The Wonderful World of NIR

Near Infrared Spectroscopy, or NIR, is the analysis of near infrared light that when pointed at a sample, will absorb, or reflect the contents based on the samples molecular structure. Common molecular structures that are favorable to NIR are carbon-hydrogen bonds that are associated with fats, nitrogen-hydrogen bonds that are associated with proteins, or oxygen-hydrogen bonds that are associated with water. Once a sample is placed in the NIR to read, the NIR can measure multiple parameters. The main use of NIR measurements is to determine if the samples conform to a model or if it deviates and further testing is required. NIR can also predict results such as the percent of protein, moisture, or fat within a sample, all under one minute! 

Each ingredient that enters our North American plants at dsm-firmenich, will get scanned in duplicate and the spectra will be sent to the NIR specialist to determine if the sample conforms to the reference model or is rejected by the reference model.  For a reference model to exist, a total of 20 lots are needed to build the model. Below are two examples of an ingredient spectra and how we determine if they are either passing or failing through NIR:

When an ingredient or premix sample passes, it is stored in a file and used for future modeling of spectra. If a sample fails, an in-depth review of the spectra is conducted to determine if this product is a true failure, or there was slight variation from the model. In addition, a review of the Certificate of

Analysis for ingredients or the batch reports for premixes will be completed. If there is significant variation, additional tests or assays will be conducted validate the ingredient or premix. NIR is another great quality control method we have here at dsm-firmenich.

Published on

21 March 2022

Tags

  • Swine
  • Poultry
  • Ruminants
  • Premixes

Share

You are being redirected.

We detected that you are visitng this page from United States. Therefore we are redirecting you to the localized version.