How do I find the minimum number of standards needed to calibrate a spectrophotometer using CLS? We never talked about CLS in class so I'm not sure of my approach. Would 1 standard solution per analyte be enough? We'll be plotting absorbance vs. concentration then finding the residuals from the best fit line.
That depends on the nature of the analyte and matrix. Ideally (Beer's Law) A should be linearly proportional to concentration but that may not be the case. As an example of this in the water industry barium chloride is often added to samples to determine how much sulfate ion they contain. Spectrophotometers are used to determine how much barium sulfate is in suspension. The relationship between A and sulfate concentration is not linear in this case.
What you should do is make up a series of standards spanning the range of concentrations you are interested in, measure the absorptions and plot A vs. c. Now start fitting curves, starting with linear and progressing to higher order until doing so does not make the residuals look any noisier. That gives you the order of the response curve you need and the number of standards required is one more than the order.
Ideally, the matrix wont interfere and Beer's law applies. In that case when c = 0, A = 0 and you only need one standard.