Allow 10–15 minutes after ignition before you start timing a reaction. During that window the arc settles, the lamp housing and reflector reach thermal equilibrium, and optical output climbs to its steady value. Starting sooner typically means running the first 10–20 minutes at 10–20 % below steady-state irradiance, with output still drifting upward.
A xenon short-arc lamp is not a resistor that switches on at a fixed output. Three things are still moving in the first minutes:
| What you observe | What it does to your data |
|---|---|
| Irradiance is low and still rising for the first 10–20 min | The gas-production curve shows an apparent "slow start" that is read as an induction period |
| Light spot drifts slightly | Effective illuminated area changes, so rates from different runs are not comparable |
| Spectral balance shifts marginally | Apparent quantum yield (AQY) calculations inherit the error, because photon counts are wrong |
A common misreading: a group sees the first 30 minutes running slow and then linear, and writes "an induction period was observed" into the paper. Repeating the run with a blank (no catalyst) and logging irradiance for 30 minutes will show whether the lamp or the chemistry is responsible. In most cases it is the lamp.
Yes, mildly. A 300 W lamp in a well-ventilated housing is usually stable within 10–15 minutes. Higher-power lamps, and lamps in housings with limited airflow, need longer because more thermal mass has to come to equilibrium. The practical rule is not a fixed number but a criterion: warm up until irradiance changes by less than about 2 % over a 5-minute window.
No. Running above the rated operating point shortens bulb life and does not make the arc or the reflector stabilise faster. Warm-up time is set by thermal mass, not by drive power.
Yes. A re-strike restarts the arc from a different position, and the housing will have cooled. Treat any re-ignition as a fresh start and warm up again.
Much less. LEDs reach stable output within seconds to a couple of minutes, and their spectrum does not shift with arc geometry. The thermal-equilibrium argument still applies to the fixture, but the wait is far shorter.
You cannot, reliably. A silicon photodiode or thermal power meter at the reactor plane costs far less than a wasted week of experiments and gives you the normalisation data you need anyway.
