A photoanode is an n-type semiconductor in contact with an electrolyte. Majority electrons diffuse into solution until equilibrium bends the conduction and valence bands upward near the surface, carving out an electron-depleted space-charge layer of width W (the Mott-Schottky relation, same form as a one-sided Schottky junction):
W = sqrt( 2·εr·ε₀·(Vbi + Vapp) / (q·Nd) )
Absorbed photons with energy above the bandgap create electron-hole pairs. Only pairs generated inside (or within a diffusion length of) the field-swept depletion layer are efficiently separated before recombining — the Gärtner picture used here as a simplified collected-fraction:
fcollected ≈ 1 − exp(−α·W)
Jphoto ≈ Jmax · (suns) · fcollected
Holes swept to the surface oxidize water (2H₂O + 4h⁺ → O₂ + 4H⁺); electrons flow through the external wire to a cathode where they reduce protons (4H⁺ + 4e⁻ → 2H₂), so H₂ bubbles form twice as fast as O₂ by Faraday's law, J/(2F) mol·s⁻¹·cm⁻² of H₂ for an assumed 4 cm² active electrode.
- Applied bias — a positive anodic bias widens W (deeper field, more collected carriers) until the photocurrent saturates; too negative a bias collapses band bending and J drops to zero.
- Illumination — scales the photon flux (and hence J) linearly, as long as the reaction isn't already collection-limited.
- Donor density Nd — heavier doping compresses W at fixed bias (more built-in charge per unit depth), trading depletion width for lower series resistance.
- Material — sets bandgap, absorption coefficient α and a representative saturation photocurrent Jmax: TiO₂ absorbs only UV (small α, low Jmax), Fe₂O₃ absorbs strongly but has a small hole-diffusion length in practice, BiVO₄ sits between the two — all illustrative, order-of-magnitude values for teaching the trend, not a specific published I–V curve.