Question: Question: Can DIY-spectrometer be used for analysis of soil

Hello, what are you interested in specifically measuring, in nutrients? Nitrogen, potassium, phosphorus? I believe there are some colorimetric reagents that can be used in conjunction with a spectrometer. In addition, samples of known concentration will help you get quantitative about your measurements.

check out this tool, using reagents: here may be what you are looking for?

Oh yeah -- actually that's an open source method too, by Joshua Pearce's lab; you can find some of the parts here: http://www.thingiverse.com/thing:958405/

@stoft and/or @gretchengehrke, i'm curious about your input on this -- does the beer lambert law hold for opacity of a reagent? I guess it depends on if the reagent color change is linearly related to the concentration of the target compound (nitrate).

If it does, this is another reason to push our #cuvette-frame design forward, as we'd likely be able to do a similar photometric analysis of the reagent color change using either a fixed-exposure camera or two or more standard control samples at known nitrate concentrations.

It is my understanding that Beer's law is not the relationship between reagent's effect on color, but is the property of light passing through a uniform substance and the attenuation of that light with distance -- which is linear assuming the substance is uniform. So, if you had some chemical reaction, which was not linear with the concentration of reagent, that non-linearity would appear in a plot of the measured attenuation (from the spectral data) as the concentration changed. The Beer's Law linearity you are relying on is that given that the sample containers are all "identical" then the path-length for light is "identical" and therefore (with all else held constant - like the light source, camera exposure, etc) then the measurements of relative intensity will be directly related (linear or non-linear) with the concentration. [Note: Realize that there are many possible components to the issue of measurement stability and that the default PLab stuff is only a first step. You might find some of my PLab notes on spectrometer stability, noise, time-averaging, calibration and uncertainty of interest.] It's all in the numbers to estimate measurement limits vs your requirements.

I think that makes sense. I didn't mean Beer Lambert would be related to the reagent's affect on the color, but merely that if the reagent reacts linearly with the substance, then Beer Lambert would ensure (all else being equal, as you point out) linearity of transmission of light through the reagent-reacted substance. But controlling all the other variables is not simple!

@all who gave a comment/ answer:

Thanks a lot for your advice, sorry for my late reply. Since I regard myself as spectrometry-newbie reading the comments felt a bit like “biting off more than one can chew”, however the plan is complete / ready now: At first I'll learn about that colorimetric reagents (which I hadn't in mind before, thinking measuring mere reflection-effects would work as well). Secondly I'll probably learn about Beer's law or similar concentration measurement principles. (The nutritients I'm interested in measuring at the moment are in particular: Nitrogen, potassium, phosphorus but in the long run I'm eager to see which other nutritients can be spectrometrically measured).
Of course I'll also have a detailed look at the elaborate ready to use devices that were mentioned above for they seem to be a good alternative. Thanks again for sharing your knowledge and thus jump-starting my hobby-project.

Hi, all - i've been collecting up some info on reagents on this page; i'll repost my question and @stoft's answer there too:

https://publiclab.org/wiki/reagents