A photocatalytic activity evaluation system earns its value through reproducibility: two runs of the same experiment should produce the same number. That property is designed in at construction time, not improvised later. This article lists what to define at the design stage, with emphasis on the gas detection link.
Fix the irradiation geometry, lamp, filters, distance and reaction plane; the reactor volume and headspace ratio; the carrier gas identity, flow and path; the stirring rate; the temperature control; and the sampling or detection method. Every one of these silently varies between labs that claim to run "the same" experiment.
Detection limit matched to expected production, commonly sub-ppm for trace products; a linear range covering induction to peak; response time fast enough to resolve the kinetics you care about; selectivity against the carrier and other products; and baseline stability across multi-hour runs. Ask vendors for detection-limit definitions, is it 3-sigma on a clean matrix or on your real matrix?
Log lamp hours, measured irradiance and calibration dates alongside the gas data. An activity number without its measurement context is not reproducible, and reviewers increasingly ask for it.
Blank runs with no catalyst should show flat baselines; duplicate runs should agree within your tolerance; standard injections should recover within a stated margin. Passing these three tests before publishing any result is the difference between data and anecdote.
Perfectlight provides complete photocatalysis evaluation solutions with matched online gas detection and calibrated light sources.
