pccp

View Article Online View Journal

PCCP Physical Chemistry Chemical Physics

Accepted Manuscript

This article can be cited before page numbers have been issued, to do this please use: P. S. Sherin, Y. P. Tsentalovich, E. Vauthey and E. Benassi, Phys. Chem. Chem. Phys., 2018, DOI: 10.1039/C8CP02183J. Volume 18 Number 1 7 January 2016 Pages 1–636

PCCP Physical Chemistry Chemical Physics www.rsc.org/pccp

This is an Accepted Manuscript, which has been through the Royal Society of Chemistry peer review process and has been accepted for publication. Accepted Manuscripts are published online shortly after acceptance, before technical editing, formatting and proof reading. Using this free service, authors can make their results available to the community, in citable form, before we publish the edited article. We will replace this Accepted Manuscript with the edited and formatted Advance Article as soon as it is available. You can find more information about Accepted Manuscripts in the author guidelines.

ISSN 1463-9076

PERSPECTIVE Darya Radziuk and Helmuth Möhwald Ultrasonically treated liquid interfaces for progress in cleaning and separation processes

Please note that technical editing may introduce minor changes to the text and/or graphics, which may alter content. The journal’s standard Terms & Conditions and the ethical guidelines, outlined in our author and reviewer resource centre, still apply. In no event shall the Royal Society of Chemistry be held responsible for any errors or omissions in this Accepted Manuscript or any consequences arising from the use of any information it contains.

rsc.li/pccp

Page 1 of 13

Please do not adjust margins Physical Chemistry Chemical Physics View Article Online

DOI: 10.1039/C8CP02183J

Published on 04 May 2018. Downloaded by Kaohsiung Medical University on 04/05/2018 17:46:23.

ARTICLE Ultrafast excited state decay of natural UV filters: from intermolecular hydrogen bonds to conical intersection Received 00th January 20xx, Accepted 00th January 20xx DOI: 10.1039/x0xx00000x www.rsc.org/

a,b,c,*

Peter S. Sherin,

a,b

Yuri P. Tsentalovich,

c

Eric Vauthey and Enrico Benassi

c,d,*

Kynurenines (KNs) are natural UV filters of the human eye lens, protecting the eye tissues from the solar UV radiation. Key points of their effective protection are the intramolecular charge transfer (ICT) in the excited state and the fast dissipation of absorbed light energy into heat via the intermolecular H-bonds. Herein we report that the introduction of unsaturated double bond in the amino acid side chain, operating as ICT-enhancing electron donor group, drastically accelerates the internal conversion (IC) due to a conical intersection (CI) between the potential energy surfaces of excited and ground states. Our photophysical study of a deaminated KN (carboxyketoalkene, CKA), an intrinsic product of KN thermal decomposition, demonstrates an unusually fast excited state decay in a broad range of solvents of different polarity and proticity. The detailed analysis of interactions in the excited state by different computational techniques revealed that the changes in molecular structure – the twist of carbonyl group from the plane of aromatic ring followed by the formation of two mutually orthogonal conjugated substructures – are responsible for the CI of the excited and ground state potential energy surfaces. Intermolecular H-bonds facilitate the transition to CI, but do not play a crucial role in the fast decay of the excited state. An extremely fast and efficient IC in CKA opens the way for the design of new types of organic UV filters and their applications in material sciences, cosmetics and medicine.

Introduction Protection from the solar UV radiation is a matter of great importance for many living organisms on our planet. The mechanisms of UV protection include the synthesis and application of UV filters – compounds absorbing UV light and dissipating the electronic energy via benign channels, mainly heat. A fast energy dissipation without initiation of chemical reactions is the protection mechanism of human tissues exposed to UV light, i.e. skin and eyes. Amongst these protecting systems, molecules with effective deactivation of photo-excited states, such as melanins in skin1-3 and kynurenines (KNs) in eye lenses,4-6 are known. Despite investigations on the photophysics and photochemistry of biological UV filters over the last decade,1-3,7-10 many aspects of their action mechanisms are still not fully understood. From the viewpoint of photoprotection, a rapid nonadiabatic transition between electronic excited (S1) and ground (S0) states via a conical intersection (CI) with the restoration of the initial S0 state is the most effective deactivation pathway of

excited states without formation of long-lived intermediates capable to react with surrounding molecules, like triplet states or radicals. The deactivation of the excited states via CI was 11 however, this reported for many molecular systems; mechanism was shown to not always lead to the complete restoration of the initial state. Nucleic acids, the building blocks of DNA, are ones of the well-known examples of effective excited state deactivation via CI with almost 100% 12 efficiency of the ground state recovery. Besides CI, other possible mechanisms of excited state deactivation may operate in biological tissues, in skin and eyes in particular. The biopolymer eumelanin, one of the three basic types of melanin – the main pigment of human skin, shows an efficient excited state decay via multiple pathways 2 involving intra- and inter-unit proton transfer. The photoprotection of human lens and retina is provided by KNs present in the lens, which efficiently transform the light energy 9,13 into heat via intermolecular H-bonds. Unfortunately, natural UV filters are not completely free from harmful side effects. For example, melanins exhibit 14 phototoxic properties and might be potentially carcinogenic. Urocanic acid, another epidermal UV chromophore, isomerizes under UV light and the accumulation of cis-isomers may induce 15 various forms of skin cancer via immunosuppression. KNs contribute to the age-related changes of the whole eye lens 16,17 tissue as well as to the cataract progression via either 18 direct reactions of photoexcited triplet states with proteins or the formation of thermally or photochemically active 19-24 products.

J. Name., 2013, 00, 1-3 | 1

This journal is © The Royal Society of Chemistry 20xx

Please do not adjust margins

Physical Chemistry Chemical Physics Accepted Manuscript

Journal Name

Please do not adjust margins Physical Chemistry Chemical Physics

Page 2 of 13 View Article Online

DOI: 10.1039/C8CP02183J

The deactivation efficiency of photoexcited KNs strongly depends on the medium. Two key steps are responsible for the 9,10,13,25 fast S1 → S0 non-radiative transition in KNs as well as in 26,27 its chromophoric core, the o-aminoacetophenone: (i) intramolecular charge transfer (ICT) from amino to carbonyl group upon photoexcitation, and (ii) effective dissipation of absorbed light energy into the surrounding environment via intermolecular H-bonds. Though such H-bond-induced nonradiative decay (HBIND) has been observed for many organic 28 molecules, a general mechanism of this phenomenon is still missing. Significant progress in understanding of this process 29 was achieved in the recent work. The authors demonstrated that the efficiency of HBIND mechanism is directly related to the ability of the solvent molecules to form an extended Hbonded network and to the presence of H-bond donating and 29 accepting functional groups in the photoactive molecule. Under physiological conditions KNs undergo a spontaneous deamination with the formation of α,β-unsaturated 30 carboxyketoalkene (CKA, Chart 1). In the present work, we demonstrate that the transformation of a KN molecule, occurring in vivo within the lens tissue, can switch the mechanism of the efficient S1 → S0 decay from HBIND to a process involving a CI. In particular, the unsaturated bond in α position with respect to the carbonyl group in CKA plays a crucial role in the excited state dynamics of CKA, enhancing ICT and leading to the intersection of S1 and S0 potential energy surfaces. This drastic change in the mechanism of nonradiative decay significantly accelerates the deactivation of the S1 state and makes it highly efficient even in the absence of intermolecular H-bonds.

(EtOH), dimethylsulfoxide (DMSO) and chloroform; the spectra in other solvents are given in Fig. S1 in the Electronic Supplementary Information (ESI).† In all solvents, CKA exhibits -1 large red shift (ap. 2300 cm ) of the absorption band as compared to KN (grey curve, Fig. 1A). This indicates a stronger ICT character of the electronic excited state in CKA than in KN due to the presence of the unsaturated bond, which is electron-donating in the case of CKA. The position of the maximum of CKA absorption band is almost independent on the solvent polarity and the ability to donate H-bond, Fig. 1A and S1.† Probably the absence of solvatochromism in CKA absorption spectra should be attributed to the presence of intramolecular H-bond between carboxyl and amino groups, as confirmed by calculations (see below). (A)

CKA H2O EtOH DMSO Chloroform KN H 2O

(B)

300

Chart 1. Chemical structures of KN and CKA.

In this work deep insight is obtained using state-of-the-art computational methods, which were used to rationalize the mechanisms of photoinduced processes in KNs along with the experimental results. In particular, various aspects of solutesolvent interactions as well as intramolecular interactions within the photoexcited molecule were analysed and rationalized. The synergy of experimental and computational investigations, in which numerous techniques and methodologies were employed, gave the possibility to eventually draw a clear and sound model to describe the mechanism of the S1 → S0 transition in CKA.

Results Steady-state optical properties. Photostability of CKA The absorption spectra of CKA were recorded in various solvents. CKA is not soluble in non-polar solvents; it has limited solubility in low polar solvents and good solubility in polar solvents. Fig. 1A shows spectra of CKA in water (H2O), ethanol

350 400 450 500 Wavelength / nm Fig. 1. (A) Measured and (B) calculated (including convolution of vibronic progression) absorption spectra of CKA in H2O, EtOH and DMSO. The absorption spectrum of KN in water is shown as grey curve in (A) for reference.

In all used solvents, CKA exhibits very weak emission, which could not be reliably recorded due to a presence of small amounts of impurities with significantly higher fluorescence quantum yields. The absence of detectable fluorescence indicates a fast radiationless S1 → S0 transition in a broad range of solvents with different polarity and proticity. Prolonged irradiation of CKA in aqueous solutions and various organic solvents does not lead to any changes in its absorption spectrum. The photodegradation yield (Φdeg) was measured for CKA in aqueous solution under anaerobic conditions (see Experimental section in ESI† for details and Fig. S2† for the concentration profile of CKA degradation under -5 used conditions); the obtained value of Φdeg = (1.5±0.3)×10 is 8,10 close to values reported for KN and its derivatives. Traces of impurities may exhibit photochemical activity towards CKA, giving a contribution in the total degradation of CKA. Thus, the measured Φdeg value should be interpreted as the upper limit

2 | J. Name., 2012, 00, 1-3




Journal Name

Absorption


ARTICLE

Page 3 of 13


DOI: 10.1039/C8CP02183J

ARTICLE indicates

high

Fluorescence (norm.)

1.0

580 nm 620 nm 660 nm 720 nm

(A)

0.8 0.6

Amplitude

The temporal evolution of CKA excited state could only be studied in polar solvents due to the low solubility of CKA in weakly polar solvents. Fluorescence time profiles were recorded with CKA in H2O, EtOH and DMSO at different wavelengths after 430 nm excitation. Fig. 2A shows some of the measured time profiles for CKA in aqueous solution; the same data for CKA in EtOH and DMSO are displayed in Fig. S3.† CKA fluorescence decays within less than one picosecond in water and several picoseconds in organic solvents. The obtained results are well reproduced by a global data analysis using a convolution of a Gaussian-like instrument response function with a sum of exponential functions: two for aqueous solution and three for organic solvents. For all solvents one additional exponential function was used to take into account small contribution from impurities; this slow component will be omitted further on because of its low amplitude. The best fits are present as smooth lines; the corresponding time constants are listed in Table 1.

signal intensity. The A1(λ) spectrum in aqueous solution (Fig. 2A, inset) and A1(λ), A2(λ) spectra in organic solvents (Fig. S3B, S3D†) exhibit positive and negative bands, indicating the decay on a blue side and the rise on a red side of the emission band. These spectral changes and characteristic time constants are typical for the dynamics Stokes shifts observed with other 31 organic dyes in water and organic solvents. Therefore, these components should be attributed to solvent relaxation. Similar Ai(λ) were observed in the case of KN and its derivatives in 9,13,22 both aqueous environment and organic solvents. The A3(λ) in all solvents should be attributed to the decay of S1 state population. Table 1. Time constants (τi, i = 1,2,3,4) obtained from a global fit of the fluorescence up-conversion and TA data measured with CKA in various solvents.

Solvent PBS, pH 7.4 H2O, pH 6.2 D2O D2O/H2O (16/1 v/v) MeOH

A1 (< 0.15 ps) A3 (0.60 ps)

1.0

EtOH 0.5

EtOH/H2O (16/1 v/v)

0.0 0.4

ACN DMF

550 600 650 700 750

0.2

Wavelength / nm

DMSO

0.0 0 8

3

6 ∆A x 10


for CKA photodegradation yield that photochemical stability of this molecule. Time-resolved fluorescence dynamics

1

2 3 Time delay / ps

0.5 ps 1.0 ps 1.5 ps 2.0 ps 3.1 ps 10 ps 1.7 ns

(B)

a

(C)

SAS1 SAS3 SAS4

2 0 500

600

700 400

500

b

τ2 / ps − − − − − − 0.80 b 1.4 a 1.4 b 1.2 a 1.9 b 1.1 b 1.7 b 2.2 a 2.1 b 1.7 a 1.7 b

τ3 / ps 0.60 a 0.67 b 0.77 a 0.68 b 0.91 b 0.86 b 2.8 b 4.4 a 4.6 b 4.1 a 4.5 b 4.3 b 7.5 b 8.9 a 9.9 b 7.3 a 7.8 b

τ4 / ps − 2.3 b − 1.6 b 2.2 b 2.7 b − − − − − − − − − − −

– fluorescence up-conversion measurements (λex = 430 nm); – TA measurements (λex = 400 nm).

Transient absorption (TA) dynamics

4

400

DMSO/H2O (16/1 v/v)

10 20 30 40 50

τ1 / ps 700 nm in aqueous solution (Fig. 2B) and strong dips at 520-580 nm in organic solvents (Fig. S4A, S4C†) correspond to the regions where the signal from the S1 → S0 stimulated emission is stronger or comparable with that from the S1 → Sn absorption. The same spectral features were 9,10,13,22 observed earlier for KN and its derivatives. The observed TA evolution was analysed globally, assuming a series of three successive exponential processes (A → B → C → 0). The resulting species-associated spectra (SAS),

J. Name., 2018, 00, 1-3 | 3




Journal Name



DOI: 10.1039/C8CP02183J

decay of the S1 state. Indeed, SAS4(λ) exhibits a negative band with the maximum around 380 nm, which corresponds to the bleach of the thermally-equilibrated CKA ground state, and a positive band with the maximum around 460 nm, red shifted with respect to the steady state absorption. These two features are typically observed for highly vibrationally excited 37,38 ground states, and this dark component may, therefore, be attributed to the S0(νn) → S1 transition. The time profiles of CKA fluorescence and TA intensity at the maxima of the corresponding bands in various solvents are shown in Fig. 3. The similarity of the fluorescence and TA dynamics reflects the fact that both methods report on the evolution of only one species – the S1 state of CKA. The temporal evolution of KN fluorescence at its emission 9 maximum (from ref. ) is shown for comparison. KN

1.0 Fluorescence / a.u.

attributed to species produced/decaying after optical excitation. SASi(λ), i = 1,3,4 for CKA in aqueous solution and i = 1,2,3 in organic solvents are shown in Fig. 2C and Fig. S4†, respectively. It should be noted that cross-phase modulation and coherence, induced by the temporal overlap of the pump and probe pulses around time zero, generate strong artifacts in TA spectra within first ≈ 0.2 ps after the excitation (data not shown). This makes difficult to obtain the real dynamics in this time frame as compared with fluorescence up-conversion technique. Therefore, for some solvents these ultrafast components could not be resolved. The SAS1(λ) should be interpreted as the TA spectrum of the Franck-Condon state observed immediately after excitation; the subsequent changes in shapes of SASi(λ) reflects the evolution of the total TA signal due to the dynamic Stokes shift of the S1 → S0 stimulated emission. The first time constant τ1 should be attributed to the solvent inertial motion, i.e., small motion of the solvent molecule or part of it in its own free volume; the obtained τ1 values agree with the values 31-33 published in literature for various organic dyes. The second component may be assigned to three processes: diffusive 31 reorientational motion of solvent molecules, vibrational 34-36 relaxation or a conformational relaxation of CKA in the 9,10,13 excited state, similarly to what was found with KNs. For most organic solvents, the observed τ2 values (Table 1) are similar to those corresponding to diffusive motion, viz. 1-2 31 ps. The case of ACN is an exception due to short 31 characteristic time for its diffusion motion, viz. 0.63 ps that is almost twice shorter than the obtained τ2 = 1.1 ps for CKA in ACN. This indicates vibrational relaxation and conformational relaxation could contribute to this stage of the evolution of CKA excited states in ACN and most likely possesses characteristic times close to the values for diffusion motion in other organic solvents. Therefore, the contributions of these three processes cannot be discriminated in the case of CKA, whereas conformational relaxation is well separated in the 9 case of KN. In aqueous solution, these processes are also mixed with the decay of excited state, τ3 ≈ 0.7 ps, which is faster than the characteristic time for diffusive motion in 32,33 water, viz. 0.8 ps. Thus, the evolution of CKA excited state during the first picoseconds after the optical excitation represents a combination of solvent, vibrational and/or conformational relaxations leading to the equilibration of the excited molecule toward the relaxed excited state S1r. The SAS3(λ) resembles the TA spectrum observed after the dynamic Stokes shift, thus, this component should mainly be assigned to the S1r state decay. In aqueous solutions, the decay of the main TA band at ca. 580 nm is followed by a slower decay of the signal at ca. 460 nm (blue curve in Fig. 2C). This component with a time constant τ4 ≈ 1.5 ps was not observed in fluorescence measurements. The introduction of additional exponential function in the analysis of the fluorescence data did not enhance the resulting fit and did not show the presence of an extra component. Thus, this component observed in TA dynamics should be attributed to a non-emitting species, probably a hot ground state populated after the ultrafast

H2O

(A)

0.8

CKA H 2O EtOH DMSO

0.6 0.4 0.2 0.0 0

2

4

6

8

(B)

1.0

10

H2O D2O MeOH EtOH ACN DMF DMSO

0.8 0.6 0.4 0.2 0.0 0

2

4

6

8

10

Time delay / ps

Fig. 3. (A) Fluorescence and (B) TA time profiles recorded at the maxima of the corresponding bands for CKA in various solvents. Smooth lines: best fits from the global data analysis. Fluorescence time profile of KN in aqueous solution (from 9) is shown as gray curve in (A) for reference.

The fast decay of S1 states of CKA within the first tens of picoseconds after the excitation does not exclude the formation of a long-lived species with low yield. To study the TA dynamics of CKA excited states on a nanosecondmicrosecond time scale, nanosecond Laser Flash Photolysis (LFP) experiments were carried out. Varying the energy of laser pulses at 355 nm in the range 5-150 mJ, no TA signals were detected in the 250-650 nm spectral region for CKA in neither aqueous solution nor organic solvents. It should be noted that the high intensities of the laser radiation might result in the biphotonic ionization of solute molecule with the absorption of a second photon by the molecule in the triplet state, as was 39 observed for KN and 3-hydroxy-kynurenine. The absence of signals from solvated electron or any other long-lived species

4 | J. Name., 2012, 00, 1-3




Journal Name

Absorption


ARTICLE

Page 5 of 13


DOI: 10.1039/C8CP02183J

ARTICLE


indicates a negligible yield of CKA triplet states or radicals. Thus, LFP results unambiguously confirm that the decay of S1 state of CKA occurs without formation of a long-lived species. Influence of intermolecular H-bonds on the rate of S1 → S0 transition To evidence the participation of intermolecular H-bonding in the decay process of S1 state of CKA, the excited state dynamics in water was compared to that in deuterated water. Indeed, if intermolecular H-bonds are involved in the decay of excited states, the substitution of H2O with D2O should cause an increase of the lifetime of the S1 state. Comparison of TA experiment results in H2O and D2O shows an increase of all τi values by a factor of 1.4 (see Table 1), confirming the influence of intermolecular H-bonds on the whole temporal evolution of CKA S1 state. We also inspected the effect of the addition of small quantities of water to organic solvents on the observed dynamics of excited state decay. The fluorescence and TA dynamics were recorded with CKA in D2O, EtOH and DMSO in the presence of 6% of H2O by volume (H2O mole fractions were 0.06, 0.17 and 0.28 for H2O / D2O, H2O / EtOH and H2O / DMSO mixtures, respectively). The obtained data (Table 1) showed a moderate influence of water on the rate of S1 state decay, ca. 10-20% even for the case of H2O / DMSO when the ratio of molar concentration is close to 1 / 4. Thus, intermolecular H-bonds accelerate the S1 → S0 transition, but do not play a crucial role in the deactivation of CKA excited states. This observation does not exclude a contribution of HBIND, which is also not very efficient in the presence of small quantities of H-bond donating solvent due to the absence of H9 bond network as it was previously shown for KN and other 29 HBIND molecules. Table 2. The time constants (τi, i = 2,3) obtained from a global fit of the TA data measured with CKA in solvents of various viscosities (λex = 400 nm).

Solvent DMSO Dimethyl phthalate Ethylene glycol DMSO / glycerol

η / cP (20°C) 2.2 19.1 19.8 50

τ2 / ps 2.1 0.84 1.5 1.4

τ3 / ps 9.9 7.4 5.3 7.4

Influence of viscosity on the rate of the S1 → S0 transition The observed ultrafast decay of CKA excited state may be associated with the large amplitude motion of some groups, for instance the chain moiety, in the S1 state. If so, the rate of the S1 → S0 transition should depend on the solvent viscosity. To check this, we recorded the TA dynamics of CKA in several solvents of different viscosity at 20°C, varying from 2 (DMSO) to 50 cP (DMSO/glycerol). No remarkable differences in the TA dynamics were observed as compared to other organic solvents (Fig. S4†). The analysis of the TA data was performed assuming two successive exponential steps (B → C → 0) similarly to the aforementioned procedure (see section TA dynamics). The resulting time constants (Table 2) do not show

an influence of viscosity on the lifetime of S1 state of CKA (τ3). Thus, no large amplitude motions of CKA bulky groups are involved in the decay of S1 state. In other words, the S1 → S0 transition in CKA is an intramolecular process proceeding without significant displacement of solvent molecules. Ground state geometry To go deeper in the mechanism of the ultrafast excited state decay of CKA, sets of computational experiments were carried out. The first purpose of the computational investigation was to determine the geometry of the different forms and isomers of CKA in the ground state (see Scheme 1 for structures), along with their energies, in the gas and solvated phases (H2O, EtOH, DMSO, see Experimental section in ESI† for computational details). Solvent effects were taken into account in two ways: (1) by means of the implicit Polarizable Continuum Model in its Integral Equation 40 Formalism (IEF-PCM), and (2) by means of explicit microsolvation models, wherein one or more solvent molecules were explicitly placed close to the CKA molecule, and the whole system was treated within IEF-PCM to include bulk solvation effects. In the cases of explicit micro-solvation, the number of solvent molecules were 1 (for DMSO, EtOH, and H2O), 2 (for H2O) or 3 (for H2O), placed – after analysis of molecular electrostatic potentials – according to chemical – common sense, i.e. close to the –NH2, >C=O, and –COO groups (see Fig. S6†). O OH N H

H

trans-CKAH, no intramolecular H-bond

O O 5

O

N H

O H

cis-CKAH

O-

O

OH

4 3

OH 2 1

N 2 H

O 1 H

trans-CKAH

N H

O H

trans-CKA

Scheme 1. Chemical structures of various form of CKA.

The obtained results demonstrate that CKA in solution is practically only present in the deprotonated form (Table S1†). − Both CKAH and CKA shows that in the gas phase and solution, (cis-trans) trans isomer is largely more stable than cis form (ΔE ~ 1.2-2.7 eV; Table S2†). Independently of the environment (gas phase, aprotic or protic solvent), a rotamer with an intramolecular H-bond between the amino and carbonyl group (Scheme 1, Fig. 4 and S6†) has always lower energy than the rotamers without intramolecular H-bond. Furthermore, a high electronic energy barrier between these rotamers (e.g., in case ‡ of aqueous solution, ΔE ~ 2.8 eV, see Fig. S8†) impedes the population of rotamers without intramolecular H-bond at the room temperature. In the followings, only results of the − deprotonated form with intramolecular H-bond, trans-CKA will, therefore, be shown and discussed. In the ground state, CKA shows a nearly flat structure (Fig. 4, left) with a small deflection of the carbonyl group from the

J. Name., 2018, 00, 1-3 | 5




Journal Name



DOI: 10.1039/C8CP02183J

Journal Name


(1) (2) (3)

plane of the aromatic ring, δ(C C C O) ≈ 7°, and the dihedral angle between carbonyl group and unsaturated bond, (5) (4) (3) δ(C C C O) ≈ 14° (see Scheme 1 for atom numeration); solvent does not significantly affects the geometry of the ground state (Table S4†). The length of intramolecular H-bond is 1.90 Å that indicates relatively strong H-bond. The presence of the intramolecular H-bond is confirmed by the Reduced Density Analysis (vide postea). Upon photoexcitation, the first electronic transition S0 → S1 has ππ* character, which involves a significant transfer of electron density from the amino group to the carbonyl group and unsaturated bond, resulting in a pronounced ICT character of this transition, as illustrated in Fig. 5.

population of the ICT (geometrically distorted) minimum of S1 PES is not related with H-bonding or specific solvent effects, but it is merely due to the high polarity of the solvent. In the gas phase, which is a reasonable approximation for low polarity solvents, like n-hexane, the geometry of the LE minimum is close to that of the optimized ground state. This remarkable change in the S1r state geometry is an element of peculiarity of CKA with respect to KN, for which no significant changes were observed after S1 relaxation in the solvated 25 phase. In solution, the optimized S1r state minimum is characterized by tremendous migration of charge from the aromatic system to the chain moiety, see Fig. 5; more details can be found in Fig. S9†. These configuration changes on the way from the Franck-Condon region to ICT S1r minimum are induced by the changes in the electron density distribution, and may be interpreted as the splitting of one conjugated system into two smaller conjugated moieties. In the case of explicit micro-solvation with water molecule placed near the carboxyl group, the distance between the explicit water molecule and COO− group in the S1r state increases as compared with S0 (Fig. S6, S7†). This points to the weakening of the H-bond interactions for –COO− in the ICT S1r state and the decrease of hydrophilicity of this group. Most likely, this effect may be attributed to the ICT-induced conjugation of the amino acid backbone.

Fig. 4. Two different views of the (TD-)DFT optimized geometries of trans CKA− in aqueous solution (IEF-PCM model) of the ground state (S0, left) and the relaxed singlet excited state (S1r, right). Legend of colors: white (hydrogen), grey (carbon), blue (nitrogen) and red (oxygen). The lengths of the relevant H-bonds are indicated in Å.

First singlet excited state geometry

In order to investigate the photophysical properties of CKA, we investigated the S1(ππ*) potential energy hypersurface (PES), and in particular its minima that may be populated upon photoexcitation and relaxation. Therefore, we optimized the geometry of CKA S1r minimum both in the gas and solvated phases (H2O, EtOH and DMSO). In the gas phase, the S1 electronic excited state relaxes towards a local minimum, close to the Frank-Condon region, which has locally excited (LE) character. As a result, the CKA geometry exhibits insignificant changes, among which a planarization of structure including carbonyl group and (5) (4) (3) unsaturated bond, δ(C C C O) ≈ 0°, should be mentioned (data not shown). On the contrary, solvent effects (even within the framework of implicit IEF-PCM) favors the population of the ICT minimum, which is structurally characterized by a large twist of the carbonyl group from the aromatic plane, (1) (2) (3) δ(C C C O) ≈ 60°, and, again, planarization of the structure, (5) (4) (3) which includes C=O and C=C bonds, δ(C C C O) ≈ 3°. As examples, two projections of molecular structure of CKA in water in the S0 and S1r states are shown in Fig. 4; some geometrical parameters are summarized in Table S4.† These changes were observed in all solvents used (see Fig. S7† and Table S4† for details), and we can therefore conclude that the

Fig. 5. Frontier molecular orbitals (HOMO and LUMO, top and bottom, respectively) for the optimized geometry of the ground (S0, left) and the relaxed singlet excited states (S1r, right) of trans CKA− in aqueous solution (IEF-PCM model).

Conical intersection between the ground and excited states

Comparing with KN,9,25 CKA exhibits significant topological changes upon excitation (first of all, the large twist of the carbonyl group, see above), and the remarkable decrease of the S1 state lifetime in all used solvents. Therefore, the acceleration of the S1 → S0 transition may be presumed to be associated with the topological changes in CKA upon excitation.

6 | J. Name., 2012, 00, 1-3




ARTICLE

Page 7 of 13


DOI: 10.1039/C8CP02183J

ARTICLE

In order to verify this hypothesis, we performed calculations at Complete Active Space Self-Consistent Field (CASSCF) level of theory for S0 and S1 states of CKA in aqueous solution and in DMSO. Fig. 6 shows the PES profiles along Linearly Interpolated Internal Reaction Coordinate (LIIRC) corresponding to the twist of the carbonyl group and the subsequent reorganization of the chain moiety with unsaturated bond for both solvents. As depicted in Fig. 6, a CI between the PES’ of S0 and S1 is clearly found in both water

and DMSO. This indicates that the twist of the carbonyl group is the probable channel of the ultrafast decay of the S1 state of CKA. Although the considered solvents are both polar, the excited PES profiles are sensitive to the environment: the path to CI is barrierless in aqueous solution, whereas it shows a barrier (~ 0.17 eV) in DMSO solution. This finding is in agreement with the experimentally observed increase of the S1 state lifetime in organic solvents as compared to water.

Fig. 6. PES profiles (in eV) of electronic ground state S0 (filled circles) and the lowest excited state S1 of trans CKA− (empty circles) in water (left) and DMSO (right), along the linearly interpolated reaction path from the ground state equilibrium geometry to the conical intersection. Solvent effects are included through explicit micro-solvation, consisting of two and one solvent molecules for water and DMSO, respectively, further surrounded by implicit solvent (IEFPCM). Insets depict initial (optimized ground state) and final geometries (at the conical intersection).

Time-resolved dynamics of S1 state population and geometry

The observed S1/S0 CI explains the ultrafast S1 → S0 transition in aprotic solvents, like DMSO. However, the shortening of the S1 state lifetime of CKA in protic solvents may originate from two channels: (a) the deactivation of S1r 9,13 state via HBIND mechanism as it was observed for KNs and 28 other organic molecules or (b) the acceleration of the structural relaxation of CKA due to solvent-solute interactions, including intermolecular H-bonds. The aforementioned calculations cannot discriminate the contributions of different channels. We, therefore, performed dynamic calculations, 41 using a modified variant of the method of Persico et al. for 42 semiclassical Surface-Hopping (SH) dynamics; see Calculation part in ESI† for details. The SH approach proposes a combination of Quantum Mechanical (QM) and Molecular Mechanical (MM) methods to describe the evolution of solute (CKA) at QM level surrounded by explicit solvent molecules (H2O, EtOH and DMSO) treated at MM level. The time evolution of the S1 state population of CKA embedded in a drop of solvent was SH simulated for 1200 sets of different initial conditions with 10 femtoseconds time steps. The goodness of the level of theory used in these calculations is confirmed by the computed CKA absorption spectra, averaged along the equilibration dynamics, see Fig. 1B. The very good agreement between experimental and calculated UV-Vis spectra (Fig. 1) support the correctness of the semiempirical Hamiltonian parameterization for the description of the CKA excited state.

Fig. 7A shows the simulated time evolution of the S1 state population averaged over 1200 SH trajectories. Comparison with Fig. 3 shows good qualitative agreement between the calculations and the experimental data, i.e. the acceleration of S1 state population decay from DMSO to water. The averaged time profiles were fitted with mono exponential functions; the c c resulting time constants, τ , are listed in Table 3. The τ value in aqueous solution agrees extremely well with the experimentally observed τ3 value, whereas in the cases of organic solvents the computed values are 2-3 times shorter than the experimental ones (Table 1). These differences may originate from the parameterization of solute-solvent interactions, since they are responsible for the dissipation of electronic excitation energy into vibrational modes of the solvent molecules (thermal reservoir). One advantage of the SH dynamics simulations is the possibility to follow the changes of molecular geometry along the photophysical process. Fig. 7B shows the time profile for (1) (2) (3) the dihedral angle, δ(C C C O), describing the twist of the carbonyl group. The rate of this twist is solvent dependent, accelerating from aprotic DMSO to protic water, and coincides with the decay of S1 state population (Fig. 7A). The opposite (1) (2) (3) (4) picture is observed for the dihedral angle δ(C C C C ), (4) demonstrating the twist of the C atom of chain moiety from the plane of aromatic ring (Fig. S10A†). The formation of alternative π-system, consisting of C=O and C=C groups (5) (4) (3) (δ(C C C O), Fig. 7C), is solvent independent and significantly slower than the twist of the carbonyl group. Therefore, the planarization of these bonds follows the main

J. Name., 2018, 00, 1-3 | 7





Journal Name



DOI: 10.1039/C8CP02183J

Journal Name

change, viz. the twist of the carbonyl group, and completes the reorganization of CKA geometry in the S1r state before the S1r → S0 transition. These results indicate that the acceleration of carbonyl group twist in protic solvents is responsible for the observed decrease of S1 state lifetime of CKA from aprotic to protic solvents.

S1 state population

(A) 0.8 0.6

Table 3. Calculated fluorescence quantum yield (Φcf), time constants (τc) and rate constants of radiative (kr) and non-radiative (knr) decay obtained from SH simulations of the excited-state dynamics of CKA in various solvents.

H2O EtOH DMSO

0.4 0.2 0.0 0.0

2.5

5.0

7.5

10.0

12.5

15.0

70

50 40 30

(B)

20 10

(3) (5) (4)

Φcf × 10-5 1.32 2.03 2.95

τc / ps 0.71 2.0 3.0

kr / 107 s-1 1.9 1.0 1.0

knr / 1012 s-1 1.41 0.50 0.33

Discussion

60

20

Solvent H2O EtOH DMSO

Time / ps

(1)

(2)

(3)

δ(C C C O) / degs

80

δ(C C C O) / degs


1.0

most part of conventional spectrofluorometers. This additionally supports that even very low concentration of impurities could completely cover the signal from CKA. The c radiative rate constants (kr), estimated as kr = Φf / τ , are 7 -1 comparable to those of KN and its derivatives, (1-5)×10 s , 9,13,22 depending negligibly on the solvent. The nonradiative c rate constants, knr = (1 – Φf) / τ , demonstrate extremely high values due to the S1/S0 CI.

0.0

2.5

5.0

7.5

10.0

12.5

Time / ps

15.0

(C)

15

10

5

0 0.0

2.5

5.0

7.5 Time / ps

10.0

12.5

15.0

Fig. 7. Time profiles of (A) the excited state S1 population (including two-exponential fits, dashed lines), (B) the dihedral angle δ(C(1)C(2)C(3)O), (C) the dihedral angle δ(C(5)C(4)C(3)O) of CKA in water, EtOH and DMSO.

Additionally, we inspected the distances between the N, (1) H and O atoms forming the intramolecular H-bond, i.e. (1) between nitrogen, proton H of the amino group and the (1) carbonyl oxygen (Fig. S10B†). The length of the bond N-H in the excited state does practically not change, whereas the (1) distance between O and H increases significantly (by a factor of 1.4) due to the twist of the carbonyl group in all solvents used (Fig. 10B and Table S4†). Therefore, structural relaxation to the S1r state lead to noticeable weakening of the intramolecular H-bond between amino and carbonyl groups. The fluorescence quantum yields (Φf) for CKA in the three solvents were also computed from the SH dynamics (Table 3). -5 The theory predicts extremely low values Φf = (1-3)×10 , which are at the border of the sensitivity threshold for the

In previous works, we studied the photophysical and photochemical properties of KN.7-9,13,18,25,39 The major decay channel of KN in the singlet-excited state in aqueous solution is HBIND, transforming the electronic energy into the heat via the H-bonds between the solute and the solvent network.9,13 Intermolecular H-bonds play a key role in the deactivation of photoexcited KN: the S1 state lifetime strongly depends on the solvent proticity; it increases from approximately 30 ps in water to 250-500 ps in alcohols reaching the value of 2.3 ns in DMSO.9 In a recent theoretical study of KN photophysics by Tuna et al.,43 an intramolecular proton transfer between the amino and carbonyl groups was suggested as the origin of the fast excited state decay of KN in aqueous solution. However, this mechanism does not explain the drastic increase of the S1 state lifetime in aprotic solvents, observed for both KN9 and its chromophoric core, o-aminoacetophenone.26 Therefore, the suggested intramolecular H-transfer cannot be the major mechanism for the S1 → S0 transition in KN. The S1 state decay of CKA is much faster than that of KN in a broad range of both protic and aprotic solvents: it varies from 0.7 ps in water to 9 ps in DMSO. In other words, the S1 state decay of CKA is faster than that of KN by a factor of ~ 40 in aqueous solution and by more than 2000 in DMSO. As it follows from the experimental data (Table 1), the solvent proticity affects S1 → S0 transition dynamics of CKA, but this transition remains very fast even in aprotic solvents, in the absence of intermolecular H-bonds. Therefore, HBIND cannot be the general mechanism responsible for S1 → S0 decay. Moreover, HBIND should probably not contribute significantly to the excited-state decay of CKA in protic solvents because the excited-state lifetimes of CKA (see Table 1) are substantially faster than those of KNs in protic solvents reported before, 10-200 ps.9,10,13,22 The excited state decay of CKA does not depend on the solvent viscosity (Table 2). Thus, the effective deactivation of CKA excited state should be attributed to volume-preserving intramolecular changes, i.e. changes requiring the

8 | J. Name., 2012, 00, 1-3




ARTICLE

Page 9 of 13


DOI: 10.1039/C8CP02183J

ARTICLE

displacement of a negligibly small solvent volume, occurring on the way from the Franck-Condon region to the ICT minimum of the S1 state. In general, at least two intramolecular mechanisms may be proposed to account for the fast S1 → S0 transition. The first mechanism is the decay of S1 state via a CI accessible via structural changes caused by ICT upon excitation. The second mechanism involves a partial or complete intramolecular proton transfer from the amino to the carbonyl group of CKA. The resulting species might be highly unstable, and readily decay via back proton transfer. However, an intermolecular H-bonding with solvent molecules should impede the intramolecular proton transfer: a deceleration of the S1 → S0 transition and an increase of the S1 lifetime could be expected in protic solvents. This contradicts the experimental observations: the CKA S1 lifetime in H2O is shorter than that in DMSO by an order of magnitude (Table 1). Thus, the experimental data do not support the hypothesis of the intramolecular proton transfer.

Fig. 8. PES profiles (in eV) of electronic ground state S0 (filled squares) and the Franck-Condon excited states S1 and S2 (filled circles and triangles, respectively) of trans CKA− in DMSO, along the normalized reaction coordinate for H-atom transfer from the nitrogen of amino group to the oxygen of carbonyl group. Solvent effects are accounted by implicit solvent (IEF-PCM). Insets depict initial and final geometries.

To check for the feasibility of an H-atom transfer mechanism, we did additional calculations for the FranckCondon (FC) and the relaxed S1 states of CKA. The energies of the S0 state and of the S1 and S2 FC states were computed as a function of the H-atom position between the amino nitrogen atom and the carbonyl oxygen atom of CKA in H2O and DMSO using the IEF-PCM approximation for the solvents. The obtained results (Fig. 8 and Table S5†) clearly show the existence of large barriers for the proton transfer from nitrogen to oxygen atoms for both ground and excited states (ap. 5 eV for S0 state and 3 eV for the S1 and S2 FC states) in both aprotic and protic solvents. Though ICT probably decreases the barrier in the FC excited S1 and S2 states, the proton transfer remains energetically unfavorable in the Franck-Condon S1 state of CKA in both aprotic and protic solvents.

The intra- and inter-molecular H-bond interactions the in S0 and S1r states of CKA were also studied in different solvents by 44 means of the Bader’s Atoms in Molecules (AIM) and Reduced 45 Density Gradient (RDG) theories, see ESI† for details. In particular, these methods were recently extended and applied to study the attractive and repulsive intramolecular interactions in the excited states, in particular H-bond 46 interactions. The classical approach, the AIM theory, is not ideal for investigating and visualizing non-covalent interactions (NCIs). Therefore, we additionally apply the NCI Index as a more general description of intramolecular interactions based on density changes instead of its local values, as was done in the AIM method. The RDG value demonstrates the probability of NCI formation. The attractive/repulsive character of this interaction may be identified by the sign of the RDG second derivative along the second main axis of variation λ2 (eigenvalue), which is negative for attractive interactions and positive for repulsive ones. This approach has been developed to reveal non-covalent interactions (e.g., hydrogen bonding) in 3D space. Regions where the electron density and RDG are low correspond to regions where NCIs occur. Isosurfaces of the RDG at low densities can be used to visualize the position and nature of NCIs. NCI isosurfaces, therefore, illustrate the nature of the intramolecular interactions. A continuous color-coding scheme based on the λ2is used to provide qualitative information about the interaction strength. For topological analysis, we used geometries and electron densities of CKA including explicit microsolvation, previously optimized for S0 and S1r states at TD-DFT level of theory (Fig. S6 and S7†). Part of the results of calculations for CKA in H2O, EtOH, and DMSO are presented in Fig. 9; other examples are given in Fig. S11.† In the ground S0 state, the attraction between amino group and carbonyl oxygen is dominating over the repulsion, leading to the formation of intramolecular Hbond. However, in the relaxed excited state, these interactions become weaker and almost equal. Thus, this topological analysis demonstrates that the intramolecular H-bond between the amino group and the carbonyl oxygen present in S0 state practically disappears in S1r state. Therefore, both the experimental data and calculation analysis of interactions in excited states indicate that the intramolecular proton transfer in the excited state is not responsible for the ultrafast S1 decay. Theoretical calculations speak in favor of the CI as the mechanism for the ultrafast S1 → S0 transition in CKA. This mechanism involves the twist of the carbonyl group, causing the increase of the angle δ(C(1)C(2)C(3)O) from 7° (in S0 state) to 60° (in S1r state), see Fig. 4 (IEF-PCM model). This ICT-induced reorganization of CKA geometry leads to the intersection of PES of S1r and S0 states (Fig. 6). The driving force of this C=O twist is the large ICT-induced change in the dipole moment of CKA in the excited state:47 the dipole moments in the ground state and the Franck-Condon and relaxed excited states are μg = 7.9 D, μFC = 10.9 D, μr = 6.6 D, respectively (solvent was accounted by IEF-PCM model; the obtained values are almost independent on the solvent). The measured τ3 value of CKA in water is close to those reported for excited molecules 12,35,36 decaying via a CI accessible without a significant barrier.

J. Name., 2018, 00, 1-3 | 9





Journal Name



DOI: 10.1039/C8CP02183J

Journal Name

According to the calculations, access to the CI involves a small barrier in DMSO but not in water, where it is barrierless (Fig. 6). This is in a good agreement with the experimental findings: the S1 state lifetime in DMSO is greater by order of magnitude than in H2O (Table 1). Very likely, the barrier is associated with the intramolecular H-bond between the amino group and the carbonyl oxygen, which hampers the twist of the carbonyl group. In protic solvents the intramolecular Hbond interactions in the excited state are probably compensated by intermolecular ones, which result in a decrease (in alcohols) or even the disappearance (in water) of

the barrier. Thus, most probably, intermolecular H-bonds reduce the barrier on the way from the Franck-Condon region to the CI. The rate of the S1 → S0 transition in CKA does not depend on solvent viscosity (Table 2). The change of CKA geometry probably occurs without significant motion or displacement of solvent molecules. Such negligible viscosity dependence was already observed in several cases where the excited-state decay involves a CI reached through significant structural 36,38,48 changes.

Fig. 9. NCI isosurfaces (s = 0.5 a.u. and a blue-green-red color scale from -0.01 a.u. < sign (λ2) ρ(r) < +0.01 a.u.) of the S0 and relaxed S1 states of CKA in different solvents. Colour code for the NCI isosurfaces: red (strongly attractive), yellow-green (weakly attractive), green-turquoise (weakly repulsive), and blue (strongly repulsive).

From a photophysical point of view, CKA is an ideal molecular UV filter providing very efficient excited state decay in a broad range of solvents without formation of reactive species. Therefore, this molecule could be attractive for practical applications, such as sunscreen formulations. Unfortunately, the thermal instability of CKA prevents its direct use: under physiological conditions it can undergo covalent 30,49 attachment (via Michael addition) to nucleophilic groups. 30 In the absence of nucleophiles, CKA cyclizes slowly with the 21-24 formation of several photochemically active products. Therefore, the thermal stability of CKA should be significantly improved prior to its practical application.

The results of the present work show that the replacement of saturated bond with the unsaturated C=C one in KN molecule results in significant increase of the ICT character of the S0 → S1 transition and in the switch of the IC mechanism from a solvent-assisted deactivation to a deactivation via a S1/S0 CI. This leads to the drastic decrease of the S1 state lifetime. One of the most frequent cases of CIs, described in 35,50,51 the literature, is the E → Z (or Z → E) isomerization, which includes the rotation/twist around a single or double bond of a molecule. This mechanism was reported for many 35,51 natural and artificial UV screens as well as for green 50,52 fluorescent proteins and retinal proteins. Though the twist

10 | J. Name., 2012, 00, 1-3





ARTICLE

Page 11 of 13


DOI: 10.1039/C8CP02183J

ARTICLE

of the carbonyl group is the crucial step on the way to the CI, the mechanism operating in CKA could not be described as the typical E → Z isomerization. Indeed, the twist of the carbonyl group does not lead to completely orthogonal orientation of the two moieties (Fig. 4 and Table S4†), and isomerization does not occur. This conclusion is supported by the superior photostability of CKA (Fig. S2†). Thus, the mechanism reported in this work could be rationalized in terms of ICT-induced “flipping” of two moieties in the excited state leading to the fast transition S1 → S0 via CI with the restoration of CKA ground state in its initial conformation.

1 2 3 4 5 6 7

Conclusions

8

The combined use of several experimental and computational techniques allowed disclosing the mechanism of the extremely fast S1 → S0 transition in the deaminated KN in solution. The presence of an unsaturated C=C bond in the chain moiety results in a large structural rearrangement of CKA in the excited state, consisting of the twist of the carbonyl group and subsequent planarization of the backbone. These ICT-induced changes in the geometry of CKA in the excited state result in the split of the extended conjugated π-systems in two smaller moieties that, in turn, moves CKA excited state to a CI between excited and ground states PES. As a result, rapid decay of CKA excited states takes place independently on the solvent ability to form intermolecular H-bonds. This substantial improvement of the photostablity of kynurenine-based UV filters could be exploited for design and synthesis of new efficient molecular UV filters.

9

Acknowledgements

14

This work was supported by Fonds National Suisse de la Recherche Scientifique (200020-165890) and University of Geneva in femtosecond fluorescence and transient absorption measurements, by Research Council grant of the President of Russian Federation (project MK-1515.2017.3) in nanosecond transient absorption measurements and by Siberian Branch of Russian Academy of Science (Program “Integration and development” II.1, project 0333-2018-0009) in photostability measurements. E.B. acknowledges Russian Foundation for Basic Research for mobility grant (14-33-50581). Authors are grateful to Dr. Jakob Grilj for the assistance with femtosecond TA measurements and Siberian Supercomputer Center of Institute of Computational Mathematics and Mathematical Geophysics of Siberian Branch of the Russian Academy of Sciences for offering computational resources and technical assistance.

10 11

12

13

15 16 17

18 19 20 21 22 23

Notes The authors declare no competing financial interest.

24 25

References

J. D. Simon, D. Peles, K. Wakamatsu, S. Ito, Pigm. Cell Mel. Res. 2009, 22, 563−579. A. Corani, A. Huijser, T. Gustavsson, D. Markovitsi, P.-Å. Malmqvist, A. Pezzella, M. d’Ischia, V. Sundström, J. Am. Chem. Soc. 2014, 136, 11626–11635. L. A. Baker, S. E. Greenough V. G. Stavros, J. Phys. Chem. Lett. 2016, 7, 4655–4665. R. van Heyningen, Nature 1971, 230, 393–394. A. M. Wood, R. J. W. Truscott, Exp. Eye Res. 1993, 56, 317– 325;. Yu. P. Tsentalovich, T. D. Verkhovod, V. V. Yanshole, A. S. Kiryutin, L. V. Yanshole, A. Z. Fursova, D. A. Stepakov, V. P. Novoselov, R. Z. Sagdeev, Exp. Eye Res. 2015, 134, 15–23. Yu. P. Tsentalovich, O. A. Snytnikova, P. S. Sherin, M. D. E. Forbes, J. Phys.Chem. A 2005, 109, 3565-3568. P. S. Sherin, Yu. P. Tsentalovich, O. A. Snytnikova, R. Z. Sagdeev, J. Photochem. Photobiol. B: Biol. 2008, 93, 127-132. P. S. Sherin, J. Grilj, Yu. P. Tsentalovich, E. Vauthey, J. Phys. Chem. B 2009, 113, 4953–4962. Yu. P. Tsentalovich, P. S. Sherin, L. V. Kopylova, I. V. Cherepanov, J. Grilj, E. Vauthey, Invest. Ophthalmol. Visual Sci. 2011, 10, 7687−7696. (a) W. Domcke, D. R. Yarkony, H. Köppel, H. World Scientific: London, 2004; Vol. 15. (b) W. Domcke, D. R. Yarkony, H. Köppel, World Scientific: London, 2011; Vol. 17. (c) D. R. Yarkony, Chem. Rev. 2012, 112, 481−498. (a) R. V. Bensasson, E. J. Land, T. G. Truscott, Oxford University Press: Oxford, 1993. (b) C. E. Crespo-Hernandez, B. Cohen, P. M. Hare, B. Kohler, Chem. Rev. 2004, 104, 1977– 2020. (c) T. Gustavsson, Á. Bányász, E. Lazzarotto, D. Markovitsi, G. Scalmani, M. J. Frisch, V. Barone, R. Improta, J. Am. Chem. Soc. 2006, 128, 607–619. (d) T. Gustavsson, R. Improta, D. Markovitsi, J. Phys. Chem. Lett. 2010, 1, 2025– 2030. (e) B. Kohler, J. Phys. Chem. Lett. 2010, 1, 2047–2053. (f) R. Improta, F. Santoro, L. Blancafort, Chem. Rev. 2016, 116, 3540–3593. P. S. Sherin, J. Grilj, L. V. Kopylova, V. V. Yanshole, Yu. P. Tsentalovich, E. Vauthey, J. Phys. Chem. B 2010, 114, 11909. S. Premi, S. Wallisch, C. M. Mano, A. B. Weiner, A. Bacchiocchi, K. Wakamatsu, E. J. H. Bechara, R. Halaban, T. Douki, D. E. Brash, Science 2015, 347, 842–847. S. Beissert, K. Loser, K. Photochem. Photobiol., 2008, 84, 29– 34. A. Korlimbinis, J. A. Aquilina, R. J. W. Truscott, Exp. Eye Res. 2007, 85, 219–225. (a) N. R. Parker, J. F. Jamie, M. J. Davies, R. J. W. Truscott, Free Rad. Biol. Med. 2004, 37, 1479–1489. (b) J. Mizdrak, P. G. Hains, R. J. W. Truscott, J. F. Jamie, M. J. Davies, M.J. Free Rad. Biol. Med. 2008, 44, 1108–1119. O. A. Snytnikova, P. S. Sherin, L. V. Kopylova, Yu. P. Tsentalovich, Rus. Chem. Bull. 2007, 56, 732–738. J. Mizdrak, P. G. Hains, D. Kalinowski, R. J. W. Truscott, M. J. Davies, J. F. Jamie, Tetrahedron 2007, 63, 4990–4999. Yu. P. Tsentalovich, O. A. Snytnikova, M. D. E. Forbes, E. I. Chernyak, S. V. Morozov, Exp. Eye Res. 2006, 83, 1439–1445. P. S. Sherin, N. P. Gritsan, Yu. P. Tsentalovich, Photochem. Photobiol. Sci. 2009, 8, 1550–1557. E. A. Zelentsova, P. S. Sherin, O. A. Snytnikova, R. Kaptein, E. Vauthey, Yu. P. Tsentalovich, Photochem. Photobiol. Sci. 2013, 12, 546−558. P. S. Sherin, E. A. Zelentsova, E. D. Sormacheva, V. V. Yanshole, T. G. Duzhak, Yu. P. Tsentalovich, Yu. P. Phys. Chem. Chem. Phys. 2016, 18, 8827–8839. E. D. Sormacheva, P. S. Sherin, Yu. P. Tsentalovich, Free Rad. Biol. Med. 2017, 113, 372-384. E. Benassi, P. S. Sherin, Int. J. Quantum Chem. 2011, 111, 3799−3804.

J. Name., 2018, 00, 1-3 | 11





Journal Name



DOI: 10.1039/C8CP02183J

Journal Name

26 T. Yoshihara, H. Shimada, H. Shizuka, S. Tobita, Phys. Chem. Chem. Phys. 2001, 3, 4972–4978. 27 H. Shimada, A. Nakamura, T. Yoshihara, S. Tobita, Photochem. Photobiol. Sci. 2005, 4, 367–375. 28 (a) S. R. Flom, P. F. Barbara, J. Phys. Chem. 1985, 89, 4489– 4494. (b) T. Yatsuhashi, H. Inoue, J. Phys. Chem. A 1997, 101, 8166–8173. (c) P. Fita, M. Fedoseeva, E. Vauthey, J. Phys. Chem. A 2011, 115, 2465–2470. (d) A. Wallabregue, P. Sherin, J. Guin, C. Besnard, E. Vauthey, J. Lacour, Eur. J. Org. Chem. 2014, 29, 6431–6438. (e) S. Richert, S. M. Vasquez, M. Grzybowski, D. T. Gryko, A. Kyrychenko, E. Vauthey, J. Phys. Chem. 2014, 118, 9952–9963. 29 B. Dereka, E. Vauthey, Chem. Sci., 2017, 8, 5057-5066. 30 (a) L. M. Taylor, J. A. Aquilina, J. F. Jamie, R. J. W. Truscott, Exp. Eye Res. 2002, 75, 165–175. (b) L. V. Kopylova, O. A. Snytnikova, E. I. Chernyak, S. V. Morozov, Yu. P. Tsentalovich, Exp. Eye Res. 2007, 85, 242–249. 31 M. L. Horng, J. A. Gardecki, A. Papazyan, M. Maroncelli, J. Phys. Chem. 1995, 99, 17311–17337. 32 W. Jarzeba, G. C. Walker, A. E. Johnson, M. A. Kahlow, P. F. Barbara, J. Phys. Chem. 1988, 92, 7039–7041. 33 R. Jimenez, G. R. Fleming, P. V. Kumar, M. Maroncelli, Nature 1994, 369, 471–473. 34 S. A. Kovalenko, R. Schanz, H. Hennig, N. P. Ernsting, J. Chem. Phys. 2001, 115, 3256−3274. 35 T. Kumpulainen, B. Lang, A. Rosspeintner, E. Vauthey, Chem. Rev. 2017, 117, 10826–10939. 36 C. Nançoz, G. Licari, J. S. Beckwith, M. Soederberg, B. Dereka, A. Rosspeintner, O. Yushchenko, R. Letrun, S. Richert, B. Lang, E. Vauthey, Phys. Chem. Chem. Phys. 2018, 20, 7254– 7264. 37 T. Häupl, R. Lomoth, L. Hammarström, J. Phys. Chem. A 2003, 107, 435−438. 38 J. Grilj, E. N. Laricheva, M. Olivucci, E. Vauthey, Angew. Chem. Int. Ed. 2011, 50, 4496−4498. 39 O. A. Snytnikova, P. S. Sherin, Yu. P. Tsentalovich, J. Photochem. Photobiol. A: Chem. 2007, 186, 364–368. 40 J. Tomasi, B. Mennucci, E. J. Cancès, Mol. Struct.: Theochem. 1999, 464, 211-226. 41 (a) G. Granucci, M. Persico, A. Toniolo, J. Chem. Phys. 2001, 114, 10608. (b) C. Ciminelli, G. Granucci, M. Persico, Chem. Eur. J. 2004, 10, 2327-2341. 42 J. C. Tully, J. Chem. Phys. 1990, 93, 1061. 43 D. Tuna, N. Došlić, M. Mališ, A. L. Sobolewski, W. Domcke, J. Phys. Chem. B 2015, 119, 2112–2124. 44 (a) R. F. W. Bader, H. Essen, J. Chem. Phys. 1983, 80, 19431960; (b) R. F. W. Bader, Oxford: Oxford University Press; 1990. (c) R. F. W. Bader, Chem. Rev. 1991, 91, 893-892. 45 (a) H. J. Bohórquez, C. F. Matta, R. J. Boyd, Int. J. Quant. Chem. 2010, 110, 2418-2425; (b) E. R. Johnson, S. Keinan, P. Mori-Sánchez, J. Contreras-Garcia, A. J. Cohen, W. Yang, J. Am. Chem. Soc. 2010, 132, 6498-6506; (c) R. A. Boto, J. Contreras-Garcia, J. Tierny, J.-P. Piquemal, Mol. Phys. 2016, 114, 1406-1414. (d) J. Andres, S. Berski, J. Contreras-Garcia, P. Gonzalez-Navarrete, J. Phys. Chem. 2014, 118, 1663-1672. (e) Y. Zhang, H. He, K. Dong, S. Zhang, S. RSC Adv. 2017, 7, 12670-12682. 46 (a) I. Koskin, E. Mostovich, E. Benassi, M. Kazantsev, J. Phys Chem. C 2017, 121, 23359–23369. (b) K. I. Lugovik, A. K. Eltyshev, E. Benassi, N. P. Belskaya, Chem. Asian J. 2017, 12, 2410-2425. (c) K. I. Lugovik, A. V. Popova, A. K. Eltyshev, E. Benassi, N. P. Belskaya, Eur. J. Org. Chem., 2017, 28, 41754187. (d) A. I. Eliseeva, O. O. Nesterenko, P. A. Slepukhin, E. Benassi, N. P. Belskaya, J. Org. Chem. 2017, 82, 86–100. (e) K. D. Gavlik, E. S. Sukhorukova, Yu. M. Shafran, P. A. Slepukhin, E. Benassi, N. P. Belskaya, Dyes Pigm. 2017, 136, 229-242. (f) K. I. Lugovik, A. K. Eltyshev, P. O. Suntsova, P. A. Slepukhin, E. Benassi, N. P. Belskaya, Chem. Asian J., 2018, 13, 311-324.

47 The electric dipole moment for a charged species is illdefined since it depends on the choice of the coordinate origin. However, in Quantum and Computational Chemistry, a “working definition” is commonly assumed, which refers to the center of the nuclear charge of the system. The electric dipole moment for ground and excited states was computed at the (TD−)DFT level, employing the aforemenƟoned working definition. 48 K.L. Litvinenko, N. M. Webber, S. R. Meech, J. Phys. Chem. A 2003, 107, 2616-2623. 49 (a) B. D. Hood, B. Garner, R. J. W. Truscott, J. Biol. Chem. 1999, 274, 32547–32550. (b) A. Korlimbinis, R. J. W. Truscott, Biochemistry 2006, 45, 1950–1960. 50 S. Gozem, F. Melaccio, H. L. Luk, S. Rinaldi M. Olivucci, Chem. Soc. Rev., 2014, 43, 4019–4036. 51 L. A. Baker, B. Marchetti, T. N. V. Karsili, V. G. Stavros, M. N. R. Ashfold, Chem. Soc. Rev., 2017, 46, 3770–3791. 52 (a) R. S. Liu, A. E. Asato, Proc. Natl. Acad. Sci. U. S. A. 1985, 82, 259−263. (b) K. Addison, J. Conyard, T. Dixon, P. C. Bulman Page, K. M. Solntsev, S. R. Meech, J. Phys. Chem. Lett. 2012, 3, 2298−2302. (c) E. Buhl, M. Braun, A. Lakatos, C. Glaubitz, J. Wachtveitl, J. Biol. Chem. 2015, 396, 1109−1115.

12 | J. Name., 2012, 00, 1-3





ARTICLE

Page 13 of 13

Physical Chemistry Chemical Physics View Article Online

DOI: 10.1039/C8CP02183J

Graphic Abstract for

1.0

0.6

Energy

Fluorescence / a.u.


S1

H2O EtOH DMSO

0.8

hν

fluo

. twist of C=O . formation of two orthogonal

0.4 0.2

conjugated sub-structures

S0

0.0 0

5 10 15 20 Time delay / ps

Reaction coordinate

Conical intersection!

Unsaturated bond in the side chain confers ultrafast decay of excited states via conical intersection independently on solvent properties.


Ultrafast excited state decay of natural UV filters: from intermolecular hydrogen bonds to conical intersection