Abstract
Two-dimensional electronic spectroscopy (2DES) made it possible to directly interrogate the interplay of coherent excitonic dynamics and environmental fluctuations in pigment–protein complexes by mapping excited-state couplings and their time evolution in the frequency domain[1, 2]. In the Fenna–Matthews–Olson (FMO) complex, a landmark 2D Fourier-transform electronic spectroscopy study reported “direct evidence” for “remarkably long-lived electronic quantum coherence” and associated “quantum beating signals” among excitons at 77 K, with beatings persisting for 660 fs[3, 4]. Subsequent work extended these observations to physiological temperatures, reporting that “the same quantum beating signals observed at 77 K persist to physiological temperature,” with an excited-state coherence e-folding lifetime of 130 fs at 277 K and coherence observed beyond 300 fs[2]. In parallel, physicists and chemical physicists developed open-quantum-system models showing that non-Markovian dynamics can sustain wave-like motion for several hundred femtoseconds even at 300 K and that conventional Markovian Redfield approaches may be unreliable when reorganization energies are not small compared to electronic couplings[5].
However, a major re-interpretation has emerged. Ambient-temperature photon-echo 2D spectra were argued to set an upper bound of about 60 fs for electronic dephasing, and long-lived oscillations were attributed to vibrational coherence rather than inter-exciton (purely electronic) coherence[6]. A broad synthesis likewise concludes that “interexciton coherences are too short lived to have any functional significance,” and that long-lived oscillations “originate from impulsively excited vibrations” (often ground-state Raman-active modes)[7]. The current physics-driven picture is therefore nuanced: quantum coherence in photosynthesis is experimentally real and theoretically unavoidable, but its functional role depends on which coherences are measured (optical, inter-exciton, vibronic, or vibrational) and on the microscopic structure of the system–bath interaction and spectral density[7, 8].
Introduction
A working physics definition of quantum biology is “the identification and study of quantum phenomena in biological systems,” and the field is described as being “dominated by a search for functional quantum mechanics hidden in complex biosystems”[9]. Within this broad agenda, photosynthetic light harvesting became a focal point because ultrafast experiments suggested coherent quantum dynamics in pigment–protein complexes, while theoretical analysis had to confront strong coupling between electronic excitations and nuclear motion in protein environments[10, 11]. A canonical model system for this physics program is the FMO complex, long used to study how electronic couplings enable efficient energy transfer from an antenna to a reaction center; indeed, visible-range 2D spectroscopy was explicitly developed to “directly measure electronic couplings” in FMO[12]. Early 2D measurements already demonstrated that excitation energy does not “simply cascade stepwise down the energy ladder,” but follows distinct pathways depending sensitively on the spatial character of delocalized excited-state wavefunctions—an inherently quantum-mechanical statement about the nature of the relevant eigenstates and couplings[12].
From a physicist’s perspective, FMO provides an experimentally constrained test bed for open-quantum-system theories in a regime where several simplifying approximations may fail. A widely cited concern is that, because of “strong coupling (100 cm) between the electronic excitations and nuclear motion in the protein environments” around FMO, perturbative, Markovian, and independent-bath approximations can break down, motivating non-perturbative and non-Markovian treatments[11]. The same review logic emphasizes that the closest “classical” comparator is the Förster model, which treats transfer as an incoherent rate and “neglects all coherences or superpositions between sites,” but that this can be insufficient in the strong-coupling regime[11].
Because “the net result is that coherence contributes, but in a subtle way,” a central task for physics-oriented quantum biology has become separating (i) what is directly established by spectroscopy and microscopic modeling from (ii) what is inferred about biological function[9]. In what follows, the FMO literature is organized around the experimentally driven coherence claims (2DES and related techniques), the theoretical frameworks used to model them (master equations, spectral densities, and non-Markovian methods), the environment-assisted transport paradigm, and the vibronic/vibrational reinterpretation that has reshaped the field’s consensus since the mid-2010s[7].
The 2007 2DES results
Two-dimensional electronic spectroscopy provides a frequency–frequency correlation map of excited-state structure and couplings, and it can resolve dynamical signatures such as coherent beatings by tracking how spectral features evolve with a “population” (waiting) time[1, 2]. In the 2007 FMO work, 2D Fourier-transform electronic spectroscopy was used to extend earlier 2DES investigations and “obtain direct evidence for remarkably long-lived electronic quantum coherence playing an important part in energy transfer processes” in FMO[3]. The central experimental signature was that “quantum coherence manifests itself in characteristic, directly observable quantum beating signals among the excitons” in FMO at 77 K, which was interpreted as wavelike energy transfer[3]. Crucially, the same paper highlighted that “the energy transfer mechanism is often described by semiclassical models that invoke ‘hopping’ of excited-state populations,” and positioned the 2DES beating as evidence that such models omit essential coherent dynamics[3].
The timescale emphasized in the original interpretation was that “the quantum beating lasts for 660 fs,” which was framed as surprising relative to “the general assumption that the coherences responsible for such oscillations are destroyed very rapidly”[4]. Within the same discussion, the authors argued that reproducing such long-lived coherence requires that “the protein must have a more active role in a realistic bath model,” i.e., that environment-induced fluctuations cannot be treated as simple, uncorrelated noise acting independently on each chromophore[4]. They also included an explicit diagnostic distinguishing electronic quantum beating from vibrational wavepacket motion: “if this oscillation were due to vibrational wavepacket motion, the exciton peak would be expected instead to oscillate in frequency but maintain constant volume”[4].
Although this 2007 claim was a catalyst, it immediately implied a difficult inverse problem: the experiment observes nonlinear optical response functions, not density-matrix elements directly, so mechanistic inference requires a model of how system–bath interactions generate the observed oscillatory cross-peak signals[4]. This is precisely the space in which physicists’ tools—quantum dynamics in structured environments, spectral densities, and non-Markovian master equations—became central to the field[5, 11].
Room-temperature coherence claims
A key question raised by the early work was whether similar coherence signatures persist at physiological temperature. A 2010 2D Fourier-transform electronic spectroscopy study reported that “the same quantum beating signals observed at 77 K persist to physiological temperature” and that the phase and frequency agreement indicates the “same quantum coherence at all temperatures”[2]. In the same report, an “e-folding lifetime” of 130 fs for the excited-state coherence was observed at 277 K, together with coherence “lasting beyond 300 fs,” which the authors connected to the possibility that evolution could exploit environmentally assisted quantum transport mechanisms[2]. They also proposed a microscopic explanation consistent with correlated noise: beating survives because “the energies of the excited states involved fluctuate such that the energy gap remains largely constant”[2].
Independent 2DES-based analyses sought to quantify dephasing rates for specific coherences at low temperature. A method was presented for “determining the dephasing rates of individual coherences by analysing quantum beating in the cross-peaks of 2D spectra,” with the claim that two “zero-quantum coherences” have lifetimes “of the order of a picosecond” at 77 K[13]. In the same work, explicit fit values were reported: a component with τ = 1/γ_p had Γ_1 = 1/τ_1 = 9 cm-1 and a component with τ = 1/γ_p had Γ_2 = 1/τ_2 = 14 cm-1[13]. These correspond to lifetimes of 1100 fs and 700 fs, respectively, while beating could still be visible at 1800 fs[13]. By contrast, a one-quantum (optical) coherence was reported to have τ = 100 fs, corresponding to about 100 fs[13].
A proposed physical explanation for the disparity between long-lived inter-exciton (zero-quantum) and short-lived optical coherences was that “transition energy fluctuations are correlated across the complex,” potentially due to spatially uniform protein dielectric fluctuations[13]. In such a correlated-noise picture, phase evolution of a zero-quantum coherence is insensitive to common-mode fluctuations because “introducing the same fluctuation in both transition energies ω_a and ω_b … does not affect the time propagation” of the relevant density-matrix elements, given phase evolution e-i(ω_b - ω_a)t[13]. This line of reasoning directly ties experimental observables (cross-peak beatings) to open-quantum-system structure (correlated vs uncorrelated bath couplings), and it motivates theoretical treatments that go beyond simplistic dephasing models[11, 13].
Methods and observables
What 2DES can and cannot uniquely identify
From a strictly physical standpoint, the interpretation of 2DES oscillations is underdetermined unless one can exclude vibrational contributions and disentangle pathway interferences. A later microscopic-simulation effort stated explicitly that “nonlinear spectroscopy cannot unambiguously distinguish coherent electronic dynamics from underdamped vibrational motion,” emphasizing that rigorous microscopic simulations are required to interpret the same type of signals that drove the early coherence claims[14]. Consistent with this caution, theoretical and experimental studies developed polarization and symmetry-based pulse sequences to separate “coherent quantum oscillations” from “incoherent energy dissipation,” using “fundamental symmetries of multidimensional optical signals” to design pulse sequences that distinguish the two contributions[15].
In the same symmetry-driven 2D photon-echo analysis, the bath was modeled by an overdamped Brownian oscillator spectral density with bath relaxation time about 100 fs and reorganization energy about 55 cm-1 for each bacteriochlorophyll, and coherent signatures were inferred to decay rapidly: “coherences decay within 150 fs,” while “C signals show incoherent relaxation”[15]. Moreover, the “coherent regime” was stated to last about 200 fs, with exciton oscillations having a 60–100 fs period and corresponding frequencies of roughly 100–300 cm-1[15]. These results illustrate a recurring theme: depending on the observable and analysis method, coherence times extracted from 2D signals can range from <60 fs to >1 ps, placing heavy weight on modeling assumptions about spectral density structure, disorder, and pathway separation[13, 15].
Atomistic inputs and spectral densities
A major physics contribution has been the attempt to connect experimentally observed dephasing and relaxation to atomistic models of the environment via spectral densities. One simulation program combined molecular dynamics, electronic-structure calculations, and spectral simulation to provide “an approach, without any free parameters,” in which one obtains trajectories for a time-dependent Hamiltonian containing “time-dependent vertical excitation energies … and their mutual electronic couplings”[16]. In that work, the predicted 2D spectra at 300 K were described as indicating “almost total loss of long-lived coherences” when extrapolating low-temperature observations to room temperature[16]. The same approach found that the site-energy distribution is “non-Gaussian” and that the absorption line shape is “largely determined by the non-Gaussian site energy distribution”[16].
Related atomistic studies focused on extracting spectral densities for FMO in different solvents and temperatures. Simulations in a glycerol–water mixture at 310 K and 77 K were used to “determine spectral densities that compare well with earlier experimental estimates,” and the approach emphasized a QM/MM treatment where “each BChl is treated separately” and the environment is included via partial charges in the force field[17]. At 77 K, slow solvent dynamics were reported to indicate “the presence of static disorder,” meaning disorder on time scales beyond those relevant for constructing spectral densities from bath correlation functions[17]. The same work reported that the amplitude of the resulting spectral densities is “about a factor 2–3 smaller than the earlier results,” and stressed that “electrostatic interactions of the pigments with their environment is of key importance”[17].
Theoretical frameworks
Open-system regimes and the limits of Redfield theory
A central theoretical message from the FMO literature is that the physical regime is neither purely coherent nor purely incoherent. In one prominent hierarchy-based quantum-dynamics treatment at physiological temperature, it was emphasized that in typical photosynthetic excitation-energy-transfer (EET) systems “the reorganization energies are not small in comparison to the electronic coupling,” so “the Redfield equation approach might lead to erroneous insights or incorrect conclusions regarding the quantum coherence and its interplay with the protein environment”[5]. Within that framework, the numerical results were reported to show “quantum wave-like motion” persisting for several hundred femtoseconds at physiological temperature and “coherent wave-like motions” observable up to 350 fs at 300 K[5].
The same model exhibited pronounced non-Markovian sensitivity: in a regime described as “strong non-Markovian,” the hierarchy-based equation yielded wave-like motion persisting for 550 fs at 300 K, which “cannot be reproduced” by the conventional Markovian Redfield equation[5]. In that theoretical interpretation, quantum delocalization was argued to help “overcome local energetic traps,” and the complex was explored as a possible “rectifier” for unidirectional energy flow by taking advantage of quantum coherence and a protein-tuned site-energy landscape[5].
A complementary view in the review literature contrasts the quantum-coherent picture with Förster theory: Förster is described as a closest classical model because it treats excitation transfer as an incoherent rate and “neglects all coherences,” while the strong exciton–vibration coupling calls for more sophisticated dynamics models than those predicting incoherent hopping[9, 11]. This sets up the physicists’ modeling agenda: build models that interpolate between coherent Hamiltonian dynamics and incoherent hopping while remaining faithful to experimentally or atomistically constrained spectral densities[11, 17].
Hierarchical equations of motion and non-Markovian modeling
Multiple lines of work highlight the need for non-Markovian methods. A HEOM-focused study remarked that common Redfield and Lindblad master equations “do not consider the non-markovian behaviour” of protein vibrations, which can be modeled as a phonon bath interacting with bacteriochlorophylls[18]. In that setting, HEOM was used to solve the dynamics for an FMO monomer at room temperature and to trace coherence and entanglement measures, including observations of transient entanglement among specific bacteriochlorophyll sites that “turn off before 0.5 ps”[18]. While such entanglement analyses are model dependent, they underscore that the open-system state can contain nontrivial quantum correlations on sub-picosecond timescales, and that these dynamics are sensitive to parameters such as reorganization energies and coherence “dropping” around 0.2 ps in certain parameter settings[18].
Atomistic open-system approaches also sought to reproduce experimental timescales without relying on assumed static correlations. One study combined molecular dynamics, time-dependent density functional theory, and open-quantum-system approaches to simulate EET dynamics, introduced a “novel … approach to add quantum corrections,” and reported coherence beatings lasting about 400 fs at 77 K and 200 fs at 300 K, with quantitative comparison to HEOM and other methods[19]. Notably, that work reported that “cross correlation of site energies does not play a significant role” in the energy transfer dynamics, suggesting that long-lived beatings do not necessarily require strong site-energy cross correlations in the underlying Hamiltonian fluctuations[19].
Vibronic spectral density structure and purely electronic coherence
A distinct theoretical mechanism for sustaining long-lived electronic coherence in dissipative exciton dynamics emphasizes the low-frequency behavior of the spectral density. In one study, long-lasting beatings in calculated 2D spectra were noted “ranging from 1.2 ps at 4 K to 0.3 ps at 277 K,” and an “alternative mechanism” was proposed to yield “long-lasting and purely electronic coherence” despite strong dissipative coupling[8]. The key argument was that “careful modeling of the continuous part of the spectral density towards zero frequency is essential” because it determines the pure-dephasing rate γ_p, and that for a super-Ohmic onset “J(0) = 0 and the puredephasing term vanishes γ_p = 0,” so decoherence arises only through relaxation[8]. Correspondingly, cross-peak oscillations were predicted to remain visible at 277 K for the super-Ohmic case, while they become largely reduced or absent for a Drude–Lorentz form[8].
Taken together, these theoretical results explain why the coherence debate has remained technically difficult: measured oscillations reflect a convolution of electronic couplings, disorder, vibronic structure, and bath spectral density, and competing assumptions about low-frequency spectral weight can qualitatively change predicted dephasing even when overall energy-transfer efficiency remains high[8].
Environment-assisted quantum transport
A central conceptual development from physics and quantum information theory is that noise can enhance, rather than simply suppress, transport through disordered networks. One study showed that “even at zero temperature, transport of excitations across dissipative quantum networks can be enhanced by local dephasing noise,” and described the mechanism in terms of dephasing-induced broadening of site energies, causing “broadened lines of neighboring sites [to] begin to overlap” so that population transfer is enhanced as resonant modes become available[20]. In the same analysis, it was emphasized that rapid transfer “cannot be explained from a purely coherent dynamics” and that the speed-up arises from dephasing that may even be local[20].
A complementary framework generalized continuous-time quantum walks to “nonunitary and temperature-dependent dynamics in Liouville space derived from a microscopic Hamiltonian,” within the Lindblad formalism[21]. In that approach, “interplay between the free Hamiltonian evolution and the thermal fluctuations in the environment” was argued to increase FMO transfer efficiency “from about 70% to 99%,” using a universal measure for transport efficiency and its susceptibility[21]. A later conceptual analysis provided a “universal origin” for environment-assisted quantum transport (ENAQT) in dephasing environments, stating that ENAQT arises due to two competing processes: a tendency of dephasing to make population uniform and the formation of an exciton density gradient defined by source and sink[22]. In that framework, the exciton current versus dephasing displays a non-monotonic dependence with a maximum at finite dephasing strength, “signalling the appearance of ENAQT,” and this is explicitly framed as remarkable because dephasing is dissipative yet can enhance current and energy flow[22].
The broader review literature similarly states that pure dephasing noise can “enhance both the rate and yield” of excitation-energy transfer compared to “perfectly coherent evolution,” and it provides an interference-based explanation: pure dephasing breaks phase coherence so that tunneling amplitudes no longer cancel, leading to complete transfer to a sink in the illustrative model[10]. It also articulates a “phonon antenna” principle: “matching the energy level splitting to the maximum of the spectral density of the environmental fluctuations” can optimize energy transport, directly linking the design problem to spectral density engineering and the structure of the excitonic Hamiltonian[10].
An important nuance is that ENAQT does not require long-lived entanglement. One dephasing-assisted transport analysis stated that “the presence of entanglement does not play an essential role for energy transport and may even hinder it,” reframing the transport advantage in terms of interference and dephasing rather than entanglement as a resource[23]. In steady-state Lindblad models of FMO-like networks, it is likewise found that “there are time independent coherences” even in non-equilibrium steady state, that these coherences can affect transport positively or negatively, and that adding dephasing “reduces, but does not destroy” coherent transport; moreover, “excitation transfer … can be improved by the addition of external noise” in that framework[24].
Vibronic reinterpretation
Vibronic and vibrational explanations for long-lived oscillations
A major post-2010 development has been the argument that long-lived oscillations in 2D spectra often originate from vibrational coherence, not long-lived inter-exciton electronic coherence. A vibronic-exciton model explicitly treated one vibrational mode per monomer and predicted oscillations in FMO 2D spectra with “1.3 ps dephasing times at 77 K,” tracing long-lived coherences to “superpositions of vibronic exciton states located on the same pigment”[25]. The same work emphasized that vibronic exciton coherences can be “remarkably long lived” with only minor damping on a 2 ps timescale, and described a bi-phasic decay in which an initial 200 fs decay is associated with coherences localized on different pigments, while long-lived oscillations reflect coherences localized on the same pigment[25]. Mechanistically, correlated fluctuations arise because system–bath interaction is independent of the vibrational mode state, so vibronic levels can experience “highly correlated fluctuations,” yielding slow dephasing; “intensity borrowing” is invoked to explain strong transition dipoles in vibronic states[26].
At the experimental–interpretive level, an ambient-temperature photon-echo 2D study of FMO argued that spectra “do not provide evidence of any long-lived electronic quantum coherence,” but instead “confirm the orthodox view of rapidly decaying electronic quantum coherence on a time scale of 60 fs”[6]. The reported logic used the antidiagonal cut of the 2D spectrum to estimate a homogeneous linewidth of , corresponding to an electronic dephasing time , which “sets a principle upper limit” for decay of any oscillations originating from beatings between excitonic transitions[6]. In the same analysis, oscillations in a specific region were linked to vibrational coherence: “oscillations … are related to vibrational coherence,” and their frequencies, lifetimes, and amplitudes were said to match molecular vibrational modes “and not long-lived electronic coherences”[6]. The authors therefore concluded that “any electronic coherence vanishes within a dephasing time window of 60 fs” and that “no long range coherent energy transport” is needed to explain overall efficiency[6].
A temperature-dependent study extending to very low temperatures argued that “important electronic quantum coherence only occurs at … ~20 K,” with electronic coherences persisting to 200 fs (near the antenna) and marginally up to 500 fs (near the reaction-center side), and that coherence decays faster with temperature to become irrelevant above 150 K[27]. In that work, previously reported long-lived beatings were attributed to vibrational origins: “they result from mixing vibrational coherences in the electronic ground state,” and “enhancement” previously assigned to electronic coherence was argued to be “purely caused by resonant beating of vibrational molecular modes in the electronic ground state”[27, 28]. A strong system–bath coupling picture was supported by an inferred reorganization energy of 120 cm, described as sufficient to reduce electronic coherence lifetimes and produce intermittent localization of the electronic wavefunction[28].
These results align with a broader synthesis that “interexciton coherences are too short lived to have any functional significance” and that observed long-lived coherences “originate from impulsively excited vibrations” observed in femtosecond spectroscopy[7]. For the FMO protein in particular, that synthesis reports calculated dephasing times of interexciton and optical coherences “in the range of 50 and 75 fs,” and argues that long-lived quantum beats are “inconsistent with interexciton coherence” and instead show signatures of Raman-active vibrational modes on the ground-state surface[7].
Vibronic mixing as a controllable design parameter
While the reinterpretation down-weights long-lived purely electronic coherence, it does not eliminate quantum structure from photosynthetic function. A separate experimental line emphasizes that biological control can tune vibronic mixing to steer energy transfer. In one 2DES study, energy transfer was measured in wild-type and mutant FMO under reducing and oxidizing conditions, and it was found that under reducing conditions energy transfer through two pathways is equal “because the exciton 4–1 energy gap is vibronically coupled with a bacteriochlorophyll-a vibrational mode,” while oxidation detunes the resonance, steering excitons preferentially through an indirect pathway and increasing quenching likelihood[29]. A Redfield model was used to show that the complex achieves this behavior by tuning a specific site energy via redox state of internal cysteine residues[29].
A closely related study reported that many excited-state coherences are “exclusively present in reducing conditions” and absent or attenuated in oxidizing conditions, and that their presence correlates with vibronic coupling that produces faster and more efficient energy transfer under reducing conditions[30]. The growth of multiple beating frequencies across hundreds of wavenumbers was used to argue that the beats are excited-state coherences with “mostly vibrational character,” and the results were summarized as suggesting that excitonic energy transfer proceeds through a coherent mechanism, with coherences serving as a tool to disentangle coherent relaxation from transfer driven by stochastic fluctuations[30].
Synthesis of competing timescales
The FMO coherence debate is often summarized as a clash of timescales extracted from different experiments and models. The table below collects representative coherence-related timescales and their stated interpretation in the cited literature.
The diversity of timescales in this table does not necessarily reflect experimental inconsistency; rather, it reflects that distinct coherence types (optical vs inter-exciton vs vibronic vs vibrational), distinct analysis pipelines (linewidth-based homogeneous dephasing vs cross-peak beating fits), and distinct environmental models (spectral density near , static disorder, correlated fluctuations) emphasize different physics and can yield distinct effective dephasing parameters[6, 8, 13].
Connections beyond FMO
Although FMO has been a paradigmatic system, related physics appears across photosynthetic complexes. In a plant photosystem II reaction center, 2DES combined with Redfield modeling was used to “elucidate the role of coherence” in charge separation by combining experiment and theory, and “quantum beats” were reported to be present for at least 1 ps at both room temperature and 80 K[32]. The oscillation frequencies were said to correspond to intramolecular chlorophyll vibrations and to match energy differences between exciton–charge-transfer (exciton–CT) states, supporting a resonance picture between vibrational modes and the electronic manifold[32]. In that study, dynamics were summarized as illustrating a “solid correlation between electronic coherence and ultrafast and efficient electron transfer,” and vibronic coherence was proposed to contribute essentially to high quantum efficiency[32].
Independent evidence that correlated protein environments can preserve electronic coherence comes from a two-color photon-echo experiment on a bacterial reaction center. In that system, the data revealed “long-lasting coherence between two electronic states” formed by mixing of bacteriopheophytin and accessory bacteriochlorophyll excited states, and it was argued that the coherence “can only be explained by strong correlation between the protein-induced fluctuations” in neighboring chromophores’ transition energies[33]. The conclusion was that correlated protein environments preserve electronic coherence and allow coherent spatial motion of excitation, enabling efficient energy harvesting and trapping[33].
These broader cases support a general viewpoint articulated in commentary: while “detection of coherent energy transport has fuelled claims that quantum effects make photosynthesis more efficient,” experiments indicate that “interplay between electronic and vibrational motion also sustains coherence” in charge separation, pushing the field toward vibronic and vibrational mechanisms rather than purely electronic long-range coherence as the central functional candidate[34].
Implications and open questions
A recurring implication across the physics literature is that function should not be equated with long-lived inter-exciton coherence. One synthesis states that “coherence contributes, but in a subtle way,” and argues that “more sophisticated theoretical models” are required because energy does not simply hop incoherently from molecule to molecule, implying a role for coherent effects that is not reducible to a single long-lived coherence time constant[9]. The same source also emphasizes that light-harvesting complexes are tuned so that “electronic energy gaps … match closely with vibrational energy gaps,” and that such evolutionary selection suggests that optimization of frequency resonances has functional importance—an idea consistent with phonon-antenna and vibronic-mixing pictures[9].
However, the degree to which any observed quantum-walk-like behavior is essential remains debated. A recent review notes that the “existence of quantum walk in the energy transfer is still under discussion,” and also cautions that rate enhancement from quantum random walks is “not guaranteed,” citing counterexamples in the literature and highlighting that trajectory-based simulations with quantum electrons and classical nuclei can suffice to describe FMO efficiency in some analyses[35]. This reinforces the need to specify which quantum signatures are being claimed (coherence, interference, vibronic mixing) and what classical comparator is being used[11, 35].
At the methodological frontier, microscopic modeling continues to evolve. One recent preprint reports “non-perturbative, accurate microscopic model simulations” and claims “long-lived excitonic coherences at 77 K and room temperature” on picosecond time scales, while simultaneously emphasizing that coarse-graining the spectral density “completely suppresses all oscillatory features” of coherence dynamics at 300 K, thereby underestimating quantum effects under realistic vibrational environments[14]. The same work reports that at both 77 K and 300 K “narrow peaks appear across the entire vibrational frequency range” of intra-pigment modes, which it uses as a vibronic fingerprint of structured phonon environments influencing excitonic dynamics[14]. In view of the earlier caution that nonlinear spectroscopy cannot uniquely distinguish electronic from vibrational coherence, such microscopic simulations are best seen as part of an integrated experimental–theoretical inference pipeline rather than as stand-alone resolution of the coherence debate[14].
Conclusions
Physics-driven quantum biology has transformed the study of photosynthetic energy transfer by turning a classical rate-process problem into a quantitatively constrained open-quantum-system problem, enabled by 2DES and related ultrafast nonlinear spectroscopies that map excited-state couplings and reveal oscillatory signatures[1, 2]. In FMO, early 2DES work reported quantum beatings persisting for 660 fs at 77 K and argued that such long-lived coherence challenges semiclassical hopping models and requires an active, structured protein bath[3, 4]. Follow-up experiments reported coherence signatures persisting to physiological temperatures with characteristic lifetimes on the order of 100 fs and observable beatings beyond 300 fs, motivating a large theoretical literature on correlated noise, non-Markovian dynamics, and spectral density engineering[2].
At the same time, rigorous reassessments have shown that many long-lived oscillations in 2D spectra can be explained by vibrational coherence and vibronic mixing rather than long-lived inter-exciton electronic coherence. Ambient-temperature photon-echo analyses infer electronic dephasing on the order of 60 fs and attribute observed long-lived oscillations to vibrational coherence, and comprehensive reviews likewise state that interexciton coherences are too short lived to be functionally significant and that long-lived signals originate from impulsively excited vibrations[6, 7].
The most defensible current conclusion, consistent with the cited sources, is therefore a layered one. First, quantum coherence in photosynthetic complexes is experimentally observed and theoretically expected, but its nature (electronic vs vibronic vs vibrational) is system- and observable-dependent[3, 7, 25]. Second, the functional role of quantum mechanics is more plausibly located in how protein environments and structured spectral densities enable efficient transfer via mechanisms such as ENAQT, phonon-antenna resonance matching, and tunable vibronic mixing, rather than in sustained long-range electronic coherence at room temperature[10, 20, 29]. Finally, resolving remaining ambiguities requires combined strategies: spectroscopy designed to separate pathways and coherence types, and microscopic simulations that respect the highly structured spectral densities and strong coupling regimes that invalidate overly coarse-grained or purely Markovian treatments[11, 14, 15].
