Honey: Adulteration Detection

Honey: Adulteration Detection

Miel

1. The Problem

Natural bee honey commands a premium price thanks to its nutritional composition, characteristic taste and natural qualities, and that premium makes it a prime target for adulteration. Dishonest producers blend honey with low cost syrups, most commonly glucose syrup (GS), rice syrup (RS) and high fructose corn syrup (HFCS), to keep the sweet taste at a fraction of the cost. They also inflate production by excessively feeding bees with these sugars during the main nectar flow, which can even be done while the harvest is still in the hive. Honey adulteration with low cost syrups remains a global concern, including at adulteration levels as low as 1 to 10 percent.1,2

The standard method to detect these syrups, the carbon isotope ratio (CIR) AOAC 998.12 technique, only identifies sugars derived from C4 plants such as corn or cane, is destructive, requires specialised laboratory processing, and is known to produce false positives in honeys such as pine honey.3 Chromatographic, elemental profiling and thermal approaches share the same drawbacks: destructive, costly and slow for routine screening.

2. The Production Chain

Honey follows a short but high risk chain, and every stage offers an opportunity for rapid authenticity screening:

  • 1. Beekeeper and apiary: bees collect nectar and mature it into honey inside the hive; European law defines honey as the natural sweet substance produced by bees, with no added sugars permitted.
  • 2. Harvest and extraction: frames are uncapped and the honey is extracted by centrifuge. This is where the harvest can be verified with portable spectroscopy before it is ever blended or bottled.
  • 3. Filtering and processing: the honey is strained and optionally pasteurised before packing.
  • 4. Packer and wholesale: batches from different beekeepers are blended and packed for export and retail. This is the classic control point for adulteration screening of incoming lots, where fast spectroscopic tools are most valuable.
  • 5. Retail: jars reach supermarkets, specialty shops and export markets, where official controls may also sample the product.

3. Where NIR Can Contribute

Adulteration screening with infrared spectroscopy can be applied at the control points described in the production chain above:

  • At the harvest and extraction: beekeepers and cooperatives can verify the purity of their own harvest before selling it, protecting themselves from syrup feeding practices or accidental blending.
  • At the packer and wholesale: incoming batches from many beekeepers are blended before bottling, so this is the classic control point. Portable or benchtop instruments can screen every incoming lot for syrup additions in seconds.
  • At official control laboratories and retail: fast, non destructive screening supports enforcement of labelling rules and gives retailers a tool to verify the products they buy.

None of these checks require destroying the sample or a laboratory workflow: a spectrum is all that is needed, and the result is available in seconds.

Proof of concept: our own study

Our own research demonstrates that this is feasible at the low adulteration levels that matter commercially. In our study, authentic honey samples collected from beekeepers across Argentina between 2015 and 2021, representing diverse floral sources, geographical origins and seasonal harvests, were adulterated in the laboratory with rice syrup, glucose syrup and high fructose corn syrup at concentrations from 1 to 10 percent. Spectra were acquired with a Fourier transform infrared microscope (Thermo Nicolet iN10) covering the mid and near infrared range down to 650 wavenumber units.

PLS-DA per adulterant1 to 10%excellent to perfect discrimination
CNN-ANN combined0.81precision, all three adulterants
Screening GUI3 to 4 sper spectrum, 100% consistent

Three dedicated PLS-DA models (one per adulterant: RS, GS and HFCS) were built and validated on independent test sets, achieving excellent to perfect discrimination even at the lowest adulteration levels. A CNN-ANN model was then trained on the combined dataset to detect all three adulterants in a single robust model, delivering a balanced accuracy of 0.79 together with the strongest specificity (0.85) and precision (0.81), the profile of choice for official control laboratories where minimising false positives is critical.

All models were integrated into a user friendly interface that accepts individual spectral files or entire folders, aligns and normalises the spectra automatically, and returns the adulteration verdict in 3 to 4 seconds per spectrum with 100 percent consistency against the original model outputs. National laboratories can therefore screen honey authenticity rapidly, non destructively and without chemometric expertise.

4. Other Applications

The same spectroscopic and machine learning toolbox supports further authenticity work in the honey chain:

  • Botanical and geographical origin: the composition of honey reflects its floral source and origin, and spectral fingerprints can be used to authenticate declared origins and detect mislabelling.
  • Comparative screening: infrared methods have been used since the early 2000s to quantify syrup additions in honey, and current reviews consolidate the analytical techniques available for honey authenticity work.2

5. References

  1. Kelly JD, et al. Detection of Honey Adulteration by Addition of Fructose and Glucose Using near Infrared Transflectance Spectroscopy. Journal of Near Infrared Spectroscopy. 2003. Available from: JNIRS 2003
  2. Recent Advances in Analytical Techniques for the Determination of Authenticity and Adulteration of Honey and Other Sweeteners. Bentham Books. 2020. Available from: Bentham 2020
  3. Akyüz E, et al. Elucidating the false positive tendency at AOAC 998.12 C 4 sugar test for pine honey samples. Journal of Food Composition and Analysis. 2022. Available from: Food Comp Anal 2022
  4. Council Directive 2001/110/EC relating to honey. EUR-Lex. Available from: eur-lex.europa.eu