Food Science & Safety

NMR can Detect Adulterants in Milk Powders

Approximately 9.2 billion pounds of milk powders (MPs) were produced between 2013 and 2015. This huge quantity demonstrates the importance of these products in many people’s diet as they contain essential nitrogen-rich proteins.

However, some companies add high nitrogen containing compounds to MPs to pass quality assurance (QA) testing. This fraudulent behavior allows companies to add less nitrogen-rich proteins to the products, thus increasing their profits. This conduct can have disastrous effects; in 2008 melamine was added to MP, which caused 300,000 people to fall ill and 6 babies died.

At the time, melamine was an unknown adulterant and so these MPs were able to pass QA because only specific compounds were tested for. Since then, the food safety industry has moved towards non-targeted methods to detect known and unknown adulterants.

This approach uses chemometric tools, such as infra-red (NIR) and Raman spectroscopy, to scan multiple authentic samples and generate a reference set. The variance is then calculated, and confidence limits are established to allow unknown samples to be classed as authentic or non-conforming. Using conformity index (CI) analysis, even small differences between data points can enable adulterants to be detected.

Recently, a proof of concept study has shown that solution-state 1H NMR spectroscopy combined with CI analysis can perform non-targeted detection of several MP adulterants. This technique also has the advantage of high sensitivity, minimal sample preparation, and high sample throughput.

Using a Bruker Avance III 600 MHz spectrometer, authentic MP samples from 46 global sources were analyzed to create a reference set. CI analysis was then performed to determine the variance in the data and a maximum CI (MaxCI) limit of ≥7.38 was decided upon. A range of adulterants were then analyzed and any samples with a MaxCI ≥7.38 were classified as non-conforming.

This method was highly sensitive to small nitrogen-rich organic molecules melamine and dicyandiamide (DCD), which are hazardous to human health. Both compounds have peaks in regions that are unhindered by MP signals (5.95 ppm and 6.58 ppm, respectively), which enable concentrations of 0.005% w/w and 0.05% w/w, respectively, to be detected.

Adulterants that do not pose significant health risks were also assessed. Sucrose and maltodextrin also have signals in regions free from MP metabolite peaks (5.16 ppm and 5.46–5.47 ppm, respectively) and were able to be identified at concentrations as low as 0.5% w/w. Urea (0.5% w/w) and ammonium sulfate (5% w/w) were also detected, but as these adulterants had no peaks in unhindered regions the method was less sensitive. Protein adulterants were also tested but due to their poor solubility in DMSO were unable to be identified.

The sensitivity achieved in the detection of the hazardous adulterant melamine is much higher in comparison to non-targeted methods combined with NIR and Raman, indicating this approach has great promise in the QA of MPs.


  1. Bergana MM., et al. (2019). Non-Targeted Detection of Milk Powder Adulteration by 1H-NMR Spectroscopy and Conformity Index Analysis. Journal of Food Composition and Analysis. https://doi.org/10.1016/j.jfca.2019.01.016.
  2. bbc.co.uk. (2010). China Dairy Products Found Tainted with Melamine. https://www.bbc.co.uk/news/10565838