state-specific (SS) electron-rich (left side) to electron-deficient (right side) group
[25–27] PCM were also applied to describe electronic spectra while the electron and hole of DCB-1d have center-symmetric
in different solvents. The detailed calculations of absorption feature. This leads to a large excited molecular dipole (μe) for
and emission spectra using TD-DFT coupled with SS-PCM C153 and zero μe for DCB-1d, as shown in Table 1.
are provided in the Supporting Information. All DFT and TD- The solvatochromic capability is usually estimated by the
DFT calculations were performed using the Gaussian 09 pack- solvent-shift range (Δν) from a non-polar solvent, such as
age [28] with the PBE0 exchange-correlation functional at 6- cyclohexane, to a polar solvent, such as methanol, acetonitrile
31G(d) level. Molecular vdW surface and ESP analysis were and acetone. From Table 1, we can see that both C153 and
performed using the Multiwfn 3.5 developed by Lu and co- DCB-1d undergo solvatochromic behaviors. The Δνemi of
workers [29]. emission is nearly twice the Δνabs of absorption. For the

S1

S1
S0
S0

Fig. 4 Histograms of electrostatic potential (ESP) on the van der Waals mapping images of S0 and S1 molecules. Some local minimum and max-
(vdW) surfaces of C153 (left) and DCB-1d (right). Calculated ESPs: blue imum ESP values are labeled on the vdW surface
first singlet excited state (S1), red ground state (S0). Inset Colored ESP
J Mol Model (2019) 25: 60 Page 5 of 8 60

Table 2 Experimental solvatochromic ranges (Δν = νnonpolar−νpolar, in cm−1), calculated using the Lippert-Mataga descriptors, Πe−Πg and SSDg of
eight molecules for their absorption spectra

∆νabs=ν(nonpolar )
No Structure (μe-μg)μg ∏e-∏g SSDg 103
-ν(polar)

-2650=27930(CCl4)
1 -3.32 0.0011 -10.75
-30580(H2O) [2]

+1597=25979(C6H12)
2 38.49 0.0139 77.67
-24382(CH3CN) [19]

+2700=22620(C6H14)
3 18.33 0.0015 11.05
-19920(DMSO) [2]

+3780=20640(C5H12)
4 34.84 0.0065 27.42
-16860 (CH2I2) [2]

+4000=25640(C5H12)
5 71.68 0.0165 58.02
-21640 (CH2I2) [2]

6 24.72 0.0086 34.14 +4030=30170(C6H14)

-26140 (H2O) [2]

+4240=30000(C6H14)
7 -8.15 0.0069 74.24
-25760(CF3CH2OH) [2]

+4710=21460(C6H14)
8 68.99 0.0133 62.77
-16750(/H2O) [2]
Here (μe-μg)μg in Debye2 , Π and SSDg in the atomic unit are calculated using the DFT and TD-DFT methods with the optimized S0 geometry

well-studied C153, the Lippert-Mataga expression (see Eq. 5) developed to approximately evaluate the solvent-shift range
can be used to explain why Δνemi is larger than Δνabs. This is shown in Eq. 8. From the calculated SSD for C153, we find
ascribed to the larger excited-state dipole (μe) values in terms that the SSD of the emission process is nearly twice that of the
of the ground-state dipole (μg). For the center-symmetric SSD of the absorption process, which explains why Δνemi is
DCB-1d molecule, μg and μe are zero. As such, the Lippert- almost twice Δνabs. Very interestingly, the SSD value of the
Mataga expression fails to explain the solvatochromism of DCB-1d emission is nearly identical to that for absorption of
DCB-1d. And the point-dipole approximation cannot estimate the C153 molecule, which explains why the Δνabs of C153 is
the solute–solvent interactions. To address this issue, a higher very close to Δνemi of DCB-1d, as observed experimentally.
multipole, such as the quadrupole, is needed. Instead of the To understand why SSDe is larger than SSDg, we checked the
higher multipole expansion, the EPS-based SSD was distribution of ESPs on the vdW surfaces of two molecules in
60 Page 6 of 8 J Mol Model (2019) 25: 60

based solely on the electron densities of S0 and S1 with the

optimized ground state geometries in vacuum, which avoids
the iterative calculations of the SS-PCM approach. Using
Lippert-Mataga descriptors, i.e., (μe−μg)μg for absorption
and (μe−μg)μg for emission, we find that these descriptors
can well explain the fact that the emission of C153 has twice
the solvatochromic range than that of its absorption spectra.
But the zero value of the Lippert-Mataga descriptor for DCB-
1d cannot explain its solvatochromic range in emission spec-
tra. When adopting the molecular charge separation descriptor
(Π) to represent intermolecular interactions, Πe−Πg describes
the difference interactions between solvent with the ground
and excited solute. From Table 1, we note that Πe−Πg of
center-symmetric DCB-1d is not zero, and the Πe–Πg of
Fig. 5 Relationship of the SSDe and the experimental solvatochromic C153 is twice as large as DCB-1d, which fits well as C153
range (Δν) of the nine molecular emission spectra has twice the solvent shift of emission spectra than DCB-1d.

their ground-state and excited-state, respectively. From Fig. 4, Validation of SSDe and SSDg
it can be seen that the ESP distributions of the excited-state of
C153 and DCB-1d are notably wider than those of their As discussed above, the SSD can describe the solute-solvent
ground-state ESPs. This indicates that solute–solvent interac- interaction well, and qualitatively estimate the solvatochromic
tions in the excited state are stronger than those in the ground effect beyond the point-dipole approximation. In order to ver-
state, which results in a larger SSD for emission according to ify their performance for the estimation of the polarity-
Eq. 8. sensitive transition energy, we calculated the SSD of the ab-
Besides these methods to qualitatively evaluate polarity- sorption and emission for some well-studied solvatochromic
sensitive fluorescent behavior, TD-DFT coupled with PCM, molecules in the literature [1, 12, 19, 30–32]. Table 2 lists the
including the linear-response (LR) and state-specific (SS) solvent-shift ranges (Δν), Lippert-Mataga descriptor and SSD
model, is widely used to calculate solvent-shift ranges (Δν) of the absorption of eight molecules. Table 2 shows that SSDg
for absorption and emission spectra. In Table 1, we listed the can qualitatively describe the solvatochromic effect.
Δνabs and Δνemi obtained from TD-DFT coupled with the Generally, the negative SSD value relates to blue-shift
LR-PCM and SS-PCM approximation. The main difference solvatochromism while positive SSD value suggests a red-
between LR-PCM and SS-PCM is how the polarization re- shift effect. Usually, the large SSD can lead to a strong
sponse charges are treated. In the SS-PCM approach, the sol- solvent-shift range. However, there are some abnormal mole-
vent polarization response is decomposed into fast and slow cules. For example, the SDD of molecule 2 is very large while
responses, which represent the response of solvent electron its Δνabs is very small. Since the SSD implicitly considers
and orientation, respectively. The excitation energy is the dif- only the electrostatic solute-solvent interaction, the effect of
ference between the corresponding free energies, i.e., with the other strong interactions, such as hydrogen bonds, cannot be
state-specific ground-state and excited-state QM continuum included in the SSD approach. Such strong interactions may
models. Thus, the excitation energies require an iterative pro- lead to a slight deviation in estimating solvatochromic effects.
cedure to considering the effective solvent-induced interac- As for the Δν of the emission spectra shown in Fig. 5, the
tions for the ground-state and excited-state calculations. In good linear relationship between the SSDe and Δνemi means
the LR-PCM approach, the absorption/emission energies are that the SSDe could correctly describe polarity-sensitive fluo-
determined directly by the poles of the frequency-dependent rescent behavior. As a comparison, the Lippert-Mataga de-
linear response functions of molecules in the ground state, i.e., scriptor fails to estimate the solvatochromic effect, especially
the solvent is always fully equilibrated with ground-state elec- for DA molecules with center-symmetry such as DCB deriva-
tron densities. Consequently, the iterative scheme of SS-PCM tions, as evidenced by the results listed in Table 3.
is no longer required. From Table 1, we find that the SS-PCM
gives a larger Δν of emission than of absorption, while the
LR-PCM presents an incorrect picture. Compared with the Conclusions
experimental data, the SS-PCM correctly describes the
Δνemi of C153, but overestimates Δνemi of DCB-1d. The In summary, molecular surface ESPs coupled with the solvent
SSDe and SSDg can relate well to the solvent-shift range for reaction field were applied to implicitly describe solute–
both emission and absorption spectra. The calculations are solvent interactions beyond the point-dipole approximation.
J Mol Model (2019) 25: 60 Page 7 of 8 60

Table 3 Experimental solvatochromic ranges (Δν = νnonpolar−νpolar, in cm−1) of eight molecules for their emission spectra

∆νabs=ν(nonpolar )
No Structure (μe-μg)μg ∏e-∏g SSDg 103
-ν(polar)

+800=23530(C5H12)
1 61.00 0.0135 97.60
-22730(H2O) [2]

+1036=19310(C6H12)
2 0.00 0.0075 102.16
-18274(CH2Cl2) [12]

+1274=21615(C6H12)
3 0.00 0.0067 96.40
-20341(CH2Cl2) [12]

+1548=19357(C6H12)
4 0.00 0.0065 89.32
-17809(CH3COCH3) [12]

+2220=21234(toluene)
5 35.77 0.0080 141.42
-19014(CH3OH) [1,32]

+3404=21753(C6H12)
6 72.62 0.0139 164.43
-18349(CH3CN) [19]

+3475=20369(toluene)
7 235.34 0.0117 217.17
-16894(H2O) [1,30]

+5498=23044(toluene)
8 189.61 0.0150 276.72
-17546(CH3OH) [1 , 31]

The Lippert-Mataga descriptor (μe-μg)μe (in Debye2 ) is calculated using the DFT and TD-DFT methods. The Πe-Πg and SSDe in atomic unit) is
calculated using the surface ESP surface values obtained by Multiwfn

The ESP-based descriptor SSD was utilized to evaluate well-studied fluorescent molecules. The SSD can well de-
solvent-shift ranges for both molecular absorption and emis- scribe the solvochromic behavior of those molecules, taking
sion spectra. Our results demonstrate that both SSDe and into consideration the dipole changes during electron state
SSDg can successfully estimate polarity-sensitive absorption transition. More importantly, it can explain the solvent shift
and emission features. A good linear relationship between of the center-symmetric DA molecules, which will be helpful
SSDe and solvatochromic effect was found in a study of nine in the molecular design of polarity-sensitive probes.
60 Page 8 of 8 J Mol Model (2019) 25: 60

Acknowledgments The authors thank Dr. Yun Wang helpful discus- electrostatic potentials and average local ionization energies. J
sions. This work is supported by the National Key R&D program of Mol Model 16(11):1679–1691. https://doi.org/10.1007/s00894-
China (Grant No. 2017YFB0203404), National Natural Science 010-0692-x
Foundation of China (21173138), and Shaanxi Innovative Team of Key 17. Murray JS, Politzer P (2011) The electrostatic potential: an over-
Science and Technology (2103KCT-17). view. Wiley Interdiscip Rev Comput Mol Sci 1(2):153–163. https://
doi.org/10.1002/wcms.19
Publisher’s note Springer Nature remains neutral with regard to jurisdic- 18. Brinck T, Murray JS, Politzer P (1992) Quantitative determination
tional claims in published maps and institutional affiliations. of the total local polarity (charge separation) in molecules. Mol
Phys 76(3):609–617. https://doi.org/10.1080/00268979200101561
19. Horng ML, Gardecki JA, Papazyan A, Maroncelli M (1995)
References Subpicosecond measurements of polar solvation dynamics: couma-
rin 153 revisited. J Phys Chem 99(48):17311–17337
1. Klymchenko AS (2017) Solvatochromic and fluorogenic dyes as 20. Stone AJ (2013) The theory of intermolecular forces. Oxford
environment-sensitive probes: design and biological applications. University Press, Oxford
Acc Chem Res 50(2):366–375. https://doi.org/10.1021/acs. 21. Tomasi J, Mennucci B, Cammi R (2005) Quantum mechanical
accounts.6b00517 continuum solvation models. Chem Rev (Washington, DC)
2. Reichardt C (1994) Solvatochromic dyes as solvent polarity indi- 105(8):2999–3094. https://doi.org/10.1021/cr9904009
cators. Chem Rev 94(8):2319–2358 22. Leach AR (2001) Molecular modelling: principles and applications.
3. Xu T, Wang W, Yin S (2018) Electrostatic polarization energies of Pearson, Harlow
charge carriers in organic molecular crystals: a comparative study
23. Lu T, Chen F (2012) Quantitative analysis of molecular surface
with explicit state-specific atomic polarizability based AMOEBA
based on improved marching tetrahedra algorithm. J Mol Graph
force field and implicit solvent method. J Chem Theory Comput
Model 38(9):314–323
14(7):3728–3739. https://doi.org/10.1021/acs.jctc.8b00132
24. Improta R, Barone V, Scalmani G, Frisch MJ (2006) A state-
4. Xu T, Wang W, Yin S (2018) Explicit method to evaluate the ex-
specific polarizable continuum model time dependent density func-
ternal reorganization energy of charge-transfer reactions in
tional theory method for excited state calculations in solution. J
oligoacene crystals using the state-specific polarizable force field.
Chem Phys 125(5):054103
J Phys Chem A 122(45):8957–8964. https://doi.org/10.1021/acs.
25. Improta R, Scalmani G, Frisch MJ, Barone V (2007) Toward effec-
jpca.8b08998
5. Luzhkov V, Warshel A (1991) Microscopic calculations of solvent tive and reliable fluorescence energies in solution by a new state
effects on absorption spectra of conjugated molecules. J Am Chem specific polarizable continuum model time dependent density func-
tional theory approach. J Chem Phys 127(7):074504
Soc 113(12):4491–4499
6. Zeng J, Hush N, Reimers J (1993) Solvent effects on molecular 26. Mewes J-M, You Z-Q, Wormit M, Kriesche T, Herbert JM, Dreuw
spectra. III. Absorption to and emission from the lowest singlet A (2015) Experimental benchmark data and systematic evaluation
(n, π*) state of dilute pyrimidine in water. J Chem Phys 99(3): of two a posteriori, polarizable-continuum corrections for vertical
1508–1521 excitation energies in solution. J Phys Chem A 119:5446–5464
7. Hush NS, Reimers JR (2000) Solvent effects on the electronic spec- 27. You Z-Q, Mewes J-M, Dreuw A, Herbert JM (2015) Comparison of
tra of transition metal complexes. Chem Rev 100(2):775–786 the Marcus and Pekar partitions in the context of non-equilibrium,
8. McRae E (1957) Theory of solvent effects on molecular electronic polarizable-continuum solvation models. J Chem Phys 143:204104
spectra. Frequency shifts. J Phys Chem 61(5):562–572 28. Frisch MJ, Trucks GW, Schlegel HB, Scuseria GE, Robb MA,
9. Amos A, Burrows B (1973) Solvent-shift effects on electronic spec- Cheeseman JR, Scalmani G, Barone V, Mennucci B, Petersson
tra and excited-state dipole moments and polarizabilities. Adv GA, Nakatsuji H, Caricato M, Li X, Hratchian HP, Izmaylov AF,
Quantum Chem 7:289–313 Bloino J, Zheng G, Sonnenberg JL, Hada M, Ehara M, Toyota K,
10. Onsager L (1936) Electric moments of molecules in liquids. J Am Fukuda R, Hasegawa J, Ishida M, Nakajima T, Honda Y, Kitao O,
Chem Soc 58(8):1486–1493. https://doi.org/10.1021/ja01299a050 Nakai H, Vreven T, Montgomery Jr JA, Peralta JE, Ogliaro F,
11. Lippert E (1957) Spektroskopische bestimmung des Bearpark M, Heyd JJ, Brothers E, Kudin KN, Staroverov VN,
dipolmomentes aromatischer verbindungen im ersten angeregten Kobayashi R, Normand J, Raghavachari K, Rendell A, Burant
singulettzustand. Z Elektrochem Ber Bunsenges Phys Chem 61(8): JC, Iyengar SS, Tomasi J, Cossi M, Rega N, Millam JM, Klene
962–975 M, Knox JE, Cross JB, Bakken V, Adamo C, Jaramillo J, Gomperts
12. Zhang JN, Kang H, Li N, Zhou SM, Sun HM, Yin SW, Zhao N, R, Stratmann RE, Yazyev O, Austin AJ, Cammi R, Pomelli C,
Tang BZ (2016) Organic solid fluorophores regulated by subtle Ochterski JW, Martin RL, Morokuma K, Zakrzewski VG, Voth
structure modification: color-tunable and aggregation-induced GA, Salvador P, Dannenberg JJ, Dapprich S, Daniels AD, Farkas
emission. Chem Sci 8:577 O, Foresman JB, Ortiz JV, Cioslowski J, Fox DJ (2009)
13. Brinck T, Murray JS, Politzer P (1992) Surface electrostatic poten- GAUSSIAN 09, revision D.01. Gaussian, Inc., Wallingford
tials of halogenated methanes as indicators of directional intermo- 29. Lu T, Chen F (2012) Multiwfn: a multifunctional wavefunction
lecular interactions. Int J Quantum Chem 44(S19):57–64. https:// analyzer. J Comput Chem 33(5):580–592
doi.org/10.1002/qua.560440709 30. Vázquez ME, Blanco JB, Imperiali B (2005) Photophysics and
14. Politzer P, Murray JS, Edward Grice M, Desalvo M, Miller E biological applications of the environment-sensitive fluorophore
(1997) Calculation of heats of sublimation and solid phase heats 6-N, N-dimethylamino-2, 3-naphthalimide. J Am Chem Soc
of formation. Mol Phys 91(5):923–928. https://doi.org/10.1080/ 127(4):1300–1306
002689797171030 31. Kucherak OA, Didier P, Mély Y, Klymchenko AS (2010) Fluorene
15. Politzer P, Murray JS, Peralta-Inga Z (2001) Molecular surface analogues of prodan with superior fluorescence brightness and
electrostatic potentials in relation to noncovalent interactions in solvatochromism. J Phys Chem Lett 1(3):616–620
biological systems. Int J Quantum Chem 85(6):676–684. https:// 32. Holmes-Farley SR, Whitesides GM (1986) Fluorescence proper-
doi.org/10.1002/qua.1706 ties of dansyl groups covalently bonded to the surface of oxida-
16. Bulat FA, Toro-Labbé A, Brinck T, Murray JS, Politzer P (2010) tively functionalized low-density polyethylene film. Langmuir
Quantitative analysis of molecular surfaces: areas, volumes, 2(3):266–281