I have Mercury-vapor lamp emitting UV light. I am looking for the nice solution to cut off spectral lines for UVA or UVC, to make it one day to generate ozone and another day to use it for disinfection as UVC band lamp. I mean using precoated tubes and put mercury vapor lamp inside or using coated glass as a shield.
TL;DR - You need two different lamps, one that has a phosphor to absorb the 254 nm light, and one that doesn't. Alternatively, you can use one lamp, and use a filter to select the wavelengths.
Mercury vapour lamps emit a significant amount of both UV-C light (around 254 nm) and UV-A light (around 365 nm). This includes fluorescent bulbs used for lighting a room. The ultraviolet emission is a fundamental property of the element, not some kind of engineering accident. Light at these wavelengths is harmful, so you don't want it flying all over the place. The solution? Coat the inside of the lamp with a substance that absorbs the UV light. In some cases (but not all), the glass of the lightbulb itself can absorb the UV light.
Except now you're turning lots of luminous energy into heat, which is wasteful. A smarter solution is to coat the lamp with a phosphor, which absorbs the UV-C and UV-A light and reemits the energy at longer wavelengths. Depending on the phosphor, the reemitted light will have a different spectrum.
If you look at a 254 nm UV-C bulb and a 365 nm UV-A bulb (while they're turned off..), you'll notice that the 254 nm lamp is completely see-through, whereas the 365 nm lamp often has a white powdery coating the inside. This powder is a phosphor, which is chosen to absorb the 254 nm UV-C light, while letting the 365 nm UV-A light through (as well as other lower energy visible wavelengths). Some 365 nm UV-A lamps also have a coating to absorb the any visible light generated, such that relativity pure invisible 365 nm light shines through. This is popularly called a "black light".
Note that this does NOT imply every mercury lamp with a phosphor coating blocks the UV-C light it generates.
Light and Ozone Production
Ozone is generated not by UVC or UVA but by Vacuum UV (VUV) in the 100-200nm wavelength range 1.
Low pressure mercury lamps generate strong VUV (184nm) and UVC (254nm) light, and medium pressure lamps generate light at various frequencies in the range of 200nm-600nm 2.
When it's said that UVC light is disinfecting or germicidal, it works through the absorption of light energy by DNA which then breaks down the molecules. Ozone (O3) is effective at disinfection because it is a reactive species of oxygen (causing faster oxidation than O2) 3. The germicidal effectiveness of light by DNA disruption peaks at around 270nm, but while the 254nm light produced by mercury lamps is not ideal, it is still about 75% as effective as 270nm light. The 180nm light band produced by mercury lamps is about 1% as effective as 270nm through that mechanism 4.
UVA light can produce germicidal effects, but in this study, it took 3 minutes for UVC to produce total irradiation and 60 minutes for UVA.
While medium pressure mercury vapor type lamps, including the HID type and flourescent tube type, produce UVC, the borosilicate or silicate glass they are generally made from absorbs >99% of the UVC light, protecting humans from the effects of those wavelengths 6. To use the lamp for germicidal purposes, it must be made from fused quartz or fused silica glass that allows UVC to pass 7.
From personal experience, medium pressure mercury vapor lamps for germicidal applications do not produce significant ozone. It's obvious if you're in the vicinity of a lamp that can produce ozone as you will smell it quickly.
If your mercury lamp is producing VUV and UVC, it might be difficult to block UVC while not blocking VUV, but if you just want the ozone, cover the light, blocking all the light, but provide airflow between the lamp and the shade to move the ozone where you want it to go. If you are interested in blocking UVC and letting UVA through, a borosilicate glass (e.g. Pyrex(TM)) tube would be sufficient.
Blocking UVA and letting through UVC has you blocking longer wavelengths and passing shorter wavelengths. This is done with UV passing, visible light blocking filters that create black lights. I don't know of a similar material that blocks UVA while passing UVC. A material like this would also probably have some safety concerns as it would likely block much of the visible light too, yielding an invisible sunburn/eyeburn causing machine.
Phosphors, as used in flourescent lights, generally absorb shorter wavelength light and use that energy to emit longer wavelength light known as the Stokes shift. Although some materials can absorb a longer wavelength and emit a shorter wavelength, I don't know if that is something that has been commercially exploited 8. There are phosphors that can absorb VUV and emit 290nm, germicidal, wavelength light, and ones that can absorb VUV and UVC to emit UVA wavelengths.