Figure 1. Temperature profiles by product, dosage and storage temperature.

Figure 2. Changes in major oenological parameters between the baseline and 31 days of contact time.

Figure 3. Changes in phenolic chemistry between the baseline and 31 days of contact time.

Color

Figure 4. Evolution of wine color and hue throughout the experiment.

Temperature had a clear effect on wine color, with higher temperatures (70°F and 90°F) producing increased color intensity. This effect can be primarily attributed to the loss of free SO₂ at elevated temperatures, which enhances visible color rather than reflecting true phenolic changes. Despite these shifts in intensity, the hue remained remarkably stable across all treatments, indicating that temperature and dosage had minimal influence on overall color balance.

Oak Aromatics Chemistry

Figure 5. Concentration profiles of eight selected oak-derived volatiles at 16 and 31 days across the applied temperature conditions.

Product Contrast
The two oak chip profiles exhibited clearly differentiated extraction patterns. Contact with High Mocha oak chips produced comparatively higher levels of furfural and vanillin, resulting in a more toasted-vanilla aromatic profile, whereas contact with Pure 2 Vanilla emphasized lactone-driven sweetness. Because oak lactones are typically more abundant in American oak (Quercus alba)—the species used to produce Pure 2 Vanilla—its stronger lactone expression and rapid equilibrium are consistent with expectations. Furthermore, it was noteworthy the chemical analysis remained consistent with the expected sensory profile for each oak product, which are both designed to offer specific and repeatable sensory results.

Extraction of Volatiles
The control wine behaved as expected: in the absence of oak-chip addition,
oak-derived compound levels remained stable throughout the contact period. Concerning the experimental treatments involving oak addition, the results showed that most compounds reached their maximum extraction after 16 days of wine–chip contact, regardless of dosage or temperature. Vanillin exhibited only a minor increase between days 16 and 31; extraction continued but at a much slower rate. By day 16, vanillin concentrations were already above its sensory threshold in red wine
(0.32 mg/L), indicating that vanilla notes can be enhanced within a short timeframe. Guaiacols are known to be extracted predominantly during the early stages of the process. Structurally, vanillin and guaiacol share similar basic chemistry functional groups, explaining their comparable extraction kinetics. As illustrated in Figure 6, vanillin includes an aldehyde group; removal of the double-bonded oxygen and hydrogen yields the guaiacol structure.

Figure 6. Comparison of molecular structures for vanillin and guaiacol.

In contrast, furfural and furaneol continued to increase beyond 16 days, indicating that longer contact times favor the extraction of these molecules. Furanic compounds may exhibit distinct extraction dynamics compared to other oak-derived volatiles, likely related to their formation during toasting and their distribution within the wood structure. From a practical standpoint, when a pronounced vanilla note is desired, limiting contact to roughly 16 days can help prevent masking by furanic compounds; conversely, extended contact can be used strategically to develop greater toasted complexity.

Throughout the study, as illustrated in Figure 5, the dosage consistently proved to be the primary factor influencing extraction, while temperature exhibited a lesser and compound-specific effect. To further quantify these relationships, a Partial Least Squares (PLS) model was applied to evaluate the individual contributions of contact time, temperature and dosage.

Ranking of Experimental Parameters
PLS (Partial Least Squares regression) is a multivariate statistical method used to model the relationship between multiple input variables (e.g., temperature, dosage rate, contact time) and multiple response variables (e.g., concentrations of volatile compounds). In this study, PLS helped identify which experimental factors had the greatest impact on compound extraction. The Variable Importance in the Direction (VID) values, derived from the PLS model, quantifies the relative influence of each factor.

Figure 7. Mean Variable Importance in the Direction (VID) across 14 compounds, by experimental factors.

Analysis of the VID values calculated for 14 oak-derived volatile compounds provided a clear ranking of the influence of the three experimental parameters: temperature, dosage rate and contact time, as shown in Figure 7. Dosage rate was the most influential factor, with VID values above 0.8 for most compounds, confirming its central role in modulating concentration levels. Temperature showed a moderate but meaningful influence, ranking second, and should not be overlooked when optimizing the rapid extraction of desired oak-related compounds. Finally, although contact time contributes to extraction, its effect was more limited under the tested conditions, consistent with the kinetics of certain compounds that plateau shortly after initial exposure.

This objective ranking highlighted the key levers for optimizing oak extraction, underscoring the primary role of dosage while reinforcing that thermal effects, often underestimated, deserve closer attention.

Table 2. Variable Importance in the Direction (VID) values for each compound and experimental factor in High Mocha and Pure 2 Vanilla profiles.

Color gradient: warm colors (yellow to red) indicate positive VID values while green shades reflect negative values. Increased color intensity represents increased magnitude for positive or negative VID values.

The analysis of VID values in Table 2 revealed notable differences in how each experimental factor influences the extraction of oak-derived compounds in the two oak chip profiles studied: High Mocha and Pure 2 Vanilla.

For High Mocha, the dosage was the most influential parameter overall, with a strong average VID (0.74), followed by temperature (0.25). The contact time showed a slightly negative influence (–0.05), suggesting that extending extraction duration (after 31 days) does not significantly improve compound release in this profile.

For Pure 2 Vanilla, the pattern was similar but less marked: dosage remained the main driver (VID = 0.66), while temperature had a moderate role (VID = 0.11). Interestingly, contact time showed a small positive average VID (0.02), suggesting a slightly more progressive extraction profile compared to High Mocha.

These findings confirm that oak chip dosage was the primary driver of compound extraction, while temperature, though often underestimated, plays a non-negligible and variable role depending on the product.

The analysis of VID values by compound revealed varying levels of sensitivity to tank temperature across chemical families, reflecting its role in modulating volatile compound extraction. Compounds naturally present in oak, such as eugenol and both cis- and trans-lactones, exhibited the highest sensitivity to temperature, as indicated by consistently elevated VID values in both product matrices.

Lignin-derived compounds—namely vanillin and guaiacol—showed a moderate influence of temperature, with VID values around 0.2 to 0.3.

Among hemicellulose degradation products, 5-hydroxymethylfurfural (5-HMF) showed relatively high VID values (above 0.4), particularly in the Pure 2 Vanilla matrix, suggesting a significant contribution of temperature to the extraction of
this volatile.

In contrast, furaneol, maltol and furfuryl alcohol showed negative or very low VID values for temperature—most notably in the Pure 2 Vanilla set—indicating an inverse or negligible association with temperature under the tested conditions.