Engineering opposite electronic polarization of singlet and triplet states increases the yield of high-energy photoproducts

Significance

All known natural reaction centers employ high-energy singlet states as electron donors, but artificial photosystems struggle to overcome energy losses associated with lower-energy triplet-state formation, which critically limits solar energy conversion efficiency. Our results illustrate a design principle—seemingly at work in the photosynthetic reaction center of Rhodobacter sphaeroides—that employs oppositely polarized singlet and triplet excited states to avert intersystem crossing and dramatically increase the yield of high-energy photoproducts, critical for enhanced efficiency solar cells.

Abstract

Efficient photosynthetic energy conversion requires quantitative, light-driven formation of high-energy, charge-separated states. However, energies of high-lying excited states are rarely extracted, in part because the congested density of states in the excited-state manifold leads to rapid deactivation. Conventional photosystem designs promote electron transfer (ET) by polarizing excited donor electron density toward the acceptor (“one-way” ET), a form of positive design. Curiously, negative design strategies that explicitly avoid unwanted side reactions have been underexplored. We report here that electronic polarization of a molecular chromophore can be used as both a positive and negative design element in a light-driven reaction. Intriguingly, prudent engineering of polarized excited states can steer a “U-turn” ET—where the excited electron density of the donor is initially pushed away from the acceptor—to outcompete a conventional one-way ET scheme. We directly compare one-way vs. U-turn ET strategies via a linked donor–acceptor (DA) assembly in which selective optical excitation produces donor excited states polarized either toward or away from the acceptor. Ultrafast spectroscopy of DA pinpoints the importance of realizing donor singlet and triplet excited states that have opposite electronic polarizations to shut down intersystem crossing. These results demonstrate that oppositely polarized electronically excited states can be employed to steer photoexcited states toward useful, high-energy products by routing these excited states away from states that are photosynthetic dead ends.

Photosynthetic energy transduction is limited by the energy of the initially formed excited state, which typically relaxes on an ultrafast timescale, releasing a significant portion of energy of the absorbed photon as heat. Indeed, rapid internal conversion between higher and lowest-lying singlet excited states (S_n → S₁) is responsible for a significant drop in the thermodynamic efficiency of natural photosystems (1), as is ultrafast intersystem crossing to lower-energy triplet states in their artificial counterparts (2). The ability to control the fate of the initially prepared excited state on this ultrafast timescale is notoriously challenging, in part because many close-lying vibronic excited states encumber the engineering of large energy gaps to control dynamics. However, transition rates between states are also governed by coupling interactions at surface crossings, ultimately proportional to wave function overlap (3). Here, we test the idea that electronic polarization of electronically excited states can control excited-state reaction trajectories on the ultrafast timescale, by tuning wave function overlap between competing states.

Excited-state polarization of an electron donor has classically been engineered in molecular donor–bridge–acceptor systems to promote an electron transfer (ET) reaction by pushing electron density of the donor toward an acceptor upon absorption of the incident photon (4, 5). This intuitive, one-way ET scheme is a positive design strategy: It promotes the formation of the final, charge-separated state by directly influencing the transition rate via an enhanced electronic coupling interaction between the initial and final state. However, when the density of states is congested, states competing with the target state need to be explicitly avoided. We wondered if electronic polarization could also be employed as a negative design strategy. Intriguingly, we indeed find that engineering of oppositely polarized excited states can promote population of the charge-separated state, even if the initial excited-state electron density is pushed away from the acceptor (U-turn ET). This work illustrates a means to purposefully traverse a dense excited-state singlet and triplet manifold on the ultrafast timescale: we show that antiparallel electronic polarizations of singlets and triplets can be engineered in the same chromophore to maximally preserve the energy of the incident photon in a final charge-separated state, critical for efficient artificial photosystems.

Results and Discussion

To explore excited-state electronic polarization as a positive and negative design strategy, we needed a model electron donor (D) possessing electronically excited states with substantial yet oppositely oriented polarizations. The existence of oppositely polarized charge-transfer (CT) excited states within the same chromophore is unusual. In this regard, the supermolecule RuPZn is exemplary (Fig. 1 and SI Appendix, Fig. S1) (6–8). The RuPZn chromophore consists of ruthenium(II)polypyridyl (Ru) and (porphinato)zinc(II) (PZn) units connected via an ethyne bridge. This connectivity aligns the Ru and PZn transition dipoles in a head-to-tail arrangement, enforcing extensive excited-state interpigment electronic communication that gives rise to significant CT character in the lowest three singlet excited states (S₁, S₂, and S₃) (6–11). The unusual frequency dispersion of the RuPZn molecular hyperpolarizability, highlighted in part by the large dynamic hyperpolarizability (β_λ) at 1,300-nm incident irradiation (Fig. 1B), and the small β_λ at 800 nm, indicate that S₁ has an electric dipole moment that is antiparallel to that of S₂ and S₃ (6, 9–12). S₂ and S₃ feature electron density pushed from Ru toward PZn, while S₁ features electron density oppositely polarized from PZn toward Ru (Fig. 1B). Two- and three-level models for β_λ (6, 9–12), along with energetic considerations of the CT transitions (6), support these assignments (SI Appendix). Indeed, the dipole magnitudes of S₁, S₂, and S₃ are substantial, calculated to span a 13–43-D range (12). The corresponding triplet manifold of RuPZn is similarly polarized, belying its long (44-μs) triplet-state lifetime, due to attenuated Franck–Condon overlap with the ground state (7). Although the excited states of RuPZn are highly delocalized and described by extensive configuration interaction (6, 9–12), an appealing picture that rationalizes these unusually strong singlet and triplet electronic polarizations can be painted by comparing the one-electron redox potentials of RuPZn to the energy of its excited states (Fig. 1C). For instance, a red photon possesses enough energy to formally oxidize the PZn unit and reduce the terpyridyl (tpy) unit, which is consistent with the CT character of S₁; a blue photon possesses enough energy to formally oxidize Ru²⁺ and reduce the PZn unit, consistent with the CT character of S₃.

Fig. 1.

The excited states of RuPZn have oppositely oriented electric dipole moments. (A) The electronic absorption spectrum of RuPZn displays distinct absorption manifolds into S₁ (red), S₂ (green), and S₃ (blue) excited states. Arrows demark the excitation (pump) wavelengths of RuPZn used in the pump–probe experiments of this report, which are also the two-photon-resonant wavelengths of the dynamic hyperpolarizability (β_λ) experiments (boxes) in B. (B) β_λ-spectrum (gold) of RuPZn determined from experimentally measured values (boxes) (6, 11, 12). See SI Appendix for more details. (C) One-electron oxidation and reduction potentials (vs. standard calomel electrode) of RuPZn in acetonitrile solvent. Colored arrows show that the one-electron oxidations and reductions track with the ground- to excited-state energies. These one-electron transitions qualitatively describe the CT character of the corresponding excited states, as well as the electronic polarizations described in B.

In the context of ET, selective optical excitation of a RuPZn electron donor (D → ¹D*) with blue or red light generates excited singlet states that are oppositely polarized: Such states may be used to trigger one-way or U-turn ET reactions (Fig. 2), which can be subsequently tracked via transient absorption pump–probe experiments. An ET reaction must occur on a timescale similar to that of the electronically excited-state lifetime, which is typically short-lived. For example, to employ both S₃ and S₁ states of RuPZn as distinct donor states, ET from S₃ must occur in less than 1 ps because S₃-to-S₁ internal conversion occurs in 0.7 ps (vide infra). We therefore directly linked a naphthalene diimide (NDI) electron acceptor to the PZn unit of RuPZn (Fig. 2 and SI Appendix). The steric hindrance between the PZn unit and NDI carbonyl groups imparts a large torsional angle between the planes of PZn and NDI (13, 14); consequently, the DA molecule RuPZn−NDI retains the chromophoric properties inherent to RuPZn (SI Appendix, Fig. S2). Such intimately linked donor–acceptor (DA) complexes are known to have mean-square electronic couplings on the scale of tens of meV (13). DA couplings on this scale place their ET reactions well within the adiabatic regime, without causing sufficient ground-state mixing such that the D- and A units would lose their individual spectroscopic identities.

Fig. 2.

Photoinduced ET schemes. The molecular structure of the DA molecule, RuPZn−NDI is shown, along with one-way, and U-turn ET schemes. One-way or U-turn ET events are phototriggered by excitation of the donor (D) chromophore (RuPZn) into its S₃ or S₁ singlet excited states, respectively. Structures of ancillary groups R₁ and R₂ are shown in SI Appendix, Scheme S1.

The pump–probe transient dynamics of RuPZn−NDI following electronic excitation into its oppositely polarized, RuPZn-localized singlet states (¹D*A) show that U-turn ET produces a substantially higher yield of the high-energy, singlet charge-separated state (¹D⁺A⁻) than does one-way ET (Fig. 3). A narrowband, blue pump pulse (400 ± 5 nm) prepares a one-way ET reaction from S₃ (Fig. 3A), whereas a red pump (650 ± 5 nm) poises a U-turn ET event from S₁ (Fig. 3C). These pump pulses partially depopulate the electronic ground state of RuPZn−NDI, observed as negative signal (ground-state bleach); the formation of transient excited states is revealed by positive signal (excited-state absorption). Excited-state evolution is monitored via the rise and decay of distinct spectral features, as a function of pump–probe time delay.

Fig. 3.

Excited-state evolution of RuPZn−NDI upon triggering one-way and U-turn ET events. (A and C) Pump–probe transient absorption difference (ΔAbs.) spectra of RuPZn−NDI in acetonitrile solvent following S₀ → S₃ and S₀ → S₁ excitation, respectively, at three time delays; color-coded labels denote the state that makes the dominant contribution to the transient spectrum. S₀ → S₃ excitation poises a one-way ET event, while S₀ → S₁ excitation poises a U-turn ET event. Arrows point in the direction of spectral evolution. The inverted, scaled electronic absorption spectrum of RuPZn−NDI is shown at the top of A and C. (B and D, Upper) Two kinetic traces of the pump–probe data in A and C. Gold traces primarily track the RuPZn-localized triplet state (³D*A); purple traces primarily track A⁻ absorption of the singlet and triplet charge-separated states (¹D⁺A⁻ and ³D⁺A⁻, respectively). (B and D, Lower) Lengths of the horizontal bars denote the time windows where these states are populated. Time delay is a log scale after the axis break.

One-way ET (Fig. 3 A and B) from S₃ is outcompeted by intersystem crossing to the RuPZn-localized triplet state (³D*A, cyan spectrum), which has the spectroscopic hallmark of a broad, high-oscillator-strength, near-IR absorption (Fig. 4B and SI Appendix, Fig. S3) (7, 8, 10). Ultrafast formation of ³D*A from ¹D*A leads to slow, thermodynamically viable triplet ET (orange arrows), producing the lower-energy triplet charge-separated state (³D⁺A⁻, orange spectrum), which has spectral signatures analogous to ¹D⁺A⁻ that include prominent 480- and 610-nm absorption bands associated with A⁻ (15). At 1-ps time delay (black spectrum), the large ratio of the 930-nm signal (predominantly ³D*A) to the 480-nm signal (predominantly ¹D⁺A⁻) indicates that ultrafast intersystem crossing cripples one-way ET to ¹D⁺A⁻. Indeed, the rise of the 480-nm absorption on a 130-ps timescale (Fig. 3B and SI Appendix, Figs. S3–S5), which occurs concomitantly with the loss of the 930-nm absorption, shows a large yield of ³D⁺A⁻ instead. These ³D⁺A⁻ states, produced in high yield, decay through charge recombination with a time constant of 30 ns (SI Appendix, Fig. S6).

Fig. 4.

Intersystem crossing is controlled by excited-state polarization. (A) Excitation spectrum (black) of RuPZn in acetonitrile solvent, monitoring ¹D* fluorescence at 710 nm. The spectrum is scaled such that signal at 638 nm is equal in magnitude to the absorbance band (gray). (B and C) Pump–probe transient absorption difference (ΔAbs.) spectra of RuPZn in acetonitrile solvent following S₀ → S₃ and S₀ → S₁ excitation, respectively, at two time delays, with dominant character indicated by the color-coded state labels. The inverted, scaled electronic absorption spectrum of RuPZn is shown at the top of B and C. (D) S₀ → S₁ bleaching band minimum shift of RuPZn in acetonitrile solvent, as a function of pump–probe time delay. The blue shift of this band tracks S₁ state solvation (t_s) and intersystem crossing (t_f) timescales. Time constants are from a biexponential fit. Numbers in parentheses are the relative amplitudes. (E) Time-resolved fluorescence of RuPZn, integrated over the 660–760-nm window; t_f is the emission lifetime. (F) Kinetics of intersystem crossing (ISC) and internal conversion (IC), derived from time-resolved and steady-state data in A–E.

Conversely, photopreparation of S₁, which pushes electron density of ¹D* away from A, predominantly achieves ultrafast ET to ¹D⁺A⁻ (Fig. 3C). The spectral signatures of ¹D⁺A⁻ overwhelm the 1-ps transient spectrum (black); a weak transient spectrum at 10.5-ps time delay (cyan) indicates extremely attenuated ³D*A formation, from which ³D⁺A⁻ is formed (orange spectrum). The fast rise and decay of ¹D⁺A⁻, observed at the 480-nm signature of A⁻ during time delays less than 10 ps, overshadows the slower (>100 ps) rise of A⁻ associated with ³D⁺A⁻ (Fig. 3D, purple trace). Upon S₀ → S₁ excitation, most of the excited ¹D*A population is funneled away from ³D*A and into ¹D⁺A⁻, shown also by the severely diminished 930-nm signature of ³D*A at time delays greater than 10 ps (Fig. 3D, gold trace), relative to that found after S₀ → S₃ excitation (Fig. 3B, gold trace). Indeed, the ratio of the quantum yields of formation of ¹D⁺A⁻ relative to ³D⁺A⁻ is 20:1 via U-turn ET from S₁, and only 1:2 via one-way ET from S₃. These data indicate that excited-state polarization of the asymmetrical supermolecular RuPZn chromophore is key to the vastly disparate ET yields shown here (Fig. 3; see also control experiments described in SI Appendix, Fig. S7).

The pivotal electronic factor governing singlet ET efficiency in RuPZn−NDI is spatial wave function overlap between the singlet and triplet excited states of D (¹D*A and ³D*A). In the isolated RuPZn molecule, the ¹D* states undergo intersystem crossing to a highly polarized, metal-to-ligand CT triplet state (³D*) in high yield (7, 8). While high yields of triplet-state formation are a general and characteristic feature of metal-polypyridyl chromophores, the highly polarized nature of the triplet ³D* state is a unique characteristic of the asymmetrical RuPZn supermolecule. Pump–probe transient absorption data indicate that the intersystem crossing timescale of RuPZn is subpicosecond from S₃ (Fig. 4B and SI Appendix, Fig. S1D) and ∼20 ps from S₁ (Fig. 4 C–E and SI Appendix, Fig. S1D), consistent with S₃ and ³D* possessing similar Ru-to-PZn CT character. The identical polarizations of the S₃ and ³D* electronic states ensure that the nuclear potential energy surfaces are strongly split in their crossing region (Fig. 5C), in contrast to the potential surfaces corresponding to the oppositely polarized S₁ and ³D* states (Fig. 5D). By pushing electron density toward the acceptor, one-way ET from S₃ of RuPZn−NDI is intercepted by intersystem crossing to a similarly polarized ³D*A. On the other hand, the opposite dipole moments of S₁ and ³D*A attenuate their spatial wave function overlap, dramatically slowing the intersystem crossing rate. Consequently, U-turn ET becomes not only viable from a longer-lived S₁ but also maximizes the yield of ¹D⁺A⁻ (Figs. 5 and 6).

Fig. 5.

Electronic polarization determines the fate of the initially prepared excited state. (A) Light absorption that polarizes donor (RuPZn) electron density toward the acceptor (NDI) in a one-way ET mechanism results in a low yield of singlet ET products (¹D⁺A⁻), due to fast intersystem crossing (ISC) to similarly polarized, donor-localized triplet state (³D*A). Slow triplet ET from ³D*A of RuPZn−NDI then produces triplet ET products (³D⁺A⁻). (B) In U-turn ET, donor (D) electron density is pushed away from the acceptor (A), which slows ISC to ³D*A, which is characterized by an oppositely oriented electric dipole moment. U-turn ET thereby maximizes the yield of ¹D⁺A⁻. (S_0, ground electronic state; S₁ and S₃, first and third singlet excited states, respectively. Note that ¹D⁺A⁻ corresponds to the relaxed, charge-separated singlet state.) (C and D) Qualitative potential energy diagrams describing one-way (C, S₀ → S₃ excitation) and U-turn (D, S₀ → S₁ excitation) ET schemes in RuPZn−NDI, where E denotes relative state energies and Q a nuclear polarization coordinate. Both S₁ and S₃ manifest adiabatic curve crossings with ¹D⁺A⁻. Identical S₃ and ³D*A electronic polarizations give rise to strong mixing between these states and manifest ultrafast intersystem crossing. As S₁ and ³D*A are oppositely polarized, these states are weakly mixed. Hence, S₁ excitation drives larger charge-separated product yields because ISC is suppressed.

Fig. 6.

Kinetics and yields of one-way vs. U-turn ET as a function of excitation wavelength. Time constants are derived from steady-state and time-resolved data. Yields of singlet and triplet ET products upon S₀ → S₃ (blue) and S₀ → S₁ (red) excitation are shown as percentages.

Photoinduced charge separation from both the S₁ and S₃ states occurs on the ultrafast timescale, congruent with these processes being adiabatic (Fig. 5 C and D). In the adiabatic regime, the probability of transitioning from the initial to final state is unity, and the rate does not depend on the electronic coupling. Indeed, in this regime, U-turn ET in RuPZn−NDI (0.8 ps) is slightly faster than one-way ET (1.1 ps), despite what would otherwise be considered a stronger electronic coupling for the latter (Fig. 6). Moreover, these ultrafast CS timescales indicate the excited vibrational state distribution is not thermalized. As such, the ET rate cannot be characterized using the high-temperature Marcus expression for the Franck–Condon factor (that would otherwise link an activation free energy to driving force and reorganization energy) (16). Nevertheless, we expect U-turn ET to elicit a larger reorganization energy than one-way ET, which may impact Franck–Condon factors associated with the transition. As solvent plays a large role in outer-sphere ET reactions, we expect the role of reorganization energy to become more pronounced with increasing solvent polarity.

Intuitively, engineering electronic-state wave functions to “look alike” has long been a design goal to maximize the transition rate between states (3); however, the success of U-turn ET shown here emphasizes a consideration of equal importance for engineering state wave functions to “look different,” to minimize access to undesirable reaction pathways. The ³D*A states of RuPZn−NDI eventually transition to the lower-energy triplet charge-separated state ³D⁺A⁻ (Figs. 3 and 5). It is interesting to note that for some organisms that exploit light-driven ET reactions—such as the purple bacterium Rhodobacter sphaeroides—triplet ET is not thermodynamically viable (17, 18); in this case, the formation of long-lived ³D*A in such organisms is not only energetically wasteful, it is physically harmful to the organism if the excited-state energy is not dissipated as heat.

The photophysics of RuPZn−NDI sheds light on the ostensibly counterproductive U-turn ET mechanism mustered by the photosynthetic reaction center of R. sphaeroides. In the reaction center, ET proceeds from an electronically excited bacteriochlorophyll “special pair” (¹P*, P_MP_L) to a monomeric bacteriochlorophyll acceptor (B_L) exclusively down the l branch (Fig. 7) (19, 20). While fast ET from ¹P* ensures very little intersystem crossing to ³P*, when forward ET is blocked, the triplet yield of R. sphareoides remains essentially zero (21–23). Boxer has shown that S₁ of P_MP_L has a substantial amount of P_M⁻P_L⁺ CT character (24), i.e., ¹P* is polarized away from B_L, toward the M branch. Furthermore, the ³P* excited triplet state of P_MP_L is delocalized between both bacteriochlorophylls of the special pair, with excess electron density on P_L, i.e., polarized toward B_L (25), underscoring that ¹P*A and ³P*A of the reaction center are oppositely polarized. We have shown (Figs. 3–5) that singlet and triplet spatial wave functions, sculpted this way, minimize intersystem crossing, safeguarding against ³P*A formation from ¹P*A. The opposite polarization of ³P* and ¹P* shuts down this intersystem crossing pathway, without impacting the ET rate (vide infra) since the transfer mechanism is adiabatic.

Fig. 7.

The bacteriochlorophyll special pair (green, P_LP_M) in the photosynthetic reaction center of R. sphaeroides (Protein Data Bank 1AIJ) has oppositely polarized excited singlet and triplet states. The initially prepared singlet excited state undergoes U-turn ET to the monomeric bacteriochlorophyll (blue, B_L) down the l branch to maximize the yield of singlet ET products (¹D⁺A⁻) and avoid formation of the deleterious donor-localized triplet state (³P*A).

Since high-efficiency solar energy conversion devices must operate in the ultrafast kinetic regime, we anticipate that molecular engineering of chromophores that possess oppositely polarized singlet and triplet states may be used to: (i) drive high quantum yield formation of charge-separated states, and (ii) maximally preserve the energy of the absorbed photon in the resultant charge-separated state (1). Indeed, current state-of-the-art “sensitizers” for dye-sensitized solar cells (2, 26), light absorbers for solar fuel cells (27), and materials for organic photovoltaics (28), suffer from high intersystem crossing yields to low-energy triplet states that limit solar energy conversion efficiency. Asymmetrical supermolecules, such as RuPZn, where molecular design may achieve oppositely polarized singlet and triplet states (Fig. 1 B and C) that mimic the engineering behind the reaction center special pair, offer an unprecedented level of control over donor intersystem crossing rates, while enabling fast ET events that drive the formation of high-energy, long-lived, charge-separated singlet states. Strongly coupled chromophores that possess internal oxidations and reductions driven by photons of different colors could act as photoswitches to effect desirable photophysical and photochemical processes. Here we have shown RuPZn as an exemplar that links molecular structure to unusual excited-state electronic polarizations that can be used as positive and negative design elements to control ultrafast excited-state dynamics.

Methods

Ultrafast (Femtosecond–Nanosecond) Pump–Probe Transient Absorption Spectroscopy.

Ultrafast transient absorption spectra were obtained using standard pump–probe methods (7, 8, 29–32). Following all pump–probe transient absorption experiments, electronic absorption spectra verified that the samples were robust. All reported pump–probe experiments were repeated at least three times with separately prepared samples.

Global Analysis of Pump–Probe Data.

Pump–probe transient dynamical data, collected between 410 and 1,100 nm, were globally fit via in-house MATLAB code that accounts for probe chirp (<1 ps) and a Gaussian instrument response. Further details of global fitting can be found in SI Appendix.

Time-Resolved Emission.

Magic-angle polarization time-resolved emission data were recorded using a Hamamatsu C4780 picosecond fluorescence lifetime measurement system, which utilizes a Hamamatsu Streakscope C4334 photon-counting detector, a Hamamatsu C4792-01 synchronous delay generator, and a Stanford Research Systems DG535 electronic delay generator. RuPZn was excited at 650 nm by the Ti:sapphire laser fed into the optical parametric amplifier of the pump–probe setup described above; the polarization of emission was set to the magic angle (54.7°) for these experiments. Hamamatsu HPD-TA software was used to acquire emission data in the single-photon counting mode, and its fitting module was used to fit the emission lifetime by deconvolution with the experimentally determined instrument response function (irf). The irf was measured using a scattering sample (cream dissolved in water or silica in water). Sample concentrations were adjusted to give an optical density of <0.1 at the excitation wavelength.

Calculation of Quantum Yield Ratios of ¹D⁺A⁻ to ³D⁺A⁻.

The yield of singlet ET products in RuPZn–NDI upon S₀→S₁ excitation was calculated as (1/0.8)/(1/0.8 + 1/20), with 0.8 ps being the ET time constant and 20 ps being the time constant of intersystem crossing. The yield of triplet ET products took 1/20 as the numerator, as all ³D*A triplet states proceed to the triplet ET products. To calculate the yield ratio for singlet and triplet ET products of RuPZn−NDI upon S₂ or S₃ excitation, we compared the ratios (R) of the 447 nm (ground-state bleach) transient absorption signal at 250-fs and 10-ps time delay for RuPZn and RuPZn−NDI. Due to the similar, ultrafast timescales of singlet ET in RuPZn−NDI following all excitation conditions, the population remaining at 10-ps time delay is predominately RuPZn-localized ³D*A states. The yield of singlet ET was calculated as 1 − R_RuPZn/R_RuPZn–NDI. The calculated yields of singlet ET for S₀ → S₃, S₀ → S₂, and S₀ → S₁ excitation are 32, 34, and 95%, respectively. The ¹D⁺A⁻:³D⁺A⁻ ratios are then 1:2 (S₀ → S₃ excitation), 1:2 (S₀ → S₂ excitation), and 20:1 (S₀ → S₁ excitation).

Synthetic Materials.

All manipulations were performed under argon prepurified by passing through an O₂ scrubbing tower packed with Schweizerhall R3-11 catalyst and a drying tower packed with Linde 3-Å molecular sieves. Air-sensitive solids were weighed in a Braun 150-M glovebox. Standard Schlenk techniques were employed to manipulate air-sensitive solutions. Tetrahydrofuran was purchased from Sigma-Aldrich (inhibitor free, HPLC grade) and all other solvents utilized in synthesis described in this work were purchased from Fisher Scientific (HPLC grade). Triethylamine and acetonitrile were dried over calcium hydride and distilled. All other reagents were used as received (Aldrich or Fisher). Chromatographic purification (silica gel 60, 230–400 mesh, EM Science, and Bio-Beads S-X1, 200–400 mesh, BioRad) of all newly synthesized compounds was accomplished on the benchtop.