Mimicry of carotenoid photoprotection in artificial photosynthetic reaction centers: Triplet-triplet energy transfer by a relay mechanism

Devens Gust,* Thomas A. Moore,* Ana L. Moore,* Darius Kuciauskas, Paul A. Liddell, and Brian D. Halbert

Center for the Study of Early Events in Photosynthesis, Department of Chemistry and Biochemistry, Arizona State University, Tempe, AZ 85287-1604

Abstract

     Two artificial photosynthetic reaction centers consisting of a porphyrin (P) covalently linked to both a carotenoid polyene (C) and a fullerene derivative (C60) have been prepared and found to transfer triplet excitation energy from the fullerene moiety of C-P-3C60 to the carotenoid polyene, yielding 3C-P-C60. The transfer has been studied both in toluene at ambient temperatures and in 2-methyltetrahydrofuran at lower temperatures. The energy transfer is an activated process, with Ea = 0.17 eV. This is consistent with transfer by a triplet energy transfer relay, whereby energy first migrates from C-P-3C60 to the porphyrin, yielding C-3P-C60 in a slow, thermally activated step. Rapid energy transfer from the porphyrin triplet to the carotenoid gives the final state. Triplet relays of this sort have been observed in photosynthetic reaction centers, and are part of the system that protects the organism from damage by singlet oxygen, whose production is sensitized by chlorophyll triplet states. These triads are unique in that they also demonstrate  stepwise photoinduced electron transfer to yield long-lived C· + -P-C60· - charge-separated states. Electron transfer occurs even at 10 K. Charge recombination of C· + -P-C60· - yields 3C-P-C60, rather than the molecular ground state. These photochemical events are reminiscent of photoinduced electron transfer in photosynthetic reaction centers.

________________________________________________________________________

1. Introduction

     Artificial photosynthetic reaction centers typically abstract some aspects of natural photosynthetic function and embody them in simpler, synthetic systems for detailed study. While many attempts to prepare artificial reaction centers have focused on photoinduced electron transfer,[1-8] which is at the heart of photosynthesis, other features of the natural process are also important. For example, components of the photosynthetic apparatus are subject to photodamage, especially at high light levels, and organisms have developed various photoprotective strategies to mitigate such destruction.[9-17] In this report, we discuss mimicry of one such strategy.
    One process leading to photodamage is singlet oxygen sensitization. Normal excitation of the special pair of bacteriochlorophylls in bacterial reaction centers is followed by photoinduced electron transfer to generate a bacteriopheophytin radical anion. This anion in turn reduces a quinone moiety QA, which passes an electron on to a second quinone acceptor QB. The oxidized special pair is reduced by a cytochrome, regenerating the photocatalyst. After a second photoinduced electron transfer, QB is fully reduced to the hydroquinone form, leaves the reaction center, and is replaced by a fresh QB molecule. Under conditions of high light, the special pair can receive excitation energy from the antenna system and carry out the initial electron transfer process before the quinone portion of the electron transport chain is regenerated. Under such conditions, the charge-separated state comprising the special pair radical cation and bacteriopheophytin radical anion cannot evolve in the normal way, and instead recombines to generate the triplet state of the special pair.
    Triplet states of chlorophylls are long-lived, energetic species that can react with other molecules to cause photodamage. A major mechanism for such damage is sensitization of singlet oxygen. This sensitization is an energy transfer process that occurs by an electron-exchange mechanism, requiring diffusion-induced collisions between the chlorophyll triplet and ground-state oxygen. Singlet oxygen is an extremely reactive species that, left unchecked, can devastate the organism, causing its demise. Photosynthetic organisms have therefore developed photoprotective strategies. A major one involves the quenching of chlorophyll triplet states by carotenoid polyenes. The triplet energy of carotenoids (~0.63 eV[18]) is much below that of most chlorophylls, and also below the energy of singlet oxygen (0.98 eV). Thus, if a carotenoid can quench the chlorophyll triplet state rapidly enough to compete successfully with oxygen for this species, no singlet oxygen will be produced. Bacterial reaction centers contain a carotenoid polyene, and photoprotection by this mechanism has been observed in them.
    Electron exchange energy transfer requires orbital overlap, and rapid transfer thus requires that the donor and acceptor be in van der Waals contact, or otherwise coupled relatively strongly. However, X-ray diffraction studies of bacterial reaction centers have revealed that the carotenoid is relatively far from the special pair, which is the chlorophyll triplet state of lowest energy in the reaction center. In fact, it is located in van der Waals contact with the "accessory" bacteriochlorophyll on the inactive branch of the reaction center.[19-21] How then does the carotenoid quench the triplet state of the special pair? It has been proposed that the process involves a triplet relay.[22-24] Triplet energy is transferred from the special pair to the adjacent bacteriochlorophyll in a thermally-activated step, and the resulting bacteriochlorophyll triplet is rapidly quenched by the carotenoid. This arrangement has the advantage of ensuring photoprotection while preventing undesirable reactions such as electron transfer between the oxidized or excited special pair and the polyene.
    In this paper, we report the observation of a triplet relay mechanism in a pair of carotene (C) porphyrin (P) fullerene (C60) triad artificial reaction centers (structures 1 and 2).


2. Results

2.1. Photochemistry in polar solvents

    The preparation and the photochemistry of triad 1 in polar solvents has been previously reported,[25] and will be summarized here. The absorption spectrum of 1 in 2-methyltetrahydrofuran solution is essentially a linear combination of the spectra of model carotenoporphyrin 3 and fullerene[26] 4, with maxima at 409, 450 (sh), 475, 506, 575, 630 and 705 nm. Porphyrin 5 has fluorescence maxima at 630 and 699 nm, whereas fullerene 4 has maxima at 714 and ~800 nm.
    The fluorescence emission of 1 features spectral shapes characteristic of both the porphyrin and the fullerene components, but is strongly quenched relative to that of the model compounds. Time resolved fluorescence measurements were performed on ~1 X 10-5 M solutions of 1 in 2-methyltetrahydrofuran using the single photon timing technique. Global analysis (l ex = 600 nm) of decays at 772, 787 and 795 nm, where emission is mainly from the fullerene, yielded one significant exponential decay component of 0.032 ns for C-P-1C60. Excitation at 590 nm, where the porphyrin is the major absorber, gave porphyrin fluorescence decays in the 625 - 640 nm region whose global analysis yielded one significant lifetime of 0.010 ns for C-1P-C60. This approaches the time resolution of the spectrometer. The singlet-state lifetimes of 3 and 4 are 7.22 ns and 1.35 ns, respectively, and the short lifetimes for 1 are thus consistent with the strong steady-state fluorescence quenching.
    By analogy with other porphyrin-fullerene dyads,[27-29] the very short lifetime of C-P-1C60 is interpreted in terms of photoinduced electron transfer from the porphyrin to yield C-P· + -C60· - according to step 2 in Figure 1, which shows the relevant transient states of 1 and their interconversion pathways. The short lifetime of C-1P-C60 is consistent with decay by a combination of photoinduced electron transfer to give C-P· + -C60· - (step 3) and singlet-singlet energy transfer yielding C-P-1C60 (step 1). Indeed, transient absorption studies of a porphyrin-fullerene dyad closely related to 1 on the subpicosecond time scale have demonstrated formation of the P· + -C60· - charge-separated state.[30]
    The C-P· + -C60· - charge-separated state could evolve by electron transfer from the carotenoid (step 4 in Figure 1) to give C· + -P-C60· - . This possibility was investigated by nanosecond transient absorption spectroscopy. Excitation of 1 in deoxygenated 2-methyltetrahydrofuran led to the observation of a strong carotenoid radical cation transient absorption at ~940 nm, indicative of C· + -P-C60· - . The charge-separated state has a lifetime of 170 ns and is formed with an overall quantum yield of 0.14. Interestingly, the decay of the carotenoid radical cation is correlated with the growth of an absorption band at 550 nm (t = 170 ns) which is characteristic of the carotenoid triplet state. The quantum yield of 3C-P-C60 is 0.13, and its (oxygen sensitive) lifetime is 4.9 m s. Thus, 3C-P-C60 is formed by radical pair recombination of C· + -P-C60· - (step 9), and recombination to the ground state by step 8 is minimal.
    Triad 1 in 2-methyltetrahydrofuran thus demonstrates two-step photoinduced electron transfer to generate a long-lived, energetic charge-separated state, mimicking the photoinduced electron transfer processes observed in natural reaction centers. In toluene solution, however, the behavior is quite different.

2.2. Photochemistry in toluene

      The absorption spectrum of triad 1 in toluene is similar to that in 2-methyltetrahydrofuran, with maxima at 412, 455 (sh), 485, 510, 575, 628, and 706 nm. Time-resolved emission studies[25] have shown that C-1P-C60 decays with a time constant of only 12 ps. Under these conditions, the lifetime of the singlet state of model carotenoporphyrin 3 is 7.34 ns. The fullerene emission in the 800-nm region grows in with a time constant of 12 ps, signaling singlet-singlet energy transfer from the porphyrin to the fullerene moiety. The porphyrin is acting as an antenna for the fullerene, whose extinction coefficient in the visible is quite small. The fluorescence emission from the fullerene component of triad 1 in toluene decays with a lifetime of 1.52 ns. The corresponding lifetime for model fullerene 4 in this solvent is 1.40 ns. Thus, in toluene, C-P-1C60 is not quenched relative to the model compound.
    More information about the fate of the excited states of 1 in toluene was obtained from transient absorption studies. Samples were excited at 590 nm, where most of the absorption is due to the porphyrin moiety, with ~200 fs, 30 m J pulses and interrogated with pulses from a continuum-generated white-light probe beam. No transients characteristic of the carotenoid radical cation in the 950-nm region or of the fullerene radical anion in the 1050-nm region were observed.
    The fluorescence and picosecond absorption results indicate that in toluene, C-1P-C60 decays entirely via singlet-singlet energy transfer to the fullerene to yield C-P-1C60 (step 1 in Figure 1), and that no photoinduced electron transfer occurs from either of these two excited singlet states. The unperturbed lifetime of C-P-1C60 in this solvent, coupled with the fact that the yield of intersystem crossing to the triplet state in fullerenes is essentially unity, suggests that the only significant product of decay of the excited porphyrin and fullerene singlet states of 1 in toluene should be the fullerene triplet state, C-P-3C60. Transient absorption studies on the ns time scale were undertaken to investigate this possibility. Excitation of an argon purged, ~1 X 10-4 M solution of 1 in toluene at 300 K with a 5-ns laser pulse at 590 nm gave rise to a transient absorption with a maximum at ~550 nm (Figure 2). This transient grows in after the laser flash with a time constant of 85 ns (Figure 3) and decays with a lifetime of ~4.2 m s, which decreases when oxygen is admitted to the cuvette. This transient is assigned to the carotenoid triplet state, 3C-P-C60, by comparison with spectra of related compounds.[31] The yield of 3C-P-C60 was determined by the comparative method using the meso-tetraphenylporphyrin triplet state (f T = 0.67, (e T-e G)440 = 6.8 X 104 L mol-1cm-1) as the standard and (e T-e G) = 1.4 X 105 L mol-1 cm-1 for the carotenoid. The yield was unity, within experimental error.
    Thus, excitation of 1 in toluene is expected to produce the fullerene triplet state, but the species observed by nanosecond transient absorption is the carotenoid triplet state. What is the genesis of this state? It cannot involve intersystem crossing in the carotenoid. The yield of this process in carotenoid polyenes of this type is essentially zero, the carotenoid moiety is not appreciably excited at 590 nm, and the rise time is inconsistent with this mechanism. The triplet also cannot arise from the recombination of the C· + -P-C60· - charge-separated state, as was observed in 2-methyltetrahydrofuran, because no evidence for formation of a charge-separated state in toluene was found for 1 or for a related porphyrin-fullerene dyad, as discussed above. It seems likely that the carotenoid triplet state is produced via triplet-triplet energy transfer from the fullerene triplet state, formed through normal intersystem crossing of C-P-1C60.
    The fullerene triplet state of 4 has an absorption maximum at ~690 nm.[28,32] Under the experimental conditions used for observation of 3C-P-C60 in 1, the extinction coefficients of the carotenoid and fullerene triplet states at 690 nm coupled with the recovery of the detection system from fluorescence-induced saturation at 690 nm precluded observation of the decay of C-P-3C60. Therefore, the assumption that triplet-triplet energy transfer is the mechanism of formation of 3C-P-C60 could not be verified.
    Consequently, the spectroscopic properties of a second, closely-related fullerene-containing triad, 2, have been investigated.[33] Excitation of this triad in toluene solution results in singlet-state photophysics essentially identical to those for 1, as is expected given the similar structures of the porphyrin and fullerene moieties. Transient absorption studies on the picosecond time scale similar to those described above again showed no evidence for formation of carotenoid radical cations or fullerene radical anions. Thus, photoinduced electron transfer does not occur. Nanosecond transient absorption studies in argon-purged toluene similar to those described above with excitation at 625 nm led to the observation of the carotenoid triplet absorption with a maximum at 550 nm. The triplet decays with a lifetime of 4.6 m s in the deaerated solvent, and the lifetime decreases when oxygen is admitted to the cuvette. The quantum yield of the carotenoid triplet state is unity, based on 590-nm excitation, just as is the case for triad 1. The 550-nm carotenoid triplet absorption of triad 2 has a rise time of 110 ns (Figure 3), which is similar to the 85 ns rise time observed for triad 1.
    Nanosecond transient absorption measurements of 2 at 690 nm revealed an absorption corresponding to the fullerene triplet state (l max ~690 nm [28]). The transient formed within the time response of the instrument (<10 ns, Figure 3), as would be expected for triplet formation by intersystem crossing from C-P-1C60 with a lifetime of ~1.4 ns. The decay of the transient at 690 nm was biexponential, with lifetimes of 110 ns (85%) and 4.6 m s (15%). The lifetimes of both components decreased when oxygen was admitted to the sample. The minor decay component is assigned to the carotenoid triplet state, which has a small absorption in this region. The amplitude of this component of the decay is only ~0.4% that of the carotenoid triplet-triplet absorption at 550 nm. The major component of the decay at 690 nm is ascribed to C-P-3C60. The fact that its time constant is identical to that found for the rise of 3C-P-C60 confirms that the fullerene triplet is the precursor of the carotenoid triplet state, and that the carotenoid triplet state was formed via triplet-triplet energy transfer. Based on these results, a similar path for formation of the carotenoid triplet state is assumed for triad 1. The fact that the yield of carotenoid triplet state in both molecules is unity, within experimental error, verifies that no other decay pathways for the porphyrin or fullerene excited singlet manifolds can compete with singlet-singlet energy transfer from the porphyrin to the fullerene followed by intersystem crossing in C-P-1C60.

2.3 Photochemistry at low temperatures

    What is the mechanism of the triplet-triplet transfer? Two pathways are conceivable. One of these is a single-step transfer from the fullerene to the carotenoid (step 16 in Figure 1), whereas the second involves an energy transfer relay such as those seen in photosynthetic reaction centers. In the relay mechanism, endergonic energy transfer from the fullerene to the porphyrin (step 11) is followed by rapid, exergonic transfer to the carotenoid (step 12). Excitation of the porphyrin moiety of carotenoporphyrin dyad 3 is followed by intersystem crossing to the porphyrin triplet state, which transfers energy to the carotenoid to form 3C-P within 10 ns. Thus, if the two-step pathway is followed in 1 and 2, the initial step 11 must be the rate-limiting step. The temperature dependence of the triplet energy transfer was investigated in order to help distinguish between the two mechanistic possibilities.
    Excitation of 1 in a 2-methyltetrahydrofuran glass at 77K gave rise to several transient species (Figure 4).[25] The C· + -P-C60· - state was detected via observation of the carotenoid radical cation absorption at 980 nm (Figure 4b, F ~ 0.10). None of the transient decays at 77K are true single exponentials, probably due to conformational or environmental heterogeneity. The decay at 980 nm can be fitted as two exponentials with a major component (73%) having a 1.5 m s lifetime and a minor, ~7 m s component. The 1.5 m s decay is accompanied by a corresponding rise of the carotenoid triplet absorption at 550 nm with a major 1.4 m s component. The 3C-P-C60 species (F ~ 0.07) decays with a lifetime of 10 m s (Figure 4c). A transient with a maximum absorbance at ~700 nm, corresponding to C-P-3C60, forms with the excitation pulse in a yield of 0.17 and decays with a major decay component of 81 m s (Figure 4a).
    These results show that in the glass at 77 K, C-P-3C60 is formed within 10 ns of excitation with a quantum yield of 0.17 and decays with a rate constant of 1.2 X 104 s-1. At 590 nm, ~19% of the absorbed light excites the fullerene directly. The yield of intersystem crossing in fullerene model compounds is essentially unity. At 77 K, C-P-1C60 is not quenched by electron transfer, but decays by intersystem crossing. Thus, the observed fullerene triplet state arises mainly from light absorbed directly by the fullerene moiety. The majority of the C-1P-C60 states decay by photoinduced electron transfer step 3 to give C-P· + -C60· - , which goes on to yield C· + -P-C60· - via step 4 with an overall yield of 0.10. The final C· + -P-C60· - charge-separated state slowly recombines to yield the carotenoid triplet (k8 + k9 = 6.7 X 105 s-1), which decays with k15 = 1.0 X 105 s-1. The triplet lifetime of a model fullerene similar to the fullerene moiety of 1 and 2 is 89 m s at 77 K in a 2-methyltetrahydrofuran glass, whereas the lifetime of the C-P-3C60 is 81 m s under the same conditions. Thus, triplet-triplet energy transfer in 1 does not occur at 77 K, and C-P-3C60 decays to the ground state by the usual pathways.
    The sharp contrast between the decay behavior of C-P-3C60 at ambient temperature and at 77 K suggested that a more detailed study of the temperature dependence of this decay would be fruitful. A ~1 X 10-3 M solution of 1 in 2-methyltetrahydrofuran was excited with a ~5 ns, 625-nm laser pulse and the transient absorption kinetics at 700 nm were determined at various temperatures. Illustrative decays at 16, 130 and 190 K are shown in Figure 5. The decays as determined under these conditions are two exponential, with a major, long component and a minor shorter component. In each case, the short part of the decay correlates with the decay at the carotenoid radical cation absorption maximum in the 950 nm region. This part of the decay results from transient absorption by the C· + -P-C60· - charge-separated state, which decays to 3C-P-C60. The longer component of the decay at 700 nm is assigned to the fullerene triplet state. Thus, C-P-3C60 decays as a single exponential with time constants of 3.8 m s, 55 m s, and 70 m s at 190, 130 and 16 K, respectively. Figure 6 shows a plot of the rate constant for decay of C-P-3C60 against 1000/T for 12 temperatures. It will be noted that the decay process is strongly activated down to about 150 K, and temperature invariant thereafter.


3. Discussion

    The results presented above show that in toluene solution, neither triad 1 nor the closely related 2 undergo photoinduced electron transfer, although they do so in high yield in more polar solvents. The only significant pathways for decay of both C-1P-C60 and C-P-1C60 result ultimately in the formation of 3C-P-C60 with unity quantum yield. This occurs via the transfer of singlet excitation from the porphyrin to the fullerene followed by normal intersystem crossing in the fullerene to yield C-P-3C60. Triplet energy is then efficiently transferred to the carotenoid.
    The remaining question concerns the mechanism of the transfer. The possibilities are a single-step transfer by path 16 in Figure 1, or a two-step relay involving slow transfer from the fullerene to the porphyrin via step 11 followed by rapid transfer by step 12 from the porphyrin to the carotenoid. The studies of 1 in 2-methyltetrahydrofuran allowed direct observation of the fullerene triplet state and its decay properties as a function of temperature. The temperature dependence shown in Figure 6 suggests an energy transfer process with a significant energy of activation which is effective down to about 150 K. At lower temperatures, the decay of C-P-3C60 is activationless. The activated process is exactly what one would expect for a triplet relay whose first step is endergonic. By contrast, single-step, exergonic triplet energy transfer from the porphyrin moiety to the carotenoid of a carotenoporphyrin with a structure related to 3 occurs within a few ns even at 77 K.[34]
    The activation energy Ea for triplet-triplet energy transfer may be determined from the temperature data. The solid line in Figure 6 is a fit of the rate data (kobs) to eq 1, which is

                                                                                  (1)

simply the Arrhenius equation with an additional rate constant, k0, that describes the activationless portion of the decay. The fit yields an Arrhenius activation energy of 0.17 eV, a k0 of 1.3 X 104 s-1, and a pre-exponential factor A of 9.5 X 109 s-1. This value of Ea must be a maximum for the energy difference between the fullerene triplet and the more energetic porphyrin triplet state. The energy of the lowest-lying triplet state of fullerene 4 is reported to be 1.50 eV.[35] The triplet energy of free base octa-alkylporphyrins is typically ~1.6 eV.[36] This suggests that the minimum energy of activation for the proposed triplet energy transfer relay should be about 0.1 eV, which is consistent with the measured activation energy for the transfer in 1.
    These results suggest that the triplet relay does indeed operate in triads 1 and 2 down to about 150 K. At this temperature, the activated process becomes too slow to compete with deactivation of the fullerene triplet by other mechanisms, presumably return to the ground state by intersystem crossing and phosphorescence. The natural photosynthetic triplet energy transfer relay also involves an endergonic step, and its rate is also strongly temperature dependent. The rate of transfer slows considerably below about 50 K, and at 10 K, triplet transfer to the carotenoid is not observed.[22,23,37,38]


4. Conclusions

    These fullerene-containing triads are thus able to mimic the photophysics of the triplet energy transfer relay that has been postulated to operate in bacterial reaction centers [22-24] and possibly in PSII reaction centers.[15,16] Although model systems have been prepared in the past that demonstrate both single step[4,5,31,34,39-42] and, in one example, two-step[43] triplet-triplet energy transfer to carotenoids, the fullerene systems are the first to also demonstrate other important properties reminiscent of natural reaction centers. For example, triad 1 features singlet-singlet energy transfer from the porphyrin to the fullerene moiety (antenna function), and photoinduced multistep electron transfer to give a long-lived, energetic charge-separated state in high yield. Electron transfer occurs even at low temperatures (down to at least 10 K); this behavior is unusual in reaction center models. Finally, charge recombination of C· + -P-C60· - occurs to give the carotenoid triplet state, rather than the ground state. This behavior, too, is rare in model reaction centers.


5. Experimental section

    Ultraviolet-visible spectra were measured on a Shimadzu UV2100U UV-VIS Spectrometer, and fluorescence spectra were measured on a SPEX Fluorolog using optically dilute samples and corrected. Fluorescence decay measurements were performed on ~1 X 10-5 M solutions by the time-correlated single photon counting method. The excitation source was a cavity-dumped Coherent 700 dye laser pumped by a frequency-doubled Coherent Antares 76s Nd:YAG laser.[44] The instrument response function was 35 ps, as measured at the excitation wavelength for each decay experiment with Ludox AS-40.
    Transient absorption measurements on the picosecond time scale were made using the pump-probe technique. The sample was dissolved in purified solvent and the resulting solution was circulated by magnetic stirring in a cuvette having a 2-mm path length in the beam area. Excitation was at 590 nm with 150-200 fs, 30 m J pulses at a repetition rate of 540 Hz. The signals from the continuum-generated white-light probe beam were collected by an optical spectrometric multichannel analyzer with a dual diode array detector head.[45]
    Nanosecond transient absorption measurements were made with excitation from an Opotek optical parametric oscillator pumped by the third harmonic of a Continuum Surelight Nd:YAG laser. The pulse width was ~5 ns, and the repetition rate was 10 Hz. The detection portion of the spectrometer has been described elsewhere.[46]


Acknowledgments

    This work was supported by the National Science Foundation (CHE-9709272). This is publication 343 from the ASU Center for the Study of Early Events in Photosynthesis.