Can a Chemical Fingerprint Identify Authentic Perfume?

Perfume contains not only ingredients that create its scent but also alcohol that dissolves them and ingredients that help maintain the product. Could measuring this mixture and comparing the resulting signal patterns provide clues for distinguishing authentic perfume from other products?

In 2021, Barbara Pacholczyk-Sienicka and colleagues used an analytical method called NMR to compare chemical fingerprints of perfumes. Among the samples studied, how clearly products differed depended on which part of the signals was examined. Small differences also appeared between authentic samples of the same perfume, raising questions about how to establish identification criteria.

Measuring the liquid’s composition rather than smelling its fragrance

The researchers aimed to explore whether authentic perfumes could be distinguished from counterfeits and other products by combining NMR, which directly measures perfume constituents, with statistical methods that compare many measurements together.

They used proton nuclear magnetic resonance, or ¹H NMR, which focuses on hydrogen atoms. When a sample in a strong magnetic field is exposed to radio waves, hydrogen atoms in its molecules produce signals at positions reflecting their surrounding structures. In a mixture such as perfume, these signals overlap to form a graph of peaks of different sizes. The combination of peak positions and sizes can be compared as a chemical fingerprint.

This differs from an experiment in which people smell and evaluate airborne fragrance. Alongside fragrance ingredients, solvents and additives in the liquid are measured. The authors saw potential in comparing the overall signal pattern without identifying the ingredient responsible for every peak.

The authors identified simple sample preparation and the ability to examine less volatile and heat-sensitive components as advantages of NMR. They also noted that overlapping signals from multiple constituents present an analytical challenge in mixtures such as perfume.

Measuring 28 samples and comparing patterns in 26

The study involved 28 samples corresponding to six branded fragrances: eight authentic samples, six classified by the authors as counterfeit, and 14 inspired products drawing on existing fragrances. This article likewise treats the latter two categories separately. Classification as inspired does not by itself establish illegal counterfeiting.

The authentic samples included two each of Light Blue and J’adore and one each of the other four fragrances. The researchers first checked similarity between the pairs of authentic samples. Subsequent pattern comparisons used one authentic sample per fragrance, for a total of 26 samples.

Small amounts of perfume were diluted in the measurement solvent and prepared under the same conditions. Five measurements were repeated for each tube. These were repeated measurements of the same sample, not tests of five separately purchased bottles or five manufacturing batches.

Principal component analysis, or PCA, was used to compare the many signals. It condenses differences in measurements into a small number of axes and represents samples as points on a plot. Here, the arrangement of points indicates how similar or different the selected signal patterns are.

Across a broad measurement range, solvent differences dominated

When signals over a broad range were first analyzed together, the most prominent differences mainly involved solvents such as alcohols. Perfume contains large amounts of ingredients that dissolve fragrance materials, so their strong signals heavily influenced the comparisons.

The researchers therefore divided the graph’s horizontal axis into three ranges. Here, ppm is a scale describing where a signal appears. It does not indicate the concentration of an ingredient in parts per million.

The 0.6–3.0 ppm range did not distinguish authentic from inspired products, and the 3.0–6.0 ppm range did not produce clear separation either. In the latter range, some samples appeared as outliers, but signals associated with water and alcohols had a strong influence and were insufficient for clearly distinguishing fragrance types and categories.

The authors suggested that solvent-related differences might involve formulations intended to reduce production costs or the quality of the ethanol used. However, these were interpretations of measurements, not findings from direct investigation of manufacturers’ formulation intentions or raw-material sourcing.

Clues to formulation differences in the 6–8.5 ppm range

The most useful range was 6.0–8.5 ppm, where signals from constituents with aromatic rings and related structures occur. Focusing on this region made differences between authentic fragrance compositions and their corresponding counterfeits clearer on the plot.

Some inspired products, however, showed patterns quite similar to their corresponding authentic perfumes. The paper does not suggest that all non-authentic products should be treated as a single group. Both differences from counterfeits and similarities to inspired products appeared in the combinations of constituents.

Signals contributing to differences between samples in this range included the musk ingredient galaxolide, benzyl salicylate, and diethyl phthalate, used as a solvent and fixative. The comparison therefore concerns not only the stars of a fragrance but formulation characteristics encompassing fragrance materials and supporting ingredients.

Comparisons of individual signals did not show that every ingredient distinguished every category. For samples corresponding to Light Blue, for example, the galaxolide signal differed significantly between authentic and counterfeit products, but not between authentic and inspired products. The presence or quantity of one ingredient cannot serve as a criterion applicable to every perfume.

Accounting for both similarity and small differences among authentic products

Comparisons between authentic products are important for identification. The two Light Blue samples were very similar, and the two J’adore samples were also similar overall, although J’adore showed small differences. The authors concluded that future work should compare authentic products from different manufacturing batches.

This means more than simply increasing sample numbers. To turn these findings into identification criteria, researchers first need to know how much signals can vary even within authentic products. Calling a product counterfeit merely because it differs slightly from one reference bottle risks treating normal authentic-product variation as suspicious.

Authentic samples from different batches should be measured under consistent conditions to establish the range of normal variation. Criteria should then be tested on samples not used to develop them, checking whether authentic products are wrongly excluded or other products overlooked. Such validation is needed before practical identification. This interprets the paper’s call for batch comparisons from the perspective of developing decision criteria.

The PCA in this study describes similarities and differences among collected samples. It did not demonstrate a percentage accuracy for identifying an unknown perfume brought in for testing. Nor did it evaluate how good fragrances smelled or their safety for people.

Summary

The study showed that signal patterns produced by perfume constituents can offer clues to product identification. In particular, shifting attention from a broad range strongly affected by solvent signals to 6–8.5 ppm made formulation differences among the studied samples clearer.

The authors present comparison of NMR fingerprints as a potentially useful way to authenticate perfume. Turning that potential into reliable decisions requires understanding the range of patterns authentic products can show. Both techniques for finding differences from authentic products and data explaining differences within authentic products are foundations for establishing criteria.

References

Authors: Barbara Pacholczyk-Sienicka, Grzegorz Ciepielowski, Łukasz Albrecht
Article: The First Application of ¹H NMR Spectroscopy for the Assessment of the Authenticity of Perfumes
Journal: Molecules 26(11), 3098 (2021)
Published: May 22, 2021
DOI: 10.3390/molecules26113098
Full original article: Publisher’s article PDF
Supplementary material: Publisher’s supplementary material (ZIP)
Original article license: Creative Commons Attribution 4.0 (CC BY 4.0)