When looking at the literature, there is quite much variation on whether UV-B lamps are filtered with cellulose diacetate or not.
- Which UV-B light sources can be used in biological research without interposing a UV-C and short wavelength UV-B blocking filter?
The normal UV-B lamps like q-panel UVB-313 and Philips colour code “12” are broadband and emit UV-C, UV-B and UV-A radiation and a significant amount of visible light all the way to the orange-red region. Most, but not all, UV researchers are aware of the damage or spurious responses that UV-C can cause and filter these lamps with cellulose di-acetate.
Narrow band UV-B lamps like those with Philips colour code “1” have a much narrower peak of emission, and is easy to think that no filtration is needed. However, they do emit UV-C albeit less than the broad spectrum lamps. The difficulty is that a very low irradiance of UV-C may be difficult to measure but still effective on organisms. So, it does not cost much to play it safe and also filter these lamps. Otherwise we need to assume that the UV-C and short wave UV-B do not affect our experiments. Not proven assumptions always weaken any evidence when testing an hypothesis.
My own recommendation: Do filter with cellulose diacetate all UV-B lamps.