Can Rare Earth Elements Be Considered as Markers of the Varietal and Geographical Origin of Wines?

The possibility of establishing the varietal and territorial affiliation of wines by the content of rare earth elements (REE) in them was studied. ICP–OES and ICP–MS with subsequent chemometric processing of the results were applied to determine the elemental image of soils containing negligible REE amounts, grapes grown on these soils, and wine materials of Cabernet Sauvignon, Merlot, and Moldova varieties produced from these grapes. To stabilize and clarify wine materials, the traditional processing of wine materials with various types of bentonite clays (BT) was used, which turned out to be a source of REE in the wine material. Discriminant analysis revealed that the processed wine materials were homogeneous within one denomination and that those of different denominations were heterogeneous with respect to the content of REE. It was found that REE in wine materials were transferred from BT during the processing, and thus they can poorly characterize the geographical origin and varietal affiliation of wines. Analysis of these wine materials according to the intrinsic concentrations of macro- and microelements showed that they formed clusters according to their varietal affiliation. In terms of their influence on the varietal image of wine materials, REE are significantly inferior to macro- and microelements, but they enhance their influence to a certain extent when used together.


Introduction
The confirmation of the quality and authenticity of dry wines and the control of the processes of their blending require the use of new analytical approaches allowing the consideration of wines as a whole based on the analysis of their "images". In this case, the identification of the origin of wines is controlled by the characteristic components or elemental profiles (electrophoretic, chromatographic, and elemental) inherent to the studied sample of wine [1][2][3][4]. Due to the specificity of the obtained profiles, this approach makes it possible to determine the quality and affiliation of wines, to detect counterfeit products, and to identify wines by variety and origin [5][6][7].
In fact, the mineral composition is one of the main indicators of the classification of the varietal and geographical affiliation of wines due to the stability of the elemental composition of soils of grapes growing [8,9]. The relationship between the mineral composition in the soil-grape-wine chain makes it possible to establish the varietal and geographical origin of wine with high reliability due to the stability and specificity of the elemental composition of the grape variety cultivated in a particular area [8][9][10].

Results and Discussion
A comparison of the elemental compositions of BT (Table 1) and tested wine materials (Table 2) proves that REE are introduced in the wine materials during the processing of rare materials by BT. The levels of almost all REE in the studied untreated wine materials were below the limits of quantification (LOQ), despite the high sensitivity of the spectrometer [32] ( Table 2), although their contents in the final products were reliably recorded. It seemed important to determine the sources of REE entry into the analyzed wines and to establish a possible correlation of their contents with the quality of soils in the grape-growing region. The presence of Y, La, and Ce in untreated wine materials (Table 2) may be associated with anthropogenic impact, namely the introduction of soil improvers, fertilizers, pesticides, air pollution, or irrigation water [27,33]. The total content of REE in the BT samples combined by groups differs ( Table 1). The average minimum content of the total REE is observed in the BT samples of the fourth (7.93 µg/g), then the second (10.03 µg/g), the first (11.83 µg/g), and the third (12.32 µg/g) groups. BT significantly increases the content of REE in wine materials and there is a multiple increase in the concentrations of Ce (4-6), Y (2)(3)(4)(5)(6)(7)(8)(9)(10)(11)(12)(13)(14)(15)(16)(17)(18)(19), and La (2-24) relative to their initial content in the wine materials, probably due to ion exchange ( Table 2). The concentrations of Tm and Lu in the wine material of the Cabernet Sauvignon variety, Pr, Ho, Tm, and Lu in Merlot and Pr, Tb, Ho, and Tm in Moldova change the least. Notably, the greatest contribution of REE from BT to wine material is observed for the Cabernet Sauvignon variety, followed by Merlot and Moldova ( Figure 1). The general nature of the increase in the content of REE in wine materials is quite similar-the greatest contribution to the increase in the content of REE in wine materials of all varieties is made by BT of the third (BT13, BT20, BT21, BT22, and BT27) and the first (BT19, BT23, and BT32) groups, as well as BT15 from the second group, the smallest-BT1 (Figure 1).  Table 3 shows the total contents of REE and trace elements (Ag, Al, As, Ba, Be, Bi, Cd, Co, Cs, Cu, Fe, Gd, Ge, Ho, Li, Mo, Mn, Nb, Ni, Pb, Ru, Sb, Sn, Sr, Ta, Th, Ti, Tl, U, V, W, Y, Zn, and Zr) after processing wine materials by BT of various groups. More detailed data on the effect of stabilization and clarification by BT on the elemental image of various varieties of wine materials are given in the supplementary material to the article. There is a notable fact of an increase in the concentrations of REE and microelements in wine materials after treatment by BT of all groups. In addition, the introduction of BT of the same group during the clarification of different varieties of wine materials affects their elemental image in different ways.  Table 3 shows the total contents of REE and trace elements (Ag, Al, As, Ba, Be, Bi, Cd, Co, Cs, Cu, Fe, Gd, Ge, Ho, Li, Mo, Mn, Nb, Ni, Pb, Ru, Sb, Sn, Sr, Ta, Th, Ti, Tl, U, V, W, Y, Zn, and Zr) after processing wine materials by BT of various groups. More detailed data on the effect of stabilization and clarification by BT on the elemental image of various varieties of wine materials are given in the Supplementary Material to the article. There is a notable fact of an increase in the concentrations of REE and microelements in wine materials after treatment by BT of all groups. In addition, the introduction of BT of the same group during the clarification of different varieties of wine materials affects their elemental image in different ways.
Maximal concentrations of Pr and Eu were observed in Cabernet Sauvignon wine after clarification by BT of the first group; after clarification by BT of the second group-Y, Dy, Ho, Er, Tm, Yb, and Lu; and after BT of the third group-La, Ce, Nd, Sm, Gd, and Tb. Maximal concentrations of Eu were observed in Merlot wine material clarified by BT of the first group; by BT of the second group-Y, Tb, Ho, Er Tm, Yb, and Lu; and by BT of the third group-La, Ce, Pr, Nd, Sm, Gd, and Dy. In the wine material of Moldova, the maximum concentrations of La, Ce, Pr, Eu, and Ho were found after clarification by BT of the first group; Y, Er, and Lu after clarification by BT of the second group; and La, Nd, Sm, Gd, Tb, Dy, and Yb-after clarification by BT of the third group. The smallest change in the elemental composition of REE in all wine materials was caused by the BT of the fourth group.
Possible relationships between REE concentrations in bentonites and processed wine materials were studied by canonical correlation analysis (Table 4). For all three varieties of wine materials, the canonical correlation coefficients R were close to 1 (R = 0.99), indicating a high relationship between the REE content in bentonites and processed wine materials. The pair correlation coefficients r between the concentrations of REE in BT and wines shown in the table indicate strong relationships (r > 0.75), which are statistically significant (p < 0.05), and the content of REE in processed wine materials is directly proportional to their content in BT. The possible varietal difference/similarity of samples of processed Cabernet Sauvignon, Merlot, and Moldova wine materials by the REE content in them was studied by discriminant analysis ( Table 5). The predictors (independent variables) of the discrimination model were REE concentrations in BT-treated wine materials (32 samples each) and the grouping (dependent) variable was the wine material variety. The main criterion for evaluating the effectiveness of discrimination is the Wilks' Lambda value, the total value of which for discriminant analysis, taking into account all the variables involved, is presented in the upper part of Table 5. Opposite each of the variables is the Wilks' Lambda value for analysis in the case when this variable is not used. The Partial Lambda value characterizes the Wilks' Lambda ratio after and before adding the corresponding variable. The Partial Lambda characterizes the single contribution of the corresponding variable to the separating force of the model.
The proximity of the Wilks' Lambda (0.169) to zero in the upper box of Table 5 indicates successful discrimination, showing that the processed wine materials form clusters according to their varietal affiliation. Such discrimination confirms that wine materials within one denomination are homogeneous and have similarities in terms of the REE content. At the same time, wine materials of different denominations are heterogeneous with respect to the content of REE in them. Based on the Wilks' Lambda value, which is the result of the exclusion of the corresponding metal from the discrimination model, the metal contribution to the separation procedure can be assessed: the larger its value, the higher the contribution, and hence the role of this metal in the model. Table 5 shows the sequence of REE in descending order in terms of their contribution to the discrimination model and the formation of a varietal cluster structure of wine materials. It can be seen that the largest contribution to the formation of the varietal cluster structure of wine materials based on the REE content is provided by Yb, then Lu, Sm, etc., and the smallest contribution is provided by Ho. At the same time, the contribution of sx elements (Yb, Lu, Sm, Gd, Er, and Dy) to discrimination is statistically significant-the significance level of the Fisher p-test is less than 0.05 (highlighted in bold). A graphical illustration of the presence of a varietal cluster structure is a scatterplot of canonical values (Figure 2) with the presentation of wine materials as points on a plane. The areas of grouping the canonical values of wine materials have different configurations and colors depending on the variety. This representation allows the visual assessment of the degree of similarity/difference between the varieties through distances according to the principle: the smaller the distance, the greater the similarity. The resulting graphic illustration shows that the distances between all varieties of wine materials are small despite the presence of a cluster structure, which means that inter-varietal differences are not sufficiently pronounced, especially between Cabernet Sauvignon and Merlot. Therefore, the use of REE as markers of the studied varieties of Kuban wine materials is problematic as it can lead to erroneous final results of their identification.
The cluster structure was significantly improved by adding macroelements, i.e., Ca, Mg, K, and Na, to the list of predictors in addition to REE as varietal markers of processed wine materials (Figure 3). In this case, the degree of homogeneity of the groups and the similarity between the cluster structures of wine materials improved. However, using the combination of REE and macroelements as varietal markers is insufficient for a correct analysis.
In a previous work [34], we studied the role of macro elements, i.e., Ca, Mg, K, and Na, and microelements, i.e., Li, Co, Zr, Mo, Cd, Cu, Zn, Be, Ge, Nb, Rb, and Pb, in maintaining inter-varietal differences in wine materials. The combination of macro-and microelements led to a significant increase in intra-varietal similarity and inter-varietal differences in wine materials (Figure 4). Molecules 2023, 28, x FOR PEER REVIEW 7 of 14 The cluster structure was significantly improved by adding macroelements, i.e., Ca, Mg, K, and Na, to the list of predictors in addition to REE as varietal markers of processed wine materials (Figure 3). In this case, the degree of homogeneity of the groups and the similarity between the cluster structures of wine materials improved. However, using the combination of REE and macroelements as varietal markers is insufficient for a correct analysis.  In a previous work [34], we studied the role of macro elements, i.e., Ca, Mg, K, and Na, and microelements, i.e., Li, Co, Zr, Mo, Cd, Cu, Zn, Be, Ge, Nb, Rb, and Pb, in maintaining inter-varietal differences in wine materials. The combination of macro-and microelements led to a significant increase in intra-varietal similarity and inter-varietal differences in wine materials (Figure 4). In a previous work [34], we studied the role of macro elements, i.e., Ca, Mg, K, and Na, and microelements, i.e., Li, Co, Zr, Mo, Cd, Cu, Zn, Be, Ge, Nb, Rb, and Pb, in maintaining inter-varietal differences in wine materials. The combination of macro-and microelements led to a significant increase in intra-varietal similarity and inter-varietal differences in wine materials (Figure 4).  On the other hand, as a result of ion exchange, REE are included in the component composition of the wine material and, given the fact that their distribution is also different depending on the wine material variety, we considered their influence in maintaining inter-varietal differences in the composition with other elements. For this purpose, macroand microelements were added to the REE predictors ( Figure 5). Notably, the elemental images of wine materials after adding REE to the discrimination procedure retain the nature of the varietal difference in the elemental images of wine materials. The addition of REE had a slight effect on the elemental images of wine materials, increasing their intra-varietal similarity and inter-varietal differences. The weak influence of REE in the formation of the image of wine material is most likely due to the initial low content of REE in the soils of the grape-growing region. On the other hand, as a result of ion exchange, REE are included in the component composition of the wine material and, given the fact that their distribution is also different depending on the wine material variety, we considered their influence in maintaining inter-varietal differences in the composition with other elements. For this purpose, macroand microelements were added to the REE predictors ( Figure 5). Notably, the elemental images of wine materials after adding REE to the discrimination procedure retain the nature of the varietal difference in the elemental images of wine materials. The addition of REE had a slight effect on the elemental images of wine materials, increasing their intravarietal similarity and inter-varietal differences. The weak influence of REE in the formation of the image of wine material is most likely due to the initial low content of REE in the soils of the grape-growing region.

Research Objects
The studies were carried out with wine materials produced from Cabernet Sauvignon, Merlot, and Moldova grape varieties. The grapes were harvested in September

Research Objects
The studies were carried out with wine materials produced from Cabernet Sauvignon, Merlot, and Moldova grape varieties. The grapes were harvested in September 2019 in the Temryuk region (Cabernet Sauvignon variety) and Krasnodar (Merlot and Moldova varieties) in the Krasnodar Territory, Russia. The processing of grapes and the production of wine materials were carried out in accordance with the general rules of the production of wines [35]. The procedure for processing grapes and obtaining wine materials is described in more detail in [29].
The stages of clarification and stabilization of wine materials by BT were studied using 32 samples of bentonite clays produced in different countries, with varying degrees of dispersion and trademarks ( Table 6). All BT samples, except Khakass (BT13), Dagestan (BT15), and Crimean (BT27) origin are traditionally used in commercial wine-making technologies. Materials for claying wine materials (BT13, BT15, and BT27) were prepared from clays selected from deposits in the republics of Dagestan, Khakassia, and Crimea, considering the requirements for the production of wines [36]. In total, the studies were carried out with 99 samples from Cabernet Sauvignon, Merlot, and Moldova grape varieties, which included 3 initial and 96 samples of wine materials treated with bentonite clays (32 of each variety). Inorganic

Wine Clarification and Stabilization Using BT
The clarification and stabilization of wine materials by BT was carried out according to the generally accepted technology in the scientific center "Winemaking" of the Federal Scientific Center for Horticulture, Viticulture, and Winemaking, Krasnodar, Russia [37]. A more detailed procedure for the clarification and stabilization of wine materials is described in [29].

Elemental and X-ray Diffraction Analysis of BT
To study the possible effect of the procedure of clarification and stabilization by BT on the content of REE in wine materials, an X-ray phase analysis of the investigated fining agents was carried out on a Shimadzu XRD-7000 diffractometer (Shimadzu, Kyoto, Japan). According to the results of the X-ray phase analysis, the BT samples were divided into four groups, considering their qualitative and quantitative composition [29]. The first group included samples BT2, BT5, BT9, BT18, BT19, BT23, BT25, BT26, BT28, BT30, and BT32, which were based on sodium montmorillonite, the phase composition of which includes up to 3% of calcite (CaCO 3 ). The second group included BT1, BT7, BT10, BT12, BT14, and BT15 with a base of sodium-calcium montmorillonite and up to 3% of quartz. The third group, i.e., BT3, BT4, BT6, BT8, BT13, BT16, BT17, BT20, BT21, BT22, BT24, BT27, and BT31, contained various forms of montmorillonite with quartz and calcite contents of more than 5%. The fourth group included BT11 and BT29. Sample BT11 was a mixture of sodium-calcium montmorillonite, 10% quartz, and BT29 in addition to sodium-calcium montmorillonite it contained approximately 4% of a non-clay mineral albite and silicon oxide in the form of cristobalite.
The elemental composition of BT was determined by inductively coupled plasma mass spectrometry (ICP-MS) on an iCAP RQ spectrometer (Thermo Scientific, Waltham, MA, USA). Sample introduction was performed using a borosilicate nebulizer MicroMist (Glass Expansion, Melbourne, Australia). The spray chamber was cooled to 2.8 • C by a Peltier element sample aerosol solution. The compromising balance between the sensitivity and minimization of the matrix effect was achieved by the construction of the spectrometer interface (Ni-sampler 1.1 mm diameter, Ni-scimmer 0.5 mm diameter, and 3.5 mm scimmer insert). Samples were prepared for analysis using a MARS 6 microwave system (CEM, Charlotte, NC, USA) considering the recommendations of the system manufacturer (digestion of clay) [38]. Then, 0.2 g of each BT sample was transferred to the vessel accompanying the microwave system employed and 5.0 mL of concentrated HF, 3.0 mL of concentrated HNO 3 , and 1.0 mL of concentrated HCl were added. The mixture in the flask was gradually heated up to 200 • C for 15 min, then kept at this temperature for 10 min. To eliminate the loss of volatile elements, the flasks were opened at a temperature below 40 • C, then the samples were transferred into 50 mL flasks and made up to the mark with deionized water.

Determination of REE in Wines
The concentrations of REE in wine materials were determined by ICP-MS. A 10-fold dilution of wine samples with deionized water was chosen taking into account the literature data on their multi-element composition and the sensitivity of the analytical instruments used [39][40][41].
Considering the possible REE contents in wine materials, calibration curves were constructed using a set of standard solutions of Y, La, Ce, Pr, Nd, Sm, Eu, Gd, Tb, Dy, Ho, Er, Tm, Yb, and Lu with the analyte concentrations in a 0.001-20 µg/L range. The analysis conditions and operating parameters of the spectrometers are given in Table 7. The procedure for determining macro-and microelements in wine materials by the ICP-OES method is described in [29].

Statistical Analysis
The relationship between the content of REE in BT and processed wine materials was established by correlation analysis. The degree of similarity/difference between samples of processed and unprocessed wine materials in terms of the REE content was assessed by discriminant analysis. The calculations were carried out using the STATISTICA program (v.13) [42].

Conclusions
The discriminant analysis of the elemental composition of wine materials obtained from grape varieties grown on soils practically free of REE has confirmed that they are introduced during the processing of wine materials and poorly characterize their varietal and regional affiliation. The resulting graphic illustration shows that, despite the presence of a cluster structure, the distances between all varieties of wine materials are small, which means that inter-varietal differences are not sufficiently pronounced, especially between Cabernet Sauvignon and Merlot. The processed wine materials within one denomination were homogeneous, and the wine materials of different denominations were heterogeneous in terms of their REE content. However, these differences are not sufficient for a reliable classification of wines based on the content of REE.
Therefore, the use of REE as markers of wine material varieties obtained from grape varieties grown on soils practically free of REE is problematic since it can lead to erroneous final results of their identification. In these cases, REE are significantly inferior to macroand microelements in terms of their ability to represent the varietal image of wine materials.
It can be assumed that the use of ICP-OES and ICP-MS in combination with modern chemometric methods will reliably determine the quality and affiliation of wines by their elemental image and the relationship of components, detect counterfeits, and identify them by variety and origin.
Supplementary Materials: The following supporting information can be downloaded at: https:// www.mdpi.com/article/10.3390/molecules28114319/s1, Table S1: Content of REE in untreated and treated with bentonite clays wines.
Author Contributions: Z.T. conceived and designed the experiments; contributed reagents, materials, analysis tools, or data; and wrote the paper. M.B. analyzed and interpreted the data and wrote the paper. A.A. performed the experiments; analyzed and interpreted the data; and wrote the paper.