The effect of ultrasound on Syrah wine composition as affected by the ripening or sanitary status of the grapes

Several studies have demonstrated that the application of ultrasounds to crushed grapes improves the chromatic and sensory characteristics of the resulting wines by facilitating the extractability of compounds from grapes to the must-wine. The objective of this work was to determine whether the application of ultrasounds to grapes of different maturity levels or different sanitary status leads to the same positive outcome as regards chromatic characteristics, phenolic and aroma compounds as well as sensory properties. The results showed that, independent of grape ripening or sanitary status, the application of ultrasounds to Syrah crushed grapes leads to wines with better chromatic characteristics at the moment of bottling, increasing colour intensity, total phenol content and tannin concentration up to 12%, 18% and 43%, respectively, in the wines from less ripen grapes and 13%, 23% and 30% in the wine from partially rotten grapes. The concentration of volatile compounds was less affected by sonication than the chromatic characteristics, with small decreases in the main families of volatile compounds, although the sensory analysis showed clear differences between control wines and those made from sonicated grapes, which were generally preferred and score higher for most of the sensory parameters evaluated.


Introduction
Several studies have demonstrated that the application of ultrasounds (US) to crushed grapes improves the chromatic characteristics of the resulting wines [1,2], increases their polysaccharide content [3,4] and may favour an increase in the concentration of some aroma compounds [5,6]. Grape skin cell walls are one of the main barriers in the extraction phenolics from grapes to must, while their extractability is aided by the erosion of the skin cell walls caused by ultrasound-generated cavitation bubbles.
It is also known that phenolic extractability increases naturally during grape ripening due to the degradation of the cell walls during the maturation process, cell walls being thicker and more rigid in the less ripe grapes [7]. The greater phenolic extraction observed in riper berries was demonstrated to be clearly associated with greater de-pectination and de-esterification of the grape berry cell walls [8]. Based on these observations, our first objective was to determine how the application of US to grapes of different degrees of maturity (and, therefore, with cell walls of differing composition and rigidity) affects the extractability of phenolics and other compounds of interest and, therefore, the phenolic composition, aroma and sensory properties of the resulting wines and whether this technique could be a useful tool for increasing the sensory quality of wines made from grapes with a low sugar content.
Moreover, sometimes, due to adverse climatic conditions during the late grape ripening process, microbial contamination may appear in the grapes, and the effect of US in this situation has not been discussed. The main organisms responsible for such microbial contamination are mainly fungi, the most common being Botrytis cinerea (grey mould), and others such as Aspergillus spp. and Penicillium spp. [9]. Other authors have stated that even low levels of infection can have a detrimental effect on wine quality [10]; indeed, this problem could be more pronounced in the case of red winemaking because of the longer skin contact time during maceration.
The presence of some rotten grapes may cause a number of microbiological and chemical problems during wine fermentation and may affect wine stability during storage and ageing. Some studies have shown that wines made from partially rotten grapes exhibit an altered colour resulting from the activity of oxidase enzymes, together with higher dry extract due to the formation of glycerine, polysaccharides, uronic acids and aldonic acids, and a substantially increased volatile acidity [11]. However, how the wine quality is affected may be linked to the extent of infection and it may exist a limit on the percentage of bunch infection over which the final wine quality is clearly affected [10,12]. Rotten grapes present a more degraded skin due to the enzymatic action of microbes, which, while facilitating phenolic extraction, may also be accompanied by the extraction of microbial enzymes capable of oxidating and degrading grape phenolic compounds, especially anthocyanins. Taking the above into account, our second objective was to test the effect of sonication of partially rotten grapes on the final wine quality, and whether the use of this technique is favourable or detrimental to wines made from batches that contain partially rotten grapes.

Grapes
Syrah red grapes were harvested from a commercial vineyard in Jumilla (Spain) and transported the same day to the winery for processing. The trials were carried out with grapes harvested at two different dates and therefore at two ripening levels (12 and 13.5º Baume). At the second harvest date, one lot of healthy grapes and one lot of grapes containing 25% of rotten fruit (determined by visual inspection) were harvested from the same vineyard and processed at the winery.

Winemaking (microvinifications)
Each lot of grapes (200 kg) was destemmed and crushed. The crushed grapes, healthy 12º and 13.5º Baume grapes, as were the 13.5º batches containing partially rotten grapes (SW12S, SW13.5S and SW13.5PR), were exposed to a pilot-scale power ultrasound system (MiniPerseo, Agrovin S.A., Alcazar de San Juan, Spain) operating at a frequency of 30 kHz. This system can treat 400 kg of crushed grapes per hour and operates at 2500 W with a power density of 8 W/cm2. This equipment is equipped with one sonoreactor, a hexagonal pipe (1 m length) in which the sonoplates are coupled and through which the crushed and destemmed grapes circulate, and is the place where the sonication occurs. The design of the equipment and the process limit the temperature increase to less than 2 •C. A batch of crushed grapes from each lot (healthy grapes of 12º and 13.5º Baume and partially rotten 13.5º Baume grapes) was not treated (control vinifications, CW12S, CW13.5S and CW13.5PR). Small stainless-steel tanks (10 kg capacity) were filled with the control and ultrasound-treated crushed grapes. Must homogeneity in each tank was achieved weighting separately the solid parts and the liquid and filling each 10 L vessels with the same quantity and proportion to assure the same solid/liquid ratio in each vessel. All vinifications were made in triplicate. Potassium metabisulfite was added to reach a final SO2 concentration of 70 mg/kg of grape. Total acidity was corrected to 5.5 g/L and selected yeasts were added (Viniferm CT007, acidity Agrovin, Spain, 20 g of dry yeast/100 kg of grapes). The fermentation was conducted at 22ºC and during maceration step (72 h), the cap was punched down twice a day. At the end of this period, the wines were pressed in a 75 L pneumatic press. Free-run and press wines were combined and stored at room temperature (22 ºC). After alcoholic fermentation, the wines were racked, cold-stabilised and bottled into 375 mL bottles and closed using corks. Must and wines were analysed immediately after the ultrasound treatment and two weeks after bottling.

Physico-chemical analysis
The musts were characterised just after crushing (control musts) and after sonicating (sonicated musts) and the wines at the moment of bottling.
The musts were characterised by measuring the soluble solid content, pH, and titratable acidity according to European Community methods [13]. Gluconic acid was determined by enzymatic method carried out via an automated analyser (Miura One, TDI, Spain). Wines were characterised by measuring the alcohol content, pH, dry extract, and titratable acidity [13]. The acetaldehyde, glycerol, glucose + fructose contents, acetic acid and gluconic acid were determined by enzymatic methods carried out by automated analyser (Miura One, TDI, Spain).

Spectrophotometric determinations
Colour intensity (CI) was calculated as the sum of absorbance at 620, 520 and 420 nm. Total and SO 2 non-decolorable anthocyanins were determined spectrophotometrically [14]. Total phenols (TP) were calculated by measuring wine absorbance at 280 nm [15]. MCPT (methyl cellulose precipitable tannins) were calculated using the methylcellulose precipitation method [16]. The tannin concentration expressed in mg/L was calculated and the ( −)-epicatechin was used as an external standard.

Analysis of tannins using the phloroglucinolysis method
Wine samples were prepared as detailed in the methodology that can be found in Busse-Valverde et al. [17]. Briefly, 5 mL of wine was evaporated in a centrivap concentrator (Labconco, USA), redissolved in 3 mL of water and then passed through a C18-SPE column (1 g, Waters, Milford, MA). The cartridge was washed with 20 mL of water, and compounds of interest were eluted with 10 mL of methanol, evaporated, and then dissolved in 1 mL of methanol. The analyses of proanthocyanidins were carried out by depolymerising the molecules using the phloroglucinol reagent. The depolymerised samples (10 μL injection volume) were analysed by HPLC. The elution conditions can also be found in Busse-Valverde et al. [17]. Proanthocyanidin cleavage products were estimated using their response factors relative to ( +)-catechin, which was used as the quantitative standard.

Isolation of wine and grape volatile compounds by SPME
For the isolation of volatile compounds by SPME, a 50/30-micron divinylbenzene-carboxen-polydimethylsiloxane (DVB/CAR/ PDMS) fibre was used and the volatile compounds were isolated and analysed by GC-MS following the methodology described by Gómez-Plaza et al. [18].

Sensory analysis
Descriptive sensory analysis: Wines were subjected to sensory discrimination testing using a descriptive and a triangle test. Prior to the sensory analysis, the wine from the three different replications for each experiment was pooled to provide a representative sample and to avoid differences among the replications. Ten students and staff members with experience in wine sensory analysis and interest in the project were selected. The assessors attended several training sessions to determine and define appropriate descriptors for formal assessment of the wines. During these sessions, standards for aroma attributes and basic tastes were presented.
Samples were presented to judges in 40 mL aliquots coded covered using standard wine glasses in maximum of three sets of two wines per session in a sensory room that was kept at 20 °C and free of unusual odours. Each panellist sat in an individual isolated booth illuminated with white light. The presentation order of wine samples within each tray was randomised across judges. Assessors were forced to take a 60 s break between wines. The intensity of each attribute was rated on a scale of zero to 10, with a score of zero indicating that the descriptor was not perceived. Data from all the judges for all samples were used in the analysis.
Triangular analysis: Judges were presented with three wine samples, two of which were identical, in coded standard wine glasses and in random order. Each assessor selected the sample they considered different (forced election). The statistical significance of the number of correct judgments versus the total number of judgments was subsequently determined. Panellists were also asked to indicate the preferred wine sample.

Statistical analysis
ANOVA and MANOVA analyses were conducted, using the Tukey test to find means that were significantly different from each other. Also, a principal component analysis was performed. All the statistical analyses were carried out using the statistical package Statgraphics Centurion XVI. Table 1 shows the main compositional data of the musts. The sugar content of the partially rotten grapes was slightly lower than that of the control healthy grapes. Barata et al. [12] stated that sour rot increased the Brix and glucose-fructose content for grape musts containing 30% of rotten grapes. Other authors [19] have found that the presence of up to 50% sour rot contributed to an increase in sugar content, due to grape dehydration, and Miele and Rizzon [20] also observed higher ºBrix in the must of grapes with ripe rot. However, our findings differed. It has been stated that the growth of some microorganisms such as Gluconobacter sp. and Acetobacter sp. and oxidative yeasts associated with sour rot may Table 1 Must composition CW 12, control must from grapes harvested at 12ºBaume; SW 12, must from grapes harvested at 12ºBaume and sonicated; CW 13.5S, control must from healthy grapes harvested at 13.5ºBaume; SW 13.5 S, must from healthy grapes harvested at 13.5ºBaume and sonicated; CW 13.5PR, control must from partially rotten grapes harvested at 13.5ºBaume; SW 13.5 PR, must from partially rotten grapes harvested at 13.5ºBaume and sonicated; TA titratable acidity, GA gluconic acid. Different letters within the same column indicate significant differences (p < 0.05) . On the other hand, sonication did not change the sugar content of the musts. Must pH did not change with maturity, ripening status or sanitary status although a slight increase was observed due to sonication. Titratable acidity decreased from 12º Baume grapes to 13.5ºBaume grapes, as expected. The highest acidity value was found in the must from partially rotten grapes. The increase in acidity observed in sonicated grapes could be due to the increased extraction of the intracellular content of the cells. However, pH did not decrease as might have been expected, due to the increase in titratable acidity, probably because the extraction of some ions, such as K + , was favoured, an effect described by Pérez-Porras et al. [2], which could lead to a higher salification of the acids.

Results and discussion
On the other hand, the higher titratable acidity value measured in the must from partially rotten grapes may have been due to the formation of other acids, such as gluconic acid, whose presence was confirmed. Gluconic acid was detected in partially rotten grapes as a result of the activity of oxidative yeasts and bacteria that typically colonise sour rotten berries [9,12,21]. Moreover, acetic acid bacteria could also have led to the accumulation of other sugar degradation products such as lactic, propionic and butyric acids [15], which would have affected the acidity values, although these acids were not measured.
Surprisingly, the gluconic acid content decreased following sonication. No clear explanation of this fact was found, although we may hypothesise that two phenomena may have occurred due to the sonication of the must. The first hypothesis is that the greater extraction of mineral compounds and metals that sonication may facilitate could have chelated part of the gluconic acid since it is known that gluconic acid is better than EDTA and many other common food grade chelators at binding with multivalent cations [22]. The other hypothesis is the possible oxidation of gluconic acid during the sonication process, since sonication may create reactive radical compounds that could affect the stability of the acid, favouring its degradation.
The physico-chemical composition of the studied wines is presented in Table 2. As expected, the alcohol content increased with ripening and was not affected by sonication. Only slight differences were observed in the pH values, although the wines made from both the 12º and 13.5ºBaume sonicated grapes presented higher pH and lower titratable acidity values, even after the latter being corrected in the must. The lower titratable acidity of these wines could be due to the higher precipitation of tartrate salts due to the increased presence of cations, especially Ca and K, as previously observed [2]. The highest value of titratable acidity was observed in the control wine made from partially rotten grapes.
The dry extract increased with sonication, as was expected, since sonication increases the extraction of cations, acids, and polysaccharides [2][3][4] and also in the wines from partially rotten grapes, a phenomenon previously described [11,12] and, therefore, the highest dry extract was observed in SW13.5S, CW13.5PR and SW 13.5PR wines.
Glycerol levels were also higher in the wines made from the more mature grapes and in the wine made with grapes that included rotten berries, as found previously [9,19,23]. The content was not affected by sonication in SW12 and decreased the level found in wines made from SW13.5S and SW13.5PR grapes. Sonication did not lead to a significant increase in acetaldehyde although the values were slightly higher in the wines from the most mature (both healthy and partially rotten) sonicated grapes. The concentration of acetic acid was within normal ranges, even in the wine made from partially rotten grapes, indicating that neither the percentage of rotten grapes used in the experience nor the application of sonication favoured the presence of this volatile acid.
In relation with these results, Barata et al. [12] also observed that wines from rotten grapes showed higher dry extract values, a greater reducing sugar content, and total and volatile acidity and reducing sugar residue could be used by spoilage microorganisms, such as those associated Table 2 Wine composition CW 12, control wine from grapes harvested at 12ºBaume; SW 12, wine from grapes harvested at 12ºBaume and sonicated; CW 13.5S, control wine from healthy grapes harvested at 13.5ºBaume; SW 13.5 S, wine from healthy grapes harvested at 13.5ºBaume and sonicated; CW 13.5PR, control wine from partially rotten grapes harvested at 13.5ºBaume; SW 13.5 PR, wine from partially rotten grapes harvested at 13.5ºBaume and sonicated; %Alc alcohol percentage, TA titratable acidity, DE dry extract, Acetald acetaldehyde, G + F glucose + fructose; AA acetic acid, GA gluconic acid. Different letters within the same column indicate significant differences (p < 0.05) with sour rot, or lactic bacteria and alter wine chemical and sensory parameters. These results only partially coincide with those of this study since we found that all the wines presented reducing sugar values lower than 1 g/L, the highest value being detected in the wine made from partially rotten grapes (0.71 g/L, which is below the 2 g/L value that indicate the end of alcoholic fermentation) and sonication further decreased this value. Gluconic acid is hardly metabolised by fermentation yeasts but it can be metabolised by some microorganisms, one possible consequence being an increase in wine volatile acidity, as described for a S. pombe mutant [11]. In our experiment, although its concentration fell during fermentation, no increase in acetic acid, the main contributor to volatile acidity, was detected.
The turbidity of the wine samples and the quantities of formed lees were also evaluated ( Table 3). The lees volume was higher in the wines made from sonicated grapes than in their control counterparts, which lends weight to the hypothesis that sonication induces greater generation of suspended grape material due to the more intense degradation of the grape cell walls. As regard turbidity, both the wines made from the less ripe grapes presented low turbidity, and no differences due to the application of ultrasound were observed. Sonication decreased the turbidity of the wines made from the healthy mature grapes but increased the turbidity of the wine made from sonicated partially rotten grapes, whose control wine showed low turbidity. The decrease in turbidity observed in SW13.5S, which reached values similar to those of the wines made from the less ripe grapes may have been due to the easier precipitation of the suspended plant material. Sonication facilitated tannin extraction from the grapes, while it is known that tannins present high affinity for the polysaccharides and proteins arising from mesocarp disaggregation, promoting greater formation of high-molecular weight insoluble complexes that precipitate in the wine [24][25][26], decreasing its turbidity.
In the case of wines made with partially rotten grapes, sonication increased both the lees and turbidity, perhaps due to the presence of glucans. One important oenological issue concerning grape infection is the difficulty of clarifying the resulting wine. The fungi responsible for bunch rot often produce high-molecular weight polysaccharides that cause turbidity and processing difficulties [9]. These polysaccharides include [β1 − 3] and [β1 − 6 glucans], which serve to protect the pathogen against host-defence mechanisms, but are also the products of the plant cell wall-degrading enzymes produced by the fungi during infection. Table 4 shows the chromatic and phenolic composition of the different wines. As can be seen, the colour intensity, total phenols and the concentration of total anthocyanins and tannins were higher in the control wines made from the more mature grapes than in the wines made from less ripe grapes. No differences were observed for the wines made from the partially rotten grapes, except that the polymeric anthocyanin and tannin values increased. Table 3 Volume of lees and wine turbidity measured at the end of alcoholic fermentation CW 12, control wine from grapes harvested at 12ºBaume; SW 12, wine from grapes harvested at 12ºBaume and sonicated; CW 13.5S, control wine from healthy grapes harvested at 13.5ºBaume; SW 13.5 S, wine from healthy grapes harvested at 13.5ºBaume and sonicated; CW 13.5PR, control wine from partially rotten grapes harvested at 13.5ºBaume; SW 13.5 PR, wine from partially rotten grapes harvested at 13.5ºBaume and sonicated. Different letters within the same column indicate significant differences (p < 0.05) Several authors have described a loss of phenolic compounds in wines made from rotten grapes due to the action of microbial oxidase enzymes, especially those of anthocyanins [9,10]. However, in agreement with our results, Barata et al. [12] found that wines made with grapes presenting sour rot have a higher level of anthocyanins and total phenols, and a more intense colour as a consequence of the higher skin to free juice ratio. Indeed, the total phenol content increased by about 13 and 19% when the percentage of rotten grapes was 30 and 50%, respectively. Deytieux-Belleau et al. [27] reported that the skin tissues of rotten grapes are degraded by fungal extracellular enzymatic activities and that these activities might enhance phenolic release, notably from the grape skin, leading to higher extraction during vinification.
Sonication promoted an increase in colour intensity, especially in the wine made from less mature grapes, although the increase in the wine made from sonicated healthy 13.5ºBaume grapes was not significant compared with its control. These differences were more accused at the moment of pressing and at the end of alcoholic fermentation (Supplementary Table 1) but the differences slightly decreased at the moment of bottling. The increase in colour intensity reached 16%, 8.1% and 14% in wines made from sonicated 12º Baume, 13.5º Baume and 13.5º Baume partially rotten grapes. The increase in total phenol values due to sonication was significant in any of the cases. The concentration of total anthocyanins decreased in the wines made from sonicated healthy grapes, due to reactivity of monomeric anthocyanins with others wine components like tannins, phenolic acids and other small molecules such as pyruvic acid and acetaldehyde increasing the concentration of those non-decolourable by SO 2 , although at the moment of pressing it could be clearly seen how sonication improved the extraction of anthocyanins in all the samples.
Tannins (MCPT) were the most favoured compounds by sonication and extraction was most intense in the wines made from partially rotten grapes. Ky et al. [10] also reported better extraction of phenol compounds during maceration and the winemaking process when the grapes used contained rotten berries, because of the disruption of the skin cell contents induced by the above mentioned fungal extracellular enzymatic activities. This already disrupted skin helps to increase the effect of sonication, in a similar way to that observed when pectolytic enzymes and sonication were applied simultaneously [18].
The same increase was observed when the concentration of the tannins that can be depolymerised by phloroglucinol was measured ( Table 5). The mean degree of polymerisation of these tannins decreased in all the sonicated samples, accompanied by an increase in their galloylation percentage. It has been suggested that sonication also affects the seed structure and favours the extraction of the tannins they contain. Seed tannins undergo a lower degree of polymerisation and higher degree of galloylation than skin tannins, which would explain the changes observed in the structure of the tannins from wines made with sonicated grapes. Also, the precipitation of high-molecular weight structures should be considered.
The extractability of high-molecular weight skin tannins increases as maturation progresses [29]. In the case of the control wines, the degree of tannin polymerisation was highest in the wine from the most mature grapes, confirming these previous results concerning the increased extraction of high-molecular weight skin tannins. This was especially true in the wines from partially rotten grapes, due to the already degraded skin. Other authors [10] described a decrease in polymeric proanthocyanidins in wines made from botrytised grapes as shown by the significantly lower level of mDP in the tannin polymeric fraction, although these authors followed a different methodology of analysis than this used in the work.
To obtain a clearer picture of the effect of the degree of grape maturity on the outcome of sonication, a multivariate analysis of variance was performed, and interaction graphs were constructed (see supplementary Tables 2-7 and supplementary Fig. 1-Fig. 6). The p value for all the sonication vs maturation level interactions were higher than 0.05, except for colour intensity and total anthocyanins, indicating that, for Syrah grapes, sonication was effective for extracting phenolic compounds and for improving the resulting wine's chromatic characteristics, and that this effect was, in general, not dependent on the maturity status of the grapes. Only colour intensity was more favoured by sonication of the less ripe grapes, while total anthocyanin levels decreased more sharply in the wine from the most mature sonicated grapes, since polymerisation was favoured. However, our experience with different grape varieties indicates that the Table 5 Concentration (mg/L) and composition of wine tannins determined by phloroglucinolysis CW 12, control wine from grapes harvested at 12ºBaume; SW 12, wine from grapes harvested at 12ºBaume and sonicated; CW 13.5S, control wine from healthy grapes harvested at 13.5ºBaume; SW 13.5 S, wine from healthy grapes harvested at 13.5ºBaume and sonicated; CW 13.5PR, control wine from partially rotten grapes harvested at 13.5ºBaume; SW 13.5 PR, wine from partially rotten grapes harvested at 13.5ºBaume and sonicated; mDP mean degree of polymerisation, EGC epigallocatechin, ECG epicatechin gallate. Different letters within the same column indicate significant differences (p < 0.05)

Samples
Tannins ( results observed in this study may be variety-related since a similar study in Monastrell grapes indicated that the sonication was significantly more effective in extracting grape phenolic compounds from mature grapes than from less ripe grapes [29]. The same analysis was conducted comparing the effect of sonication in healthy or partially rotten grapes (sonication vs sanitary status of the grapes, see supplementary Tables 8-13 and supplementary Fig. 7 to Fig. 12) and the results showed that the interactions were significant for most of the chromatic characteristics and phenolic compounds (except polymeric anthocyanins and MCPT). This suggests that the effect of sonication is more evident in partially rotten grapes, increasing colour intensity and total phenols more than when ultrasounds were applied to healthy grapes. Although the extraction of tannins from grapes is usually strongly favoured by sonication, no differences were observed in the effect of sonication on MCPT between healthy and partially rotten grapes, probably due to the already disrupted grape skins in the partially rotten grapes.
The results of the semiquantitative analysis of the main families of volatile compounds for the control wines and wines made from sonicated grapes, at the two different levels of maturity and sanitary status are shown in Fig. 1. The identified compounds were grouped into higher alcohols, monoterpenes and norisoprenoids, esters, and fatty acids.
The results pointed to only small differences in the main volatile compounds. The concentration of higher alcohols slightly decreased due to sonication in SW12 but was not significantly affected by sonication in the wines made from 13.5º Baume grapes, both healthy and partially rotten.
As regards ester compounds, it is well known that esters contribute to the aroma of young wines, being responsible for floral and fruity odours. Although no significant differences were observed in the sum of these compounds, the quantities of esters were slightly higher in the control than in the sonicated wines.
Fatty acid production is associated with the initial must composition and the fermentation conditions. Their concentration was higher in SW12 and SW13.5S (although not statistically different from the control wine) and very similar in CW13.5PR and SW13.5PR, although contrary to that, the formation of some fatty acids due to fungal and bacterial infection has been reported [15].
Terpenes and norisoprenoids are considered to be closely related to the variety and are important for the expression of varietal characteristics in wine, due to their low olfactory threshold [30,31]. Wines from sonicated grapes, especially SW13.5S showed slightly higher concentrations of these compounds than the control wines but slightly decreased in the wine made with sonicated partially rotten grapes. Other authors have found that sonication may promote the liberation of the bound aromatic compounds present in the must, and hence increase their concentration in wines [5]. Oliver Simancas et al. [6] also studied the effect of US on the content of varietal volatile compounds (free and glycosidically bound) in musts and on the overall aroma of wines, finding that sonication led to an increase in the concentration of free varietal compounds such as C6 alcohols, terpenes and norisoprenoids in musts and also in wines made from sonicated grapes. However, in general, we found no relevant differences in the concentration of wine volatile compounds due to sonication or due to the presence of rotten grapes. With regards to the effect of grape microbial contamination, other authors have indicated that sour rot influences secondary metabolism of yeast, decreasing the levels of some volatile compounds, especially those due to amino acid metabolism [12] and a decrease of geraniol, nerol and linalool concentrations has been observed in sour rotten Riesling grapes [21].
To increase our knowledge as regards how the sonication and/or the presence of partially rotten grapes affects all the studied variables and the similarity of the wines, a multivariate statistical analyses (a principal component analysis) was conducted (Fig. 2). Multivariate statistics allows the simultaneous analysis of more than one variable and may help to understand how the samples relate to each other.
To determine how the samples were grouped and which variables were those that promoted the separation of the samples, a principal component analysis (PCA) was conducted. As expected, all the wines made from sonicated grapes were grouped together and were located in the positive part of PC1 (which explained 47% of variance), due to their higher values of polymeric anthocyanins, total tannins, the percentage of tannin galloylation and total phenols. Control wines, located in the negative part of PC1, presented higher loads in aromatic compounds, tannins with higher mDP and higher content of prodelphinidins and higher total anthocyanin content. PC2 separated the samples according to maturity and sanitary status, explaining 40% of total variance, and in a similar way in control and sonicated wines. Maturity increased all the measured parameters in the wines and the same was observed for the wines made from partially rotten grapes.
The results described above point to wide differences in the phenolic compound content and chromatic characteristics of wines following the use of ultrasounds in the winery, while variations in wines volatile compounds were not so evident. However, sensory analysis is perhaps the most important tool for evaluating whether the chemical changes measured in wines due to ultrasounds have any effect on the appreciation of a wine. The sensorial analysis results are shown in Fig. 3, in which each control wine is compared with its respective wine made with sonicated grapes. The results showed that the wines scored quite similar although, in general, sonication improved the scores for most of the evaluated attributes. When comparing wines from the less ripe grapes (12ºBaume), significant differences in colour intensity and the fruity aroma scores were found, the values being higher in the wine made from sonicated grapes. Both wines made from mature healthy grapes were also very similar; in this case, no significant differences could be found in any attribute.
In wines made from partially rotten grapes, both the control wine and that made from sonicated grapes were scored highly, indicating that they were well accepted by the panellists. Sonication led to wines that differed in colour intensity, aroma quality and mouthfeel intensity, the scores being higher in the wine made from sonicated grapes. The results of the triangular analysis (Table 6) showed that sonicated wines from heathy grapes, at both maturity levels, could be distinguished from their respective controls and, when panellist were forced to select, wines from sonicated grapes were preferred. Oliver Simancas et al. [6], working with Monastrell wines from sonicated grapes, found that all the wines made from sonicated grapes had better scores in the olfactory attributes with respect to the control wines. On the other hand, panellist could not distinguish the wines from partially rotten grapes, so sonication cannot be considered detrimental for these wines.
The results of the sensorial analysis demonstrated that the fermentation of grape musts containing up to 25% of grapes with sour rot provided wines with similar scores to those of wines made with healthy grapes. Furthermore, sonication led to an improvement in wine quality in the case of healthy grapes and had no detrimental effect on wines made from partially rotten grapes. The generation of earthy offodours as consequence of bunch rot has been described [32,33], but this problem was not detected in our wines. Other authors [12] also performed a sensory evaluation of wines produced with different sour rot percentages, finding that 40% rot could be regarded as the threshold above which wines become depreciated. Different results were observed by Ky et al. [10] who found that sensory analyses underlined a loss of wine sensory quality that was perceptible from a threshold as low as 5% of Botrytis-affected grapes.

Conclusions
When Syrah grapes, at two different maturation levels, were sonicated and vinified, the results showed that wines made from sonicated grapes improved their chromatic and sensory characteristics and were preferred by the members of a sensory analysis panel. Sonication of the less ripe grapes significantly improved wine phenolic and sensory characteristics. This is an important issue because it shows that the sonication of grapes with a less than optimum ripening level may help to overcome the lack of phenolic maturity in these grapes.
When the effect of applying ultrasound to partially rotten grapes was studied to determine whether the wine quality is affected, the results showed that even in this case, the application of US was not detrimental for wine quality and neither wine (one made from control grapes and the other from sonicated grapes) could be differentiated in a triangular sensory analysis. These results confirm the interest of applying ultrasounds in oenology to improve the extraction of compounds of interest and the quality of the final wines. Funding Open Access funding provided thanks to the CRUE-CSIC agreement with Springer Nature. This research was funded by the Ministerio de Ciencia, Innovación y Universidades from the Spanish Government and Feder Funds, grant number RTI2018-093869-B-C21.

Data availability
The data that support the findings of this study are available from the corresponding author upon reasonable request.

Conflict of interest
The authors declare no conflict of interest. The funders had no role in the design of the study; in the collection, analyses, or interpretation of data; in the writing of the manuscript; or in the decision to publish the results. This manuscript is original and has not been submitted to, nor is under review at, another journal, whether in part or in its entirety. We assign the copyright and the consent for publication to European Food Research and Technology.

Compliance with ethics requirements This research does not involve human or animal tests.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http:// creat iveco mmons. org/ licen ses/ by/4. 0/. Table 6 Results of the triangular sensory analysis CW 12, control wine from grapes harvested at 12ºBaume; SW 12, wine from grapes harvested at 12ºBaume and sonicated; CW 13.5S, control wine from healthy grapes harvested at 13.5ºBaume; SW 13.5 S, wine from healthy grapes harvested at 13.5ºBaume and sonicated; CW 13.5PR, control wine from partially rotten grapes harvested at 13.5ºBaume; SW 13.5 PR, wine from partially rotten grapes harvested at 13