Clarification

Some array spectrographs (with bells and whistles) can perform better in the UV than others ...

A recent posting (now updated with a caveat at the end) created bit of a stir with colleagues in Germany. They were concerned that by using their Gigahertz instrument in the manner I reported, we hadn’t cast it in a good light.

That certainly wasn’t my intention, but I guess they have a case - that I should try to redress here. Their particular instrument is not just a “standard diode array spectrograph”. It includes optional extras - which can make all the difference. If you’re in-the-know, you can make use of optical filters that can be switched into the beam to sequentially reduce the stray light problem. With those filters and some accompanying fancy data processing they’ve developed, much better results can be achieved, as discussed here, here, and here.

There’s an excellent series of slides on the topic prepared by Sebastian Lorenz and his team at the Federal Office for Radiation in Vienna. The relevant plot- from the 7th page - is reproduced below. Data from this group’s array spectrograph manufactured by Gigahertz (red line, labelled ‘BTS’) shows quite remarkable agreement with that from a reference double monochromator (blue points, labelled ‘DM’), even for wavelengths less than 294 nm! The steeply dropping values at wavelengths below 300 nm are a far cry from those reported by Herndon et al. (and others who’ve fallen into the same trap).

We didn’t use those bells and whistles to obtain the data showed in that last post because we were simply trying to illustrate the more general problem of using standard diode array spectrographs to attempt to measure the sunlight reaching Earth’s surface at wavelengths in the UVC region.

I’ll try to persuade my colleagues at Lauder to see if they can put some time into getting similar agreement with their state-of-the-art instruments there. But time is money, and that’s in short supply there. In any case, that exercise would be restricted to wavelengths longer than 280 nm.

Even with all those improvements implemented by Gigahertz (and clearly not implemented by groups like Herndon’s), the reliable range of valid sunlight data transmitted to Earth’s surface is limited to wavelengths longer than 290 nm. With the current technology available, any reported readings in the UVC region aren’t worth the paper they’re written on.