ARTICLE pubs.acs.org/JPCB
Singlet and Triplet State Spectra and Dynamics of Structurally Modified Peridinins Marcel Fuciman,† Miriam M. Enriquez,† Shanti Kaligotla,† Dariusz M. Niedzwiedzki,† Takayuki Kajikawa,‡ Kazuyoshi Aoki,‡ Shigeo Katsumura,‡ and Harry A. Frank*,† †
Department of Chemistry, University of Connecticut, U-3060, 55 North Eagleville Road, Storrs, Connecticut 06269-3060, United States ‡ Department of Chemistry, Kwansei Gakuin University, 669-1337, Hyogo, Japan ABSTRACT: The peridininchlorophyll aprotein (PCP) is a lightharvesting pigmentprotein complex found in many species of marine algae. It contains the highly substituted carotenoid peridinin and chlorophyll a, which together facilitate the transfer of absorbed solar energy to the photosynthetic reaction center. Photoexcited peridinin exhibits unorthodox spectroscopic and kinetic behavior for a carotenoid, including a strong dependence of the S1 excited singlet state lifetime on solvent environment. This effect has been attributed to the presence of an intramolecular charge transfer (ICT) state in the molecule. The present work explores the effect of changing the extent of π-electron conjugation and attached functional groups on the nature of the ICT state of peridinin and how these factors affect the excited singlet and triplet state spectra and kinetics of the carotenoid. In this investigation three peridinin analogues denoted C-1R-peridinin, C-1-peridinin, and D-1-peridinin were synthesized and studied using steady-state absorption and fluorescence techniques and ultrafast timeresolved transient absorption spectroscopy. The study explores the effect on the singlet and triplet state spectra and dynamics of removing the allene group from the peridinin structure and either replacing it with a rigid furanoid ring, replacing it with an epoxide group, or extending the polyene chain into the β-ionylidine ring.
’ INTRODUCTION Several marine algae make use of the peridininchlorophyll aprotein (PCP) complex for light harvesting.13 This pigmentprotein complex contains a highly substituted carotenoid, peridinin (Figure 1), and chlorophyll a (Chl a) in a stoichiometric ratio of eight peridinins to two Chl's. The chromophore pigments are bound to the protein with the eight peridinins divided into two equivalent and symmetric clusters, each of which surrounds and is in van der Waals contact with a single Chl a.1 Peridinin is not only adept at light harvesting in this protein complex, but is also efficient at protecting the system from photodegradation due to reactive oxygen species.411 The present work seeks to explore the relationship between the complex structure of peridinin and its light-harvesting and photoprotective roles by examining the spectra and kinetics of the excited singlet and triplet states of a series of systematically modified peridinin molecules. The overall goal of the work is to understand why nature selected such a synthetically labor-intensive carotenoid as peridinin for inclusion into the PCP complex and how the structure of this molecule relates to its proclivity for accomplishing highly efficient singlet and triplet energy transfer to and from Chl a. It is important to point out that even though peridinin contains a number of functional groups that render it essentially r 2011 American Chemical Society
asymmetric (Figure 1), optical transitions taking place in the molecule follow the quantum mechanical selection rules governing absorption and emission of molecules having C2h symmetry. Accordingly, one-photon transitions to and from the ground state, S0, and the lowest-lying excited state, S1, are forbidden due to the fact that both states are characterized by the Ag irreducible representation. The selection rules dictate that a change in both symmetry (g T u) and parity (þ T ) is required for such transitions to be allowed.1220 Therefore, a one-photon transition between S0 (11Ag) and S2 (11Buþ) is allowed, and this is the transition responsible for the strong optical absorption bands of these molecules in the visible spectral region. For many carotenoids and polyenes, the forbiddenness of the S0 T S1 transitions manifests itself in a lack of sensitivity on solvent environment of the energy and lifetime of the S1 state. However, when a carbonyl functional group is present as part of the π-electron conjugated chain, as is the case with peridinin (Figure 1) and several other carotenoids and polyenes, a profound effect of solvent polarity on the lifetime of the lowest excited singlet state is observed.2127 The effect disappears when Received: October 21, 2010 Revised: March 17, 2011 Published: March 31, 2011 4436
dx.doi.org/10.1021/jp110092c | J. Phys. Chem. B 2011, 115, 4436–4445
The Journal of Physical Chemistry B
Figure 1. Structures of peridinin and C-1-R-, C-1-, D-1-peridinin analogues.
the carbonyl group is removed.21,22,28 This behavior has been rationalized by the presence of an intramolecular charge transfer (ICT) state electronically coupled to the S1 state.19,21,22 It is thought that the energy of the ICT state relative to that of the S1 state can be altered by changing the solvent polarity. Thus, when the ICT state is stabilized by polar solvents such that it lies below the S1 state, the ICT state can accept population from S1 thereby shortening its lifetime.19,22 However, the precise nature of the ICT state and its electronic coupling to S1 has yet to be elucidated. Despite the profound dependence of the S1 lifetime on solvent polarity reported for peridinin and other carbonyl-containing carotenoids, no such dependence of the lifetime of the lowest excited triplet state (T1) of these molecules has ever been reported.29 It is a well-known fact that the triplet state of a carotenoid must lie lower in energy relative to that of Chl a to quench Chl triplet states before they can sensitize the formation of the 1Δg excited singlet state of oxygen.30 In this way carotenoids protect the photosynthetic apparatus from photodynamic damage due to singlet oxygen. However, the triplet state energy of peridinin is not known, nor is it well understood how the triplet state wave function is configured in the molecule or in the PCP complex when it is in proximity to Chl a. Some investigators have proposed that peridininChl a interactions give rise to a delocalized triplet state electronic configuration that produces a triplet trap sufficiently low in energy to quench Chl triplet states before they can sensitize the formation of singlet oxygen.6,31 Others have suggested that a triplet state is localized on a single peridinin, and that this is the controlling factor for photoprotection.9,10,32 Whatever the case, a knowledge of how the triplet state spectra and kinetics are affected by altering the structure of peridinin will
ARTICLE
help in revealing the conditions under which the triplet state photophysics of the molecule can be modulated. To explore the nature of the ICT state in carbonyl-containing carotenoids, and to address the controlling features of singlet and triplet excited state spectra and dynamics, we synthesized a series of functionalized peridinin analogues (Figure 1). Naturally occurring peridinin has a C37 carbon skeleton and eight linearly conjugated carboncarbon double bonds, one of which lies between carbons 70 and 80 as part of an allene functional group (Figure 1). Peridinin, like the other molecules studied here, also has a carbonyl group in conjugation with the π-electron system via a lactone ring attached to the linear chain at carbon positions 9 and 11. The synthetic analogues are shown along with peridinin in Figure 1. They are C-1-R-peridinin, which has one less double bond than peridinin due to a furanoid ring substitution connecting carbon positions 50 and 80 , C-1-peridinin, which has the same number, N, of carboncarbon double bonds as peridinin (N = 8) but lacks the allene functional group between carbons 60 and 80 and instead has an epoxide group at carbon positions 50 and 60 , and D-1-peridinin, which has nine carboncarbon double bonds but one of its conjugated double bonds occupies a position in a terminal β-ionylidine ring. In all other ways concerning functional groups, the molecules are structurally identical to peridinin (Figure 1). The trends in the excited state spectroscopic and kinetic properties exhibited by this systematic series of peridinin analogues dissolved in solvents having a wide range of polarities are consistent with a model in which the ICT state is formed from both S1 and S2, but decays to the ground state independently from S1. The rate of formation of the ICT state from S1 is strongly dependent on solvent polarity, and therefore, altering the solvent is shown to have a profound effect on the S1 lifetime. The triplet state lifetime is found here to be insensitive to solvent polarity, suggesting that either no ICT state analogous to that affecting the singlet state photophysics of peridinin and analogues is present in the triplet manifold, or if a triplet ICT state does exist, it lies sufficiently high in energy relative to the lowest excited triplet, T1, that it plays no role in controlling the rate of T1 f S0 intersystem crossing or in the photoprotection of Chl a in the PCP complex.
’ EXPERIMENTAL METHODS Sample Preparation. C-1-R-, C-1-, and D-1-peridinins were synthesized as previously described33,34 and provided as dry samples. Prior to the spectroscopic measurements, the molecules were dissolved in acetonitrile and purified using high-performance liquid chromatography (HPLC) carried out on a Millipore Waters 600E HPLC instrument equipped with a Waters 2996 single diode-array detector. Either a YMC-Carotenoid C30 column (250 mm 4.6 mm, 5 μm particle size) or a Nova-Pak C18 column (300 mm 3.9 mm, 4 μm particle size) employing an isocratic solvent protocol consisting of acetonitrile/methanol/water (87/10/3, v/v/v) was used. The flow rates were 0.5 mL/min for C-1-R-peridinin and 1.0 mL/min for C-1- and D-1-peridinins. The eluting HPLC peak of each molecule was collected, dried under gentle stream of N2 gas, and stored at 80 °C until ready for the spectroscopic measurements. Chlorophyll a was purchased from Sigma-Aldrich and used without further purification. Spectroscopic Methods. Steady-State Absorption and Fluorescence. Steady-state absorption spectra of C-1-R-, C-1-, and D-1-peridinins in n-hexane, methyl tert-butyl ether (MTBE), 4437
dx.doi.org/10.1021/jp110092c |J. Phys. Chem. B 2011, 115, 4436–4445
The Journal of Physical Chemistry B ethyl acetate, 2-propanol, and methanol were recorded using a Varian Cary 50 UVvisible spectrophotometer. The polarities of these solvents computed from the formula P(ε) = (ε 1)/ (ε þ 2), where ε is the dielectric constant of the solvent, are 0.229 (n-hexane), 0.350 (MTBE), 0.626 (ethyl acetate), 0.852 (2-propanol), and 0.913 (methanol). The polarizabilities of these solvents computed from the formula P(n) = (n2 1)/ (n2 þ 2) where n is the index of refraction, are very similar: 0.228 (n-hexane), 0.226 (MTBE), 0.226 (ethyl acetate), 0.230 (2-propanol), and 0.203 (methanol). Fluorescence measurements were carried out using a JobinYvon Horiba Fluorolog-3 Model FL3-22 equipped with double monochromators having 1200 grooves/mm gratings, a Hamamatsu R928P PMT emission detector, and a 450 W ozone-free Osram XBO xenon arc lamp. The fluorescence of the molecules in various solvents was detected at a right angle relative to the excitation. The excitation wavelengths were 433452 nm for C-1-R-peridinin, 472482 nm for C-1-peridinin, and 478 486 nm for D-1-peridinin. Both excitation and emission monochromator slit widths were set to correspond to a band pass of 5 nm. Correction factors were applied to all the fluorescence spectra to compensate for the wavelength-dependent characteristics of the emission spectrometer and the signal detector response. Ultrafast Transient Absorption Spectroscopy. Transient absorption spectra of the modified peridinins in various solvents were measured using a femtosecond transient absorption spectrometer system based on an amplified Ti:sapphire laser pumped at a 1 kHz repetition rate as previously described.35 Pump pulses with a duration of ∼60 fs were obtained from an OPA-800C optical parametric amplifier. Probe laser pulses were derived from a white light continuum generated by a 3 mm sapphire plate. The probe beam was overlapped with the pump beam at the sample at magic angle (54.7°) polarization. A fiber optic spectrometer from Ocean Optics (Model S2000) was used for detection. Transient absorption spectra were recorded, and the signals were averaged over 5 s, on samples adjusted to an optical density ranging from 0.4 to 0.6 at the excitation wavelength in a 2 mm cuvette. The pump wavelengths were 480 nm for C-1R-peridinin, 482 nm for C-1-peridinin, and 485 nm for D-1peridinin. The energy of the pump beam was set between 0.8 and 1 μJ/pulse in a spot size having a 1 mm diameter. This corresponds to a pump beam intensity of ∼2 1014 photons/ cm2 per pulse. Steady-state absorption spectra were recorded before and after laser excitation to confirm the integrity of the samples. Surface Explorer software (version 1.0.6) was used to correct for the chirp of the probe pulse and to subtract peaks associated with scattered excitation light in the transient absorption spectra. ASUFit v 3.0 software was used for global fitting on the basis of a sequential decay model analysis. Triplet State Transient Absorption Spectroscopy. Measurements of the triplet state spectra and lifetimes were carried out using an Edinburgh Instruments LP920 flash photolysis spectrometer on solutions placed in a 1 cm path length cuvette at room temperature as previously described.29 The experiments were performed on solutions containing Chl a as a triplet state donor at a concentration of ∼12 μM in n-hexane or acetonitrile. The carotenoid concentrations ranged from 0 to 20 μM. A single drop of pyridine was added to all the solutions to avoid aggregation of Chl a. All transient profiles were the average of 20 scans. Analysis of the kinetics data was carried out using Origin 7.5 and ASUFit v 3.0 software.
ARTICLE
Figure 2. Steady-state absorption spectra of peridinin and C-1-R-, C-1-, and D-1-peridinin analogues taken in different solvents at room temperature. All spectra were normalized.
’ RESULTS Steady-State Absorption. Steady-state absorption spectra corresponding to the S0 f S2 transition of peridinin and C-1-R-, C-1-, and D-1-peridinin derivatives in n-hexane, MTBE, ethyl acetate, 2-propanol, and methanol are shown in Figure 2. The spectral line shapes demonstrate that, for a given solvent, the S0 f S2 transition lies highest in energy for C-1-R-peridinin and lowest in energy for D-1-peridinin, with C-1-peridinin and peridinin being in the middle. This is consistent with C-1-R-peridinin having the shortest, and D-1-peridinin having the longest, πelectron conjugated chain (Figure 1). The absorption spectra of all the molecules in n-hexane show reasonably well-resolved vibrational bands, but as the solvent polarity increases, the vibronic band resolution decreases to the point where there is little, if any, vibronic structure evident in the line shapes. However, in the most polar solvent, methanol, C-1-R-peridinin, C-1-peridinin, peridinin, and D-1-peridinin display a minor feature attributable to vibronic structure near the absorption maximum (Figure 2E). The λmax values for the S0 f S2 transitions are summarized in Table 1. Steady-State Fluorescence. The fluorescence spectra of the molecules in the various solvents are shown in Figure 3. The longest molecule in the series, D-1-peridinin, in all solvents, exhibits fluorescence bands originating from both the S2 and S1 states. (See dashed lines in Figure 3.) The band near 520 nm is associated with the S2 f S0 transition, and the longer wavelength band peaking between 750 and 800 nm belongs to the S1 f S0 transition. The ratio of the intensity of S2 to S1 emission from D-1-peridinin decreases with increasing solvent polarity until the molecule is dissolved in methanol, in which case the ratio increases slightly. C-1-peridinin also shows evidence of S2 emission that, for the most part, decreases with increasing solvent polarity, but for this molecule, in all solvents, S1 emission dominates. (See dark solid line in Figure 3.) The fluorescence spectra of C-1R-peridinin in all solvents show essentially no evidence of S2 emission. (See thin solid line in Figure 3.) Similar to the steady-state absorption spectra (Figure 2), the fluorescence 4438
dx.doi.org/10.1021/jp110092c |J. Phys. Chem. B 2011, 115, 4436–4445
The Journal of Physical Chemistry B
ARTICLE
Table 1. Wavelengths of Maximum Absorption of the Ground and Excited State Spectra of C-1-R-peridinin, C-1peridinin, perdinin, and D-1-peridinin λmax/nm molecule
solvent
C-1-R-peridinin n-hexane
C-1-peridinin
peridinina
D-1-peridinin
a
S0 f S2 S1 f S0 S1 f Sn ICT f Sn 434
665
491
623
MTBE
434
654
512
615
ethyl acetate
436
676
536
603
2-propanol
454
662
545
590
methanol n-hexane
451 482
696 718
545 517
563 652
MTBE
450
726
521
647
ethyl acetate
452
732
521
646
2-propanol
476
740
522
643
methanol
471
755
526
635
n-hexane
456
716
516
657
MTBE
457
722
521
649
ethyl acetate 2-propanol
458 472
728 727
530 546
640 628
methanol
472
743
545
592
n-hexane
460
766
540
687
MTBE
461
768
540
683
ethyl acetate
464
775
543
681
2-propanol
476
781
545
676
methanol
471
779
545
622
λmax values for peridinin were taken from ref 37. Figure 4. Transient absorption spectra of C-1-R-, C-1-, and D-1peridinins in n-hexane, MTBE, ethyl acetate, 2-propanol, and methanol taken at room temperature at various times after photoexcitation of the molecules into the S2 state.
Figure 3. Fluorescence emission spectra of C-1-R-, C-1-, and D-1peridinin analogues taken in different solvents at room temperature. All spectra were normalized.
spectra of C-1-peridinin and C-1-R-peridinin in nonpolar solvents show vibronic features, the resolution of which is diminished as the solvent polarity increases; e.g., compare dark solid lines for C-1-peridinin in Figure 3A and 3E. The wavelengths
corresponding to the λmax values for the S1 f S0 transitions in the molecules are summarized in Table 1. Ultrafast Excited State Transient Absorption. Transient absorption spectra of the excited singlet states of the molecules taken in the various solvents at different delay times after excitation on the long-wavelength sides of the S0 f S2 transitions are shown in Figure 4. At short wavelengths the spectra exhibit negative peaks associated with either photobleaching of the ground state absorption or stimulated emission from the S2 state. All the transient spectra show evidence for the ICT state in the form of broad transient ICT f Sn spectral profiles that are red-shifted relative to the well-known sharper S1 f Sn transition of carotenoids.21,27,36,37 For C-1-R-peridinin in 2-propanol (Figure 4D) and methanol (Figure 4E), the S1 f Sn transition is barely discernible because it is buried under the strong, broad line shape of the ICT f Sn transition. However, C-1-peridinin and D-1-peridinin show in all solvents, but especially in n-hexane (Figure 4F,K), transient absorption line shapes where a prominent, sharp short-wavelength feature belongs to the S1 f Sn transition. For all of the molecules, as the polarity of the solvent increases, the spectra gain an increasing contribution from the broad, red-shifted ICT f Sn band; e.g., see the series of spectral traces for D-1-peridinin shown in Figure 4KO. The values of the wavelength of the S1 f Sn transition and the largest band in the ICT f Sn spectral profiles are given in Table 1. Global Fitting Analysis of the Excited State Dynamics. The dynamics of the excited states of the molecules were determined 4439
dx.doi.org/10.1021/jp110092c |J. Phys. Chem. B 2011, 115, 4436–4445
The Journal of Physical Chemistry B
Figure 5. Evolution associated difference spectra (EADS) obtained from global fitting of the transient absorption data sets shown in Figure 4.
by global fitting the spectral and temporal data sets. A sequential decay model was used in the analysis resulting in evolution associated difference spectra (EADS) as shown in Figure 5. The spectral profiles of the kinetic components illustrate the timedependent spectral evolution of the excited states. Table 2 summarizes the results of the analysis. For all of the molecules except C-1-R-peridinin in n-hexane, three kinetic components were required to fit the data sets based on a consideration of the results of a chi square (χ2) test and minimization of residuals. The first component (black traces in Figure 5), denoted τ1 in Table 2, is in all cases either shorter than or comparable to the ∼150 fs time resolution of the laser system. At long wavelengths it corresponds to a rise in amplitude which is attributable to the buildup of the S2 f Sn transition. At shorter wavelengths there is a decrease in amplitude associated with photobleaching of the ground state absorption or stimulated emission from the S2 state. These observations indicate that this kinetic component is associated with the lifetime of the S2 state which is instantaneously populated by the pump laser excitation.38 Due to the very rapid lifetime of this component and the limitations in the time response of the laser system, no effect of solvent or functional group substitution on the value of this component could be discerned. A second component denoted τ2 (red traces in Figure 5) has a much longer lifetime than τ1, ranging from approximately 1 to 13 ps depending on the molecule and the solvent. The third component, denoted τ3 (blue traces in Figure 5), required for a good fit to the data sets, also shows a
ARTICLE
strong dependence on molecular structure and solvent. For example, the values of τ3 for the molecules dissolved in n-hexane span a range from 630 ps for C-1-R-peridinin to 50 ps for D-1peridinin. This component is associated with either S1 f Sn or ICT f Sn transitions. Global analysis could not separate the contributions of these transitions due to their very similar decay kinetics (see below), with the exception of C-1-R-peridinin in n-hexane, where they were able to be resolved. For this sample only, a fourth EADS component was required for a good fit (Figure 5A). Figure 6 shows the solvent dependence of the decay kinetics of the S1 f Sn spectral bands monitored at the wavelength of maximum transient absorption. The solid lines represent best fits obtained from the global fitting analysis. Note from the data presented in Table 2 that while the τ3 values of the molecules in the nonpolar solvent, n-hexane, are strikingly different, the values for all of the molecules converge to 14 ( 5 ps in the polar solvent, methanol. Previous studies on apocarotenals, apocarotenoic acids, and peridinins having different extents of π-electron conjugation have shown similar convergence behaviors in polar solvents.2426,37 Figure 7 presents an overlay of the kinetics of the short (S1 f Sn) and long (ICT f Sn) wavelength transient absorption signals taken from C-1-R-peridinin, C-1-peridinin, and D1-peridinin in methanol. For all three molecules in this solvent, the decay kinetics of the S1 f Sn transient absorption signals were observed to be slightly slower than those associated with the ICT f Sn signals. Triplet Absorption Spectra and Kinetics. Transient absorption spectra of the excited triplet states of peridinin, C-1-R-, C-1-, and D-1-peridinin using Chl a as the triplet state donor in n-hexane and acetonitrile are shown in Figure 8. Chl a in solution by itself (dashed lines in Figure 8A,B) exhibits bleachings of the Soret (∼440 nm), Qx (∼625 nm), and Qy (662 nm) absorption bands after photoexcitation. Additionally, Chl a displays a small positive peak at ∼470 nm corresponding to transient T1 f Tn absorption. Solutions containing the peridinins in addition to Chl a showed additional peaks between 460 and 540 nm associated with triplettriplet absorption of the carotenoids. Just like the steady-state absorption spectra, the triplettriplet absorption spectra of the carotenoids taken in n-hexane (Figure 8A) exhibit a red shift of the T1 f Tn band in going from C-1R-peridinin to D-1-peridinin consistent with an increasing extent of π-electron conjugation (Figure 1). Upon changing the solvent from n-hexane to acetonitrile, the triplettriplet absorption spectra broaden substantially. (Compare panels A and B in Figure 8.) Panels C and D of Figure 8, which were constructed by subtracting the contribution of Chl a to the spectra, show the triplettriplet absorption bands of the carotenoids and the broadening in acetonitrile more clearly. The time response profiles measured at the wavelengths of Chl a Qy band bleaching at 662 nm and at the maxima of the T1 f Tn triplettriplet absorption spectral bands for the carotenoids were analyzed to obtain the triplet state lifetimes and energy transfer rates using a kinetic model previously described.29,39 The lifetimes of the carotenoids were obtained directly from the kinetic traces, e.g., from Figure 9A,B for C-1-R-peridinin. The energy transfer rate constants were obtained from plots of the observed rate constant, kobs, versus the concentration of the carotenoid, e.g., see Figure 9C for C-1-R-peridinin in n-hexane and acetonitrile probed at 662 nm. All of the triplet state kinetic data are summarized in Table 3. 4440
dx.doi.org/10.1021/jp110092c |J. Phys. Chem. B 2011, 115, 4436–4445
The Journal of Physical Chemistry B
ARTICLE
Table 2. Dynamics of the Excited States of C-1-R-peridinin, Peridinin, C-1-peridinin, and D-1-peridinina lifetime molecule C-1-R-peridinin
C-1-peridinin
peridinind
solvent b
n-hexane
τ1/fs
2.8 ( 0.8
