Concerted Electron-Nuclear Motion in Proton-Coupled Electron Transfer-Driven Grotthuss-Type Proton Translocation.

Photoinduced proton-coupled electron transfer and long-range two-proton transport via a Grotthuss-type mechanism are investigated in a biomimetic construct. The ultrafast, nonequilibrium dynamics are assessed via two-dimensional electronic vibrational spectroscopy, in concert with electrochemical and computational techniques. A low-frequency mode is identified experimentally and found to promote double proton and electron transfer, supported by recent theoretical simulations of a similar but abbreviated (non-photoactive) system. Excitation frequency peak evolution and center line slope dynamics show direct evidence of strongly coupled nuclear and electronic degrees of freedom, from which we can conclude that the double proton and electron transfer processes are concerted (up to an uncertainty of 24 fs). The nonequilibrium pathway from the photoexcited Franck-Condon region to the E2PT state is characterized by an ∼110 fs time scale. This study and the tools presented herein constitute a new window into hot charge transfer processes involving an electron and multiple protons.

Reversibly coupling the electrochemical potential of redox reactions to the generation of proton-motive force (PMF) is the central principle of bioenergetics in all living organisms.[1] Among a variety of mechanisms for generating PMF, two fundamental processes may be used in series. First, redox reactions involving protons may use proton-coupled electron transfer (PCET) processes, which, in addition to providing kinetically efficient, low-activation energy reaction pathways for protons and electrons, provide a thermodynamically efficient mechanism for coupling electrochemical potential to proton activity.[2−4] Second, by coupling PCET to a Grotthuss-type proton translocation, protons can be efficiently transferred over the nanoscale distances required to span the biological membrane. Examples of PCET can be found in a wide range of biochemical reactions, and Grotthuss-type processes have been suggested as mechanisms for proton transfer in many biological systems.[2−14]

A detailed understanding of how nature accomplishes the efficient conversion of redox potential to PMF via PCET and a Grotthuss-type process is necessary for it to be adapted to reengineered photosynthesis for improved yields.[15−17] In addition, proton control is technologically important for improvements in a wide variety of catalysts.[18,19] Grotthuss-type processes have also been identified as being key to extremely high charge/discharge rates in proton batteries.[20] As the full range of proton-based energy-linked processes in biology becomes better understood and biomimicry becomes more influential in technology, the value of exchanging chemical potential between redox potential and PMF is likely to also find its way into technological applications, enabling cleaner and more efficient processes.

Synchronizing the ensemble by ultrafast optical triggering and simultaneously observing the vibrational markers for electron and proton transfers has the potential to provide new insight into the coupling of PCET and Grotthuss-type processes. Recently, we investigated photodriven PCET in a biomimetic model system with the emerging two-dimensional electronic-vibrational (2DEV) spectroscopic technique[21,22] and electronic structure calculations in which electron transfer and proton transfer were found to occur on a 120 fs time scale accompanied by substantial inner sphere reorganization.[23] In that system, a nonequilibrium PCET pathway was observed and found to correspond to the evolution from the photoexcited, partly charge transferred state to the fully charge separated state accompanied via nuclear rearrangement and solvation. Because the nonequilibrium pathway proceeded without a barrier, it was possible to dynamically track the evolution of the reaction (with the presence of a barrier, only the initial and final states would be significantly populated).

Herein, we investigate a more complicated system in which a similar photodriven PCET process is coupled to a Grotthuss-type proton wire so that a second proton translocation occurs and can be characterized dynamically. The construct is porphyrin-benzimidazole-phenol-pyridine [PF15-BIP-Pyr (Scheme )]. Upon photoexcitation of the porphyrin moiety, electron transfer (ET) from the phenol to the porphyrin is coupled to both proton transfer from the phenolic proton to benzimidazole and proton transfer from benzimidazole to pyridine (Scheme ), leading to the formation of a one-electron, two-proton transfer (E2PT) product (vide infra). In this work, we leverage the application of 2DEV spectroscopy to elucidate the coupled PCET and Grotthuss-type dynamics leading to the formation of the E2PT product. We focus exclusively on the nonequilibrium pathway and find that the proton translocations are facilitated by a low-frequency vibration of the BIP-Pyr moiety as supported by a recent theoretical investigation.[24] In addition, the presence of the low-frequency modulation allows for a comparison of the phase associated with the spectroscopic signatures tracking electron and proton transfer and provides evidence that E2PT product formation is a concerted process. 2DEV spectroscopy also allows for the correlation between the electronic and nuclear degrees of freedom to be followed directly and is therefore highly sensitive to electronic-vibrational (vibronic) mixing, which is found to play a significant role in the dynamics of this proton wire.

Scheme 1

Molecular Structure of the Photoinduced Proton Wire (PF15-BIP-Pyr)

The red arrows indicate the double proton transfer (2PT), and the blue arrow indicates the electron transfer (ET) following photoexcitation (hν).

Steady state infrared spectroelectrochemical (IRSEC) measurements were performed to examine changes in the IR spectra due to oxidation and reduction of the system (Figure ) and to facilitate assignments in the 2DEV spectra discussed below. Upon oxidation, new bands at 1636, 1614, and 1535 cm–1 appeared (Figure a). These frequencies are similar to those observed in pyridinium salts, C5H5NH+ (1635, 1610, and 1535 cm–1), and are assigned to the PyrH+ ring stretching modes.[25,26] Therefore, these bands serve as markers for double proton transfer (2PT). Upon reduction (Figure b), a band at 1593 cm–1 emerges, which is assigned to the formation of a porphyrin radical anion as observed in previous studies.[23,27] This band then serves to track ET. The oxidative IRSEC measurements indicate the location of the translocated protons in the E2PT product, and the reductive IRSEC measurements identify the porphyrin radical anion in the E2PT product.

Figure 1

IRSEC spectra of PF15-BIP-Pyr upon (a) oxidation and (b) reduction. The black curves show the neutral species, and the colored curves show oxidized (red) and reduced species (blue). Spectra were recorded in a dry deuterated acetonitrile solution with 0.1 M TBAPF6. The applied potentials indicated in the figure are vs a silver wire reference electrode (see section 1.2 of the Supporting Information for experimental details).

The midpoint potentials (E1/2) for the PF15-BIP-Pyr+/PF15-BIP-Pyr and PF15-BIP-Pyr/PF15-BIP-Pyr– redox couples were estimated to be 0.85 and −0.83 V versus SCE, respectively, by cyclic voltammetry measurements (Figure S2). The E1/2 of a reference compound that can undergo only one-electron, one-proton transfer (E1PT) (BIP+-Pyr/BIP-Pyr) was measured to be 1.01 V versus SCE.[28] Assuming the center-to-center distance between the phenol and porphyrin moieties (l) to be 6.4 Å and the excited state (S1) energy (E00) to be 1.92 eV, the driving force (ΔG) can be estimated to be −300 and −140 meV for the E2PT and E1PT products, respectively (see section 2.1 of the Supporting Information for further details). The differential driving force (between E1PT and E2PT states) obtained electrochemically is consistent with quantum chemical calculations performed on the full PF15-BIP-Pyr molecule (see section 3 of the Supporting Information and Table S2). Therefore, only the E2PT product is expected to be populated at thermodynamic equilibrium in these measurements, just as previously determined for BIP-Pyr in a recent study,[28] and as expected on the basis of the IRSEC results presented above.

We now turn to the ultrafast 2DEV measurements (see section 1.4 of the Supporting Information for experimental details). Briefly, in a 2DEV experiment, visible excitation pulses prepare an ensemble of electronic/vibronic states that evolve as a function of waiting time, T, which are tracked via an IR detection pulse. The resulting spectra are then presented as (visible) excitation frequency–(IR) detection frequency correlation plots at given waiting times, allowing for the correlation between electronic and vibrational degrees of freedom to be followed directly. This correlation can be quantitatively described by the center line slope (CLS) of the spectral features, which is related to a cross correlation of vibrational and electronic dipoles.[21,29] As a result, this quantity is highly sensitive to both dynamical changes in the electronic structure and the mixing of vibronic states.[23,30−32]

Figure shows representative 2DEV spectra for PF15-BIP-Pyr. Immediately after photoexcitation, ground state bleach (GSB) features were observed at 1505, 1524, 1537, and 1564 cm–1, while photoinduced absorption (PIA) features were observed at 1514, 1530, 1544, 1586, 1604, 1627, 1648, 1665, 1682, and 1694 cm–1. Below ∼1540 cm–1, the observed spectral features reflect those of PF15.[23] However, distinct bands at 1586, 1604, and 1627 cm–1 were observed in the 2DEV spectra of this species, PF15-BIP-Pyr, and are ascribed to ET (1586 cm–1) and 2PT (1604 and 1627 cm–1) on the basis of the IRSEC measurements, indicating the formation of the E2PT product following photoexcitation. Additional PIAs were observed at >1630 cm–1 and are very likely related to the E2PT product on the basis of similarities in the observed dynamics. Although only slight spectral evolution was observed in this region in the IRSEC measurements (Figure ), the E2PT product in the 2DEV measurements is formed on the excited state potential surface, so we do not expect complete agreement. The dynamics of the features at 1586, 1604, and 1627 cm–1, corresponding to the most significant changes in the IRSEC measurements, will be the focus for the remainder of the discussion. We further note that the observed appearance of these bands within ∼90 fs (our instrument response function) suggests, on the basis of the free energy difference between E1PT and E2PT products (more driving force for E2PT by ∼160 meV), that the E1PT is only transiently formed (<90 fs), if at all. As we will show, analysis of the 2DEV spectral dynamics reveals that it is very unlikely the E1PT product is ever formed.

Figure 2

(a–d) 2DEV spectra of PF15-BIP-Pyr in deuterated acetonitrile at 150, 200, 300, and 500 fs, respectively. Positive signals (red/yellow contours) indicate ground state bleaches, and negative signals (blue contours) represent photoinduced absorptions (PIAs). Contour levels are drawn in 5% intervals. Arrows indicate the main PIA features of interest at 1586 cm–1 (ET, blue), 1604 cm–1 (2PT, pink), and 1627 cm–1 (2PT, red).

To understand the dynamics of the ultrafast, nonequilibrium formation of the E2PT product, we first focus on the excitation frequency (ωexc.) peak evolution at specific detection frequencies corresponding to either the 2PT or ET signals (Figure a; see Figure S5 for raw data). We find that the ωexc. peak evolution for these features exhibits significant evolution within the first few hundred femtoseconds (Figure a). This is in stark contrast to features that are specific to PF15 (1505 and 1514 cm–1),[23] not relevant to the E2PT product, which do not show any dynamics along ωexc. during the waiting time (Figures S5 and S6). The evolution of the 2PT and ET signals, however, is striking in that the ωexc. peak evolution appears to undergo an ∼200 cm–1 blue-shift, contradictory to the expected energetic relaxation from the Franck–Condon region due to either electronic or solvent rearrangement. The observed dynamics can be rationalized by viewing the observed ωexc. peak evolution not as a blue-shift but rather as a result of modulation by a low-frequency vibrational mode. This is reminiscent of a previous 2DEV study that revealed that a low-frequency vibrational mode can influence the ωexc. peak evolution if it is coupled to both the excited electronic state and the probed vibrational mode.[33] Indeed, we find here that the dynamics of the ωexc. peak evolution fit well and consistently to a damped cosine function (results listed in Table ) and not to a mono- or biexponential function. From Table , it is apparent that the three bands yield nearly identical fits, where the average values for the frequency and damping time of this low-frequency mode were found to be 55 ± 2 cm–1 and 112 ± 12 fs, respectively.

Figure 3

(a) ωexc. peak evolution of PF15-BIP-Pyr at 1586 cm–1 (ET, blue), 1604 cm–1 (2PT, pink), and 1627 cm–1 (2PT, red). Solid black lines indicate the fit results (listed in Table ), and the residuals for each fit are shown above panels a and b. (b) Center line slope (CLS) dynamics of PF15-BIP-Pyr at 1586 cm–1 (ET, blue), 1604 cm–1 (2PT, pink), and 1627 cm–1 (2PT, red). Solid black lines indicate the fit results (listed in Table ). In panels a and b, the time range in which visible and IR pulses overlap, <90 fs, is indicated by the shaded area.

Table 1

Parameters from the Fit of ωexc. Peak Evolution (see Figure a) to a Damped Cosine Function [f = A0+ A1 cos(ω1t + φ1)e–1]

	A0	A1	ω1 (cm–1)	φ1 (deg)	τ1 (fs)
1586 cm–1 (ET)	17090	675	56 ± 3	57 ± 7	140 ± 30
1604 cm–1 (2PT)	17127	1121	51 ± 5	57 ± 4	86 ± 15
1627 cm–1 (2PT)	17068	923	58 ± 2	49 ± 3	110 ± 13

Table 2

Parameters from a Fit of CLS Dynamics (see Figure b) to a Damped Cosine Function [f = A0 + A1 cos(ω1t + φ1)e–]

	A0	A1	ω1 (cm–1)	φ1 (deg)	τ1 (fs)
1586 cm–1 (ET)	0.27	0.73	48 ± 8	220 ± 10	180 ± 50
1604 cm–1 (2PT)	0.39	0.61	61 ± 9	210 ± 13	150 ± 70
1627 cm–1 (2PT)a	0.11	–0.89	–	–	50 ± 23

The CLS dynamics for this mode were unable to be fit to a damped cosine function.

In a first-principles molecular dynamics study of the nonequilibrium PCET dynamics of a BIP construct by Goings and Hammes-Schiffer, the vibrational modes most crucial to the proton transfer process were predicted.[24] In particular, a slow bending mode in the plane of the BIP subunit, found to dominate the inner sphere reorganization, was identified as the most important vibration for facilitating proton transfer. For the BIP-Pyr species (the one under investigation here but without the PF15 group), the frequency of this vibrational mode was determined to be 45 cm–1, specifically corresponding to an in-plane bending motion that modulates the PT distances following oxidation. Despite the fact that this investigation features the inclusion of a photoactive, electron-accepting PF15 group and is in solution rather than in the gas phase, we nevertheless identify the presence of a low-frequency vibration with a similar frequency, 55 ± 2 cm–1, in the dynamics of the nonequilibrium PCET pathway. We therefore assign the observed vibration as an in-plane bend localized on the BIP-Pyr that, when damped, promotes 2PT (where we note that this mode is quite delocalized over the BIP moiety but does not involve the PF15 portion of the molecule as is evident from the lack of any ωexc. peak evolution displayed by modes specific to the PF15 moiety as discussed above and shown in Figures S6 and S7). We emphasize that the 55 cm–1 mode specifically promotes the formation of the E2PT product (i.e., involves double proton translocation), which suggests that the E1PT product is not formed or at most is only transiently populated during the reaction.

An additional important distinction to note between the previous theoretical study and this experimental work involves the ET process. In the work of Goings and Hammes-Schiffer, the oxidation process was treated by instantaneously removing an electron from the system. However, in the photoactive E2PT species investigated here, the nonequilibrium pathway does not undergo instantaneous ET. Rather, initial photoexcitation corresponds to a π to π* electronic transition on the porphyrin, followed by an ultrafast, nonequilibrium evolution of protons and electron density giving rise to the E2PT process (see section 3 of the Supporting Information, Figure S11, and the discussion below). We also find experimentally that both the ET and 2PT features are modulated by the low-frequency mode. In fact, the frequencies and phase shifts extracted from the fit to a damped oscillator are nearly identical (Table ). The agreement in the phase shift in particular constitutes strong evidence that the transfers of the electron and two protons are concerted, even on ultrafast time scales. In addition, the error in the phase shifts can be used to estimate the degree of concertedness of E2PT product formation. On the basis of the average frequency of the low-frequency mode, 55 ± 2 cm–1, and the largest error in the phase, ±7°, we can estimate that the error in the delay between the ET and 2PT features is ∼24 fs, which we interpret to mean that the ultrafast motion of the electron and two protons is concerted within an uncertainty of ∼24 fs.

Turning to the CLS, we can understand how the coupling of this vibration to the E2PT process gives rise to vibronic mixing and influences the correlation between the electronic and vibrational degrees of freedom that are directly interrogated. Figure b shows the CLS dynamics of the ET and 2PT features. The CLS dynamics observed are strikingly similar to the ωexc. peak evolution, notably the modes at 1586 and 1604 cm–1. The discrepancy in the CLS dynamics of the feature at 1627 cm–1, which we show for the sake of completeness in Table , could be due to differences in the degree of coupling between this mode and the low-frequency mode that is known to drastically influence the CLS.[30] Another possibility is that peak overlap from weak positive features just above and below 1627 cm–1 could obscure the CLS dynamics of this mode. Both of these possibilities are supported by the fact that there is agreement between the ωexc. peak evolution for this feature and the modes at 1586 and 1604 cm–1 (Table ) despite the discrepancy in the CLS dynamics. We now focus on the modes at 1586 and 1604 cm–1. Utilizing a damped cosine fit function, we obtain an average frequency for the CLS dynamics of 55 ± 9 cm–1 and a damping time of 165 ± 60 fs, which are in agreement with the values obtained for ωexc. peak evolution (see Table for complete fit details). This indicates that the same vibrational mode is driving the CLS as was observed in the ωexc. peak evolution. Because the CLS is related to a cross correlation of electronic and vibrational dipoles, we can understand the non-zero CLS in this case as arising from vibronic mixing, which significantly alters the transition moments.[23,30] In contrast, the CLSs of the bands specific to PF15 (1505 and 1514 cm–1),[23] which are not relevant to the E2PT product, were essentially constant over our observation time window (Figure S4). We conclude that the ultrafast Grotthuss-type process in this system is driven by highly correlated electronic and nuclear evolution arising from vibronic mixing.

Interestingly, the CLS dynamics for this species differ significantly from those observed previously for PF15-BIP (an E1PT only system). In that case, the CLS dynamics reflected a monotonic rise on a time scale estimated to be 120 fs that was assigned to the evolution from the Franck–Condon region of the locally excited (LE) state to the minimum of the E1PT state, accompanied by a twisting of the dihedral angle between BIP and porphyrin moieties from 60° to 90°.[23] This dihedral twisting was found to result in an increase in the electronic dipole moment, which in turn drove the observed increase in the CLS. To rationalize and gain further insight into the differences in the PCET reaction in PF15-BIP versus PF15-BIP-Pyr, dipole moments of the E2PT state of the latter were calculated as a function of the dihedral angle between BIP and porphyrin moieties over the same range (see section 3 of the Supporting Information for further details). One can see that the electronic dipole moment again shows a concomitant increase with the dihedral angle and reaches a maximum at 90°. However, in addition, near the Franck–Condon region (i.e., with the dihedral angle fixed to that of the ground state, ∼60°), 2PT is found to correspond to ∼80% ET. Thus, the E2PT system has a larger degree of charge transfer character near the Franck–Condon region compared to that in PF15-BIP (∼50%).[23] Furthermore, the E2PT state is calculated to be exergonic even without substantial dihedral twisting (see section 3 of the Supporting Information for further details). For these reasons, the low-frequency vibration modulating the PT distances is likely more critical for triggering the concerted 2PT and ET reaction than the dihedral twisting, though we find that the latter is still necessary for complete ET (see section 3 of the Supporting Information). Together, we find that significant nuclear rearrangement is necessary for PCET; however, the specific rearrangement appears to be system-specific.

The PF15-BIP-Pyr molecule illustrates an optically triggered Grotthus-type mechanism for the transfer of two protons and one electron. As in the case of one-proton, one-electron transfers (PCETs),[23] we find an ultrafast, nonequilibrium channel for E2PT in this work, enabling the reaction to be followed in real time (cartoon representation shown in Figure ). No evidence is found for a single proton transfer intermediate, and the process is facilitated by a low-frequency mode as suggested by Goings and Hammes-Schiffer in a recent theoretical study of a related molecule.[24] This modulation of the interproton distances produces a damped cosinusoidal oscillation of the mid-IR modes assigned to the 2PT and ET components of the E2PT process, clearly demonstrating the involvement of vibronic coupling in the dynamics. Fits of the ωexc. peak evolution for the modes tracking 2PT and ET reveal identical (within error) phases for the low-frequency oscillation. This, in turn, enables an estimate of the maximum possible delay between the ET and 2PT processes, i.e., the degree to which the ET and 2PT processes are concerted. The error bounds on the phase in Table have a maximum value of ±7°, which given the frequency of ∼55 cm–1, translates to a maximum delay of ∼24 fs between the ET and 2PT processes. A more precise delineation of “concertedness” is likely not warranted because the electron distribution progressively evolves until it roughly corresponds to one whole electron transferred to the porphyrin moiety, as also discussed in our recent work.[23] Although not reported at this time, photochemically driven translocations over nanoscale distances should be possible in extended versions of this genre of constructs, and further spectroscopic studies are likely to significantly advance our understanding of the dynamics governing such systems if an optical trigger can be incorporated.[24,34−39]

Figure 4

Cartoon schematic of the ultrafast, nonequilibrium dynamics of the Grotthus-type proton wire as a function of the double proton transfer (2PT) and electron transfer (ET) coordinates. Following photoexcitation (yellow arrow), the system is led to the unrelaxed Franck–Condon region where concerted 2PT and ET lead to formation of the E2PT product on the ∼110 fs time scale. The concerted electron-nuclear motion is promoted by the damping (purple end of the arrow) of a low-frequency vibration (green end of the arrow) localized in the BIP-Pyr moiety.