Photodeposit 0.5–1.0 wt% Pt on most oxide photocatalysts. Dissolve H2PtCl6·6H2O, which is 37.7 wt% Pt, add 10–20 vol% methanol as hole scavenger, purge 30 min with argon, then irradiate at 100–300 mW/cm2 for 1–2 h. One wt% Pt usually raises the H2 rate 5–20 times; above 2 wt% the gain reverses.
Loading is platinum mass divided by catalyst mass: 1 wt% on 1.00 g of TiO2 means 10.0 mg of Pt. Because the hexachloroplatinic acid salt carries only 37.7 wt% Pt, reaching 10.0 mg of metal takes 26.5 mg of precursor. The most common error in the literature is weighing 10 mg of the salt, dissolving it and reporting “1 wt% Pt”, when the real loading is 0.38 wt%.
For a loading series, prepare one stock rather than weighing six times. 265 mg of precursor in 100 mL of water gives 1.0 mg Pt per mL, so a 0.2 / 0.5 / 1.0 / 2.0 wt% series on 1.00 g of catalyst is 2, 5, 10 and 20 mL of stock, each made up to the same final volume. Add the stock dropwise over 5 min with stirring; pouring it in creates a local concentration spike and larger particles.
| Method | Typical Pt size | Where Pt sits | Working range | Main risk |
|---|---|---|---|---|
| Photodeposition, 10–20 vol% methanol, 1–2 h | 2–5 nm | At electron accumulation sites: defects, edges, specific facets | 0.1–2 wt% | Incomplete reduction; dark adsorption before illumination |
| Incipient wetness + H2 at 300 °C | 3–10 nm | Spread over the whole accessible surface | 0.5–5 wt% | Sintering above 400 °C; chloride retention |
| NaBH4 reduction in slurry | 2–8 nm | Random, including nuclei formed in the liquid | 0.5–5 wt% | Homogeneous nucleation wastes metal off-support |
| Commercial 1 wt% Pt/TiO2 | 2–4 nm | Vendor-defined | Fixed | Unknown pretreatment and chloride history |
The reason to choose photodeposition is placement, not size. Photoexcited electrons arrive at specific surface sites, so Pt grows where reduction actually happens, which shortens the charge-transfer path from the support to the metal. Above roughly 3 wt% the process becomes slow and uneven, and impregnation is the better tool.
Dark adsorption. PtCl62- is an anion and oxide surfaces are positive below pH 6.3, so 20–40 % of the platinum can leave solution within 30 min in the dark. Material adsorbed in the dark is not photodeposited even if it is reduced later, so run a dark control and report the adsorbed fraction separately.
Incomplete reduction. Check the supernatant by ICP-OES. More than 10 % of the nominal Pt still in solution means the run was too short, the donor was consumed, or light did not reach the whole slurry.
Chloride left behind. At 1 wt% on 1 g the precursor brings about 11 mg of chloride. Left in 100 mL that is over 100 ppm, enough to shift selectivity and attack stainless fittings. Three 100 mL washes bring it below 10 ppm.
Pt on the wrong surface. A grey mirror on the reactor wall or the stir bar is lost catalyst. Check the glass after each run and prefer quartz or a coated immersion well.
Aggregation above 2 wt%. Particles grow, platinum area per gram saturates, and at 3–5 wt% the metal begins absorbing incident light. Rates typically peak between 0.5 and 1.5 wt% and drop 20–50 % at 3 wt%.
Three measurements make a loading claim defensible: ICP-OES of the digested composite for actual wt%, ICP of the supernatant for reduction yield, and CO pulse chemisorption or STEM for dispersion. A nominal 1 wt% that measures 0.7 wt% by ICP is normal, and it should be reported as measured rather than nominal.
0.5 to 1.5 wt% covers most oxide photocatalysts, with the optimum usually near 1 wt%. Below 0.2 wt% there are too few active sites; above 2 wt% particles grow and screen light, and rates fall 20 to 50 % at 3 wt%.
For placement, yes: it puts Pt where electrons arrive, typically 2 to 5 nm particles at defects. For loadings above 2 wt%, or for supports that cannot generate electrons, impregnation followed by H2 at 300 degrees Celsius is faster and more even.
Check four things: actual Pt by ICP, chloride residue, metal deposited on the reactor wall, and whether the activity test used a donor at all. A nominal 1 wt% that measures 0.3 wt% explains a flat result without any mechanistic surprise.
Yes for oxides. Methanol at 10 to 20 vol% is standard, and ethanol, TEOA and Na2S/Na2SO3 also work. Without a donor, photogenerated holes oxidise the lattice or solvent slowly and reduction stays incomplete after two hours.