Specify a photothermal source in W/cm², not mW/cm²: plan on 1–5 W/cm² at the catalyst bed and keep at least 30% of the delivered radiant power beyond 780 nm. A 300 W xenon lamp focused to a 10 mm spot reaches 1.5–3 W/cm². Ordinary photocatalysis runs at 100 mW/cm², so optics, filters and cooling all change.
In ordinary photocatalysis light is a photon supplier and heat is an artefact: 100 mW/cm² (one sun) is enough to generate carriers, and warming is suppressed. In photothermal catalysis the same beam must also hold the bed at 150–350 °C, where the activation energy for CO2 hydrogenation sits. Absorbed power, not photon count, sets that temperature.
| Parameter | Ordinary photocatalysis | Photothermal catalysis | Why it changes |
|---|---|---|---|
| Irradiance at sample | 100 mW/cm² (1 sun) | 1–5 W/cm² | 0.6–3 W/cm² absorbed holds 150–350 °C |
| Spectral window | 300–420 nm, filtered | 300–1100 nm, NIR retained | 780–1100 nm heats without generating carriers |
| Infrared share of delivered power | Under 10% | 30–60% | IR is the heat source, not a contaminant |
| Bed temperature | 25–60 °C | 150–350 °C | Reaction window, not a side effect |
| Temperature readout | Bath thermocouple | Bed-surface thermocouple plus IR camera | Gas reads 30–80 °C low at 1–3 W/cm² |
Most photocatalysis hardware is built to remove heat. A water-jacketed IR filter, a 420 nm long-pass filter or a cold mirror all strip 780–1100 nm so the sample stays isothermal. Put them in a photothermal beam and 40–60% of the radiant power disappears; the bed never reaches temperature.
Xenon arcs suit this work because roughly half their radiant output already sits beyond 780 nm. A 365 nm LED array is the opposite case: no infrared, so heating comes only from non-radiative relaxation inside the catalyst, and achievable temperature is limited by the powder’s absorption coefficient rather than by lamp power. If you must cap the upper wavelength, use a 1100 nm long-pass filter rather than a heat filter — and air-cool it, because a filter absorbing several watts over a 25 mm aperture runs hot and shifts its cut-off.
Raising irradiance from 1 to 2 W/cm² on a 60% absorbing bed does not double the temperature rise. Radiative loss scales with the fourth power of absolute temperature, so the same doubling typically moves equilibrium temperature by 60–90 °C, not 150 °C. Above about 3 W/cm² the curve flattens and extra lamp power mostly heats the housing.
Measurement is the weak point in most setups. A thermocouple in the gas stream reads 30–80 °C below the bed surface at 1–3 W/cm² and 20 mL/min, which is enough to make a photothermal claim untestable. Put the junction in contact with the bed, shield it from direct radiation with thin foil, and cross-check once against an IR camera whose emissivity is set from that reading.
For a 10 mm fixed bed, a 300 W short-arc xenon with an ellipsoidal reflector and a 10 mm spot is the practical baseline: 1.5–3 W/cm² at the window plane, adjustable by lamp current, since moving the lamp changes intensity and uniformity together. Keep uniformity within ±10% across the bed; at 5 W/cm² a 10% gradient becomes a 20–30 °C gradient and the kinetics you extract are an average of two experiments.
Higher irradiance means a smaller spot, not a bigger lamp. Focusing a 300 W source to 5 mm roughly quadruples power density but halves the usable area, so the bed must shrink with it or the edges run cold.
Usually not as delivered. A 300 W xenon with a water IR filter gives about 100 mW/cm² at 15 cm, which will not pass 120 °C on a normal bed. Remove the filter and refocus to a 10 mm spot and the same lamp reaches 1.5–3 W/cm².
Not for rate measurements. An 808 or 980 nm laser gives 10–100 W/cm² on a sub-millimetre spot, which suits single-particle mechanism work but not kinetics: the Gaussian profile runs the bed centre and edge at different temperatures. A focused xenon source gives 1–5 W/cm² over 10–20 mm with ±10% uniformity.
Run the same bed in the dark at the same measured bed temperature. If the light-on rate exceeds the dark rate at matched temperature, the difference is photochemical; if they coincide, the light is only a heater.
Emissivity. A dark catalyst bed sits at 0.85–0.95, polished stainless at 0.2–0.3, so an uncorrected camera reads the metal far too cold. Set emissivity from one thermocouple cross-check, then use the camera for mapping, not for absolute values.
