3.1. Fermentation
The fermentation stage in the three different batches occurred at various time intervals. The first batch underwent a total of 20 days of fermentation before distillation. The extended time was due to the fermentation not progressing correctly during the first week, as the yeast used (Fermentis SafSpirit C-70, Lille, France) requires temperature conditions between 30 °C and 40 °C. The average ambient temperature in the laboratory where the fermentation tanks were located fluctuated between 15 °C to 20 °C. Therefore, the fermentation of the first batch occurred slowly during the first week because it did not have the correct thermal conditioning, as the Propagate Pro heating pad presented technical issues that were not resolved until the first week had passed. The tank was covered with aluminum, and the fixed PropagatePro heating pad was used at 40 °C for 6 days after initial inoculation and sealing. No re-inoculation was carried out; only the tank conditions were adapted. Once 48 h had passed under these conditions, the internal temperature of the tank, monitored with the thermocouple of the heating pad, reached 30 °C and oscillated between 30 °C and 35 °C until the end of fermentation. This temperature increase had direct implications on the fermentation, as can be seen in
Figure 3, where the slope for the graphs of SC, alcohol per volume (% ABV), and specific gravity became more pronounced around days 5–7 of the fermentation process when the heating pad was included. Figures depicting the variables monitored for each batch are point-to-point data representations, with each sampling point explicitly marked. This method was chosen to accurately reflect the discrete nature of the data collected at specific intervals. By connecting each data point directly, a clear visualization of the changes in specific gravity, alcohol by volume, pH, and Brix is shown over time. Even in the pH graph, similar behavior can be observed in terms of the slope, which, in this case, was a negative slope, given that the metabolic activity of the yeast had a noticeable acceleration. Organic acids are formed from the oxidation of aldehydes, mainly acetic acid [
31], in the sub-reactions present in the fermentation processes. Consequently, it is normal for the pH to decrease, and the pH decreased more when the tank was thermally treated. The %ABV obtained for this first batch was 9.54% ABV out of a total of 21 L of fermented must, a value within the expected range given the capacity of the yeast used.
The second batch underwent a longer fermentation process, lasting 34 days, facilitated by the yeast Oenoferm Freddo, which operates optimally between 13 and 17 °C. This temperature range aligns with the laboratory’s average room temperature, ensuring ideal fermentation conditions. The process was allowed to extend because the goal was to obtain the highest amount of alcohol possible. Because the constant monitoring showed that the SC continued to decrease considerably, the process was continued. Since the pre-fermentation must was not filtered to outline the differences between the first batches, the initial SC was higher than the first. Additionally, as refined sugar was used as an adjustment instrument for SC in the must, the initial value for this parameter was higher in this batch than the rest (29 °Bx). For this reason, it was decided to take advantage of the highest possible quantity of available sugars for fermentation. The pH in this batch considerably decreased during the first seven days, as seen in
Figure 4. After this time, the pH remained between 3.3 and 3.5. A re-inoculation could have been possible here because the pH presented slight increases that would have been controlled if re-inoculated. However, the fermentation process was not interrupted, and once the SC stopped decreasing, the fermentation was stopped entirely using potassium metabisulfite (0.3 g/L). The %ABV obtained according to the calculations made from the specific gravities corrected for standard temperature was 14% for a total of 21 L of fermented must. Using the difference between initial and final SC readings, this value is 17%ABV.
The third batch was the smallest, using 8 kg of processed and liquefied yacon root without adding water to concentrate the number of sugars available for fermentation. The yeast used, namely SaffSpirit HG-1, does not require thermal conditioning; its best temperature range is 25 °C to 35 °C; however, fermentation can occur without problems at lower temperatures (15 °C to 25 °C) with slower kinetics. If it were decided to use the PropagatePro heating pad, fermentation would occur at temperatures higher than the ambient temperature of the laboratory (15 °C to 20 °C). It is possible that a lower yield and alcohol production could have been obtained according to the manufacturer’s specifications (Fermentis LeSaffre, Lille, France). The fermentation behavior was rapid, as shown in
Figure 5. The SC decreased substantially during the first week of fermentation, and the alcohol production per volume was higher. The batch started with 12.2 °Bx; by day 10, the refractometer reading was 4.8 °Bx. On day 12 of fermentation, it was decided to re-inoculate the same amount of yeast and nutrients as at the start of the fermentation process to use the most significant quantity of sugars available from the yacon. However, this re-inoculation did not obtain the expected results because the yeast had reached its maximum alcohol production capacity in the tank, and there was no evidence of further alcohol production, so fermentation was stopped with potassium metabisulfite. Due to the high specific density of this batch, no readings of the controlled variables were taken using the hydrometer, as the aim was to preserve the most significant amount of must before filtration for alcohol production and subsequent distillation. According to the readings obtained and corrections for temperature and specific gravity, the must has 4.2% ABV, but when calculated with the difference in SC during fermentation, this value was 6.8% ABV. The batch was filtered twice through cheesecloth filters to prepare it for distillation, avoiding sediments that could impair the alcohol obtained. The information in
Table 2 states that the main sugars found in yacon root are fructose, glucose, and sucrose, as this batch was not adjusted using additives, such as refined sugars or pure dextrose.
A comparison between density, the weight of the must, theoretical ethanol volume, and fermentation yield for the three batches is shown below in
Table 4.
The fermentation yield results, as presented above in
Table 5, highlight the efficiency of the fermentation process for each batch. The density of the must was calculated using Equation (2). These densities indicate the concentration of soluble solids, which are critical for determining the potential alcohol yield. The weight calculation was essential for scaling the fermentation process accurately. The theoretical ethanol volume was calculated from the stoichiometric conversion of sugars (dextrose for Batches 1 and 3 and sucrose for Batch 2) using Equations (1) and (2). The actual fermentation yields were 45.3% for Batch 1, 64.2% for Batch 2, and 79.4% for Batch 3, as calculated using Equation (4). These results suggest that Batch 3, which relied solely on the sugars naturally present in yacon without additional dextrose, was the most efficient, possibly due to optimal hydrolysis and fermentation conditions. In contrast, Batch 1 had the lowest yield, possibly due to filtration losses and less efficient fermentation dynamics. The high yield in Batch 2 indicates the effective utilization of the added sucrose. These findings underscore the importance of optimizing each process step, from hydrolysis to fermentation, to maximize ethanol production.
3.2. Distillation
Distillation for Batch 1 was carried out in two sessions. On the first day of distillation, the heads and hearts were obtained up to a cut of 60% ABV. On the second day, the remaining heart cuts and the tails of the distillate were obtained up to a cut of 15% ABV, where the process was stopped due to sour aromas and off-flavors of the cut. For this distillation, 20 L of must at 9.5% ABV was used, according to temperature corrections made to the specific gravity samples collected during the fermentation stage. Based on this, 1.9 L of alcohol at 100% ABV would have been obtained in a distillation where all cuts were recovered. However, the calculation based on the data obtained from the collection of SC samples during fermentation showed that the ferment had obtained 11.1% ABV at the time of distillation, considering the following formula: (initial SC − final SC) × 0.9. Thus, the total amount of alcohol at 100% ABV obtained from this batch would have been 2.34 L. When starting the distillation, there were problems with setting up the distiller as the cooling system had leaks, so the head cuts were obtained after 4 h of heating the must to around 90 °C on the thermocouple of the distiller’s pot. Exactly 2% of the heads were set aside due to methanol, which was approximately 420 mL. Next, cuts between 80% ABV and 65% ABV were obtained. The must was stored in airtight conditions for later continuation. In the second distillation, the first cut obtained was at 65% ABV; from there, the ABV of the distillate decreased drastically. Again, a last heart cut was made with a global %ABV of 45%. Finally, nearly 400 mL of tails were obtained. There was a leak in the distillate output duct in both distillations, with minimal dripping, which was stored during the distillation. The final volume was 100 mL after 10 h of distillation at 60% ABV. It was decided to store it as it had no unpleasant odor or taste and could be used for the final product. The distillation efficiency was calculated by determining the amount of alcohol at 100% ABV recovered in each cut, divided by the expected amount based on the specific gravity corrections of the samples. The distillation efficiency obtained was 79.5%. Details can be seen in
Table 6 and
Figure 6.
The separated cuts received an aroma and flavor analysis to determine which ones would be used in the aging and final product-obtaining process. It was established that the cuts used for the liquor would be the seven cuts made from the hearts, so the calculation was made to determine the amount of water that would be necessary to dilute it to 55%ABV, as shown in
Table 6, and then introduce it into the white oak barrel, as shown in
Figure 7. A sample of 100 mL was set aside for later comparison by sommeliers with the aged sample. After a month, the barrel was opened and emptied, and the liquid inside had a mild change in its ABV%; it was 55% ABV at the start of the aging process, and it ended up at 50%ABV, which is expected that some of the alcohol evaporated. Using the mean diameter for the barrel, volume and area were calculated, which led to finding the value for the specific surface. The theoretical value of the aging time on wooden barrels with various geometries and known specific surfaces is calculated using a base value of specific surface for 700 L Brazilian barrels [
32], and the aging time obtained is 73.3 days. Afterward, the distillate was diluted with bottled water down to 40% ABV and bottled in 750 mL tinted glass bottles.
For the second distillation, approximately 6 L of must have to be discarded; in addition to the filtration process before starting the distillation, 2 to 3 L of must were wasted because the capacity of the distiller is 20 L. Unlike the first batch, a continuous process took place, as it was not desired to split the distillation process between various days. This batch was expected to obtain 14% ABV, meaning 2.8 L of alcohol at 100% ABV would be expected to be retrieved when recovering all the distillation cuts. If the SC readings were to be considered, this value would be 17% ABV; thus, theoretically, 3.4 L of pure alcohol should have been recovered. This time, there were no problems in setting up the cooling system of the distiller, and the first drop of heads was received 3 h after starting the distillation. A total of 1.5% of the heads was collected, which amounts to 300 mL at 85% ABV. The distillation continued by obtaining the hearts from 84% ABV to 50%. At this point, a cut was made after 10 h of distillation, and during the last 2 h, tails were obtained starting at 40% ABV. Distillation was stopped shortly after due to the aromatic and taste profiles of the tails obtained, which were not the desired ones for the final product since they had an earthy aroma, a greasy consistency, and high turbidity.
Furthermore, this decision to stop distillation was linked to a surprisingly high number of hearts having been retrieved, which was much more helpful for dilution, aging, and later bottling. However, as seen in
Table 7 and
Figure 8, this decision was not detrimental to the distillation efficiency since, in the case of this second batch, most of the alcohol content was extracted in the heart cuts, which included a 1016 mL cut at 75% ABV. Like in the first batch, there was a liquid escape issue at the outlet where the alcohol hydrometer was placed; this ‘cut’ was previously named residues. In this case, the residues were 140 mL after 12 h of continuous distillation. These residues had 70% ABV. The distillation efficiency was calculated as in the previous batch; for this batch, 2.64 L of alcohol at 100% ABV was retrieved from the 2.8 L of alcohol at 100% ABV expected, based on the must fermentation calculations. As with Batch 1, a sensory evaluation was conducted, especially on the tails cut and the residues, to determine if they would be used in the final product. Ultimately, it was decided not to include both cuts and to use only the hearts of the distillation (see
Table 8) for subsequent dilution and bottling. Unlike Batch 1, the hearts were diluted to 50% ABV to rest after being combined for aging since the barrel was occupied with the first batch. During the dilution process, a cut belonging to the heart of the distillation was measured with the hydrometer, and due to a handling error, the cut was lost. Nevertheless, this cut was still considered for distillation efficiency. After one month of aging, the distillate was diluted with bottled water to 40% ABV and placed in transparent glass bottles of 750 mL.
The distillation of the third batch consisted of two successive distillations, although differently from the first batch, as the process was interrupted and later continued for the latter. For Batch 3, a first distillation was carried out using the 20 L distiller. However, since there was only 5.6 L of must, 5 L of water was added to avoid an excessive contact surface between the initial small amount of must and the heater. In addition, distilling with a small quantity of must in the distiller is generally avoided; a minimum requirement of half its capacity should be used to prevent excessive gas flow inside and, thus, avoid accidents. In this case, 5.6 L of must (excluding the water) was distilled at 4.2% ABV according to measurements made by adjusting SC to find the specific gravity, as hydrometer readings were not taken. However, based on the refractometer measurements, the %ABV calculated from the difference between the initial and final SC was 6.8% ABV. The first distillation was continuous, and no cuts were made as the aim was to obtain all the alcohol in the must. The distillation lasted 5 h, the minimum run time compared to the other two batches. A total of 1550 mL of distillate at 20% ABV was obtained, which is 310 mL of pure alcohol at 100% ABV. The distillation efficiency in this first distillation can be calculated as follows:
The amount of alcohol expected in this batch was calculated based on the theoretical %ABV computed using the difference in SC, not the correction for specific gravity. This is because the %ABV calculated from specific gravity gave an expected alcohol amount lower than what was obtained in practice, making the distillation efficiency exceed 100%, which is illogical. For the second distillation, the 2 L distiller was used, and unlike the first distillation, it was necessary to make the pertinent cuts to obtain the desired product. A total of 1% of the distilled volume was extracted as part of the head cut, which amounts to 15.5 mL at 80% ABV. Next, four heart cuts were made from the alcohol obtained at between 80% ABV and 50% ABV, and 370 mL were recovered within this ABV range. Finally, the tails were cut from 35% ABV until the distillation was finished. A total of 210 mL of tail cuts were retrieved in this second distillate. The distillation efficiency achieved was the highest among the three batches because most of the alcohol was recovered in the second distillation; however, the first distillation obtained an acceptable efficiency.
Similarly, by performing two complete distillations separately, the distillation process for the second run is facilitated, as it is easier to accelerate the process and recover all the alcohol in the 2 L distiller, as shown by these efficiency calculations. The details of this second distillation for the third batch can be seen in
Table 9 and
Figure 9. Regarding the sensory evaluation to determine which cuts were most suitable for the final product, it was decided to include only three cuts of hearts (see
Table 10) since the tail cuts and the fourth lowest cut from the hearts (50% ABV) had unpleasant aromas and flavors, as well as a whitish appearance, indicating turbidity. These selected cuts amounted to 310 mL at 70% ABV after being mixed; afterward, they underwent dilution with bottled water to bring the distillate down to 50% ABV. The distilled product was aerated for one week by tightly sealing the sample with a coffee filter, allowing the distillate to breathe and settle. It typically remained sealed for two more weeks until a final dilution with bottled water was made to bring it to 40% ABV so that it could be bottled.
3.3. Sensory Evaluation and Organoleptic Analysis
The tasting sheet used by the sommeliers can be found in
Appendix A as
Figure A1. Regarding the sensory testing conditions, the evaluations adhered to ISO standards, utilizing specifically designed glasses to enhance the olfactory experience, which is essential for assessing the aroma and overall quality of spirits. The testing environment was meticulously controlled for lighting, temperature, and odors to eliminate any external influences on the assessments. To minimize olfactory fatigue, each sommelier was provided adequate spacing between sessions and palate cleansing with water, ensuring the sensory evaluations’ reliability and consistency. A quantitative and qualitative assessment was conducted based on the visual aspect of the distillate (clarity and presence of bubbles, if any), aromas in terms of intensity and main notes, as well as the mouthfeel, body, and texture. The presence or absence of bubbles in distilled spirits can indicate alcohol content and purity. Bubbles, often called “pearls” in traditional Mexican mezcal production, can form when a liquor stream is poured into a vessel. These bubbles remain stable for tenths of seconds if the alcohol content is around 50%. The extended lifetime of these bubbles is due to changes in surface tension, density, viscosity, and the presence of surfactants. Although bubbles are not always expected in all spirits, their presence can indicate specific beverage properties. In our study, the absence of bubbles in the yacon-based spirits suggests a different alcohol content or purity compared to traditional methods that might display such characteristics [
33]. On the other hand, the quantitative evaluation considered the visual aspect, purity and clarity of the sample, intensity and complexity of the aromas, as well as the flavors’ intensity, complexity, and persistence. This quantitative evaluation was scored on a scale of 100 points for each aspect mentioned above. In the qualitative assessment, Batch 1 received comments regarding the liquid’s clarity and the absence of bubbles; the main flavor notes mentioned were herbal, spicy, vanilla, root, and caramel. The mouthfeel of this batch was described as having a medium texture with an alcoholic body. Regarding its quantitative rating, it received an average score of 92 among the three sommeliers. As for Batch 2, it was noted to have a clear visual aspect without bubbles and an intense aroma intensity with main herbal, caramel, and toasted root notes, along with a secondary note of licorice. In this case, the mouthfeel was also described as having a medium texture. The quantitative results of this batch received an average score of 93.6 among the three evaluators. Regarding Batch 3, the visual aspect was clear without bubbles, with moderate aromas and main herbal and alcoholic notes, along with earthy secondary notes. The mouthfeel was rated as having a light to medium texture and was noted for its dense texture. The body of the distillate, according to the evaluators, was alcoholic. The average rating for this batch in the quantitative evaluation was 87.
One major limitation of this study is the absence of replicated trials, as the experiments involved only three batches with different processes, potentially affecting the reliability and generalizability of the results. Variability in fermentation conditions, including temperature and yeast strain, further complicates the isolation of yacon’s effects as a fermentation substrate. Additionally, the small production scale may not accurately reflect commercial scenarios, introducing challenges not accounted for in this study. The sensory evaluation, while insightful, was limited to a small panel of sommeliers, and a larger, more diverse panel would provide a more comprehensive assessment of the beverage’s attributes. Addressing these limitations in future research will be crucial for validating the findings and advancing the commercial viability of Yacon-based alcoholic drinks.