Determination of geographic origin of caffeine by stable isotope analysis

Support vector regression for prediction of stable isotopes and trace elements using hyperspectral imaging on coffee for origin verification

The potential of using rapid and non-destructive near-infrared - hyperspectral imaging (HSI-NIR) for the prediction of an integrated stable isotope and multi-element dataset was explored for the first time with the help of support vector regression. Speciality green coffee beans sourced from three continents, eight countries, and 22 regions were analysed using a push-broom HSI-NIR (700-1700 nm), together with five isotope ratios (δ13C, δ15N, δ18O, δ2H, and δ34S) and 41 trace elements. Support vector regression with the radial basis function kernel was conducted using X as the HSI-NIR data and Y as the geochemistry markers. Model performance was evaluated using root mean squared error, coefficient of determination, and mean absolute error. Three isotope ratios (δ18O, δ2H, and δ34S) and eight elements (Zn, Mn, Ni, Mo, Cs, Co, Cd, and La) had an R2predicted 0.70 - 0.99 across all origin scales (continent, country, region). All five isotope ratios were well predicted at the country and regional levels. The wavelength regions contributing the most towards each prediction model were highlighted, including a discussion of the correlations across all geochemical parameters. This study demonstrates the feasibility of using HSI-NIR as a rapid and non-destructive method to estimate traditional geochemistry parameters, some of which are origin-discriminating variables related to altitude, temperature, and rainfall differences across origins.

Stable isotope and trace element analyses with non-linear machine-learning data analysis improved coffee origin classification and marker selection

BACKGROUND

This study investigated the geographical origin classification of green coffee beans from continental to country and regional levels. An innovative approach combined stable isotope and trace element analyses with non-linear machine learning data analysis to improve coffee origin classification and marker selection. Specialty green coffee beans sourced from three continents, eight countries, and 22 regions were analyzed by measuring five isotope ratios (δ¹³ C, δ¹⁵ N, δ¹⁸ O, δ² H, and δ³⁴ S) and 41 trace elements. Partial least squares discriminant analysis (PLS-DA) was applied to the integrated dataset for origin classification.

RESULTS

Origins were predicted well at the country level and showed promise at the regional level, with discriminating marker selection at all levels. However, PLS-DA predicted origin poorly at the continental and Central American regional levels. Non-linear machine learning techniques improved predictions and enabled the identification of a higher number of origin markers, and those that were identified were more relevant. The best predictive accuracy was found using ensemble decision trees, random forest and extreme gradient boost, with accuracies of up to 0.94 and 0.89 for continental and Central American regional models, respectively.

CONCLUSION

The potential for advanced machine learning models to improve origin classification and the identification of relevant origin markers was demonstrated. The decision-tree-based models were superior with their embedded variable identification features and visual interpretation. © 2023 The Authors. Journal of The Science of Food and Agriculture published by John Wiley & Sons Ltd on behalf of Society of Chemical Industry.

Near-infrared reflectance spectroscopy accurately predicted isotope and elemental compositions for origin traceability of coffee

Stable isotope ratios and trace elements are well-established tools that act as signatures of the product's environmental conditions and agricultural processes; but they involve time, money, and environmentally destructive chemicals. In this study, we tested for the first time the potential of near-infrared reflectance spectroscopy (NIR) to estimate/predict isotope and elemental compositions for the origin verification of coffee. Green coffee samples from two continents, 4 countries, and 10 regions were analysed for five isotope ratios (δ¹³C, δ¹⁵N, δ¹⁸O, δ²H, and δ³⁴S) and 41 trace elements. NIR (1100-2400 nm) calibrations were developed using pre-processing with extended multiplicative scatter correction (EMSC) and mean centering and partial-least squares regression (PLS-R). Five elements (Mn, Mo, Rb, B, La) and three isotope ratios (δ¹³C, δ¹⁸O, δ²H) were moderately to well predicted by NIR (R²: 0.69 to 0.93). NIR indirectly measured these parameters by association with organic compounds in coffee. These parameters were related to altitude, temperature and rainfall differences across countries and regions and were previously found to be origin discriminators for coffee.

A review of the newly identified impurity profiles in methamphetamine seizures

Forensic intelligence of synthetic illicit drugs suffers a problem of continuous introduction of new synthetic methods, modification of the existing routes of manufacture, and adulterations practiced by criminal networks. Impurity profiling has been indispensable in methamphetamine intelligence based on precursors, synthetic routes, and chemical modifications during trafficking. Law enforcement authorities maintain the credibility and integrity of intelligence information through constant monitoring of the chemical signatures in the illicit drug market.

Changes in the synthetic pattern result in new impurity profiles that are important in keeping valuable intelligence information on clandestine laboratories, new synthetic routes, trafficking patterns, and geographical sources of illicit Methamphetamine.

This review presents a critical analysis of the methamphetamine impurity profiles and more specifically, profiling based on impurity profiles from Leuckart, Reductive amination, Moscow, Emde, Nagai, Birch, Moscow route; a recent nitrostyrene route and stable isotope signatures. It also highlights the discrimination of ephedrine from pseudoephedrine sources and the emerging methamphetamine profiling based on stable isotopes.

Isotope-abundance variations and atomic weights of selected elements: 2016 (IUPAC Technical Report)

There are 63 chemical elements that have two or more isotopes that are used to determine their standard atomic weights. The isotopic abundances and atomic weights of these elements can vary in normal materials due to physical and chemical fractionation processes (not due to radioactive decay). These variations are well known for 12 elements (hydrogen, lithium, boron, carbon, nitrogen, oxygen, magnesium, silicon, sulfur, chlorine, bromine, and thallium), and the standard atomic weight of each of these elements is given by IUPAC as an interval with lower and upper bounds. Graphical plots of selected materials and compounds of each of these elements have been published previously. Herein and at the URL http://dx.doi.org/10.5066/F7GF0RN2, we provide isotopic abundances, isotope-delta values, and atomic weights for each of the upper and lower bounds of these materials and compounds.

A novel high-temperature combustion interface for compound-specific stable isotope analysis of carbon and nitrogen via high-performance liquid chromatography/isotope ratio mass spectrometry

RATIONALE

In aqueous samples compound-specific stable isotope analysis (CSIA) plays an important role. No direct method (without sample preparation) for stable nitrogen isotope analysis (δ(15) N SIA) of non-volatile compounds is known yet. The development of a novel HPLC/IRMS interface based on high-temperature combustion (HTC) for both δ(13) C and δ(15) N CSIA and its proof of principle are described in this study.

METHODS

To hyphenate high-performance liquid chromatography (HPLC) with isotope ratio mass spectrometry (IRMS) a modified high-temperature combustion total organic carbon analyzer (HTC TOC) was used. A system to handle a continuously large amount of water (three-step drying system), favorable carrier and reaction gas mix and flow, an efficient high-temperature-based oxidation and subsequent reduction system and a collimated beam transfer system were the main requirements to achieve the necessary performance.

RESULTS

The proof of principle with caffeine solutions of the system succeeded. In this initial testing, both δ(13) C and δ(15) N values of tested compounds were determined with precision and trueness of ≤0.5 ‰. Further tests resulted in lower working limit values of 3.5 μgC for δ(13) C SIA and 20 μgN for δ(15) N SIA, considering an accuracy of ±0.5 ‰ as acceptable.

CONCLUSIONS

The development of a novel HPLC/IRMS interface resulted in the first system reported to be suitable for both δ(13) C and δ(15) N direct CSIA of non-volatile compounds. This highly efficient system will probably open up new possibilities in SIA-based research fields. Copyright © 2016 John Wiley & Sons, Ltd.

Development of a methodology using gas chromatography-combustion-isotope ratio mass spectrometry for the determination of the carbon isotope ratio of caffeine extracted from tea leaves (Camellia sinensis)

RATIONALE

Compound-specific isotope analysis (CSIA) of the extracted caffeine can be used to determine the authenticity of the origin of tea. Elemental analysis-isotope ratio mass spectrometry (EA-IRMS), which is widely used to measure the carbon isotope ratio of caffeine, has a strict requirement for the purity of the extracted caffeine. To obtain high-purity caffeine from tea leaves, the conventional extraction process has to be repeated and usually takes about 5-6 h. To improve the measurement of the carbon isotope ratio of caffeine, a more rapid and accurate measuring method is needed.

METHODS

An analytical protocol was developed for the determination of the carbon isotope ratio of caffeine from tea leaves using gas chromatography-combustion-isotope ratio mass spectrometry (GC-C-IRMS) combined with our extraction process. The procedure to extract caffeine and determine its carbon isotope ratio takes around 1.5 h.

RESULTS

The standard deviation of the method is less than 0.1‰ (1σ). The measured carbon isotope ratios were not influenced by the amount of caffeine injected (0.08-0.62 µg) or by the extraction yield of caffeine from the tea leaves.

CONCLUSIONS

The carbon isotope ratios of caffeine from eight tea cultivars were determined using the protocol.

Caffeine in your drink: natural or synthetic?

Owing to possible adulteration and health concerns, it is important to discriminate between natural and synthetic food ingredients. A new method for compound-specific isotope analysis (CSIA) by coupling high-temperature reversed-phase liquid chromatography to isotope ratio mass spectrometry (HT-RPLC/IRMS) was developed for discrimination of natural and synthetic caffeine contained in all types of drinks. The analytical parameters such as stationary phase, column inner diameter, and column temperature were optimized for the separation of caffeine directly from drinks (without extraction). On the basis of the carbon isotope analysis of 42 natural caffeine samples including coffee beans, tea leaves, guaraná powder, and maté leaves, and 20 synthetic caffeine samples from different sources by high-temperature reversed-phase liquid chromatography coupled to isotope ratio mass spectrometry, it is concluded that there are two distinguishable groups of caffeine δ(13)C-values: one between -25 and -32‰ for natural caffeine, and the other between -33 and -38‰ for synthetic caffeine. Isotope analysis by HT-RPLC/IRMS has been applied to identify the caffeine source in 38 drinks. Four mislabeled products were detected due to added but nonlabeled synthetic caffeine with δ(13)C-values lower than -33‰. This work is the first application of HT-RPLC/IRMS to real-world food samples, which showed several advantages: simple sample preparation (only dilution), high throughput, long-term column stability, and high precision of δ(13)C-value. Thus, HT-RPLC/IRMS can be a very promising tool in stable isotope analysis of nonvolatile compounds.

Determination of the geographical origin of green coffee by principal component analysis of carbon, nitrogen and boron stable isotope ratios

In this study we show that the continental origin of coffee can be inferred on the basis of coupling the isotope ratios of several elements determined in green beans. The combination of the isotopic fingerprints of carbon, nitrogen and boron, used as integrated proxies for environmental conditions and agricultural practices, allows discrimination among the three continental areas producing coffee (Africa, Asia and America). In these continents there are countries producing 'specialty coffees', highly rated on the market that are sometimes mislabeled further on along the export-sale chain or mixed with cheaper coffees produced in other regions. By means of principal component analysis we were successful in identifying the continental origin of 88% of the samples analyzed. An intra-continent discrimination has not been possible at this stage of the study, but is planned in future work. Nonetheless, the approach using stable isotope ratios seems quite promising, and future development of this research is also discussed.

18O pattern and biosynthesis of natural plant products

Oxygen atoms in plant products originate from CO(2), H(2)O and O(2), precursors with quite different delta18O values. Furthermore their incorporation by different reactions implies isotope effects. On this base the resulting non-statistical 18O distributions in natural compounds are discussed. The delta18O value of cellulose is correlated to that of the leaf water, and the observed 18O enrichment (approximately +27 per thousand) is generally attributed to an equilibrium isotope effect between carbonyl groups and water. However, as soluble and heterotrophically synthesised carbohydrates show other correlations, a non-statistical 18O distribution - originating from individual biosynthetic reactions - is postulated for carbohydrates. Similarly, the delta18O values of organic acids, carbonyl compounds, alcohols and esters indicate water-correlated, but individual 18O abundances (e.g. O from acyl groups approximately +19% above water), depending upon origin and biosyntheses. Alcoholic groups introduced by monooxygenase reactions, e.g. in sterols and phenols, show delta18O values near +5 per thousand, in agreement with an assumed isotope fractionation factor of approximately 1.02 on the reaction with atmospheric oxygen (delta18O=+23.5 per thousand). Correspondingly, a "thermodynamically ordered isotope distribution" is only observed for oxygen in some functional groups correlated to an origin from CO(2) and H(2)O, not from O(2). The individual isotopic increments of functional groups permit the prediction of global delta18O values of natural compounds on the basis of their biosynthesis.