Identifying missing ingredients in a blend for energy drinks


A sports nutrition company was committed to improving quality control throughout its organisation. They deployed Chemometric Brain to analyse each batch of raw material and the final blend of ingredients to produce energy drinks.

They used both quantitative and qualitative analysis to ensure every batch has the right composition and the qualitative analysis, which only takes a few seconds, highlighted an issue.


Blend of ingredients to produce energy drinks.


One of the manufactured batches (BATCH 1; red circle in the first image) was not compliant and showed anomalies with respect to the rest of the batches.


As we can see in the following picture, when validating BATCH 1, Chemometric Brain detected the defective batch, thus placing it outside the confidence zone defined by the specific model for the target product.

The customer thought that the problem could be the lack of an ingredient in the blend: taurine or maltodextrin. To verify this, they added both ingredients, separately and together, and after recording the new NIR spectra, the batches were re-evaluated using Chemometric Brain.

The validation of the modified batches (marked in the second image) on the model available for this product showed that the composition of the original blend was incorrect. In the following figure you can see that the samples containing maltodextrin are the furthest ones from those included in the model (BATCH 1_MALTO; red circle in the image). In contrast, the samples containing a mix of maltodextrin and taurine are closer to the confidence region defined by the model, but they still seem to be non-conforming (BATCH 1_MALTO-TAUR; orange circle in the image). Finally, batches to which taurine had been added are the only ones conforming since they appear in the central zone of the mesh (BATCH 1_TAUR; green circle in the image).


By using Chemometric Brain, it was possible to identify that the missing ingredient in the blend was the taurine. Therefore, our customer was able to detect early variations in the composition of manufactured batches through a quick and simple analysis, which can be performed by any technician with minimal training and avoiding human error. The problem was solved without the need for external analysis, thus saving time and money.

This system is fully replicable to any food company and allows to increase quality control and food safety.