Mechanistic Photochemistry of Methyl-4-hydroxycinnamate

Jul 11, 2016 - Herein we computationally studied the excited-state properties and decay dynamics of methyl-4-hydroxycinnamate (OMpCA) in the lowest th...
78 downloads 8 Views 3MB Size
Article pubs.acs.org/JPCA

Mechanistic Photochemistry of Methyl-4-hydroxycinnamate Chromophore and Its One-Water Complexes: Insights from MSCASPT2 Study Xiao-Ying Xie, Chun-Xiang Li, Qiu Fang, and Ganglong Cui* Key Laboratory of Theoretical and Computational Photochemistry, Ministry of Education, College of Chemistry, Beijing Normal University, Beijing 100875, China S Supporting Information *

ABSTRACT: Herein we computationally studied the excited-state properties and decay dynamics of methyl-4hydroxycinnamate (OMpCA) in the lowest three electronic states, that is, 1ππ*, 1nπ*, and S0 using combined MS-CASPT2 and CASSCF electronic structure methods. We found that one-water hydration can significantly stabilize and destabilize the vertical excitation energies of the spectroscopically bright 1ππ* and dark 1nπ* excited singlet states, respectively; in contrast, it has a much smaller effect on the 1ππ* and 1nπ* adiabatic excitation energies. Mechanistically, we located two 1ππ* excited-state relaxation channels. One is the internal conversion to the dark 1nπ* state, and the other is the 1ππ* photoisomerization that eventually leads the system to a 1ππ*/S0 conical intersection region, near which the radiationless internal conversion to the S0 state occurs. These two 1ππ* relaxation pathways play distinct roles in OMpCA and its two one-water complexes (OMpCA-W1 and OMpCA-W2). In OMpCA, the predominant 1ππ* decay route is the state-switching to the dark 1nπ* state, while in onewater complexes, the importance of the 1ππ* photoisomerization is significantly enhanced because the internal conversion to the 1 nπ* state is heavily suppressed due to the one-water hydration.



INTRODUCTION Sunscreens are an efficient tool in providing photoprotection against the harmful effects of ultraviolet (UV) radiation.1,2 Acute and chronic UV exposure usually leads to sunburn, photocarcinogensis, and photoaging. There exist several photoprotection methods, for example, sun avoidance, seeking shade, use of protective clothing, and application of sunscreens, among which the use of sunscreens remains one of the most prevalent protection strategies in our daily life. However, their safety and efficacy still do not satisfy various realistic needs. A lot of studies have called into question the potential for vitamin D deficiency with sunscreen use. The possibility of adverse biological effects from various ingredients of sunscreens has recently been developed as well.3−5 So, understanding their atomistic photoprotection mechanism is meaningful, which can provide much valuable knowledge for the design of better sunscreens.6−12 Because of sharing a same molecular skeleton within the commonly used sunscreen, excited-state dynamics of various methylcinnamate compounds, for example, p-coumaric acids, methoxy methylcinnamate, and sinapoyl ester derivatives, have attracted a lot of experimental and theoretical studies in the past decade.13−17 For example, Buma and co-workers employed high-resolution spectroscopic techniques to study the excitedstate dynamics of para-methoxy methylcinnamate.17 Miyazaki et al. studied the S1 state dynamics of para-, ortho-, and meta© 2016 American Chemical Society

methoxy methylcinnamate under supersonic jet-cooled conditions using both laser-induced fluorescence and mass-resolved resonant two-photon ionization spectroscopy and low-temperature matrix−isolation Fourier transform infrared spectroscopy.7 Very recently, Miyazaki et al. and Cui et al. have also employed time-dependent density functional theory (TDDFT) and MS-CASPT2 methods to explore the excited-state photoisomerization and decay pathways of para-methoxy methylcinnamate compound, respectively.7,12 In this work, we focus on the excited-state dynamics of another kind cinnamate compound, that is, methyl-4hydroxycinnamate (OMpCA). Experimentally, Buma et al. have used high-resolution gas-phase ultravisible−infrared spectroscopic techniques to explore the spectroscopic properties of the lower excited singlet states of OMpCA in neutral form and showed that the system can adopt four stable conformations with similar excited-state properties.18,19 In addition to local spectroscopic properties, they have also explored the excited-state dynamics of OMpCA and its onewater clusters using high-resolution resonance-enhanced multiphoton ionization and laser-induced fluorescence excitation spectra.20 They concluded that one-water hydration has a large Received: June 11, 2016 Revised: July 10, 2016 Published: July 11, 2016 6014

DOI: 10.1021/acs.jpca.6b05899 J. Phys. Chem. A 2016, 120, 6014−6022

Article

The Journal of Physical Chemistry A

space of 12 electrons in 9 orbitals, that is, five occupied π molecular orbitals and one occupied n orbital, three unoccupied π* molecular orbitals, is used (see Supporting Information). Equal state weights for the lowest five roots are also used in the CASSCF calculations. In addition, an imaginary shift of 0.2 au is used to avoid the intruder-state issue;31 the Cholesky decomposition technique with unbiased auxiliary basis sets is used for accurate two-electron integral approximations;32 the ionization potential-electron affinity (IPEA) shift was not applied.33 Vertical excitation energies are also computed using the MS-CASPT2 method. The 6-31G* basis set34,35 is used for all computations. All CASSCF optimizations for conical intersections are conducted using Gaussian03;36 all CASSCF optimizations for minima and minimum-energy paths, and all MS-CASPT2 single-point energy refinements, are performed using MOLCAS8.0.32

influence on the excited-state lifetime and that the decay pathway involving the dissociative 1πσ* state is less important. Instead, the internal conversion to an optically dark 1nπ* state is a rapid excited-state relaxation channel. Recently, Buma and co-workers further reported a high-resolution three-color nanosecond multiphoton ionization pump−probe spectroscopic study on the excited-state dynamics of OMpCA and its one-water clusters. They further confirmed that the main decay channel of the lowest excited 1ππ* singlet state of OMpCA is the internal conversion to the dark 1nπ* state, while one-water complexation completely changes the photodynamics of OMpCA. In this situation, the S1 bond isomerization plays a dominant role.21 Nevertheless, Ebata et al.22 suggested that both OMpCA and its water complexes decay to the ground state due to the photoisomerization. Clearly, on the experimental side, the photophysical and photochemical mechanism of OMpCA and its one-water complex is elusive and even incompatible in certain aspects. To resolve these mechanistic issues, ab initio electronic structure calculations are desirable. De Groot et al. performed TD-DFT, CASSCF, and EOM/CCSD computations on vertical excitation energies of OMpCA.18 Ebata et al. performed a symmetry-adapted cluster-configuration interaction (SAC− CI) study on the S1 photoisomerization of OMpCA and its one-water complex, but they only focused on the initial stage (180°−120°), possibly due to the use of single-reference electronic structure method.22 Tan et al. employed the spincomponent scaled coupled-cluster singles-and-doubles model (SCS-CC2) to compute the vertical and adiabatic excitation energies of OMpCA and its one-water complex.21,23 However, to our best knowledge, there is no reported comprehensive electronic structure calculations aiming at elucidating the mechanistic photophysics and photochemistry of OMpCA and its water complexes. In this work, we employed CASSCF and MS-CASPT2 methods to explore the photophysical and photochemical mechanism of OMpCA and its two one-water clusters and found that the one-water hydration effects are sitespecific, which can visibly affect the vertical excitation energies, the internal conversion to the optically dark 1nπ* state, and eventually the photophysical and photochemical mechanism of OMpCA.





RESULTS S0 Minima. At the B3LYP/6-31G* level, we optimized four conformers for OMpCA, which are the same as those obtained by de Groot et al.18 Consistent with previous DFT and CCSD electronic structure calculations, the conformer I is predicted as the most stable one (see Supporting Information). Similarly, we optimized two most stable one-water complexes, referred to as OMpCA-W1 and OMpCA-W2 hereinafter, in which the water molecule is bound with the terminal hydroxyl and carbonyl groups, respectively. In the following, we only focus on the photophysics and photochemistry of these three most stable structures, which are referred to as S0-MIN, S0−W1-MIN, and S0−W2-MIN hereinafter. Local Spectroscopic Properties. Table 1 collects the MSCASPT2 computed vertical excitation energies to the 1nπ* and Table 1. MS-CASPT2//CASSCF Computed Vertical Excitation Energies (in eV) to the 1nπ* and 1ππ* Excited Singlet States at the Franck−Condon Points of OMpCA and Its Two Water Complexes OMpCA-W1 and OMpCA-W2 ππ*

OMpCA OMpCA-W1 OMpCA-W2

1

1

4.32 3.91 3.91

4.74 5.17 5.29

nπ*

ππ* excited singlet states at the Franck−Condon points of OMpCA and its two water complexes OMpCA-W1 and OMpCA-W2. For all these three complexes, the 1ππ* excited singlet state is predicted to be the lowest one at their Franck−Condon points, and the vertical excitation energies are computed to be 4.32, 3.91, and 3.91 eV for OMpCA, OMpCA-W1, and OMpCA-W2 at the MS-CASPT2 level, respectively. One-water hydration significantly reduces the vertical excitation energy to the 1ππ* excited singlet state. The reduction is up to ca. 0.4 eV at the MS-CASPT2 level. However, the hydration at the O atom of the hydroxyl group or of the carbonyl group does not make visible difference (see Table 1). The 1nπ* excited singlet state is a spectroscopically dark state at the Franck−Condon point. At the MS-CASPT2 level, the 1 nπ* excited singlet state is predicted to be much higher than the 1ππ* state, by 0.42 eV for OMpCA, by 1.26 eV for OMpCA-W1, and by 1.38 eV for OMpCA-W2. Thus, onewater hydration significantly destabilizes the vertical excitation energy to the 1nπ*state. Furthermore, it is apparent that the hydration at the O atom of the carbonyl group has a little stronger influence than that at the O atom of the terminal hydroxyl group. 1

COMPUTATIONAL DETAILS

S0 conformers are first optimized using the B3LYP method.24−27 Minima, minimum-energy conical intersections (MECIs), minimumenergy paths (MEPs), and linearly interpolated internal coordinate (LIIC) paths are computed using the state-averaged complete active space self-consistent field (SA-CASSCF) method in which equal state weights are used. MECI optimizations are performed using the gradient-projection optimization algorithm.28 Minima and MECIs are optimized without any geometric constraints; MEPs are obtained using the constrained optimization technique, in which only the selected one reaction coordinate, for example, dihedral angles, is fixed at certain value, and all the other degrees of freedom are fully optimized. In all SA-CASSCF calculations, an active space of 10 electrons in 8 molecular orbitals is employed, which is referred to as CASSCF(10,8) hereinafter. Specifically, for the 1ππ* state, five occupied π molecular orbitals and three unoccupied π* molecular orbitals are used, while for the 1nπ* state, one occupied π molecular orbital is replaced by the n orbital of the O atom of the carbonyl group (see Supporting Information). To obtain more accurate potential energy profiles, the MS-CASPT2 method29,30 that provides more correlation energy is exploited to reevaluate the energies of all CASSCF optimized geometries and reaction paths. In single-point MS-CASPT2 calculations, a larger active 6015

DOI: 10.1021/acs.jpca.6b05899 J. Phys. Chem. A 2016, 120, 6014−6022

Article

The Journal of Physical Chemistry A We analyzed molecular orbitals relevant to the S0→1nπ* and S0→1ππ* electronic transitions at the S0 minimum of OMpCA (see Figure 1). The S0→1ππ* electronic excitation is mainly

Figure 2. CASSCF(10,8)/6-31G* optimized S0 and S1 minima and S1/S0 conical intersection of OMpCA (bond length in angstrom and dihedral angle in degree). See Supporting Information for their full Cartesian coordinates. Table 2 collects their relative energies refined by the MS-CASPT2 method.

Figure 1. MS-CASPT2 computed molecular orbitals involved in the S0→1ππ* (first two rows) and S0→1nπ* (third row) electronic transitions.

composed by two electronic configurations. The first one corresponds to an electron hopping from highest occupied molecular orbital (HOMO−1) to lowest unoccupied molecular orbital (LUMO; weight: 0.374); the other one does to an electron transition from HOMO to LUMO+1 (weight: 0.286). Further analysis shows that HOMO−1 and LUMO+1 are primarily localized within the phenyl group, while HOMO and LUMO are extensively delocalized over the whole molecular space. Thus, both electronic configurations have clear chargetransfer character. In comparison, the S0→1nπ* electronic excitation mainly stems from an electronic transition from HOMO−3 to LUMO. The HOMO−3 molecular orbital is nearly exclusively from the lone-pair orbital of the O atom of the CO carbonyl group, and the lone-pair orbital of the remote O−H group has little contribution. Molecular orbitals relevant to the S0→1nπ* and S0→1ππ* electronic transitions of OMpCA-W1 and OMpCA-W2 are similar to those of OMpCA and thus compiled in Supporting Information. Minima. At the CASSCF level, we optimized the S0, 1nπ*, and 1ππ* minima of OMpCA, which are denoted as S0-MIN, NP-MIN, and PP-MIN in Figure 2. All these three structures have a planar conformation. Compared with the geometric parameters of S0-MIN, the C2−C3 bond length of PP-MIN is shortened heavily, 1.467 Å of S0-MIN versus 1.409 Å of PP-MIN; both C3−C4 and C4− C5 bond lengths are increased and decreased subtly, respectively. In addition, it can be found that the CO bond length is nearly not changed, 1.190 Å of S0-MIN versus 1.193 Å of PP-MIN at the CASSCF level. In stark contrast, the C3−C4, C4−C5, and C5−O6 bond lengths are changed remarkably from S0-MIN to NP-MIN. For example, the C3−C4 bond is elongated to 1.433 Å of NP-MIN from 1.315 Å of S0-MIN (see the other changes in Figure 2). Of course, in comparison with those of the S0 minimum, the largest structural change of the 1nπ* minimum is related to the CO bond length, which is increased to 1.380 from 1.190 Å at S0-MIN. As shown in Figures 3 and 4, the hydrogen-bonding interactions at the terminal hydroxyl group and the carbonyl group do not cause remarkable structural changes for the S0,

Figure 3. CASSCF(10,8)/6-31G* optimized S0 and S1 minima and S1/S0 conical intersection of OMpCA-W1 (bond length in angstrom and dihedral angle in degree). See Supporting Information for their full Cartesian coordinates. Table 2 collects their relative energies refined by the MS-CASPT2 method.

Figure 4. CASSCF(10,8)/6-31G* optimized S0 and S1 minima and S1/S0 conical intersection of OMpCA-W2 (bond length in angstrom and dihedral angle in degree). See Supporting Information for their Cartesian coordinates. Table 2 collects their relative energies refined by the MS-CASPT2 method.

nπ*, and 1ππ* minima of OMpCA-W1 and OMpCA-W2 compared with those without water hydration (i.e., OMpCA) except the number of the hydrogen bonds. In OMpCA-W1, only a hydrogen bond is formed; in contrast, there are two hydrogen bonds in OMpCA-W2.

1

6016

DOI: 10.1021/acs.jpca.6b05899 J. Phys. Chem. A 2016, 120, 6014−6022

Article

The Journal of Physical Chemistry A On the energetic side, the 1ππ* minima still have the lowest adiabatic excitation energies as shown in Table 2. At the MS-

ψS = 0.79|222222000| − 0.45|222221100| 0

and

Table 2. MS-CASPT2 Refined Single-Point Energies (kcal/ mol) of All CASSCF Optimized Stationary Points and Conical Intersections of OMpCA and Its Two Water Complexes OMpCA-W1 and OMpCA-W2 OMpCA OMpCA-W1 OMpCA-W2 a

ππ*

1

in which 2, 1, and 0 represent doubly-, singly-, and emptyoccupied molecular orbitals, respectively. In contrast, at the conical intersections of OMpCA-W1 and OMpCA-W2, the S1 and S0 wave functions primarily comprise a Slater determinant

ππ*/S0 a

S0

1

1

1

0.0 0.0 0.0

87.9 86.1 86.9

89.7 89.3 92.0

68.6/67.7 67.5/63.2 70.4/63.4

nπ*

ψS = −0.43|222222000| − 0.82|222221100|

ψS = −0.92|222221100| 0

ψS = 0.89|2222222000|

S0 and S1 energies are reported for the conical intersections.

1

and ψS = 0.93|222221100|

CASPT2 level, they are predicted to be 87.9, 86.1, and 86.9 kcal/mol, which are 1.8, 3.2, and 5.1 kcal/mol lower than those of the 1nπ* minima. Further examination shows that the hydration at different sites only makes a little difference on the adiabatic excitation energy of the 1ππ* state, for example, 87.9 kcal/mol of OMpCA versus 86.1 kcal/mol of OMpCA-W1 versus 86.9 kcal/mol of OMpCA-W2. In contrast, the hydration effect on the 1nπ* state is a little complicated. The hydration at the terminal hydroxyl group has almost no influence on the 1 nπ* state, 89.7 kcal/mol of OMpCA versus 89.3 kcal/mol of OMpCA-W1; in contrast, the hydration at the carbonyl group destabilizes the 1nπ* state by 2.3 kcal/mol. This difference can be understood by taking into account that the lone-pair electron involved in the 1nπ* electronic transition stems from the O atom of the carbonyl group, not from the O atom of the terminal hydroxyl group (see above discussion). In other words, the hydrogen-bonding interaction inhibits to certain extent the n→π* electronic transition. Conical Intersections. In addition to minima in the 1nπ* and 1ππ* states, we optimized the 1ππ*/S0 minimum-energy conical intersections of OMpCA, OMpCA-W1, and OMpCAW2 at the CASSCF level. Their schematic structures are shown in Figures 2−4 as well as some selected geometric parameters. These 1ππ*/S0 conical intersections have some similar geometric parameters. For example, all systems have a much twisting C2−C3−C4−C5 dihedral angle, which is computed to be 97.3°, 97.4°, and 104.6° for OMpCA, OMpCA-W1, and OMpCA-W2. The C3−C4 bond length is 1.474 Å in OMpCA, 1.462 Å in OMpCA-W1, and 1.452 Å in OMpCA-W2; the C5− O6 bond length is 1.283 Å in OMpCA, 1.285 Å in OMpCAW1, and 1.255 Å in OMpCA-W2; the C4−C5 bond length is 1.363 Å in OMpCA, 1.374 Å in OMpCA-W1, and 1.363 Å in OMpCA-W2. However, the C2−C3 bond length of OMpCAW2, 1.389 Å, is much shorter than those of OMpCA and OMpCA-W1 (1.444 and 1.433 Å at the CASSCF level). At the MS-CASPT2 level, the S1 and S0 potential energies of these 1ππ*/S0 conical intersections are computed to be 68.6/ 67.7, 67.5/63.2, and 70.4/63.4 kcal/mol for OMpCA, OMpCAW1, and OMpCA-W2, respectively. The S1−S0 energy gap of the 1ππ*/S0 conical intersection of OMpCA is still maintained to a small value (0.9 kcal/mol at the MS-CASPT2 level), whereas the energy gaps of OMpCA-W1 and OMpCA-W2 become a little larger, which are 4.3 and 7.0 kcal/mol at the MS-CASPT2 level (see Table 2). We also analyzed the wave functions of the S1 and S0 states at these 1ππ*/S0 conical intersections. At the conical intersection of OMpCA, the S1 and S0 wave functions are mainly composed of two Slater determinants

0

ψS = −0.94|2222222000| 1

respectively. Figure 5 shows the HOMO and LUMO orbitals at the conical intersection of OMpCA.

Figure 5. Two singly occupied molecular orbitals composed of two main electronic configurations at the 1ππ*/S0 conical intersection. See text for discussion.

Relaxation Paths. To explore the excited-state decay dynamics of OMpCA, OMpCA-W1, and OMpCA-W2, we studied several possible 1ππ* excited-state deactivation channels with respect to the rotations of the central C2−C3−C4−C5, C1−C2−C3−C4, and C3−C4−C5−O7 dihedral angles, and the linearly interpolated internal coordinate (LIIC) paths connecting the 1ππ* minima and the 1nπ* minima. These paths are optimized at the CASSCF level (except LIIC paths), and single-point energies are further refined at the MS-CASPT2 level. The top-left panel of Figure 6 shows the S1 minimum-energy rotation path with regard to the rotation of the C2−C3−C4− C5 dihedral angle. In this route, the S1 and S0 energies gradually increase along this reaction coordinate in the first phase (i.e., from 180°−130°). Then the S1 system overcomes a small barrier of 7.6 kcal/mol at 130°. Afterward, both states move very close to each other but do not cross. It is noteworthy that in OMpCA, there exist two low-lying 1 ππ* excited states, which is similar to the situation in trans-pcoumaric acid.37 The lower one is V′(1ππ*) with a relatively small oscillator strength, while the other one, V(1ππ*), has a larger oscillator strength (see Supporting Information). At the SCS-CC2/cc-pVTZ level, the oscillator strengths of V′(1ππ*) and V( 1 ππ*) are computed to be 0.221 and 0.627, respectively;21 hence, both states can be populated at the Franck−Condon point. Our current study is focused on the photoisomerization process in the adiabatic S1 state because relevant experimental excited-state dynamics studies are performed in the lowest bright spectroscopically state.20,22 Further electronic structure analysis shows that the S1 state is of V′(1ππ*) character in the beginning, while it is of V(1ππ*) 6017

DOI: 10.1021/acs.jpca.6b05899 J. Phys. Chem. A 2016, 120, 6014−6022

Article

The Journal of Physical Chemistry A

Figure 6. MS-CASPT2//CASSCF computed S1 minimum-energy rotation and LIIC paths of OMpCA. See text for discussion.

Figure 7. MS-CASPT2//CASSCF computed S1 minimum-energy rotation and LIIC paths of OMpCA-W1. See text for discussion.

the internal conversion to the dark 1nπ* singlet state is efficient due to a small barrier of ca. 2.9 kcal/mol at the MS-CASPT2 level. Therefore, for OMpCA, the dominant 1ππ* nonadiabatic relaxation pathway is the internal conversion to the dark 1nπ* singlet state. When separately adding a water molecule at the hydroxyl and carbonyl groups, the overall shapes of these three minimumenergy rotation paths do not change significantly (see Figures 7 and 8). For example, the barriers for the rotations along the C2−C3−C4−C5 dihedral angle are computed to be 5.9 and 7.5

character in the end. The barrier in the adiabatic S1 state is actually an avoided crossing between the diabatic V′(1ππ*) and V(1ππ*) states (MS-CASPT2, see Figure 6). In comparison, the S1 minimum-energy rotation paths with respect to the rotations of the C1−C2−C3−C4 and C3−C4− C5−O7 dihedral angles are mechanistically unimportant because of large S1→ S0 energy gaps and barriers in the entire reaction paths (see bottom-left and -right panels). The topright panel shows the MS-CASPT2 computed LIIC path connecting both 1ππ* and 1nπ* minima, which illustrates that 6018

DOI: 10.1021/acs.jpca.6b05899 J. Phys. Chem. A 2016, 120, 6014−6022

Article

The Journal of Physical Chemistry A

Figure 8. MS-CASPT2//CASSCF computed S1 minimum-energy rotation and LIIC paths of OMpCA-W2. See text for discussion.

efficiently. Figure 9 illustrates the two main S1 excited-state relaxation routes of OMpCA, OMpCA-W1, and OMpCA-W2, respectively. However, the importance of these two S1 excited-state relaxation pathways is variational from OMpCA to OMpCAW1 and OMpCA-W2. For OMpCA, the photoisomerization route is heavily suppressed by the efficient internal conversion to the dark 1nπ* singlet state because the former has a much higher barrier than the latter, 7.6 versus 2.9 kcal/mol at the MSCASPT2 level. For OMpCA-W1 and OMpCA-W2, the photoisomerization route is comparable to the internal conversion to the dark 1nπ* singlet state, because their barriers are close to each other, 5.9 versus 5.1 kcal/mol for OMpCAW1 and 7.5 versus 7.6 kcal/mol for OMpCA-W2. Previous experiments by Buma et al.21 show that the dominant decay channel of the lowest 1ππ* excited singlet state of OMpCA is the internal conversion to the lower 1nπ* state, which is supported by our present MS-CASPT2//CASSCF electronic structure calculations. As shown in the top panel of Figure 9, the state-switching from the 1ππ* to 1nπ* state is associated with a small barrier of 2.9 kcal/mol, which is much lower than that associated with the S1 photoisomerization (7.6 kcal/mol at the MS-CASPT2 level). Buma et al. also argued that when adding a single water molecule, the dominant 1ππ* decay path of OMpCA is radically changed to the 1ππ* photoisomerization.21 This can also be understood on the basis of the present computational results. When a water molecule is bound to the terminal hydroxyl group, the photoisomerization barrier is decreased a little, from 7.6 kcal/mol of OMpCA to 5.9 kcal/mol of OMpCA-W1; in contrast, the barrier of switching to the dark 1nπ* state is increased remarkably, by ∼2.2 kcal/mol at the MS-CASPT2 level (see the middle panel of Figure 9). Therefore, the importance of the photoisomerization in the S1 relaxation is significantly enhanced, which is eventually comparable with that of the state switching to the dark 1nπ* state.

kcal/mol for OMpCA-W1 and OMpCA-W2, which are close to 7.6 kcal/mol for OMpCA. Nonetheless, the barriers for the internal conversion to the dark 1nπ* singlet state are increased, 5.1 and 7.6 kcal/mol for OMpCA-W1 and OMpCA-W2 versus 2.9 kcal/mol of OMpCA at the MS-CASPT2 level. Thus, for OMpCA-W1 and OMpCA-W2, the rotation pathway along the C2−C3−C4−C5 dihedral angle and the internal conversion to the dark 1nπ* singlet state becomes comparably important from energetics viewpoint (i.e., similar barriers). As discussed above, adding one water at the hydroxyl and carbonyl groups does not drastically affect the barriers associated with the rotation paths along the C2−C3−C4−C5 dihedral angles but intensively increases the barriers of the internal conversion to the dark 1nπ* singlet state. How to understand these features? This is because the intermolecular hydrogen bonding inhibits the corresponding n→π* electronic transition but has little effect on the π→π* electronic transition. Furthermore, it is easy to figure out that the barrier of the internal conversion to the dark 1nπ* singlet state of OMpCAW2 is higher than that of OMpCA-W1. This can be also understood taking into account that the lone-pair electrons involved in the n→π* electronic transition originate not from the oxygen atom of hydroxyl group but of the carbonyl group. Discussions and Conclusion. OMpCA, OMpCA-W1, and OMpCA-W2 share similar excited-state properties at their Franck−Condon points. They have a spectroscopically bright 1 ππ* singlet excited state. Upon irradiation, this 1ππ* excited singlet state is populated; then, the system first relaxes to its nearby 1ππ* minimum. From this S1(1ππ*) minimum, there exist two main S1 excited-state relaxation pathways. One is the radiationless internal conversion process to the spectroscopically dark 1nπ* singlet state; the other is the S1 photoisomerization with respect to the central C2−C3−C4−C5 dihedral angle. Along the photoisomerization path, the system arrives at an S1(1ππ*)/S0 conical intersection region, near which the nonadiabatic S1→ S0 internal conversion happens 6019

DOI: 10.1021/acs.jpca.6b05899 J. Phys. Chem. A 2016, 120, 6014−6022

Article

The Journal of Physical Chemistry A

CASPT2 levels; see Supporting Information). In this work, we first explored the photochemistry of this one-water complex, OMpCA-W2. The bottom panel of Figure 9 schematically depicts its two competitive S1 relaxation pathways, that is, the 1 ππ* photoisomerization and the state switching to the dark 1 nπ* state. Their importances are also comparable, which is similar to OMpCA-W1. However, one can find that both photoisomerization and state-switching barriers are increased from OMpCA-W1 to OMpCA-W2, for example, from 5.1 to 7.6 kcal/mol for the state switching to the dark 1nπ* state. This is because the 1nπ* state originates from the n→π* electronic transition, in which the lone-pair electron stems from the O atom of the carbonyl group (see Figure 1). The hydration at the carbonyl group inhibits the n→π* electronic excitation, thereby destabilizing the 1nπ* state and finally increasing the barrier. In one word, the S1 relaxation dynamics of OMpCAW2 is similar to that of OMpCA-W1 but associated with a little higher barriers, but its contribution to the photodynamics of one-water complexes of OMpCA should not be completely excluded. Finally, one can find that OMpCA shares an overall similar photophysical and photochemical mechanism with p-methoxy methylcinnamate (p-MMC).12 In these two chromophores, the internal conversion from the initially populated bright 1ππ* to dark 1nπ* state is dominant decay path in vacuo, whereas in their one-water complexes, the S1 photoisomerization becomes significantly pronounced. Nevertheless, there are many different aspects between OMpCA and p-MMC; for example, the vertical excitation energy to the 1ππ* state of OMpCA is much lower than that of p-MMC (see literature for others).7,12 In addition, p-MMC only has a one-water complex; in contrast, OMpCA has two. Thereby, the photodynamics of OMpCA should be more complicated. In summary, we have employed combined MS-CASPT2// CASSCF electronic structure methods to computationally study the photophysics and photochemistry of OMpCA and its two one-water complexes in the lowest three S0, 1nπ*, and 1 ππ* states. We have found that the one-water hydration at either hydroxyl or carbonyl group to different extent destabilizes the 1nπ* state and thus increases the 1ππ*→ 1 nπ* state-switching barrier. In contrast, the hydration has a smaller influence on the 1ππ* state and the 1ππ* photoisomerization barrier. We have also confirmed that the internal conversion to the dark 1nπ* state is the major excited-state relaxation channel and that the 1ππ* photoisomerization is almost negligible in OMpCA; however, the importance of the 1 ππ* photoisomerization is enhanced very much in OMpCAW1 and OMpCA-W2. Our present theoretical work provides valuable mechanistic insights for understanding the photophysics and photochemistry of methylcinnamate compounds, which share a same molecular skeleton with the commonly used sunscreens.

Figure 9. Suggested photophysical and photochemical mechanisms of OMpCA, OMpCA-W1, and OMpCA-W2 based on the present MSCASPT2//CASSCF electronic structure calculations. See text for discussion.

Ebata and co-workers22 estimated the lifetime of the S1→ S0 origin of OMpCA to be ∼100 times shorter than that of OMpCA-H2O (8−9 vs 930 ps). This is qualitatively consistent with our present results. For OMpCA, the dominant S1 decay pathway is the state switching from the 1ππ* to 1nπ* state with a barrier of 2.9 kcal/mol. In contrast, the barriers related to both 1ππ* photoisomerization and state switching to the dark 1 nπ* state are increased to 5.9 and 5.1 kcal/mol for OMpCAW1 and 7.5 and 7.6 kcal/mol for OMpCA-W2, respectively. In addition to OMpCA-W1, there as well exists another kind of one-water complex, namely, OMpCA-W2. In the former, one water is bound to the terminal hydroxyl group; in the latter, the water is bound to the carbonyl group. Previous electronic structure calculations only consider OMpCA-W1. OMpCA-W2 is never studied computationally until now; however, the contribution of this kind of one-water complex to the excitedstate dynamics of one-water complexes of OMpCA cannot be excluded, because OMpCA-W1 and OMpCA-W2 are energetically close to each other (within 1 kcal/mol at B3LYP and MS-



ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.jpca.6b05899. Illustrated molecular structures, tabulated calculated relative energies of four conformers of OMpCA and OMpCA-H2O, illustrated active orbitals, illustrated molecular orbitals, active space in the CASSCF 6020

DOI: 10.1021/acs.jpca.6b05899 J. Phys. Chem. A 2016, 120, 6014−6022

Article

The Journal of Physical Chemistry A



(14) Karpkird, T. M.; Wanichweacharungruang, S.; Albinsson, B. Photophysical Characterization of Cinnamates. Photochem. Photobiol. Sci. 2009, 8, 1455−1460. (15) Kockler, J.; Oelgemöller, M.; Robertson, S.; Glass, B. D. Photostability of Sunscreens. J. Photochem. Photobiol., C 2012, 13, 91− 110. (16) Stavros, V. G. A Bright Future for Sunscreens. Nat. Chem. 2014, 6, 955−956. (17) Tan, E. M. M.; Hilbers, M.; Buma, W. J. Excited-State Dynamics of Isolated and Microsolvated Cinnamate-Based UV-B Sunscreens. J. Phys. Chem. Lett. 2014, 5, 2464−2468. (18) de Groot, M.; Gromov, E. V.; Köppel, H.; Buma, W. J. HighResolution Spectroscopy of Methyl 4-Hydroxycinnamate and Its Hydrogen-Bonded Water Complex. J. Phys. Chem. B 2008, 112, 4427− 4434. (19) Tan, E. M. M.; Amirjalayer, S.; Smolarek, S.; Vdovin, A.; Rijs, A. M.; Buma, W. J. Conformational Heterogeneity of Methyl 4Hydroxycinnamate: A Gas-Phase UV-IR Spectroscopic Study. J. Phys. Chem. B 2013, 117, 4798−4805. (20) Smolarek, S.; Vdovin, A.; Tan, E. M. M.; de Groot, M.; Buma, W. J. Spectroscopy and Dynamics of Methyl-4-Hydroxycinnamate: the Influence of Isotopic Substitution and Water Complexation. Phys. Chem. Chem. Phys. 2011, 13, 4393−4399. (21) Tan, E. M. M.; Amirjalayer, S.; Bakker, B. H.; Buma, W. J. Excited State Dynamics of Photoactive Yellow Protein Chromophores Elucidated by High-Resolution Spectroscopy and ab initio Calculations. Faraday Discuss. 2013, 163, 321−340. (22) Shimada, D.; Kusaka, R.; Inokuchi, Y.; Ehara, M.; Ebata, T. Nonradiative Decay Dynamics of Methyl-4-Hydroxycinnamate and Its Hydrated Complex Revealed by Picosecond Pump-Probe Spectroscopy. Phys. Chem. Chem. Phys. 2012, 14, 8999−9005. (23) Tan, E. M. M.; Amirjalayer, S.; Mazzella, P.; Bakker, B. H.; van Maarseveen, J. H.; Bieraugel, H.; Buma, W. J. Molecular Beam and ab initio Studies of Photoactive Yellow Protein Chromophores: Influence of the Thioester Functionality and Single Bond Rotation. J. Phys. Chem. B 2014, 118, 12395−12403. (24) Vosko, S. H.; Wilk, L.; Nusair, M. Accurate Spin-Dependent Electron Liquid Correlation Energies for Local Spin Density Calculations: A Critical Analysis. Can. J. Phys. 1980, 58, 1200−1211. (25) Lee, C.; Yang, W. T.; Parr, R. G. Development of the ColleSalvetti Correlation-Energy Formula into a Functional of the Electron Density. Phys. Rev. B: Condens. Matter Mater. Phys. 1988, 37, 785−789. (26) Becke, A. D. Density-Functional Exchange-Energy Approximation with Correct Asymptotic Behavior. Phys. Rev. A: At., Mol., Opt. Phys. 1988, 38, 3098−3100. (27) Becke, A. D. A New Mixing of Hartree-Fock and Local DensityFunctional Theories. J. Chem. Phys. 1993, 98, 1372−1377. (28) Bearpark, M. J.; Robb, M. A.; Schlegel, H. B. A Direct Method for the Location of the Lowest Energy Point on a Potential Surface Crossing. Chem. Phys. Lett. 1994, 223, 269−274. (29) Andersson, K.; Malmqvist, P.-Å; Roos, B. O.; Sadlej, A. J.; Wolinski, K. Second-Order Perturbation Theory with a CASSCF Reference Function. J. Phys. Chem. 1990, 94, 5483−5488. (30) Andersson, K.; Malmqvist, P.-Å; Roos, B. O. Second-Order Perturbation Theory with a Complete Active Space Self-Consistent Field Reference Function. J. Chem. Phys. 1992, 96, 1218−1226. (31) Försberg, N.; Malmqvist, P.-Å. Multiconfiguration Perturbation Theory with Imaginary Level Shift. Chem. Phys. Lett. 1997, 274, 196− 204. (32) Aquilante, F.; Lindh, R.; Pedersen, T. B. Unbiased Auxiliary Basis Sets for Accurate Two-Electron Integral Approximations. J. Chem. Phys. 2007, 127, 114107. (33) Ghigo, G.; Roos, B. O.; Malmqvist, P.-Å A Modified Definition of the Zeroth-Order Hamiltonian in Multiconfigurational Perturbation Theory (CASPT2). Chem. Phys. Lett. 2004, 396, 142−149. (34) Hariharan, P. C.; Pople, J. A. The Influence of Polarization Functions on Molecular Orbital Hydrogenation Energies. Theor. Chem. Acc. 1973, 28, 213−222.

computations, and Cartesian coordinates of all optimized structures. (PDF)

AUTHOR INFORMATION

Corresponding Author

*Phone: 86-010-58806770. E-mail: [email protected]. Notes

The authors declare no competing financial interest.



ACKNOWLEDGMENTS This work was supported by the National Natural Science Foundation of China (Nos. 21522302 and 21520102005); G.C. is also grateful for financial support from the “Recruitment Program of Global Youth Experts” and ”Fundamental Research Funds for Central Universities”.



REFERENCES

(1) Huncharek, M.; Kupelnick, B. Use of Topical Sunscreens and the Risk of Malignant Melanoma: A Meta-Analysis of 9067 Patients from 11 Case-Control Studies. Am. J. Public Health 2002, 92, 1173−1177. (2) van der Pols, J. C.; Williams, G. M.; Pandeya, N.; Logan, V.; Green, A. C. Prolonged Prevention of Squamous Cell Carcinoma of the Skin by Regular Sunscreen Use. Cancer Epidemiol., Biomarkers Prev. 2006, 15, 2546−2548. (3) Salamone, L. M.; Dallal, G. E.; Zantos, D.; Makrauer, F.; DawsonHughes, B. Contributions of Vitamin D Intake and Seasonal Sunlight Exposure to Plasma 25-Hydroxyvitamin D Concentration in Elderly Women. Am. J. Clin. Nutr. 1994, 59, 80−86. (4) Kielbassa, C.; Roza, L.; Epe, B. Wavelength Dependence of Oxidative DNA Damage Induced by UV and Visible Light. Carcinogenesis 1997, 18, 811−816. (5) Wagner, C. L.; Greer, F. R. Prevention of Rickets and Vitamin D Deficiency in Infants, Children, and Adolescents. Pediatrics 2008, 122, 1142−1152. (6) Karsili, T. N. V.; Marchetti, B.; Ashfold, M. N. R.; Domcke, W. ab initio Study of Potential Ultrafast Internal Conversion Routes in Oxybenzone, Caffeic Acid, and Ferulic Acid: Implications for Sunscreens. J. Phys. Chem. A 2014, 118, 11999−12010. (7) Miyazaki, Y.; Yamamoto, K.; Aoki, J.; Ikeda, T.; Inokuchi, Y.; Ehara, M.; Ebata, T. Experimental and Theoretical Study on the Excited-State Dynamics of ortho-, meta-, and para-Methoxy Methylcinnamate. J. Chem. Phys. 2014, 141, 244313. (8) Kumasaka, R.; Kikuchi, A.; Yagi, M. Photoexcited States of UV Absorbers, Benzophenone Derivatives. Photochem. Photobiol. 2014, 90, 727−733. (9) Baker, L. A.; Horbury, M. D.; Greenough, S. E.; Coulter, P. M.; Karsili, T. N. V.; Roberts, G. M.; Orr-Ewing, A. J.; Ashfold, M. N. R.; Stavros, V. G. Probing the Ultrafast Energy Dissipation Mechanism of the Sunscreen Oxybenzone after UVA Irradiation. J. Phys. Chem. Lett. 2015, 6, 1363−1368. (10) Baker, L. A.; Horbury, M. D.; Greenough, S. E.; Ashfold, M. N. R.; Stavros, V. G. Broadband Ultrafast Photoprotection by Oxybenzone across the UVB and UVC Spectral Regions. Photochem. Photobiol. Sci. 2015, 14, 1814−1820. (11) Ignasiak, M. T.; Houée-Levin, C.; Kciuk, G.; Marciniak, B.; Pedzinski, T. A Reevaluation of the Photolytic Properties of 2Hydroxybenzophenone-Based UV Sunscreens: Are Chemical Sunscreens Inoffensive? ChemPhysChem 2015, 16, 628−633. (12) Chang, X.-P.; Li, C.-X.; Xie, B.-B.; Cui, G. L. Photoprotection Mechanism of p-Methoxy Methylcinnamate: A CASPT2 Study. J. Phys. Chem. A 2015, 119, 11488−11497. (13) Promkatkaew, M.; Suramitr, S.; Karpkird, T. M.; Namuangruk, S.; Ehara, M.; Hannongbua, S. Absorption and Emission Spectra of Ultraviolet B Blocking Methoxy Substituted Cinnamates Investigated Using the Symmetry-Adapted Cluster Configuration Interaction Method. J. Chem. Phys. 2009, 131, 224306. 6021

DOI: 10.1021/acs.jpca.6b05899 J. Phys. Chem. A 2016, 120, 6014−6022

Article

The Journal of Physical Chemistry A (35) Ditchfield, R.; Hehre, W. J.; Pople, J. A. Self-Consistent Molecular-Orbital Methods. IX. An Extended Gaussian-Type Basis for Molecular-Orbital Studies of Organic Molecules. J. Chem. Phys. 1971, 54, 724−728. (36) Frisch, M. J.; Trucks, G. W.; Schlegel, H. B.; Scuseria, G. E.; Robb, M. A.; Cheeseman, J. R.; Montgomery, J. A., Jr.; Vreven, T.; Kudin, K. N.; Burant, J. C. et al. Gaussian 03, Revision D.02; Gaussian, Inc: Wallingford, CT, 2004. (37) Gromov, E. V.; Burghardt, I.; Köppel, H.; Cederbaum, L. S. Impact of Sulfur vs Oxygen on the Low-Lying Excited States of transp-Coumaric Acid and trans-p-Coumaric Thio Acid. J. Phys. Chem. A 2005, 109, 4623−4631.

6022

DOI: 10.1021/acs.jpca.6b05899 J. Phys. Chem. A 2016, 120, 6014−6022