Fix a datum, check that the lamp is coaxial with the housing bore, center the beam on the reactor window, set the working distance against an irradiance target, then repeat the check with the lamp hot. On a 300 W class source the hot check is not optional: the arc shifts as the housing warms.
Reference the beam to a physical datum, never to the edge of the bench. The useful pair is the reactor window plane, marked with a card taped flat against the window frame, and the optical axis height marked on the lamp housing. Once both exist, every adjustment becomes a measured deviation from a known state, and you can return to that state after cleaning, a bulb change or a move across the corridor.
Two datums are enough. Anything between them — lenses, filters, fiber couplers — is positioned relative to the housing face and the window plane, not to the table. Mark the distances on the rail with a paint pen; tape slides and memory is worse.
Each step has an observable outcome, which is what makes the procedure repeatable by someone who was not there when the bench was first built.
| Step | Action | What you check | Failure it prevents |
|---|---|---|---|
| 1. Datum | Mark the window plane and the axis height; record distances on the rail | Card sits flat and square against the frame | Every later step measured from a reference that moves |
| 2. Coaxial check | Hold a target card at two distances from the exit port | Spot center stays on the same axis at both distances | Lamp seated off-axis, asymmetric spot, hot edge of the beam |
| 3. Beam centering | Move the spot onto the window or the aperture center | Aperture clips equally on all four sides | Clipped beam and a one-sided gradient across the window |
| 4. Working distance | Set the gap to the value that gives your target irradiance on a meter | Meter reading and spot diameter both match the target | Over- or under-flux, and excess infrared load on the window seal |
| 5. Hot re-check | Repeat steps 2 to 4 at operating temperature | Same center, same reading within the meter's repeatability | Silent drift across a long run |
Step 4 gets skipped when a bench is rebuilt in a hurry. The distance is whatever value reproduces the irradiance you intend to report; it belongs in the method, not in someone's head. Two people setting different gaps will differ by more than their catalysts do.
Misalignment costs irradiance in three separate ways, and only one of them looks like an obviously smaller spot. The first is distance: for a compact arc the irradiance falls with the square of the working distance, so at the short gaps used in laboratory reactors a 10% change in distance changes the flux by roughly 20%. Nothing looks wrong; the meter simply reads lower.
The second is lateral offset. The beam is brightest on axis, so sliding the reactor sideways moves the window into a dimmer region even though the spot still covers it. The sample sees a gradient, and the average flux depends on where the stirrer sits. The third is clipping at the aperture or the window frame: a beam that is cut on one side delivers a flux that varies across the illuminated area, which quietly biases any quantum efficiency you calculate from an average meter reading.
A fourth loss is incidence angle: oblique light delivers flux reduced by the cosine of the angle, so a tilted window that still catches the spot still under-delivers.
Use a fluorescent target card at the window plane with the lamp at low power. Mark the center, rotate the card 90° and mark again; if the centers disagree, the spot is elliptical and the lamp is not seated correctly. Do this before the reactor is mounted.
For centering, replace the card with an aperture smaller than the window and move the source until the clipped ring is even all around. Confirm with a meter at the center and at four points near the window edge. Symmetry tells you about centering; spread tells you about source uniformity, which alignment cannot fix.
Burn paper locates the focal region but leaves ash, and ash on a window is a transmission loss until it is cleaned off. Use the lowest power that marks, keep it off coated optics, and treat it as a one-time check.
Repeat steps 2 to 4 once the lamp has reached operating temperature. The housing, lamp mount and reflector all expand, and convection deflects the beam slightly, so a beam centered cold sits a little off hot. Record both sets of numbers; a later change then shows up as a deviation, not a suspicion.
Run the same check after any bulb change, after the bench is moved, and after the reflector is cleaned or adjusted. An aligned source stays aligned only until someone opens the housing.
Work at low lamp power with a fluorescent or white card at the window plane. Mark the spot center, rotate the card 90° and mark again to confirm the shape, then transfer the mark to the window frame. Outside the visible range, use a card coated for that range or a pinhole scanned across the beam with a meter.
Yes. A replacement lamp is never seated in exactly the same position, and the arc gap and its orientation relative to the reflector differ slightly between units. Repeat the coaxial check and the centering step after every bulb change, and repeat the hot re-check, because the new lamp will expand differently from the old one.
Usually because the alignment is drifting thermally, or because the window is fogging, or because the lamp output is settling as it ages. Compare a cold reading taken before the run with a hot reading taken at the end, and check the window for haze. If the spot center has moved, the mount is the problem; if it has not, the source or the window is.
