Oxygen is produced during photosynthesis when water molecules are split apart inside a protein complex called photosystem II, which sits in the membrane of chloroplasts. Light energy powers the removal of electrons from water, and the leftover oxygen atoms combine to form the O₂ that you breathe. The process is more violent and precise than the phrase “water splitting” suggests, involving a cluster of metal atoms that ratchets through four consecutive light-driven steps before releasing a single molecule of oxygen. Understanding how this works reveals one of the most remarkable chemical feats in biology.
Light Kicks Off the Process
Photosynthesis uses two large protein complexes embedded in the thylakoid membrane, called photosystem I and photosystem II. Oxygen production happens exclusively in photosystem II. When a photon of light is absorbed by the pigments surrounding photosystem II, the energy funnels inward toward a special pair of chlorophyll molecules known as P680. That name refers to the wavelength of light it absorbs most strongly, 680 nanometers, which falls in the red part of the visible spectrum.
Once P680 receives the energy, it becomes excited and almost immediately hands off an electron to a nearby molecule called pheophytin. This transfer is astonishingly fast. At room-relevant temperatures it happens in well under a picosecond, which is less than a trillionth of a second.1PubMed Central. Charge separation in the reaction center of photosystem II studied as a function of temperature Energy also bounces between P680 and nearby accessory chlorophyll molecules on timescales measured in femtoseconds, a thousand times faster still.2PubMed. Excitation energy transfer and charge separation in the isolated Photosystem II reaction center Recent work using advanced spectroscopy found that charge separation proceeds through a single pathway, with a pheophytin molecule acting as the primary electron acceptor while two chlorophyll molecules work together as the electron donor.3PubMed Central. Charge separation in the photosystem II reaction center resolved by multispectral two-dimensional electronic spectroscopy
The result of this ultrafast chemistry is that P680 ends up missing an electron. In its oxidized form, often written as P680⁺, it becomes one of the strongest biological oxidants known. It is so eager to reclaim an electron that it can rip one away from water, a molecule most chemical reactions cannot break apart under mild conditions. That electron-hungry state of P680 is the driving force behind all oxygen production on Earth.
The Metal Cluster That Splits Water
P680⁺ does not tear apart water directly. The actual water-splitting chemistry takes place at a small inorganic cluster embedded in photosystem II, called the oxygen-evolving complex. This cluster contains four manganese atoms, one calcium atom, and five bridging oxygen atoms, written as Mn₄CaO₅.4PubMed Central. Structural changes of the oxygen-evolving complex in photosystem II during the catalytic cycle It is a tiny piece of mineral-like chemistry sitting inside a living protein, and nothing else in biology quite resembles it.
The choice of manganese is not accidental. Manganese can cycle through multiple oxidation states, meaning it can hold onto or release electrons flexibly. The calcium atom plays a structural and functional role, helping to position water molecules where they need to be. Quantum-chemical calculations have shown that the electronic structure of this cluster is unusually complex, with multiple low-energy states that allow it to shuffle electrons during catalysis.5PubMed. Entangled quantum electronic wavefunctions of the Mn₄CaO₅ cluster in photosystem II The cluster also binds manganese in a ratio of roughly four atoms per complex, a stoichiometry confirmed by biochemical purification studies.6PubMed Central. Manganese-binding proteins of the oxygen-evolving complex
Think of the oxygen-evolving complex as a kind of molecular battery. Each time P680⁺ pulls an electron away from the cluster, the cluster becomes slightly more oxidized. After four such electron removals, the cluster has accumulated enough oxidizing power to rip apart two water molecules at once, producing one molecule of O₂, four protons, and four electrons. Those four electrons refill the cluster, resetting it to its starting state, ready for the next round.
Four Flashes to Make One Oxygen
In the late 1960s, a researcher named Bessel Kok discovered something striking: if you give dark-adapted chloroplasts brief flashes of light, oxygen is not released after every flash. Instead, it appears in bursts after every fourth flash. This pattern revealed that the oxygen-evolving complex cycles through five intermediate states, labeled S₀ through S₄. Each flash of light advances the complex by one step. When it reaches S₄, water is split, oxygen is released, and the complex resets to S₀. This cycle is known as the Kok cycle, or the S-state cycle.
Researchers have now captured structural snapshots of several of these intermediate states using X-ray crystallography performed at room temperature with incredibly short X-ray pulses from a free-electron laser. These snapshots show that the manganese cluster physically rearranges at each step. One particularly revealing finding came from the S₂ to S₃ transition, where a new oxygen atom, labeled Ox, was seen binding between the calcium atom and one of the manganese atoms. A water molecule bound to calcium, called W3, appears to deliver this new substrate oxygen to the active site.7PubMed Central. Structures of the intermediates of Kok’s photosynthetic water oxidation clock
The final step, S₃ to S₄ and then back to S₀, is where the oxygen-oxygen bond actually forms. This is the moment two oxygen atoms join together and O₂ is born. Capturing this fleeting transition has been one of the grand challenges in photosynthesis research. In 2023, a team reported room-temperature crystallographic snapshots of this final step, revealing coordinated structural changes across both the active site and the surrounding protein that guide the reaction to completion.8PubMed Central. Structural evidence for intermediates during O(2) formation in photosystem II The data also helped narrow down which oxygen atoms in the cluster form the O₂ molecule, an issue that has been debated for decades.
Getting Water In and Protons Out
For the water-splitting reaction to keep running, fresh water molecules need to reach the manganese cluster, and the protons released during water oxidation need to escape into the surrounding fluid. The oxygen-evolving complex is buried deep inside the photosystem II protein, so the protein has evolved a system of internal channels to handle this traffic.
Computational simulations of fully hydrated photosystem II have identified three major water channel systems on the side of the membrane facing the interior of the thylakoid, each named for the oxygen or chloride atom nearest its terminus at the cluster. These are called the O1, O4, and Cl1 channels (older literature sometimes refers to them as the “narrow,” “large,” and “broad” channels). All three remain filled with water throughout the simulation, but they differ in length, width, and branching.9PubMed Central. Functional Water Networks in Fully Hydrated Photosystem II
Comparative studies across different photosynthetic organisms suggest that these channels have specialized roles. The O1 channel appears to serve primarily as the inlet for fresh substrate water. The tighter Cl1 and O4 channels seem to function mainly as proton-exit pathways. The O4 channel, in particular, features a highly ordered chain of hydrogen-bonded water molecules connecting directly to the cluster, which strongly suggests it acts as a proton wire, especially during certain steps of the Kok cycle.10PubMed Central. Evolutionary diversity of proton and water channels on the oxidizing side of photosystem II and their relevance to function The protons released through these channels contribute to a concentration gradient across the thylakoid membrane, which the cell then uses to drive ATP production through a separate enzyme, ATP synthase.11PubMed Central. Proton gradient across the chloroplast thylakoid membrane governs the redox regulatory function of ATP synthase
Where the Electrons Go After Water Is Split
The four electrons stripped from water do not just vanish. They are passed along a chain of carriers within the thylakoid membrane, eventually reaching photosystem I, where a second round of light absorption boosts their energy high enough to reduce molecules the cell uses for carbon fixation. This electron transport chain is classically depicted as a Z-shaped diagram, reflecting the two successive energy boosts from the two photosystems.
An alternative view, based on experiments tracking how the electron carrier ferredoxin gets reduced, has proposed that photosystem II can reduce ferredoxin on its own in some circumstances, without requiring photosystem I.12PubMed. Divergent pathways of photosynthetic electron transfer: The autonomous oxygenic and anoxygenic photosystems Whether this represents a major physiological pathway or a minor side reaction is still debated. For the purpose of understanding oxygen production, the key point is simpler: the electrons removed from water during oxygen evolution are the ultimate source of the electrons that power the rest of photosynthesis. When you eat a plant, the chemical energy stored in its sugars traces back to the moment those electrons were ripped from water.
The Danger of Making Oxygen
Producing oxygen is essential for life on Earth, but it comes with a cost for the organism doing it. Oxygen is reactive. And the very machinery that produces it, photosystem II, is constantly exposed to intense light energy. When the system absorbs more light than it can use, the excess energy can interact with oxygen to create singlet oxygen, a highly reactive form that damages proteins, lipids, and pigments. In photosystem II, singlet oxygen is produced when molecular oxygen reacts with the excited triplet state of chlorophyll in the reaction center.13PubMed. Singlet oxygen production in photosystem II and related protection mechanism
Plants have evolved multiple lines of defense. Beta-carotene, the same pigment that makes carrots orange, sits right next to the chlorophyll in photosystem II and can quench singlet oxygen before it causes damage. Tocopherol, a form of vitamin E, does the same. A molecule called plastoquinol also acts as a singlet oxygen scavenger within the membrane, though it gets consumed in the process.14PubMed. Plastoquinol as a singlet oxygen scavenger in photosystem II If these defenses fail, the cell can even flip a genetic switch, activating defense genes that ramp up production of protective molecules.
Despite all this, photosystem II still gets damaged regularly. The D1 protein at the heart of the reaction center is the most frequently replaced protein in the entire thylakoid membrane. When light damages it, a dedicated repair cycle kicks in: the damaged D1 is chopped up by specialized enzymes called FtsH and Deg proteases, removed, and replaced with a freshly synthesized copy.15PubMed. D1 fragmentation in photosystem II repair caused by photo-damage of a two-step model Blue light appears to cause more of this kind of damage than red light, suggesting that the primary target of the most harmful photodamage is the manganese cluster itself, not the chlorophyll. This makes sense: the cluster is handling some of the most extreme chemistry in biology, and even small disruptions to its structure can knock it offline.
Conserved Across Billions of Years
Virtually every organism that produces oxygen, from cyanobacteria in a pond to the oak tree in your yard, uses the same basic photosystem II architecture with the same Mn₄CaO₅ cluster. The core structure and the fundamental mechanism of light-driven water oxidation are remarkably well conserved across all oxygen-producing phototrophs.16PubMed Central. Photosystem II: commonality and diversity with emphasis on the extrinsic subunits What has changed over evolutionary time is the set of accessory proteins that sit on the outer surface of the complex and help stabilize water oxidation. Different lineages, such as cyanobacteria, green algae, red algae, and land plants, have swapped, added, or lost these extrinsic subunits while keeping the catalytic core intact.
The evolutionary origin of this machinery stretches deep into Earth’s history. Geological evidence from stromatolites and chemical biomarkers suggests that cyanobacteria capable of oxygenic photosynthesis had evolved by at least 2.7 billion years ago, and possibly much earlier. Some uranium-lead data from rocks roughly 3.8 billion years old are consistent with this metabolism having arisen near the very start of the geological record.17PubMed Central. When did oxygenic photosynthesis evolve? Ancestral-state reconstructions suggest that the earliest oxygen-producing organisms lived in terrestrial environments like soils and freshwater, not in the open ocean.18PubMed Central. Distribution of early-branching Cyanobacteriia and the potential habitats that gave rise to the earliest oxygenic phototrophs
For a long stretch after oxygenic photosynthesis appeared, atmospheric oxygen remained low. The Great Oxygenation Event, when free oxygen first accumulated to significant levels in the atmosphere roughly 2.4 billion years ago, may have been triggered by an ecological shift: cyanobacteria gaining dominance over older forms of bacteria that performed photosynthesis without producing oxygen.19PubMed Central. The Great Oxygenation Event as a consequence of ecological dynamics modulated by planetary change In other words, the chemistry of water splitting existed long before the atmosphere reflected it. The planet’s oxygen story is not just about the evolution of the reaction, but about when the organisms performing it gained enough of a foothold to reshape the global environment.
How Scientists Watch Water Being Split in Real Time
One reason our understanding of oxygen production has advanced so quickly in the past decade is a revolution in experimental technique. The traditional approach to determining protein structures, X-ray crystallography, required freezing samples and exposing them to X-rays for extended periods. For photosystem II, this was a problem: the intense X-ray beam could damage the manganese cluster, producing images of a broken machine rather than a working one.
The solution came from X-ray free-electron lasers. These machines produce X-ray pulses so short, on the scale of femtoseconds, that they capture a diffraction snapshot of the protein before the radiation has time to cause damage. Researchers developed a method of streaming tiny photosystem II crystals past the laser beam in a jet of liquid, collecting thousands of diffraction images per hour. By illuminating the crystals with visible-light laser pulses at precise intervals before the X-ray snapshot, they can advance the Kok cycle one step at a time and capture the structure at each intermediate state.20PubMed Central. Serial time-resolved crystallography of photosystem II using a femtosecond X-ray laser
This approach has yielded structures at multiple time points during the catalytic cycle, including snapshots taken just microseconds after a light flash, catching the enzyme in the act of rearranging its atoms to form O₂.21Structural Dynamics. Structural changes in Photosystem II using time-resolved X-ray free-electron (XFEL) crystallography and crystalline molecular dynamics The level of detail is extraordinary: researchers can now track individual water molecules moving into position, see bonds forming and breaking at the cluster, and watch the protein backbone flex to accommodate each catalytic step. This kind of molecular movie was unthinkable twenty years ago.
Isotope Experiments and What They Reveal
Another way scientists probe the water-splitting reaction is by using isotopically labeled water. If you dissolve heavy-oxygen water, H₂¹⁸O, into a photosystem II preparation and then measure the oxygen gas that comes out with a mass spectrometer, you can learn something about how the enzyme handles its substrates. At 50% enrichment of H₂¹⁸O, you would expect equal amounts of ¹⁸O₂ and ¹⁶O₂ if the enzyme treats both isotopes the same. But experiments found a clear deviation: the measured isotope effect ranged from about 1.14 to 1.30, higher than theoretical predictions. The effect also varied strongly with temperature and showed a discontinuity around 11°C.22PubMed. 18O isotope effect in the photosynthetic water splitting process
The researchers interpreted this as evidence that water does not arrive at the catalytic site as single isolated molecules. Instead, small clusters of water molecules may enter the site together, and the slightly different physical properties of heavy water (like its slower diffusion) could affect how readily it reaches the reaction center. This finding matters because it adds another layer to our understanding of how the protein controls access to its active site. The channels described earlier are not just passive tunnels; the protein environment shapes which water molecules arrive and when.
Copying Nature for Clean Energy
The water-splitting reaction in photosystem II has inspired a whole field of research into artificial photosynthesis. If we could build a synthetic catalyst that splits water using sunlight the way plants do, we could produce hydrogen fuel (from the protons) and oxygen without burning fossil fuels. The appeal is obvious: water is abundant, sunlight is free, and the only byproduct is oxygen.
The catch is that nature’s catalyst uses manganese, calcium, and a handful of common elements, all cheap and earth-abundant, to perform a reaction that industrial chemistry typically accomplishes only with expensive rare metals like iridium or ruthenium. Understanding exactly how the Mn₄CaO₅ cluster achieves this feat has become a guiding principle for catalyst design. Researchers developing biomimetic systems explicitly use the structural and mechanistic features of the natural cluster as design criteria, aiming to build synthetic catalysts from first-row transition metals that could be manufactured at global scale.23PubMed Central. Artificial photosynthesis: understanding water splitting in nature
Progress has been real but slow. Synthetic manganese-oxide catalysts can oxidize water, but none yet match the natural enzyme’s combination of speed, efficiency, and durability. One persistent challenge is that the protein environment in photosystem II does far more than just hold the metal cluster in place. It fine-tunes the energetics of each step, channels water to the right spot, removes protons at the right moment, and protects the cluster from the very reactive intermediates it produces. Reproducing all of those functions in a synthetic scaffold remains one of the toughest problems in chemistry. But every new structural snapshot from X-ray laser experiments brings researchers a step closer to understanding which features of the natural system are truly essential and which can be simplified in a device meant for a rooftop rather than a leaf.