NIR can predict Ca/p/salt , YES or No

kongfanli's picture

Dear all

I work in the NIR Center of C.P. group in China. We are feed factory, and our group used NIR for more than 10 years.  we have the feed calibrations with the Ca, P and Salt parameter, they are mainly provided by lime stone (formula: CaCO3), dicalcium phosphate (formula: CaHPO4·2H2O), and pure Salt. The calibration seems OK, P calibration is the best and Salt calibration is the worst in the 3 parameter, But now we have some problems, recently we did a test to add manually CaHPO4 ,CaCO3 and Salt to the feed in different proportions, we used the present calibration to predict these feed, the results do not change observably. My boss now want to delete these parameter in the calibration, but I think it’s so pity to do this.

Can you give me some advice why our results is like this, because our calibration ?we collected data? Or any other reasons?

Next question is whether NIR can predict Ca/P/Salt?

And the last, do you have any information or literatures about NIR predict Ca/P/Salt in feed.

Thank you so much for your help




td's picture

Hello Fanii,

Welcome to the Forum.

I am not surprised that you are having problems with your salt calibration. NaCl does not absorb in the IR or NIR regions. However you can get salt calibrations from its interaction with water. You need to hydrate the samples and then dry under standard conditions or (if you can handle liquid samples) extract the salt with H2O and prepare a calibration for the salt in the usual way.

Hope this is helpful.

Best wishes,

Tony Davies

kongfanli's picture


thank you very much for your advice ,but we want to konw the content of Ca/P/Saltt in the feed, not pure salt or other two raw material. do you understand me ?

hwswien's picture

Dear Fanli,

I congratulate your boss that he wants to delete the calibrations. The only compound which absorbs is the crystal water in the dicalcium phospate. I do not know what you measured so far, certainly not the concentrations of these salts. Possibly some correlations of the salt concentations with another parameter.

Kind regards,



Heinz W. Siesler

Dept. of Physical Chemistry

University of Duisburg-Essen

D 45117 Essen, Gernmany


Drajay.mane's picture

Dear Sir

Please add more on issue noted below


1.Which Molecules/ elements get absorbed in NIR region ( 800 to 2500nm)

as many companies in India markating NIR Machines claiming Analysis of Phosphorus, Salt, Calcium, Ash, Acid insoluble ash is possible with NIR.

As per my knowledge only O-H, C-H And N-H GET Absorbed.

Thanks and regards

Dr. Ajaykumar

Animal Nutritionist

Belgaum, Karnataka



hlmark's picture

It's true that the majority of NIR absorbances are due to -H vibrations, and also true that a major source of these vibrations are due to -OH. That said, water is a premier source of -OH, in addition to organic molecules containing -OH in a functional group. Water can also occur in inorganic substances as water of hydration, and there are likely to be some shifts in the -OH vibrational frequency depending on the structures the water finds itself in. This can give rise to the possiiblity to distinguish those various salts via their indirect effect on the water of hydaration they contain. In the experiment described, that started this discussion, apparently the pure salts were added to the feed mixtures. It seems likely that these added salts (Ca, P, etc) were anhydrous, and therefore did not contain the "normal" amount of water in the "normal" environment, that a hydrated sample of these same salts normally would.




Drajay.mane's picture

Thank You Sir

For your Input

khaled's picture

Dear All,

Just for information, I was once ever made calibration for Fish Meal And Meat and Bone Meal, for Calcium and Phosphorus and NIR predict very well. Good Also in validation.

My Finished feed calibration also good, Calcium and Phosphorus can predict very well. Even though we add Calcium Carbonate more on the Feed, NIR responds very well.

But this does not happen with the Salt calibration.

Salt Calibration is not so good, and when we add Salt more in the feed, the NIR do not respond at all. So we delete calibration for Salt.

But we use calibration for Calcium and Phosphorus.

Kindly please suggest what is happening, is it because the chelate organic?

I read in some paper NIRS can predict Mineral very well in chelate organic.

Many thanks for your input,

Best regards,

Khaled Sungkar

NIRS Manager at Feed mill company






AUNIR's picture

Dear Fanli

I agree with you, it is a shame to remove these calibrations from your prediction set. In many cases for QC on the final feed all we need to be able to identify is

Has the mineral been added?

Has the mineral been omitted?

or has there been a double addition of the mineral?

NIR calibrations are certainly capable of this. However, there is one caveat you need large datasets to ensure the calibrations can pick up all variations in formulations. We have developed calibrations with over 15000 samples and these have proven very capable.

I have uploaded the performance statistics of these calibrations for your information.

Kind regards