Grätzel's TiO2 sponge let a one-molecule dye layer make a 7% solar cell

In 1991 the Nature paper opened with a stark number: electricity from commercial solar cells cost about ten times more than conventional generation. Dye sensitization was an old idea — chlorophyll on a zinc oxide electrode had shown photon-to-electricity conversion in the early 1970s — but a flat electrode holds only one monolayer of dye, which absorbs about 1% of incident light. The whole approach looked hopeless.

Brian O'Regan and Michael Grätzel, who had begun the work together at UC Berkeley in 1988, changed the geometry instead of the chemistry. They made a 10-µm optically transparent film of titanium dioxide particles a few nanometres across, sintered into a sponge with ~1,000× the surface area, and coated its internal surface with a single monolayer of charge-transfer dye. That one molecular layer now harvested 46% of the incident solar energy flux, injected electrons into the TiO2 conduction band with over 80% photon-to-current efficiency, and was regenerated by an iodide electrolyte.

The cell converted 7.1–7.9% of simulated sunlight — commercially realistic for 1991 — and 12% of diffuse daylight, sustained currents above 12 mA/cm², and survived at least five million turnovers without decomposition. Because electrons travel through the bulk conduction band rather than across junction surfaces, the semiconductor could be low-purity and cheap: the paper's own framing was a cell 'created from low-to medium-purity materials through low-cost processes.'

Separate light absorption from charge transport

Silicon PV priced itself out of bulk generation because ultra-pure crystals and junction-grade processing dominated cost; the dye cell was designed to tolerate cheap, imperfect materials.

The bottleneck was absorption, not chemistry: a monolayer catches only so much light, so a thousand-fold surface increase fixed two decades of failure without a new dye.

Charge separation happens the instant the dye injects an electron into TiO2's conduction band, ahead of recombination — the physics that lets a molecular device compete with a p–n junction.

The iodide electrolyte continuously restores the bleached dye, so stability came from regeneration chemistry: five million turnovers per molecule instead of photodegradation.

When a component is expensive or fragile, decouple its job from the structure's: a one-molecule layer absorbs, cheap bulk conducts. Nanoscale roughness turns that decoupling into a device.

The 1991 paper became one of the most cited in photovoltaics (27,000 citations on Nature's own page) and spawned a worldwide Grätzel-cell field. It did not displace silicon in power plants — electrolyte stability and the 25-year lifetime building integration demands stayed hard, so commercial DSSCs settled into niches like portable and indoor-light chargers. Its deepest legacy is architectural: perovskite solar cells, the fastest-improving photovoltaic technology of the 2010s, began as perovskite 'dyes' painted onto the same mesoporous TiO2 sponge.