Food fraud and determination of food authenticity have been topics of utmost importance for the food industry, causing major concerns to consumers and to all stakeholders involved in the food production and food industry ^[1]. One of the most common types of food fraud is the intentional and economically motivated adulteration of foods, which can include the replacement of a high-value nutrient with a low-value one, the dilution of a high-value liquid ingredient with a cheaper liquid, the addition of foreign materials in order to enhance food quality or organoleptic attributes, or hiding the low quality of food ingredients or products ^[1]. The adulteration of food products not only compromises its authenticity and reduces product quality, but also may seriously jeopardize consumer’s health. Although all foods can be susceptible of potential adulteration, there are certain products that are considered more prone to fraudulent acts, particularly if they are produced and supplied in complex market chains, if they are considered as luxurious food commodities, or on the other hand, if they have low margins of profit. These products include spices, honey, olive oil, wine, milk, and dairy products, coffee, and tea, among others ^[2][3].

Nowadays, several physicochemical and instrumental analytical methodologies exist to detect food adulteration and determine its authenticity. These techniques include target analyses that identify specific compounds and assess if they are within the limit stated on the label or established by law and non-target analyses that afford a fingerprint of the whole sample, which can then be used to evaluate its authenticity or detect a possible adulteration ^[1][4][5]. The most commonly used techniques are physicochemical analytical methodologies, gas and liquid chromatography frequently hyphenated with mass spectrometry, immunochemical and DNA-based techniques, isotope ratio, elemental analysis, and spectroscopic techniques that include UV-Visible, infrared, Raman, and nuclear magnetic spectroscopy (NMR) ^[4][5]. All of these techniques have their own pros and cons. For example, classical physicochemical methods are time-consuming, some of them involve many analytical steps and require large volumes of organic solvents that generate toxic wastes. Although very selective and specific, the modern chromatographic methodologies, NMR, and mass spectrometry require the acquisition and maintenance of very expensive instrumentation and highly qualified laboratory technicians, and frequently also involve the need of very complex sample pre-treatment,  being less suitable for routine or large-scale analyses ^[5]. Furthermore, as most adulterants are unknown, it is very unlikely that they could be detected using target methodologies ^[4]. In the last years, with the urgent need to develop new tools, researcher’s attention has been focused on the application of untargeted methodologies to food analyses. Untargeted methods such as spectroscopic (IR, Raman, NMR), hyperspectral imaging, and chromatographic techniques (GC-MS and HPLC-MS) provide a molecular fingerprint of the whole food matrix.

In particular, mid-infrared spectroscopy has emerged as a potential analytical tool and has been considered an alternative to other more expensive and complex methods, being suitable for implementation in factories during the production process as well as in quality control laboratories. Since it is considered a high-throughput approach, some advantages of this non-destructive methodology include the speed of analyses, simple or no-sample preparation, no waste generation, fast acquisition of spectra, the possible detection of unexpected adulterants or unexpected deviations to the reference samples ^[6].

However, food is a very complex matrix that contains a high number of components, giving rise to a multitude of spectral information and large data sets. Therefore, it is mandatory to employ chemometric tools that allow the extraction of relevant information and the conception of models that could be used to perform exploratory studies and define important features of samples or predict analyses on new samples ^[7][8]. The classic univariate statistic methods, such as analysis of variance (ANOVA), are focused on the reductionist approach (e.g., one variable at times), and food compounds or properties are analyzed independently of the entire food matrix. Conversely, modern chemometric methods are multivariate analyses that allow the treatment of multidimensional and complex data sets, revealing properties that are important through their various interferences and interactions in the whole food matrix ^[8][9]. Multivariate methods can be applied either to qualitative and quantitative analysis, letting relevant information be extracted from complex data, allowing the creation of empirical models that could be used to perform exploratory studies, and describing important characteristics of samples or predictive analyses on new samples ^[7].

Several academic research groups worldwide clearly have been proved that the MIR spectroscopy associated with attenuated total reflection acquisition mode and different chemometric tools could be broadly applied to address quality, authenticity, and adulteration issues ^{[3][10][11][12][13]}. Nevertheless, the majority of these studies are purely academic. In order to be further applied as a potential analytical tool in official food control procedures, several challenges must be overcome ^[4][14]. The implementation of validated standard methods is of utmost importance aiming at assuring a high level of reproducibility across different laboratories, equipment, and analyses, allowing data to be comparable, which is mandatory in official food control procedures^[14]. In this context, validation guidelines are urgently needed to standardize all the steps of method development in MIR spectroscopy. Choosing an adequate source of samples is also essential for method development. The origin of samples should be known, preferably from reputable producers rather than from commercial outlets or markets. Certified reference materials should be used whenever possible ^[4]. For both method development and validation, it is also imperative to collect a sufficient number of representative samples in order to cover all sample variations and develop a robust model that could be further employed for legal and regulatory purposes. Furthermore, the choice of the most appropriate chemometric methodology is another issue that must be addressed. A plethora of multivariate analyses has been employed for processing the vast collection of spectral data. Some of these software tools are expensive and data analysis requires time for the complex statistical treatments and specific abilities to interpret the results.

In conclusion, MIR spectroscopy is a rapid and valuable tool that could be useful as a preliminary screening of a commodity, particularly when fast and robust methods are necessary in order to accept or reject a product or a batch before its introduction in a food chain. Nevertheless, further confirmatory analysis may be required using already validated target methods that accomplished the current legislation for official food control.