Technical Comments

Comment on “Engineering coherence among excited states in synthetic heterodimer systems”

See allHide authors and affiliations

Science  06 Jun 2014:
Vol. 344, Issue 6188, pp. 1099
DOI: 10.1126/science.1250926


Hayes et al. (Reports, 21 June 2013, p. 1431) used two-dimensional (2D) electronic spectroscopy to study molecular heterodimers and reported a general mechanism for the prolongation of electronic coherences, consistent with previous interpretations of 2D spectra for light-harvesting systems. We argue that the dynamics attributed to electronic coherences are inconclusive based on experimental inconsistencies arising from limited sample characterization and insufficient control measurements.

Hayes et al. (1) report on a series of experiments on synthetic dimers to investigate the physics of electronically coupled pigments using multidimensional spectroscopy. By synthesizing rigidly linked heterodimers with tunable resonant energy gaps, they prepare systems that partially mimic the environment found in photosynthetic proteins, where long-lived coherences attributed to electronic couplings have previously been observed at ambient temperature using two-dimensional (2D) spectroscopy (24). The complication is that the amplitudes, frequencies, and decay rates of these coherences can also be interpreted as vibronic and not interpigment in origin. Studies of model systems, such as the halofluorescein assemblies in this work, are important tests for distinguishing these effects. The authors report “long” interpigment dephasing times for all heterodimers and infer that the rigidity of the molecules plays an essential role in suppressing dephasing, concluding that it is a general design principle employed in photosynthetic systems to preserve electronic coherence.

The approach is welcome and could address recent hypotheses regarding the interplay of vibronic coupling and electronic coherence in energy transfer between pigments (5, 6). However, there are major issues with both the collection and the analysis of the four-wave mixing signals that affect the assignment and subsequent interpretation of interpigment dynamics. Despite the emphasis on the simple and tractable nature of the studied heterodimers, they are insufficiently characterized, leading to uncertainty surrounding the nature of the measured coherences.

Experimental control measurements were performed on the monomeric dyes to attempt to isolate dynamics associated with the dimeric species. This, however, does not then confirm their assignment to electronic coherences. Dimerization via covalent bonds will lead to changes in relative peak amplitudes and the appearance of new peaks in resonant Raman spectra as compared with the monomers (7, 8). Without complementary infrared or Raman studies, the comparison of monomers and heterodimers using 2D spectroscopy is insufficient.

This is borne out in numerous ambiguities in the oscillatory dynamics. In figure S3 of (1), monomers B′ and C′ demonstrate oscillatory modes corresponding to roughly 600 cm−1 and 750 cm−1, yet these peaks are completely absent in monomer A′. In the mixture of monomers B′ and C′ shown in figure S4 of (1), these peaks are still present but are now accompanied by a new intense peak in the vicinity of 700 cm−1. These substantial differences in what should amount to simple control measurements cast considerable doubt on the limited interpretation of the Fourier components discussed in the text.

With these unassigned oscillatory features, and given the variability and congestion in the measured power spectra from molecule to molecule (and from monomers to dimers), the strict assignment of certain peaks to electronic coherences is unconvincing. This issue stands out dramatically for dimers AB and BC, where the “electronic” frequencies are additionally present in the mixtures of monomers in figure S4 of (1), further invalidating this method for reliably isolating contributions from electronic or nuclear dynamics.

Much of the above ambiguity could be resolved with a clearer understanding of the coupling induced by dimerization, which in the present work is completely absent. In the limit of very weak coupling (implied by the authors when choosing to neglect shifts in energy levels), the dimer absorption spectra should resemble those of an equimolar mixture of monomers, yet this is only the case for dimer BC [figure 2 of (1)], which recalls a vibronic dimer in the weak coupling regime (9, 10). These absorption lineshapes raise questions about the justification for neglecting the electronic coupling entirely. Furthermore, vibronic coupling is also neglected, although it has also been shown to modify the optical response of even weakly coupled dimers (11).

Deeper insight into these energy levels, and the transitions that link them, is precluded by the homodyne-detected nature of the data. The main strength of 2D spectroscopy is to separate signals that are spectrally overlapped in more conventional third-order spectroscopies; however, by using homodyne detection, the authors cannot fully benefit from this decongestion to isolate spectral features which reflect interpigment coherences. Previous work has shown that the analysis of even phased absorptive 2D spectra of electronic dimers requires great care (12) due to strongly overlapped excited state absorption and emission peaks, acting to suppress the signatures of interpigment coherence. This suppression results in more pronounced contributions from vibrational activity to the homodyne signal. Taken together, there is no spectroscopic evidence that excludes vibrational wave-packet motion as the explanation for the observed dynamics.

Even with these open questions surrounding the oscillatory analysis, the recovered interpigment coherences are characterized as long-lived, although they persist for merely 60 fs (90 fs for dimer AB). For dimer BC, this would imply that the electronic coherence has decayed to e−3 of its initial amplitude upon a single period of oscillation (T ~ 165 fs). With such short dephasing times, the modes assigned to electronic coherence must have initial amplitudes much larger than the long-lived vibrational modes (we estimate by at least an order of magnitude) to result in power spectral densities of comparable magnitude after apodization. However, in the time domain data shown in figures 3 and S5 of (1), no signs of such large-amplitude oscillations are observed. Moreover, given the small energy gap between B′ and C′, it is surprising that the interpigment coherence in dimer BC nevertheless dephases on a time scale that is typical for optical transitions in monomeric dyes (13, 14) if the combination of rigidity and exchange narrowing should indeed suffice to prolong coherences between excited states in such assemblies.

The conclusions drawn by Hayes et al. were cast favorably in the context of the long-lived oscillatory features observed in biological systems, but we find this comparison wanting. Instead, these results approach those observed in a rigid homodimer (15), where the interpigment dephasing times were found to be roughly identical to the homogeneous lifetime for the electronic transition (50 fs). Rather than shed new light on the very controversial issue of quantum coherence effects (16), especially in regard to light-harvesting, this work oversimplifies the relevant physics on what should be a better defined system, leading to unjustified conclusions generalizing the role of coherence in molecular assemblies.

References and Notes

  1. Acknowledgments: R.J.D.M. acknowledges financial support by the Max Planck Society and the Natural Sciences and Engineering Research Council of Canada.
View Abstract

Navigate This Article