The halflife is a good explanation of the strength of the radiation generated by the tritium in the tube, but saying that the brightness falls off in the same manner is making assumptions about the phosphors used to line the tube and convert the radiation energy to visible light energy. The light dimming curve would reflect the radiation strength, conversion efficiency of the phosphor at different radiation intensities and the deterioration of the phosphor itself over time. Hydrogen is easier to contain than helium in some respects but there will still be some minor leakage straight through the container walls, and any leakage will also decrease the radiation intensity. Does anyone have real data on the actual dimming curve of these?