When I designed my Light Counter LC2 system, I wanted to use the best possible sensor for platinum/palladium printing. On the surface, my design requirements were straightforward:
- The sensor had to have peak sensitivity at about 365nm
- I did not want the sensor to respond to visible light nor UV-B or UV-C, so as to avoid misleading measurements
- It had to be affordable
Achieving this turned out to be quite difficult. Here’s why.
The light we see with our eyes is part of a wider spectrum of electromagnetic energy. As shown in the following diagram, visible light has a wavelength that ranges from 400nm (violet) to about 700nm (red). Ultraviolet light has a wavelength between about 100nm and 400nm (just to the left of the visible spectrum in the diagram).
It’s well understood that platinum/palladium printing and other iron-based processes require ultraviolet light for the exposure; also that this should be UV-A, not UV-B or UV-C. The UV-A band is from about 315nm to 400nm. So far so good, because lots of electronic sensors are marketed as being suitable for UV-A.
However, sensor selection starts to get complicated when you consider the practicalities of iron-based printing.
The first complication is that iron is only really interested in one specific wavelength in the UV-A band: 365nm. UV light of different wavelengths won’t have an effect.
The second complication is that pretty much all electronic ‘UV’ sensors are sensitive to a wide range of light, not just the 365nm wavelength that we want for printing. Here’s an example response chart from a sensor that I tried quite early on. While its response at 365nm is fairly strong, its response between 400nm and 500nm is even stronger, which means it will give a misleading reading if any violet, blue or turquoise light is present during exposure.
The third complication is that different UV light sources emit different wavelengths of light. The chart below shows the spectrum emitted by a 300W Osram Vitalux bulb. There is a strong UV-A peak at about 365nm, which is excellent for printing. But the bulb also emits plenty of violet (410nm), blue (440nm), green (545nm) and yellow (680nm) light. If the sensor is responsive to these wavelengths, then you will get a misleading reading. The same is true for light sources that emit significant amounts of UV-B or UV-C light.
This problem gets worse if we measure a light source with the wrong sensor. For example, if we measured this Vitalux bulb with the Gallium Phosphide sensor shown earlier, we would have big problems due violet and blue light. The sensor is more sensitive to violet light than it is to 365nm UV; and has about the same sensitivity to blue as it has to 365nm UV. Therefore the sensor reading would be dominated by the violet and blue light rather than the 365nm UV.
It gets worse again if you are using the same sensor to measure different light sources. If two light sources emit the same amount of 365nm UV, then you would want the sensor to give the same measurement for them both. But if one is ‘blue heavy’ and the other isn’t, then a blue-sensitive sensor will give different readings for the two light sources.
Eventually I realised that the best design choice was to use a sensor with peak sensitivity close to 365nm, and then to pair it with a narrow-band 365nm filter.
Here is the response of my final UV sensor design. Its peak response is very close to 365nm, which is good, and it has no meaningful response to visible light or UV-B. As a result it should work well with any UV light source.