167
Part Four · Quantum biology

The coherence debate in photosynthesis

In April 2007 a Nature paper by Gregory Engel, Graham Fleming and colleagues reported quantum beats lasting more than 660 femtoseconds in the FMO complex, cooled to 77 K. The interpretation: excitations explore many paths at once, like a quantum search, and that might explain photosynthesis’s efficiency. It launched the modern field of quantum biology and hundreds of follow-up papers. Over the next decade, careful experiments and theory attributed the long-lived beats mostly to vibrations, not electronic coherence, and showed that electronic coherence dies too fast to matter. This chapter reconstructs how that happened, because it is a model of science correcting itself.

On the plate: two kinds of beating in a spectroscopy signal.

Two-dimensional electronic spectroscopy

Hit a sample with three ultrashort laser pulses and detect a fourth, emitted field. Fourier transforming over the delay between the first two pulses and over the detection time gives a 2D map: which frequencies are excited, and which are emitted, after a waiting time TT. Cross peaks link pigments that share energy. If a superposition of two states is created, cross-peak amplitudes oscillate in TT at the frequency difference between them: quantum beats. Their decay time measures how long the superposition lasts. The catch is that superpositions of vibrational levels, on the ground or excited electronic state, beat too. Established

Electronic or vibrational?

Show , or , with electronic dephasing time 100 fs. Electronic coherences between excitons are broadened by the same fluctuating protein environment that broadens absorption lines; their dephasing time follows from the linewidth, T2=1/(πc Δν~)T_2 = 1/(\pi c\,\Delta\tilde\nu), about 100 fs for a 100 cm⁻¹ line. Vibrational coherences of an intramolecular mode are weakly damped and can beat for picoseconds. They can also mix with electronic states, “vibronic” coherence, producing long-lived beats in cross peaks. Distinguishing them needs polarization control, isotope substitution, mutants, and careful modelling. Established

The record

2007: Engel and colleagues, beats over 660 fs at 77 K, interpreted as electronic coherence. 2010: Gitt Panitchayangkoon and colleagues report beats near physiological temperature, about 300 fs at 277 K. 2012 onward: several theory groups show that vibronic mixing can produce long-lived beats. 2017: Hong-Guang Duan, R. J. Dwayne Miller and colleagues, in PNAS, find no long-lived electronic coherence in FMO at room temperature: the long-lived beats are vibrational, and electronic coherence dephases within about 60 fs. 2017–18: Erling Thyrhaug, Jürgen Hauer and colleagues identify the coherences in FMO and find electronic coherence gone within about 240 fs even at 77 K. 2020: Jianshu Cao and seventeen co-authors from groups across the field, experimentalists and theorists, publish a joint review in Science Advances concluding that long-lived electronic coherence plays no significant functional role in natural light harvesting, while noting open questions about vibronic effects. Debated, leaning strongly against a functional role for long-lived electronic coherence.

What you now have

How 2D electronic spectroscopy detects coherence, why electronic and vibrational beats look alike, the linewidth–dephasing relation, and the 2007–2020 record of claim, challenge and revision.

· · ·

Three questions to carry with you.

What produces quantum beats in 2D spectra?tap
Superpositions of states, electronic or vibrational, oscillating at their energy difference.
Why do vibrational beats last longer?tap
Intramolecular vibrations are weakly damped; electronic coherences feel the full protein fluctuations.
What did the 2020 joint review conclude?tap
Long-lived electronic coherence has no significant functional role in natural light harvesting.

And one to work out. A transition has a homogeneous linewidth of 100 cm⁻¹. What is T2=1/(πc Δν~)T_2 = 1/(\pi c\,\Delta\tilde\nu) in femtoseconds?

T₂ ≈

Sources: G. S. Engel et al., Nature 446, 782 (2007). G. Panitchayangkoon et al., Proc. Natl Acad. Sci. USA 107, 12766 (2010). H.-G. Duan et al., Proc. Natl Acad. Sci. USA 114, 8493 (2017). E. Thyrhaug et al., arXiv:1709.00318; Nat. Chem. 10, 780 (2018). J. Cao et al., Sci. Adv. 6, eaaz4888 (2020). Checked by code: 1/(π × 2.998 × 10¹⁰ × 100) = 106 fs.

Next · Chapter 168
The radical-pair mechanism

A field 100× weaker than a fridge magnet can change the outcome of a chemical reaction.

Turn the page