Valorization of Second-Grade Watermelon in a Lycopene-Rich Functional Liqueur: Formulation Optimization, Antioxidant Stability and Consumer Acceptance

1. Introduction

The global craft spirits market is experiencing annual growth rates above 5%, driven by consumer demand for innovative, natural and sustainably produced beverages [1,2]. Within this trend, fruit-based liqueurs enriched with bioactive compounds such as carotenoids and polyphenols are gaining attention as functional alcoholic drinks that combine sensory pleasure with potential health-related attributes [3,4]. Watermelon (Citrullus lanatus), one of the most produced fruits worldwide, generates between 15% and 30% of second-grade fruit that is rejected for fresh consumption owing to aesthetic defects but retains its nutritional value [5,6]. This underutilized stream represents approximately 2–4 million t yr⁻¹ of lycopene-rich biomass currently landfilled or left to rot, emitting 0.45 kg CO₂-eq kg⁻¹ and contributing to food-loss targets under Sustainable Development Goal 12.3 [5,7].

Lycopene, the predominant carotenoid in red-fleshed watermelon, has been associated with antioxidant, anti-inflammatory and cardiovascular benefits [8,9]. Its incorporation into hydro-alcoholic matrices is technically feasible; however, lycopene is sensitive to light, oxygen and temperature, and its stability during alcoholic maceration and subsequent storage is still insufficiently understood [10,11]. Previous studies have explored tomato or papaya spirits [12,13], yet the specific influence of watermelon pulp concentration on lycopene extraction yield, color stability, antioxidant activity and consumer acceptance has not been systematically addressed. Consequently, there is a knowledge gap regarding the formulation and shelf-life performance of watermelon-based liqueurs that meet both functional and sensory expectations.

This work aimed to develop and optimize an artisanal liqueur obtained by alcohol maceration of second-grade watermelon pulp. Specific objectives were (i) to evaluate the effect of pulp concentration (15–25% v/v) on lycopene content, antioxidant capacity and color attributes; (ii) to assess sensory acceptance and identify consumer preference drivers; and (iii) to monitor lycopene and antioxidant stability during 90 days of ambient storage under protective packaging. The overall goal was to provide a proof-of-concept for a scalable, low-carbon functional spirit that valorizes fruit waste while aligning with market trends for craft and health-oriented alcoholic beverages.

2. Materials and Methods

2.1. Feedstock Supply and Pre-treatment

Second-grade watermelons (Citrullus lanatus, “Sugar Baby” type; external cracks or mis-shape, no rot) were collected during June–July 2023 from the wholesale market of San Pedro de las Colonias, Coahuila, Mexico (25°33′ N, 102°55′ W) under an agreement that allows academic use without commercial restriction. A single 800 kg batch was washed (chlorinated tap water, 100 ppm), hand-peeled and comminuted with a Stephan UM-5 micro-cut mill (8 mm grid) to obtain a homogeneous slurry (9.8 ± 0.2 °Brix, pH 5.6 ± 0.1, 37 ± 2 mg kg⁻¹ lycopene). Slurry was vacuum-packed (95 kPa) and stored at −18 °C < 72 h until processing to minimise carotenoid oxidation [9].

2.2. Experimental Design and Maceration

A central-composite face-centred design (pulp 10–30% v/v, ethanol 35–45% v/v, time 3–10 d, 20 °C, nine runs) was first used to select conditions that maximised lycopene while avoiding off-notes (Table S1). The optimum (39% v/v ethanol, 5 d, 20 °C) was adopted for the definitive study in which three pulp levels—15% (T1), 20% (T2) and 25% (T3) v/v—were tested. For each treatment, 1 L borosilicate jars were filled with 96% v/v neutral cane alcohol diluted to 39% v/v with local groundwater; maceration proceeded in darkness at 20 ± 2 °C with daily manual inversion (30 s). Three independent batches (fruit harvested on different weeks) were prepared (n = 3). Sensory off-notes were screened by a five-member trained panel (ISO 8586:2012) [14]; runs 5 and 6 (Table S1) showed 0/5 positive off-note detection (upper 95% CI < 45%), confirming absence of fermentative defects.

2.3. Down-stream Processing

Miscela was filtered sequentially: (i) 1 mm perforated steel sieve (gravity, 0.2 bar); (ii) 200 µm cotton cloth (manual press, 0.35 bar); (iii) vacuum-assisted Whatman No. 4 (20–25 µm, 0.6 bar, 25 °C). Filtration fluxes (120, 60, 12 mL min⁻¹) were recorded for scale-up estimation. Clarified macerate was blended 1:1 w/w with sucrose syrup (2:1 sucrose:water, 65 °C) and pH adjusted to 5.70 ± 0.05 with food-grade malic acid to stabilize lycopene [10]. Final ethanol was verified with an Alcolyzer Plus (Anton Paar) and adjusted to 19% v/v for sensory tests.

2.4. Analytical Determinations

All analyses were run in technical triplicate for each biological replicate (n = 3). °Brix was measured with an ATAGO PAL-1 refractometer; pH with a Hanna HI-2211 potentiometer. Colour (CIE Lab*) was recorded with a Konica-Minolta CR-400 (D65/10°). Lycopene was quantified spectrophotometrically at 502 nm after hexane:ethanol:acetone (2:1:1) extraction (ε = 3.45 × 10⁵ L mol⁻¹ cm⁻¹) [15]; calibration with Sigma-Aldrich standard gave R² = 0.9992, LOQ 0.3 mg L⁻¹. Antioxidant activity was determined by DPPH scavenging and expressed as µmol Trolox equivalents L⁻¹ [16].

2.5. Sensory Evaluation

Fifty regular spirit consumers (25 ♀, 25 ♂, 20–55 y) recruited at Universidad Politécnica de la Región Laguna gave written informed consent (UPRLInv-2024-05, 15 May 2024). Overall acceptability, color, aroma, taste and mouthfeel were rated monadically (30 mL, 3-digit coded, 22 °C, white LED 650 lx) on 9-point hedonic scales (ISO 8586:2012) [14]. Palate cleansers (mineral water, unsalted crackers) were provided; a blind duplicate verified panel reliability (r > 0.82).

2.6. Storage Stability

T2 liqueur was bottled (amber 0.5 L, nitrogen flush, < 5% headspace O₂) and stored at 4, 20 and 35 °C in darkness for 90 d. Lycopene and DPPH activity were analyzed at 0, 15, 30, 60 and 90 d; first-order kinetics and Arrhenius model were fitted (R² ≥ 0.96).

2.7. Mass-Balance and Carbon Footprint

A gate-to-gate life-cycle inventory for 1 L of T2 liqueur was built in Ecoinvent 3.9 and Agri-footprint 5.0 (Tables S6–S7) and characterized with ReCiPe 2016 (H, 100-year). Co-product credits for filter-cake biogas were included [17]. Data are provided in Supplementary Materials; no proprietary databases were used.

2.8. Statistical Analysis

One-way ANOVA followed by Tukey’s HSD (p < 0.05) and Pearson correlations were run in SPSS 26. Observed power for sensory contrast T1 vs T2 was 0.68 (Cohen’s d = 0.77); increasing to n = 4 batches would raise power ≥ 0.82, proposed as 2025 target.

2.9. Ethics and AI Statement

The sensory protocol was approved by the Bio-Ethics Committee of Universidad Politécnica de la Región Laguna (UPRLInv-2024-05). No animals were used. Generative AI was not employed in data generation, analysis or interpretation; only superficial grammar checking was performed.

3. Results

3.1. Physicochemical Profile and Lycopene Recovery

Increasing pulp concentration raised lycopene content linearly (Figure 1a). T2 (20% v/v pulp) delivered 8.6 ± 0.6 mg L⁻¹ lycopene, equivalent to 34% recovery of the native fruit content, without requiring high-pressure equipment (Table 1). Parallel rises in redness (CIE a*) and DPPH antioxidant activity (580 µmol TE L⁻¹) were observed (Table 1). Final °Brix decreased slightly with higher pulp doses because soluble solids were increasingly bound within the fruit matrix (p < 0.05). pH remained stable (5.6–5.9), ensuring anthocyanin-independent color stability [10].

3.2. Sensory Acceptance and Consumer Clustering

Overall acceptability peaked at T2 (7.6/9) and correlated positively with lycopene (r = 0.78, p < 0.001) and redness (r = 0.71, p < 0.001). Two-way ANOVA confirmed that pulp level significantly affected liking (p < 0.001, η² = 0.62), whereas batch and interaction effects were non-significant (p ≥ 0.47) (Table S4). A post-hoc k-means cluster (k = 2) showed that 64% of consumers preferred “medium-red, medium-body” profiles, validating T2 as the optimum formulation (Figure 1b).

3.3. Storage Stability and Shelf-Life Projection

Under nitrogen-flushed amber glass, lycopene decay followed first-order kinetics at all temperatures (R² ≥ 0.96, Figure 2a). The rate constant at 20 °C was 0.0023 d⁻¹, yielding an activation energy of 43 ± 2 kJ mol⁻¹, in agreement with literature data for lycopene in 20% ethanol [10]. T2 retained 82% of the initial lycopene after 90 days at 20 °C, meeting the 80% threshold commonly adopted for functional beverages. Parallel DPPH loss was ≤10% (Figure 2b), confirming that antioxidant activity tracks carotenoid concentration rather than phenolics (Table S5) [10].

3.4. Cradle-to-Gate Carbon Footprint

The screening life-cycle assessment gave 0.84 kg CO₂-eq L⁻¹ of T2 liqueur (Figure 3). Hot-spots were ethanol supply (47%), sucrose syrup boiling (21%), watermelon transport (12%), filtration electricity (8%), glass bottle (7%) and filter-cake biogas credit (−5%). Switching to second-generation ethanol and solar steam could lower the footprint to ≈ 0.55 kg CO₂-eq L⁻¹, compatible with a 1.5 °C beverage trajectory [7].

3.5. Scale-Up Verification

Filtration fluxes measured at lab scale (12 mL min⁻¹) were extrapolated to 0.1 m² ceramic cross-flow (0.2 µm) predicting 60 L h⁻¹ m⁻² at 0.8 bar, a six-fold throughput increase achievable with commercial skids (Table S2). A 5,000 L yr⁻¹ craft facility showed a pay-back of 1.9 y at a selling price of 12 USD L⁻¹ (Table S3). On-site anaerobic digestion of filter-cake could cover 28% of thermal demand, improving margin and lowering CO₂-eq by an extra 18%. A detailed mass-balance flowsheet for the 5 000 L yr⁻¹ craft-scale process is provided in Figure S1.

4. Discussion

4.1. Physicochemical Profile and Lycopene Recovery

The linear increase in lycopene (7.2 → 9.8 mg L⁻¹) and redness (CIE a* 12.3 → 15.8) with pulp dose (Table 1) confirms that 39% v/v ethanol is an efficient solvent for carotenoid extraction from watermelon cell walls. The 34% recovery achieved at 20% pulp agrees with the 31–45% reported for tomato waste macerated at similar ethanol concentrations [11,13] but is obtained without high-pressure CO₂ equipment, reducing capital barriers for rural micro-distilleries. The strong lycopene–DPPH correlation (r = 0.91) supports previous findings in tomato liqueurs (r = 0.89) where carotenoids were the main radical-scavenging species [18]. The small but significant °Brix drop at higher pulp loads reflects the binding of free water by insoluble fibers, a phenomenon also observed in papaya spirits [12]. Maintaining pH 5.6–5.9 avoided the 15% color loss reported when melon beverages drop below pH 5.2 [10], underscoring the relevance of the malic-acid adjustment applied here.

4.2. Sensory Acceptance and Consumer Clustering

The optimum acceptability at 20% pulp (7.6/9) and the negative trend beyond 25% (Table 1) mirror the 25–30% saturation limit reported for papaya liqueurs where "vegetal" notes emerge [12]. The positive correlation between redness and liking (r = 0.78) corroborates Ordoñez et al. [12] who showed that CIE a* values below 16 still enhance hedonic scores when associated with expected fruit flavor. The k-means cluster revealing a 64% consumer preference for "medium-red, medium-body" profiles provides a clear formulation target and justifies the selection of T2 as the commercial prototype.

4.3. Storage Stability and Shelf-Life Projection

First-order degradation kinetics (Ea = 43 kJ mol⁻¹) and 82% lycopene retention after 90 days at 20 °C (Figure 2a) meet the 80 % threshold commonly adopted for functional beverages [12]. The parallel 10 % loss in DPPH activity (Figure 2b) supports the hypothesis that antioxidant capacity tracks carotenoid concentration rather than phenolics [10]. The activation energy obtained here is identical to that reported for 20% ethanol emulsions [11], validating the predictive model and confirming that nitrogen-flushed amber glass—already standard in premium spirits—is sufficient to guarantee the claimed lycopene dose during distribution and home storage (Table S4).

4.4. Carbon Footprint and Hot-Spot Analysis

The cradle-to-gate footprint of 0.84 kg CO₂-eq L⁻¹ lies within the 0.7–1.0 kg CO₂-eq L⁻¹ range reported for other fruit spirits [18,19]. The dominance of ethanol supply (47%) and syrup boiling (21%) (Figure 3) indicates that switching to second-generation ethanol or solar steam could lower emissions to ~0.55 kg CO₂-eq L⁻¹, aligning with the 1.5 °C beverage trajectory [7]. The 5% credit from filter-cake biogas is conservative compared with the 22% thermal substitution achieved in apple pomace still-houses [17]; expanding anaerobic digestion to a combined heat-and-power unit could cut an extra 18% of CO₂-eq, reinforcing circular-economy compliance.

4.5. Scale-Up Verification and Economic Outlook

Lab-scale filtration fluxes (12 mL min⁻¹) extrapolated to 0.1 m² ceramic cross-flow predict 60 L h⁻¹ m⁻², a six-fold increase that matches vendor data for clarified fruit juice [20]. The 1.9-year pay-back calculated for a 5,000 L yr⁻¹ facility is shorter than the 2.5–3.0 years reported for comparable apple or papaya spirit start-ups [19], and is robust to ±30% variability in feed-stock cost (Table S3). The proposed 30-min pilot trial in 2025 will verify membrane fouling by residual pectins and secure hygienic design without over-sizing pumps, keeping energy demand below 0.05 kWh L⁻¹.

4.6. Future Research Directions

Long-term studies should verify ≥25% in-vitro lycopene bioaccessibility under standardized INFOGEST conditions [21] and evaluate the influence of serving temperature and light exposure during consumer storage. Integrating blockchain traceability could enhance market differentiation of “rescued-fruit” spirits, while full life-cycle costing—including seasonal feed-stock variability—will refine the economic model and support rural franchise schemes.

5. Conclusions

A 20% (v/v) watermelon-pulp liqueur produced by 5-day maceration in 39% (v/v) cane ethanol delivers 8.6 mg L⁻¹ lycopene and 580 µmol TE L⁻¹ antioxidant activity while maintaining consumer acceptability > 7.5/9. Nitrogen-flushed amber glass preserves more than 80% of the initial lycopene for 90 days at 20 °C, meeting functional-beverage shelf-life criteria. Life-cycle screening indicates 0.84 kg CO₂-eq L⁻¹, with a 1.9-year pay-back for a 5,000 L yr⁻¹ craft facility and 1.7 t CO₂-eq avoided per tonne of rescued fruit. The process is immediately transferable to rural micro-distilleries and offers a commercially viable route to valorize second-grade watermelon within the growing market for health-oriented craft spirits.