Summary

Recent advances in destemmer technology have reduced the amount of non-grape material entering fermentation bins. However, stem jacks are usually still present at some level. Post-destemming sorting of jack stems as a way to limit green character and harsh stem tannins is a common practice in Virginia wineries. However, post-destemming sorting is labor intensive and time consuming. The purpose of this experiment was to determine if the time and effort spent sorting jack stems made a perceptible difference in overall wine quality. Experiments were conducted using both Cabernet Franc and Petit Verdot. For each variety, one TBin of fruit was fermented from grapes that were sorted post-destemming while a second TBin of fruit was fermented from grapes that were not sorted. General wine chemistry was very similar between treatments for both varieties. In each, the unsorted fruit produced wine that had slightly higher tannins. The Cabernet Franc wines were not significantly different in a triangle test. The Petit Verdot wines were distinguishable in a triangle test, however there were no differences in sensory scores for fruit intensity, fruit character, herbaceous/green character, bitterness or astringency. 

 

Introduction

During his service as the Virginia Tech Enology Extension Specialist, Bruce Zoecklein made many contributions to quality wine production in Virginia. Many of the winemaking techniques and standard protocols at Virginia wineries are based on his advice during this time, and his Enology Notes remain one of the best resources for quality wine production in Virginia and beyond. 

In accordance with Zoecklein’s advice, many Virginia wineries employ post-destemming sorting to remove jack stems as standard protocol. In Enology Notes #1071, Zoecklein cautions that in the dry 2005 vintage, the crop may suffer abrupt stops in Brix increases due to low moisture reserves, over-cropping, or insufficiently developed root systems. He states that this may lead to immature cap stem phenols evidenced by difficult to pick berries, and resulting in immature skin and seed tannins. Among other recommendations, he includes this caution:

Use care in destemming. A high concentration of immature cap stems suggests the need for post-destemming sorting. This may be a requirement for consistent premium red wines in this region, and is always a good idea. This may be essential if the red must contains a high concentration of immature cap stems or jacks¹.

In Enology Notes #1172, while discussing phenolic compounds in red wine processing, he said:

We know that prefermentation sorting is important with regard to phenolic composition of wines, as is gentle fruit handling. …An area traditionally overlooked is post-destemming sorting to remove cap stems or jacks. Stem tannins are chemically different from skin tannins, and impart a different sensory profile. Stem tannins, including cap stem tannins, are more astringent and harsher than skin tannins. Jack stems in the fermentor can be a problem, if the concentration is high and the stems are green or not lignified, resulting in increased tannin intensity and astringency².

With these recommendations in mind, all Vitis vinifera red grapes get sorted at Fabbioli Cellars as part of an SOP developed decades ago. However, sorting can be time consuming and expensive especially when considering the opportunity cost of labor during harvest. Several aspects of production flow have been put in place to ensure that the best quality fruit enters the winery. Fruit is hand harvested with care, pre-sorting in the vineyard to avoid damaged or diseased fruit. A full time crew is paid by the hour rather than by the lug in order to encourage quality picks. When fruit comes into the winery, it is destemmed and crushed, then sorted by 4-6 people before it is deposited in TBins for fermentation. The primary aim of sorting at this stage is to remove any stems or “funk” that gets through. 

In recent years, Doug Fabbioli has begun to question if recent advances in vineyard management in Virginia and improvement in destemmer technology may have diminished the need for post-destemming sorting, allowing for better production efficiency without a loss in quality. The purpose of this experiment was to compare wine quality made from fruit that was sorted post-destemming from that which was not sorted. The experiment was initially planned for Cabernet Franc because of the potential green character that might come from the jacks. During harvest, many red lots were picked on the same day in anticipation of a rain event, leading to a backlog of lots to be sorted. This presented an additional opportunity for experimentation with Petit Verdot.

 

Methods

Fruit was hand harvested, chilled overnight, then destemmed and lightly crushed. One 1.5 Ton bin per variety was filled without sorting. One 1.5 Ton bin per variety was sorted post-destemming prior to placement in the fermentation vessel. Sorting was done by a group of 4-6 people. Each bin received equal weight of fruit. All other cellar operations were the same between bins of the same variety, including additions, inoculations, pressing times, and barrel character. Oak chips (1 lb/ton), 25 ppm SO² and 75 mL/ton Scottzyme Color Pro were added at the time of destemming. Juice samples were taken for analysis immediately after processing. The following day, Cabernet Franc was inoculated with 25 g/hL CSM yeast rehydrated in 30 g/hL GoFerm. Petit Verdot was inoculated with 25 g/hL D80 yeast rehydrated in 30 g/hL GoFerm. Fermentations were co-inoculated for malolactic fermentation with VP41 one day after yeast inoculation. Cabernet Franc was chaptalized to raise the sugar by 1° Brix after the start of fermentation (approximately 3 days post-destemming). At the completion of malolactic fermentation, 1.2 g/L tartaric acid was added to Cabernet Franc and 1.9 g/L was added to Petit Verdot. All barrels were treated with 50 ppm SO² without racking.

Sensory analysis was completed by a panel of 18 wine producers for Cabernet Franc and 17 wine producers for Petit Verdot. Wines were presented blind in randomly numbered glasses. Tasters were presented with three wines, two of one type and one of another, and asked to identify which wine was different (a triangle test). To account for order effects, there were four tasting groups with the unique wine in the triangle test balanced between groups. Tasters were then asked to score each wine on a scale of 0 to 10 for fruit intensity, fruit character, herbaceous/green character, bitterness and astringency. They were also given open ended questions to describe the wines. Results for the triangle test were analyzed using a one-tailed Z test. Descriptive scores were analyzed using repeated measures ANOVA. Lastly, tasters were given 5 different tangible fabrics (in order from softest to roughest; soft suede, suede, velvet, sandpaper and burlap) and were asked to select the fabric that most closely represented the astringency of the wine. Fabrics were chosen to include an increasingly rough tactile perception. 

 

Results

Additional effort for sorting

For the experiment, each treatment was destemmed and crushed into 1.5 ton fermenters. The sorted fruit was destemmed to a plastic slide between the destemmer and TBin. As fruit moved over the slide, members of the crew picked out jacks and any other MOG. Approximately one lug of stem material was removed per lot. Sorting in this way required approximately 30 extra minutes per harvest bin, with up to 6 people. When fruit was not sorted, the work was approximately 5 times faster.

Cabernet Franc

Cabernet Franc was harvested with relatively low Brix (Table 1), indicating immature fruit, which, according to Zoecklein’s guidelines, made it a good candidate for sorting out jack stems. General chemistry of the must at harvest (Table 1) as well as post-fermentation chemistry (Table 2) were very similar between treatments. Slightly higher Brix in the fermentation of unsorted fruit led to slightly higher ethanol in the resulting wine. The wine made from unsorted fruit also had slightly higher anthocyanins and tannins (Table 3, Figure 1). Higher tannin profiles might be a consequence of  higher ethanol, as much as from higher stem load in the fermenter.

In a triangle test of sorted vs. unsorted Cabernet Franc, 7 out of 18 respondents were able to distinguish which wine was different, indicating the wines were not significantly different (Z=0.25, p= 0.40). Those able to distinguish the wines scored the wine from unsorted fruit higher for fruit character than wine from sorted fruit. Fruit character was defined as a range from bright/fresh/red fruit (low scores) to dark/dried/black fruit (high scores), meaning the unsorted fruit produced wine with darker fruit character. This wine also received significantly higher scores for astringency. There were no significant differences in scores for the remaining descriptors (fruit intensity, herbaceous/green, and bitterness) (Table 6). When asked to compare the astringency of wine to 5 different tactile fabrics, wine made from sorted fruit was categorized to indicate softer tannins while wine made from unsorted fruit was categorized with descriptors consistent with coarser texture (Figure 3). 

Petit Verdot

Petit Verdot fruit was harvested just before a rain event, at the same time as many other lots, so juice chemistry was not taken. The finished chemistry of the wines is very similar between treatments (Table 4), with ethanol values indicating this, too, may have been relatively low Brix (potentially underripe) fruit. Samples were taken from Barrel 1 for sensory analysis to minimize differences in free SO² between barrels. The wine made from unsorted fruit had higher tannin but lower anthocyanins than the wine made from sorted fruit (Table 5, Figure 2).

In a triangle test of Petit Verdot made from sorted vs. not sorted fruit, 15 out of 17 respondents were able to distinguish which wine was different, indicating the wines were significantly different (Z=4.54, p< 0.001). However, there were no significant differences in scores for any of the specific descriptors (Table 7). When asked to compare the astringency of wine to 5 different tactile fabrics, both wines were most often categorized with velvet texture (Figure 4). 

 

---

References

(1) Zoecklein, B. Red Fruit Maturity. Enology Notes 2005, 107.

(2) Zoecklein, B. Tannins, Color and the 2006 Season, Continued: Phenolic Compounds and Red Wine Processing. Enology Notes 2006, 117.

Table 1: Juice chemistry for Cabernet Franc (October 3, 2023)(Vinterra)

 

Table 2: Wine chemistry of Cabernet Franc wine made from sorted and not sorted fruit (ICV Labs, March 2023)

 

Table 3: Phenolic composition of Cabernet Franc wine made from sorted and not sorted fruit (mg/L) (ETS Labs, March 2023)

 

Table 4: Wine chemistry of Petit Verdot wine made from sorted and not sorted fruit (ICV Labs, March 2023)

 

Table 5: Phenolic composition of Petit Verdot wine made from sorted and not sorted fruit (mg/L) (ETS Labs, March 2023)

 

Table 6: Mean descriptive scores for five attributes of sorted vs. not sorted Cabernet Franc

 

Table 7: Mean descriptive scores for five attributes of sorted vs. not sorted Petit Verdot

Figure 1: Comparison of phenolic measurements for Cabernet Franc (ETS Labs)

 

Figure 2: Comparison of phenolic measurements for Petit Verdot (ETS Labs)

 

Figure 3: Fabric selected by respondents to represent astringency of each Cab Franc treatment

 

Figure 4: Fabric selected by respondents to represent astringency of each Petit Verdot treatment

 

Summary

Recent advances in destemmer technology have reduced the amount of non-grape material entering fermentation bins. However, stem jacks are usually still present at some level. The winemaking goals at Blenheim are to produce fruit-forward red wines that bottle within 12 months of harvest and are consumed within 1-2 years post bottling. The standard protocol at Blenheim has been to sort out jack stems post destemming to minimize harsh tannins and bitterness entering the wine. However, destemming is labor intensive and time consuming. The purpose of this experiment was to determine if time and effort spent sorting jack stems make a perceptible difference in overall wine quality. Experiments were conducted using both Cabernet Franc and Petit Verdot. For each, one TBin of fruit was fermented from grapes that were sorted post-destemming while a second TBin of fruit was fermented from grapes that were not sorted. There was a slight increase in pH, acetic acid, and tannin but no change in methoxypyrazine in the finished wine when Cabernet Franc grapes were not sorted. The wines were not significantly different in a triangle test. Those that could tell the difference between the wines rated the wine made from sorted fruit as having higher astringency while there were no differences in bitterness ratings. The wine made from unsorted Petit Verdot also had higher pH than the sorted treatment, with no differences in the other chemical measures. The wines were not significantly different in a triangle test. Those that could tell the difference between the wines rated the wine made from sorted fruit as having higher fruit intensity with no other differences in descriptors. In this experiment, sorting fruit came at the cost of considerably slower destemmer speed (2 tons/hour vs. 5 tons/hour) as well as labor (3-5 people at the sorting table). Each winemaker must decide for themself if these differences are worth the cost.

 

Introduction

The winemaking goal for the red wine program at Blenheim Vineyards is to produce fruit forward, early bottling wines that are usually consumed within 1-2 years of release. To this end, all red grapes are destemmed but not crushed before fermentation to remove the potential for bitterness, harsh astringency, and vegetal character that may be imparted by stems. 

In 2021, Blenheim purchased a new destemmer. One of the main criteria for this purchase was to maximize the efficiency of stem removal and minimize the amount of stem jacks entering the fermentation bins. Several characteristics of destemmers are aimed at maximizing wine quality, including material (rubber beaters vs. metal beaters), size of basket perforations, rotational direction of the beaters vs. the basket (opposite is more efficient but more prone to stem and berry damage), and speed of rotation^1–3. In a 2021 experiment testing if destemmer speed affected wine quality in Cabernet Franc and Petit Verdot, wines produced by fast destemming could not be distinguished from those produced by slow destemming in a triangle test and there were no significant differences in sensory scores for astringency, bitterness, or fruitiness (link to experiment here). 

Discussion of the destemmer trial during the sensory session included a debate surrounding the impact of jacks on wine quality. Specifically, does the amount of jacks produced by normal destemming of relatively healthy fruit make a difference in wine quality? At Blenheim, harsh tannins or bitterness if introduced by jacks at harvest do not have long aging time to soften. Will these attributes decrease the quality of the wine if not sorted out? 

The purpose of this experiment was to test if post-destemming sorting of jack stems makes a perceptible difference in wine quality. The experiment included both Cabernet Franc and Petit Verdot because each variety presents unique challenges. A primary concern with Cabernet Franc is its tendency toward “green character”. Stems contain hexenal and other C6 compounds that might lead to a “leafy” perception. Stems can also contain methoxypyrazine, which can contribute to the perception of green bell pepper⁴. Petit Verdot tends to have leggier stems that sometimes get tangled in the destemmer and produces more jacks than Cab Franc, with potentially more effect on wine quality. Ripeness may also be an important factor in how many jacks are produced, with riper fruit destemming more easily and producing fewer jacks. A companion trial was run at Fabbioli Cellars (link to study here). 

 

I think this will be the only way that I will be able to sleep soundly at night thinking about all of the jacks that end up in fermentations! (Kirsty Harmon)

 

Methods

For each variety, fruit was harvested and processed with the same operations and additions between treatments with the sole exception of post-destemming sorting. Sorting was done after fruit was destemmed, prior to fruit entrance into TBins. All fruit was destemmed onto the sorting table. Control fruit traveled down the sorting table without any MOG removal. For treatment lots, sorting was done by 3 (CF) - 5 (PV) people. The amount of material removed was weighed after sorting was complete.

Cabernet Franc

Fruit was hand harvested on 9/7/23 from the Deer Ridge Vineyard (Monticello AVA), cooled overnight then destemmed but not crushed into TBins with the addition of 50 mg/L SO₂ (added as 35 g KMBS). Fruit in each bin was inoculated for fermentation with 15 g/hL (70 g) EC1118 yeast. Bins were punched down twice daily with no other additions. Bins were pressed at the completion of alcoholic fermentation on 10/3, then racked off gross lees into barrels on 10/4 and allowed to go through malolactic fermentation. There were two barrels per treatment. Malolactic fermentation was completed on 10/31. On 11/22, an acid trial was completed and 1.25 g/L tartaric acid was added to each barrel along with 100 mg/L SO₂. 

Sensory analysis was completed by a panel of 18 wine producers. Wines were presented blind in randomly numbered glasses. Tasters were presented with three wines, two of one type and one of another, and asked to identify which wine was different (a triangle test). To account for order effects, there were four tasting groups with the unique wine in the triangle test balanced between groups. Tasters were then asked to score each wine on a scale of 0 to 10 for fruit intensity, fruit character, herbaceous/green character, bitterness and astringency. Fruit character was defined as a range from bright/fresh/red (low scores) to dark/dried/black (high scores). They were also given open ended questions to describe the wines. Results for the triangle test were analyzed using a one-tailed Z test. Descriptive scores were analyzed using repeated measures ANOVA. Lastly, to assess differences in tannin texture, tasters were given 5 different tangible fabrics (in order from softest to roughest; soft suede, suede, velvet, sandpaper and burlap) and were asked to select the fabric that most closely represented the texture of the wine. Fabrics were chosen to include an increasingly rough tactile perception. 

Petit Verdot

Fruit was hand harvested on 9/21/23 from the Claim Vineyard (Monticello AVA), cooled overnight then destemmed but not crushed into TBins with the addition of 50 mg/L SO₂ (added as 35 g KMBS). Fruit in each bin was inoculated for fermentation with 15 g/hL (70 g) EC1118 yeast. Bins were punched down twice daily with no other additions. Bins were pressed at the completion of alcoholic fermentation on 10/11, then racked off gross lees into barrels on 10/13 and allowed to go through malolactic fermentation. There were two barrels per treatment. Malolactic fermentation was completed on 11/7. On 11/28, an acid trial was completed, but no acid addition was needed. Wine was treated with 100 mg/L SO₂ on 12/2.

Sensory analysis was completed by a panel of 17 wine producers. Wines were presented blind in randomly numbered glasses. Tasters were presented with three wines, two of one type and one of another, and asked to identify which wine was different (a triangle test). To account for order effects, there were four tasting groups with the unique wine in the triangle test balanced between groups. Tasters were then asked to score each wine on a scale of 0 to 10 for fruit intensity, fruit character, herbaceous/green character, bitterness and astringency. Fruit character was defined as a range from bright/fresh/red (low scores) to dark/dried/black (high scores). They were also given open ended questions to describe the wines. Results for the triangle test were analyzed using a one-tailed Z test. Descriptive scores were analyzed using repeated measures ANOVA. Lastly, to assess differences in tannin texture, tasters were given 5 different tangible fabrics (in order from softest to roughest; soft suede, suede, velvet, sandpaper and burlap) and were asked to select the fabric that most closely represented the texture of the wine. Fabrics were chosen to include an increasingly rough tactile perception. 

 

Results

Additional effort for sorting

Fruit was destemmed at a speed of 2 tons/hour onto the sorting table, which conveyed fruit into TBins for fermentation. When fruit is not sorted, the destemmer is usually set to 5 tons/hour. Regardless of treatment, all fruit was destemmed at the slower speed onto the table but jacks were removed from the “sorted” treatment only. For each of the “sorted” treatments, less than one pound of material was removed by the end of the run. Figure 1 shows a comparison of must from each treatment. Sorting did not remove all imperfections, but there were noticeably more jacks in the TBin containing unsorted fruit than the TBin containing sorted fruit.

Cabernet Franc

Cabernet Franc was harvested with balanced chemistry, indicating good ripeness with little difference in fruit between treatment and control (Table 1). There was also very similar general chemistry in the finished wines (Table 2). The fermentation with unsorted fruit completed with slightly higher acetic acid and alcohol levels, however these are small differences, based on a single fermentation only, that may have been caused by factors other than sorting. Notably, there was no difference in methoxypyrazine (IBMP) between treatments. The threshold of detection for IBMP in wine is estimated to be 6-15 ng/L, so these values were all likely below threshold. There were no consistent differences in anthocyanins between fermentations with sorted and unsorted fruit (Table 4, Figure 2). The fermentation with sorted fruit produced wine with a lower concentration of tannin than the fermentation with unsorted fruit. This effect may have been due to slightly higher alcohol bin with unsorted fruit extracting slightly more tannin.

In a triangle test of Cabernet Franc with and without jack inclusions, 8 out of 18 respondents were able to distinguish which wine was different, indicating the wines were not significantly different (Z= 0.75, p= 0.23). There were no significant differences in scores for fruit intensity, fruit character, herbaceous/green character, or bitterness (Table 5). The mean scores for astringency were nearly significantly higher for the wine made from sorted fruit vs. unsorted fruit. When asked to compare the astringency of wine to 5 different tactile fabrics, responses were dispersed between fabrics. The highest number (3 tasters) selected sandpaper to describe the astringency of the unsorted fruit, and suede to describe the astringency of wine made from sorted fruit (Figure 3). Overall, wine made from sorted fruit was rated as having slightly coarser texture.

Petit Verdot

Petit Verdot fruit chemistry was also very similar between treatment and control bins (Table 1), with balanced and ripe chemistry. Finished wine chemistry was also very similar between treatments (Table 3). The fermentation with unsorted fruit had noticeably higher pH than the sorted fruit. Potassium is found in stems, which may lead to higher pH values. A change in potassium as small as 10% can lead to a shift of up to 0.1 pH units⁵. The wine made from sorted fruit had slightly higher concentration of total anthocyanins but there was no difference in tannin concentration between wines (Table 4). 

In a triangle test of Petit Verdot with and without jack inclusions, 6 out of 17 respondents were able to distinguish which wine was different, indicating the wines were not significantly different (Z=-0.09, p= >0.50). There were no significant differences in scores for any of the descriptors (Table 6), though scores for fruit intensity of wine from unsorted fruit were nearly significantly higher than those for wine from sorted fruit. When asked to compare the astringency of wine to 5 different tactile fabrics, the overall perception of coarseness was nearly the same though there was higher variation in responses for wine made from sorted fruit (Figure 4). 

---

References

(1) Christmann, M.; Freund, M. Chapter 15 Advances in Grape Processing Equipment. In Managing Wine Quality Volume 1: Viticulture and wine quality; CRC Press, Woodhead Publishing Limited: New York, 2010.

(2) Catania, P.; Pasquale, C. D.; Morello, G.; Vallone, M. Influence of Grape Transport and Destemming Systems on the Quality of Chardonnay Wines. 2016, 8.

(3) Phillips, C. Product Review: Destemmers. Wine Business Monthly 2005, June, 1.

(4) Blackford, M.; Comby, M.; Zeng, L.; Dienes-Nagy, Á.; Bourdin, G.; Lorenzini, F.; Bach, B. A Review on Stems Composition and Their Impact on Wine Quality. Molecules 2021, 26 (5), 1240. 

(5) Boulton, R. The General Relationship Between Potassium, Sodium, and PH in Grape Juice and Wine. American Journal of Enology and Viticulture 1980, 31 (2), 5.

 

Figure 1: Comparison of must from sorted and not sorted treatments.

 

Table 1: Juice chemistry for Cabernet Franc and Petit Verdot with and without sorting (Imbibe Solutions)

 

Table 2: Wine Chemistry for Cabernet Franc with and without sorting (ICV Labs, March 2023)

 

Table 3: Wine Chemistry for Petit Verdot with and without sorting (ICV Labs, March 2023)

 

Table 4: Phenolic compounds found in Cabernet Franc and Petit Verdot with and without sorting (mg/L) (ETS Labs, March 2023)

 

Figure 1: Comparison of anthocyanins and tannins for two treatments of Cabernet Franc (ETS Labs March 2023) 

 

Table 5:  Mean descriptive scores for five attributes of sorted vs. not sorted Cabernet Franc

 

Figure 3: Fabric selected by respondents to represent astringency of each Cab Franc treatment

 

Table 6: Mean descriptive scores for five attributes of sorted vs. not sorted Petit Verdot

 

Figure 4: Fabric selected by respondents to represent astringency of each Petit Verdot treatment