Through-bottle spectroscopy as a tool for quality control and anti-counterfeiting of Brandy and Cognac

Abstract

Counterfeiting of premium spirits poses significant economic and health risks, that could be tackled by robust, accurate, portable and non-destructive through-bottle measurements. Here, we demonstrate the capability of focus-matched inverse spatially offset spectroscopy, combining fluorescence and Raman signals, for authenticating Cognac and Brandy. The technique accurately identifies age classification, bottling year, spoilage due to elevated storage temperatures, and distinguishes between Cognac brands. Critically, our method effectively differentiates genuine Cognacs from counterfeit products, correctly identifying 98% of counterfeit samples. This shows the promise of through-bottle spectroscopy as a powerful tool for supply chain integrity and consumer protection in the high-value spirits market.

Figures (8)

• Figure 1: Comparison of the experimental Removal of Glass Signal between (a) FM-iSORS and (b) conventional iSORS. Spectra obtained from a sealed bottle of Cognac, where each spectrum is acquired for a different bottle position. The legend for each panel shows the separation along the optical axis between the bottle surface, and the focal point of lens L$_{2}$ in Figure \ref{['setup']}. In FM-iSORS, increasing this separation has the effect of suppressing the signal due to glass which forms a peak centered at 1380 cm${^{-1}}$. The Raman shift region corresponding to the dominant Raman peaks of ethanol and glass are highlighted in purple and green respectively. The ratio of glass-to-ethanol signal $G_{p}/E_{p}$ for FM-iSORS shows a decrease with increased focal distance, while the signal measured due to Raman scattering from the ethanol remains constant. In contrast, the ethanol peak is significantly weaker and does not substantially change with bottle position for iSORS, while the glass peak only reduces by $\sim$ 15% as the bottle moves away from the collection lens.
• Figure 2: Spectral analysis of VS, VSOP and XO age classifications from a single producer. (a) The average spectrum (solid line) and standard error (shaded area) of each age classification shows visible differences in the broad fluorescence profile. (b) Principal component loadings of the first four principal components of the spectra. The variance captured by each principal component is included within the relevant panel. Red shaded regions show the positions of the two dominant Raman peaks of ethanol at 880 cm$^{-1}$ and at 1460 cm $^{-1}$.
• Figure 3: Identification of age classification is unaffected by bottle substitution. In (a), we present a box plot of PC1 values for the different age classifications of Cognac (VS, VSOP and XO), showing how they can be discriminated by fluorescence alone. Each point represents the PC1 score for a single measurement. In (b), we show box plots comparing the PC1 scores obtained from the original test blends with other bottles refilled with VS blend. Data are presented as PC1 score values. Red data and boxes denote low end VS bottles while blue and black boxes represent VSOP and XO respectively. Each box represents the 25% to 75% range. Solid red boxes represent data from substituted bottles of different glass types (GT). Data were analysed with PCA and one-way ANOVA with Tukey's multiple comparison test. Asterisks indicate statistical significance between treatment groups. ****P<0.0001 ns: no significance.
• Figure 4: Clustering of spectra from different production years. Principal component score plots for VS bottles show clustering based on the year of production. (a) PC1-PC2 scatter plots for individual VS bottle measurements shows overlapping clusters. (b) However, averaged measurements of each bottle shows separation by year along the PC2 axis. Data in (b) are presented as mean $\pm$ SEM. Colours denote the year of production, while symbol shapes denote different bottles.
• Figure 5: Cognac stored at higher temperatures shows spectral changes. (a) Through-bottle spectra of VS and (b) principal component analysis for VS, VSOP and XO Cognac stored at different temperatures: 20° C (green), 25° C (red) and 35° C (magenta). All data were recorded with 100 mW power and 2s integration time. Inset in (a) shows the region around the dominant 880 cm${^{-1}}$ ethanol peak highlighting the strong overlap between samples stored at acceptable temperatures. In (b), VS, VSOP and XO are represented in red, blue and black respectively. Data were analysed with PCA and one-way ANOVA with Tukey's multiple comparison test. Asterisks indicate statistical significance between treatment groups. ****P<0.0001 ns: no significance. Cognac stored at 35° C is consistently shown to have a statistically significant shift in the PC2 value.