Extrapolating Standard Addition with PLS-R

Report
Question
views

Please explain why do you think this question should be reported?

Report Cancel

In a univariate experiment one can add  known amounts of an analyte to a sample. The equation of the linear regression line y=ax+b can then be extrapolated to where y=0. The concentration of the analyte is then equal to the absolute value of x.

I like to use this method with near-infrared spectrometrie (NIRS) initialy with only 1 analyte. It is relatively straight forward to use PLS-R for standard addition. But how would I calculate the initial analyte concentration?

Kind regards, Kees

About the Author

Thread Reply

  1. External Admin
    31/03/2020 at 12:35 pm

    Please briefly explain why you feel this answer should be reported .

    Report Cancel

    Standard addition data is not really suited for inverse regression analysis like PLS. An alternative would be to attempt instead a multivariate curve resolution (MCR) model of the spectral data utilizing whatever constraints you might know about in your application: for example, including what a “zero” spectrum looks like, enforcing non-negativity, enforcing known bands of the analyte (and interferences), etc. One idea would be to try an MCR where one of the spectral profiles is the difference between the highest spiked spectrum and the original unknown- and assign this as a “pure S” constraint in the MCR. Unfortunately- if you only have a few spiked samples in the dataset (plus the original unknown) there’s not a lot of population for MCR to extract all the relevant spectral components.

    Find more information about MCR in the Unscrambler Help menu: Help -> Contents -> MCR .

  2. 01/04/2020 at 10:37 am

    Please briefly explain why you feel this answer should be reported .

    Report Cancel

    Thanks for the reply.

    I am familiar to a certain extend with MCR, but I have never considered it. I will certainly try it.

     

Leave an answer