# How to use a spectrometer to determine NPK

I'm trying to determine if it's possible to use a low-grade spectrometer to read NPK values of a solution.

I believe I should be able to take readings of the individual compounds (e.g., calcium nitrate) and then use that to "subtract" from a reading of the mixed/final solution and basically deduce how much of each compound is present - is that correct?

In Journal of Research of NBS, Section A: Physics and Chemistry, 1972, 76A, 469-482 (free PDF) the uv spectrum of potassium nitrate in aqueous solution is given. The spectrum exhibits a maximum at $\lambda = \pu{302 nm}$ with a molar absorption coefficient $\epsilon$ as low as $\pu{7 l\,mol\,cm^{-1}}$.
Even if this wavelength wasn't below your detection range, the low $\epsilon$ wouldn't get you very far anyway.