Sensory and Spectroscopic Characterization of Argentinean Wine and Balsamic Vinegars: A Comparative Study with European Vinegars

Depósito de investigación de la Universidad de Sevilla https://idus.us.es/ “This is an Accepted Manuscript of an article published by Elsevier in: FOOD CONTROL on 2020, available at: https://doi.org/10.1016/j.foodchem.2020.126791” Journal Pre-proofs Sensory and spectroscopic characterization of Argentinean wine and balsamic vinegars: a comparative study with European vinegars R. Ríos-Reina, S.M. Azcarate, J.M. Camiña, R.M. Callejón PII: S0308-8146(20)30653-1 DOI: https://doi.org/10.1016/j.foodchem.2020.126791 Reference: FOCH 126791 To appear in: Food Chemistry Received Date: 22 November 2019 Revised Date: 18 March 2020 Accepted Date: 10 April 2020 Please cite this article as: Ríos-Reina, R., Azcarate, S.M., Camiña, J.M., Callejón, R.M., Sensory and spectroscopic characterization of Argentinean wine and balsamic vinegars: a comparative study with European vinegars, Food Chemistry (2020), doi: https://doi.org/10.1016/j.foodchem.2020.126791 This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition of a cover page and metadata, and formatting for readability, but it is not yet the definitive version of record. This version will undergo additional copyediting, typesetting and review before it is published in its final form, but we are providing this version to give early visibility of the article. Please note that, during the production process, errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain. © 2020 Published by Elsevier Ltd. 1 Sensory and spectroscopic characterization of Argentinean wine 1 and balsamic vinegars: a comparative study with European 2 vinegars 3 R. Ríos-Reinaa, S.M. Azcarateb,*, J.M. Camiñab, R.M. Callejóna,* 4 a Área de Nutrición y Bromatología, Fac. Farmacia, Univ. Sevilla, C/P. García Gonzalez 5 no. 2, E-41012 Sevilla, Spain. 6 b CONICET- Facultad de Ciencias Exactas y Naturales, Universidad Nacional de La 7 Pampa, and Instituto de Ciencias de la Tierra y Ambientales de La Pampa (INCITAP), 8 Santa Rosa, Argentina. 9 *corresponding author: rcal[email protected]; [email protected] 10 11 ABSTRACT 12 In Argentina, vinegars are cheap agro-food products without exhaustive regulation and 13 the production of high-quality vinegars has not been exploited yet. In fact, Argentinean 14 vinegars have not been studied. In this context, a first study of Argentinean balsamic 15 and wine vinegars was carried out by a sensory and spectroscopic characterization 16 and by a comparison with well-recognized European vinegars. For that, ultraviolet- 17 visible and fluorescence spectroscopies were applied together with principal 18 component analysis (PCA) and parallel factor analysis (PARAFAC) performed on each 19 data set, respectively. Results showed differences between acetification processes, 20 origin countries and a wide variability within Argentinean production. The sensory 21 characterization on Argentinean wine vinegars was performed by triangular and 22 ordering preference tests showing statistically significant preferences toward the 23 traditional and the rapid vinegars. This work highlights the effect of production on 24 quality in order to provide added value to the Argentinean vinegars. 25 26 Keywords: Wine and Balsamic Vinegars, Industrial and Traditional Production, UV-vis 27 Spectroscopy, Sensory Evaluation, Exploratory Analysis. 28 29 30 2 31 1. Introduction 32 Consumers’ preferences towards high-quality products are felt in all types of food, 33 being increased in recent years in the vinegar world market (Paneque, Morales, 34 Burgos, Ponce & Callejón, 2017). Thus, a change in the appreciation of these products 35 has happened, particularly in balsamic and wine vinegars, which are increasingly 36 valued as a condiment. Whereby they have been differentiated and protected under 37 Protected Designation of Origin (PDO) as some high-quality wine vinegars in Europe 38 (Council Regulation (EC) No 510/2006) or some balsamic vinegars in Italy (Council 39 Regulation (EC) No 813/2000). 40 Argentina has been over the years an outstanding country respect to the production of 41 great wines, which possess recognized PDO such as “Luján de Cuyo” (INV, 2005) and 42 “San Rafael” (INV, 2007). However, the quality vinegar production has not already 43 been exploited adequately. Currently, a wide variety of vinegars are produced in the 44 country from different raw materials - including wine, alcohol and apple- and through 45 different processes of production. However, statistic information on production rate of 46 each type of vinegar has not been faithfully documented yet. 47 In order to obtain high production rate and low cost in the local market, wine vinegars in 48 Argentina are mainly elaborated by rapid industrial fermentation systems obtaining 49 products with a poor quality and few organoleptic nuances. From scarce information 50 providing mostly by producers in the region, it is estimated that only 19% of the total 51 production of vinegars are wine vinegars, with lack of a national production of high 52 quality at industrial scales. Otherwise, balsamic vinegar production is increasing and 53 reaching to 2% the last years (Grigolato, Massó del Valle & Schlatter, 2014). In the 54 internal market, the balsamic vinegars are commercialized as “type balsamic” 55 inasmuch as they are not produced through traditional method. In addition, different 56 3 ingredients such as honey, syrup, and molasses, among others, can be incorporated in 57 the final product. 58 Regarding to national regulations, the legal materials referred to both wine and 59 balsamic vinegars are scarce. The Codex Alimentarius Commission regulates the basic 60 specifications that a wine vinegar must meet in order to be marketed in the country 61 (CODEX 2008); however, a distinction between the different types of wine vinegar 62 production, or a statement that serves as an indicator of quality, have not been 63 established in this law. Respect to balsamic vinegars, even Codex does not contain a 64 specific definition about them. Therefore, the quality established requirements are 65 basic and equals for all vinegars marketed in Argentina. The lack of information and 66 regulation concerning to higher quality vinegar production has undoubtedly led to the 67 limited exploitation of this sector. 68 One of the most important quality parameters for consumers, which have a strong 69 economical relevance, are the odors and flavors (Starowicz, Koutsidis & Zieliński, 70 2018). Hence, sensory evaluation plays an important role in food quality and 71 complement a full study of a food product. One way to obtain consumer’s opinion is to 72 evaluate products by hedonic sensory test. Thus, various studies have been developed 73 in order to clarify how sensory characteristics of vinegars contribute to consumer 74 preference and satisfaction (Chung et al., 2012; Tesfaye et al., 2010; Cocchi et al., 75 2006; Tesfaye, Morales, García-Parrilla & Troncoso, 2009a). 76 However, as several hundreds of compounds from different families contribute to 77 vinegar flavor and quality, being mainly determined by the acetification system, the raw 78 material used and in some cases the wood aging (Ho, Lazim, Fazry, Zaki & Lim, 2017), 79 it is also necessary to investigate reliable analytical methods to establish criteria for 80 determining vinegar’s quality. In this sense, different techniques such as gas 81 chromatography-mass spectrometry (GC-MS), high-performance liquid 82 chromatography (HPLC) or capillary electrophoresis, have been developed for the 83 4 individual estimation of some compounds in vinegar to characterize and differentiate 84 them (Callejón, Ríos-Reina, Morales, Troncoso, Thomas & Camin, 2018). Although 85 these conventional methods provide high-quality information, they are expensive, 86 destructive, and time-consuming, requiring skilled operators and having a high 87 environmental impact. 88 In this context, other analytical platforms, such as spectroscopies, have emerged as 89 competitive fingerprinting techniques in vinegar’s quality control. Thus, rapid methods 90 such as near infrared spectroscopy (NIR) and Fourier transform mid infrared 91 spectroscopy (FTIR) have demonstrated its usefulness in the identification and 92 verification of raw materials and final products in broad types of vinegars (Lohumi, Lee, 93 Lee & Cho, 2015; Durán Guerrero, Castro Mejías, Natera Martín, Palma Lovillo & 94 García Barroso, 2010; Dong, Zheng, Jiao, Lang & Zhao, 2016; Sinanoglou et al., 2018; 95 Ríos-Reina, Callejon, Oliver-Pozo, Amigo & García-González, 2017a; Ríos-Reina, 96 García-Gonzalez, Callejon & Amigo, 2018; De la Haba, Arias, Ramírez, López, & 97 Sánchez, 2014). On the other hand, multidimensional fluorescence spectroscopy 98 (MFS) has also been adopted for wine vinegar authentication furnishing relevant 99 information about vinegar fluorescent components (Callejon, Amigo, Pairo, Garmón, 100 Ocaña & Morales, 2012; Ríos-Reina, Elcoroaristizabal, Ocaña-González, García- 101 González, Amigo & Callejón, 2017b). More recently, UV-Vis spectroscopy has been 102 scarcer implemented in vinegars, focusing only in vinegars produced from different raw 103 materials and blends of them (Torrecilla, Aroca-Santos, Cancilla & Matute, 2016; Xie, 104 Bu, Peng & Li, 2011). However, neither of these two last techniques have been 105 implemented in balsamic vinegars. The basic reasons of interest in these 106 methodologies are their advantages as fastness, non-sample preparation, and no- 107 solvent employment. Additionally, they are easy to handle so that specially trained 108 workers are not needed. These techniques are able to provide non-specific qualitative 109 information (spectral fingerprints) generating complex signals and huge amounts of 110 5 data in a short time, as a consequence, advanced and efficient data analysis strategies 111 have been raised (Medina, Pereira, Silva, Perestrelo, & Câmara, 2019). In this context, 112 the use of multivariate statistical analysis (chemometrics) for the evaluation and 113 interpretation of the spectral information has been reported (Efenberger-Szmechtyk, 114 Nowak, & Kregiel, 2018). 115 The wide applicability of spectroscopic techniques to a great deal of foods combined 116 with chemometrics, added to their advantageous aforementioned characteristics, make 117 them appropriate for developing vanguard analytical strategies for vinegar 118 characterization. At present and to our knowledge, characterization of Argentinean both 119 balsamic and wine vinegars and the differentiation of those produced by traditional and 120 industrial systems have not been reported in the literature. Accordingly, the aim of this 121 work was to carry out an exploratory analysis on spectral data obtained from MFS and 122 UV-Vis spectroscopy in order to characterize Argentinean balsamic and wine vinegars. 123 In order to assess the potential of these techniques combined with multivariate data 124 analysis in the differentiation of Argentinian vinegars, a detail description of the 125 absorbance and fluorescence bands of the different compounds that might differ 126 between different vinegars was accomplished. Furthermore, a comparison with well- 127 recognized European vinegars was carried out to evaluate similarities and differences 128 among them respect to the production systems. Finally, a sensory test was developed 129 to evaluate the organoleptic properties of the different types of Argentinean wine 130 vinegars. 131 2. Materials and methods 132 2.1. Samples 133 A total of 67 commercial vinegars from Argentina, Spain and Italy were acquired in 134 local wineries and stores. Specifically, 19 wine vinegars were obtained from different 135 Spanish wine vinegar PDOs (Vinagre de Jerez, Vinagre de Montilla-Moriles, and 136 Vinagre de Condado de Huelva), all of them aged in the minimum allowed period 137 6 (between 6 months and 2 years) and elaborated throughout traditional submerged 138 method in oak barrels, and 12 Spanish wine vinegars produced by quick acetification 139 process were studied. Regarding to Argentinean wine vinegars, 14 samples marketed 140 with similar prices were acquired in local market, except one of them which was 141 obtained from a local winery. The analyzed balsamic vinegars were industrial samples; 142 9 Italian samples and 13 Argentinean samples. More information is shown in Table 1. 143 2.2. Sample preparation 144 In order to acquire suitable spectroscopic information about vinegar samples and to 145 avoid detector saturations, prior to the analytical determinations by UV-Vis 146 spectroscopy, a dilution test with ultrapure water (Millipore Synergy System, Billerica, 147 MA, USA) was carried out. At this point, due to the differences in color intensity among 148 wine and balsamic vinegars as well as due to the way they are elaborated, both types 149 of vinegars were analyzed separately. Spectra of samples were collected at different 150 dilutions and subsequently analyzed, showing for each individual group –wine or 151 balsamic vinegarssimilar dilution effect (data not shown). Thus, the optimal dilutions 152 selected were the following vinegar/ultrapure water ratios: v/v: 1/10 and 1/50 for wine 153 vinegars and balsamic vinegars, respectively. These ratios allowed the observation of 154 unsaturated spectral curves with sharper peaks that could provide information for the 155 planned analysis. 156 2.3. Spectroscopic fingerprint measurements 157 2.3.1. UV-Vis analysis 158 UV-Vis spectroscopy measurements were performed using a spectrophotometer UV- 159 Vis Ocean Optics CHEMUSB4 coupled with a coupled charge detector (CCD). The 160 samples were placed in a quartz cuvette with a path length of 10 mm. The absorbance 161 as a function of wavelength was measured with a resolution of 2 nm in a working range 162 from 280 to 700 nm by triplicate. Ultrapure water was used as reference scan. 163 2.3.2. Fluorescence analysis 164 7 Fluorescence measurements were recorded using a Varian Cary-Eclipse fluorescence 165 spectrophotometer (Varian Iberica, Madrid, Spain), equipped with two Czerny-Turner 166 monochromators, and a Xenon discharge lamp pulsed at 80 Hz with a half peak height 167 of 2 ms (peak power equivalent to 75 kW). A high-performance R298 photomultiplier 168 tube detector was used for collecting the fluorescence spectra. Wine vinegar samples 169 were analysed directly without dilution, whilst, balsamic vinegars were diluted 10 times 170 with ultrapure water before the measurements to avoid impediment to the path of light. 171 A 1-cm path length standard quartz cell (Hellma Analytics, Müllheim, Germany) was 172 used to perform the measurements in a Peltier thermostatic cuvette holder (25.00 ± 173 0.05 °C). The spectrometer was interfaced to a computer with Cary-Eclipse software 174 for spectra acquisition and exportation. 175 The fluorescence Excitation-Emission Matrices (EEMs) were obtained by varying the 176 excitation wavelength (ʎex) ranging between 250 and 650 nm (every 5 nm) and 177 recording the emission spectra (ʎem) from 300 to 700 nm (every 4 nm). For these 178 measurements, excitation and emission slits were set at 5 nm and the scan rate was 179 fixed to 1200 nm min-1. In order to avoid a possible instrument wavelength drift, the 180 spectrofluorometer was calibrated every day using the water Raman peak as 181 reference. EEMs were recorded in triplicate for each wine vinegar type. 182 2.3. Sensory analysis 183 Seven Argentinean wine vinegars with color and olfactory strongly marked differences 184 were selected to carry out a preliminary screening in order to evaluate similarities, 185 differences and preferences among them. For that, each wine vinegar was analyzed for 186 sensory differences in terms of olfactory sensations by a panel of 48 consumers (32 187 females and 16 males, aged between 24 and 65). With this aim, first triangular tests 188 (IS0 4120 - 1983) were carried out, in which the panelists had to mark or choose which 189 sample was different, which sample had higher quality and which sample was 190 preferred. Samples were coded with three-digit numbers and distributed in a 191 14 the best model. The obtained model was enough robust, explaining more than 99% of 349 the variance with a core consistency over zero. Figure III. Supplementary Material 350 includes the PARAFAC loadings (excitation and emission spectra) of the extracted 351 factors in the left side, and the scores plot of the third and fourth factors (F3 and F4) in 352 the right side of the figure. 353 As could be seen in the scores plot of Figure III. SM., F3 and F4 were responsible of 354 the differentiation between acetification processes, as the STWV samples with a 355 traditional production method and recognized under a PDO were separated by F4 from 356 the wine vinegars made by a quick acetification process (SIWV and AIWV), mainly 357 explained by F3, regardless of the origin of the samples. This separation is showed by 358 a dash line in Figure III. SM. 359 The third factor (F3 represented in yellow) showed excitation and emission maxima 360 around 350 nm and 400 nm, respectively with a shoulder with high intensity around 425 361 nm. In contrast, the fourth factor (F4, in purple) was a peak centered at 450/550 nm of 362 excitation/emission wavelengths. These wavelengths agreed with the trend observed 363 for SIWV and AWV samples in contrast to STWV by looking at their landscapes (Fig. 364 1c). The fluorescent loading patterns of the modelled factors can be also linked to 365 some fluorophores previously described in the literature (Callejón, Amigo, Pairo, 366 Garmón, Ocaña & Morales, 2012; Ríos-Reina, Elcoroaristizabal, Ocaña-González, 367 García-González, Amigo & Callejón, 2017b; Ríos-Reina, Ocaña, Azcarate, Pérez- 368 Bernal, Villar-Navarro & Callejón, 2019; Airado-Rodríguez, Durán-Merás, Galeano- 369 Díaz & Wold, 2011; Sádecká, Tóthová & Májek, 2009; Azcarate, De Araújo Gomes, 370 Alcaraz, Ugulino De Araújo, Camiña & Goicoechea, 2015). Thus, phenolic compounds 371 showed excitation/emission wavelengths around 330/420 nm, matching well with F3, 372 and mainly presented in the samples obtained by a quick industrial acetification 373 process. These compounds are naturally present in wines, and due to the simple and 374 quick production method without aging in wood barrels, these compounds remain 375 15 without many modifications or losses. However, the F4 wavelengths could be linked to 376 compounds originated from wooden casks or formed during the Maillard reactions such 377 as HMF (Ríos-Reina, Elcoroaristizabal, Ocaña-González, García-González, Amigo & 378 Callejón, 2017b; Ríos-Reina, Ocaña, Azcarate, Pérez-Bernal, Villar-Navarro & 379 Callejón, 2019). These results reaffirm the relation between this factor with PDO wine 380 vinegars (STWV), due to these vinegars are made by a traditional method, using wood 381 barrels and being aged for at least 6 months, which modifies their composition 382 including these kind of compounds, matching also with the model obtained by UV-Vis 383 spectra. 384 Moreover, regarding the rapid wine vinegars (AWV and SIWV), a separation between 385 Spanish and Argentinean samples could be also noticed by both spectroscopic 386 techniques. For this reason, the next step was to develop PCA models with UV-Vis 387 spectra and PARAFAC models with fluorescence spectra to assess the differentiation 388 between samples with the same acetification but different production country. 389 3.2.2. Differences between wine vinegars from different country 390 It could be seen an evident difference in the intensity between industrial wine vinegars 391 from different countries by a simple observation of the UV-Vis spectra (Fig. 1a). Thus, 392 AWV samples showed the highest intensity respect to SIWV, which presented less UV 393 intensity, whereas in the visible range these last samples did not present any 394 absorption bands. Accordingly, in order to visualize these differences and to evaluate 395 their causes, a PCA model was developed considering SIWV and AWV (excepting 396 DOS_20 and DON_26 samples). 397 Figure 2A shows the scores plot displaying the two first PCs where it could be 398 observed that most of the samples were clearly separated into two groups according to 399 their production country. Thus, PC1, which explained 89.65% of the total variance, 400 clearly described the separation of SIWV, in the negative side, from AWV in the 401 16 positive side. Furthermore, it could be also observed that throughout the PC2, 402 explaining the 8.70% of total variance, the samples were scattered within each group. 403 These results showed again the higher variability in the wine vinegars elaborated in 404 Argentina. Furthermore, it could be also observed that three SIWV samples namely 405 TBOR, HACT and BBOR were grouped with AWV samples. 406 Insert here Figure 2 407 Both situations can be explained based on the raw material used, due to the wide 408 variety of grapes used in wine and vinegar production. Generally, for industrial wine 409 vinegars, a blend of different raw wine varieties can be accomplished, such as red, 410 rosé or white wines. In this sense, it is very simple to establish the type of wine used in 411 Spanish production inasmuch this information is usually indicated in the labels; 412 conversely, AWV labels do not specify the raw wine used. Undoubtedly, although many 413 of them are elaborated from red wines considering that they are the major national 414 production, however, and taking in account the different colors of each AWV sample 415 (Fig. I SM), they can be also elaborated from rosé and white wine or, more probably, 416 from a blend of them. These considerations could explain the Argentinean production 417 variability and the marked differences with the Spanish samples. In addition, owing the 418 three SIWV above mentioned were produced from red wine, this could be the reason 419 for which they were grouped with AWV samples. 420 The UV-Vis data is plainly profitable to differentiate wine vinegars that have been 421 elaborated from different raw wines. PC1 and PC2 loadings (Figure 2B) showed that 422 the separation of both groups seemed to be explained by a stronger absorption band at 423 300 nm. Moreover, an additional important band in PC2 around 380 nm could reveal 424 that the variability found in each country could be attributable, among the main 425 reasons, to the varieties of grape or alike the raw wine used in the production. 426 Considering other works, the phenolic composition of wine vinegars had demonstrated 427 to be useful to differentiate and predict the applied elaboration method and even to 428 17 assess the geographical origin of wines used as substrate (Callejón, Amigo, Pairo, 429 Garmón, Ocaña & Morales, 2012; Ríos-Reina, Elcoroaristizabal, Ocaña-González, 430 García-González, Amigo & Callejón, 2017b). 431 Insert here Figure 2 432 Concerning the fluorescence spectra, a 4-factor PARAFAC model was obtained with 433 SIWV and AWV wine vinegars. In addition, in order to test the optimum number of 434 factors and to achieve a good variance and core consistency, one independent 435 PARAFAC model for each kind of vinegar was made. Extracted factors were plotted 436 together as is shown in Figure 2C. The results explained that 4 factors were needed 437 due to each group of samples were explained by 3 factors, and when these factors 438 were plotted together, it could be seen that the first and third factors (yellow and blue 439 lines in Fig.2C) matched quite well between the two models, while the excitation 440 wavelengths for the second factor (F2, plotted in orange) were different between AWV 441 and SIWV models. 442 Moreover, when the scores of F1 and F3 are plotted together (Fig. 2D), it could be 443 seen a good grouping and separation of the SIWV samples, with higher scores for F1 444 and F3, from the AWV, except for three SIWV samples that were overlapped. These 445 three samples were the same overlapped in the UV-vis PCA model (Fig. 2A). F3 for 446 both models had an excitation and emission maxima around 350 nm and 400 nm, 447 respectively, and F1 showed an excitation maximum around 375 nm and an emission 448 maximum around 450 nm for SIWV vinegars and slightly upper (i.e. 475 nm) for AWV 449 vinegars. According to these wavelengths, fluorophores could match with the presence 450 of coumarins, tannins and other unknown fluorescent compounds originating from 451 wooden casks, as well as with phenols and flavonols (Ríos-Reina, Elcoroaristizabal, 452 Ocaña-González, García-González, Amigo & Callejón, 2017b; Airado-Rodríguez, 453 Durán-Merás, Galeano-Díaz & Wold, 2011). 454 18 According to both UV-Vis and fluorescence results, it was concluded that the 455 concentrations of phenolic compounds varied in different wine vinegars as a function of 456 several factors, being the grape variety one of the most important. In general terms, 457 some phenols contained in AWV, such as benzoic acids, prevailed in concentration 458 over SIWV. Otherwise, AWV information in the visible range, could be explained by the 459 contribution of anthocyanins (475-545 nm in UV-Vis spectra), which could be 460 responsible of the color. Therefore, the phenolic composition was the largest 461 contributing factor for the wine vinegar’s discrimination, indicating the potential 462 application of this technique for variety establishment purposes. 463 3.2.3. Differences between Argentinean wine vinegars 464 Ultimately, in order to evaluate the national Argentinian production of wine vinegars, a 465 PCA on AWV samples was carried out. Scores plot of the first two PCs explaining the 466 98.41% of the total variance and displayed in the Figure 3A, established three well- 467 differentiated groups: the group H which contained two samples (DOS_20 and 468 DON_26) that coincided with those that were grouped with STWV samples when 469 differences between acetification process in wine vinegars were evaluated in Section 470 3.2.1; and other two groups, a group M which included three AWV samples (GRE_15, 471 HAL_21 and CON_22) and a group L which included the rest of AWV. The main 472 differences between the groups could be evaluated from Figure 3B, which shows the 473 loadings plot of the first two PCs. Once again DOS_20 and DON_26 samples were 474 separated from the rest throughout of PC1 by a band around 325 nm, that could 475 involve those compounds related to traditional process and contributing to higher signal 476 intensities. The PC2 loadings displayed two peaks responsible for the separation of the 477 other two groups, indicating again the variability inside the same region. Thus, the 478 current grouping could be explained by the used grape varieties as well as the 479 implemented production. Moreover, when a 3-factor PARAFAC model was made with 480 19 these samples (Fig 3C), and the scores plot of F1 and F3 was plotted together (Fig 481 3D), it could be seen again the same three groups observed by UV-Vis. 482 Insert here Figure 3 483 3.3. Balsamic vinegar exploratory data analysis 484 Due to balsamic vinegars present characteristics very different to wine vinegars, they 485 were studied separately. Figure 1B showed the balsamic vinegar’s absorption bands. 486 The most of them had a strong spectral band that encompassing wavelengths from 280 487 nm to 680 nm. The spectra showed a similar shape in all samples differing in their 488 intensity. Thereby, Italian balsamic vinegar showed significantly higher intensities than 489 Argentinean balsamic vinegars in this spectral range; therefore, they could be good 490 indicators for the discrimination among them. 491 Thus, in order to assess the specific properties of each group of balsamic vinegars, a 492 PCA was performed with the UV-Vis spectra of these samples. The scores plot 493 obtained by selecting the first three PCs (Fig. 4A) showed similarities and differences 494 among the samples, being PC1, which described 97.05% of the variance, the PC 495 mainly responsible of the differentiation between Argentinian and Italian balsamic 496 vinegars. This PC1 exhibit a maximum peak in the loadings plot around 360 nm mainly 497 responsible of the sample’s discrimination (Figure 4B), that could be associated to the 498 contribution in the characteristic color of these vinegars of some raw materials specially 499 used in balsamic vinegar production, such as condensed musts, cooked musts, wine 500 vinegars and caramel in different ratios. 501 Insert here Figure 4 502 Accordingly, the typical components of these vinegars are acetate, HMF, malic acid, 503 and glucose. Some authors have demonstrated that the concentration of HMF, a 504 product that comes from sugar, was higher in balsamic vinegars than in white and red 505 wine vinegars (Theobald, Muller & Anklam, 1998). For that, the main differences 506 20 observed between ABV and IBA could reside basically in the types and proportions of 507 the raw materials used for their production. 508 Regarding fluorescence data, a 4-factor PARAFAC model was made with balsamic 509 vinegars. As in the above sections, in order to test the global model as well as to study 510 the differences in the extracted fluorophores for each class, one PARAFAC model was 511 also made separately for each kind of balsamic vinegar (i.e. ABV and IBV). As it could 512 be seen in Figure 4C, the optimum number of factors needed was 3 and 4 in IBV and 513 ABV PARAFAC models, respectively. Thus, Argentinian samples needed one more 514 factor than the Italian’s, showing a maxima of excitation/emission wavelengths at 515 325/350 nm. 516 Furthermore, it could be seen that excitation/emission wavelengths for F1 and F2 were 517 similar to the above factors observed in the wine vinegar’s PARAFAC models, might 518 being explained by some fluorophores that are characteristic of the raw wine or grapes. 519 However, F4 for ABV and F3 for IBV models showed excitation/emission maxima 520 slightly different form the other factors in this study, showing a maximum at 500 nm of 521 excitation and 550 nm of emission wavelengths. These wavelengths agreed with those 522 observed in a previous study about the addition of grape-must caramel to wine 523 vinegars, a common practice in balsamic vinegar production (Ríos-Reina, Ocaña, 524 Azcarate, Pérez-Bernal, Villar-Navarro & Callejón, 2019). Moreover, F2 in ABV and F1 525 in IBV models, with maxima excitation/emission wavelengths centered at 400/500 nm, 526 could match with HMF (Ríos-Reina, Elcoroaristizabal, Ocaña-González, García- 527 González, Amigo & Callejón, 2017b; Ríos-Reina, Ocaña, Azcarate, Pérez-Bernal, 528 Villar-Navarro & Callejón, 2019), above described as a compound that comes from 529 sugar and with higher presence in balsamic vinegars than in wine vinegars. 530 Regarding the scores plot of F2 vs F4 (Fig. 4D), it could be seen a separation between 531 both origins, Argentina and Italy, and also between some ABV, mainly based in the 532 proportions of wine, grape-must caramel and sugars used for their production, 533 21 reinforcing the UV-Vis results. Thus, samples with higher F4 and F2 scores could have 534 more grape-must caramel and sugars than those placed in the left and bottom corner 535 of the plot. 536 Finally, the geographical origin of the samples should be also taken into account, due 537 to it could also contribute to the differentiation between balsamic vinegars. This 538 differentiation can be explained by several compounds, which were above mentioned 539 in fluorescence and UV-Vis spectra, tightly related to wine and wine vinegar 540 discrimination. 541 3.4. Sensory analysis of Argentinean wine vinegars 542 In order to know the consumers’ opinion respect to the wine vinegars marketed in 543 Argentina, two sensory tests were conducted. From three groups visualized in the 544 scores scatter plot of Argentinean wine vinegars, seven samples were selected: two 545 samples of group H (A1, A2), two samples of group M (A3 and A4) and three samples 546 of group L (B1, B2 and B3) (Figure 4A). Foremost, from a previous test performed by 547 some expert members, it was possible to anticipate that the four wine vinegars 548 designated as A seemed to be of better quality than those three named as B when the 549 differences in color and olfactory sensations were assessed. 550 For this reason, a first step was to evaluate if the judges’ panel was able to distinguish 551 A samples from B samples. Thus, a discriminative triangle test was implemented in 552 order to know if there were noticeable sensory differences between two products 553 without having to specify the nature of the possible difference. 554 For this test, all A samples were compared with one B sample, existing for each case 555 twelve possible combination of the samples with their respective duplicates (e.g., 556 B1A1B1, A1B1A1, B1A2B1, A2B1A2, B1A3B1, A3B1A3, B1A4B1, A4B1A4 –exactly the same 557 sequences were repeated front of B2 and B3 samples-), which were presented to the 558 judges in randomly order and then, they identified the different sample through 559 22 olfactory sensation. Certainly, there were clear sensory differences between both wine 560 vinegar groups being perfectly differentiated for all judges in all sequences for the 561 samples A1 and A2 from B samples. Regarding to A3 and A4 samples, they were 562 correctly differentiated from B samples in 83.3% and 72.9% of the cases, respectively. 563 In addition, all judges, who were able to differentiate correctly all examined sequences, 564 reported and coincided that the samples A1, A2, A3 and A4 were of better quality and 565 they preferred them. On the other hand, it could be observed that the judges who not 566 properly distinguished the samples A3 and A4, they confused them in the sequences 567 that sample B3 was included. In these cases, they had difficulties to choose the better- 568 quality sample and, in some sequences, they did not prefer none of them. 569 The next step was to assess the degree of consumer’s wine vinegar preference. For 570 that, a preference test directed by order test was performed. Taking into account the 571 obtained results in triangle test the selected samples for this study were those 572 statistically preferred by the judge’s panel namely A1, A2, A3 and A4. Moreover, the 573 sample B3 was also included considering the confusions presented by part of the panel 574 when this sample was evaluated. Thus, the consumer had to order the samples 575 according to the preference degree. For that, the five wine vinegar samples (A1, A2, 576 A3, A4 and B3) were presented to each judge and then, according to their preferences 577 they ordered them from the most to the least preferred. Table I. Supplementary 578 Material summarizes the trends in the preference orders presented for the panel. From 579 these results, in order to establish if a specific sample is preferred in front of others, the 580 rank sums for each sample were calculated and then, they were compared using the 581 Kramer test (see Table 2). For 5 items ranked by 48 panelists the critical value at p < 582 0.05 is 42.3. Then, items whose rank sums differ by more than this amount may be 583 considered significantly different (Lawless & Heymann, 2010). The obtained results 584 revealed there were not any significant differences between A1 and A2. However, 585 these two samples had a significant preference respect to A3, A4 and B3 samples. 586 23 Moreover, A4 was significantly most preferred than A3 and B3 and, A3 and B3 were 587 equally preferred. 588 Overall, both A1 and A2 wine vinegars were highly preferred compared to the rest of 589 the studied vinegars. The A1 wine vinegar is the sample (DON_26) which was 590 produced in traditional way in a local winery, along with another brand of the market, 591 A2 (CAS_13), that belongs to a prestigious Argentine company. These two samples 592 were preferred significantly compared to other vinegars with less acceptance. Despite 593 significant differences between A1 and A2 were not presented, A1 sample was 594 selected in the first place 87% more than A2 selected in second order. 595 Furthermore, in this instance, the panel had the possibility to offer a brief description of 596 the olfactory sensation of each tested sample. From gathering contributed information 597 by each judge, the predominant odors for sample A1 were raisins, wood, figs, peach, 598 nuts, chocolate, ethyl acetate, glue, sweet and liquor and defining it with high aromatic 599 complexity. For sample A2, some coincident characteristic such as raisins, wood, 600 peach, sweet and liquor attributes were also described adding other attributes as sweet 601 cider, wine, whiskey, licorice, ripe fruit and toasted notes. Apparently, these peculiar 602 attributes were highly accepted by consumers inasmuch as these two samples were 603 significantly the most preferred. Besides, comparing both samples, the most judges 604 considered that sample A1 had a greater intensity and aromatic complexity. 605 Otherwise, regarding consumers’ descriptions, the less preferred wine vinegar sample 606 B3 presented aggressive characteristics and unpleasant smell with specific descriptors 607 as yeast, cheese, fermented fruit, sweet, apple, sour acid, pungent, bitter, and rancid, 608 according to literature (Tesfaye, Morales, Callejón ,Cerezo, González, García-Parrilla 609 &Troncoso 2010), and it was very poor aromatically, so it was rejected against other of 610 better quality. Concerning sample A3, the described characteristics were glue, plum, 611 vegetable, meat, caramel, unpleasant, pungent, bitter, and vanilla while for sample A4 612 were figs, ripe fruit, sweet, raisin, honey, metallic, yeast, vegetable, itchy. It is 613 30 Food Science and Nutrition, 58, 1767-1779. 776 https://doi.org/10.1080/10408398.2017.1284742. 777 Tesfaye, W., Morales, M. L., García-Parrilla, M. C., & Troncoso A. M. (2009a). 778 Improvement of Wine Vinegar Elaboration and Quality Analysis: Instrumental 779 and Human Sensory Evaluation Wine Vinegar Elaboration and Quality Analysis. 780 Food Reviews International, 25, 142–156. https://doi.org/10.1111/j.1745- 781 459X.2009.00253.x. 782 Tesfaye, W., Morales, M. L., Callejón, R. M., Cerezo, A. B., González, A. G., García- 783 Parrilla, M. C., & Troncoso, A. M. (2010). Descriptive sensory analysis of wine vinegar: 784 Tasting procedure and reliability of new attributes. Journal of Sensory Studies, 25, 216- 785 230. https://doi.org/10.1111/j.1745-459X.2009.00253.x. 786 Theobald, A., Muller, A., & Anklam, E. (1998). Determination of 5- 787 hydroxymethylfurfural in vinegar samples by HPLC. Journal of Agricultural and 788 Food Chemistry, 46, 1850–1854. https://doi.org/10.1021/jf970912t. 789 Torrecilla, J. S., Aroca-Santos, R., Cancilla, J. C., & Matute G. (2016). Linear and non- 790 linear modeling to identify vinegars in blends through spectroscopic data. LWT - 791 Food Science and Technology, 65, 565-571. 792 https://doi.org/10.1016/j.lwt.2015.08.027. 793 Xie, H. D., Bu, L. J., Peng, X. W., & Li, Z. X. (2011). Ultraviolet Spectroscopy Method 794 for Classifying Vinegars. Advanced Materials Research, 346, 865–74. 795 https://doi.org/10.4028/www.scientific.net/AMR.346.865. 796 Yu, J., Wang, H., Zhan, J., & Huang, W. (2017). Review of recent UV-Vis and infrared 797 spectroscopy researches on wine detection and discrimination. Applied 798 Spectroscopy Reviews, 53, 65-86. 799 https://doi.org/10.1080/05704928.2017.1352511. 800 801 Figure Captions 802 31 803 Figure 1: UV-Vis absorption spectra (A and B) and fluorescence landscapes (C and 804 D) of the different studied vinegars: Spanish traditional wine vinegars –STWV- (red 805 line), Spanish industrial wine vinegars –SIWV- (purple line), Argentinean wine vinegars 806 –AWV- (blue line); and Balsamic vinegars, Argentinean –ABV- (green line) and Italian – 807 IBV- (pink line). 808 809 Figure 2. Scores (A) and loadings (B) plots of the first two PCs obtained by PCA on the 810 spectra of the Spanish industrial wine vinegars (SIWV) and Argentinean industrial wine 811 vinegars (AWV). Excitation and Emission spectra (PARAFAC loadings) of the main 812 fluorophores (C) and scores plot of Factor 1 vs Factor 3 (D) obtained with SIWV and 813 AWV samples. 814 815 Figure 3. Scores showing three groupings H, M, L (A) and loadings (B) plots about the 816 first two PCs showing the differentiation of argentinean wine vinegars (AWV) by UV-vis 817 spectra. (C) and (D) shows the excitation and emission spectra (PARAFAC loadings) of 818 the main fluorophores for AWV and scores plot of the first vs thrid extracted factors. 819 820 Figure 4. Scores (A) and loadings (B) plots about the first three PCs showing the 821 differentiation between balsamic wine vinegars (ABV and IBV) by UV-vis spectra. (C) 822 and (D) shows the excitation and emission spectra (PARAFAC loadings) of the main 823 fluorophores for ABV and IBV and scores plot of the extracted Factor 2 vs Factor 4. 824 825 32 826 Table 1. Samples information included in the study 827 828 Vinegars Class Origin Code N* Acetic grade (%) Wine Non-PDO Argentinean AWV 12 5-6 Spanish SIWV 13 6 PDO Spanish STWV 19 6 Balsamic Non-PDO Argentinean ABV 14 4-6 Italian IBV 9 6 *N = number of samples. 829 830 831 832 833 834 835 836 837 838 839 840 841 842 843 844 845 846 33 847 Table 2: Data analysis comparing rank sums of the samples for Kramer test. 848 849 Sample/ Rank sum A1 A2 A3 A4 B3 60 91 197 137 235 A1 60 0 -31 -137 -77 -175 A2 91 31 0 -106 -46 -144 A3 197 137 106 0 60 -38 A4 137 77 46 -60 0 -98 B3 235 175 144 38 98 0 850 851 852 853 854 34  It was the first time that Argentinian wine and balsamic vinegars were studied. 855  UV-vis and fluorescence spectroscopies were used for a vinegar characterization. 856  Argentinian wine and balsamic vinegars were also characterized by sensory analysis. 857  Differences between acetification, origin and Argentinean productions were shown. 858  These results encourage the production of high-quality vinegars and new regulations. 859 860 35 Credit Author Statement 861 R. Ríos-Reina: Experimental, Data analysis and Writing. S.M. Azcarate: Experimental, 862 Supervision, Writing and Reviewing. J.M. Camiña: Methodology, Wrirting and Editing. R.M. 863 Callejón: Supervision, Writing, Reviewing and Editing. 864 865 36 866 867 37 868 869 38 870 871 39 872