2. Destructive Techniques
2.1. Physicochemical Methods
For detection of juice fraud, physicochemical analysis is frequently used by the quality control staff in production plants, and quality parameters are assessed, such as total soluble solids (TSS), titratable acidity (TA), pH, organic acid content, and mineral content. Adulteration by water dilution and sugar addition can be detectable through measurement of TSS. More accurate detection of sugar addition can be obtained using a reducing sugar test, based on the fact that reducing sugars often account for 80 to 90% of TSS. Additionally, fruit juices adulterated with pulp wash usually have lower concentration of primary amines than authentic juice. Hence, the formol or formaldehyde test can be used to distinguish 100% juices and adulterated ones
[15]. In case of wine, some physicochemical methods can be used for disclosing adulteration, including measurement of TSS using hydrometry and/or refractometry, quantification of ethanol with hydrometry, determination of organic acid concentration by measuring TA, determination of volatile acidity using steam distillation and titration, and determination of reducing sugars using reducing sugars test with copper
[16].
Despite its low cost and broad application in quality control, the physicochemical methods exhibit limited detection of low-level adulteration. Wang et al. (2016)
[17] prepared adulterated lemon juices by adding a solution containing citric acid (5%,
w/
w) and sucrose (6%,
w/
w). Total TA, °Brix value, and pH values were measured to distinguish between the lemon juice samples. The results indicated that these traditional quality indicators of juice were not able to distinguish the pure juice and the adulterated ones
[17]. In another study by Vitalis et al. (2020)
[18], TSS was measured using a digital pocket refractometer to identify adulterants in tomato concentrate at relatively low concentration (0.5–10%). Their finding also revealed that the used method was not suitable for detection of adulteration below a certain level
[18].
Overall, physiochemical methods are simple in measurement and appropriate for initial monitoring of adulteration in production plants. However, they are not good choices for detection of sophisticated types of fraud and low adulteration levels.
2.2. Isotope Analysis
Geographical origin, variety, methods of cultivation, and production methods create the unique identities of isotope ratios in authentic juice and wine. Thus, isotopic methods based on the analysis of isotope ratios can be used for juice and wine authentication.
Isotope analysis is widely applied to detect juice adulteration with sugars, organic acids, and water. For instance, to detect adulteration in lemon and lime juices, Guyon et al. (2014)
[19] optimized an analytical protocol to determine δ
13C values of organic acids and sugars in 35 samples collected from different geographical origins using high performance liquid chromatography linked to isotope ratio mass spectrometry (HPLC-IRMS). The average δ
13C values of a mixture including citric acid, glucose, and fructose were found to be −25.40 ± 1.62‰, −23.83 ± 1.82‰ and −25.67 ± 1.72‰, respectively. These ranges of δ
13C values were then used to verify the adulteration in commercial lemon and lime juices. The results revealed that 10 in 30 commercial juice samples contained added organic acids or sugars as they had δ
13C values outside the reference ranges
[19]. Additionally, in a study by Bononi et al.
[20], the δ
13C values for organic acids and sugars in 20 genuine lemon juices derived from two regions of Italy were determined using HPLC-IRMS. After measuring these genuine samples, four isotopes were used to identify the natural range of these components in Italian lemon juice. The exogenous addition of these compounds to commercial lemon juice (42 samples) was investigated using these isotopes. Finally, the adulteration of lemon juices was detected due to more positive values of δ
13C resulting from the addition of citric acid and sugar
[20]. In a recent study by Wu et al.
[21], δ
13C and δ
18O values of 21 fruit and vegetable juices were analyzed to detect the addition of water and sugar to not-from-concentrate juice. The authors confirmed that their method was able to determine more than 20% added water and more than 7% extraneous sugar in juice samples
[21]. Nevertheless, the generalization of their results was limited because the number of studied samples was quite small.
In detection of wine adulteration, the site-specific natural isotopic fractionation with
2H nuclear magnetic resonance (SNIF-NMR) method based on isotopic ratio of deuterium/hydrogen (D/H) and with IRMS based on determination of
13C/
12C ratio of ethanol and
18O/
16O ratio of water have become the official methods for assessment of wine fraud in the European Union (EU)
[30][22]. This method has been applied in wine authentication by various authors. For example, Geana et al.
[22][23] initially investigated 23 authentic wines and found their δ
13C values in the range of −29.19‰ to −25.19‰ and their δ
18O values in the range of 0.71‰ and 4.38‰. Then, the investigation was continued on 29 commercial wines, and their authenticity was verified using the identified range of authentic wines. Among the commercial products, 16 samples were identified as ‘’good wine’’, and the remaining samples were identified as “adulterated wine’’ and ‘’suspect wine”
[22][23]. As obtained data from ‘’suspect wine” was quite close to natural range of authentic wines, more information on weather conditions at ripening and harvesting time is required to draw an accurate conclusion. Using the same analytical method but by combining with chemometrics, Wu et al.
[23][24] determined the
13C/
12C ratio of ethanol and glycerol and the
18O/
16O ratio of water in wine to discriminate 600 imported wine samples in China. Three multivariate methods, including artificial neural network (ANN), discrimination analysis (DA), and random forest (RF), were used to develop classifiers. The results showed that ANN outperformed the two remaining methods with an accuracy of up to 93.1%; RF was found unsuitable for wine origin traceability in their study
[23][24]. In another study by Wu et al.
[24][25], the δ
13C values of wine ethanol and glycerol and the δ
18O values of the wine water were analyzed in an attempt to develop a classification tool for the verification of geographic origin of 240 French red wines, using machine learning models (ANN and DA). The results showed that only ANN method with an accuracy of 98.2% was suitable for differentiating red wines
[24][25]. Obviously, using chemometrics provided researchers a better understanding about the discrimination performance of the applied method.
Although isotope analysis has high sensitivity, high accuracy, and low detection limit, its accuracy can be reduced due to the instability of isotope ratios during processing and storage of products. Additionally, high equipment price is another limitation of this technique.
2.3. Elemental Analysis
Natural juice is characterized by a certain range of mineral levels. Therefore, elemental analysis can be also applied for detection of juice adulteration based on element markers
[31][26]. In a study by Schmutzer et al.
[26][27], elemental profiles of 23 commercial orange juices were analyzed to evaluate their authenticity using inductively coupled plasma mass spectrometry (ICP-MS). Though some of the juice samples were labeled as “100% fruit juice”, the results revealed that all juices had a ratio of K to Mg of less than 50; it meant that they were adulterated with exogenous sugar
[26][27]. In addition, a combined data of elemental and isotope analysis was used by Cristea et al.
[25][28] who adopted ICP-MS to discriminate commercial and freshly squeezed apple and orange juices. The supervised classification method of linear discriminant analysis (LDA) was applied to evaluate the differences between juice samples. A satisfactory classification accuracy above 90% (validation) was obtained for both apple and orange juices. The contents of K and Na were the most important variables for discrimination of apple juice, while Na content provided the most contribution to the result. In all cases, the isotopic ratio of oxygen was the most significant variable
[25][28].
The elements in wine are derived from endogenous (grape variety and maturity and climatic conditions) and exogenous sources (external impurities from different winemaking procedures); therefore, the wine authentication can be performed based on the analysis of minerals. The effectiveness of elemental analysis was confirmed by Geana et al.
[27][29] who used ICP-MS for the analysis of elemental composition to differentiate 60 wine samples from three main wine production regions of Romania. They pointed out that the contents of five analyzed elements, including Mn, Cr, Sr, Ag, and Co, were the most useful for differentiating wines. The principal component analysis (PCA) model with three first PCs could separate the wine samples using 83% of the total variance of the acquired data
[27][29]. Their classification result could be more accurate, if supervised models were applied. In addition, a combination of ICP-MS with multivariate statistical analysis (PCA and LDA) was used by Azcarate et al.
[28][30] for differentiation of wine from different Argentinean regions. Based on the evaluated elemental profile (Ba, As, Pb, Mo, and Co), the proposed method allowed correct discrimination in terms of the geographical regions. Accordingly, the PCA result explained 95.95% of the variance of the total obtained data; and LDA model reached an accuracy higher than 96%
[28][30]. Nevertheless, their findings would be strengthened when the correlation between the element profile of soil and wine was considered. In another study, inductively coupled plasma optical emission spectrometer (ICP-OES) was adopted by Rodrigues et al.
[29][31] to distinguish 111 sparkling wines from four countries (Brazil, Argentina, France, and Spain), based on elemental profile (Al, B, Ba, Ca, Cu, Fe, K, Li, Mg, Mn, Na, and Sr). A result of 94% accurate classification was achieved by the authors, using the three key elements of B, K, and Na
[29][31].
Like the isotopic method, this technique has the advantages of high accuracy and low detection limit. However, the high costs of sample pre-treatment and the high requirement for the experimental operation are the disadvantages of elemental analysis.
2.4. Chromatographic Techniques
Chromatography is a reliable analytical approach that is suitable for the identification of adulterants in food. By utilization of chromatographic techniques, a targeted sample containing a mixture of various compounds is separated and detected. For quality control of beverages, the most used detection methods are flame ionization detector (FID) and mass spectrometry (MS)
[32]. The validation of authentication of products can be performed through determination of specific marker compounds or fingerprinting analysis. The sensitivity and high separation efficiency of chromatography in authentication have been proven through many studies; however, the technique faces some limitations such as complex procedure for sample pretreatment and operation and possible loss of instable compound
[33].
2.4.1. Gas Chromatography (GC)
Numerous studies on the application of gas chromatography (GC) for detection of fruit juice adulteration have been reported so far. For example, Yamamoto et al.
[34] used γ-terpinene and linalool as chemical markers for detecting the addition of Shiikuwasha juice to calamondin juice with gas chromatography–mass spectrometry (GC-MS), and the lowest detection level of 1% was reported. Willems and Low
[35] developed a method using capillary GC with FID for detecting the addition of pear juice in apple juice using oligosaccharide and arbutin as marker. Consequently, a low adulteration level of 0.5–3% was achieved by the proposed method
[35]. In another study by Nuncio-Jáuregui et al.
[36] on the identification of pomegranate juice adulteration with peach and grape juice, the method of headspace solid phase microextraction (HS-SPME) was optimized to extract the volatile compounds. GC-MS was then used to isolate and identify volatile profile. Based on the variation of specific volatile compounds (acetic acid, isoamyl butyrate, 1-hexanol, and linalool for added grape juice samples; butyl acetate, isobutyl butyrate, benzyl acetate, and isoamyl butyrate for added peach juice sample), the authentication of pomegranate juice was achieved with the lowest detectable levels of 10% for peach juice and 50% for grape juice
[36]. Recently, the authenticity of premium organic orange juices was confirmed by Cuevas et al.
[37] using fingerprinting analysis (HS-SPME coupled with GC-MS) combined with chemometric methods. As a result, the mid-level data fusion partial least squares-discriminant analysis (PLS-DA) model was found appropriate for authentication with the accuracy of 100%
[37].
When it comes to wine, this type of beverage contains a variety of compounds that originate from grapes, the alcoholic fermentation, and the aging of wine. Several of these compounds have effects on wine aroma and can be used as a “fingerprint” to authenticate wines. Many studies on GC technique have been conducted, focusing on differentiating wines according to geographic origin and grape variety, and detecting wine adulteration. For example, Welke et al.
[38] discriminated five types of wines with different grapes based on the volatiles obtained using a combination of HS-SPME and comprehensive two-dimensional gas chromatography with time-of-flight mass spectrometry detection (GC×GC/TOFMS). This two-dimensional GC system allowed the authors to obtain better separation capabilities compared to one-dimensional GC. Twelve extracted volatile compounds were useful to distinguish the wine samples with accuracy of 100% using LDA model
[38]. In another study, Langen et al.
[39] utilized a heart-cut multidimensional GC-MS system to determine α-ionone, β-ionone, and β-damascenone in various authentic and commercial wines. Their finding revealed that an elevated concentration of these compounds in wine samples can be served as indicators of adulteration (suggested thresholds: content of α-ionone > 0.003 μg/L, content of β-ionone > 1 μg/L, and content of β-damascenone > 10 μg/L). Moreover, the enantiomeric ratio of α-ionone could be used as an adulteration marker because the addition of exogenous α-ionone resulted in the change of this ratio
[39]. Sagandykova et al.
[40] developed a method using SPME-GC–MS for the detection of semi-volatile additives (propylene glycol and sorbic and benzoic acids) in commercial wines by optimization of SPME method. The optimized approach showed good performance in terms of linearity (coefficient of determination, R
2 > 0.98) when a method of standard addition was performed. The linear ranges for the detection of propylene glycol and sorbic and benzoic acids were 0–250 mg/L, 0–125 mg/L, and 0–250 mg/L, respectively. Using the developed method, three in twenty-five wine samples were found adulterated with propylene glycol and sorbic and benzoic acids
[40].
2.4.2. Liquid Chromatography (LC)
According to the literature, the application of LC on assessing the addition of adulterants to fruit juice can be accomplished based on the analysis of juices’ phenolic profile and anthocyanin profile. For example, Abad-García et al.
[41] analyzed polyphenols profile of citrus juices (orange, tangerine, lemon, and grapefruit juices) for the assessment of authenticity, using a reversed-phase high performance liquid chromatography (HPLC) with photodiode array detection. The acquired polyphenolic profiles were then used to develop LDA and PLS-DA classification models. As a result, the LDA model obtained 100% accuracy using four best variables (naringenin-7-
O-rutinoside-4′-
O-glucoside, naringenin-7-
O-rutinoside, hesperetin-7-
O-rutinoside, and apigenin-6,8-di-
C-glucoside). For PLS-DA model, 100% accuracy was also reported with the most significant variables of naringenin-7-
O-rutinoside-4′-
O-glucoside, apigenin-6,8-di-
C-glucoside, isosakuranetin-7-
O-rutinoside, and naringenin-7-
O-rutinoside
[41]. Phloridzin, a phenolic compound, is not naturally available in grapes; while this compound is present in apples in a higher proportion compared to other fruits. Based on this fact, Spinelli et al.
[42] used HPLC with a photodiode array detector to analyze phloridzin for detecting the addition of apple juice to purple grape juice. Their results indicated that the proposed method allowed the detection of apple juices in adulterated grape juices with phloridzin content of 4.54 to 8.39 mg/L
[42].
Like in juice, phenolic compounds are also found in grapes as well as in must and wine. Since the compositions of polyphenolic compounds in wine vary widely, depending on the grape varieties, the winemaking process, and the climatic conditions; they can be used as markers for authenticating wine
[46][43]. For instance, HPLC method based on analysis of anthocyanin profile was utilized to discriminate different wines (Brazilian tropical wines, Brazilian temperate wines, and temperate Chilean wines) by de Andrade et al.
[43][44]. As a result, the concentrations of nine anthocyanins were determined and used as the classification factor for PCA. The PCA results showed that the two first PCs accounted for 88% of the total variance. The contents of petunidin-3-glucoide, peonidin-3-glucoside, and malvidin-3-glucoside contributed the most to PC1; while the contents of peonidin-3-glucoside coumarate, peonidin-3-glucoside-acetate, and malvidin-3-glucoside primarily represented PC2
[43][44]. In another study, Pavloušek et al.
[44][45] succeeded in classifying 43 different wines with a HPLC method for analyzing non-flavonoid phenolic compounds. Results indicated that the canonical discriminant analysis allowed them to classify the wine with 95.4% correction rate
[44][45]. Similarly, Geana et al.
[22][23] applied a HPLC system to characterize the anthocyanin profile for authenticating wine. An LDA model was developed to differentiate the wine samples with respect to concentrations of anthocyanins and defined anthocyanins ratios. Differentiation results of LDA achieved 100% accuracy using two discriminant factors. The most significant anthocyanins for discrimination included individual anthocyanins of delphinidin-3-
O-glucoside, petunidin-3-
O-glucoside, peonidin-3-
O-glucoside, malvidin-3-
O-glucoside, and peonidin-3-
O-(6-p-coumaroyl) glucoside
[22][23]. Recently, Zhi et al.
[45][46] developed an analytical method that combined three-way HPLC with diode array detection and chemometrics (PCA-LDA) to distinguish wines according to their vintages. The proposed method could prevent the loss of analytes and thus increase the accuracy of analysis. As a result, an discrimination rate above 90% was achieved
[45][46].
The differences in polyphenolic profile of juices are sometimes not caused by adulteration but by climatic conditions, the environment, the processing technology, and the degree of fruit ripeness. Furthermore, the oxygen stability of anthocyanins and betacyanins can significantly change due to the activities of native polyphenol oxidases when these compounds are used as markers
[31][26]. Therefore, these factors need to be taken into account in the analysis of polyphenols profile.
2.5. DNA-Based Techniques
Since DNAs of food products maintain their stability under conditions of environment and cultivation and production process, DNA-based methods have become reliable means for food authentication
[47]. DNA polymorphisms resulting from natural mutations of genetic code can be used to identify plant species. In DNA-based techniques, the DNA fragment of interest is extracted from objective samples, and then the specific genetic polymorphisms are amplified to obtain the amplicons, followed by an analysis of the obtained amplicons to reveal the characteristics of polymorphisms
[48]. For authentication and detection of food adulteration, the most commonly used DNA-based methods are polymerase chain reaction (PCR), real-time PCR, high resolution melting (HRM) analysis, microarrays, and next generation sequencing
[49].
In the study of juice adulteration, orange juices have gained the most interest from researchers due to their market value. Most of the studies focused on disclosure of juice-to-juice adulteration. For instance, a PCR restriction fragment length polymorphism (RFLP) assay and a PCR heteroduplex assay allowed the detection of grapefruit and mandarin juice in orange juice. Specifically, the PCR heteroduplex assay showed a better limit of detection of 2.5%, while a limit of detection of 10% resulted from using the RFLP assay
[50]. Additionally, the detection of adulterated orange juice with mandarin juice performed by Aldeguer et al.
[51], using a single nucleotide polymorphism (SNP) at the trnL–trnF intergenic region of the chloroplast chromosome as marker. As a result, a limit of detection of 5% added mandarin was achieved in both fresh and reconstituted orange juices
[51]. A similar study was conducted by Pardo and Miguel Angel
[52] who successfully determined the addition of mandarin in orange juice with the detection limit of 1%.
In the case of wine adulteration, DNA analysis is the most applied approach for identifying the varietal origin of grapes for wine production. In grapevine varietal identification, OIV has approved the use of simple sequence repeat (SSR) as nuclear molecular markers. However, SSR markers have revealed some limitation in terms of DNA quality relating to the possible inhibition of PCR reactions by large amount of polyphenols, polysaccharides, and proteins in must and wine and the degradation of DNA by alcoholic fermentation
[53]. Recently, small molecular markers like SNP have been used to deal with DNA degradation. For instance, Boccacci et al.
[54] used SNP genotyping assays to authenticate varietal origin of “Nebbiolo” musts and wines. Based on 1157 genotypes, two SNPs were sufficiently adopted for authentication. The developed assays allowed the authors to identify the must mixtures and wine mixtures at the sensitivity of 1% and 10–20%, respectively
[54].
Overall, the advantages of DNA-based techniques are their sensitivity, quick analysis time, and convenience for large-scale measurement. Additionally, these approaches are also not affected by geography. However, the application of these techniques faces several problems, such as the degradation of DNA in fruit juices due to thermal processing under acid conditions and the removal of pulp in the clarification process of clarified juice leading to the difficulty in collecting DNA. Moreover, the heavy workload, the complex procedure for selecting molecular markers, and the high cost are other problems that must be taken into account
[55].