Valorization of Postharvest Mandarin Residues in the Production of Pectins and Distillates with Antioxidant Capacity

Abstract

Non-marketable postharvest residue from Satsuma Owari mandarins (Citrus unshiu)—discarded due to having a puffy peel (MC) or being damaged (MD)—was evaluated against a commercial mature reference (MM) using an integrated cascade biorefinery. Along the liquid pathway, must fermentation (6.1–9.0% v/v) and fractional distillation yielded a citrus brandy (~30% v/v). Vinasse emerged as the stream of highest strategic relevance, maximizing bioactive response (up to 1339.1 µg GAE/mL) with a naringenin-dominated phytochemical profile. Along the solid pathway (peel), sequential recovery of essential oil and pectins was performed. Essential oil yields (0.575–0.85 mL/100) decreased significantly with physical peel damage, underlining flavedo integrity. Concurrently, recovered pectins (18.6–23.3%) exhibited a degradation-driven structural transition: MM retained galacturonic acid-rich linear domains (89.8%), whereas in MD they had evolved toward a branched architecture with higher rhamnogalacturonan I (30.6%). All fractions were low-methoxyl (DM < 50%), ensuring functionality in ionic gelling systems. In conclusion, this reproducible strategy successfully transforms postharvest losses into functional ingredients and chemical distillates, providing a robust foundation for circular economy models within the citrus industry.

1. Introduction

Mandarins are one of the most commercially significant citrus resources globally. For the 2024–2025 production cycle, worldwide production was estimated at 37.51 million metric tons, with Peru contributing approximately 0.57 million tons (~2% of the international supply) [1]. The Satsuma Owari mandarin (Citrus unshiu) cultivar stands out due to its exceptional nutritional attributes, characterized by a profuse profile of L-ascorbic acid, carotenoids, and phenolic compounds. Furthermore, this variety exhibits superior processing aptitude compared to other traditional citrus fruits, registering optimal juice yields between 52.9% and 60.1% under standard operational conditions [2].

Nonetheless, this cultivar presents marked postharvest vulnerability. Its intrinsic propensity for developing chilling injury and accelerated loss of organoleptic attributes drastically shortens its commercial shelf life [3]. Consequently, a substantial volume of biomass is systematically segregated during the harvesting, sorting, and packaging phases because it fails to satisfy the stringent aesthetic and physicochemical standards required for the fresh market, such as homogenous external coloration, specific maturity indices (soluble solids-to-acid ratio), juice yield, or tissue integrity against pests and physiological disorders [4,5,6].

Among the most critical physiological disorders in loose-skinned cultivars is puffing (or albedo breakdown), a destabilizing condition that causes separation between the flavedo and the pulp, decreasing macroscopic fruit firmness, dehydrating juice segments, and exponentially increasing susceptibility to mechanical damage [7]. Concurrently, opportunistic pathogens such as Trichothecium roseum trigger severe postharvest diseases like pink rot, thereby compromising the safety and economic viability of the value chain [8]. Globally, aggregate losses in the citrus sector reach up to 34%, distributed transversally from preharvest stages to commercial distribution [9]. This critical volume of discards, comprising fruits that fall outside market specifications but remain biochemically viable, constitutes a strategic raw material for developing industrial cascade valorization processes.

The effective implementation of a biorefinery requires acknowledging the phytochemical heterogeneity of agri-food waste streams, as distinct residual matrices do not share the same bioconversion potential. In this study, the biomass categorization was not arbitrary but rather designed to capture the primary non-marketable waste streams generated throughout the citrus value chain. The mature reference mandarin (MM) serves as the biochemical baseline against which “puffy” mandarins (MCs)—characterized by collapsed albedo but intact flavedo—and mandarins with damaged peel (MDs)—highly susceptible to microbial spoilage—were evaluated, directly reflecting real-world commercial rejection criteria (visual defects, loss of firmness, and mechanical damage). Evaluating how each physiological condition impacts the yield and quality of pectin, essential oil, and vinasse provides a critical operational foundation to determine whether these heterogeneous batches can be co-processed en masse or require segregation and differential handling at the plant scale.

Based on this segmentation, the mandarin peel constitutes a complex matrix due to the coexistence of hydrophobic volatile metabolites and structural macromolecules distributed throughout its tissue architecture (Figure 1). The flavedo accumulates oil glands rich in d-limonene, a monoterpene in high demand across the food, pharmaceutical, and cosmetic industries due to its functional versatility, bioactive properties, and applications as a biopesticide agent [10]. Emerging extraction technologies, such as hydrodistillation assisted by acidic deep eutectic solvents (A-DES), have optimized these processes in Citrus reticulata, reaching yields of 3.19 ± 0.02% under optimized conditions [11]. Conversely, the albedo fraction represents a profuse source of pectins, constituting between 20% and 30% of the biomass on a dry weight basis. Studies of analogous cultivars such as the ‘Ponkan’ mandarin confirm that these polymeric fractions possess a high galacturonic acid (GalA) content and technofunctional properties comparable to commercial standards [12,13]. The projection of these polysaccharides as gelling, thickening, or emulsifying agents, or as functional macromolecular scaffolds, is strictly dictated by their molecular architecture. Therefore, the quantification of key attributes such as GalA content, degree of methoxylation (DM), degree of acetylation (DA), neutral monosaccharide profile, and average molar mass is indispensable to precisely determine their technological performance [11,12,13,14].

Simultaneously, the liquid fraction or must concentrates fermentable sugars, organic acids, and secondary metabolites that can be channeled through alcoholic fermentation and fractional distillation processes. The literature reports total phenolic concentrations in citrus juices and fruit wines ranging from 0.75 to 2.61 mg GAE/mL, as well as flavonoid contents in fresh juices between 0.19 and 0.52 mg RE/mL [15,16]. During fermentative metabolism, this phytochemical profile undergoes profound modifications induced by oxidation phenomena, adsorption onto yeast cell membranes, hydrolytic release from the cellular matrix, and bioconversion of bound glycosides into free aglycones—processes whose kinetics depend on the microbial strain and operational conditions [14,16].

The subsequent distillation of the base wine follows a fractionation pattern analogous to that of traditional spirit distillates, segregating streams of heads, hearts (citrus brandy), and tails, while simultaneously generating vinasse as the majority liquid residue [17,18]. While the head and tail fractions concentrate restrictive volatile congeners such as acetaldehyde and furfural—which limits their suitability for direct consumption but opens up non-food valorization opportunities through rectification into industrial alcohol [18]—the residual vinasse retains the highest density of polyphenols, flavonoids, and non-volatile compounds from the original matrix. This converts the effluent into a strategic platform for recovering high-value antioxidants, requiring rigorous compositional characterization to validate its technofunctional viability [18,19,20]. Nonetheless, any transposition of this stream toward nutraceutical or food schemes demands prior acidity neutralization, removal of interfering organic load, and strict evaluation of safety, macromolecular stability, and organoleptic acceptability [18].

Despite extensive research into isolated processing units—such as the standalone extraction of essential oils, independent purification of pectins, or isolated fermentation of juice/must—a fundamental knowledge gap persists. There is a lack of empirical evidence regarding how postharvest physiological degradation transversally affects an integrated, multi-product processing system. Specifically, it remains unknown how distinct physiological defects (such as puffiness or peel damage) simultaneously alter volatile extraction efficiency, topochemical and conformational transitions of pectin, fermentation kinetics, volatile congener distribution during distillation, and residual antioxidant capacity in the vinasse. The central novelty of this work lies in implementing a bidirectional cascade biorefinery architecture that couples the solid peel pathway with the liquid juice pathway, mapping these processes directly as a function of the fruit’s initial physiological condition.

Accordingly, the objective of this research was to evaluate the comprehensive valorization potential of postharvest residues from Satsuma Owari mandarins (Citrus unshiu), classified into three differentiated commercial conditions: mature reference (MM), puffy (MC), and damaged peel (MD). Through this cascade approach, the direct influence of raw material quality on pectin yield and molecular architecture, the chemical composition of the resulting distillates, and the phytochemical and antioxidant profiles of the residual liquid fractions was elucidated.

2. Materials and Methods

2.1. Raw Materials, Sample Classification, and Reagents

Nine independent experimental batches of Satsuma Owari mandarins (Citrus unshiu) were harvested from the “Don Germán” orchard, located within the Regional Institute of Development of the Universidad Nacional Agraria La Molina in Cañete, south of Lima, Peru. To ensure statistical rigor and capture seasonal variability, three distinct sampling campaigns were conducted on 20 July 2025, 30 July 2025, and 20 February 2026. During each campaign, 20 kg of fruit was collected for each of the three physiological conditions evaluated: mature reference mandarin (MM), residual puffy mandarin (MC), and residual damaged-peel mandarin (MD). Each batch was processed independently through separation, fermentation, distillation, essential oil recovery, pectin extraction, and analytical characterization. Consequently, the three batches per physiological condition served as statistical experimental replicates.

Fermentation and distillation were performed at the VAINQUIRE-CEMTRAR laboratory. The commercial wine yeast Saccharomyces cerevisiae (AB Biotek, Casteggio, Italy) was utilized for inoculation. For the spectrophotometric assays, Folin–Ciocalteu reagent, 2,4,6-tripyridyl-s-triazine (TPTZ), and 2,2-diphenyl-1-picrylhydrazyl (DPPH) were used. Analytical standards for gallic acid, Trolox, and quercetin were supplied by Sigma-Aldrich (USA). Monosaccharide standards used for pectin characterization, including D-galacturonic acid (GalA), D-glucose (Glc), D-mannose (Man), D-galactose (Gal), L-rhamnose (Rha), L-fucose (Fuc), L-arabinose (Ara), and D-xylose (Xyl), were purchased from Merck. All other chemical reagents used in this study were of analytical grade.

Phenolic compound profiling of the vinasse streams was performed via UHPLC-ESI-Orbitrap-MS using an array of reference standards encompassing alkaloids, catechins, flavonoids, and anthocyanins. Comprehensive specifications regarding these reference materials, including manufacturers and purities, are detailed in Section 2.3.4.

2.2. Valorization Scheme for Postharvest Mandarin Residues

Mandarins from the three categories were washed and manually peeled. Subsequently, fruits were manually fractionated to segregate the peel from the juice (must). Peel fractions were coded as C-MM, C-MC, and C-MD and were allocated for the sequential recovery of essential oils (via steam distillation) and pectin. The peeled fruits were utilized as the fermentation substrate for bioethanol conversion and subsequent distillation.

The biorefinery scheme comprised two complementary processing pathways applied to the three mandarin categories. The liquid phase pathway initiated with must fermentation to produce citrus wine. These wines were subsequently distilled to yield a premium citrus brandy, while simultaneously recovering head, tail, and vinasse byproducts. The solid-state pathway involved the sequential utilization of C-MM, C-MC, and C-MD for essential oil recovery and downstream pectin extraction (Figure 2). The resulting fractions were characterized according to matrix-specific requirements, assessing chemical composition, phenolic content, antioxidant capacity, and the structural attributes of the recovered pectins.

2.2.1. Fermentation and Distillation of Mandarin Must

Mandarin juices (M-MM, M-MC, and M-MD) were extracted by processing peeled fruits in a horizontal pulper (Vulcano, Peru), followed by filtration through a 3 mm mesh screen to remove coarse suspended solids. The identical fermentation protocol was applied across all three evaluated categories (MM, MC, and MD), maintaining a constant juice volume, inoculum ratio, yeast activation temperature, and process termination criterion. The inoculum was prepared by activating 5 g of Saccharomyces cerevisiae in 1.2 L of filtered juice at 32 °C for 30 min. Notably, no exogenous sugars were added, allowing for exclusive evaluation of the endogenous fermentative potential of each mandarin category.

Fermentation progress was monitored daily by measuring total soluble solids (°Brix) using a portable refractometer (model LB32T, Guangdong, China). The process continued until soluble solids dropped below 4 °Brix, which was set as the operational endpoint. During this period, gentle agitation was performed on day 3, and residual solids were removed on day 7. Upon fermentation completion, the broth was filtered to remove yeast sediment, yielding the citrus wine, which was stored at 25 °C prior to distillation; density stability was confirmed using a hydrometer.

Distillation was conducted in a 30 L copper pot still, adhering to operational criteria established by INDECOPI [21] for artisanal pisco production. Fractionation was governed by a comprehensive assessment integrating volumetric parameters, continuous monitoring of alcohol by volume (ABV), and organoleptic flow evaluation. Initially, the head fraction was separated, corresponding to 1% of the total boiler charge volume. Subsequently, the citrus brandy (heart fraction) was collected within an operational target range of 25% to 35% ABV. Tail separation was initiated upon detecting a drop in ABV below the target reference range or the onset of characteristic sensory off-notes. The final ABV of all recovered fractions was verified in accordance with Peruvian National Standard NTP 211.052:2018 for alcoholic beverages [22].

2.2.2. Recovery and Conditioning of Distillation Byproducts

During the distillation of the citrus wine, byproduct fractions corresponding to heads, tails, and the residual pot liquor (vinasse) remaining in the still were recovered. The head fraction, coded as HD, comprised approximately 1% of the initial citrus wine volume loaded into the alembic. The tail fraction (TD) was defined as the terminal distillate stream; its collection cutoff was determined by the progressive decline in alcoholic strength and concomitant changes in the organoleptic profile of the distillate flow. The vinasse, coded as V, constituted the majority liquid residue remaining after distillation.

The HD, TD, and V fractions were volumetrically quantified. Subsequently, samples were filtered and stored at 4 °C until analysis. These streams were characterized by their alcoholic strength, volatile acidity, volatile organic compounds, total phenolic content, total flavonoid content, and antioxidant capacity. For the vinasse fraction, comprehensive chemical profiling was also conducted via liquid chromatography.

2.2.3. Peel Essential Oil Recovery and Downstream Pectin Extraction

Mandarin peels from the three evaluated conditions were sectioned into fragments of approximately 2 cm × 3 cm. For essential oil recovery, 280 g of fresh peel was weighed and combined with 500 mL of water in a 1000 mL round-bottom flask. The mixture was subjected to steam distillation for 2 h. The recovered essential oil was separated, quantified, and expressed as yield in mL/100 g of fresh peel.

Following oil extraction, the spent residual peels were collected for drying. This process was performed in a forced-air oven (model UF160, Memmert, Büchenbach, Germany) at 40 °C for 24 h. The dried biomass was comminuted using a knife mill and sieved through a 40-mesh screen to obtain a fine, homogeneous peel flour.

For pectin extraction, 10 g of the peel flour was weighed and transferred to a 250 mL round-bottom flask containing 100 mL of 0.6 M phosphoric acid. The mixture was heated under reflux at 90 °C with continuous magnetic stirring at 300 rpm for 2 h. Following extraction, the hot slurry was filtered through Whatman No. 41 filter paper. The recovered filtrate was blended with three volumes of 96% (v/v) ethanol and allowed to stand overnight at 4 °C to induce pectin precipitation.

The macromolecular precipitate was isolated via filtration using Whatman No. 41 paper, washed three times with 96% (v/v) ethanol, and subsequently lyophilized for 24 h (model BK-FD18P, BioBase, Jinan, China). The purified pectin was obtained as a dry powder and stored under desiccating conditions until further characterization.

2.3. Analytical Determinations

2.3.1. Volatile Organic Compound Profiling and Volatile Acidity

Volatile compounds were analyzed in the citrus wine, the citrus brandy, and their respective byproducts. Ethyl acetate, furfural, acetaldehyde, and methanol were quantified. Additionally, higher alcohols (isopropanol, propanol, butanol, isobutanol, 3-methyl-1-butanol, and 2-methyl-1-butanol) were determined in accordance with the guidelines of the NTP 211.035:2019 [23].

Chromatographic analysis was performed using a gas chromatograph (model 7890A, Agilent Technologies, Santa Clara, CA, USA) equipped with a flame ionization detector (FID) and an Agilent J&W 19091F-105 specialized capillary column. The detector temperature was maintained at 220 °C, and an oven temperature program ranging from 160 to 200 °C was applied. Helium served as the carrier gas at a constant flow rate of 35 mL/min. Sample preparation involved blending 9 mL of the liquid sample with 1 mL of 3-pentanol as an internal standard. The injection volume was 1 µL. Concurrently, volatile acidity was determined as acetic acid equivalents via neutralization with sodium hydroxide according to NTP 212.031:2019 [24].

2.3.2. Total Phenolic and Total Flavonoid Contents

The total phenolic content (TPC) of each matrix was determined using the Folin–Ciocalteu colorimetric method described by Singleton et al. (1999) [25], with minor modifications. Briefly, a 50 µL aliquot of the sample was mixed with 1000 µL of Folin–Ciocalteu reagent (1:10 dilution) and allowed to react for 5 min. Subsequently, a 7.5% Na₂CO₃ solution was added to reach a final volume of 2 mL, followed by incubation at 45 °C for 15 min. Absorbance was read at 765 nm on a Genesys 30 spectrophotometer (Thermo Scientific, Waltham, MA, USA). Quantification was conducted using a gallic acid standard calibration curve over a concentration range of 0.001 to 0.01 µg/mL. The results are expressed as µg of gallic acid equivalents per mL of sample (µg GAE/mL).

The total flavonoid content (TFC) was determined via the aluminum chloride colorimetric method adapted from Pourmorad et al. (2006) [26], with minor modifications. An 800 µL aliquot of the sample was blended with 40 µL of 10% (w/v) AlCl₃ and 40 µL of 1 M CH₃COOK. The final volume was adjusted to 2000 µL with deionized water, and the mixture was incubated for 30 min at room temperature. Absorbance was measured at 415 nm using a Genesys 30 spectrophotometer. Quantification was performed using a quercetin calibration curve within a range of 0.02 to 0.2 µg/mL. The results are expressed as µg of quercetin equivalents per mL of sample (µg QE/mL).

2.3.3. Determination of Antioxidant Capacity

Antioxidant capacity was evaluated using ferric reducing antioxidant power (FRAP) and DPPH radical scavenging assays. The FRAP assay was performed according to Benzie and Strain [27], following the adaptations described by Arévalo-Ortiz et al. [18], with minor modifications. The working FRAP reagent was prepared fresh by mixing acetate buffer (0.3 M, pH 3.6), TPTZ solution (10 mM), and FeCl₃·6H₂O (20 mM) in a 10:1:1 (v/v/v) ratio. For the analysis, 1000 µL of freshly prepared FRAP reagent was added to 1000 µL of sample, and the mixture was incubated at 37 °C for 15 min. Absorbance was recorded at 593 nm on a Genesys 30 spectrophotometer. A Trolox calibration curve (0.05–0.5 mmol/L) was used for quantification, and the results are reported as µmol of Trolox equivalents per mL of sample (µmol TE/mL).

The DPPH radical scavenging activity was determined according to Mitrevska et al. [28], following the adaptations of Arévalo-Ortiz et al. [18], with minor modifications. A 50 µL aliquot of the sample, previously diluted in methanol as required by the matrix, was mixed with 1950 µL of a fresh 0.1 mM DPPH solution. The mixture was incubated in the dark at room temperature for 30 min, and the absorbance was measured at 515 nm on a Genesys 30 spectrophotometer. Trolox was used as the reference standard within a range of 0.22 to 2.24 µmol/mL. The results are expressed as µmol of Trolox equivalents per mL of sample (µmol TE/mL).

2.3.4. Phytochemical Profiling of Vinasse Streams

The vinasse streams were characterized via liquid chromatography following an adapted protocol from Arévalo-Ortiz et al. [18]. The analytical platform comprised a UHPLC system (Dionex Ultimate 3000, Thermo Scientific, Sunnyvale, CA, USA) equipped with a binary gradient pump, an autosampler, and a thermostated column compartment, coupled to an Exactive Plus mass spectrometer with an Orbitrap analyzer.

Ionization was performed using an electrospray ionization (ESI) interface operated in positive ion mode with a capillary voltage of ±3.5 kV. Molecular identification was executed in full-scan mode. The process involved extracted ion chromatogram (EIC) generation for the protonated molecules [M+H]⁺ of the targeted analytes. Mass accuracy with Δppm < 2, isotopic ratios, and fragmentation patterns were rigorously verified for identity confirmation. Targeted analyte quantification was carried out using calibration curves prepared with certified reference materials, including standards for alkaloids, catechins, flavonoids, phenolic acids, anthocyanins, ursolic acid, and hesperidin.

2.3.5. Characterization of Essential Oils and Recovered Pectins

Essential oil yield was calculated based on the volume recovered via steam distillation and expressed as mL of essential oil per 100 g of fresh peel. Pectin molecular structure was elucidated via Fourier-transform infrared spectroscopy (FTIR) using a Nicolet iS10 spectrometer (Thermo Scientific, Waltham, MA, USA) operated in attenuated total reflectance (ATR) mode. Spectra were recorded in the wavenumber range of 400 to 4000 cm⁻¹, with a resolution of 1 cm⁻¹ and 16 scans. The obtained spectra were compared with the established pectin literature, focusing primarily on absorption bands associated with hydroxyl groups, C–H stretching, ester groups, free carboxyl groups, and characteristic glycosidic skeletal vibrations.

The degree of methoxylation (%DM) was estimated by integrating the absorption bands corresponding to esterified carbonyl groups (1745 cm⁻¹) and free carboxylate groups (1630 cm⁻¹). This calculation followed the methodology described by Wang et al. [29] for galacturonic acid units, applying the standard area ratio equation:

$$\%DM={\displaystyle \frac{A_{1745}}{A_{1745}+A_{1630}}}\times 100$$

Pectin was further characterized by determining its moisture, ash, protein, and total phenolic contents, as well as its monosaccharide composition, degree of acetylation, weight-average molecular weight, and polydispersity index. Monosaccharide composition was resolved via acid hydrolysis, reduction with NaBH₄, and subsequent acetylation of the pectin fragments; the resulting alditol acetates were quantified through gas chromatography. The weight-average molecular weight and polydispersity index were determined via high performance size exclusion chromatography (HPSEC) following the approach outlined by Reichembach and de Oliveira Petkowicz [30].

2.4. Experimental Design and Statistical Analysis

As the primary objective of this study was to evaluate the comprehensive valorization potential of mandarin postharvest residue, the residual fruits (MCs and MDs) were contrasted with reference mature mandarins (MMs) within a methodological framework designed to verify the analytical consistency and robustness of the system, rather than a conventional treatment differentiation approach. This allowed for assessment of these byproducts as stable and reproducible raw materials for biorefinery processes.

The experimental design incorporated three sampling campaigns across different chronological periods to capture the natural variability inherent to the physiological conditions of the fruit. The first and second campaigns, performed consecutively over a short-term interval, served as immediate consistency replicates to control for internal experimental error. Conversely, the third campaign, harvested after a 7-month gap, was deliberately used as a seasonal validation block. This extended temporal window reflects the real-world supply dynamics of an industrial-scale plant, where residual biomass is subject to interannual climatic, edaphoclimatic, and phenological fluctuations. Thus, the rationale for including the third campaign was to rigorously demonstrate that the physicochemical properties, macromolecular composition, and bioconversion yields of the residual streams maintain technical stability and reproducibility despite seasonal variability.

To formally incorporate this temporal structure into the statistical model, the sampling period factor was integrated as a blocking effect (randomized block design). This approach mathematically isolated variance induced by seasonality or batch origin from the residual error variance, ensuring that effects attributed strictly to the fruit’s physiological condition (MM, MC, and MD) were unbiased by the harvest window.

Due to the intrinsic heterogeneity of the samples collected across extended time intervals, the quantitative data were transformed using the log (X + 1) function to mitigate heteroscedasticity and approximate normal distributions, which were formally verified via the Shapiro–Wilk test. Once parametric assumptions were validated, one-way analysis of variance (ANOVA) was performed, followed by Tukey’s honest significant difference (HSD) post hoc test with α = 0.05 for individual parameter mean comparisons. The corresponding values are reported in tables and clustered column charts as the mean ± standard deviation of the untransformed raw data.

Additionally, Principal Component Analysis (PCA) was performed of the transformed data matrix of the active observations (citrus wine, citrus brandy, heads, tails, and vinasse). The aim of performing this multivariate analysis was to explore the interrelationships among the active variables—encompassing total phenolic and total flavonoid contents, in vitro antioxidant capacity (FRAP and DPPH), alcohol by volume, volatile acidity, and volatile organic compounds (ethyl acetate, furfural, acetaldehyde, higher alcohols, and methanol), all analyzed in triplicate—and to correlate these active variables with active observations. Data processing and statistical modeling were executed using the XLSTAT software version 2025.1.3 (Addinsoft, New York, NY, USA).

3. Results and Discussion

3.1. Citrus Brandy Production from Postharvest Mandarin Residue

The bioconversion of mandarin must via alcoholic fermentation and subsequent fractional distillation yielded citrus wine, citrus brandy, and three distinct distillation byproducts: heads, tails, and vinasse. This sequence constitutes the liquid phase pathway of the comprehensive valorization scheme. Characterizing this pathway enables evaluation of how raw material conditions (MM, MC, and MD) modulate fermentative kinetics, distillate recovery efficiency, and the generation of secondary streams with downstream valorization potential.

3.1.1. Fermentation and Distillation Performance

The initial physicochemical parameters of the mandarin musts exhibited noticeable variations depending on the physiological condition of the raw material (

Table 1. Results of mandarin fermentation and distillation. Results are expressed as the mean ± standard deviation, n = 3.

	Total Soluble Solids °Brix	Percentage of Initial Volume (%)	Alcoholic Strength % (v/v)
Products
Must (M)
- M-MM	8.4 ± 2.3	100.0 ± 0.0	0.0 ± 0.0
- M-MC	11.0 ± 5.7	100.0 ± 0.0	0.0 ± 0.0
- M-MD	10.0 ± 4.2	100.0 ± 0.0	0.0 ± 0.0
Citrus wine (CW)
- CW-MM	<4.0 ± 0.0	99.0 ± 0.0	6.1 ± 0.1
- CW-MC	<4.0 ± 0.0	99.0 ± 0.0	6.1 ± 0.1
- CW-MD	<4.0 ± 0.0	99.0 ± 0.0	9.0 ± 0.1
Citrus brandy (CB)
- CB-MM	nd	6.4 ± 2.7	31.1 ± 3.5
- CB-MC	nd	5.8 ± 4.5	30.0 ± 4.2
- CB-MD	nd	7.1 ± 3.8	29.4 ± 4.1
By-products
Head distillate (HD)
- HD-MM	nd	0.9 ± 0.1	43.7 ± 7.6
- HD-MC	nd	0.9 ± 0.2	48.3 ± 12.6
- HD-MD	nd	1.1 ± 0.4	46.3 ± 11.0
Tail distillate (TD)
- TD-MM	nd	2.5 ± 0.2	17.7 ± 3.6
- TD-MC	nd	2.1 ± 0.1	17.5 ± 4.0
- TD-MD	nd	2.1 ± 1.5	15.7 ± 3.2
Vinasse (V)
- V-MM	nd	90.0 ± 3.3	<0.5 ± 0.0
- V-MC	nd	88.7 ± 6.1	<0.5 ± 0.0
- V-MD	nd	88.9 ± 4.8	<0.5 ± 0.0

MM: mature reference mandarin; MC: residual puffy mandarin; MD: residual damaged-peel mandarin; nd: not detected.

). Total soluble solids (TSSs) ranging from 8.4 to 11.0 °Brix and total acidity values of 0.39% to 0.69% w/v (expressed as citric acid) were observed across the evaluated categories (MM, MC, and MD). These baselines align with values reported in the literature for mature Satsuma Owari fruit, where TSSs typically range from 9.1 to 13.9 °Brix, and total acidity varies between 0.50% and 0.90% [2].

In citrus matrices, soluble solids, titratable acidity, and their quantitative ratio are critical indicators of fruit quality, maturity, and industrial processing aptitude, all of which are governed by species, cultivar, and preharvest environmental factors [3,5,31]. Furthermore, pulp development involves complex physiological dynamics characterized by concurrent water and carbohydrate accumulation alongside organic acid catabolism, directly dictating the chemical profile of the must available for fermentation [32]. The stability observed across these core quality parameters confirms that, despite being classified as postharvest discards, the residual biomass retains sufficient biochemical integrity to undergo industrial bioconversion [2].

The baseline TSS concentration in the must directly dictates alcoholic fermentation, as it reflects the availability of fermentable sugars for Saccharomyces cerevisiae. Unlike conventional grape musts, fruit-based formulations frequently feature lower sugar profiles and more pronounced acidity levels—conditions that typically mandate prior technical corrections, such as chaptalization or deacidification [33]. Throughout the development of citrus wine, the initial sugar concentration serves as the primary carbon source for cellular proliferation and subsequent ethanol synthesis; consequently, the original variety and composition of the must restrict the ultimate alcoholic strength achieved [16].

Following fermentation, TSS levels dropped below 4 °Brix across all three categories, confirming efficient sugar assimilation by the yeast. The resulting citrus wines exhibited ethanol contents of 6.1% to 9.0% v/v (Table 1). These ethanol metrics fall squarely within the standard range for fruit wines (5% to 13% v/v), which are categorized as non-distilled fermented beverages [33]. However, these values remain lower than those reported for citrus wines subjected to pre-fermentative sugar adjustments [16].

Contrasting these findings with conventional viniferous matrices such as “Red Globe” table grapes—which reach TSS values of 14.7–16.4 °Brix at harvest [34]—provides clear context for the lower alcoholic strength of the citrus wine. This discrepancy reflects the explicit valorization approach of this study, which aims to exploit the intrinsic potential of the residual biomass without relying on energy-intensive external corrections prior to fermentation.

Thermal fractionation of the citrus wine enabled separation and recovery of a principal alcoholic fraction (citrus brandy) and three secondary streams. The citrus brandy accounted for 5.8% to 7.1% of the initial volume loaded into the still, reaching final alcoholic strengths of 29.4% to 31.1% v/v (Table 1). Despite the lower initial sugar density relative to grapes, these yields demonstrate the technical feasibility of recovering a spirit distillate from residual citrus biomass. Distillation of fruit base wines represents an established strategy for the valorization of agricultural sorting losses and production surpluses, wherein operational parameters directly dictate the physicochemical profile of the final products [18,35].

The head fraction accounted for 0.9% to 1.1% of the initial citrus wine volume for MM, MC, and MD, respectively, yielding corresponding alcohol contents of 43.7% to 48.3% v/v (Table 1). The tail fraction represented 21% to 25% of the initial charge volume, with ethanol contents of 15.7 to 17.7% v/v (Table 1). Concurrently, vinasse constituted the majority liquid stream, accounting for volumetric fractions of 88.7 to 90.0%, with residual ethanol levels below 0.5% v/v. The elevated proportion of vinasse underscores that distillation inherently generates a predominantly liquid residue stream, requiring robust compositional characterization to justify its subsequent downstream valorization within the proposed integrated biorefinery scheme.

3.1.2. Chemical Quality of Mandarin Citrus Wine and Citrus Brandy

The chemical quality of the citrus wine and citrus brandy was determined by quantifying ethyl acetate, furfural, acetaldehyde, higher alcohols, methanol, and volatile acidity (

Table 2. Chemical quality of mandarin citrus wine and mandarin citrus brandy. The results are expressed as the mean ± standard deviation in mg/L of sample analyzed, n = 3.

Products	Ethyl Acetate (mg/L)	Furfural (mg/L)	Acetaldehyde (mg/L)	Propanol (mg/L)	Butanol (mg/L)	Isobutanol (mg/L)	2- y 3-Methyl-1-Butanol (mg/L)	Methanol (mg/L)	Acetic Acid (mg/L)
Citrus wine (CW)
CW-MM	15.7 ± 6.5 d	nd	75.7 ± 34.2 b	22.4 ± 12.9 b	0.0 ± 0.0	7.2 ± 5.4 b	34.8 ± 24.8 b	145.3 ± 39.4 b	857.8 ± 524.3 a
CW-MC	101.0 ± 34.0 b,c	nd	83.4 ± 9.2 b	41.7 ± 2.7 a,b	0.0 ± 0.0	16.9 ± 0.2 a,b	67.8 ± 16.2 a,b	275.7 ± 88.0 b	1543.0 ± 189.3 a
CW-MD	28.2 ± 6.6 d	nd	135.6 ± 41.5 a,b	20.3 ± 16.5 b	0.0 ± 0.0	7.7 ± 6.4 b	37.8 ± 29.3 b	145.2 ± 48.3 b	1834.3 ± 111.4 a
Citrus brandy (CB)
CB-MM	189.8 ± 68.2 b,c	36.2 ± 5.8 a	176.5 ± 56.7 a,b	222.4 ± 120.7 a	8.8 ± 1.2 a	91.6 ± 82.4 a,b	351.4 ± 280.1 a	1155.1 ± 631.6 a	998.5 ± 115.2 a
CB-MC	34.7 ± 22.1 c,d	13.9 ± 0.9 b	161.1 ± 86.7 a,b	196.8 ± 18.9 a	9.4 ± 4.4 a	59.6 ± 9.2 a,b	235.1 ± 26.2 a	1171.5 ± 491.7 a	1358.5 ± 410.5 a
CB-MD	1015.5 ± 375.4 a	31.1 ± 5.5 a	376.8 ± 142.8 a	211.4 ± 72.7 a	7.9 ± 2.7 a	83.3 ± 44.8 a	346.5 ± 132.8 a	1281.1 ± 671.4 a	1990.6 ± 1177.9 a

^a–d Superscript letters denote statistical groupings based on Tukey’s test (p < 0.05). Values sharing at least one common letter within the same parameter are not significantly different, indicate process homogeneity. MM: mature reference mandarin; MC: residual puffy mandarin; MD: residual damaged-peel mandarin; nd: not detected.

). These analytical parameters allow for evaluation of both the fermentative kinetics of the must and the mass transfer and concentration rates of volatile compounds during distillation, thereby identifying critical control points for technological valorization of the fractions.

In fermented citrus matrices, the global aromatic profile is dictated by the interaction of esters, alcohols, aldehydes, and terpenes, whose relative abundance depends on the nature of the raw material, the operational conditions, the yeast strain, and the specific processing method employed [36,37,38].

Ethyl acetate is a key ester in fermented matrices due to its prominent contribution to the fruity aroma profile [38]. In the citrus wine, its concentration ranged from 15.7 to 101.0 mg/L (Table 2), with the CW-MC sample exhibiting the statistically highest value. Following distillation, its concentration in the citrus brandy increased markedly in the reference and damaged-peel categories (CB-MM and CB-MD), with CB-MD exhibiting a statistically significant difference relative to the other distillate fractions for this parameter. This phenomenon is attributable to tissue stress activating alcohol acetyltransferase enzymes, thereby upregulating synthesis up to 1015.5 mg/L [39]. Conversely, the CB-MC fraction exhibited the opposite behavior, displaying an outlying value of 34.7 mg/L. This behavior stems from a molecular entrapment mechanism wherein released polyphenols form hydrogen bonds that bind the ester within the liquid phase of the still, thereby hindering its evaporation [37], consistent with the intermolecular behavior reported for similar acetate esters [40].

Furfural was not detected in the citrus wine and was quantified exclusively in the citrus brandy, with levels fluctuating between 13.9 and 36.2 mg/L (Table 2), indicating statistically homogeneous behavior across all three categories. Its appearance serves as a direct marker of thermal transformation and sugar dehydration—primarily derived from hemicellulosic fractions—which is a characteristic phenomenon during the advanced stages of pot distillation [18]. This aligns closely with the accumulation of this compound reported in mandarin byproducts exposed to high thermal regimens [36].

In the citrus wine, acetaldehyde concentrations ranged from 75.7 to 135.6 mg/L (Table 2) and were statistically comparable to one another. In the citrus brandy, these levels increased substantially to a range of 161.1 to 376.8 mg/L (Table 2), preserving the same distribution pattern observed in the base wine. This concentration effect is driven by its high volatility and subsequent recovery in the distillate fraction [18]. The higher numerical value recorded in the MD variant suggests the presence of prior oxidative processes or fermentative deviations tied to physical degradation of the raw material.

Higher alcohols displayed a clear trend toward thermal carryover and concentration in the distilled fraction. In the citrus wine, propanol levels were between 20.3 and 41.7 mg/L (Table 2) and increased in the brandy to a range of 196.8 to 222.4 mg/L (Table 2); in both matrices, variations among the conditions lacked statistical significance. Isobutanol and the amyl alcohol fraction (2- and 3-methyl-1-butanol) exhibited homologous behaviors, aligning with the fractionation dynamics established for viticultural distillates [18]. Synthesis of higher alcohols is primarily governed by fermentative metabolism of amino acids, followed by their volatility-driven mass transfer into the spirit. Arévalo-Ortiz et al. [18] reported higher concentrations of propanol, butanol, isobutanol, and 3-methyl-1-butanol/2-methyl-1-butanol in grape distillates than those observed in our mandarin citrus brandy, with the most pronounced difference appearing in the amyl alcohol fraction. This divergence is likely related to the baseline profile of nitrogen that yeast can assimilate from the raw materials and the specific fermentation parameters.

Methanol levels in the fermented fractions oscillated between 145.2 and 275.7 mg/L (Table 2). Following distillation, these values scaled critically to an interval of 1155.1 to 1281.1 mg/L (Table 2), showing statistical homogeneity among treatments in both processing stages. This sharp increase confirms that methanol constitutes a critical control point in the distillation of citrus matrices. Its biological origin stems from the enzymatic demethoxylation of structural cell wall pectins by endogenous pectin methylesterase during pre-fermentation and early processing [2]. Consequently, the presence of methanol demands rigorous monitoring during distillate fraction cutting (head/body/tail segregation) if the citrus brandy is intended for human consumption as a spirit beverage. Evaluations for compliance with international regulatory frameworks, such as Regulation (EU) 2019/787 [41], must strictly account for the specific product category, analytical expression limits, and moisture/ethanol bases used.

Volatile acidity, expressed as acetic acid equivalents, ranged from 857.8 to 1834.3 mg/L in the citrus wine (Table 2). In the citrus brandy, volatile acidity values stabilized between 998.5 and 1990.6 mg/L (Table 2), maintaining statistical homogeneity across all experimental conditions. These high-acidity profiles demonstrate that the initial physiological and physical condition of the postharvest discard drastically impacts the chemical profile of the final products. This underscores the need to optimize raw material sorting and conditioning strategies, particularly for the MC and MD streams, to prevent excessive volatile acid generation prior to bioconversion.

When contrasted with the NTP 212.014:2011 [42], which stipulates a maximum volatile acidity threshold of 1200 mg/L for fruit wines, only the CW-MM sample was compliant, whereas CW-MC and CW-MD exceeded the limit, although there is no significant difference between their values. Conversely, using Regulation (US) 27 CFR 5.145 [43] as a reference framework—which tolerates up to 2000 mg/L for certain distillate categories—the citrus brandy values remained technically below or at the limit of said standard.

The analytical parameters evaluated demonstrate that mandarin citrus brandy has viable technical potential for valorization as a distilled spirit. Nonetheless, its use as a high-quality spirit beverage depends on standardization of the initial must, optimization of fraction cuts during distillation, and rigorous control of regulated compounds such as methanol, acetaldehyde, and volatile acidity. The results validate the feasibility of the process but would require these technological variables prior to commercial scaling.

3.1.3. Phenolic Compounds, Flavonoids, and Antioxidant Capacity of Citrus Wine and Citrus Brandy

The bioactive potential of the citrus wine and citrus brandy was evaluated through total phenolic content (TPC), total flavonoid content (TFC), and in vitro antioxidant capacity assays (FRAP and DPPH) (Figure 3). Measuring these parameters enables comparison between the fermented and distilled fractions, facilitating estimation of the retention or mass transfer rates of functional compounds during distillation.

TPC values in the citrus wine exhibited variations depending on the raw material category, ranging between 721.9 and 1123.3 µg GAE/mL (Figure 3a). These results fall within or slightly above the magnitude range reported for mandarin juices and fermented products. Specifically, the MM value aligns with the range described by Kelebek et al. [44] for mandarin wines. Conversely, the contents in MC and MD were higher than that in MM, although no statistically significant differences were observed among the three categories (Figure 3a).

This variability is consistent with findings from Zhang et al. [37], who demonstrated high phenolic heterogeneity across different Citrus reticulata tissues and genotypes. A wide diversity of these compounds has been previously characterized in citrus matrices and byproducts: Kelebek et al. [44] identified hydroxybenzoic acids, hydroxycinnamic acids, and flavanones (predominantly vanillic acid, ferulic acid, and hesperidin); Kaur et al. [45] reported gallic, chlorogenic, p-coumaric, and ferulic acids, alongside hesperidin and quercetin, in pomace; and Mamy et al. and Chen et al. [46,47] quantified complex profiles in the pericarp via HPLC (including caffeic and chlorogenic acids, nobiletin, tangeretin, naringenin, narirutin, hesperetin, sinensetin, 4′,5,7,8-tetramethoxyflavone, and 5-O-desmethylnobiletin). These precedents provide context for the values obtained, although the specific phenolic contribution heavily depends on the cultivar, tissue type, processing method, and analytical technique. The higher TPC level observed in the MC citrus wine could be associated with the puffy condition of the fruit and a relatively higher contribution of phenol-rich tissues (such as peel, albedo, or pulp residues). Zhang et al. [48] demonstrated that the peel and juice have, on average, higher phenolic contents than other fractions. Furthermore, Zhan et al. [7] reported that the puffing disorder in Citrus reticulata induces differentiated metabolic shifts in phenolics and antioxidant capacity between the peel and pulp. This interpretation aligns with the general role of secondary metabolites in plant defense mechanisms against environmental stress [49].

Mechanistically, this stress-associated accumulation can be linked to upregulation of the phenylpropanoid pathway, particularly increased phenylalanine ammonia-lyase (PAL) activity. PAL is commonly activated in citrus tissues by chilling injury, mechanical wounding, and postharvest physiological disorders such as “puffing.” This results in de novo biosynthesis of phenolic precursors, which are released into the juice during pulp disruption and subsequently carried through fermentation, providing a biosynthetic rationale for the numerically higher TPC values observed in MCs [50].

In the citrus brandy, TPC decreased drastically to a range of 2.45 to 4.54 µg GAE/mL across the evaluated categories, displaying statistical homogeneity among treatments (Figure 3a). This sharp reduction is consistent with the predominantly non-volatile nature of citrus phenolic compounds, which tend to remain in the residual liquid phase (vinasse). In this regard, Tripoli et al. [51] describe these compounds as complex structures (glycosides or aglycones) with negligible mass transfer during distillation. Therefore, the detected values suggest a highly limited yet measurable carryover of low-molecular-weight reducing molecules into the distillate.

TFC in the citrus wine ranged from 56.88 to 78.43 µg QE/mL (Figure 3b). Unlike TPC, flavonoid contents were comparable between MMs and MCs, with the highest numerical value recorded in MDs; nevertheless, the three categories remained statistically similar (Figure 3b). The elevated value in MDs suggests that physical peel damage is associated with a relatively higher release of flavonoid compounds or substances reactive to the colorimetric method employed. This distribution is linked to the characteristic tissue heterogeneity of the species [48] and variations tied to the maturity stage [52].

Flavonoids represent a major fraction of citrus phenolics. Tripoli et al. and Hu et al. [51,52] identified profiles comprising flavanones (hesperidin, narirutin, didymin), flavones, flavonols, and polymethoxylated flavones (tangeretin, nobiletin, rutin, eupatorin, 5-O-desmethylnobiletin, among others), whose final concentrations are strictly dictated by the tissue, variety, and genotype of the fruit [48,53].

In the citrus brandy, TFC dropped to trace amounts and exhibited statistically homogeneous levels (0.190 to 0.323 µg QE/mL; Figure 3b). This behavior confirms the non-volatility of the complex flavonoids described by Tripoli et al. [51], verifying that they are almost entirely retained within the residual processing streams.

FRAP values for the citrus wine were statistically similar across all categories, spanning a range of 3.451 to 3.502 µmol TE/mL (MD) (Figure 3c). In citrus juices, vitamin C (ascorbic acid) accounts for a substantial proportion of the total antioxidant activity [54]. Similarly, global antioxidant capacity is heavily influenced by fruit maturity and the tissue origin of the must components [7,52].

As expected, the distillates (citrus brandy) exhibited considerably lower and homogeneous FRAP values (0.086 to 0.088 µmol TE/mL; Figure 3c). This drastic drop confirms the poor transfer of hydrophilic antioxidants (phenols, flavonoids, and vitamin C) into the distillate. This result aligns with reports of unaged spirits such as grappa or white rum, which exhibit negligible antioxidant capacities compared to wood-aged liquors [55] or in contrast to their respective base wines and vinasses [18,51]. From a technological standpoint, the primary reducing potential of the liquid pathway is concentrated in the citrus wine and the non-distilled streams.

DPPH radical scavenging activity in the citrus wine displayed no significant statistical differences among the evaluated categories, exhibiting values from 1.717 to 1.848 µmol TE/mL (Figure 3d). This uniformity confirms that the DPPH assay responds to the presence of phenolics, their specific chemical structures (electron- or hydrogen-donating capacity), and the synergistic action of other antioxidants within the matrix—variables that are heavily dependent on tissue type, maturity, and the physiological state of the fruit [7,44,48,52,54].

In the citrus brandy, DPPH activity decreased markedly to a statistically homogeneous range of 0.029 to 0.039 µmol TE/mL (Figure 3d). This sharp reduction is consistent with the TPC, TFC, and FRAP trends, corroborating that thermal distillation isolates volatile compounds while segregating the antioxidant fraction into the undistilled bottoms [44,55].

Collectively, the results from these analytical assays demonstrate that the citrus wine retains a substantial fraction of the mandarin’s original bioactive and antioxidant potential. Conversely, the citrus brandy exhibits a significantly lower residual capacity due to the low volatility and negligible mass transfer of polar phenolic and flavonoid compounds [7,44,51,52,55]. Therefore, within a comprehensive valorization framework for the liquid pathway, antioxidant interest is not focused on the main distillate; rather, it is strategically located in the base wine and, predictably, in the vinasse. The downstream characterization of the latter is imperative to justify its subsequent recovery as a byproduct rich in bioactive compounds.

3.2. Valorization of Distillation Byproducts

In addition to the citrus brandy as the primary product, distillation of the citrus wine generated three secondary processing streams with distinct valorization potentials: heads, tails, and vinasse. While the heads and tails constitute secondary distilled fractions, the vinasse represents the non-volatile liquid residue remaining in the pot still’s reboiler.

This segregation framework aligns with fractional distillation models utilized in traditional matrices, such as pisco production [17], as well as biorefinery schemes for the valorization of postharvest residues [18]. In these systems, the heart fraction is isolated as the commercially valuable product, whereas the adjacent fractions and liquid effluents are redirected toward specific ecological or industrial upcycling pathways.

Based on this premise, the heads and tails concentrate volatile compounds and prevent their direct incorporation into the main distillate; however, they retain high technical value for rectification processes or industrial alcohol recovery. Conversely, the vinasse constitutes the bulk liquid fraction; its pronounced acidity, residual volatile load, and eventual polyphenol concentration necessitate an evaluation targeted at the recovery of bioactive compounds with antioxidant properties [17,18].

3.2.1. Chemical Quality of Distillation Byproducts

The chromatographic and physicochemical profiles of the secondary processing streams revealed distinctly differentiated compositions (

Table 3. Chemical quality of mandarin distillation byproducts. The results are expressed as the mean ± standard deviation in mg/L of sample analyzed, n = 3.

By-Products	Ethyl Acetate (mg/L)	Furfural (mg/L)	Acetaldehyde (mg/L)	Propanol (mg/L)	Butanol (mg/L)	Isobutanol (mg/L)	2- y 3-Methyl-1-Butanol (mg/L)	Methanol (mg/L)	Acetic Acid (mg/L)
Head distillate (HD)
HD-MM	1037.0 ± 439.1 b	nd	454.1 ± 104.0 a,b	473.3 ± 170.8 a	20.6 ± 8.3 a	302.7 ± 168.7 a	1250.9 ± 384.5 a	1525.2 ± 857.1 a	677.3 ± 138.8 a
HD-MC	5238.6 ± 1895.1 a	nd	392.4 ± 8.5 a,b	591.2 ± 137.3 a	10.2 ± 5.5 a	406.2 ± 132.6 a	1627.2 ± 161.9 a	764.4 ± 374.0 a	1390.5 ± 1212.6 a
HD-MD	11,874.4 ± 4235.8 a	nd	1444.2 ± 550.2 a	389.5 ± 64.7 a	15.8 ± 10.0 a	276.0 ± 78.7 a	1313.4 ± 178.8 a	1660.1 ± 991.1 a	1042.9 ± 552.5 a
Tail distillate (TD)
TD-MM	32.6 ± 22.6 c,d	24.3 ± 20.5 a	143.7 ± 91.0 b,c	54.6 ± 32.1 b	nd	8.1 ± 3.0 b	29.0 ± 9.7 b,c	1010.6 ± 779.0 a	1206.2 ± 194.8 a
TD-MC	19.1 ± 2.1 c,d	nd	83.1 ± 4.3 c	97.5 ± 47.1 b	nd	15.6 ± 7.0 b	43.0 ± 26.6 b	946.1 ± 353.7 a	1735.2 ± 494.4 a
TD-MD	18.5 ± 0.5 d	20.9 ± 6.2 a	98.0 ± 3.0 c	87.2 ± 42.4 b	nd	10.8 ± 6.8 b	30.9 ± 11.4 b,c	967.3 ± 541.7 a	2508.5 ± 1738.9 a
Vinasse (v)
V-MM	49.4 ± 39.5 c	nd	119.8 ± 39.0 c	0.0 ± 0.0 d	nd	nd	12.4 ± 5.4 c,d	86.2 ± 26.4 b	1270.5 ± 183.4 a
V-MC	16.4 ± 0.0 d	nd	102.9 ± 16.6 c	10.5 ± 0.0 c	nd	nd	6.6 ± 0.1 d	105.3 ± 38.2 b	1486.2 ± 4.6 a
V-MD	34.0 ± 8.9 c,d	nd	129.7 ± 89.3 c	0.0 ± 0.0 d	nd	nd	0.0 ± 0.0 e	80.6 ± 59.6 b	2137.2 ± 1706.1 a

^a–d Superscript letters denote statistical groupings based on Tukey’s test (p < 0.05). Values sharing at least one common letter within the same parameter are not significantly different, indicate process homogeneity. MM: mature reference mandarin; MC: residual puffy mandarin; MD: residual damaged-peel mandarin; nd: not detected.

), underpinning the necessity of implementing independent management strategies for each stream.

As the initial distillate fraction, the heads concentrated the highest magnitudes of low-boiling volatile compounds, such as ethyl acetate and acetaldehyde. Ethyl acetate reached maximum values (5238.6 to 11,874.4 mg/L) in the HD-MC and HD-MD fractions, respectively, compared to the HD-MM fraction (1037.0 mg/L); however, acetaldehyde concentrations remained statistically homogeneous across all evaluated conditions.

This accumulation of light hydroalcoholic volatiles coincides with the distillation profiles described by Palma et al. and Arévalo-Ortiz et al. [17,18], who identified acetaldehyde and ethyl acetate as the major and limiting components of the initial fraction. Furthermore, Hidalgo et al. [56] demonstrated that the fermentative profile of the base wine directly dictates the mass transfer of these congeners into the distillate.

The pronounced dispersion and elevated values observed in HD-MC and HD-MD suggest that utilizing residual raw materials induces high volatile variability during fermentation, which directly impacts the head fraction. Given that the concentration of these compounds defines the organoleptic profile and commercial quality of distilled spirits [57], the head fraction must be excluded from the final citrus brandy and directed exclusively to controlled rectification processes or industrial applications.

The final distillate fraction, the tails, was established as a stream chemically opposite to the heads, characterized by the depletion of light volatiles and a higher proportion of acidic compounds. Consequently, it presented a drastic decline in ethyl acetate, exhibiting values between 18.5 and 32.6 mg/L (Table 3). Conversely, volatile acidity (as acetic acid) increased substantially, ranging from 1206.2 to 2508.5 mg/L (Table 3). Furfural was detected exclusively in the tails of MM and MD, remaining absent in TD-MC (Table 3). For all these parameters, the values exhibited statistical homogeneity among the treatments.

The late appearance of furfural and heavy organic acids is consistent with the literature [17,56], where furfural is described as a marker of the thermal degradation of pentoses and hemicellulosic components due to prolonged heating in the pot still’s reboiler. Due to its low alcohol content by volume and the presence of these undesirable compounds, the tail fraction is classified as a secondary byproduct suitable for rectification or external energy valorization.

Vinasse constituted the quantitatively dominant effluent of the process, accounting for approximately 90% of the initial volume charged into the still, with a marginal alcohol content (<0.5% v/v; Table 1). Chemically, it presented a distinct profile compared to the distilled fractions (p < 0.05), as validated by Tukey’s test for comparing means (Table 3). It exhibited minimal or undetectable methanol levels (80.6 to 105.3 mg/L; Table 3), differentiating it from the distilled streams. Acetaldehyde remained at measurable concentrations ranging from 102.9 to 129.7 mg/L (Table 3), distinguishing this effluent from the head fraction. Additionally, the vinasse exhibited high volatile acidity values, ranging from 1270.5 to 2137.2 mg/L (Table 3). Statistical analysis indicated that these parameters were homogeneous among the three raw material conditions.

According to Palma et al. and Poblete et al. [17,58], agroindustrial vinasses are characterized by high chemical oxygen demand (COD), acidic pH, intense coloration, and a significant load of residual polyphenols. Therefore, rather than being discarded as a conventional residue effluent, mandarin vinasse should be considered a matrix with bioactive potential. Although its high acidity and residual acetaldehyde pose operational challenges, the implementation of stabilization pretreatments, conditioning, or advanced oxidation processes [58] could pave the way for the selective recovery of natural antioxidants or the development of functional inputs.

The overall chemical balance confirms the heterogeneous nature of the obtained byproducts: the head fraction concentrates light volatiles, whereas the tails carry over acidity and thermal degradation products, allowing for integrated alcohol recovery via rectification. In contrast, the vinasse stands out due to its substantial volume and its role as a reservoir for non-volatile compounds with antioxidant potential. The specific characterization of this matrix via TPC, TFC, FRAP, and DPPH assays is discussed in detail in the subsequent section.

3.2.2. Bioactive Compounds and Antioxidant Capacity of Distillation Byproducts

The TPC, TFC, and antioxidant capacity (FRAP and DPPH) of the secondary processing streams were evaluated to determine their distinct valorization pathways based on matrix-specific compositional variations (Figure 3). While the distilled fractions (heads and tails) exhibited baseline or marginal levels across all assays, the vinasse concentrated a massive load of bioactive compounds and antioxidant activity. This asymmetric distribution is consistent with the non-volatile nature of polyphenols and reducing agents, which were retained in the residual liquid phase of the pot still’s reboiler rather than being partitioned into the vapor phases of the distillate. This behavior positions vinasse as a promising source of secondary metabolites for reintegration into agroindustrial circular economy frameworks [59].

TPC analysis confirmed marked chemical segregation among the byproducts. The head (HD) and tail (TD) fractions exhibited low and statistically homogeneous concentrations, ranging between 2.306 and 3.331 µg GAE/mL for HD and between 1.816 and 2.867 µg GAE/mL for TD (Figure 3a). This minimal presence is expected in a stream dominated by low-boiling volatile compounds; nonetheless, the subtle detection of TPC in light distillates is typically attributed to microscopic physical carryover during boiling or to the non-specific nature of the Folin–Ciocalteu assay, which responds to the global reducing capacity of the medium rather than exclusively to phenolic structures [60].

In contrast, the vinasse accumulated the highest levels of TPC, ranging from 958.623 to 1339.109 µg GAE/mL (Figure 3a). These results confirm that the liquid residue retains the bulk of the antioxidants present in the initial must. This profile is consistent with traditional distillery vinasses (such as those from grape pisco), which simultaneously represent an effluent with a high organic load and acidity and a matrix with dual relevance to the environment and to biorefineries, suitable for the selective recovery of polyphenols via conventional or assisted extraction techniques [17,19,20,58].

TFC followed a homologous trend, further demonstrating the complete exclusion of these hydrophilic macromolecules from the vapor streams. In the head fraction, values fluctuated moderately between 1.275 and 2.080 µg QE/mL, whereas in the tail fraction, they dropped to baseline levels (0.119 to 0.206 µg QE/mL). Conversely, the analytical response in the vinasse increased substantially, reaching a range of 58.769 to 94.149 µg QE/mL (Figure 3b), displaying statistical homogeneity due to intrinsic batch variability.

Notably, the higher retention of TFC in the V-MD category suggests that processing fruits with physical peel damage may promote the hydrolysis or massive release of stable flavonoids into the must, resulting in their subsequent accumulation within the distillation residue. In mandarin matrices, these values are associated with characteristic polyphenols such as hesperidin, naringin, hesperetin, and rutin, reaffirming the value of this citrus byproduct as an extraction platform [59].

Indirect antioxidant capacity assays validated the high bioactive potential concentrated in the reboiler residue. In the FRAP assay, distilled streams yielded homogeneous values ranging from 0.074 to 0.086 µmol TE/mL for heads and 0.079 to 0.099 µmol TE/mL for tails, whereas the vinasse exhibited a robust ferric reducing capacity of 2.989 to 3.489 µmol TE/mL, with no statistically significant differences among treatments (Figure 3c). Both the distilled streams and the vinasse showed statistical homogeneity within their respective matrix groups. The highest numerical response observed in V-MC correlates directly with the elevated TPC levels obtained for the same category, serving as a complementary indicator of the reducing efficiency of the matrix [60].

The DPPH radical scavenging activity displayed an analogous trend. The distilled fractions (HD and TD) yielded baseline values between 0.029 and 0.049 µmol TE/mL, whereas the vinasse exhibited a significantly higher inhibition capacity, reaching values between 1.136 and 1.414 µmol TE/mL (Figure 3d). In all cases, variations among raw material categories within the same matrix were statistically non-significant. According to Prior et al. [60], a multi-method analytical approach (FRAP and DPPH) is indispensable for evaluating complex matrices, as each assay operates via distinct mechanisms—single-electron transfer vs. hydrogen atom transfer—thereby providing an integrated validation of the antioxidant profile of the vinasse compared to the volatile fractions.

In summary, the functional characterization of the byproducts supports a clear stream segregation strategy. The head and tail fractions, characterized by marginal responses in TPC, TFC, FRAP, and DPPH assays, are unsuited for antioxidant extraction, reaffirming their valorization pathway toward rectification and industrial alcohol recovery. Conversely, mandarin vinasse is established as a highly bioactive residual liquid matrix. Its rich content of phenolic and flavonoid compounds, backed by notable reducing and radical-scavenging capacity, offers a viable route for upcycling via conventional or assisted extraction technologies (e.g., ultrasound or microwave-assisted extraction). This approach simultaneously mitigates the environmental impact of the effluent while transforming a critical residue stream into a high-value input for the circular bioeconomy [20,58,59].

Principal Component Analysis (PCA) of the data matrix transformed via log(X + 1) provided an integrated evaluation of the global interrelationships between the physicochemical and bioactive parameters (active variables) and the five fractions from the distillation process (active observations) (Figure 4). The first principal component (F1) explained the largest proportion of the total variance (66.53%), operating as a differentiation axis driven by component volatility. F1 exhibited highly positive loadings for alcohol by volume and volatile congeners, including ethyl acetate, acetaldehyde, methanol, and higher alcohols (propanol, butanol, isobutanol, and the methyl-1-butanol isomers). Conversely, this same axis displayed strongly negative loadings for bioactive compounds (TPC, TFC) and antioxidant capacity assays (FRAP and DPPH). This opposing distribution mathematically describes an inversely proportional relationship that reflects the intrinsic physicochemical segregation of distillation, wherein compounds with reducing power are excluded from volatile-rich streams.

Meanwhile, the second principal component (F2, 15.29%) was governed almost exclusively by furfural, whose positive loadings dominated this axis. The selective projection of the tail fraction (TD) toward the positive extreme of F2 confirms the late accumulation of this compound during the terminal stages of the process. From a mechanistic perspective, the genesis and concentration of furfural toward the end of the distillation are directly associated with thermal degradation and dehydration of the hexoses remaining in the pot still’s reboiler, which are intensified by the sustained high temperatures characteristic of this stage.

The two-dimensional biplot of the observations confirmed sharp functional segregation of the analyzed matrices. The non-distilled liquid phases, represented by the citrus wine (CW) and the vinasse (V), clustered within the negative region of F1, strongly associated with elevated levels of TPC, TFC, and antioxidant capacity. In contrast, the distilled streams positioned themselves within the positive space of F1, highlighting the absolute predominance of the volatile fraction over the hydrophilic one. Specifically, the heads (HD) displayed a positive correlation with higher alcohols, confirming the preferential volatility of these low-boiling, hydrophobic congeners during the initial minutes of distillation.

Finally, the close proximity and partial overlap of the samples derived from mandarin postharvest residues (MC and MD) relative to the reference condition (MM) demonstrate that the biological and physicochemical properties of the discard streams are homologous to those of the commercial fruit. This behavior validates the robustness of these residual raw materials against the chronological and seasonal variability of the three evaluated batches, supporting their technical viability for uniform and reproducible valorization.

3.2.3. Chemical Profiling of Vinasse via LC-MS

The chemical profile of the vinasse was analyzed via LC-MS to complement the spectrophotometric (TPC, TFC) and antioxidant capacity (FRAP, DPPH) assays. As detailed in

Table 4. Compounds identified and quantified by LC-MS in mandarin vinasse. The results are expressed as the mean ± standard deviation in mg/kg of sample analyzed, n = 3.

Compound	Retention Time (min)	Limit of Quantification (mg/kg)	V-MM (mg/kg)	V-MC (mg/kg)	V-MD (mg/kg)
Phenolic Acids
Caffeic acid	4.1	0.1	0.22 ± 0.08 a	<0.1	0.11 ± 0.01 b
Ferulic acid	5.5	0.1	<0.1	<0.1	<0.1
p-Coumaric acid	4.7	0.1	0.44 ± 0.05 a	<0.1	0.44 ± 0.05 a
p-Hydroxybenzoic acid	3.7	0.1	2.13 ± 0.25 a	<0.1	2.07 ± 0.31 a
Rosmarinic acid	5.5	2.0	<2.0	<2.0	<2.0
trans-Cinnamic acid	6.1	0.4	<0.4	<0.4	<0.4
Vanillic acid	4.5	0.1	<0.1	<0.1	<0.1
Flavonoids
Apigenin	6.7	0.1	1.03 ± 0.15 b	3.23 ± 0.25 a	1.43 ± 0.40 b
Epicatechin	3.7	0.1	<0.1	<0.1	<0.1
Epicatechin gallate	4.9	0.1	<0.1	<0.1	<0.1
Epigallocatechin gallate	4.7	0.1	<0.1	<0.1	<0.1
Luteolin	6.4	0.1	<0.1	<0.1	<0.1
Naringenin	6.2	0.1	30.87 ± 1.00 b	95.87 ± 2.51 a	31.80 ± 1.44 b
Pinocembrin	7.1	0.1	<0.1	0.29 ± 0.10	<0.1
Quercetin	6.3	0.1	<0.1	<0.1	<0.1
Rutin	5.3	0.1	<0.1	21.87 ± 1.67 a	5.20 ± 0.20 b
Triterpenoids
Ursolic acid	9.4	0.1	<0.1	<0.1	<0.1
Alkaloids and methylxanthines
Caffeine	4.1	0.1	<0.1	<0.1	<0.1
Theobromine	3.2	0.1	<0.1	<0.1	<0.1
Theophylline	3.6	0.1	<0.1	<0.1	<0.1

^a,b Superscript letters denote statistical groupings based on Tukey’s test (p < 0.05). Values sharing at least one common letter within the same compound are not significantly different, indicate homogeneity process. V: vinasse; MM: mature reference mandarin; MC: residual puffy mandarin; MD: residual damaged-peel mandarin; < indicates a concentration below the corresponding limit of quantification.

, the residual stream retains an array of phenolic acids and flavonoids metabolically linked to the citrus matrix, confirming its viability as a platform for phytochemical valorization [16,59].

Naringenin was established as the predominantly retained flavonoid across all evaluated samples, with the V-MC treatment displaying a substantial (p < 0.05), order-of-magnitude increase relative to the mature reference (V-MM) and damaged-peel (V-MD) categorized (Table 4). This phenomenon is governed by a dual biochemical and hydrothermal mechanism. First, the compartmentalized distribution of native flavonoids within citrus fruit tissues (flavedo, albedo, and juice sacs)—where glycosylated flavanones are predominantly concentrated in the albedo and segment membranes [16,61]—is altered by the physiological stress induced by puffing in puffy fruits, activating phenylpropanoid defense pathways and upregulating synthesis of these metabolites. Second, during prolonged boiling regimes in the pot still’s boiler, acidic hydrothermal conditions serve as catalysts for kinetic hydrolysis. This process efficiently cleaves the glycosidic bonds of precursor flavonoids (such as narirutin), promoting their deglycosylation into free aglycones like naringenin—a behavior characteristic of flavanones subjected to intensive industrial thermal processing [62]. Mechanistically, free aglycones possess high thermal stability owing to their planar structure and the absence of labile sugar moieties, requiring critical degradation temperatures far exceeding operational limits (~190 °C), thus explaining their persistence and high accumulation in the final vinasse residue [63].

In addition to its prominent naringenin content, the profile of the V-MC vinasse was distinguished by a higher abundance of other polyphenolic entities, yielding apigenin, rutin, and pinocembrin. In contrast, the reference control condition (V-MM) exhibited lower levels of apigenin and concentrations below the limit of quantification (LOQ) for pinocembrin and rutin. Meanwhile, V-MD presented an intermediate profile. This behavior confirms that outer fruit integrity and postharvest conditions drastically modulate the final polyphenolic profile recoverable in the liquid effluent [59].

A plausible mechanistic explanation for this enrichment in V-MC involves puffing-associated detachment between the flavedo and pulp, which increases tissue permeability and exposes flavonoid glycosides to endogenous and yeast-derived β-glucosidases during fermentation. The resulting enzymatic deglycosylation facilitates the release of naringenin and other aglycones from precursors such as narirutin, elevating their apparent concentration in the juice prior to distillation, alongside the aforementioned thermal stability effect [64].

Unlike the flavonoids, low-molecular-weight phenolic acids were present at marginal concentrations and exhibited an inverse trend (Table 4). While p-hydroxybenzoic, p-coumaric, and caffeic acids were quantifiable in the V-MM and V-MD streams, all evaluated phenolic acids fell below the LOQ in the V-MC (Table 4). This exclusion suggests that, in puffy mandarins, the metabolic pathway of the fruit prioritized condensation toward structurally more complex flavonoids; or, alternatively, that these acids underwent differential thermal degradation or polymerization during distillation.

Integration of these molecular results with the global indices (TPC, TFC, FRAP, and DPPH) provides a robust chemometric overview. While colorimetric and antioxidant assays express the overall reducing response of the matrix—influenced by the specific chemical mechanisms and kinetic limitations discussed by Prior et al. [60]—the LC-MS analysis successfully identifies the individual effectors driving this bioactivity. This consolidates naringenin, apigenin, and rutin as the primary functional pillars underlying the antioxidant potential of the vinasse.

3.2.4. Valorization Potential of Mandarin Vinasse

The analytical findings demonstrate that mandarin vinasse transcends the conventional definition of an industrial residue stream, emerging instead as an intermediate matrix of high phytochemical value. However, its transition toward a practical upcycling framework requires a comprehensive technological perspective. The elevated concentration of bioactive compounds coexists with marked acidity, a high organic load, and traces of residual volatiles (such as methanol and acetic acid). As cautioned by Poblete et al. [58] for homologous matrices (e.g., grape pisco vinasses), the intense coloration and high oxygen demand necessitate physicochemical conditioning prior to final disposal. This imparts dual significance to the stream: the recovery of high-value functional compounds and controlled environmental management.

Based on this premise, a cascade valorization strategy structured into two sequential phases is proposed. The first stage focuses on the extraction, concentration, and stabilization of the flavonoid and phenolic acid fractions, capitalizing on the notable molecular richness identified, particularly in the V-MC category. Robust precedents exist for upcycling of distillery effluents (distillery vinasse), demonstrating the technical feasibility of recovering polyphenols with high antioxidant activity via conventional extraction methods or advanced technologies such as ultrasound- and microwave-assisted extraction [19,20]. Once stripped of its polyphenolic load, the remaining liquid stream undergoes a second phase of targeted environmental treatment to optimize the reduction of oxygen demand, color, and residual acidity.

This cascade fractionation approach systematically coordinates the industrial destination of all streams generated along the distillery’s liquid pathway. Consequently, three functional destinations are established, preventing the byproducts from being treated as a homogeneous residues mass. Thus, citrus brandy is obtained as the primary commercial product for consumption; the heads and tails are isolated as secondary distilled fractions directed toward controlled industrial alcohol recovery or rectification under strict quality control criteria; and the vinasse serves as the residual liquid platform for the extraction of natural antioxidants for technological applications, followed by effluent remediation.

Finally, this circular bioeconomy model is complemented by the process’s solid pathway. The residual mandarin peel, acting as a primary byproduct, can be processed in parallel for the selective recovery of high market value aromatic and structural compounds—principally essential oils and pectins—thereby maximizing the material efficiency and overall yield of the citrus biomass [59].

3.3. Integrated Valorization of Mandarin Peels and Residual Fractions

Comprehensive valorization of residual mandarin biomass requires strategic utilization of its peel—a highly complex byproduct owing to the anatomical and chemical heterogeneity of its tissues. The flavedo, or outer layer, contains oil glands that concentrate volatile metabolites responsible for the signature citrus aroma profile, whereas the albedo and internal cellular structures are rich in cell wall polysaccharides, primarily pectins. This tissue compartmentalization establishes the development of complementary and sequential fractionation routes: the hydrodistillation of essential oils, followed by the subsequent extraction of pectins, which synergistically integrates with the previously analyzed liquid pathway.

This approach aligns with the circular economy models described by Russo et al. [59], who highlight the value of citrus residues as multi-specific sources of polymers, fibers, and phytochemicals of industrial interest. Furthermore, it validates the implementation of integrated co-recovery strategies, such as microwave-assisted extractions coupled with deep eutectic solvents (DESs), which have been successfully applied to homologous matrices to simultaneously isolate oils and pectins [65]. In the present study, the solid biomass is organized under this comprehensive valorization logic, directing each fraction toward a technological alternative consistent with its native composition.

3.3.1. Recovery of Essential Oils

Extraction of the volatile fraction from the flavedo represents the initial stage in cascade valorization of the peel. This enables the isolation of high market value hydrophobic compounds for the food, cosmetic, and pharmaceutical industries before subjecting the solid matrix to more severe hydrolytic treatments. In this work, the essential oil yields obtained via steam distillation were 0.85 ± 0.25 mL/100 g for the reference mature mandarin peel (C-MM), 0.73 ± 0.09 mL/100 g for the residual puffy mandarin peel (C-MC), and 0.575 ± 0.27 mL/100 g for the damaged residual mandarin peel (C-MD) (Figure 5).

The progressive decline in volatile recovery among the residual samples, which was particularly critical in C-MD, suggests that macroscopic mechanical damage, postharvest physiological stress, and cellular senescence compromise the structural integrity of the oil glands, thereby inducing premature volatilization or oxidative degradation of the terpenes. At the biochemical level, mechanical gland disruption can also activate endogenous oxidative enzymes, particularly lipoxygenases, which co-oxidize unsaturated lipids and terpenoid substrates. These enzymes are typically upregulated by mechanical wounding and tissue senescence in fruit, providing a plausible enzymatic mechanism that compounds the physical loss of volatile oil in the MD category [66].

From a qualitative standpoint, the chemical profile of these essential oils is dominated by monoterpene hydrocarbons, with d-limonene serving as the majority constituent, accompanied by a network of minor congeners that configure its aromatic identity, such as γ-terpinene, α-pinene, β-pinene, linalool, and decanal [67,68]. Specifically, Navarrete et al. [67] reported standardized mandarin profiles containing 70.8% d-limonene, followed by β-pinene (11.2%), γ-terpinene (8.3%), α-pinene (4.4%), linalool (2.3%), α-terpineol (1.1%), terpinolene (0.9%), and decanal (0.9%). This terpenic signature bears close congruence with reports by Wang et al. [69] for residual peels from Citrus reticulata var. depressa, where d-limonene, γ-terpinene, p-cimene, and linalool governed the volatile fraction, as well as with the characterizations of Citrus reticulata provided by Wen et al. [68], who additionally identified key odor descriptors such as 3-carene and thymol.

Beyond their volatile terpenic fraction, the oily matrix derived from Citrus unshiu peel also contains co-extracted non-volatile phenolic compounds that substantially enhance its overall functional value. Recent evidence identifies quercetin (2084.67 µg/mL) as the predominant phytocompound, which is co-extracted with essential hydroxycinnamic acids such as p-coumaric acid (47.43 µg/mL) and ferulic acid (28.37 µg/mL), thereby supporting its high intrinsic antioxidant capacity [70]. The viability of incorporating these oil fractions into commercial matrices depends upon their oxidative stability. Despite the inherent lability of monoterpenes, advanced stabilization strategies—such as nanoencapsulation or synergistic co-formulation with complementary essential oils (e.g., cinnamon or clove)—effectively mitigate photooxidation phenomena, expanding their biocidal and antioxidant potential [71].

As a final mass transfer vector within this valorization chain, these stable oils can be effectively integrated into biodegradable polymeric matrices (e.g., avocado seed starch) for the development of active packaging materials. The addition of the volatile fraction of Citrus unshiu not only preserves the antioxidant capacity of the protective material but also substantially optimizes its mechanical properties, elevating the tensile strength up to 3.3 MPa and enhancing structural rigidity, all without altering its progressive biodegradability profile (with mass losses exceeding 55% over a 25-day period) [70].

Within this context, prior recovery of the aromatic fraction not only yields a high-value industrial input but also conditions and degreases the residual albedo. This pre-conditioning step removes interfering terpenes, optimizing the solid matrix for subsequent pectin extraction and purification [65].

3.3.2. Recovery and Characterization of Mandarin Peel Pectins

The remaining solid fraction after essential oil hydrodistillation was valorized through the extraction of pectin, a structural polysaccharide of extensive industrial interest due to its gelling, thickening, and stabilizing properties. Structurally, pectin consists of a linear backbone of α-(1→4) linked D-galacturonic acid (GalA) units (the homogalacturonan domain, HG), which is interrupted and branched by rhamnose residues carrying neutral sugar side chains (the rhamnogalacturonan I domain, RG-I) [72]. To elucidate the molecular organization and purity levels of the fractions recovered from the different raw material categories, advanced compositional characterization was integrated with Fourier-transform infrared (FTIR) spectroscopy [29,30].

The FTIR spectra confirmed the typical vibrational footprint of pectic structures across all analyzed samples (Figure 6). Characteristic bands were identified at 3325 cm⁻¹, associated with hydroxyl group (O–H) stretching, and at 2900 cm⁻¹, attributed to the tensile vibrations of aliphatic C–H bonds, which match assignments described by Wang et al. [29]. In the critical carboxyl group region, a prominent band was observed around 1730 cm⁻¹, corresponding to methyl-esterified carbonyl (C=O) stretching, along with an adjacent signal near 1610 cm⁻¹ assigned to the asymmetric vibrations of the free carboxylate group (COO⁻); this behavior is fully consistent with the footprints reported by Chasquibol et al. [73].

Additionally, signals located at 1430, 1240, and 1020 cm⁻¹ were attributed to the bending and stretching vibrations of C–O and C–O–C bonds within the glycosidic backbone of the polysaccharide. The intensity ratio between the esterified and free carboxyl bands is fundamental, as it allows for qualitative and quantitative estimation of the degree of esterification, thereby anticipating the macroscopic behavior of the polymer in solution [29,73].

The detailed molecular compositional characterization presented in

Table 5. Chemical composition and molecular characteristics of pectins obtained from mandarin peel. Data are presented as mean ± standard deviation, n = 3.

	MM	MC	MD
Pectin %	19.52 ± 1.51	23.27 ± 2.87	18.56 ± 1.23
Moisture %	12.35 ± 0.01	12.11 ± 0.01	11.98 ± 0.67
Protein %	4.23 ± 0.33	3.94 ± 0.16	3.75 ± 0.29
Phenolics %	0.89 ± 0.01	0.66 ± 0.01	0.75 ± 0.02
Ash %	12.01	12.32	13.23
Monosaccharides (relative %)
Rhamnose (Rha)	1.4 ± 0.1	1.9 ± 0.1	4.1 ± 0.2
Arabinose (Ara)	1.3 ± 0.1	1.8 ± 0.1	8.6 ± 0.4
Xylose (Xyl)	0.5 ± 0.0	1.2 ± 0.1	1.2 ± 0.1
Mannose (Man)	0.5 ± 0.0	1.1 ± 0.1	3.0 ± 0.1
Galactose (Gal)	3.3 ± 0.0	11.1 ± 0.1	13.8 ± 0.6
Glucose (Glc)	3.2 ± 0.1	4.3 ± 0.1	8.7 ± 0.4
Galacturonic acid (GalA)	89.8 ± 0.1	78.6 ± 0.3	60.6 ± 1.8
Homogalacturonan (%)	88.4 ± 0.1	76.7 ± 0.2	56.5 ± 1.4
Ramnogalacturonan I (%)	7.4 ± 0.1	16.7 ± 0.1	30.6 ± 0.4
(Ara + Gal)/Rha	3.29 ± 0.1	6.79 ± 0.1	5.46 ± 0.4
Degree of methylation (%)	44.90 ± 0.10	38.73 ± 0.31	40.23 ± 0.12
Degree of acetylation (%)	2.23 ± 0.12	2.70 ± 0.01	2.03 ± 0.06
Weight-average molecular weight (Mw) (g/mol)	3.621 × 104	1.80 × 105	1.35 × 105
Polydispersity index (Mw/Mn)	1.81 ± 0.08	3.15 ± 0.17	2.78 ± 0.21

MM: mature reference mandarin; MC: residual puffy mandarin; MD: residual damaged-peel mandarin; number-average molecular weight: Mn.

demonstrated that the initial condition of the raw material had a drastic effect on pectin architecture. The GalA content declined progressively as fruit quality deteriorated, reaching a maximum of 89.8% for MM, decreasing to 78.6% for MC, and dropping to a minimum of 60.6% for MD. This trend corroborates a severe contraction of the linear HG domain, which accounted for 88.4% in MM, 76.7% in MC, and 56.5% in MD. From a commercial quality standpoint, FAO and European Union regulations establish a minimum threshold of 65% GalA for food-grade pectins [74]. While the MM and MC fractions easily exceed this standard, the MD pectin falls below the limit, suggesting either partial degradation of the linear backbone or higher co-extraction of cell wall components rich in neutral sugars.

Correspondingly, the proportion of the branched RG-I domain increased notably in the residual samples, exhibiting values of 7.4% (MM), 16.7% (MC), and 30.6% (MD). This behavior correlates with the higher relative abundance of neutral branching sugars (arabinose, Ara; galactose, Gal) and other constituent monosaccharides (Rha, Xyl, Man, and Glc) identified in the MD fraction. According to Yapo and Koffi, as well as Petkowicz [30,72], the molecular balance between the HG and RG-I domains governs critical technofunctional properties such as solubility, intrinsic viscosity, and intermolecular interaction capacity. This behavior is attributed to a selective degradation mechanism during thermal extraction. In the MD sample, operational conditions favored controlled hydrolysis that preserved the HG backbone—reaching 60.6% GalA—while the arabinogalactan and galactan side chains within the branched domain underwent partial cleavage, stabilizing the RG-I proportion at 30.6%.

As reported by Yapo et al. [72], the sensitivity of glycosidic bonds in the branched regions of pectin varies significantly depending on the isolation method, wherein the combination of temperature and duration can induce depolymerization of neutral constituents without drastically altering the galacturonic acid core. This phenomenon accounts for the observed structural rigidity and the efficient retention of functional domains within the macromolecular matrix obtained with this treatment.

Regardless of the rind condition, all fractions were classified as low-methoxyl pectins (degree of methoxylation, DM < 50%), yielding values of 44.90% in MM, 38.73% in MC, and 40.23% in MD. This structural status potentially enables them to undergo gelling processes induced by calcium-mediated ionic cross-linking, a behavior that remains conditional upon cooperative factors such as molar mass and molecular architecture [11,72]. Concurrently, the degree of acetylation (DA) remained consistently low throughout the entire experimental design (2.23% in MM, 2.70% in MC, and 2.0% in MD), falling within ranges consistent with recent Citrus reticulata characterizations reported by Chen et al. [11] (1.09–2.30%). Mechanistically, a low degree of acetylation is highly favorable from a functional standpoint, since the absence of steric hindrance from acetyl groups facilitates the alignment and packing of polymer chains during formation of the gelled network [29,30].

Regarding hydrodynamic properties, the weight-average molecular weight (Mw) values exhibited marked differences. According to Table 5, the MM pectin exhibited an Mw of 3.621 × 10⁴ g/mol, whereas the residual fractions MC and MD presented significantly higher magnitudes of 1.794 × 10⁵ g/mol and 1.349 × 10⁵ g/mol, respectively. This indicates that tissue stress or postharvest discard phenomena did not necessarily lead to scission of the polymer chains; on the contrary, using the extraction method employed, they permitted solubilization of larger macromolecular aggregates. However, the residual samples also displayed substantially higher polydispersity indices (Mw/Mn) (3.153 in MC and 2.779 in MD, compared to 1.811 in MM), denoting a more heterogeneous and polydisperse molecular population [30,72]. The functionality of a pectin does not depend solely on its molar mass, but rather on the combination of its structural attributes. Along those lines, Chen et al. [11] showed that differences in molar mass, degree of methoxylation, acetylation, and the ratio of structural domains collectively dictate the behavior of Citrus reticulata pectins within emulsion systems.

Beyond compositional characterization, these structural transitions dictate the technological performance of the recovered pectins. Calcium-mediated gelation in low-methoxyl pectins occurs via the classic “egg–box” model, wherein divalent cations coordinate linear stretches of deprotonated GalA residues within the HG domain [75]. Thus, the extensive, homogeneous HG blocks of MM pectin (88.4% HG and lower polydispersity) favor rigid, ordered networks suitable for conventional food gels, such as jams, dairy products, or confectionery. In contrast, the enrichment of RG-I side chains and neutral sugars in MC and MD interrupts HG domain continuity, reducing calcium gel firmness. However, this branched architecture with higher molar mass—coupled with the co-extraction of proteins and phenolics—confers superior surface properties upon MC and MD for the steric stabilization of emulsions, film formation, and encapsulation. Therefore, the divergence caused by postharvest deterioration does not degrade polymer value but rather diversifies its industrial applications: MM as a traditional gelling agent and MC/MD as emulsifiers and functional matrices for packaging or encapsulation.

The analysis presented in Table 5 also identified the coexistence of minor constituents integrated into the polymer matrix. Protein content ranged from 4.23% (MM) to 3.94% (MC) and 3.75% (MD), while the residual phenolic fractions were 0.89%, 0.66%, and 0.75%, respectively, alongside stable ash levels between 12.01% and 13.23%. The retention of these proteinaceous and phenolic entities adds complexity to pectin functionality; M. Chen et al. and H. Chen et al. [11,76] demonstrated that the nature of protein–polyphenol–pectin complexes plays a determining role in stabilizing oil-in-water (O/W) emulsions by modifying interfacial tension and acting as steric barriers. Consequently, the presence of these minor fractions expands the potential applications of residual mandarin pectins in advanced food technology fields to roles such as vehicle systems, stabilization, and emulsification.

In summary, the physiological condition of the mandarin rind directly modulates the yield, elemental composition, and branching profiles of the recovered pectin. This extraction pathway for structural polysaccharides successfully rounds out the valorization of the solid fraction within a cascade biorefinery scheme. The technological sequencing—oriented toward preliminary degreasing via essential oil recovery followed by controlled depolymerization of the cell wall to obtain pectins—replicates cutting-edge global trends in the byproduct bioeconomy of Citrus reticulata, Citrus paradisi, and Citrus unshiu [11,65,77]. This comprehensive utilization model not only mitigates the distillery’s environmental liability but also efficiently segments the residual biomass into high-value commercial streams, harmoniously complementing the previously analyzed liquid pathways.

3.3.3. Integration of Valorization Pathways Within a Cascade Biorefinery Framework

Unlike conventional citrus valorization approaches focused solely on sequential extraction of the peel to yield essential oils, flavonoids, and pectin [78,79] or limonene and activated carbon [80], this study coherently integrates both the solid and liquid fractions of discarded Satsuma Owari mandarins (Citrus unshiu). On the one hand, the solid fraction was valorized via sequential recovery of essential oil and pectin; on the other hand, the liquid fraction was upgraded through fermentation and distillation, generating a fine spirit (citrus brandy) and a vinasse stream with a high phenolic density predominantly characterized by naringenin. This dual-pathway experimental design fits within integrated biorefinery models for citrus waste [81] while introducing a novel perspective: the fruit’s postharvest physiological state serves as a critical modulating factor governing overall valorization performance.

In the liquid pathway, the initial juice fraction is funneled toward biotransformation for the production of citrus wine and citrus brandy. Consequently, the remaining vinasse from the distillation process retains a considerable amount of phenolic compounds and high antioxidant capacity, which underscores its potential as an intermediate matrix for the recovery of value-added phytocompounds or, alternatively, for the design of environmental conditioning strategies. This interpretation converges with findings from Mikucka and Zielińska and Mikucka et al. [19,20], who described the technical feasibility of isolating polyphenols and antioxidants from liquid distillery effluents. Furthermore, it is complemented by the perspective of Poblete et al. [58], who emphasized the critical need to treat vinasses to mitigate organic load, color, and phenolic phytotoxicity prior to final disposal or industrial reuse.

In parallel, the fruit peel serves as the core of the solid processing pathway. The anatomical heterogeneity of this matrix allows the flavedo to be directed toward the extraction of essential oils rich in volatile terpenes, while the albedo and adjacent structural tissues are preserved for the hydrolytic recovery of pectins. This selective fractionation overcomes the operational limitations of treating residual biomass as a homogeneous substrate, optimizing process allocation according to the chemical nature of each metabolite. This integrated approach validates the conceptualization of citrus byproducts as multi-specific sources capable of simultaneously providing functional polymers, essential oils, flavonoids, carotenoids, and dietary fiber intended for the food, pharmaceutical, and biotechnological industries [59].

Conventional citrus waste management is generally limited to linear pectin extraction or direct drying for animal feed—practices that underutilize the waste’s molecular potential and generate effluents with high BOD. In contrast, the proposed cascade biorefinery strategy implements a zero-waste framework that maximizes value at each processing tier. First, soluble sugars are converted into value-added beverages; second, distillation is repurposed as a hydrothermal treatment that promotes the deglycosylation and concentration of naringenin in the vinasse; and finally, the residual solid fraction is exhaustively processed to recover functional low-methoxyl pectins. This technological coupling demonstrates that even fruits affected by physiological disorders or physical damage (such as the MC category) retain structural and molecular matrices suitable for reincorporation into closed-loop supply chains, thereby mitigating environmental impacts and reinforcing agroindustrial economic viability.

From an industrial scaling perspective, this cascade biorefinery approach exhibits strong implementation potential, as it relies on unit operations standard to the agro-food industry (including pulping, filtration, fermentation, distillation, steam distillation, acid extraction, drying, and milling). Nevertheless, its economic feasibility will depend on boundary conditions such as seasonal biomass availability, efficiency in segregating the MC and MD fractions, the actual yields of each product stream, and the energy consumption associated with thermal operations and ethanol recovery during pectin precipitation. Therefore, the systemic integration of these pathways does not mandate execution within a single operational facility; rather, it offers processing flexibility consistent with models assisted by emerging technologies, such as microwaves and acidic deep eutectic solvents (A-DESs) [11,65].

As a study limitation, the inherent variability observed in the initial parameters of juices and intermediate streams must be acknowledged. This variability stems from the natural heterogeneity of discarded postharvest fruit compared to standardized industrial feedstocks. It is linked to variations in maturity stage, tissue integrity, and harvest timing across batches and may reduce estimation precision for certain quantitative variables. Nonetheless, despite this natural variability, the overarching trends identified across replicates support the consistency and robustness of the valorization scheme under the evaluated conditions. Future research should expand the number of sampled batches to validate reproducibility across broader commercial scenarios.

Collectively, the comprehensive valorization of residual mandarins establishes a robust technological platform tailored for the strategic, differentiated utilization of discarded fruit. By harmoniously coupling fermentative–distillative biotransformation, bioactive recovery from liquid effluents, and macromolecular peel fractionation, this work positions citrus waste as a high-value agroindustrial feedstock. Moreover, it provides a reproducible technical and methodological model for deploying scalable circular economy frameworks designed to minimize environmental liabilities while generating functional ingredients and intermediate products for diverse industrial applications.

4. Conclusions

This study validates the technical feasibility of transforming postharvest waste from Satsuma Owari mandarins (Citrus unshiu) into high-value-added products through an integrated cascade biorefinery model. The primary scientific contribution of this work lies in providing the first quantitative evidence that the fruit’s postharvest physiological state (MM, MC, MD) exerts a coordinated, cross-cutting effect across valorization pathways. The category simultaneously governs both the yields and molecular architectures of the liquid (vinasse, distillate) and solid (essential oil, pectin) fractions, thereby refuting the assumption that citrus waste can be processed as a homogeneous biomass and establishing a mechanistic foundation for category-differentiated processing decisions.

From a methodological perspective, this study confirms the robustness and consistency of the valorization framework to the natural variability of the raw material. Despite the intrinsic heterogeneity of the batches harvested across distant chronological periods, the application of a logarithmic transformation during statistical analysis successfully stabilized variance, demonstrating the reproducibility of key biorefinery interactions across seasonal fluctuations.

Along the liquid pathway, it is established that the antioxidant value lies strategically within the vinasse, which emerges as the resource of greatest functional relevance due to its concentration of the highest density of phenolics and flavonoids from the original must. LC-MS profiling identifies naringenin as the predominant bioactive pillar, consolidating this effluent not as a waste stream, but as a concentrated platform for the recovery of specific antioxidants.

Regarding the solid pathway, it is concluded that the physical integrity of the flavedo is indispensable for maximizing recovery of the volatile fraction, which suffers significant losses under conditions of mechanical damage (MD). Simultaneously, the pectins exhibit a structural transition linked to fruit deterioration, evolving from linear domains rich in galacturonic acid (HG) toward more branched architectures (RG-I). Nonetheless, their consistent classification as low-methoxyl pectins ensures their technological functionality for calcium-mediated gel formation.

From an applied perspective, the industrial implementation of this proposed framework will require pilot-scale process validation, mass and energy balance optimization, seasonal biomass supply chain assurance, and techno-economic assessment (TEA) of thermal operations, ethanol recovery, and vinasse conditioning.

From a sustainable development dimension, this research provides an integrated scientific approach that addresses the socioeconomic and environmental challenges of the citrus supply chain through the design of a sustainability tool based on cascade valorization. By establishing a mechanistic baseline that defines, quantifies, and monitors the qualitative variability of residual biomass, it facilitates the implementation of precise metrics to audit resource-use efficiency and estimate carbon footprint reductions. The transformation of effluents (vinasse) and solid wastes into high-value-added coproducts not only mitigates soil and water pollution associated with final waste disposal but also offers highly viable sustainability applications for local bio-industries. In this way, the developed platform provides indispensable empirical data to inform sustainability-related policies, such as extended producer responsibility schemes, incentives for an agro-industrial circular economy, and regulatory frameworks for comprehensive food waste management.

Finally, the developed cascade biorefinery offers a reproducible technical platform for advancing circular economy practices in citrus agriculture. By harmoniously coupling juice biotransformation with peel valorization and bioactive recovery from liquid effluents, this strategy mitigates the environmental footprint of postharvest losses while maximizing the molecular value of residual biomass. Future research should focus on continuous or semi-continuous scale-up, life cycle assessment (LCA), and functional validation of the extracted pectins and antioxidants within real-world food, pharmaceutical, and active packaging matrices.