Photocurrent peaks at a film thickness set by how far light penetrates and how far carriers travel: typically 100–300 nm for BiVO4 and α-Fe2O3, 1–10 µm for TiO2 nanoparticle films, with loadings of 0.1–3 mg/cm2. Past the optimum, doubling thickness usually costs 20–40 % of the current density.
Absorption decays as exp(−αx), so at a depth of 1/α only 37 % of the light survives. Carriers must then reach the back contact before they recombine, over a distance set by the diffusion length L = √(Dτ). The photocurrent is the overlap of the two: 63 % of the light is absorbed within one absorption depth, and nothing beyond about 3 L contributes.
| Material | α near the absorption edge | 1/α | Limiting carrier length | Typical optimum | Loading |
|---|---|---|---|---|---|
| TiO2, P25 film | ~2 × 105 cm−1 at 350 nm | 50 nm | 10–100 nm per particle, µm-scale percolation | 1–10 µm | 0.5–3 mg/cm2 |
| α-Fe2O3 | 1–2 × 105 cm−1 at 400 nm | 50–100 nm | 2–4 nm hole diffusion | 20–100 nm | 0.05–0.3 mg/cm2 |
| BiVO4 | ~1 × 105 cm−1 at 450 nm | 100 nm | 70–100 nm electron diffusion | 100–300 nm | 0.2–0.6 mg/cm2 |
| WO3 | ~1 × 105 cm−1 at 400 nm | 100 nm | 100 nm to 1 µm | 0.5–2 µm | 0.5–2 mg/cm2 |
The pattern explains why hematite is made thin and titania is made thick. It also explains why the same 1 mg/cm2 loading performs differently on two materials: mass is a proxy for thickness only when the density and porosity are known.
For spin coating, thickness scales roughly with ω−0.5, so halving the speed thickens the film by about 40 %. For doctor blading, the wet thickness is set by the blade gap and the dry thickness by the slurry concentration: 10 mg/mL at a 100 µm gap gives roughly 1 mg/cm2. Electrodeposition is governed by passed charge, and spray coating by passes and dilution.
Near the optimum, a 50 % thickness change typically moves photocurrent by 20–40 %, which makes thickness control the largest single reproducibility risk in a photoanode study. Fix the slurry batch, the humidity and the substrate temperature, and record the speed, gap and layer count in the methods section. Spin coating at 35 % relative humidity versus 65 % can shift thickness by 10–15 %.
Profilometry across a masked step reads 5 nm to 10 µm with ±5 nm repeatability and is the fastest route. Cross-section SEM gives an independent check and shows whether the film is dense or porous, which profilometry cannot. Gravimetry catches what both miss: weigh the substrate before and after deposition on a microbalance reading 1 µg, and divide by the coated area only, not the whole slide.
Optical absorbance is a useful cross-check when α is known, since A = αt for a non-scattering film. Nanoparticle films scatter, so the apparent thickness from absorbance runs high; use it for relative comparisons within one series.
A nanoparticle film with 20 nm particles has a roughness factor of 50–200, so 1 cm2 of geometric area carries 50–200 cm2 of active surface. Current densities are conventionally quoted per geometric cm2, and that convention is only interpretable when the thickness is stated alongside. Normalising by mass instead produces numbers that rise as films get thinner, because material beyond the collection depth is electrochemically dark.
Where a fair comparison between loadings is needed, report three numbers: geometric current density in mA/cm2, thickness in nm, and either the roughness factor from double-layer capacitance (20–60 µF/cm2 of real area) or the mass loading in mg/cm2.
Films above about 1 µm made in a single blade pass crack on drying, and cracks admit electrolyte to the substrate. Bare FTO delivers 5–20 µA/cm2 under 100 mW/cm2, small but not zero, and a cracked 0.1 mA/cm2 film can derive a measurable share of its signal from exposed substrate. Deposit in two or three passes with drying between, sinter at 450–500 °C with a 2 °C/min ramp, and run a bare-substrate control in the same cell.
There is no universal value. Use 100–300 nm for BiVO4 and 20–100 nm for α-Fe2O3, but 1–10 µm for TiO2 nanoparticle films. Run a four-point thickness series on your own material; the optimum shifts with doping, morphology and light intensity.
Weigh the substrate before and after deposition, divide the difference by the coated area rather than the whole substrate, and dry before weighing. A 1 cm2 film at 1 mg/cm2 needs a balance readable to 1 µg for 1 % precision.
Because carriers generated beyond the diffusion length recombine before reaching the contact, and the extra series resistance of a thicker film slows collection. Cracking and delamination add a third loss channel above about 1 µm in a single pass.
No. Sites deeper than roughly three absorption depths receive no light, so additional material adds mass and resistance without adding photocurrent. Loading should be quoted with thickness, otherwise the number cannot be compared between laboratories.
