Fabry-Perot measurements of barium temperature in fluorescent lamps

Summary form only given: It is generally understood that the ultimate demise of an electrode of a fluorescent lamp, and therefore of the lamp itself, is due to the cumulative loss of the work function- reducing emitter material that occurs over the course of lamp life. Evaporation and sputtering are the predominant loss mechanisms. Sputtering is very intense during the starting phase of a lamp, especially if the electrode is initially at room temperature (so-called ´instant start´). In these situations sheath potentials achieve levels of several hundred volts. It is straightforward to sense the plethora of atoms (predominantly, and therefore most easily, barium) that are liberated during this period (e.g., by optical absorption). During steady-state operation of the lamp the electrode emits electrons thermionic ally, and emitter loss is almost exclusively by evaporation. It is generally assumed that very little, if any, sputtering occurs during this time because electrode sheath potentials achieve peak levels, during the cathode phase, of ~13-15 volts. This is most likely less than the low energy threshold for sputtering, which is not precisely known, but thought to be ~16 volts (barium atoms sputtered by mercury ions). However, since the peaks are very close to the imprecisely-known threshold, and since sputter yields increase very rapidly with energy beyond the threshold, sputtering may be more significant than realized, and/or it may become significant if the lamp is operated in non-standard ways (e.g., with discharge current significantly lower than that for which it is designed). In order to determine if, and under what conditions, sputtering may occur during steady-state, we have developed an optical diagnostic that can distinguish between evaporated and sputtered barium. The diagnostic senses the 553 nm barium resonance radiation emitted from the discharge near the electrode, and it incorporates a tunable Fabry-Perot interferometer that can spectrally reso- ve the line. Since the line is Doppler broadened, the temperature of the barium is deduced. The underlying idea is that the (effective) temperature of sputtered barium is higher than the temperature of evaporated barium. The diagnostic has proven to be successful in a number of situations. Measurements have been made on argon and on Hg-argon discharges with various pressures, discharge currents, and auxiliary coil currents. Measurements are phase-revolved for a.c. driven discharges.