[Edoman]

I have an ancient Perkin Elmer 330 UV-VIS-NIR (dual beam) and the deuterium lamp has burned out. Those things are quite expensive and I am wondering whether or not it is feasible (in a cheap and nasty way) of using some LED's in place of the lamp? Plausible or not?

E

[Arkcon]

If they make LEDs that emit intensely at the short wavelengths you need a deuterium lamp for ... well, I want one.  But I suspect they're not intense enough, or cheap enough, if they exist at all.

[Edoman]

As I understand it, the most efficient LEDs peak at around 395nm, but I think that some are available working down to a little over 200nm. In my particular case, I don't actually have a current interest in using the spectrometer down in the UV; my interests are in the the VIS-NIR region.....but, this instrument goes through a self-test on start-up which fails if the UV source is not good, but I don't know what wavelength that check is made at.

E

[Arkcon]

It would probably use one of the peaks in the deuterium spectrum:  http://en.wikipedia.org/wiki/Deuterium_arc_lamp

[Edoman]

...and those peaks are up in the visible region, which makes me wonder whether the self-test could be fooled using a tungsten lamp (white light) - and it already uses one such lamp as the NIR source!

E