Between the reactor and the detector sits the least glamorous and most error-prone part of any gas analysis train: the sampling path. Dead volume, incomplete purging and cross-contamination between runs can bias results by tens of percent, and they do so silently.
Any unswept pocket, a tee, a gauge stub, a wide-bore section, delays and dilutes the signal. In kinetic work, dead volume smears the time axis; in cumulative work, it shifts apparent production. Design the path as the shortest possible swept line, and calculate the system time constant, volume divided by flow, and compare it with the kinetics you intend to resolve.
Before the run, purge long enough to replace the headspace and the train several times over. After any interruption, re-purge. Residual air also brings oxygen, which can react with products or poisons and alter the chemistry being measured.
Product adsorbed in tubing or traps from a high-yield run can desorb into the next low-yield run. Sweep between experiments until the baseline returns, and for high-adsorption products bake out or replace suspect sections. Run a blank between different catalysts to prove the train is clean.
Purge a defined multiple of system volume; sample from a swept position, not a stagnant port; keep the same geometry every run; and document the purge protocol in the methods section. Reviewers do ask.
Perfectlight's online analysis systems and sealed photocatalytic reaction systems are engineered as one continuous swept path.
