Report

Toward control of electron transfer in donor-acceptor molecules by bond-specific infrared excitation

See allHide authors and affiliations

Science  19 Dec 2014:
Vol. 346, Issue 6216, pp. 1492-1495
DOI: 10.1126/science.1259995

Abstract

Electron transfer (ET) from donor to acceptor is often mediated by nuclear-electronic (vibronic) interactions in molecular bridges. Using an ultrafast electronic-vibrational-vibrational pulse-sequence, we demonstrate how the outcome of light-induced ET can be radically altered by mode-specific infrared (IR) excitation of vibrations that are coupled to the ET pathway. Picosecond narrow-band IR excitation of high-frequency bridge vibrations in an electronically excited covalent trans-acetylide platinum(II) donor-bridge-acceptor system in solution alters both the dynamics and the yields of competing ET pathways, completely switching a charge separation pathway off. These results offer a step toward quantum control of chemical reactivity by IR excitation.

Big impact from a well-placed shake

Since the advent of ultrashort laser pulses, chemists have sought to steer reaction trajectories in real time by setting particular molecular vibrations in motion. Using this approach, Delor et al. have demonstrated a markedly clear-cut influence on electron transfer probabilities along the axis of a platinum complex. The complex comprised donor and acceptor fragments—which respectively give and take electrons upon ultraviolet excitation—bridged together by triply bonded carbon chains linked to the metal center. By selectively stimulating the carbon triple-bond stretch vibration with an infrared pulse, the authors could induce substantial changes in the observed electron transfer pathways between the fragments.

Science, this issue p. 1492

The fundamental process of electron transfer (ET) is central to many light-driven reactions. Light-induced ET in molecules starts by excitation to a so-called Franck-Condon electronic excited state, in which nonzero vibrational levels are usually populated (13). The excess vibrational energy is rapidly lost as the excited state relaxes and nuclear rearrangements begin to occur. ET may occur on time scales commensurate with the depopulation of these “vibrationally hot” states, typically 10−14 to 10−10 s (4), allowing ET to proceed from such nonthermalized states. Previous infrared (IR) spectroscopy studies with vibrational state resolution have shown a significant dependence of ET rates on the vibrational quantum number of the reactant state (5). Taken together, these phenomena offer an enticing strategy to influence ET reactions by exciting vibrations coupled with the ET process, thus perturbing nuclear-electronic (vibronic) interactions, which are known to mediate ET in systems ranging from heterogeneous assemblies to complex biological architectures (615). If achieved, this strategy provides a means to alter transition probabilities and ultimately redefine the nature and the yields of products generated in ET processes.

The challenge of directing ET is particularly complex owing to the ultrafast time scales and multicoordinate energy landscape of this class of reactions (3). Because the energies of molecular vibrations lie in the IR region, low-energy IR light can be used to target specific vibrations and potentially affect photoreactivity—an approach successfully implemented in gas-phase photochemistry (1618). With regard to manipulating photoinduced ET processes, the use of IR light complements existing strategies employing either chemical modifications—for example, of photosynthetic reaction centers (19)—or nonlinear laser methods (2022) that use ultrashort pulsed laser light in the ultraviolet (UV) or visible spectral regions.

Control of ET processes using IR excitation is underpinned by a strong theoretical framework pertaining to vibronic interactions in the nonadiabatic (weakly electronically coupled) electron tunneling medium. This framework introduces multiple vibronic coupling pathways between electron donor and acceptor, whereby the interference between the electrons propagating through the different pathways directs ET. Theoretical analyses of vibronic tunneling pathways in model molecules by Beratan and co-workers predicted that in inelastic ET reactions, where a tunneling electron may interact and exchange energy with a bridge-localized vibrational mode, IR driving of pathway-specific vibrations can effectively label and select the ET path (2325). Such experiments would thus offer the means to control ET reaction outcomes using external IR excitation.

Indeed, vibrational excitation with broadband (~120 cm−1) IR pulses that perturbed the H-bonding network linking donor and acceptor molecules was shown to slow down an intermolecular charge-separation process and decrease its yield by ~1.8% (26). These elegant experiments successfully demonstrated the viability of using IR pulses to change the course of an ET reaction.

Our alternative approach to direct ET uses a transform-limited narrowband (~12 cm−1, 1.5 ps) (27) IR pulse to achieve mode-specific excitation of intramolecular vibrations directly coupled to ET in a covalently bound donor-bridge-acceptor (DBA) molecular triad. The designed DBA triad 1, based on a platinum(II)-trans-acetylide motif (28, 29), PTZ-CH2-Ph-C≡C-Pt-C≡C-NAP (30), enforces directionality of ET between a powerful electron donor, phenothiazine (PTZ), and an electron acceptor, naphthalene-monoimide (NAP) (Fig. 1A). 1 represents an ideal substrate for IR modulation as the high-frequency C≡C stretching vibrations of the bridge at ~2100 cm−1 have previously been identified as the ET reaction coordinate in related Pt(II) acetylides (31) and are sufficiently separated in energy and space from the nearest intense IR absorption [~1700 cm−1, ν(C=O) of the acceptor] to ensure selective vibrational excitation.

Fig. 1 Schematic of the experiment.

(A) DBA triad PTZ-CH2-Ph-C≡C-Pt-C≡C-NAP (1). (B) Pump-pump-probe pulse sequence used in the IR-modulation experiments. (C) The photophysical picture of energetics and dynamics of photoinduced pathways in 1 in CH2Cl2 solvent at room temperature. There, the vibrationally hot DB+A manifold branches over three decay pathways to lower electronic states (30); “X” denotes the pathway suppressed by intermediate IR excitation. All figures in the paper follow the same color coding.

The photoinduced processes in 1 have been characterized by a combination of time-resolved infrared (TRIR) and electronic transient absorption spectroscopies (30) (see Fig. 1C). The excited-state dynamics are illustrated in Fig. 2, C and D, using TRIR data, collected with a {UVpump, IRbroadband-probe} pulse sequence. Absorption of the UV pump partially depopulates the ground electronic state, observed as negative peaks (ground-state bleach), whereas the formation of transient excited states is manifested by positive peaks (excited-state IR absorption) (32).

Fig. 2 Excited-state evolution of 1 without and with IR pump in CH2Cl2 at room temperature.

(A and B) Ground-state Fourier-transform IR (FTIR) spectra showing vibrational modes before ET is initiated. (C and D) TRIR spectra after 400-nm excitation at representative UV:IRprobe delays T. Corresponding kinetics are in fig. S2. (E to H) Difference spectra ΔAbs(IRpumpON − IRpumpOFF) at various UV:IRprobe delays T with UV:IRpump delay fixed at t = 2 ps, and IR-pump at 1908 cm−1 (pulse sequence in Fig. 1B). Arrows indicate formation or decay of a band assigned and color-coded according to the excited states involved; curved arrows represent growth followed by decay of a band.

Electronic excitation of 1 with a 400-nm, ~50-fs UV pump initiates ET from the Pt-acetylide center to the NAP acceptor and populates a vibrationally hot charge-transfer manifold, DB+A. This nonthermalized manifold decays over a range of lifetimes from <200 fs to 14 ps, branching over three separate charge transfer pathways (30): forward ET from the PTZ donor, forming the full charge-separated state 3D+BA (10% yield, blue arrows in Fig. 2, main IR-band positions 1550, 1605, and 2105 cm−1), which subsequently decays to the ground state with a 1-ns lifetime; population of a long-lived π−π* intraligand triplet state of the NAP acceptor, 3A (57% yield, 190-μs lifetime, red arrows, main IR-band positions 1590, 1638, and 1950 cm−1); and back ET reforming the ground-state DBA (GS, gray arrows in Fig. 2).

Excited-state branching over multiple ET pathways on the picosecond time scale fulfills an essential requirement to achieve the IR modulation of photoprocesses in 1. These experiments were performed as follows. Electronic excitation of 1 with a UV pump prepares the vibrationally hot DB+A. A tunable narrowband IR pump is then applied at variable time delays t after the UV pump (Fig. 1B), exciting selected vibrations in DB+A. The product states generated in all reaction pathways are quantified with a broadband IR probe (~400 cm−1) at variable time delays T. Overall, an ultrafast three-pulse {UVpump, IRnarrowband-pump, IRbroadband-probe} sequence (33) is applied to 1 to first initiate ET, then selectively excite vibrations in the bridging unit, and finally probe the changes in the charge transfer dynamics and pathways created by the intermediate IR pump (Fig. 1, B and C). The changes in IR absorption were recorded as ΔAbs(IRpumpON − IRpumpOFF), thereby solely extracting the effect of the IR excitation.

Figure 2, E to H, shows the effect of exciting the ν(C≡C) bridging vibration in DB+A on the photoprocesses in 1. Clear difference signals from all product states that evolve with time indicate that the IR pump has an effect on both the dynamics and on the relative population of transient states (34). When the IR probe arrives at early time delays T after the UV pump (Fig. 2, E and F), the transient IR spectra display pronounced difference signals that indicate a decrease in DB+A (black arrows), and an increase in 3A (red arrows) and the GS (gray arrows) relative to their population without the IR pump. Thus, IR excitation accelerates charge recombination from the gateway state DB+A. The lack of signals associated with the full charge-separated state 3D+BA at early time delays T suggests no acceleration of forward ET from the donor.

The key observation revealed in Fig. 2, G and H, is that difference signals from the product states caused by the intermediate IR pump persist long after the excitation. Thus, at an IRpump:IRprobe delay (T t in Fig. 1B) of ~98 ps (T = 100 ps, t = 2 ps), the positive signals at 1950, 1638, and 1590 cm−1 indicate that IR excitation increases the population of the 3A state. The negative signals at 2105, 1605, and 1550 cm−1 indicate that less 3D+BA is formed. By T = 2.2 ns, most of the 3D+BA (1-ns lifetime) has decayed to the GS. Consequently, the initial decrease in the amount of 3D+BA caused by the IR excitation translates to a lack of ground-state recovery on the nanosecond time scale, resulting in the observed negative GS signals. Thus, the system retains memory of the IR intervention throughout its electronic excited-state lifetime.

The observed IR-modulation effect is mode-specific and is proportional to the IR band intensity of the ν(C≡C) mode in the DB+A state. This conclusion is based on the following: (i) The magnitude of the effect at different IR-excitation frequencies follows the transient IR-absorption profile of the ν(C≡C) vibration in DB+A (Fig. 3A); (ii) the magnitude of the effect also depends on the UV:IRpump delay used. This dependence follows the decay kinetics of DB+A in TRIR experiments (Fig. 3B); (iii) IR excitation of the acceptor-localized ν(C=O) vibration in the same DB+A excited state or of a solvent IR absorption band has no effect (fig. S4), confirming mode specificity and ruling out the influence of local temperature effects.

Fig. 3

Action spectrum of IR modulation, scanning IR-pump frequency, and UV:IRpump delay t. (A) Overlay of the normalized TRIR spectrum of 1 at T = 3 ps (—) with the relative magnitudes of population change ΔAbs(IRpumpON − IRpumpOFF) for 3A (1950 cm−1) (circles) and 3D+BA (negative projection, 2105 cm−1) (triangles) when scanning the IR pump frequency from 1750 to 1950 cm−1 in 10 cm−1 steps. (B) Normalized overlay of a TRIR decay trace at 1908 cm−1 (—) with ΔAbs(IRpumpON − IRpumpOFF) for 3A (circles) and for the negative projection of 3D+BA (triangles) at different UV:IRpump delays t. (C) Source spectral data for (B).

A quantitative investigation of the effect of the IR excitation revealed that every molecule that absorbs an IR photon has its product states redefined. Without IR excitation, the yield of the full charge-separated state 3D+BA is 10% (30). Normalizing to the number of IR photons absorbed by molecules in the gateway DB+A state, we infer that the absolute yield change per IR photon absorbed by the sample is –10 ± 1% for the formation of 3D+BA and +10 ± 1% for 3A (from 57 to 67%) (35). Thus, within experimental error, the data indicate that the pathway to 3D+BA is fully suppressed; forward ET from the donor is completely switched off by selective IR excitation.

In contrast, in a control bridge-acceptor system 2 without the PTZ donor, Ph-C≡C-Pt(PBu3)2-C≡C-NAP (30, 36), the IR excitation of the ν(C≡C) in the B+A state influences only the rates of ET, but not the product yields (fig. S7). This difference between 1 and 2 suggests that product yield changes in 1 are due to a process specifically affecting the competitive ET pathways between forward electron transfer from the donor and back electron transfer from the acceptor, i.e., between charge separation and recombination across the bridge.

The observed increased rates of ET from higher vibrational levels of a charge-transfer state in both 1 and 2 are consistent with previous studies on the effect of vibrational quantum state on ET dynamics (5, 37, 38). However, the major changes in product yields and complete suppression of ET in 1 may be better rationalized within the framework of vibronic tunneling pathways (2325). The charge transfer state DB+A in 1 can be visualized as a vibronic state from which two ET pathways are available: charge recombination and charge separation. Upon IR excitation of the bridging C≡C oscillator, the latter can exchange energy with a tunneling electron (24) and perturb the interferences governing the two ET pathways, modulating ET propensity nonuniformly across the available paths and in this case achieving the observed elimination of the full charge-separated state. From a more general perspective, when ET is mediated by molecular bridges in donor-acceptor systems (39, 40), the bridge structure and its thermal and nuclear motions modulate ET probability (25, 41) by affecting a number of parameters, including electronic and vibronic coupling interactions, which in turn may be affected by IR excitation (42).

To explore specifically the effect of the IR excitation of the ν(C≡C) mode in 1 on the relative energies of the excited states involved in ET processes, we performed preliminary time-dependent density functional theory (TD-DFT) calculations for a range of CC distances of the -Pt-CC-NAP fragment (35). The CC distances were varied stepwise across the range ± 0.15 Å from the equilibrium distance calculated for the ground-state geometry, and at each distance the energies of multiple excited states were calculated. The results of these calculations for the lowest three excited states (Fig. 4) show a DB+A/3D+BA state crossover at C≡C bond distances slightly longer than that calculated for the equilibrium geometry. It is therefore feasible that the IR excitation–induced population of a higher vibrational state of the acetylide bridge mode will lead to a transient increase in the average bond length, which could drive the system across this cross-over point along the C≡C reaction coordinate, contributing to the experimentally observed suppression of 3D+BA (43).

Fig. 4 Calculated potential energy curves and corresponding electrostatic potential maps for 1 in CH2Cl2.

The curves demonstrate a crossing of states at C≡C bond lengths of the Pt-C≡C-NAP (BA) fragment ~0.04 Å away from equilibrium. Details are given in fig. S8 and table S1. Electrostatic potential maps illustrate the nature of the excited states by the distribution of electron density upon excitation. Charges are color-coded from –0.1 (red) to +0.1 (blue).

The above results demonstrate a simple and flexible strategy to influence the outcomes of light-induced electron transfer by IR excitation of intramolecular vibrations and suggest the features of molecular design required to achieve such an effect. We show that ultrafast mode-specific IR excitation of bridge vibrations coupled to electron transfer in a trans-acetylide Pt(II) donor-bridge-acceptor system in solution can control competing ET channels and ultimately switch between different product states. Direct experimental observation of such effects provides a test bed for theoretical frameworks and can contribute to the understanding of electron transfer processes under nonequilibrium conditions. Overall, directing charge transfer by vibrational excitation may provide hitherto inaccessible insight into chemical mechanisms, in a step toward quantum control of chemical reactivity.

Supplementary Materials

www.sciencemag.org/content/346/6216/1492/suppl/DC1

Materials and Methods

Figs. S1 to S8

Table S1

References (4457)

References and Notes

  1. To influence ET processes by pulsed IR excitation, both IR excitation and ET should occur on time scales shorter than or comparable to that of vibrational relaxation, i.e., typically a few picoseconds, while the minimum ~1.5-ps IR excitation pulse length is required to achieve the sub–12-cm−1 pulses necessary for selective vibrational excitation. The chemical system and the experimental setup used in this study fulfill these requirements.
  2. A broad transient signal appears as a baseline offset that decays with the lifetime of the initially formed DB+A- excited-state manifold. This signal is due to a broad electronic absorption profile of the initial charge-transfer state that extends into the mid-IR.
  3. Pumping the C≡C ν = 0→1 transition with the IR pulse in DB+A is expected to populate ν = 1. A bleach of the ν(0→1) band and a corresponding ν(1→2) transient signal anharmonically shifted to lower energies are expected, as well as similar signals for vibrational modes coupled to the C≡C mode. These signals are characteristic of two-dimensional IR (2DIR) and transient-2DIR spectra (33), which may use similar experimental setups. However, owing to vibronic interactions between different electronic states in 1, the IR pump gives rise to signals associated with electronic states other than the pumped DB+A-.
  4. Materials, methods, and additional information are available as supplementary materials on Science Online.
  5. In diad 2, the excited state populated as a result of 400-nm excitation is B+A, an equivalent to the DB+A branching state in 1. However, in 2 no electron donor is present, and therefore no forward electron transfer is possible.
  6. Excitation of the high-frequency ν(C≡C) mode will affect other vibrations in its vicinity, which may in turn affect the nature and ordering of the excited states. Presently, exploring this multidimensional energy landscape is computationally intractable.
  7. Acknowledgments: We are grateful to M. Newton (Brookhaven National Laboratory), D. Phillips (Imperial College London), A. Orr-Ewing (Univ. of Bristol), D. Clary (Univ. of Oxford), and G. Worth (Univ. of Birmingham) for comments on the manuscript; to I. Rubtsov and R. Schmehl (Tulane University) for discussions; and to the Engineering and Physical Sciences Research Council, Univ. of Sheffield, and Science and Technology Facilities Council for support.
View Abstract

Navigate This Article