See discussions, stats, and author profiles for this publication at: https://www.researchgate.net/publication/11670039
Photophysical Characterization of Natural cisCarotenoids¶ Article in Photochemistry and Photobiology · October 2001 DOI: 10.1562/0031-8655(2001)0740549PCONCC2.0.CO2 · Source: PubMed
CITATIONS
READS
9
28
4 authors, including: Per Ola Andersson
Shinichi Takaichi
Swedish Defence Research Agency
Nippon Medical School
59 PUBLICATIONS 1,148 CITATIONS
164 PUBLICATIONS 3,793 CITATIONS
SEE PROFILE
SEE PROFILE
Richard Cogdell University of Glasgow 533 PUBLICATIONS 17,554 CITATIONS SEE PROFILE
All content following this page was uploaded by Shinichi Takaichi on 09 January 2015. The user has requested enhancement of the downloaded file. All in-text references underlined in blue are added to the original document and are linked to publications on ResearchGate, letting you access and read them immediately.
Photochemistry and Photobiology, 2001, 74(4):
549–557
Photophysical Characterization of Natural cis-Carotenoids¶ Per Ola Andersson*1, Shinichi Takaichi2, Richard J. Cogdell3 and Tomas Gillbro4 Department of Chemistry and Biomedical Sciences, Kalmar University, Kalmar, Sweden; Biological Laboratory, Nippon Medical School, Kanagawa, Kawasaki, Japan; 3 Division of Biochemistry and Molecular Biology, Institute of Biomedical and Life Sciences, University of Glasgow, Glasgow, UK and 4 Department of Biophysical Chemistry, University of Umea˚, Umea˚, Sweden 1 2
Received 19 January 2001; accepted 17 July 2001
ABSTRACT
ies (3–6). The effects of the conjugation length on the excited-state energies and dynamics have specifically been considered for simple polyenes (3–6), and later also for carotenoids, usually possessing a larger number of conjugated double bonds (N†), resulting in an improved understanding of the electronic excited-state properties of carotenoids (7– 10). Several other studies, mainly during the last decade, have been carried out utilizing different spectroscopic techniques, steady state as well as time resolved, and have collectively contributed to a relatively clear photophysical picture of carotenoids (11–15). In spite of the loss of symmetry caused by methyl groups and other substituents, and by nonplanar ground-state conformations (16), all-trans-carotenoids are classified according to the C2h symmetry point group. The main one-photon absorption corresponds to the strongly optically allowed 1Ag → 1Bu (S0 → S2) transition, and thus has large transition dipole moment. From this state only a minor fraction of the excited carotenoids relax back to ground state via second excited singlet state (S2) fluorescence. S2 → S0 fluorescence quantum yields (ff) are very low, in the range of 1024 (11,17,18), which is reflected by a very short-lived S2 state, in the range of 50–200 fs depending on the solvent and the degree of conjugation (11–13,15,19). This excited singlet state is mainly deactivated by fast internal conversion to a lower-lying singlet excited state (S1) labeled 21Ag. The 21Ag → 11Ag transition is forbidden by symmetry, whereupon the excited molecules undergo an efficient internal conversion back to ground state associated with a large change in the C5C stretching frequency (3,20,21). From this latter observation, together with the fact that the S1 lifetime is only weakly affected on going from room-temperature solvents to low-temperature highly viscous environments (22,23), it is generally believed that the C5C stretching vibrations act as promoting and accepting modes between S1 and S0, and that the involvement of low-frequency, large-amplitude vibrations, such as out-of-plane skeletal twisting, bending and torsional modes is negligible. The trend that S1 lifetime contin-
By means of steady-state fluorescence spectroscopy we explore the photophysics of two lowest lying singlet excited states in two natural 15-cis-carotenoids, namely phytoene and phytofluene, possessing three and five conjugated double bonds (N), respectively. The results are interpreted in relation to the photophysics of all-transcarotenoids with varying N. The fluorescence of phytofluene is more Stokes-shifted relative to that of phytoene, and is ascribed to the forbidden S1 → S0 transition, with its first excited singlet state (S1) lying 3340 cm21 below the dipole allowed second excited singlet state (S2), at 77 K. For phytoene the S2 and S1 potential surfaces are closer in energy, probably giving rise to the mixed S2 and S1 fluorescence characteristics. The origin of phytoene fluorescence is discussed and is suggested to be due to the S1 → S0 transition; with the S1 state located 1100 cm21 below S2 at 77 K. The dependence of the fluorescence quantum yield on temperature and viscosity shows that large amplitude molecular motions are involved in the radiationless relaxation process of phytoene. The transition dipole moment of absorption and emission are parallel in phytoene and nonparallel in phytofluene.
INTRODUCTION Carotenoids, a kind of linear conjugated polyenes, have several important biological functions in photosynthetic organisms, e.g. light-harvesting, photoprotection and structure stabilization (1,2). To better understand their different biological roles, it is crucial to collect information about their electronic structures and dynamics, and further elucidate how these properties are related to the molecular structure, and also to study how they respond to environmental and temperature changes. Moreover, the simple structure of linear (un)substituted polyenes make them well suited as test compounds for theoretical models as well as spectroscopic stud¶Posted on the website on 27 July 2001. *To whom correspondence should be addressed at: Department of Chemistry and Biomedical Sciences, Kalmar University, SE-391 82, Kalmar, Sweden. Fax: 46-480-446262; e-mail: per-ola.
[email protected] q 2001 American Society for Photobiology 0031-8655/01
†Abbreviations: 3mp, 3-methylpentane; n, refractive index; N, number of conjugated double bonds; rs, steady-state fluorescence anisotropy; S1, first excited singlet state; S2, second excited singlet state; lem, emission wavelength; lexc, excitation wavelengths; ff, fluorescence quantum yield.
$5.0010.00
549
550 Per Ola Andersson et al. enoids, are the mismatched fluorescence excitation and absorption spectra obtained for phytofluene at 77 K (not for phytoene), and the nonparallel S0 → S2 absorption and S1 → S0 emission transition dipole moments of phytofluene (not for phytoene).
MATERIALS AND METHODS
Figure 1. (a) Phytoene: 15-cis-7,8,11,12,79,89,119,129-OctahydroC,C-carotene. C40H64. (b) Phytofluene: 15-cis-7,8,11,12,79,89,-Hexahydro-C,C-carotene. C40H62.
uously decreases as the conjugation length increases (23– 25), has successfully been analyzed in terms of the energy gap law (26), i.e. the rate of S1–S0 internal conversion increases exponentially with decreasing S1–S0 energy gap. For carotenoids with large N (.9) it means that S1 fluorescence is effectively absent, and only the less–Stokes-shifted S2 fluorescence is observed; whereas the polyene length gets shorter, the S1 fluorescence intensity gradually rises (7–11). However, trienes, whether substituted or not, are only extremely weakly fluorescent (27,28), whereas dienes are considered as nonfluorescent (6,29). It is argued both from experimental (27) and from theoretical (6,29,30) points of view that this abrupt loss of systematic trend is a consequence of new nonradiative decay channels that open up between S1 and S0 for N # 3, consisting of out-of-plane vibrations as promoting and accepting modes induced by nonplanar S1 state. It was recently reported that S2–S1 and S1–S0 internal conversion of cis-1,3,5-hexatriene follows the conical intersection pathway of the involved potential surfaces, followed by conformational relaxation around the C–C single bonds in the electronic ground state (30,31). However, carotenoids with 4 # N # 8 exhibit distinct S1 → S0 fluorescence with ff in the range of 1024–1022, considerably more efficient than the S2 fluorescence (23). In this paper, we focus attention on the two lowest lying excited singlet states of two carotenoids, phytoene and phytofluene, with three and five conjugated double bonds, respectively, both with 15-cis-conformations (Fig. 1). We attempt to elucidate whether these compounds follow the photophysical trends of all-trans-carotenoids outlined above. Steady-state fluorescence spectra are measured at room temperature as well as at cryogenic temperatures, and ff is determined. The fluorescence anisotropies, rs, are also studied in order to reveal the mutual direction of the absorption and emission transition dipole moments. The most striking feature is that phytofluene shows a distinct Stokes-shifted S1 → S0 fluorescence, whereas phytoene exhibits mixed S2 and S1 fluorescence characteristics ascribed to the relatively small S2–S1 energy gap. As with polyenes with short conjugation length, i.e. N # 3, it is likely that large-amplitude vibrational modes are involved in the deactivation process of the excited states, and in the vibrational coupling to the ground state in phytofluene and particularly in phytoene. The photophysical discrepancies obtained, in comparison with all-trans-carot-
Isolation and purification. Phytoene and phytofluene were isolated and purified from cells of Rhodospirillum rubrum strain S1 cultured in a medium containing 60 mM diphenylamine, which inhibits biosynthesis of colored carotenoids as described in Gillbro et al. (10). It is well known that 15-cis-phytoene is synthesized in the first step of carotenoids synthesis (32). Both purified phytoene and phytofluene fractions contained approximately 85% 15-cis-forms, which were estimated by high-performance liquid chromatography with a Novapak C18 column (33). Spectral measurements. All absorption measurements were made on a Beckman DU-70 spectrophotometer, whereas the steady-state fluorescence and fluorescence excitation spectra were performed on a SPEX Fluorolog 112 instrument (SPEX Ind., Metuchen, NJ). The spectral bandwidth of the fluorescence excitation and emission monochromators was set to 7.4 and 7.2 nm, respectively, at room temperature, and to 7.4 and 3.6 nm, respectively, for low-temperature measurements. Variations in the excitation light intensity were corrected by means of a rhodamine 101 quantum counter. All fluorescence (excitation) spectra were also corrected for the wavelength dependence of the optical components in the monochromators: to correct the fluorescence spectra, a correction curve in the wavelength range 280–1000 nm was recorded with a standard lamp, and the fluorescence excitation spectra were corrected using a correction curve obtained from different standard fluorescent samples (naphtalene, anthracene, perylene, fluorescein, flavinmononucleotide and rhodamine 101) with known fluorescent spectral profiles. The maximum absorbance of the fluorescence samples was below 0.3. The absorption spectrum was checked before and after the fluorescence measurements, and the mixed samples were always stored in darkness. For all room-temperature fluorescence experiments the solvent background signal was always measured and subtracted from the sample fluorescence. All solvents were of p.a. grade, and 3-methylpentane (3mp), which was used as solvent at cryogenic temperatures, was further distilled to avoid fluorescence of impurities. These low temperatures were achieved using a cryostat (Oxford Instruments, Witney, UK). Steady-state fluorescence and fluorescence excitation anisotropies (rs) were calculated using Eqs. 1 and 2: rs 5
IVV 2 GIVH IVV 1 2 ·GIVH
G5
IHV IHH
rs 5
IVV 2 GIHV IVV 1 2 ·GIHV
G5
IVH IHH
(1)
(19)
(2)
(29)
The subscripts on the emission intensity, I, refer to the vertical (V) or horizontal (H) settings of the excitation and emission polarizers, respectively. The settings of IVH correspond to vertical plane polarized excitation light and horizontal plane polarized emission light. The correction factor, G, given by Eqs. 19 and 29, compensates for the different instrument responses for horizontally and vertically polarized light. All fluorescence spectra, with the exception of anisotropy measurements, were obtained with the emission polarizer at the magic angle (54.78) relative to the excitation polarizer. The fluorescence quantum yields, ff, were determined for the carotenoids dissolved in n-hexane and cyclohexane at room temperature and in 3mp at cryogenic temperatures. The quantum yield was calculated from (34):
Photochemistry and Photobiology, 2001, 74(4) 551
Figure 2. (a) Normalized absorption (thinner curve) and fluorescence (broader curve) spectra of phytofluene in 3-methylpentane (3mp) at 295 K (solid curve) and 77 K (dashed curve), respectively. (b) Fluorescence and fluorescence excitation spectra of phytofluene in 3mp at 77 K. Excitation and emission wavelengths are given in the figure. (c) Normalized absorption spectrum (solid curve) measured at 77 K and normalized fluorescence excitation spectrum measured at 295 and 77 K (dashed curves), respectively, of phytofluene in 3mp. (d) Fluorescence (lexc 5 352 nm) and fluorescence excitation (lem 5 484 nm) spectra of phytofluene in 3mp obtained at 77 K are shown together with the fluorescence anisotropy measured at room temperature and at 77 K.
ff 5 fref
2 I (1 2 102Aref )nhex/3mp 2A 2 Iref (1 2 10 )nrefsolv
(3)
Here, A denotes the absorbance at the excitation wavelength (lexc) and n is the refractive index of the solvent. As reference sample for phytoene, we used naphthalene, whose fref has been determined to be 0.23 in cyclohexane (35), and for phytofluene, we used flavinmononucleotide, whose fref has been determined to be 0.26 in ethanol (36). The temperature dependence of n was neglected, whereas A was measured at different temperatures. The fluorescence quantum yield of phytofluene at 77 K was determined from Eq. 3 only for the excitation wavelength 347 nm, whereas at the other wavelengths ff was calculated relative to the 347 nm ff value, with compensation for wavelength dependence of excitation light intensity. We were not able to measure ff at excitation wavelengths shorter than 286 nm because of too weak a excitation light intensity.
RESULTS Phytofluene In the room-temperature absorption spectrum of phytofluene in 3mp (Fig. 2a), the first three peaks in the main band are
located at 366, 347 and 330 nm, in accordance with literature values (33,37). The maximum of the so-called cis-peak was found to be at 255 nm, which also agrees well with reported value of phytofluene (33). Minor absorption bands were measured at 290 and 300 nm (303 nm at 77 K), probably because of absorption of breakdown products. However, no such products were induced during the fluorescence experiments, as the absorption spectrum did not change over the experimental time. When the main absorption band was excited, the spectral shape of fluorescence excitation spectrum was nearly completely independent of emission wavelengths (lem). Only a minor enhancement of shorter-wavelength fluorescence (300–330 nm) was found using shorter lem (Fig. 2b). In a similar way, the fluorescence spectra were nearly independent on lexc; only with lexc shorter than 335 nm did minor extra bands show up between 350 and 420 nm (Fig. 2b). These indicate that the main part of the fluorescence originates from the same type of molecules, i.e. phytofluene, but
552 Per Ola Andersson et al. in the short-wavelength region, there are additional species weakly contributing to the observed fluorescence. We suggest that these impurities belong to breakdown products of phytofluene. That the major part of the observed fluorescence really originates from phytofluene is further emphasized in the following text. The emission maximum of phytofluene in 3mp at 77 K is located at 484 nm (Fig. 2a,b), which is about 2800 cm21 redshifted relative to the 0–0 emission. This difference is typical for S1 emission of linear polyenes (6,7) and for carotenoids with short conjugation length (7,10). At 77 K the 0–0 transitions of absorption and emission are located at 373 and 426 nm, respectively, corresponding to a Stokes shift of 3340 cm21. Fluorescence quantum yields (ff) were determined to be 0.52 6 0.07, with excitation at 373, 335 and 319 nm, respectively, and 0.29 6 0.07 at lexc 5 352 nm, i.e. significantly lower ff for the 0–1 transition. This difference was not observed in the room-temperature spectra of phytofluene dissolved in n-hexane, where the average value of ff was calculated to be (5.5 6 0.5) 3 1022, which agrees well with the earlier result (10). The room-temperature absorption (Fig. 2a) and fluorescence excitation (Fig. 2c) spectra match fairly well; the maximum peak (0–1 transition) is slightly more intense than the 0–0 band in the absorption spectrum (Fig. 2a), whereas the reverse is valid in the fluorescence excitation spectrum (Fig. 2c). A major difference is found for the 77 K spectra: they are largely mismatched at the 0–1 transitions (Fig. 2c), correlating to the lower ff at 352 nm. Owing to the close spectral overlap between absorption and fluorescence excitation spectra at room temperature, reproducibility of both fluorescence and fluorescence excitation spectra, large Stokes shifts and similar vibronic structure for polyenes and related carotenoids, we are convinced that the observed fluorescence (except at the short-wavelength region) originates from phytofluene and not from impurities, and, moreover, we assigned the fluorescence to the optically forbidden S1 → S0 transition. This assignment is further supported by the spectral change of absorption and fluorescence induced by decreasing the temperature from room temperature to 77 K, which is shown in Fig. 2a. The absorption spectrum shifts 510 cm21 to the red because of increasing environmental polarizability induced by lowering the temperature. The polarizability is proportional to the term ([n2 2 1]/[n2 1 2]), and the refractive index (n) is larger at lower temperature compared with higher temperature, resulting in a well-known temperature dependence of the optically allowed transition S0 → S2 (3,38). In contrast, the fluorescence spectrum is blueshifted 500 cm21 when phytofluene is moved from liquid to 77 K glasses of 3mp, in agreement with earlier observations (9). This shift is caused by hindrance of solvent relaxation in the solid phase. In Fig. 2d both fluorescence and fluorescence excitation spectra of phytofluene obtained at 77 K are shown, together with the corresponding steady-state anisotropy (rs). Roomtemperature anisotropies also are included. The average value of rs is in the range of 0.23–0.24 at 77 K and close to zero at room temperature. Phytoene In Fig. 3a phytoene fluorescence spectra at room temperature and 77 K are shown together with the fluorescence excitation
Figure 3. (a) Normalized fluorescence spectra of phytoene measured at room temperature and 77 K with lexc 5 297 nm and lexc 5 286 nm, respectively. Fluorescence anisotropy and normalized fluorescence excitation spectrum obtained at 77 K are also shown. In the room-temperature measurement, n-hexane was used as solvent, otherwise it was 3-methylpentane. (b) Normalized room-temperature absorption and fluorescence spectra of phytoene in n-hexane and cyclohexane, respectively. (c) Room-temperature fluorescence spectrum of phytoene in n-hexane and cyclohexane, respectively. The absorption at lexc 5 295 nm was 0.21 in n-hexane and 0.20 in cyclohexane.
Photochemistry and Photobiology, 2001, 74(4) 553 Table 1. Fluorescence quantum yields of phytoene in 3-methylpentane at different temperatures. The quote knr/kr is calculated from Eq. 4 and the viscosity (h) is taken from Ling and Willard (44) Ff
T (K) 77 91 102 110 120 133 295
Figure 4. Fluorescence spectra of phytoene in 3mp at different temperatures. Excitation wavelength was 286 nm.
spectrum and fluorescence anisotropy measured at 77 K. The average value of rs is 0.37–0.38. The fluorescence spectra were not dependent on lexc or lem, and the fluorescence excitation spectrum overlaps well with the absorption spectrum (Fig. 3b), confirming that the observed emission originates from phytoene. Moreover, the locations of the first three vibronic bands in the absorption spectrum (Fig. 3b) are in agreement with the literature values (33), confirming the purity of the sample. The cis-peak of phytoene was not observed because it is located at too short a wavelength. In contrast to phytofluene spectra (Fig. 2a), both the absorption and fluorescence spectra of phytoene show an approximately 400 cm21 redshift when going from liquid phase, at room temperature, to solid phase, at 77 K, indicating that the absorbing and the emitting states are the same, which will be discussed subsequently. The fluorescence spectrum is less Stokes-shifted for phytoene, compared with phytofluene. The 0–0 transition of phytoene in 3mp glasses at 77 K is found at 301.5 nm of the absorption (data not shown) and 312 nm of the emission, corresponding to a Stokes shift of 1100 cm21. At room temperature the fluorescence spectrum shows a similar Stokes-shifted emission about 1200 cm21 (Fig. 3b). The spectral shape of the fluorescence differs from that of phytofluene. The fluorescence maximum of phytoene is located about 1600 cm21 to the red, relative to the 0–0 band. This 1600 cm21 shift is surprisingly similar to properties of all-trans-carotenoids, with N varying between 9 and 13 (11,17). It is well documented, for these carotenoids, that the fluorescence originates from the strongly optically allowed 11Bu (S2) state (11,17), i.e. the absorbing and the emitting states are the same for these compounds. Fluorescence and absorption spectra of phytoene dissolved in n-hexane and cyclohexane are shown in Fig. 3b. The spectra obtained in the solvent cyclohexane are slightly more redshifted, compared with those of phytoene in n-hexane, which means that both absorbing and emitting states are responding to changes in solvent polarizability in the same manner. This is again in accordance with earlier conclusions for all-trans-carotenoids with large N (9,17,23). The fluorescence quantum yield of phytoene in n-hexane at room temperature is calculated to be (2 6 1) 3 1023. By
0.44 0.29 0.15 0.060 0.023 0.014 0.002
6 6 6 6 6 6 6
knr/kr 0.04 0.03 0.02 0.006 0.003 0.003 0.001
1.27 2.45 5.67 15.67 42.5 70.4 499
h (cP) 2 3 1015 1.4 3 108 1.2 3 104 100 1 0.01 —
exchanging n-hexane (n ø 1.375120) for cyclohexane (n ø 1.426620), ff increases nearly two times as can be seen in Fig. 3c; an effect ascribed to the increasing solvent polarizability induced by the higher solvent refractive index of cyclohexane. This fact is inconsistent with phytofluene and all-trans-carotenoids, but it is in agreement with the solvent dependence of diphenylhexatriene fluorescence originating from excited state of 21Ag symmetry (39). At cryogenic temperatures the ff of phytoene is substantially larger than at room temperature. It is calculated to be 0.44 6 0.04 at 77 K. The temperature dependence is shown in Fig. 4 and Table 1. From ff the ratio between the nonradiative (knr) and radiative (kr) rate constants were calculated at each temperature by the expression knr/kr 5 ff21 2 1
(4)
where kr is assumed to be constant (Table 1). The natural logarithm of knr/kr is plotted versus T21 and, as shown in Fig. 5, the temperature dependence nearly follows the Arrhenius equation. The deviations from linear dependence are within the experimental errors. By fitting the data with the function, ln knr /kr 5 A 2
B T
(5)
where A and B are treated as constants, resulting in an ac-
Figure 5. Arrhenius plot. The logarithm of ratio between nonradiative (knr) and radiative (kr) rate constants versus the inverse of temperature.
554 Per Ola Andersson et al. tivation energy (E) of 540 6 150 cm21, close to those obtained for previtamin D3 (40) and mini-3—a compound homologous to all-trans-b-carotene, N 5 3 (41).
DISCUSSION That the carotenoids used in this work are in a double bond cis-conformation (85%), instead of the extensively studied all-trans-conformation, has no substantial effect on the excitation energies of the S1 and S2 states. The absorption of phytoene and phytofluene, caused by the allowed transition S0 → S2, labeled 11A1 → 11B2 in C2v symmetry point group, lies in a spectral region expected from spectroscopic data of trans-carotenoids (11) and polyenes (3–7). The fluorescence of phytofluene is attributed to the forbidden S1 → S0 transition, labeled 21A1 → 11A1 in C2v, showing up in an expected spectral range (3–7). The origin and location of phytoene fluorescence are discussed below. That the cis- or trans-conformation only has minor effect on S2 and S1 energies has been observed for linear polyenes (3), and, for example, the absorption spectra of cis- and all-trans-b-carotene shift only a few nanometers from each other. Photophysics of phytofluene (N 5 5) Let us first consider the photophysics of phytofluene. The excitation energies are predictable whereas the temperature dependence of ff and rs are more unexpected. The 10-fold rise of ff, when going from room temperature to 77 K, is significantly larger than the two- to three-fold increase that occurs in the carotenoids of N $ 5 (9,23). The latter fact shows that large-amplitude molecular motions only have a minor influence on the radiationless decay from S1 and S0, as they are hindered in high viscous medium, as in 3mp glass, at 77 K (22). Instead, in accordance with linear polyenes of N $ 4 (6,29,30), it is argued that the efficient S1 to S0 internal conversion processes are mainly mediated by the C5C stretching modes, which are not affected by altering the temperature or viscosity of the solvent medium. Thus, the larger temperature and viscosity dependence of phytofluene relative all-trans-carotenoids of N $ 5 shows that other modes than C5C stretching modes contribute to the deactivation of the S1 state of phytofluene, i.e. large-amplitude modes. It is worth noting that for shorter linear polyenes of N # 3 (6,27–30), it is validated that torsional modes are involved in the relaxation of 21Ag to the ground state. The 77 K absorption and fluorescence excitation spectra are similar, except for the 0–1 transition where the fluorescence is reduced (Fig. 2c), to the nearly two times less ff for this transition compared with other vibronic transitions. In the room-temperature spectra there are only minor differences, i.e. when the 0–1 band is excited, the radiationless relaxations are less sensitive to temperature, alterations, viscosity alterations or both, indicating that C5C stretching vibrations are more involved in the radiationless relaxation from S2 via S1 back to ground state than with excitation of other vibronic bands. This fact has never been reported for all-trans-carotenoids, and it has also not been observed for phytoene; because the sample mainly contains cis-conformation (85%), we ascribe this effect to 15-cis-phytofluene. Both fluorescence and fluorescence excitation anisotropies of phytofluene in 3mp at 77 K are constant, about 0.23–
0.24, over the whole wavelength range. If there were any large fluorescence contributions from impurities, they must occur in the same wide wavelength interval as for phytofluene, otherwise rs should vary with the wavelength (with the assumption that the rs of phytofluene and the impurities differs). We found this very unreasonable. According to this, it also very unlikely that the unexpected temperature dependence of the 0–1 transition in the fluorescence excitation spectrum of phytofluene is an effect of impurities. For phytofluene in organic solvent at room temperature, rs is measured to be close to zero. By comparing ff of N 5 5 carotenoids, e.g. the short carotene (mini-5) homologous to all-trans-b-carotene exhibits a value of ff 5 (7 6 3) 3 1023 (9), it is obvious that ff 5 0.055 of phytofluene is rather high, and a relatively long S1-state lifetime in the order of nanoseconds is expected. The S1 state of mini-5 has been determined to be 2 ns (9). During the excited-state lifetime of phytofluene the dynamic behavior on a shorter time scale totally depolarizes the emitted light giving rise to rs ø 0. In solid phase, at 77 K, where the rotational motion is hindered, rs is increased to approximately 0.24 (Fig. 2d). Neither for all-trans-carotenoids with differing chain lengths and N (23) nor for cis-b-carotene, exhibiting S2 → S0 fluorescence (unpublished results), such a low rs value has never been reported for carotenoids embedded in a 77 K matrix. For these compounds the absorption and emission transition dipole moments are collinear, irrespective of the S1 or S2 origin of the fluorescence, showing an rs close to 0.4 (23). Consequently, this cannot be the case for phytofluene; the S0 → S2 and S1 → S0 transition dipole moments are nonparallel in phytofluene. The relation of the mutual angle (u) between these dipole moments is given in Eq. 6 (42). rs 5 ro
1
2
3 cos2 u 2 1 2
(6)
With ro 5 0.4, we obtained u 5 318, and because the absorption and emission is governed by the transition S0 → S2 and S1 → S0, respectively, this angle should represent the angle between their corresponding transition dipole moments. However, because the samples consist of not only cis-conformation but also trans-forms we are not able to determine u for 15-cis-phytofluene. Nevertheless, because the rs of all-trans-carotenoids are close to 0.4 (11,23) and because the samples contain 85% 15-cis, we attribute the unusually low rs to 15-cis-phytofluene, and with 100% purity probably an even lower rs should be obtained, correlating to a u . 318. Photophysics of phytoene (N 5 3) Before we discuss the photophysics of phytoene let us consider the excited-state characteristics of related molecules. Longer linear polyenes and intermediate carotenoids, 4 # N # 8, show substantial S1 → S0 fluorescence (3,7,23), whereas, all-trans-trienes, whether substituted or not, are regarded as either nonfluorescent or very weakly fluorescent in liquid phase (6,27,29). As mentioned earlier, the missing fluorescence from butadiene and the extremely weak triene fluorescence have been explained by theoretical calculations (6,29,30) to be associated with a nonradiative relaxation pathway from 21Ag to 11Ag, induced by a nonplanar stable
Photochemistry and Photobiology, 2001, 74(4) 555 21Ag conformation, which is not found for longer polyenes (N $ 4). The photophysical picture of the short polyenes can be summarized as follows: the absorption takes place upon S0 → S2 (11Ag → 11Bu) excitation, where both involved states are of similar planar conformation. The S2 state is rapidly depopulated by efficient internal conversion to the lower-lying nonplanar S1 (21Ag) state, which relaxes back to the ground state by out-of-plane vibrations as accepting modes. Another striking feature of (un)substituted polyenes is that the excitation energies of S2 or S1 decrease, whereas the energy gap between the states increases when the conjugation length gets longer (3–5). We should, therefore, expect an S2–S1 energy difference of phytoene to be less than 3340 cm21 as determined for phytofluene at 77 K. With all of this in mind we would expect phytoene to be nonfluorescent, or only very weakly fluorescent at room temperature, because of the symmetry forbidden S1 (21A1) state, but intensively fluorescent at 77 K because of hindrance of vibrational coupling modes, and a relatively little Stokesshifted fluorescence spectrum. Although ff 5 2 3 1023 for phytoene in n-hexane is comparable to those found for carotenoids with larger N, it is clearly less than for phytofluene. So, as predicted above, we expect that the nonradiative excited state relaxation is more pronounced in phytoene than in phytofluene. Furthermore, the strong temperature dependence at cryogenic temperatures of the fluorescence intensity is also predictable from related polyenes showing S1 → S0 fluorescence. These findings, together with that of a less– Stokes-shifted fluorescence of phytoene than for phytofluene, allow us to ascribe the emitting state to the forbidden 21A1 (S1), located about 1100 cm21 below the allowed 11B2 (S2) state. However, some spectral characteristics might be inconsistent with this assignment. The Stokes shift of phytoene emission is comparable to those obtained for large conjugated carotenoids possessing mainly S2 → S0 fluorescence. For instance, the S2 emission of cis- and all-trans-b-carotene is Stokes-shifted about 500–600 cm21 at 77 K (unpublished results), i.e. two times less shifted than for phytoene. Another, perhaps even larger, spectral similarity with longer conjugated carotenoids is the spectral shape of the fluorescence spectrum. The fluorescence maximum corresponds to the 0–1 transition located about 1300 cm21 lower than the 0–0 transition, as is widely observed in S2 → S0 spectra. A typical S1 → S0 fluorescence is that of phytofluene, shown in Fig. 2d, with the maximum band about 2800 cm21 to the red of the 0–0 transition. Further indication of S2 emission is the 220 cm21 redshift of both fluorescence and absorption spectra when the solvent is exchanged from n-hexane to cyclohexane, shown in Fig. 3b, as well as the redshift induced by decreasing the temperature to 77 K, which increases the solvent polarizability (Fig. 3a). Finally, differing from phytofluene, the rs is close to 0.4 in agreement with the S2 fluorescence of carotenoids possessing collinear absorption and emission transition dipole moments (23). From all these indications we could claim, in a similar way as above, that the phytoene fluorescence originates from the allowed S2 state. Which is then correct? We cannot exclude S2 fluorescence at the moment; however, we prefer the first assignment of symmetry forbidden emission for the following reasons. There is no doubt that the S2 and S1 states are very close
in energy. When the polarizability was increased, by using cyclohexane instead of n-hexane, ff increased about 1.8 times. Although there is only a minor change of the solvent polarizability, this produces a large effect in ff. In a fluorescence study of diphenylpolyenes a similar observation was made for diphenylhexatriene with a fluorescent S1 state lying about 1600 cm21 below the allowed S2 state in toluene (39). The emission rate was increased by a factor of two when the solvent was exchanged from n-hexane to dioxane (n 5 1.416520). Therefore we suggest that the forbidden state is located very close to the allowed one, but at lower energy, as for all linear polyenes and carotenoids investigated so far with N $ 3 (3–6,11). Let us now assume that our assignment is correct, with a symmetry forbidden emitting S1 state lying about 1100 cm21 below a very short-lived optically allowed S2 state. Then, because of the small S2–S1 energy gap, the intensity borrowing mechanism, strongly dependent on the S2–S1 energy difference, could make the forbidden S1 → S0 transition more allowed for phytoene in comparison with carotenoids possessing larger S2–S1 energy gap. Thus, strong mixing of S1 and S2 states might induce the extraordinary ‘‘allowed’’ character in the S1 → S0 transition. This phenomenon has been analyzed in detail for diphenylpolyenes with N varying between 2 and 4 (39). The temperature dependence of ff is more pronounced for phytoene when compared with phytofluene (Figs. 4 and 5). A 200-fold increase is found upon cooling from room temperature to 77 K. In the temperature interval between 133 and 77 K, ff increases from 0.014 to 0.44 (Table 1). It means that the radiationless decay rate is strongly dependent on temperature, and as the solvent viscosity drastically decreases in this temperature range (Table 1), it also has an effect on the relaxation rate. As mentioned earlier, large-amplitude molecular motions are eliminated in the 3mp glasses as they are strongly dependent on solvent viscosity (43). When the temperature rises and the viscosity drops, these modes get gradually more unblocked. Thus, we ascribe these large-amplitude modes to be strongly involved in the deactivation of the excited state; the C5C stretching mode plays only a minor role in this relaxation process. Similar temperature-dependent effects have been reported for previtamin D3 (40) and a compound homologous to all-trans-b-carotene, N 5 3, called mini-3 (41); but such effects have never been observed for longer carotenoids with N . 3. For mini-3 the fluorescence was attributed to the 21Ag → 11Ag transition (41). In the temperature interval 133–77 K the viscosity of 3mp alters dramatically from less than 1 to 1015 cP (44). Therefore, it is likely that the activation energy Ea 5 540 cm21 reflects a combination of a molecular potential barrier (E0) and a viscosity-induced barrier (E), i.e. Ea 5 E0 1 E, resulting in E0 , 540 cm21. This means that after 11A1 → 11B2 excitation, phytoene undergoes a rapid internal conversion to the 21A1 potential surface, from which the molecule continues to relax to a new equilibrium conformation in the ground state, unless it is prevented by frictional drag, very low thermal energy or both. Because of similar spectral shape, except for a minor shift between 77 K and roomtemperature fluorescence spectra, we emphasize that the fluorescence originates from the same electronic state irrespective of temperature, i.e. presumably from 21A1, and that
556 Per Ola Andersson et al. the 11B2 → 21A1 internal conversion, together with excited state vibrational energy redistribution, is not affected by this significant temperature change, viscosity change or both.
CONCLUSIONS The cis-conformation of phytoene and phytofluene has, relative to all-trans-carotenoids, no influence or only a minor influence on the excitation energy of the S2 and S1 states. The fluorescence from phytofluene is clearly originating from the forbidden S1 (21A) state lying 3340 cm21 below the dipole allowed 11B1 state at 77 K. In contrast, the assignment of the phytoene fluorescence is not that straightforward. However, there is no doubt that the S2 and S1 states are nearly degenerate, which probably gives rise to the observed mixed S2 and S1 fluorescent characteristics. The origin of phytoene fluorescence is uncertain, but we propose that S2 is nonfluorescent and thus the S1 fluorescence is measured with its 0–0 transition located 1100 cm21 below the S2 state (in 77 K glasses). The radiationless relaxation is more efficient in phytoene than in phytofluene, which is reflected in their differing fluorescence quantum yields when measured at room temperature. Usually, with all-trans-carotenoids, ff is independent of the excitation wavelength—a fact also manifested for phytoene and phytofluene at room temperature, but for phytofluene (not for phytoene) in 77 K glasses, excitation of the 0–1 absorption band results in a substantial decrease of ff relative to other excitation wavelengths, which is also reflected in the 77 K fluorescence excitation spectrum. Why is this kind of radiationless relaxation rate enhancement observed for phytofluene? We do not, as yet, have any satisfying explanation for this. The ff increases 10-fold, with the exception of 0–1 excitation, as compared to a two- to threefold increase as is usually observed for all-trans-carotenoids with N $ 5, upon decreasing the temperature from 295 to 77 K (11,22), strongly indicating that not only are C5C stretching modes involved in the S1–S0 internal conversion process, but also are some large-amplitude out-of-plane vibrational modes. Thus, when the 0–1 band is excited, the C5C stretching vibrations are probably acting more efficiently as accepting and promoting modes compared with other excitation wavelengths. The dependence of ff on temperature, viscosity or both is much more pronounced for phytoene, whose ff is enhanced about 200 times upon cooling to 77 K in 3mp. It is obvious that for longer conjugated carotenoids, with their weak dependence on the relaxation rates of temperature, viscosity or both, the C5C stretching modes facilitates the radiationless relaxation pathways, and when the conjugation length decreases, the degree of involvement of large-amplitude vibrational modes in the excited state deactivation process gradually increases. In an Arrhenius plot of knr/kr of phytoene, we determined the total potential barrier to be about 540 6 150 cm21. In accordance with, e.g. diphenylpolyenes, but in contrast to other carotenoids, the ff of phytoene is strongly affected by the solvent polarizability. We suggest that the small spacing between the S2 and S1 energy levels is the basic reason. By increasing the refractive index of the solvent, and thereby the solvent polarizability, the S2–S1 energy gap decreases,
thus inducing a stronger S1 emission intensity because of the intensity borrowing mechanism. Because the fluorescence anisotropy is high, 0.37–0.38 of phytoene in solid phase at 77 K, the absorbing and emitting transition dipole moments must be parallel, in agreement with all-trans-carotenoids, irrespective of 11Bu or 21Ag fluorescence. Compared to this, the rs ø 0.24 of phytofluene in 77 K glasses is anomalously low, corresponding to nonparallel S0 → S2 absorption and S1 → S0 emission transition dipole moments. Acknowledgements R.J.C. thanks the BBSRC and the Human Frontiers of Science for financial support. T.G. thanks the NFR, HFS and TFR for financial support. T.G. and P.O.A. also thank the Kempe foundation for financial support. The authors are very grateful to Eva Wikstro¨m for skillful technical assistance.
REFERENCES 1. Krinsky, N. I. (1971) Function. In Carotenoids (Edited by O. Isler, G. Guttman and U. Solms), pp. 669–716. Birkhauser, Basel. 2. Frank, H. A. and R. J. Cogdell (1996) Carotenoids in photosynthesis. Photochem. Photobiol. 63, 257–264. 3. Hudson, B. S., B. E. Kohler and K. Schulten (1982) Linear polyene electronic structure and potential surfaces. In Excited States, Vol. 6 (Edited by E. C. Lim), pp. 66–73. Academic Press, New York. 4. Tavan, P. and K. Schulten (1987) Electronic excitations in finite and infinite polyenes. Phys. Rev. B. 36, 4337–4358. 5. Kohler, B. E. (1991) Electronic properties of linear polyenes. In Conjugated Polymers (Edited by J. L. Bredas and R. Silbey), pp. 405–434. Kluwer Academic, Netherlands. 6. Orlandi, G., F. Zerbetto and M. Z. Zgierski (1991) Theoretical analysis of spectra of short polyene. Chem. Rev. 91, 867–891. 7. Cosgrove, S. A., M. A. Guite, T. B. Burnell and R. L. Christensen (1990) Electronic relaxation in long polyenes. J. Phys. Chem. 94, 8118–8124. 8. DeCoster, B., R. L. Christensen, R. Gebhard, J. Lugtenburg, R. Farhoosh and H. A. Frank (1992) Low-lying electronic states of carotenoids. Biochim. Biophys. Acta 1102, 107–114. 9. Andersson, P. O., T. Gillbro, A. E. Asato and R. S. H. Liu (1992) Dual singlet state emission in a series of mini-carotenes. J. Lumin. 51, 11–20. 10. Gillbro, T., P. O. Andersson, R. S. H. Liu, A. E. Asato, S. Takaichi and R. J. Cogdell (1993) Location of the carotenoid 2Ag-state and its role in photosynthesis. Photochem. Photobiol. 57, 44–48. 11. Koyama, Y., M. Kuki, P. O. Andersson and T. Gillbro (1996) Singlet excited states and the light-harvesting function of carotenoids in bacterial photosynthesis. Photochem. Photobiol. 63, 243–256. [Invited Review] 12. Mimuro, M., S. Akimoto, S. Takaichi and I. Yamazaki (1997) Effect of molecular structures and solvents on the excited state dynamics of the S2 state of carotenoids analyzed by the femtosecond up-conversion method. J. Am. Chem. Soc. 119, 1452– 1453. 13. Macpherson, A. N. and T. Gillbro (1998) Solvent dependence of the ultrafast S2–S1 internal conversion rate of b-carotene. J. Phys. Chem. 102, 5049–5058. 14. Bautista, J. A., R. E. Connors, B. B. Raju, R. G. Hiller, F. P. Sharples, D. Gosztola, M. R. Wasielewski and H. A. Frank (1999) Excited state properties of peridinin: observation of a solvent dependence of lowest excited singlet state lifetime and spectral behavior unique among carotenoids. J. Phys. Chem. B 103, 8751–8758. 15. Walla, P. J., J. Yom, B. P. Krueger and G. R. Fleming (2000) Two-photon excitation spectrum of light-harvesting complex II and fluorescence up-conversion after one- and two-photon excitation of the carotenoids. J. Phys. Chem. B 104, 4799–4806. 16. Sterling, C. (1964) Crystal-structure analysis of b-carotene. Acta Crystallogr. 17, 1224–1228.
Photochemistry and Photobiology, 2001, 74(4) 557 17. Gillbro, T. and R. J. Cogdell (1989) Carotenoid fluorescence. Chem. Phys. Lett. 158, 312–316. 18. Bondarev, S. L., S. M. Bachilo, S. S. Dvornikov and S. A. Tikhomirov (1989) S2–S0 fluorescence and transient SN–S1 absorption of all-trans-b-carotene in solid and liquid solution. J. Photochem. Photobiol. 46, 315–322. 19. Kandori, H., H. Sasabe and M. Mimuro (1994) Direct determination of a lifetime of the S2 state of b-carotene by femtosecond time-resolved fluorescence spectroscopy. J. Am. Chem. Soc. 116, 2671–2672. 20. Noguchi, T., S. Kolaczkowski, C. Arbour, G. H. Aramaki, G. H. Atkinson, H. Hayashi and M. Tasumi (1989) Resonance Raman spectrum of the excited 21Ag state of b-carotene. Photochem. Photobiol. 50, 603–609. 21. Hashimoto, H. and Y. Koyama (1989) Raman spectra of alltrans-b-carotene in the S1 and T1 states produced by direct photoexcitation. Chem. Phys. Lett. 163, 251–256. 22. Wasielewski, M. R., D. G. Johnson, E. G. Bradford and L. D. Kispert (1989) Temperature dependence of the lowest singletstate lifetime of all-trans-b-carotene and fully deuterated alltrans-b-carotene. J. Chem. Phys. 91, 6691–6697. 23. Andersson, P. O., S. M. Bachilo, R.-L. Chen and T. Gillbro (1995) Solvent and temperature effects on dual fluorescence in a series of carotenes. Energy gap dependence of the internal conversion rate. J. Phys. Chem. 99, 16 199–16 209. 24. Andersson, P. O. and T. Gillbro (1995) Photophysics and dynamics of the lowest singlet state in long substituted polyenes with implications to the very long-chain limit. J. Chem. Phys. 103, 2509–2519. 25. Frank, H. A., R. Farhoosh, R. Gebhard, J. Lugtenburg, D. Gosztola and M. R. Wasielewski (1993) The dynamics of the S1 excited states of carotenoids. Chem. Phys. Lett. 207, 88–92. 26. Englman, R. and J. Jortner (1970) The energy gap law for radiationless transitions in large molecules. Mol. Phys. 18, 145– 164. 27. Petek, H., A. J. Bell, R. L. Christensen and K. Yoshihara (1992) Fluorescence excitation spectra of the S1 states of isolated trienes. J. Chem. Phys. 96, 2412–2415. 28. Buma, W. J., B. E. Kohler and K. Song (1991) Lowest energy excited state of isolated cis-hexatriene. J. Chem. Phys. 94, 6367–6376. 29. Zerbetto, F. and M. Z. Zgierski (1990) The missing fluorescence of s-trans butadiene. J. Chem. Phys. 93, 1235–1245.
View publication stats
30. Fuss, W., Y. Haas and S. Zilberg (2000) Twin states and conical intersections in linear polyenes. Chem. Phys. 259, 273–295. 31. Ohta, K., Y. Naitoh, K. Tominaga, N. Hirota and K. Yoshihara (1998) Femtosecond transient absorption studies of trans- and cis-1,3,5-hexatriene in solution. J. Phys. Chem. A 102, 35–44. 32. Britton, G. (1998) Overview of carotenoid biosynthesis. In Carotenoids, Vol. 3, Biosynthesis and Metabolism (Edited by G. Britton, S. Liaaen-jensen and H. Pfander), pp. 13–147. Birkhauser, Basel. 33. Takaichi, S. (2000) Characterization of carotenes in a combination of a C18 HPLC column with isocratic elution and absorption spectra with a photodiode-array detector. Photosynth. Res. 65, 93–99. 34. Birks, J. B. (1970) Photophysics of aromatic molecules. WileyInterscience, London. 35. Berlman, I. B. (1965) Handbook of fluorescence spectra of aromatic molecules. Academic Press, New York. 36. Weber, G. and F. W. Teale (1957) Determination of the absolute quantum yield of fluorescent solutions. Trans. Faraday Soc. 53, 646–655. 37. Mouly, P. P., M. E. Gaydou and J. Corsetti (1999) Determination of the geographical origin of Valencia orange juice using carotenoid liquid chromatographic profiles. J. Chromatogr. A 844, 149–159. 38. Andersson, P. O., T. Gillbro, L. Ferguson and R. J. Cogdell (1991) Absorption spectral shifts of carotenoids related to medium polarizability. Photochem. Photobiol. 54, 353–360. 39. Bachilo, S. M., E. V. Bachilo and T. Gillbro (1998) Spectral shape of diphenylpolyene fluorescence and mixing of the S1 and S2 states. Chem. Phys. 229, 75–91. 40. Dauben, W. G., B. Disanayaka, D. J. H. Funhoff, B. E. Kohler, D. E. Schilke and B. Zhou (1991) Polyene 21Ag and 11Bu states and the photochemistry of Previtamin D3. J. Am. Chem. Soc. 113, 8367–8374. 41. Andersson, P. O., T. Gillbro, A. E. Asato and R. S. H. Liu (1995) Temperature and viscosity sensitive S1 emission from highly substituted triene. Chem. Phys. Lett. 235, 76–82. 42. Lakowicz, J. R. (1983) Principles of Fluorescence Spectroscopy. Plenum Press, New York. 43. Sundstro¨m, V. and T. Gillbro (1984) Picosecond dynamics of large-amplitude motions of triphenylmethane molecules in alcohol solution. Chem. Phys. Lett. 110, 303–307. 44. Ling, A. C. and J. E. Willard (1968) Viscosity of some organic glasses used as trapping matrices. J. Phys. Chem. 72, 3349– 3351.
