Chemical Forums
Chemistry Forums for Students => Analytical Chemistry Forum => Topic started by: KayVL on April 18, 2021, 02:11:22 PM

Hello, I am working on a project for school around FTNIR.
In quantitative measurements, why should one not perform with increasing or decreasing concentration? I think it has something to do with the influence of temperature, but I'm not sure. Can someone help me answer my question?
Thank you in advance.

In quantitative measurements, why should one not perform with increasing or decreasing concentration? I think it has something to do with the influence of temperature, but I'm not sure.
Question doesn't make sense. Why should one not perform what with increasing or decreasing concentration?

I had not worded my question correctly I am sorry.
When one wants to determine concentrations using ftnir, a model or method is needed. This model can be established by measuring a calibration set of samples. Through these measurements, a calibration line can be established from which, an unknown concentrations can be calculated. This series of samples (used to create the calibration line) must not show collinearity. Which means that they can not be measured in a linear de or increase concentration of the components. Doing so can lead to inaccurate measurements and erroneous results. I do not know why collinearity is a problem, and why it shows wrong results. I hope you understand my question.
Thank you in advance.

Do you have a link to where you read this?
(EDIT: if you are referring to some kind of multivariate calibration/analysis, this is a mathematical problem, not a spectroscopic one. The idea is basically that variables need to be independent. If they are linearly dependent on each other, then it can introduce error into the analysis. I looked around for an easy explanation, and found this, which may be helpful: https://statisticsbyjim.com/regression/multicollinearityinregressionanalysis/)

I read this in the manual I used when I did the experiment. However, there was no explanation as to why this was a problem. This is a manual that I only have on paper and not online, I'm sorry.
The software I used is indeed based on a multivariable calibration analysis. It allows for quantitative analysis of different spectra consisting of bands that have significant overlap. So as you said it will probably be a mathematical problem. I have already read through part of the link you forwarded and this will indeed be the explanation. I'll look further in to it.
I first thought that it might have to do with linear fluctuations in temperature or concentration (due to for example: heating and cooling of the sample or evaporation, because I used acetic acid wich is a volatile substance).
I'll continue reading the article.
Thank you for you help and time.

Basically, the idea is this: suppose you are trying to calibrate to a peak, where the peak height is proportional to the concentration of whatever causes the peak. That's easy to do. But NIR is an especially tricky region because all kinds of things absorb there. Suppose that something else also absorbs at the same wavelength as the substance you are measuring (and calibrating to). Could be water, could be air, could even be the substance itself because there are also lots of overtone bands and combination bands that can interfere. Call the substance you are measuring/calibrating A and the interferant B. If contribution to the peak height due to A scales linearly with the concentration of A and the contribution to the peak height due to B scales linearly with the concentration of B, then in principle these could be easy to separate by just varying either A or B while holding the other fixed. The problem is that if the contribution of the peak height due to A depends linearly on A and also to some unknown extent on the concentration of B (and vicevera), it is hard to deconvolute the signals just by varying A or B, since they are codependent. If this codependence is strong enough, it can introduce significant error into your calibration and is also hard to deconvolute because you may not know the degree of the dependence. Statistical methods are needed in this cases.

Now that you explain it like this, it makes perfect sense.
Thank you very much for explaining it to me. You have helped me a lot.
Kind regards
Kay