Glucose and Xylose Production Under a Biorefinery Approach: Essential Oil Extraction, Hydrolysis of Orange Residues, and Reaction Kinetics at Pilot Scale

Abstract

The orange juice industry generates large amounts of waste, leading to significant environmental impacts. Within the framework of a citrus biorefinery, this study evaluates an integrated pilot-scale scheme combining essential oil extraction with hydrolysis of orange waste. A self-designed modular system was used, characterized by ease of operation and maintenance, consisting of a 20 L sealed reactor and a condenser with water recirculation. Essential oil extraction was carried out by hydrodistillation, producing 35 mL of essential oil per run and a yield of 2.57 mL per 100 g of orange peel. Hydrolysis was investigated using a 2³ factorial design considering time (30 and 60 min), waste type (with and without pulp), and H₂SO₄ concentration (0 and 0.25% v/v). ANOVA results showed that the waste type was the dominant factor, while the acid concentration had no significant effect. The optimal hydrolysis condition was waste with pulp, 0% acid, and 30 min, achieving 108.5 g/L of glucose and 30.4 g/L of xylose. Under these conditions, the kinetics of glucose and xylose release were determined. The energy consumption was 45.96 MJ, equivalent to 70.61 kJ/g of glucose and 236.59 kJ/g of xylose, with corresponding costs of 0.0017 and 0.0057 USD/g, respectively. Orange waste containing pulp, obtained directly from juice-processing facilities, exhibits greater valorization potential than orange waste without pulp to produce essential oil, glucose, and xylose within a biorefinery scheme.

1. Introduction

The global citrus industry produces more than 105 million tons of fruit annually, primarily oranges, mandarins, lemons, and grapefruits [1]. According to 2023 data from FAOSTAT, Mexico ranks among the world’s top three orange producers, along with Brazil and China [2]. Between 30% and 40% global orange production is allocated to manufacture juices and concentrates. This process generates large volumes of solid agro-industrial waste, as nearly half of the processed fruit mass becomes non-edible by-products, referred to here as orange waste (OW), including peel, pulp, and seeds. Globally, the citrus industry is estimated to generate between 15 and 21 million tons of OW each year [3].

Currently, most OW and other citrus residues are disposed of in sanitary landfills, which represent the officially registered disposal route, while a significant fraction is still managed improperly in open dumps [4]. Such practices generate acidic leachates with high organic loads, posing a risk of contamination to surface and groundwater bodies. They also promote anaerobic decomposition of biomass, leading to methane and CO₂ emissions, as well as unpleasant odors that degrade air quality [5].

The environmental and economic challenges associated with waste from the juice industry have driven circular economy approaches that revalorize orange peels (OP) as a source of energy products, chemicals, and materials [6]. Biomass conversion into such products is achieved through different valorization processes, which may be designed as stand-alone operations or integrated within a single utilization scheme. When these operations are coordinated and integrated, the concept of biorefinery is applied, defined as the integration of processes that use biomass as feedstock, analogous to a petroleum refinery that fractionates crude oil into multiple value streams. When applied to citrus residues, this approach enables the production of essential oils (EO), pectin, fermentable sugars, bioethanol, biogas, and even fibrous fractions that can be used as materials or soil amendments [7].

Orange peel is an attractive substrate for biorefinery applications. On a dry basis, it contains 15–30% pectins, 12–25% cellulose, and 10–16% hemicelluloses, with lignin generally below 5% [8]. In addition, it retains soluble sugars representing 20–30% of the dry matter [9]. The remaining components include proteins (5–7%), lipids (3–5%), and secondary metabolites—primarily essential oils and flavonoids—along with ash contents of 3–6% [8]. Although composition varies with cultivar, geographic region, and processing conditions, this lignin-poor and polysaccharide-rich matrix offers high potential to produce bioenergy carriers, bioproducts, and sugar-derived compounds [9,10,11].

One of the most distinctive components of OP is its essential oil (EO; 3–5% w/w), which is mainly composed of D-limonene, accounting for 70–95% of the EO composition, together with other terpenes and minor volatile compounds [12,13,14]. D-Limonene is a high–commercial-value product with well-recognized antioxidant and antimicrobial activity [14,15]; however, it also acts as an inhibitor of fermentative processes. Concentrations of 0.05–0.1% (w/w) inhibit microbial growth, and levels above 0.2% w/w inhibit alcoholic fermentation or anaerobic digestion [16,17]. For this reason, biorefinery schemes commonly include a preliminary EO extraction step, which simultaneously generates a high-value coproduct and reduces D-limonene to non-inhibitory levels [18,19]. Technologies used for this purpose include cold pressing, hydrodistillation, green solvents, microwave or ultrasound-assisted extraction, supercritical CO₂, and pretreatments such as steam explosion, all of which can recover most of the D-limonene [20,21,22].

Pectins are structural polysaccharides extracted from the albedo through acid processes at high temperature. Citrus pectin is widely used as a gelling agent, thickener, and stabilizer in the food industry, for example in jams, dairy products, and confectionery, as well as in pharmaceutical and cosmetic applications [23].

Reducing sugars are released from cellulose and hemicellulose through hydrolysis processes and are subsequently fermented to produce bioethanol [24]. This bioethanol, obtained from residues, is a renewable biofuel that is blended with gasoline to reduce net CO₂ emissions and dependence on fossil fuels in the transport sector; in countries such as the United States of America and Brazil, such blends are part of established regulatory mandates [24]. In fact, it is estimated that global citrus residues could generate more than one billion liters of bioethanol per year if efficiently utilized [25,26].

In addition, OP enables the production of other outputs within the same valorization scheme, including methane-rich biogas generated through anaerobic digestion from the residual organic fraction [27,28,29], as well as a variety of high-value bioproducts such as organic acids, biopolymers, nutraceuticals, and enzymes [24,30,31,32,33]. Finally, the remaining solid residue can be conditioned and used as fertilizer or soil amendment, thereby closing the utilization cycle of citrus biomass [34].

Overall, the valorization of OP through biorefinery schemes allows the production of multiple high-value products from the same waste biomass, including EO, pectin, ethanol, and biogas. It is estimated that, in an optimized process, 1 ton of OP could yield up to 9 L of EO, 39 kg of pectin, 40 L of ethanol, and 45 m³ of biogas, in addition to fertilizer fractions [35,36], illustrating its remarkable economic and environmental potential.

Over the past decades, numerous laboratory-scale studies have been conducted to convert OW into biofuels and bioproducts, assessing isolated stages such as hydrolysis, fermentation, and anaerobic digestion. These processes are technically well established and have achieved satisfactory yields under controlled conditions [37,38,39,40]. However, reports at the pilot scale remain scarce and are mainly limited to EO extraction, hydrolysis, or anaerobic digestion operations carried out in reactors with volumes ranging from 20 to 100 L aimed at evaluating feasibility under conditions closer to industrial operation [41,42,43,44,45]. This scenario highlights a gap in the literature regarding integrated citrus biorefinery schemes at pilot scale [43,46].

In this work, a pilot-scale experimental design was used to evaluate the production of glucose and xylose from OW, planned for biofuel production under a biorefinery framework. The proposed scheme integrates EO extraction by hydrodistillation with the subsequent hydrolysis of OW as obtained directly from juice-processing facilities. The process was integrated with modularity, ease of operation, and straightforward maintenance. In addition, the reaction kinetics of sugar production under the best conditions were determined, and the energy consumption analysis and operating cost per gram of glucose and xylose produced were calculated.

2. Materials and Methods

2.1. Pretreatment of Orange Waste and Orange Peel

The hydrolysis process was integrated after an EO extraction pilot-scale hydrodistillation process [47]. In this study, two types of biomass were used: OP, consisting exclusively of the orange peel, and OW, corresponding to the whole residue (pulp, seeds, and peel). This work does not address hydrolysis as an isolated process; rather, it is implemented as a downstream step following EO extraction, with the aim of maximizing waste valorization and identifying the optimal operating conditions within the integrated scheme. The processing steps and operating conditions are presented in Figure 1.

The equipment used included a Retsch GM 300 mill (Retsch GmbH, Dusseldorf, Germany) with a 5 L capacity vessel for the size reduction stage. This step is critical because volatile compounds are located between the flavedo and the albedo; therefore, mechanical disruption of the tissue facilitates EO release and enhances mass transfer between the peel and the steam. The heating source was an LP gas stove equipped with a double-ring burner, selected for its versatility and independence from the electrical grid. In this system, heat input was controlled by adjusting the fuel volumetric flow rate, which was experimentally estimated by calibrating the consumption time of a known LPG volume at a fixed burner opening and expressing the result in L/h.

The extraction system employed a sealed commercial aluminum pressure reactor T-Fal^® (T-Fal, Rumilly, France) with a nominal capacity of 22 L and an effective working volume of 12 L to prevent overpressure, in which both the analog pressure gauge and the relief valve were integrated into the commercial unit. Pressure was monitored through this integrated analog gauge. Temperature was monitored using a mercury thermometer and corroborated with an FLIR TG165 infrared thermometer (TELEDYNE FLIR, Wilsonville, OR, USA). A 2/3-inch outlet was coupled to the reactor and connected to an insulated, food-grade 1/2-inch hose with a length of 0.80 m, which in turn was connected to a condenser. The stainless-steel condenser was designed as a single-pass unit with concentric tubes operated in countercurrent flow, consisting of a 1-inch inner tube jacketed by a 2-inch outer tube, with an effective heat-transfer length of 1.1 m and an estimated heat-transfer area of approximately 0.088 m². The cooling-water inlet and outlet, as well as the vapor inlet and outlet, were 3/8 inch in diameter. The cooling medium was water stored in a 208 L tank and recirculated in countercurrent flow by a Truper BOAP-1/2 peripheral pump model 10068 (Truper S.A. de C.V., Jilotepec, Mexico), with a power of 1/2 hp and providing a flow rate of 42 L/min. The cooling water in the storage tank started at ambient temperature (20–25 °C) and, by the end of the EO extraction stage, reached approximately 32–35 °C. Once the EO extraction stage was completed, hydrolysis was initiated in the same sealed aluminum pressure reactor.

2.2. Essential Oil Extraction at Pilot Scale

Essential oil extraction was carried out by hydrodistillation, the most widely used method for citrus materials [45]. OW obtained from a juice processing facility was used as the raw material and was processed immediately after collection. The OW consisted of pulp, seeds, and peel. For each run, 3 kg of waste were used on a wet basis, although the extraction conditions were maintained as previously optimized in Armenta et al. (2023) [45], including the water-to-waste ratio, milling conditions, heating time, and heat input. In processes intended exclusively for essential oil extraction, the reactor can operate with feed loads of 3.5 and 4.0 kg according to its capacity. However, in the present study, the extraction step was integrated with a subsequent hydrolysis stage, in which additional water or acidic solution had to be added once the extraction was completed. Therefore, the initial feed load was reduced to 3.0 kg to preserve sufficient working volume inside the reactor, avoid excessive pressure buildup, and maintain safe operating conditions during the integrated process. The waste was first subjected to a milling step using a Retsch GM 300 mill (Retsch, Haan, Germany), processed in four batches because the equipment capacity did not allow the full 3 kg to be handled simultaneously. During each milling batch, 500 mL of water was added to facilitate size reduction. The ground material was then transferred to a 22 L sealed reactor, where additional water was added to achieve a water-to-waste mass ratio of 2:1. The working volume was limited to a maximum of 60% of the reactor’s total capacity to accommodate volume expansion associated with temperature increase.

Heating was applied by LP gas combustion, with an estimated heat input of 33.88 MJ/h. The total heating time was 50 min; during the first 20 min, no distillate droplets were observed at the condenser outlet. The heat exchanger water recirculation pump was activated 15 min after the heat start, in anticipation of steam generation, which began at approximately 20 min, under a gauge pressure of 5 psi and an operating temperature of 105–115 °C. This temperature range was maintained during both the EO extraction and the subsequent hydrolysis stage. The effective EO extraction time was defined as 30 min from the appearance of the first drop of distillate, since longer periods did not result in significant changes in the volume of EO recovered. The EO/water distillate was directed to a separatory funnel, where the EO was separated by density difference and collected immediately after the extraction in a graduated cylinder, to minimize losses due to volatility.

Extraction yield was calculated as the volume of EO in mL, obtained per unit mass of OP in kg. For runs conducted with OW, an experimental factor was considered in which 2.2 kg of OW were required to obtain 1 kg of OP; that is, the peel fraction in the total waste corresponded to approximately 45.5% (w/w) as OP, while the remaining 54.5% corresponded to pulp and seeds. Accordingly, the yield was consistently expressed as a function of the equivalent OP mass. Once the EO extraction was completed, the pipes from the vessel to the condenser were disconnected and 1.875 L of water were added in the non-acid runs, whereas in the acid runs, 1.875 L of H₂SO₄ solution were added; this volume already included the amount required to achieve the desired final acid concentration in the total reactor mixture. The specific concentration levels evaluated are described in Section 2.3. The system was conditioned to immediately initiate the hydrolysis process described in Section 2.3.

2.3. Experimental Design for Glucose and Xylose Hydrolysis

A three-factor, two-level factorial design was applied in the pilot-scale hydrolysis study of OW to produce reducing sugars. The factors evaluated were reaction time, with levels of 30 and 60 min; the medium acidic condition analyzed at one level in the absence of acid and at another level with acid at 0.25% v/v H₂SO₄; and the type of waste, with pulp (OW) and without pulp (OP). The response variables were the glucose and xylose concentrations expressed in g/L. In addition, during the factorial design, key process parameters such as heat input, waste mass, and water quantity were kept constant. Each experimental run was performed in duplicate and randomized using Minitab^® 19 (Minitab LLC, State College, PA, USA). The factors and levels were selected based on the physicochemical properties of OP and information reported in the literature [26,48,49]. It has been reported that low concentrations of H₂SO₄ promote sugar release in citrus residues, in contrast to other lignocellulosic materials that typically require higher acid conditions; therefore, a low acid level of 0.25% v/v was evaluated. Likewise, a no-acid level was included to assess whether acid addition is necessary at the pilot scale and, if not, to reduce reagent consumption and avoid corrosion and additional maintenance requirements [50]. After hydrolysis, the reaction mixture was filtered through a mesh filter (approximately 600 μm), and 6 L of liquid hydrolysate were obtained. Only this liquid fraction was used for sugar quantification.

The statistical significance of the factors and their interactions was evaluated by analysis of variance (ANOVA) at a 95% confidence level (p < 0.05). In addition, residual plots and overlaid contour plots were generated using Minitab^® 19.

Quantification of glucose and xylose was performed using a UV–VIS 3100PC spectrophotometer (VWR, Radnor, PA, USA) by means of the 3,5-dinitrosalicylic acid (DNS) method. The DNS reagent solution was prepared with 2.5 g of 3,5-dinitrosalicylic acid, 7.5 g of sodium potassium tartrate, and 4 g of NaOH [51]. Samples of the hydrolyzed liquid were diluted 20 times prior to analysis. Subsequently, 1 mL of each dilution was mixed with 1 mL of the DNS solution, heated in a boiling water bath for 5 min, and then cooled to room temperature. Calibration curves were constructed using known concentrations of glucose and xylose, because they are the predominant monosaccharides in OP. Absorbance measurements were taken at 575 nm for all samples.

2.4. Reaction Kinetics

Kinetic experiments were performed under the conditions that gave the highest glucose, and xylose yields according to the factorial design. During the experiment, ten samples of 8 mL were collected at equally spaced reaction times throughout the hydrolysis. Each 8 mL sample was divided into four aliquots of 2 mL for subsequent glucose and xylose quantification. The temporal profiles of glucose and xylose were fitted separately using a four-parameter logistic function based on the Verhulst logistic equation, presented in Equation (1).

$$\mathrm{C}\left(\mathrm{t}\right)={\mathrm{C}}_{\mathrm{m}\mathrm{i}\mathrm{n}}+{\displaystyle \frac{{\mathrm{C}}_{\mathrm{m}\mathrm{a}\mathrm{x}}-{\mathrm{C}}_{\mathrm{m}\mathrm{i}\mathrm{n}}}{1+{\mathrm{e}}^{\left[-\mathrm{k}\left(\mathrm{t}-{\mathrm{t}}_0\right)\right]}}}$$

(1)

In this model, C(t) is the glucose or xylose concentration (g/L); t is time (min); C_min is the initial concentration; C_max is the asymptote or maximum attainable concentration; k is the apparent rate constant (1/min) that controls the slope; and t₀ is the inflection time. This model was selected based on literature since the profiles exhibit an initial slow phase, followed by an intermediate acceleration, and finally a constant production behavior like an asymptote [52,53,54]. The parameters were estimated by nonlinear least-squares regression using all experimental data points. Preliminary parameter estimation was supported by ChatGPT 5.2; all the kinetic data was input in a prompt to generate Equation (1) for glucose and xylose. The AI tool used Mean square residues (MSRE) as evaluation point with a stopping criterion of 1 × 10⁻⁸. Next, the statistical robustness of the fit was subsequently evaluated in MATLAB^® ver. 2025a (MathWorks, Natick, MA, USA) by comparing experimental and predicted values. Model fit was assessed using the determination coefficient (R²), root mean square error (RMSE), and residual analysis and finally the kinetics results were plotted combined with their logistic equation.

2.5. Energy Consumption for Glucose and Xylose

The total energy consumption of the process was determined as the sum of the electrical energy consumed by the mill, the thermal energy supplied by the fuel gas, and the electrical energy consumed by the water recirculation pump. The energy consumption of the mill and the pump was calculated from the nominal power of each equipment and its effective operating time, whereas the thermal energy from the gas was estimated based on the volumetric flow rate used, its heating value, and the duration of heat source operation. The overall process energy consumption is given by Equation (2).

$${\mathrm{E}}_{\mathrm{p}\mathrm{r}\mathrm{o}\mathrm{c}\mathrm{e}\mathrm{s}\mathrm{s}}={\mathrm{P}}_{\mathrm{g}\mathrm{r}\mathrm{i}\mathrm{n}\mathrm{d}\mathrm{e}\mathrm{r}}{\mathrm{t}}_{\mathrm{g}\mathrm{r}\mathrm{i}\mathrm{n}\mathrm{d}\mathrm{e}\mathrm{r}}+{\mathrm{C}\mathrm{i}}_{\mathrm{g}\mathrm{a}\mathrm{s}}{\dot{\mathrm{V}}}_{\mathrm{g}\mathrm{a}\mathrm{s}}{\mathrm{t}}_{\mathrm{g}\mathrm{a}\mathrm{s}}+{\mathrm{P}}_{\mathrm{p}\mathrm{u}\mathrm{m}\mathrm{p}}{\mathrm{t}}_{\mathrm{p}\mathrm{u}\mathrm{m}\mathrm{p}} \left[\mathrm{k}\mathrm{J}\right]$$

(2)

In Equation (2), ${\mathrm{P}}_{\mathrm{grinder}}$ and ${\mathrm{P}}_{\mathrm{pump}}$ represent the nominal power of the mill and the recirculation pump, respectively, expressed in kW; ${\mathrm{t}}_{\mathrm{grinder}}$ and ${\mathrm{t}}_{\mathrm{pump}}$ correspond to the effective operating times of the mill and the pump, measured in s; ${\mathrm{t}}_{\mathrm{gas}}$ is the gas usage time, measured in h; ${\dot{\mathrm{V}}}_{\mathrm{gas}}$ is the volumetric flow rate of the gas used as fuel, expressed in m³/h; and $\mathrm{C}{\mathrm{i}}_{\mathrm{gas}}$ is the lower heating value of the gas, expressed in kJ/m³, such that each term in the equation is ultimately expressed in kJ.

The total cost of the process was calculated as the sum of the electrical cost of operating the mill and the recirculation pump, the cost of the gas used as the heat source, and the cost of the water consumed. The cost of the mill and the pump was obtained by multiplying the energy consumed by each piece of equipment and the electricity price; the gas cost was estimated as the product of the volumetric gas flow rate, the usage time, and the price per liter; finally, the water cost was determined by multiplying the total volume of water used by its unit price. This relationship is expressed in Equation (3).

$${\mathrm{C}}_{\mathrm{p}\mathrm{r}\mathrm{o}\mathrm{c}\mathrm{e}\mathrm{s}\mathrm{s}}={\mathrm{E}}_{\mathrm{g}\mathrm{r}\mathrm{i}\mathrm{n}\mathrm{d}\mathrm{e}\mathrm{r}}{\mathrm{C}}_{\mathrm{e}\mathrm{l}\mathrm{e}\mathrm{c}}+{\mathrm{E}}_{\mathrm{p}\mathrm{u}\mathrm{m}\mathrm{p}}{\mathrm{C}}_{\mathrm{e}\mathrm{l}\mathrm{e}\mathrm{c}}+{\dot{\mathrm{V}}}_{\mathrm{g}\mathrm{a}\mathrm{s}}{\mathrm{t}}_{\mathrm{g}\mathrm{a}\mathrm{s}}{\mathrm{C}}_{\mathrm{g}\mathrm{a}\mathrm{s}}+{\mathrm{V}}_{\mathrm{w}\mathrm{a}\mathrm{t}\mathrm{e}\mathrm{r}}{\mathrm{C}}_{\mathrm{w}\mathrm{a}\mathrm{t}\mathrm{e}\mathrm{r}} \left[\mathrm{U}\mathrm{S}\mathrm{D}\right]$$

(3)

In Equation (3), ${\mathrm{C}}_{\mathrm{process}}$ denotes the overall process cost, expressed in U.S. dollars (USD). The quantities ${\mathrm{E}}_{\mathrm{grinder}}$ and ${\mathrm{E}}_{\mathrm{pump}}$ correspond to the electrical energy consumed by the mill and the recirculation pump, respectively, measured in kWh. ${\mathrm{C}}_{\mathrm{elec}}$ is the unit cost of electrical energy, set at 0.205 USD/kWh. The term ${\dot{\mathrm{V}}}_{\mathrm{gas}}$ represents the volumetric gas flow rate (m³/h), ${\mathrm{t}}_{\mathrm{gas}}$ is the gas usage time (h), and ${\mathrm{C}}_{\mathrm{gas}}$ is the unit price of the gas, taken as 0.000579 USD/m³. Finally, ${\mathrm{V}}_{\mathrm{water}}$ is the total volume of water used (L), and ${\mathrm{C}}_{\mathrm{water}}$ is the unit cost of water, established at 0.027 USD/L.

3. Results and Discussion

3.1. Yields of Essential Oil Extraction at Pilot Scale

Using the hydrodistillation conditions described in Section 2.2, EO extraction from OW yielded 35 ± 1 mL of EO per run. Considering that 3 kg of OW experimentally correspond to 1.36 kg of equivalent OP (at a ratio of 2.2 kg of OW per 1 kg of OP), this value corresponds to a yield of 25.7 mL EO/kg OP. In runs conducted exclusively with OP, using 3 kg of OP per run, 43 mL of EO were obtained, corresponding to a yield of 14.3 mL EO/kg OP. When these yields are compared with values reported in the literature for OP hydrodistillation, they fall within the range reported for citrus biomasses under similar operating conditions [47,50,55,56,57].

Given that the operating conditions (biomass loading, liquid-to-solid ratio, effective extraction time, and heating power) are identical to those used in [50], the EO composition is expected to be dominated by D-limonene, with concentrations exceeding 95%, as reported in the chemical composition analysis performed in that study.

3.2. Factorial Design Applied to Glucose and Xylose

Sixteen hydrolysis runs were performed to evaluate glucose and xylose production from OP and OW. The obtained values are presented in

Table 1. Factorial design applied to glucose and xylose production.

Time (min)	1: With Pulp 2: Without Pulp	Acid (% v/v)	Concentration (g/L)
			Glucose	Xylose
30	1	0	108.50	30.38
30	1	0	104.72	26.78
30	1	0.25	92.40	26.50
30	1	0.25	104.80	30.22
30	2	0	46.71	14.19
30	2	0	42.56	13.22
30	2	0.25	69.54	20.06
30	2	0.25	84.82	18.52
60	1	0	104.61	32.38
60	1	0	89.01	26.82
60	1	0.25	85.29	25.99
60	1	0.25	77.03	22.17
60	2	0	64.34	18.89
60	2	0	63.04	18.42
60	2	0.25	56.84	17.97
60	2	0.25	60.08	25.75

, which reports the concentrations of both sugars.

The ANOVA results for glucose and xylose are presented in

Table 2. Results for glucose and xylose ANOVA.

	Glucose						Xylose
Source	df	Sum of Squares	Mean Square	Contribution	F-Value	p-Value	df	Sum of Squares	Mean Square	Contribution	F-Value	p-Value
Model	7	6611.88	944.55	94.68%	20.35	0.00017	7	449.79	64.26	86.84%	7.54	0.00534
Time (A)	1	180.97	180.97	2.59%	3.90	0.08373	1	4.54	4.54	0.88%	0.53	0.48631
Waste (B)	1	4845.20	4845.20	69.38%	104.40	0.00001	1	344.23	344.23	66.46%	40.40	0.00022
% acid (C)	1	3.34	3.34	0.05%	0.07	0.79528	1	2.33	2.33	0.45%	0.27	0.61520
A × B	1	190.10	190.10	2.72%	4.10	0.07759	1	29.05	29.05	5.61%	3.41	0.10206
A × C	1	515.63	515.63	7.38%	11.11	0.01033	1	14.75	14.75	2.85%	1.73	0.22475
B × C	1	649.61	649.61	9.30%	14.00	0.00569	1	52.76	52.76	10.19%	6.19	0.03763
A × B × C	1	227.03	227.03	3.25%	4.89	0.05792	1	2.13	2.13	0.41%	0.25	0.63032
Error	8	371.26	46.41	5.32%			8	68.17	8.52	13.16%
Total	15	6983.14		100%			15	517.96		100.00%

. The factors evaluated were time (A), residue type (B), and acid percentage (C). The model explains 94.68% of the variability in glucose and 86.84% in xylose. In both cases, waste type is the dominant factor, indicating that the presence of pulp significantly increases sugar availability and release.

A factor was considered to significantly influence the response variable when p < 0.05. Acid percentage did not show a significant main effect; however, it produced significant interaction effects. For glucose, the B × C and A × C interactions are significant, whereas for xylose, the B × C interaction is also significant. Time, as a main effect, is not significant, although in the case of glucose it shows a trend and interacts with acid percentage. The triple interaction A × B × C does not reach statistical significance.

The ANOVA indicates that the model for glucose explains 94.68% of the variability in this process, while xylose explains 86.84%. In the case of xylose, this reduction is attributed to the more limited variability explained by the factors and their interactions. Xylose is released in smaller proportions and is subject to secondary phenomena, such as degradation of the sugar itself. Nevertheless, both models are significant and reliable, as the overall p-value (<0.05) validates the ANOVA applied to both response variables. In addition, the residual plots satisfy the normality assumptions, as they follow a straight-line trend in the normal probability plot, and the residuals versus fitted values plot shows a random dispersion. The residual plots for both ANOVA analyses are shown in Figure 2. The residue plot in Figure 2a,c show the normality trend of the model for glucose and xylose, while Figure 2b,d depict no clear tendency of the residues for the glucose and xylose model.

The overlaid contour plots presented in Figure 3 show regions where both sugars coexist under different combinations of reaction time, residue type, and acid concentration. The desired output concentration range was 28–32 g/L for xylose and 90–110 g/L for glucose, these values were selected as concentrations of interest for evaluating process performance under the studied conditions. In all three subfigures of Figure 3, the white region represents the range of conditions that satisfy the established criteria, whereas the gray region corresponds to combinations that do not meet it.

At 0% v/v of acid and in the presence of pulp (Figure 3a), high concentrations of glucose and xylose were achieved, demonstrating that acid addition is not required to maximize sugar release. In Figure 3b, under low-acid conditions and with pulp-containing residue, both concentrations increased with reaction time; however, when the acid concentration was increased to 0.25% v/v, the concentration decreased, indicating a negative effect of acid on yield within this range. Finally, in Figure 3c, where the plot is conditioned at 30 min, glucose decreased consistently when increasing acid concentration, while xylose remained essentially stable. Taken together, these results indicate an adverse effect of acid concentration on glucose and show that acidity is not a determining factor for xylose release.

3.3. Reactions Kinetics

Kinetic experiments were conducted under the conditions that maximized sugar release (0% acid, OW, 30 min). Figure 4 presents the sugar release kinetics curves: (a) glucose concentration and (b) xylose concentration as a function of time, integrating the experimental values from both runs. This representation allowed the temporal evolution of sugar release to be visualized graphically. Equations (4) and (5) are also plotted in Figure 4 as the trend of glucose and xylose production.

The results show an increasing trend in the concentration of both sugars, with similar overall behaviors but some differences in their release rates. During the first minutes of the process, glucose and xylose release increases, although not significantly. This initial phase suggests that heat begins to affect the structure of the residue, allowing partial decomposition of the present carbohydrates.

From minute 18 onward, sugar release becomes more pronounced. The concentrations of glucose and xylose begin to increase more sharply, indicating that hydrolysis reaches a point at which the bonds forming structural carbohydrates start to break more efficiently, suggesting the hydrolysis of structural fractions of the residue, particularly hemicelluloses and, probably, part of the more accessible cellulose fraction. This behavior continues steadily until minute 27, when the maximum concentration of both sugars is reached, remaining essentially constant until the end of the process at minute 30. The reaction kinetics exhibit asymptotic behavior in this final stage, indicating that the process has reached its maximum limit of fermentable sugar release under the evaluated conditions.

The production of glucose and xylose over time is described by a modified logistic equation, which accounts for an initial slow phase, an intermediate acceleration, and an approach to an asymptote; this behavior is presented in Equations (4) and (5). The fits resulted in R² values of 0.9741 for glucose and 0.9398 for xylose, with RMSE values of 1.0695 and 0.5212 g/L, respectively, supporting the robustness of the model with respect to the experimental averages over the 3–30 min range. In addition, the residuals showed no marked systematic trend.

$${\mathrm{C}}_{\mathrm{g}\mathrm{l}\mathrm{u}\mathrm{c}\mathrm{o}\mathrm{s}\mathrm{e}}\left(\mathrm{t}\right)=72.16+{\displaystyle \frac{25.79}{1+{\mathrm{e}}^{\left[-0.193 \left(\mathrm{t}-24.09\right)\right]}}}$$

(4)

$${\mathrm{C}}_{\mathrm{x}\mathrm{y}\mathrm{l}\mathrm{o}\mathrm{s}\mathrm{e}}\left(\mathrm{t}\right)=16.205+{\displaystyle \frac{17.020}{1+{\mathrm{e}}^{\left[-0.0686 \left(\mathrm{t}-29.181\right)\right]}}}$$

(5)

3.4. Results Discussions

Table 3. Result comparison with literature.

Reference	Glucose (g/L)	Xylose (g/L)	Method	Conditions
Current study	108.5	32.4	Hydrolysis of OW, 0% H2SO4, prior integration with EO extraction	Modular system; reaction times of 30–60 min; wet, pulp-containing residue; no enzymes or supercritical CO2; absence of acid.
Armenta et al., 2025 [26]	52.14	15.7	Dilute acid hydrolysis of OP	0, 0.5, and 1% H2SO4 at 125 °C; factorial design evaluating temperature, acid concentration, and time.
Corona Vázquez et al., 2017 [58]	18	7	Thermal hydrolysis of OP (high temperature and pressure) in the presence of H2SO4	120–180 °C, 600–1000 kPa, 30 min; agitation at 350 rpm; subsequent fermentation with immobilized yeast.
Vinotha et al., 2023 [59]	1.75 1	NA	Supercritical CO2 extraction/pretreatment followed by subsequent hydrolysis steps	Supercritical CO2 (>73 atm), high temperature; use of organic solvents and additional processing stages.
Patsalou et al., 2019 [46]	NA	NA	Distillation and acid hydrolysis	Storage at −20 °C; use of OW. EO extraction for 24 h at 70 °C. Acid hydrolysis with 0.5% v/v H2SO4 followed by enzymatic hydrolysis. Fermentation with P. kudriavzevii KVMP10 and S. cerevisiae strains.
Ayala et al., 2021 [51]	24.6	NA	Hydrolysis of fermentable sugars from OP	Moderate hydrolysis conditions of 0.5, 1, and 1.5% v/v; focus on characterization and integral valorization.
Tsouko et al., 2020 [7]	4.9 1	NA	Integrated stages for EO, pectin, ethanol, and methane production	Multi-stage sequence including EO and phenol extraction, acid hydrolysis, and an enzymatic step aimed at producing bacterial cellulose and other products.
Ortiz et al., 2013 [60]	59.88 1	NA	Delignification and acid hydrolysis	Particle size reduction to 3–5 mm, followed by NaOH delignification and subsequent acid hydrolysis.
Tejada et al., 2010 [61]	80 1	NA	Delignification and acid hydrolysis	NaOH delignification followed by acid hydrolysis with 5% v/v H2SO4.

¹ Values reported as total sugars. NA stands for Not Available.

compares the maximum glucose and xylose concentrations achieved in this study with those reported in other investigations on orange waste valorization. In the present work, concentrations of up to 108.5 g/L glucose and 32.4 g/L xylose were obtained by hydrolysis of pulp-containing orange waste with 0% H₂SO₄, integrated with a prior EO extraction step.

Table 3 comparison shows that, although several studies on citrus waste have explored thermal or dilute acid hydrolysis processes, often combined with more complex stages such as the use of supercritical CO₂, alkaline pretreatments, or sequences integrating acid and enzymatic hydrolysis, these have generally been conducted at laboratory scale. In contrast, the present work, developed at pilot scale, achieves the highest glucose (108.5 g/L) and xylose (32.4 g/L) concentrations through a low-severity thermal scheme, even in the absence of acid and without the use of enzymes or supercritical CO₂.

In the study by Corona Vázquez et al. (2017) [58], hydrolysis of orange peel reached approximately 18 g/L of glucose and 7 g/L of xylose, values mainly measured to supply the fermentation stage and evaluate bioethanol production. In the present work, under acid-free conditions (0% H₂SO₄), up to 108.5 g/L of glucose and 32.4 g/L of xylose were obtained, corresponding to approximately 500% more glucose and 360% more xylose than those reported by Corona, even though that study employed high temperatures, elevated pressure, and acid addition to promote hydrolysis.

In the study by Vinotha et al. (2023) [59], OP was subjected to pretreatment and hydrolysis followed by fermentation with a bacterial broth, without a prior EO extraction step. Carbohydrate release was reported only as total reducing sugars on the order of 1–2 g/L, without distinguishing between glucose and xylose, values that are significantly lower than those achieved in the present work.

Armenta et al. (2025) [26] and Ayala et al. (2021) [51] are the most comparable studies in terms of dilute acid concentration and hydrolysis temperatures. Even so, the concentrations obtained in the present work are 108% higher for glucose and 106% higher for xylose than those reported by Armenta, and 341% higher for glucose than those reported by Ayala. This difference is explained using an integral residue containing both peel and pulp, since the pulp concentrates a significant fraction of soluble sugars that is not available when OP alone is used.

In the studies by Ortiz et al. (2013) [60] and Tejada et al. (2010) [61], OP was first subjected to alkaline delignification with NaOH and then to hydrolysis with high-concentration H₂SO₄ (5% v/v), which increases chemical consumption and the aggressiveness of the medium toward equipment materials. Both studies report only total reducing sugars, without distinguishing glucose and xylose. When compared with the sum of glucose and xylose obtained in the present study, these values represent only 42% and 57% for Ortiz and Tejada, respectively.

The work by Patsalou et al. (2019) [46] explicitly reports the use of integral citrus waste including peel, pulp, and seeds, within a biorefinery scheme that integrates multiple processing stages. However, their study was conducted at laboratory scale, with a sequence of chemical and biological pretreatments primarily aimed at ethanol and methane production, in which reducing sugars were monitored only as substrates consumed during fermentation, without detailed reporting of the maximum sugar concentrations released in the hydrolysate. Similarly, Tsouko et al. (2020) [7] propose a multi-stage integrated process, but their focus is not on the exhaustive quantification of released sugars, but rather on the use of the sugar-rich hydrolysate as a culture medium to obtain high concentrations of bacterial cellulose.

It should be noted that the comparisons presented must be interpreted with caution, since the reported studies differ in several methodological aspects, including biomass fraction, moisture basis, reactor scale and volume, and thermal severity. In addition, many authors do not describe the hydrolysis stage in enough detail to allow a strictly equivalent comparison of all operating conditions. Under these limitations, the most comparable result from the present study is the condition without pulp and without acid, which yielded 64.34 g/L of glucose and 18.89 g/L of xylose. This condition provides a more balanced basis for comparison with studies focused mainly on orange peel and indicates that the sugar concentration obtained in the present work is consistent with the range reported in the literature. By contrast, the condition including pulp led to markedly higher sugar concentrations, which constitutes one of the main findings and novel contributions of the present study, as it highlights the advantage of using whole orange waste rather than peel alone.

Furthermore, the conditions evaluated in the present work do not correspond to an isolated hydrolysis step but are defined by integration with the prior essential oil extraction stage, which imposes specific constraints on temperature, time, and residue condition. Within this operational framework, established by the first stage of the process, the system can release higher amounts of fermentable sugars from OW compared to those studies presented in Table 3. This reinforces that the results obtained are not only high in terms of concentration, but also consistent with a truly integrated biorefinery scheme that is compatible with the practical reality of residues generated by juice processing facilities.

3.5. Process Energy Consumption and Cost

Equation (2) was solved considering a mill power of 2 kW operating for 60 s, a re-circulation pump rated at 0.373 kW operating for 1800 s, and the use of LP gas with an assumed heating value of 25 MJ/L, supplied at a flow rate of 0.0007379 m³/h for 1.33 h, the total energy consumption of the integrated EO extraction and hydrolysis process was 45,964.73 kJ, equivalent to 45.96 MJ per experimental run.

Considering the maximum sugar concentrations obtained in the hydrolysate (108.5 g/L glucose and 32.38 g/L xylose) and a hydrolysate volume of 6 L, a total of 651 g of glucose and 194.28 g of xylose were obtained. Relating these amounts to the energy consumed in the integrated extraction and hydrolysis process (45,964.73 kJ), unit energy consumptions of 70.61 kJ/g of glucose and 236.59 kJ/g of xylose were calculated.

It is important to note that the energy consumption distribution is strongly dominated by the LP gas heating stage: 98.3% of the total energy consumed in the process corresponds to the heat source, while the mill and the recirculation pump contribute only 1.7%. This imbalance is associated with the low thermal efficiency of the current heating system, whose configuration favors radial heat losses to the environment. In this regard, the heat source is identified as one of the main opportunities for process improvement, as a more efficient and better insulated system could significantly reduce energy consumption. Accordingly, energy consumption represents the first major challenge for future industrial scaleup of the proposed process. For comparison, Corona Vázquez et al. [58] reported 1.407 kWh (5065.2 kJ) for orange peel hydrolysis at 120 °C, while Patsalou et al. (2020) [11] identified 116 °C for 10 min as suitable conditions for citrus peel hydrolysates. However, direct comparison remains limited since those studies were conducted at laboratory scale, and Corona Vázquez et al. [58] used only 2.5 g of dry biomass, whereas the present work used a 3 kg wet basis feed at pilot scale. In thermal hydrolysis, the main objective is to bring the reactor mixture to the target temperature range, here 110–120 °C; therefore, overall energy consumption depends largely on the configuration, heat source, and thermal efficiency of each system. Nevertheless, at this stage of development, LP gas was selected as the heat source due to its availability, ease of operation and maintenance, and the modularity it offers for assembling and adjusting the pilot-scale system.

From Equation (3), a total cost per run of 0.9150 USD was calculated. Relating this value to the maximum masses of sugars produced in the hydrolysate (651 g of glucose and 194.28 g of xylose), the specific process cost was 0.0014 USD/g of glucose and 0.0047 USD/g of xylose, equivalent to 1.41 and 4.71 USD/kg, respectively. From Equation (3), a total cost per run of 0.9150 USD was calculated. Relating this value to the maximum masses of sugars produced in the hydrolysate (651 g of glucose and 194.28 g of xylose), the specific process cost was 0.0014 USD/g of glucose and 0.0047 USD/g of xylose, equivalent to 1.41 and 4.71 USD/kg, respectively. In addition, when this total cost per run was related to the volume of essential oil recovered, the corresponding specific cost was 0.0261 USD/mL for the pulp-containing condition (35 mL EO) and 0.0213 USD/mL for the condition without pulp (43 mL EO). These values provide an additional economic reference for the integrated biorefinery approach, since essential oil represents a coproduct obtained prior to the hydrolysis stage.

4. Conclusions

The integrated process presented here constitutes a solid contribution toward a citrus biorefinery by coupling essential oil extraction with the hydrolysis of orange waste generated by juice-processing facilities. The results showed maximum concentrations of 108.5 g/L glucose and 32.38 g/L xylose, corresponding to 651 g of glucose and 194.28 g of xylose in 6 L of hydrolysate, substantially higher than values reported in the literature, attributable to the use of an integral, pulp-containing waste whose high fermentable sugar content makes it a more attractive feedstock than peel alone, despite the latter having been favored in other studies due to its broader range of potential products. A relevant finding for biorefinery design is that high sugar concentrations were obtained even in the absence of acid, making it possible to avoid strongly corrosive conditions and, consequently, to reduce equipment deterioration and maintenance requirements associated with acidic media.

The reaction kinetics exhibited a clear three-stage pattern: an initial phase of slow sugar release, followed by an accelerated increase, and finally an approach to a maximum value at around 30 min. This behavior allows precise identification of the operating window in which fermentable sugar production is maximized within the proposed integrated scheme.

In terms of energy consumption and cost, the integrated process required 45.96 MJ per run, with a total cost of USD 0.9150, corresponding to specific costs of 1.41 and 4.71 USD/kg of glucose and xylose, respectively. These values, together with the coproduction of 35 mL of essential oil per run and the high glucose and xylose concentrations achieved, are promising for the design of an integrated citrus biorefinery. In future work, purification of the hydrolysate will allow the recovered sugars to be directed toward the production of value-added chemicals or their fermentation to bioethanol, either at laboratory grade or oriented toward energy applications, while the remaining solid residue, after appropriate characterization, may be routed to new valorization pathways.