Why is the
grass green?

Grass contains chlorophyll, held inside chloroplasts in the mesophyll cells beneath the leaf's transparent upper skin. Chlorophyll absorbs blue-violet and red light strongly and green light weakly. The green light that is not absorbed is scattered back out of the leaf or transmitted through it. Some of that light reaches your eye, so the grass looks green.

That is a real and correct answer. The measured spectra in the bench above make it quantitative: at 550 nm, chlorophyll a in solution absorbs at 2.5% of the strength it shows at 428 nm. The molecule is not slightly less interested in green light; it is almost transparent to it, in solution. What happens inside a leaf is a different and more interesting number, and it is the subject of the next level.

Two things to hold onto before going deeper. Chlorophyll is doing a job, not decorating: it converts light energy into chemical energy, and its color is a side effect of how it does that job. And the word "green" describes your visual system's response to the relative mix of wavelengths leaving the leaf, not a property the leaf possesses on its own.

The textbook picture invites a wrong inference: if chlorophyll barely absorbs green, then leaves must bounce green light away. Measured with an integrating sphere, an ordinary land-plant leaf absorbs 50% to 90% of light at 550 nm (lettuce at the low end, evergreen broad-leaved trees at the high end), against 80% to 95% for blue and red. The gap between "2.5% of peak" for the molecule and "50 to 90%" for the leaf is the whole story of this level.

Two optical effects, both named in the leaf-physiology literature, explain the gap:

• Sieve effect. Pigment is packed into chloroplasts, not spread evenly. At strongly absorbed wavelengths a photon that misses a chloroplast passes through a clear gap, so packaging reduces absorptance relative to a uniform solution. At weakly absorbed wavelengths the loss is small, because a single chloroplast was never going to capture the photon anyway.
• Détour effect. Cell walls, air spaces and organelles scatter light, lengthening its path through the leaf by roughly threefold in a standard model. For red and blue that adds only about 11% to absorptance (they were nearly all absorbed already). For green, where each pass absorbs little, a three-times-longer path multiplies the chance of absorption, and the gain is large.

The consequence is that green light penetrates deeper and is absorbed lower down. On a per-chlorophyll basis, the chloroplasts at the bottom of a leaf receive well under 20% of the green light that the top ones do, but for red and blue that number is under 10%. In strong white light, where the upper chloroplasts are already saturated, an added green photon is more likely to drive extra photosynthesis than an added red photon, because it reaches chloroplasts that still have capacity. That is the finding of Terashima and colleagues (2009) in sunflower leaves, and it is why "green light is useless to plants" is on the mistakes list below.

Human color vision also matters. Cone responses are compared against each other, so a leaf that returns 20% of green and 8% of red and blue is seen as vividly green even though it is, in absolute terms, a dark object that swallowed most of the light.

Light arrives in photons, and a photon's energy depends only on its wavelength:

$$E=\frac{hc}{\lambda }=\frac{1239.84\text{eV·nm}}{\lambda }$$

where h is Planck's constant and c the speed of light. Shorter wavelength, more energy per photon. A green photon at 550 nm carries 2.25 eV; a red photon at 680 nm carries 1.82 eV. If "not enough energy" explained the green gap, chlorophyll would have to reject red even harder. It does the opposite: the red Q_y band is its second-strongest absorption.

So the amount of energy in isolation is the wrong variable. What matters is whether the photon's energy matches an allowed transition of the absorbing molecule, and how strongly that transition couples to light. A molecule is not a bucket that fills with any photon above a threshold; it is a set of discrete electronic states with rules about which jumps are permitted and how probable each one is. That is Level 4.

One useful side fact: per mole of photons, 550 nm light delivers about 217 kJ. Photosynthesis spends two photons (one per photosystem) for every electron it moves from water to NADP⁺, so energy per photon is not the constraint. Matching is.

Chlorophyll a (C₅₅H₇₂MgN₄O₅, 893.5 g/mol) is a chlorin: a nearly planar ring of four pyrrole-type units, one of them partly reduced, joined by methine bridges around a central magnesium ion, with a long phytol tail that anchors it in protein and membrane. Around the ring, π electrons are delocalized over an 18-atom aromatic path. Delocalized electrons have closely spaced molecular orbitals, and the energy differences between the highest occupied and lowest unoccupied orbitals fall in the visible range. That is what makes chlorophyll a pigment at all.

A photon is absorbed only when three things line up: its energy matches an allowed gap between orbitals, the transition is permitted by symmetry (selection rules), and the transition dipole moment is large, meaning the electron redistribution that the jump produces couples strongly to the light's oscillating electric field. The strength of a band is measured by its oscillator strength, and the measured extinction coefficient in the bench above is the practical readout of it.

For chlorophyll a the strong bands are:

• Soret (B) band, peak 427.8 nm in ether: a transition to a higher excited singlet state with very large oscillator strength.
• Q_y band, peak 660 nm in ether: the transition to the lowest excited singlet state, polarized along the molecule's y axis, strong.
• Q_x band, near 578 and 615 nm as weak shoulders: polarized along x, much weaker in chlorophyll a because the reduced ring breaks the symmetry that would make it strong.

Between the Soret band and the Q bands, in the 480 to 600 nm window, there is no strong allowed electronic transition. Weak vibronic features and the tail of Q_x keep the window from being empty, which is why the molecule still absorbs at a few percent of peak there. Bound in a protein, each chlorophyll's bands shift by several nanometres and broaden, and the ensemble of shifted copies is what a leaf actually shows; in photosystem II the reaction-center chlorophylls absorb at 680 nm, in photosystem I at 700 nm.

The standard explanation is Gouterman's four-orbital model (1961). Take the two highest occupied π orbitals and the two lowest unoccupied ones. Four single-electron promotions are possible, and because the orbitals are nearly degenerate the promotions mix. In the mixture where the two transition dipoles add, you get an intense band: the Soret (B) band. In the mixture where they nearly cancel, you get weak bands: the Q bands. In a symmetric porphyrin the cancellation is nearly complete, so Q is faint and the molecule looks red-purple.

Chlorophyll is a chlorin. Reducing ring IV breaks the four-fold symmetry and lifts the degeneracy, so the cancellation in the y-polarized combination fails. Q_y gains oscillator strength (76% of the Soret peak in ether) and shifts to the red, while Q_x stays weak. That single structural change is why chlorophylls have two strong absorptions rather than one, and why the strong ones sit at opposite ends of the visible spectrum with a weak middle.

Kasha's rule: the blue photon's surplus is heat

A blue photon (2.9 eV at 428 nm) promotes the molecule to S₂, the Soret state. It does not stay there. Internal conversion and vibrational relaxation drop it to S₁, the Q_y state, on a sub-picosecond timescale, dumping about 1 eV into molecular vibrations and the surroundings. Kasha's rule (1950) states the general pattern: emission and photochemistry proceed from the lowest excited state of a given multiplicity, whatever state was first populated. So a blue photon and a red photon end up in the same S₁ state doing the same chemistry; the blue one simply arrived with more energy and wasted the surplus.

This has a practical consequence for the color question: a leaf gains nothing chemically from blue over red, per photon. The advantage of absorbing blue is having more photons to absorb, not having stronger ones. And it means the fate of every absorbed photon, whatever its color, is decided in S₁, where three exits compete: transfer to a neighbour (fastest, the normal path), fluorescence (a few nanoseconds in solution, mostly outcompeted in a leaf), and intersystem crossing to the triplet state, which can sensitize singlet oxygen unless a carotenoid quenches it.

In a chloroplast, chlorophyll molecules are bound in pigment-protein complexes embedded in thylakoid membranes. The workhorse antenna of plants, LHCII, binds 14 chlorophylls per protein monomer (8 chlorophyll a, 6 chlorophyll b) and 4 xanthophylls (two luteins, one violaxanthin, one neoxanthin), and assembles into trimers around the photosystem II core. Almost every chlorophyll in a leaf is an antenna pigment. Only a handful, the "special pairs" P680 in photosystem II and P700 in photosystem I, actually initiate chemistry.

When an antenna chlorophyll absorbs a photon, the excitation is not a private property of that molecule. Neighbouring pigments a nanometre apart interact through their transition dipoles, so the excited state is shared, a collective state called an exciton. Energy then migrates through the network by a mix of coherent delocalization within tightly coupled clusters and incoherent hopping (Förster transfer) between clusters, always biased downhill in energy toward the reaction center, which is tuned to sit slightly lower than the antenna. Once at P680, the excitation drives charge separation: an electron leaves the excited chlorophyll for a nearby pheophytin within a few picoseconds, and the transient optical excitation has become a longer-lived separated charge.

So photosynthetic energy conversion is two problems, not one:

1. Excitation-energy transfer: collect photons across a large area and many wavelengths and route the excitations to a small number of reaction centers before they are lost. Quantum yield of this step in healthy leaves: above 0.8, often quoted as the dark-adapted F_v/F_m ratio of about 0.83.
2. Charge separation: turn the excitation into separated charge and stabilise it faster than it can recombine. Quantum yield from P680 excitation: near unity.

Calling chlorophyll "the plant's solar panel" is therefore incomplete in a specific way: a solar panel converts where it absorbs. A leaf absorbs in hundreds of places per reaction center and converts in one, so its performance depends on the wiring between them at least as much as on the absorber.

Sunlight at a leaf is not steady. Clouds, wind-moved leaves, sunflecks in a canopy, and the leaf's own changing geometry make the input fluctuate over milliseconds to minutes, and the fluctuations are largest where the spectrum is most intense. Downstream, the reaction centers and the electron-transport chain have finite capacity. A surge above capacity is wasted or damaging; a dip below it starves the chain.

Arp, Kistner-Morris, Aji, Cogdell, van Grondelle and Gabor (Science, June 2020) treated the antenna as a network with two absorbing channels receiving a noisy spectral input, and asked which pair of absorption wavelengths minimises fluctuations in the power delivered to the reaction center while keeping power conversion efficient. The answer, for a smooth input like the solar spectrum at the ground, is two channels placed on the slopes of the input spectrum, on either side of its peak, rather than one channel at the peak. Fluctuations in the two channels then partly cancel, in the spirit of a balanced or noise-cancelling input.

Fed the appropriate input spectrum, the model reproduces the broad features of absorption in green plants (peaks either side of the green solar maximum), purple bacteria (near-infrared bands where their filtered light is strongest) and green sulfur bacteria. One principle, three very different organisms. The press summary was "why plants are not black": absorbing at the peak would maximise raw capture but maximise noise with it.

What the model does not do: it does not derive chlorophyll's spectrum from first principles, it does not include the canopy, and it has not been tested by building an antenna with a different spectrum and measuring fitness. It should be presented as a compelling organising principle that reproduces observed spectra and yields testable predictions, alongside historical constraint, electronic structure, photoprotection and canopy optics, not as the final word on why grass is green.

At the microscopic level a light-harvesting complex is an open quantum system: electronic excitations that are quantum objects, coupled to molecular vibrations, protein motion, surrounding water, thermal fluctuations and the electromagnetic field. If the pigments were isolated, an excitation would evolve as a coherent superposition spread over several molecules. Strong coupling to the environment scrambles the phases, producing decoherence and transport that looks like classical hopping.

But environmental coupling is not purely destructive. Dephasing can help an excitation escape a local energy trap, or bridge states that would otherwise be poorly connected. This is environment-assisted quantum transport (ENAQT), proposed independently in 2008 by Mohseni, Rebentrost, Lloyd and Aspuru-Guzik and by Plenio and Huelga: transport efficiency peaks at an intermediate noise level, neither fully coherent nor fully incoherent. A noisy environment, at the right noise level, is a feature.

The experimental story has a clear arc. In 2007, two-dimensional electronic spectroscopy of the FMO complex from a green sulfur bacterium, at 77 K, showed oscillating signals that were read as electronic coherence lasting at least 660 fs, far longer than expected, and the paper suggested wavelike sampling of transfer paths. In 2010 similar beats were reported at 277 K. Over the following decade the beats were re-examined: vibrational coherence, correlated vibrations and mixed vibronic states produce similar signatures, and a 2017 re-measurement of FMO concluded that electronic coherence decays within about 60 fs, too fast to steer transport. A 2020 multi-author review titled "Quantum biology revisited" reset the consensus: long-lived electronic coherence is not what makes light harvesting efficient; the productive questions are about how electronic states, vibrations, dissipation and protein structure work together.

So the frontier question is precise: do nontrivial quantum-coherent dynamics materially improve biological energy transport under physiological conditions? Current work on exciton-vibrational coupling and on which vibrational modes are protected from noise is the serious version of the question. "The exciton tries every path at once" is the pop version, and it is not supported.