Exploring Enhanced Phenolic Extraction Methods

Exploring Enhanced Phenolic Extraction Methods with Alex Remy of Atlas Wine Co.

Alex Remy from Atlas Wine Co., our long-term client and collaborator, took advantage of our new phenolics panel this November to better understand his extraction trials this year.

Alex has recently moved into a new facility and has a gentile destemmer with no crusher and no mono pump. After macrobins are filled with destemmed fruit they are loaded into open-tops with a dumping forklift. With the mostly intact berries, Alex was concerned with inadequate early maceration of fruit and looking to see if Saignee or Petite Sirah co-fermentation could be helpful in getting the color and tannin extraction he was looking for.

Alex says:

“I’m all for these “soft” destemming technologies these days that allow for you to have a high level of control over what makes it into the tank, we’ve found they can really help us reduce our extraction of astringent seed tannin, but it does mean that you need to re-adjust parts of your process to get the full benefit.”  

With the new Tastry Phenolics panel, we looked at the extraction of a few of his cab lots: 

Table 1. Sample Name, Harvest and Fermentation Metrics

Name Abbreviation Harvest date Brix at HarvestDays on Skins 
23-CS-E79-Saignee (15%)SiG-1510/5/2326.314
23-CS-E79-Petite Sirah (i.e. Durif) (6%) PS-610/5/2326.314

The three lots all came from the same vineyards. The PS-6 and SiG-15 were harvested 25 days before the CtR/Late (see table 1). After picking, CtR/Late was adjusted to 23.5 Brix. No fermentation oak products were added and no sulfites were added prior to or during fermentation. 1g/L of tartaric acid was added to all lots at approximately 20°Bx. All fermentations were supplemented with Go-Ferm PE, Fermaid O, and DAP with YAN goals of 300 mg/L. Fermentations were run with S. Cerevisiae LALVIN ICV D254®. These were production-scale fermentations of 4 tons each. Punch downs were done twice a day. Prior to fermentation, 15% of the tank’s volume in juice was removed from SiG-15, the Saignee lot. For the PS-6 lot 6% by weight Petit Sirah (i.e. Durif) was added to the fermenter.

Basic Chemistry 

Table 2. Basic Chemistry of the Trial Wines

At the time of submitting samples to Tastry all samples had finished primary fermentation and had not detectable Residual Sugar (RS.) Malic acid for all of the samples was around 1 g/L with low corresponding lactic acid. 

Looking at the final ethanol of the CtR/Late it appears that the brix adjustment (water addition) was overshot. Adjusting brix in a fermentation with some amount of shrivel can be a challenge as grapes rehydrate as they begin fermentation. Despite the high brix adjustment in CtR/Late, TA and pH were similar to the other two lots, hex color, extracted as a function of A420 A520 A620 measurements, show a significantly lighter color for the CTR-late wine. 


Across the three lots, PS-6 was measured with the highest phenolic values, followed by the SiG-15, and then the CtR/Late. With the higher % alcohol of the PS-6, higher tannins than the SiG-15 should be expected, but the addition of the likely higher colored PS, high anthocyanin was extracted as well. The lower polyphenols in CtR/Late likely come as a combination of lower alcohol (leading to lower tannin extraction) and the high degree of brix adjustment. As is often the case with commercial-scale trials happening in the thick of harvest, removing confounding variables can be a challenge.

Table 3. Phenolics Data for the Three Experimental Lots 

Analyte  CtR/LatePS-6SiG-15
Polyphenols (mg/L)141017001530
A280 (TPI)43.461.355.7
A320 (a.u.)18.42724.5
Monomeric Anthocyanins (mg/L)390578513
Polymeric Anthocyanins (mg/L)375649
Tannin (mg/L)365564493
Polymeric Anthocyanins / Tannin Ratio0.10.0990.1
Catechin (mg/L)16.62119.2
Catechin / Tannin Ratio0.0460.0370.039
Gallic Acid (mg/L)
Total  Anthocyanins (mg/L)426634563

Petit Sirah berries are medium in size while Cabernet Sauvignon berries are small. The juice-to-skin ratio of the PS-6 lot is expected to be the highest of all three lots. The PS-6 also has the highest TPI, anthocyanin, and tannin content. While higher % alcohol may have led to enhanced extraction of tannin from in this lot, the higher extraction in all other phenolics categories suggest the Petite Sirah’s inclusion in fermentation may be primarily responsible for these higher values.

Early Consumer Sentiment

At Tastry, we are still in our early days in Applying our AI algorithms to understand how the consumer sentiment of a wine changes during secondary fermentation. However, we thought it might be interesting to take a look at how consumers are predicted to perceive these wines directly after harvest.

The two key predicted consumer sentiment metrics that we produce at Tastry are: 

  1. Consumer Sentiment Performance (CSP) – CSP is the percentage of US drinking-age consumers who are predicted to find a wine at minimum appealing to their palates. We produce CSP metrics by using AI to connect our wine chemistry database with our consumer palate preference database. 
  1. Predicted 5-Star Rating (P5S) – A predicted 5-star rating is a means to replicate consumer ratings on the 5-star rating scale. Our P5S is predictive of real-world average consumer 5-star ratings to within ±0.10 stars. The AI that produces P5S is trained by connecting our wine chemistry database to hundreds of thousands of real-world consumer ratings of commercial wines. Price heavily influences P5S for wine, generally as wines increase in price P5S increases. 

Table 4. Consumer Sentiment Performance (CSP) and Predicted 5-Star Rating (P5S) of Three Experimental Wines  

SampleCSP (%)P5S  (@$20)

Table 4. shows an inverse relationship between CSP and P5S when comparing the CtR/Late sample with the enhanced extraction. Where it appears that more US consumers will find CtR/Late appealing than the other two samples, these same consumers are not predicted to give the wine as high a 5-star rating as the heavily extracted wines. Inversely, fewer people are expected to find the SiG-15 and PS-6 wines appealing, but they are expected to perform better with consumer 5-Star Ratings. 

At Tastry, we see this type of inverse relationship often. The chemistry that allows for a wide audience for a wine is often not the same chemistry that elicits high consumer ratings. More consumers are predicted to find the more ripe flavors and lower extraction of the CtR/Late wine more palatable than the PS-6 and the Sig-15, However, they would be more likely to give the higher extractions and fresher flavors of the PS-6 and the Sig-15 higher scores. 

Red wines change a lot between the end of primary fermentation, the end of secondary, and their release. We will be tracking the progress of these samples with further analysis through secondary fermentation and to provide visibility on the degree of associated changes to predicted consumer appeal. 

Palate Space

From the point of view of consumer appeal, how similar are these wines?

Figure 1. Palate Space Venn Diagram of the Three R&D Samples: PS-6, SiG-15, CtR/Late

Our palate space Venn diagram (see fig. 1) allows us to see the relative overlap in CSP (consumer appeal) across these wines. In the Venn diagram, the size of each circle corresponds to the size of the US population who is expected to find this wine at minimum appealing. Where the circles overlap you can see the percentage of customers will find both or all three wines appealing.

When it comes to consumer appeal, the consumers who like the enhanced extractions of PS-6 and SiG-15 appear to overlap strongly with 2.1% of consumers liking both wines strongly (1.01% + 1.09%= 2.1%) and only 1.47% and 0.9% of consumers liking only PS-6 and SiG-15 individually -not taking into consideration the appeal of CtR/Late. 

Unsurprisingly, there is more overlap in consumers who find both CtR/Late and SiG-15 appealing (1.77%) than in consumers who find both CtR/Late and PS-6 appealing (1.09 + 0.24 = 1.33%). The co-fermentation of a different variety in PS-6 (Petite Sirah) and all of its different and unique phenolics and aroma compounds may be responsible for the larger difference that we see.


From the same vineyard and the same fruit, three wines were produced with significant differences in predicted consumer sentiment. Enhanced extractions receive a lot of buzz promising winemakers the opportunity to get more out of their fruit; Overall, enhanced extraction methods did appear to result in more, tannic and more highly colored wines in this trial. With the heavy-handed brix adjustment and later harvest, in CtR/Late, it’s challenging to say if saignée and co-fermentation are helpful methods specifically on overcoming challenges related to decreased crushing and pre-fermentation maceration. Multiple extraction levels and ripeness levels will no doubt be beneficial come blending season. 

Our data shows that directly after harvest, enhanced extractions can have a lower appeal than a standard extraction and be appealing to significantly different consumers. This makes sense as often enhanced extractions are not released as individual lots blended strategically into larger lots to optimize the phenolic profile of each wine. Even if these smaller lots are not attractive to a larger audience they can be helpful tools in blending. 

Next March we will re-analyze these wines and run a Tastry CompuBlend®: our in-house blend recommender. With the predicted consumer sentiment data of the three wines (which we think may change come March), Tastry CompuBlend will recommend 6 of the most highly diversified blends expected to commend the highest 5-star rating and CSP. 

Do you have winery trials this year that you want a second opinion on? Send us your R&D trials from this year. Get in touch by sending a quick note to hello@tastry.com.

Share the Post: