Can Color-Changing Films Tell Perfumes Apart?

A thin film changes color when exposed to compounds rising from a perfume. Could photographing that color with a smartphone help distinguish perfumes? In 2024, Kuo Zhan, Anemar Bruno Kanj, and Lars Heinke combined three types of specialized film to classify vapors from 16 commercial products, including perfumes.

The researchers sought a reusable optical sensor that could distinguish even molecules related as mirror images. They examined whether an ordinary camera could read its responses to complex mixtures such as perfumes.

Molecules that do not overlap, like right and left hands

Some molecules contain the same combination of atoms but are mirror images that cannot be superimposed, like right and left hands. This property is called chirality. Fragrance materials include such molecules, and the two mirror-image forms can sometimes smell different.

The sensor was therefore given a structure in which uptake differs depending on a molecule’s orientation. The materials were thin films of metal–organic frameworks (MOFs), in which metals and organic molecules connect to form small pores.

The study placed two types of chiral film alongside one achiral film. Combining films responsive to this distinction with a film lacking chirality helps reveal response differences that would be difficult to detect with a single film.

Why does taking up molecules change the color?

The film sits between a gold-based reflecting surface below and a thin platinum reflecting surface above. Light reflected from the two surfaces overlaps, producing a particular color through interference. Just as with the colors of soap bubbles, the way light overlaps affects the visible color.

When vapor molecules enter the film’s pores, they change its refractive index, which describes how light travels through it, and the color of the reflected light changes too. This is not a mechanism in which a dye reacts with perfume to form a different pigment. The aim is for the film to return to its original state as molecules leave the pores.

The three films respond differently even to the same vapor. Recording each film’s color as three values, red, green, and blue, gives nine numbers across the three films. This combination was used as the signal for distinguishing samples.

First tested with a pair of mirror-image molecules

The first substances tested were the R and S forms of 2-octanol. Measurements with a dedicated optical analysis instrument showed that the chiral films differed in the magnitude of their color changes in response to the two forms. The achiral film, by contrast, responded to a similar degree to both.

Repeatedly exposing the films to vapor and then removing it confirmed that the response reversed and reappeared. This tested their behavior as reversible sensors based on molecular uptake and release.

Next, the researchers photographed the three films with an ordinary smartphone. When the resulting color values were fed into a machine-learning method that classifies similar response patterns, classification accuracy in this experiment with the R and S forms was 99.0%. This result concerned one pair of pure substances.

How well did the sensor distinguish 16 commercial products?

The researchers then examined 16 commercial products, including perfumes, eaux de toilette, and aftershaves. Vapors from the products were passed into a transparent sealed container holding the films, whose colors were photographed by a smartphone in a fixed position. The phone’s camera alone was not directly photographing odors in the air.

For each condition, color values were extracted from ten consecutive photographs. The data were divided into five parts for cross-validation, using some for training and the remainder for evaluation. Color patterns differed between perfume vapors, and the reported classification performance for the product group was 98.9%.

However, consecutive photographs do not each represent an independent perfume bottle or a measurement on a different day. The paper’s total counts of ‘data points’ also leave some ambiguity about how the multiple color values obtained from photographs were counted. This performance cannot be interpreted as accuracy validated on many independent product samples or as an identification rate in other environments.

Further work is needed to determine whether the same distinctions hold across different manufacturing batches of the same brand, changes after opening, different camera models, or different lighting.

Matching three samples with their names concealed

The researchers measured three samples selected from the 16 products without the experimenter knowing their product names. The nine color values were summarized in a two-dimensional plot and compared with the positions of previously measured products. All three were correctly matched.

This test asked which of the registered products a sample corresponded to. It did not decode the formula of a completely unknown perfume or identify its individual fragrance ingredients. Indeed, the authors state that they did not determine the molecular compositions or mixture ratios of the products used.

The paper also states that the authors’ untrained noses could not classify all the products. However, this was not a comparison with a formal sensory evaluation using standardized participant numbers and procedures, so it cannot establish that the sensor outperforms human smell in general.

Simple readout, but manufacturing challenges remain

An advantage of the study is that film responses could be quantified with an everyday camera, not just expensive analytical equipment. Adding films with different pore structures might also provide responses to a wider variety of compounds and mixtures.

The authors suggest possible future uses in detecting perfume mix-ups, counterfeits, and diluted products. However, this study did not test detection performance using actual counterfeits or perfumes with systematically varied concentrations.

Simple readout does not necessarily mean that manufacturing the entire sensor has already become inexpensive. The authors identify a need to improve reliable large-area production of chiral films and reduce the costs of the gold-based substrates and platinum top mirrors. Practical use still requires work on the material-production process as well.

The research received support from the German Research Foundation, China’s Fundamental Research Funds for the Central Universities, and the China Scholarship Council, among others. The authors declare no conflict of interest.

Summary

Three types of film whose interference colors change when they take up molecules responded differently to a pair of mirror-image molecules. Colors photographed with a smartphone enabled classification of vapors from 16 commercial products. Three samples with concealed names were also successfully matched within the known product group.

The important step is the link between materials that selectively take up molecules and optical changes that are easy to read. Expanding the range of reusable films and refining manufacturing and measurement conditions could develop this approach into a way of distinguishing complex scents for product quality management.

References

Zhan, K., Kanj, A. B., & Heinke, L. (2024). Classification and Identification of Perfume Scents by an Enantioselective Colorimetric Sensor Array of Chiral Metal-Organic-Framework-Based Fabry-Pérot Films. Chemistry – A European Journal, 30(39), e202400798. DOI: 10.1002/chem.202400798. Full original article. © 2024 The Authors. Chemistry – A European Journal published by Wiley-VCH GmbH. The original article is published under CC BY-NC 4.0.