I did NOT award myself!! I don't know what happened there.

The both curve sets appear to have some variation which I believe is simply attributable to the repeatability error of the spectrometer with >2nm resolution at best. The LED curves, with broader bandwidth, I believe more easily shows some filtering effect of glass -- mostly from the appearance of a broader and smoother curve with the quartz as might be expected. That said, I don't have an explanation for the higher signal level for glass out at 460nm. The 405nm laser plot appears to show there is no significant attenuation out to 400nm for glass. However, a 405nm comparison between quartz and acrylic cuvettes might be a more practical comparison for those using the PLab for oil sample tests -- acrylic might have more attenuation at UV wavelengths.

Hey Dave! I don't know either about the higher signal level at the 406nm peak. That makes sense about testing the acrylic and quartz together, I know the acrylic ones are cheaper but are more prone to damage from certain solvent mixes and preps. I have a wide assortment of glass when Edmond Scientifics use to sell defective and chipped lenses in bulk, but those days are over they don't do that anymore but I still have a lot of them.

Right and I'm still believing that repeatability errors with the PLab-web-cam-slit-hardware configuration all too often hints at differences which are not really there or can only be validated when there are sufficient samples variants to test. ..... I mention the acrylic as those are the ones PLab sells so many?of and others may be using them already or in the future; so they are a candidate just because they are cheap and available (even if they are unsuitable for some solvents).

You are correct, it is difficult to keep the hardware parameters the same everytime, I was checking out your prototype spectrometer and the design is spot on, although the board camera that you are using, the only specs I can find are just dimensional, MP and FOV info, I've searched until the cows came home and found some 5mp board cameras but specs are few and the ones that have them you have to get them custom manufactured for an arm and a leg.

Ha - @dhaffnersr, sorry, I think you found a bug in the "Post related work" button -- it seems it copies even the "barnstar" tag from the older post. You can delete it by removing the tag barnstar:empiricism. Sorry, and if you'd file a bug (see footer for details) that'd be helpful! Thanks.

One way to test @stoft's idea that this may just be within the repeatability limits of the hardware, rather than a difference in materials, would be to scan a glass cuvette a few times (removing it each time, say, so small differences in its placement happen) and see if the differences vary more or less than the comparison here. The same might be done with repeated scans of a quartz cuvette. It'd be a good plot to examine!

Hey Jeff! Yes I will file that bug report. I think as far as repeatability results go, your idea is a logical way to go, I'm just concerned when data is captured and processed, can the slight incontinency in the hardware be accounted for and if so, we can formulate a protocol to compensate for it.

Glass absorption tends to increase at shorter wavelengths, but it varies strongly with the type of glass. Here's a link to plots of absorption in some optical materials from Newport, a major optics company. https://www.newport.com/Optical-Materials/144943/1033/content.aspx