Boron Isotopic Ratio in Brazilian Red Wines : a Potential Tool for Origin and Quality Studies ?

For the first time, the boron isotopic ratio is reported for Brazilian red wines. The boron concentration, B/B isotopic ratio and δB were determined in 89 red wines produced in Campanha Gaúcha, Serra Gaúcha (SG), and Vale dos Vinhedos (VV) in South Brazil and Vale do São Francisco (VSF) in Northeast Brazil. A broad boron concentration range, < 0.07 to 20.4 mg L, was observed, where a few values were above the maximum allowed concentration according to the International Code of Oenological Practices (14.0 mg L). The obtained B/B ratios were quite different from those observed in the literature. The ratios were quite similar among the Southern Brazilian wine producing areas, but a large difference was observed when compared with the Northeastern wines, allowing for the differentiation between these two important Brazilian wine regions. The results have also shown that is possible, based on the boron isotopic ratio, to differentiate the Brazilian red wines from Chilean, Italian and Portuguese red wine.


Introduction
Boron has two natural isotopes, 10 B (19.9%) and 11 B (80.1%), 1 and is an oligo-element essential to plant health and growth.In wines, boron is present in a concentration of 0.1 to 10 mg L -1 , expressed as boric acid, [2][3][4] with a maximum acceptable boric acid concentration of 80 mg L -1 (14 mg L -1 B), according to the International Organisation of Vine and Wine (OIV). 5he boron isotopic variation can be expressed either as the 11 B/ 10 B ratio or as δ 11 B, [6][7][8][9][10] where δ 11 B is defined as: (1)   where ( 11 B/ 10 B) sample is the 11 B/ 10 B isotopic ratio of the sample and ( 11 B/ 10 B) standard is the 11 B/ 10 B isotopic ratio according to the NIST 951a Boron Acid Isotopic Standard.
][13][14][15][16] According to Vorster et al., 7 if the natural variation in the 11 B/ 10 B ratio of the soil is reflected in the vines, it can become an important tool in provenance studies of agricultural products, such as wines, helping to minimize adulteration and fraud in these products.
The 11 B/ 10 B ratio has been applied in wine geographical provenance studies to elucidate the country or even region inside a country.Coetzee and Vanhaecke 6 observed differences that allowed them to differentiate between wines produced in South Africa, France and Italy.Bruyn 3 was able to observe differences among the soils and wines from the South African regions of Swartland, Stellenbosh and Robertson.Combining the boron isotopic ratio with elemental analysis, Vorster et al. 7 and Coetzee et al. 8 also classified the soils and wines from the Swartland, Stellenbosh, Robertson and Walker Bay regions.These authors claim that other South African regions could be included, aiming for the development of a fingerprint for each wine-producing area of the country.
Of course, the applied equipment directly influences the obtained uncertainty.Quadrupole inductively coupled plasma mass spectrometry (Q-ICP-MS) has poor precision (relative standard deviation (RSD) = 0.1-0.5%)compared to thermal ionization mass spectrometry (TIMS) (RSD = 0.002-0.005%)and multicollector ICP-MS (RSD = 0.2-0.002%).][8] It should be pointed out that beside lead, which was also applied to Brazilian red wines, 17 other isotopes have been utilized in viticulture, Santesteban et al. 9 presented a review about this subject.Delta(D) and delta(O-18) are quite applied to mineral water provenance studies 18 and its use could be extended to detect water addiction in wines. 9rapevines are considered plants with C3 photosynthesis cycle and present a δ( 13 C) around −25‰ while sugar cane is a C4 plant with a δ( 13 C) about −11‰, therefore, sugar addition will alter the observed delta value. 9,19However, water deficit could also alter the δ( 13 C) and has to be taken into account. 9,19Particularly interesting is the application of δ( 87 Sr) since wine/soil delta values are close to one, 8 but, the same authors showed that, in order to obtain better discrimination among the wine regions, smaller uncertainty is necessary and, in this direction, the application of boron had provided better results. 8he present study represents the first study of the boron isotopic ratio in Brazilian red wines.Sixty percent of the wine produced in Brazil is bottled outside of its producing region, making it difficult to trace its identity and follow its quality. 20Another characteristic is the organization of the production system based on producer cooperatives, which includes up to 370 families, resulting in wine that is produced from vines from different families and eventually involving different cities than the designated origin.Therefore, at the consumer level, one has to trust the information printed on the label, including the location of the vines and the presence of additives.Based on these arguments, we decided to perform the present work on wines bought at the consumer level to evaluate the application of boron content and boron isotopic ratio as a tool to evaluate the wine provenance and quality.

Sampling and sample treatment
Information related to the involved wine regions, sample collection and sample treatment is described in detail by Almeida et al. 17 A total of 89 Brazilian red wines were analyzed: 10 from Campanha Gaúcha, 36 from Vale dos Vinhedos, 33 from Serra Gaúcha and 10 from Vale do São Francisco.
For the quantitative determination of boron by applying inductively coupled plasma optical emission spectrometry (ICP OES), the samples were diluted 1:100 with Milli-Q ® quality water.Based on the obtained boron content, the samples were diluted accordingly to obtain a boron concentration of approximately 20 µg L -1 , in order to avoid dead-time corrections during isotopic ratio determination. 17

Reagents and standard solutions
All solutions were prepared with Milli-Q ® quality water.Suprapur nitric acid (65%) (Merck) was used in the oxidation of the wine samples, mannitol (20%) (Sanobiol) and PA-grade ethanol (Merck) were used in the matrix effects studies.

Instrumentation
The boron content was determined by applying an Optima 4300 DV ICP OES (PerkinElmer), equipped with a cyclonic chamber-type twister (Glass Expansion).The corresponding operational parameters are described in Table 1.
The boron isotope ratios were determined using an ICP-MS (Agilent, model 7500ce), equipped with an octopole reaction system (ORS), a low-flow nebulizer (MicroMist) connected to a quartz nebulizer chamber (Scott) mounted on a Peltier block, which was kept at 2 °C, and a quartz torch with a 2.5 mm diameter injector.Nickel sampler and skimmer cones were utilized.Additional operational conditions are given in Table 1.

Isotopic ratio determination
Due to the large relative mass difference involved in the 10 B/ 11 B isotopic ratio measurement, which leads to enhanced mass discrimination effects, when using ICP-MS, the sequence proposed by Coetzee and Vanhaecke 6 was applied: standard -sample 1 -standard -sample 2standard.
The uncertainty in the mass bias correction factor (β) and the uncertainty in the calculated atom ratio were derived as proposed by Godoy et al. 21lthough the mass bias correction was applied only during sample analysis and accuracy testing and not during the first method development steps, as ethanol affects the tests, a larger deviation from the certified value was applied as a decision criteria.

Data analysis
The boron isotope ratios were analyzed using one-way analysis of variance (ANOVA).The degree of discrimination among the different regions may be an indication of the possible application of this technique to verify the existence of significant differences among the regions.

Results and Discussion
Matrix effects: carbon content Potential tailing effects of the 12 C peak on the 11 B peak can lead to a higher 11 B/ 10 B ratio, as reported by Coetzee and Vanhaecke, 6 mainly when dealing with samples with high carbon content.According to Barbaste et al., 22 the effect of ethanol on the 11 B/ 10 B ratio can be reduced by using a cooled nebulizer chamber, as utilized in the present work.
If ethanol has an influence on the 11 B peak, the 11 B/ 10 B ratio should increase with ethanol content.Solutions of NIST SRM 951a (20 µg L -1 ) were prepared containing 0.125, 0.25, 0.50 and 1.0% ethanol.As shown in Table 2, an effect was not observed even at 1% ethanol, which corresponds to a wine sample diluted 1:10.
Another possible way to investigate the 12 C effect on the 11 B/ 10 B ratio is by successive dilution of the wine sample, since wine contains organic compounds other than ethanol.As shown in Figure 1, a matrix effect was observed.At a wine dilution of 1:50, a very high 11 B/ 10 B ratio was observed due to the 12 C tailing effect on the 11 B peak.The effect is quickly reduced, and minor changes were observed at dilutions higher than 1:100.As low boron concentrations lead to poor measurement statistics, a higher dilution should be applied only when the boron content allows.

Mannitol effect
Boron loss by volatilization during sample treatment steps, by evaporation or acid dilution, is a concern in studies involving this element.One important boron characteristic is the formation of cis-diol complexes with a large variety of organic compounds; however, the observed yield is variable. 23Notwithstanding, it is known that mannitol is one of the most efficient polyalcohols for this purpose and is usually applied in the volumetric determination of boric acid. 24ased on this finding, to minimize boron loss, NIST SRM 951a solutions were prepared at different boron concentrations (20, 30 and 100 µg L -1 ) with and without the addition of mannitol (0.2% final concentration), as proposed by Bruyn. 3 It should be noted that mannitol addition has the side effect of imparting a higher carbon content to the sample, showing the previously mentioned effect on the 11 B/ 10 B ratio.Therefore, the potential volatilization effect was verified by monitoring 10 B, while the 12 C influence was monitored based on the 11 B/ 10 B ratio.It was found that mannitol addition did not have the expected effect on the 10 B intensity but did enhance the 12 C influence on the 11 B/ 10 B ratio.As a consequence, it was decided not to use mannitol addition.

Boron concentration effect
Based on our previous experience with lead, 17 we decided to fix the boron concentration at approximately 20 µg L -1 .However, as boron has a lower ionization efficiency than lead and higher optimal concentration (100 µg L -1 ) was proposed by Bruyn, 3 Coetzee and Vanhaecke, 6 Vorster et al., 7 and Coetzee et al., 8 we decided to investigate the influence of boron content, in particular, taking into account the necessity to work with a boron concentration that is as low as possible in order to minimize the 12 C effect.It was observed that between 10-30 µg L -1 , despite of the existence of statistically valid differences among the isotopic ratio results, no clear trend on the 10 B/ 11 B ratio was observed, may be due to the lack of mass bias correction.However, on the other hand, at 100 µg L -1 , the concentration effect is noticeable (Table 3).

Application to red wine samples: boron content and isotopic ratio determination
To test the achieved accuracy and precision, ten measurements on the 20 µg L -1 NIST SRM 951a solution were performed and the mass bias correction was applied.The obtained mean value and standard deviation were 0.2436 and 0.0053, respectively, which is statistically equivalent (95%) to the certified value ( 10 B/ 11 B = 0.2473 ± 0.00024).The relative standard error is approximately 0.5%, which fits with the expected value for Q-ICP-MS.
Additionally, in order to test the accuracy, the boron isotopic ratio and boron content were determined for three Valpolicella wines bought in wine shops.The obtained values (Table 4) are in agreement with those published by Coetzee and Vanhaecke 6 (mean value = 4.143, SD = 0.002).
The obtained results for the Brazilian red wines are shown in Table S1, Supplementary Information.The boron concentration ranged between < 0.070 and 20.4 mg kg -1 .However, only a few brands presented values above 14 mg kg -1 , the maximum allowed concentration proposed by the OIV. 5 Four of them originated from the same winery (V6) but were produced from different vines.This finding may be related either to some local soil-specific characteristic or to a practice adopted by the winery.The other high concentration came from winery V7, where three wines were analyzed: two presented values below the detection limit and one presented 18.7 mg kg -1 .In fact, low boron concentrations are not expected for Vale dos Vinhedos wines.From a total of 36 samples, only three presented low values, two being from winery V7, which requires further study.In this direction, Vale dos Vinhedos is in fact a sub-region of Serra Gaúcha, and several Serra Gaúcha wines presented boron concentrations below the detection limit.The great majority of wines that presented concentration ranges below the limit of detection, mainly from the region of Serra Gaúcha, according to the information on the labels, are so-called "soft table" red wines.These wines are inferior in quality to those produced with Vitis vinifera grapes and are made from American red grapes, such as Isabel.Other wines from this region are fine red varietal wines made from Cabernet Sauvignon, Cabernet Franc, Merlot, Tannat and blends of these grapes.Some were produced by cooperatives, and most of them were from Bento Gonçalves/Vale dos Vinhedos.
The range of boron content in Brazilian wines was broader than those reported by Ozbek and Akman 25 in Turkish red wines (7.94 to 10.7 mg L -1 ).For Portuguese wines, Rodrigues et al. 26 were able to distinguish between wines from three wine regions, Alentejo, Dão and Bairrada, where Bairrada had the highest boron concentration (mean value of approximately 5.0 mg L -1 ).Hernandez et al. 27 studied the boron content in wines of the region of Tacoronte-Acentejo (DOC region) and obtained a mean value of 7.0 mg L -1 .
Among the three wine regions in Southern Brazil, Campanha Gaúcha presented the narrowest 11 B/ 10 B isotopic ratio range of 4.032-4.072.This reflects the fact that the analyzed wines from this region came from only two different wineries located in the same county.
As Vale dos Vinhedos represents a Serra Gaúcha sub-region, both presented similar 11 B/ 10 B isotopic ratio ranges of 3.857-4.094and 3.799-4.088,respectively.It was not possible to distinguish these three Southern Brazilian winery regions based only on the boron concentration and 11 B/ 10 B isotopic ratio, similar to the results observed when applying lead isotopes. 25lso similar to the reported results for lead isotopes, 17 the measured 11 B/ 10 B isotopic ratio of wines produced in the Vale do São Francisco region were distinguishably different from those produced in the Southern region.Thus, the 11 B/ 10 B isotopic ratio can be applied as a tool to differentiate these two Brazilian winery regions.Higher 11 B/ 10 B isotopic ratios of 4.110-5.152were observed, and treating this last value as an outlier, the range becomes 4.110-4.173.
It was observed that the low 11 B/ 10 B isotopic ratio values of Serra Gaúcha wines is related to the boron content being lower than 1 mg kg -1 , which can be connected to the 10 B peak measurement statistics since the samples were diluted 1:100.Therefore, in further evaluations, only values associated with a boron concentration higher than this value were considered.Figure 2, where the 11 B/ 10 B isotopic ratio is plotted against the boron concentration, was built based on this guideline.The figure shows that it is possible to apply the 11 B/ 10 B isotopic ratio as a tool to differentiate between wines produced in the Southern region from those produced in the São Francisco Valley.It is also possible to verify that the 11 B/ 10 B isotopic ratio does not depend on the boron concentration in the wine.This boron differentiation among South and Northeastern regions becomes more clear when using a box plot (Figure 3).Table S2 (Supplementary Information) presents the boron concentrations and 11 B/ 10 B isotopic ratios of wines from Portugal, Italy, Spain, France, USA, South Africa, Chile and Argentina.Two French and two Spanish

Figure 1 .Table 3 .
Figure 1.Sample dilution effect on the 11 B/ 10 B isotopic ratio in a red wine sample.

Figure 3 .
Figure 3. Box plot indicating the 11 B/ 10 B isotopic ratios of Brazilian red wine according to the producing region (VV = Vale dos Vinhedos, SG = Serra Gaúcha, VSF = Vale do São Francisco).

Figure 2 .
Figure 2. Variation in the 11 B/ 10 B isotopic ratios with boron concentrations for red wines produced in the Vale dos Vinhedos, Campanha Gaúcha, Serra Gaúcha and Vale do São Francisco regions.

Table 1 .
Operating conditions, data acquisition modes and other information for the determination of boron concentration and isotope ratios by ICP OES and ICP-MS, respectively

Table 2 .
Ethanol effect on the 11 B/ 10 B isotopic ratio

Table 4 .
Boron concentration and 11 B/ 10 B isotopic ratio on Italian red wine samples