Potential of Acrocomia aculeata Pulp Waste for Fermentative Hydrogen Production and the Impact of Hydrothermal Pretreatment

Featured Application

This study presents the first evaluation of Acrocomia aculeata pulp waste as a substrate for fermentative hydrogen production. A subcritical water hydrolysis pretreatment was applied at different temperatures, and both solid and liquid hydrolyzed fractions were systematically characterized and assessed through biochemical hydrogen potential assays and a mass-energy balance. The results demonstrated that this biomass can be efficiently converted into biohydrogen without the need for hydrothermal pretreatment.

Abstract

This study provides the first comprehensive evaluation of the biochemical hydrogen production (BHP) potential of Acrocomia aculeata pulp waste, a residue abundantly generated during fruit processing in Latin America. The valorization of this underused biomass is essential to promote circular bioeconomy strategies and expand renewable energy sources in the region. The fermentative hydrogen potential of untreated pulp and of fractions obtained after subcritical water pretreatment was assessed under mesophilic conditions to quantify hydrogen yields and elucidate the energy distribution between solid and liquid phases. Pretreatments were performed at 150, 200, and 250 °C, and both fractions were individually tested. The untreated pulp achieved the highest BHP (125.1 NmL H₂/g VS fed), while pretreated solids showed decreasing values of 118.1, 71.6, and 41.6 NmL H₂/g VS fed at 150, 200, and 250 °C, respectively. The liquid fractions yielded 107.2, 79.4, and 76.0 NmL H₂/g COD fed, showing a similar decline with increasing severity. A mass-energy balance revealed that 1 ton of residual pulp could produce up to 104 m³ H₂, equivalent to 15 GJ/ha-year, while the combined solid plus liquid fractions from pretreatment at 150 °C recovered a comparable 14.5 GJ/ha-year, with 65% of hydrogen energy originating from the liquid phase. More severe conditions led to up to 40% lower total energy yields. These findings demonstrate that A. aculeata pulp waste inherently exhibits high fermentative hydrogen potential without requiring hydrothermal pretreatment, highlighting its direct applicability as a renewable substrate for sustainable biohydrogen production.

1. Introduction

Acrocomia aculeata (Jacq.) Lodd. ex Mart. (Arecaceae), regionally known as mbokajá, macaúba, bocaiúva, dendé, or coyol, is a perennial fruit-bearing palm native to the American tropics, with a documented distribution from Florida to Paraguay and northern Argentina [1]. This species typically grows in open grasslands, disturbed forest areas, and agricultural systems [2]. Its fruit yield is estimated at approximately 25 t/ha during the harvest period [3]. The palm produces large clusters of drupaceous fruits composed of the shell (epicarp), pulp (mesocarp), endocarp, and kernel [4]. Two types of oil can be extracted from the fruit—kernel and pulp oil—with combined yields of up to 5 t/ha [5]. During oil extraction, considerable residues are generated, including shell (27.21%), endocarp (20.59%), and pulp cake (27.16%), which can serve as feedstocks for biofuel production, thus adding value and promoting a sustainable and profitable cultivation model [6].

A. aculeata is currently cultivated in the Americas as a native oil-bearing species, and its cultivation could be expanded to other tropical and subtropical regions, including sub-Saharan Africa and parts of Asia, to foster social inclusion and enhance national competitiveness [7]. Although the species has been primarily investigated for biodiesel production owing to its high lipid content [7,8], its solid residues—shell, pulp, and kernel—are rich in carbohydrates, proteins, and lignocellulosic fractions, making the entire fruit a promising raw material for energy generation [9]. Moreover, since the fruits are not intended for food use, their energetic and non-energetic applications do not compete with food production [10]. Despite this potential, limited research has focused on the composition and processing of the fruit pulp or on the valorization of its by-products for sustainable bioenergy production [6,11].

The defatted pulp of A. aculeata exhibits high bioenergy potential owing to its content of structural polysaccharides, including galactoglucomannans, pectins, xyloglucans, xylans, and cellulose. These polymers provide fermentable sugars such as glucose, mannose, galactose, and xylose, which can be released through hydrolysis to produce gaseous biofuels such as biogas or biohydrogen. Its low protein, ash, and lipid contents minimize the formation of inhibitory compounds, making the pulp a sustainable and efficient feedstock for energy bioconversion [12,13]. Nevertheless, specific studies on gaseous biofuel production from A. aculeata remain scarce. Aguilar-Aguilar et al. (2025) evaluated the valorization of A. aculeata shells through thermochemical pretreatments, demonstrating that these strategies improve the accessibility of fermentable components and make the pretreated biomass more suitable for biogas generation [14]. Ampese et al. (2021) also investigated shell valorization via subcritical water hydrolysis followed by anaerobic digestion, reporting a significant increase in biogas yields [15]. Similarly, Aguilar-Aguilar et al. (2025) analyzed the biochemical methane potential (BMP) of A. aculeata fruit, applying a statistical design to determine the optimal proportions of pulp, seed, and shell [16].

Hydrogen represents another renewable gas that can be obtained from agro-industrial residues. It is characterized by its high calorific value (142 MJ/kg) and by water vapor as its only combustion by-product, making it a key vector for clean energy [17]. Among biological routes, dark fermentation stands out for its sustainability and capacity to convert low-cost organic waste into hydrogen [18]. However, to the best of the authors’ knowledge, no studies have yet reported biohydrogen production from A. aculeata pulp via dark fermentation. Moreover, although A. aculeata pulp contains abundant structural polysaccharides that could serve as fermentable substrates, the effect of pretreatment on its conversion efficiency has not yet been investigated. In this context, hydrothermal hydrolysis using subcritical water is a sustainable pretreatment method that uses water under high temperature and pressure to hydrolyze various biomasses without chemical additives. It enhances the solubilization of organic matter, preserves environmental safety, and improves the subsequent bioconversion of lignocellulosic and agro-industrial residues [19]. When implemented in sudden expansion reactors (SERs), this technology benefits from precise control of temperature, pressure, and residence time [20,21]. The rapid heating and cooling rates of SER prevent secondary reactions and degradation, enabling efficient and environmentally friendly hydrolysis of complex biomasses [22]. In this context, the present study for the first time evaluated the biochemical hydrogen potential (BHP) of A. aculeata pulp, both untreated and pretreated with subcritical water at different temperatures. The pretreatment was conducted in a pressurized hydrolysis unit equipped with a SER, which enabled rapid depressurization. The resulting hydrolysate was separated by centrifugation into liquid and solid fractions, both of which were comprehensively characterized and subsequently assessed for their fermentative hydrogen production potential. Finally, a mass-energy balance was also performed to elucidate the energy distribution between the solid and liquid fractions obtained after subcritical water pretreatment and to compare it with the untreated pulp. This study highlights the potential of A. aculeata as a renewable feedstock for biohydrogen production.

2. Materials and Methods

2.1. Raw Material Preparation

A. aculeata pulp was obtained from fruits manually harvested in Caapucú (Paraguarí, Paraguay) under conditions that prevented soil contact and preserved fruit quality. The fruits were visually inspected, and those showing signs of deterioration or mechanical damage were discarded [23]. After cracking the fruits, the pulp was carefully separated using stainless steel knives to avoid contamination. The fresh pulp was subjected to a two-step pretreatment with 95% ethanol at 70 °C and a pulp-to-ethanol ratio of 1:20 (w/v) for 2.5 h per step, in order to remove extractive compounds, mainly lipids. After each step, the mixture was separated by filtration, and the pretreated pulp was dried in a convection oven at 80 °C for 24 h to ensure complete removal of residual ethanol.

2.2. Inoculum Preparation

The inoculum used in the BHP assays was obtained from a 100 L anaerobic reactor fed with food waste from restaurants and operated under mesophilic conditions. To enrich the hydrogen-producing microbial community, a heat-shock pretreatment was applied at 90 °C for 20 min, followed by three successive subcultures, according to the methodology described by Martínez-Mendoza et al. (2022) [24]. The resulting material was stored at 4 °C until use in the experiments. Prior to the assays, the inoculum was activated through a 24 h preculture in a sealed 2.1 L anaerobic reactor containing 0.1 L of inoculum and 0.9 L of culture medium with the following composition (g/L): 10.0 lactose, 2.40 NH₄Cl, 2.40 K₂HPO₄, 1.18 MgCl₂, 0.60 KH₂PO₄, 0.11 CaCl₂, and 0.024 FeCl₂ [24]. Incubation conditions were 37 ± 1 °C and 200 rpm. After activation, the inoculum presented a pH of 5.2, and total and volatile suspended solids concentrations of 1.72 ± 0.31 g/L and 1.55 ± 0.27 g/L, respectively.

2.3. Subcritical Water Pretreatment

The subcritical water pretreatment of A. aculeata pulp was carried out in the continuous pilot plant of the FASTSUGARS process, following the procedure described by Martínez et al. (2019) [25]. Prior to pretreatment, the pulp was subjected to size reduction in a Retsch TM 300 cutting mill (Retsch, Haan, Germany) for 4 h until reaching a particle size ≤250 μm. The ground pulp was suspended in deionized water at 5% (w/w) and maintained under constant stirring to ensure homogeneity before feeding into the reactor. Heating was achieved by mixing the ambient-temperature feed with pressurized hot water through a T-junction, allowing a rapid increase to the target reaction temperature. The effluent was then instantaneously depressurized through a needle valve, inducing a sudden cooling and an immediate cessation of hydrolysis. The residence time (τ) was defined as the interval during which the pulp remained in contact with water under subcritical conditions, from the mixing point to the expansion valve. This parameter was calculated using Equation (1), which relates residence time to the reactor volume, water density, and mass feed flow rate:

$$\tau ={\displaystyle \frac{\mathrm{V}\rho }{{\mathrm{F}}_{\mathrm{m}}}}$$

(1)

where V is the reactor volume (m³), ρ is the density of water at the operating temperature and pressure (kg/m³), and F_m is the mass feed flow rate (kg/s).

The subcritical water hydrolysis was performed under three reaction conditions at an average pressure of 170.50 ± 0.39 bar and temperatures of 150, 200, and 250 °C, with residence times of 3.2, 3.2, and 2.7 s, respectively. The effective feed concentrations entering the reactor were 1.8%, 2.6%, and 2.2% (w/w). The treatment generated three reaction effluents, which were subsequently separated into solid and liquid fractions by centrifugation at 7500 rpm for 15 min. Hereafter, the codes P-S150, P-S200, and P-S250 refer to the solid fractions, and P-L150, P-L200, and P-L250 to the corresponding liquid fractions obtained at 150, 200, and 250 °C, respectively. Each fraction was subsequently evaluated for hydrogen production.

2.4. Biochemical Hydrogen Potential (BHP) Assays

Batch assays were conducted to evaluate the effect of hydrothermal pretreatment on the solid and liquid fractions obtained from A. aculeata pulp, in comparison with the untreated pulp. BHP tests were performed using an automatic AMPTS II system (Bioprocess Control AB, Lund, Sweden). To ensure accurate hydrogen quantification, CO₂ traps containing 3 M NaOH and thymolphthalein were placed between the reactors and the gas flow measurement unit. The assays were conducted according to the standardized protocol described by Carrillo-Reyes et al. (2020) [26], using the corresponding inoculum, substrate, mineral medium with macro- and micronutrients, and deionized water. Briefly, the tests were carried out at 37 °C under intermittent stirring (3 min on/1 min off) at 45 rpm, with a working volume of 0.36 L. The initial pH was adjusted to 7.5 using a 5 M NaOH solution. For the solid fractions, a substrate loading of 5 g volatile solids (VS)/L was used, while for the liquid fractions a concentration of 5 g COD/L was applied. Each reactor was inoculated with 10% (v/v) of inoculum. Blank controls (containing only inoculum and nutrient solutions) and a positive control (with glucose) were also prepared. The systems were hermetically sealed with rubber stoppers and purged with nitrogen for 60 s to remove oxygen and maintain strict anaerobic conditions. All experiments were performed in triplicate and terminated when the cumulative hydrogen production reached a stable value. At the end of the assays, liquid samples were collected for pH and organic acid analyses. The BHP was expressed as NmL H₂ per gram of VS fed for the solid fractions (P-S150, P-S200, and P-S250), and NmL H₂ per gram of COD fed for the liquid fractions (P-L150, P-L200, and P-L250).

2.5. Analytical Methods

The structural carbohydrate content of A. aculeata pulp was determined according to NREL/TP-510-42618 [27], while aqueous and organic extractives were quantified following NREL/TP-510-42619 [28]. The A. aculeata pulp and the hydrolyzed solid fractions obtained from the ultrafast SER were analyzed to determine their physicochemical composition. Moisture and ash contents were determined according to NREL/TP-510-42621 and NREL/TP-510-42622, respectively [29,30]. Protein content was determined by the Kjeldahl method for total nitrogen, using a conversion factor of 6.25, as described in NREL/TP-510-42625 [31]. Soluble and insoluble lignin fractions were quantified following NREL/TP-510-42618 [27], while solids were determined according to the standard methods [32]. The hydrolyzed liquid fractions were characterized by measuring total organic carbon (TOC) and total nitrogen (TN) using a Shimadzu TOC-Vcsh analyzer (Kyoto, Japan). Total (tCOD) and soluble chemical oxygen demand (sCOD) were determined according to standard methods [32], using potassium dichromate digestion followed by colorimetric quantification with a Hach DR 6000 spectrophotometer (Hach Company, Loveland, CO, USA). The contents of carbon, hydrogen, nitrogen, and oxygen in untreated pulp and in the solid fraction of hydrolysates were determined by elemental analysis using a Flash 200 analyzer (Thermo Fisher Scientific, Waltham, MA, USA), in accordance with the ASTM D5373-21 standard [33]. Total sugar content was determined by the phenol–sulfuric acid colorimetric method [34], while pH was measured following standard protocols [32]. Total and volatile suspended solids (TSS and VSS) in the inoculum were determined according to standard methods [32]. The gas composition generated during the BHP assays (particularly during the acceleration phase of hydrogen production) was analyzed by gas chromatography (GC) using an Agilent 8860 system (USA) equipped with a thermal conductivity detector (TCD) and a Varian CP-Molsieve 5A Capillary Column (15 m, 0.53 mm, 15 µm) interconnected with a Varian CP-PoreBOND Q Capillary Column (25 m, 0.53 mm, 10 µm) (Agilent Technologies, Santa Clara, CA, USA), according to the procedure described elsewhere [35]. Organic acids in the liquid effluents from the BHP assays were determined from samples centrifuged at 10,000 rpm for 10 min, filtered through 0.22 μm membranes, and acidified with concentrated H₂SO₄. Organic acids were determined using an HPLC system (Shimadzu LC-2050C; Oregon, Portland, OR, USA) equipped with a UV detector set at 214 nm. The chromatographic configuration included a HyperREZ XP H₂⁺ pre-column (UK) and a HyperREZ XP carbohydrate column (London, UK), operated at 55 °C with 5 mM sulfuric acid as the mobile phase at a flow rate of 0.6 mL/min [36]. The organic acids that could be determined by the applied HPLC method included lactate, acetate, formate, propionate, butyrate, isobutyrate, valerate, and isovalerate.

2.6. Data Processing

The cumulative hydrogen volumes obtained from the BHP assays were automatically normalized to standard temperature and pressure (0 °C and 1 atm) on a dry gas basis and expressed as net volumes after subtracting the average endogenous hydrogen volume quantified from the negative control condition. The experimental data were fitted to the modified Gompertz model (Equation (2)), which describes the evolution of cumulative hydrogen production as a function of time. As diauxic hydrogen production was observed, the model included a two consecutive step approach [37]. In Equation (2), H(t) represents the cumulative hydrogen volume (NmL) at time t (h), Hmax is the maximum cumulative hydrogen volume (NmL), Rmax is the maximum hydrogen production rate (NmL/h), and λ is the lag-phase duration (h) preceding the onset of active gas production. The subscripts 1 and 2 indicate that the kinetic parameters correspond to the first and second phases of the curve, respectively. The volumetric hydrogen production rate (VHPR, mL H₂/L-d) was calculated by dividing Rmax by the working volume (L).

$$H\left(t\right)=H_{max1}.exp\left\{-exp\left[{\displaystyle \frac{R_{max1}.2.71828}{H_{max1}}}.\left({\lambda }_1-\mathrm{t}\right)+1\right]\right\}+H_{max2}.exp\left\{-exp\left[{\displaystyle \frac{R_{max2}.2.71828}{H_{max2}}}.\left({\lambda }_2-\mathrm{t}\right)+1\right]\right\}$$

(2)

Additionally, a mass–energy balance was performed based on the experimental data obtained from the hydrothermal pretreatment and BHP assays. Calculations were performed on the basis of 1 ton of A. aculeata pulp. Two scenarios were considered: (i) scenario 1, evaluating hydrogen production from untreated pulp, and (ii) scenario 2, assessing hydrogen generation from the solid and liquid fractions obtained after hydrothermal pretreatment at 150, 200, and 250 °C. The mass flow of pulp fed to the SER was calculated from the total feed rate and total solid (TS) content, considering the operation time. After pretreatment and centrifugation, the masses of the solid and liquid fractions were recorded and expressed as kg of fraction per kg of pulp fed. Hydrogen production from solid fractions was estimated using BHP values expressed as NmL H₂/g VS fed and the VS content per unit of pulp (g VS/kg pulp). For liquid fractions, hydrogen production was determined from BHP values expressed as NmL H₂/g COD and the corresponding COD load (g COD/kg pulp). In both cases, the results were expressed as NL H₂/kg pulp. The equivalent energy yield (MJ/ton pulp) was calculated according to Equation (3), considering the higher heating value of hydrogen (12.7 MJ/m³) [35]. For pretreated conditions, the total energy output was obtained as the sum of the contributions from solid and liquid fractions. Finally, energy yields were scaled to a per-hectare basis (MJ/ha-year; Equation (4)), assuming an estimated annual production of 25 tons of fruit per hectare and that the pulp represents approximately 45% of the fruit weight [6,14]. All experimental values are expressed as mean ± standard deviation.

$$\mathrm{Energy} \mathrm{yield}={\mathrm{H}}_2 \mathrm{yield}\left({\displaystyle \frac{{\mathrm{m}}^3{\mathrm{H}}_2}{\mathrm{ton} \mathrm{of} \mathrm{pulp}}}\right)\times 12.74\left(\mathrm{MJ}/{\mathrm{m}}^3{\mathrm{H}}_2\right)$$

(3)

$$\mathrm{Annual} \mathrm{Energy} \mathrm{yield}=\mathrm{Energy} \mathrm{yield}\left({\displaystyle \frac{\mathrm{MJ}}{\mathrm{ton} \mathrm{of} \mathrm{pulp}}}\right)\times 25\left({\displaystyle \frac{\mathrm{ton} \mathrm{fruit}}{\mathrm{ha}·\mathrm{year}}}\right)\times {\displaystyle \frac{45\left(\mathrm{ton} \mathrm{pulp}\right)}{100\left(\mathrm{ton} \mathrm{fruit}\right)}}$$

(4)

3. Results and Discussion

3.1. Physicochemical Characterization of the Untreated Pulp

The compositional analysis of untreated A. aculeata pulp showed that the hemicellulosic fraction (35.64 ± 0.13 g/100 g DM) was slightly higher than the cellulosic fraction (31.93 ± 0.89 g/100 g DM). This pattern differs from that reported by Duarte et al. (2023) [38], who identified cellulose (45.42 g/100 g DM) as the main structural component of the same biomass, exceeding hemicellulose (15.89 g/100 g DM). Similarly, the contents of aqueous (3.63 ± 0.34 g/100 g DM) and organic extractives (8.77 ± 0.11 g/100 g DM) here determined were lower than those reported by the same authors (17.13 g/100 g DM). These discrepancies may be attributed to genetic, environmental, and agronomic factors influencing the composition of A. aculeata fruits, as well as to differences in pulp conditioning. In this study, the pulp was delipidized by ethanolic extraction prior to analysis, whereas Duarte et al. used a pulp cake supplied by an industrial pressing facility, where oil is mechanically extracted for soap production [38]. The pulp contained 28.69 ± 2.12% fat prior to ethanol defatting, which is within the range reported for A. aculeata pulp (22.5–39.6%) [39,40]. As shown in

Table 1. Physicochemical characterization of A. aculeata pulp and solid fractions obtained after subcritical water pretreatment.

Parameters	Untreated Pulp	P-S150	P-S200	P-S250
Total solids (g TS/Kg)	897.01 ± 1.12	170.35 ± 1.37	153.73 ± 1.06	104.60 ± 1.02
Volatile solids (g VS/kg)	837.91 ± 1.23	166.15 ± 1.84	150.64 ± 1.03	103.25 ± 1.05
VS/TS (%)	93%	97%	98%	99%
Insoluble lignin (g/100 g DM)	21.02 ± 0.79	39.09 ± 0.05	43.36 ± 0.05	45.04 ± 0.02
Soluble lignin (g/100 g DM)	3.85 ± 0.12	3.63 ± 0.02	3.19 ± 0.03	2.54 ± 0.03
Protein (g/100 g DM)	6.67 ± 0.06	4.53 ± 0.02	4.22 ± 0.07	4.15 ± 0.05

, the TS and VS contents determined for the untreated pulp in this study were comparable to those reported by Aguilar-Aguilar et al. [16] (95.7 ± 4.23% and 87.9 ± 3.98%, respectively). The VS/TS ratio exceeded 90%, indicating that most of the solid content consisted of organic matter and that the mineral fraction was relatively low. This high VS/TS ratio suggests a substantial fraction of biodegradable compounds potentially available for dark fermentation [41]. Similarly, the lignin content obtained in this study was slightly higher compared with that reported by Duarte et al. [38] (17.90%). Elemental analysis of the untreated pulp showed 43% C and 2.7% N (dry basis), corresponding to a C/N ratio of 16 on a mass basis, which may support active microbial growth during fermentative hydrogen production. The protein content of the untreated pulp (6.67 g/100 g DM) was consistent with values reported for A. aculeata pulp in previous studies [42].

3.2. Physicochemical Characterization of the Solid Fractions Obtained After Subcritical Water Hydrolysis

Subcritical water pretreatment influenced the physicochemical composition of the solid residues derived from A. aculeata pulp (Table 1). Since the pulp was diluted before pretreatment, the absolute values of TS and VS in the recovered fractions cannot be directly compared with those of the untreated pulp. The increase in the VS/TS ratio (from 93% in untreated pulp to 99% in P-S250) can be attributed to the preferential removal of soluble mineral components during pretreatment and washing, leading to an apparent enrichment of organic matter in the solid fractions. This change reflects a shift in the relative composition rather than an increase in the absolute amount of organic material. On the other hand, as the temperature increased, a consistent rise in the insoluble lignin content was observed (from 39.09 g/100 g DM in PS-150 to 45.04 g/100 g DM in P-S250), which can be primarily attributed to the preferential solubilization of hemicellulose and part of the cellulose under hydrothermal conditions [43]. Indeed, the soluble lignin decreased from 4.53 g/100 g DM in PS-150 to 2.54 g/100 g DM in P-S250. Finally, the protein content was similar among the different pretreated samples but lower compared to the untreated control, reflecting slightly solubilization of nitrogenous compounds into the liquid phase and/or potential thermal denaturation under subcritical conditions [44].

3.3. Physicochemical Characterization of the Liquid Fractions Obtained After Subcritical Water Hydrolysis

As shown in

Table 2. Physicochemical composition of the liquid fractions obtained after subcritical water hydrolysis of A. aculeata pulp.

Parameter	P-L150	P-L200	P-L250
Total organic carbon (g/L)	4.20 ± 0.3	5.39 ± 0.4	5.86 ± 0.4
Total nitrogen (g/L)	0.05 ± 0.005	0.07 ± 0.006	0.10 ± 0.006
Total COD (g O2/L)	11.62 ± 0.233	15.19 ± 0.186	16.97 ± 0.646
Soluble COD (g O2/L)	11.03 ± 0.280	13.25 ± 0.062	14.24 ± 0.124
pH	6.58 ± 0.05	5.79 ± 0.05	5.17 ± 0.04
Total sugars (g/L)	3.297 ± 0.289	3.790 ± 0.260	5.429 ± 0.327
Galacturonic acid (g/L)	0.352 ± 0.032	0.472 ± 0.044	0.473 ± 0.041
Acetic acid (g/L)	0.035 ± 0.003	0.115 ± 0.002	0.248 ± 0.003
Formic acid (g/L)	0.038 ± 0.001	0.048 ± 0.002	0.085 ± 0.002
5-HMF (g/L)	0.0010 ± 0.00	0.0011 ± 0.0002	0.0012 ± 0.0001

, increasing the pretreatment temperature from 150 to 250 °C promoted a consistent rise in the concentration of organic compounds in the liquid phase. TOC increased from 4.20 to 5.86 g/L, and the total nitrogen content from 0.05 to 0.10 g/L, indicating enhanced solubilization of both carbohydrate- and nitrogen-containing components from the pulp matrix. The parallel increase in total and soluble COD (from 11.62 and 11.03 g O₂/L to 16.97 and 14.24 g O₂/L, respectively) further supports the progressive transfer of organic matter into the aqueous phase under higher hydrothermal severity. The pH of the hydrolysates decreased from 6.58 at 150 °C to 5.17 at 250 °C, consistent with the formation of acidic compounds during hydrolysis. The progressive accumulation of acetic acid (from 0.035 g/L to 0.248 g/L) and formic acid (form 0.038 g/L to 0.085 g/L) indicates that deacetylation of hemicellulosic structures and partial oxidation of monosaccharides were intensified at higher temperatures. Total sugars also rose with temperature, from 3.30 g/L to 5.43 g/L, confirming the enhanced hydrolysis of structural polysaccharides under subcritical conditions. The simultaneous increase in galacturonic acid (from 0.35 g/L to 0.47 g/L) further suggests partial depolymerization of pectin-rich components from the pulp cell wall. The relatively low concentrations of 5-hydroxymethylfurfural (5-HMF < 0.0012 g/L) indicate that hexose degradation into furan derivatives was not a predominant pathway under the tested conditions. Organic acids such as acetic acid and trace furanic by-products such as 5-HMF are known to influence microbial activity, potentially leading to reduced hydrogen production in dark fermentation [45]. Despite this, the low levels detected in this study suggest that hydrolysates obtained under mild subcritical conditions could still be considered suitable substrates for biological hydrogen production (as discussed in Section 3.4).

3.4. BHP Tests

The cumulative hydrogen production profiles revealed distinct effects of the hydrothermal pretreatment severity on the dark fermentation of A. aculeata pulp (Figure 1). The untreated pulp showed the highest hydrogen yield (125.1 NmL H₂/g VS fed), while the solid fraction obtained at 150 °C exhibited a similar yield (118.1 NmL H₂/g VS fed), indicating that mild pretreatment did not compromise substrate fermentability (Figure 1,

Table 3. Biochemical hydrogen potential (BHP), kinetic parameters from the modified Gompertz model, and VHPR rates for the pulp and the solid fractions recovered from the hydrothermal pretreatment.

Parameters	Untreated Pulp	P-S150	P-S200	P-S250
BHP (NmL H2/g VSfed)	125.13 ± 4.68	118.09 ± 3.35	71.64 ± 4.54	41.64 ± 3.34
λ1 (h)	2.23 ± 0.03	4.04 ± 0.01	4.43 ± 0.83	2.44 ± 0.09
λ2 (h)	23.43 ± 1.41	23.69 ± 1.32	23.02 ± 1.76	30.71 ± 0.35
Hmax1 (NmL H2/L)	101.07 ± 11.60	60.58 ± 0.99	39.09 ± 2.01	26.59 ± 0.48
Hmax2 (NmL H2/L)	521.98 ± 29.46	551.94 ± 2.92	330.07 ± 14.52	181.91 ± 3.72
HmaxTotal (NmL H2/L)	623.05 ± 19.3	612.52 ± 14.6	369.16 ± 16.1	208.50 ± 4.6
Rmax1 (NmL H2/h)	5.09 ± 0.44	5.88 ± 0.02	4.17 ± 0.15	0.65 ± 0.02
Rmax2 (NmL H2/h)	13.27 ± 1.37	7.04 ± 0.20	4.31 ± 0.23	5.40 ± 0.06
R2	0.9995	0.9995	0.9992	0.9994
VHPRmax1 (NmL H2/L -h)	14.14 ± 1.21	16.34 ± 0.07	11.59 ± 0.43	1.79 ± 0.04
VHPRmax2 (mL H2/L -h)	36.87 ± 3.81	19.54 ± 0.57	11.98 ± 0.64	14.99 ± 0.18

). However, hydrogen yields markedly decreased at higher temperatures, reaching 71.6 and 41.6 NmL H₂/g VS fed at 200 °C and 250 °C, respectively, corresponding to reductions of about 39% and 65% relative to P-S150. These results suggest that excessive pretreatment severity can lead to the loss of fermentable carbohydrates and potentially promote the formation of inhibitory by-products such as organic acids and furan derivatives. However, in this study, the concentrations of detected compounds, including 5-HMF and acetic acid, remained below reported inhibitory thresholds, indicating that the differences in hydrogen yields are more likely linked to the nature of the fermentable substrates generated during pretreatment. Interestingly, despite its relatively high lignin content, the parenchymatous structure of A. aculeata pulp likely enabled good carbohydrate accessibility, as lignin is loosely associated with hemicellulose and cellulose in this soft matrix [38]. This intrinsic biodegradability explains the high hydrogen yields obtained from the untreated pulp, comparable to those of the mildly pretreated sample (150 °C). These findings demonstrate that efficient hydrogen production can be achieved without any pretreatment step, confirming that the native pulp structure is already suitable for microbial conversion and enabling a more cost-effective and sustainable process by avoiding the energy and equipment demands of hydrothermal pretreatment.

Regarding the liquid fractions, the cumulative hydrogen yields followed a similar trend to that observed for the solid residues (Figure 2,

Table 4. Biochemical hydrogen potential (BHP), kinetic parameters from the modified Gompertz model, and VHPR rates for the positive control and liquid fractions recovered from the hydrothermal pretreatment.

Parameters	Positive Control	P-L150	P-L200	P-L250
BHP (NmL H2/g CODfed)	110.76 ± 3.93	107.22 ± 4.06	79.41 ± 4.08	75.98 ± 4.03
λ1 (h)	3.38 ± 0.24	3.40 ± 0.19	5.00 ± 0.21	4.38 ± 0.34
λ2 (h)	-	20.41 ± 0.85	21.36 ± 0.69	24.08 ± 0.45
Hmax1 (NmL H2/L)	555.16 ± 12.30	87.77 ± 5.61	61.24 ± 10.25	49.64 ± 7.51
Hmax2 (NmL H2/L)	-	454.10 ± 19.31	337.55 ± 18.47	334.30 ± 5.77
HmaxTotal (NmL H2/L)	555.16 ± 12.30	541.87 ± 12.3	398.79 ± 15.6	383.93 ± 12.3
Rmax1 (NmL H2/h)	47.68 ± 1.87	3.91 ± 0.19	4.11 ± 0.94	1.39 ± 0.68
Rmax2 (NmL H2/h)	-	16.27 ± 0.09	15.28 ± 1.17	16.02 ± 1.10
R2	0.9992	0.9993	0.9995	0.9990
VHPRmax1 (NmL H2/L-h)	132.45 ± 5.19	10.87 ± 0.54	11.42 ± 2.61	3.87 ± 1.90
VHPRmax2 (NmL H2/L-h)	-	45.19 ± 0.24	42.44 ± 3.24	44.50 ± 3.04

). The positive control with glucose reached 110.8 NmL H₂/g COD fed (1.1 mol H₂/mol glucose utilized), which can be taken as a sign of acceptable hydrogenogenic activity of the inoculum used under the experimental conditions. Among the hydrolysates, the highest yield was obtained for P-L150 (107.2 NmL H₂/g COD fed), indicating that mild hydrothermal pretreatment effectively released soluble and readily fermentable compounds. In contrast, hydrogen yields decreased to 79.4 and 76.0 NmL H₂/g COD fed for P-L200 and P-L250, respectively, representing reductions of approximately 28–31% relative to P-L150. Although the total and soluble COD, as well as the concentration of total sugars, increased with pretreatment severity (Table 2), hydrogen yields from the liquid fractions decreased. This indicates that the additional organic matter solubilized at higher temperatures was less suitable for dark fermentation. The higher concentrations of acetic and formic acids, together with traces of 5-HMF, suggest partial sugar degradation during the hydrothermal treatment, reducing the overall fermentability of the hydrolysates [45].

The cumulative hydrogen production curves were successfully fitted to a sequential modified Gompertz model to describe the diauxic behavior observed for all assays (Figure 1 and Figure 2). The model accurately reproduced the experimental data (R² > 0.999). The kinetic parameters of each condition are summarized in Table 3 and Table 4. All liquid and solid fractions recovered from the hydrothermal pretreatment exhibited a diauxic pattern in hydrogen production. Particularly, the untreated pulp showed the highest total hydrogen potential per liter of reactor (Hmax_total = 623.1 NmL H₂/L) and short lag phases (λ₁ = 2.2 h; λ₂ = 23.4 h), with a markedly higher maximum hydrogen production rate during the second phase (Rmax₂ = 13.3 NmL H₂/h) compared to the first one (Rmax₁ = 5.1 NmL H₂/h). The kinetics of P-S150 were similar in total hydrogen potential (612.5 NmL H₂/L), whereas both hydrogen potential and production rates declined substantially at higher pretreatment severities. Regarding the liquid fractions, the first hydrogen-producing phase was characterized by short lag times (3–5 h), followed by a dominant second phase with higher hydrogen evolution rates. Increasing pretreatment temperature slightly prolonged the second lag phase (λ₂ = 20–24 h) and reduced the overall hydrogen potential (Hmax_total decreased from 542 NmL H₂/L at 150 °C to 384 NmL H₂/L at 250 °C). For both matrices, the second fermentation phase contributed the majority of hydrogen production, representing 84–90% of hydrogen in the solid fractions and 82–87% in the liquid hydrolysates from untreated and mildly pretreated samples. However, this contribution decreased markedly at higher pretreatment severities, as Hmax₂ dropped by 37% and 67% for P-S200 and P-S250, and by 38% and 41% for P-L200 and P-L250, respectively, relative to the corresponding 150 °C treatments.

In contrast to the pretreated solid and liquid fractions and the untreated pulp, the positive control exhibited a single, rapid hydrogen-producing phase with a short lag time (3 h) and a higher Rmax of 47.7 NmL H₂/h, showing no diauxic behavior (Table 4). When comparing the maximum volumetric hydrogen production rates (VHPRmax), the untreated pulp exhibited the highest value among the substrates (36.9 NmL H₂/L-h), followed by the solid and liquid fractions pretreated at 150 °C (19.5 and 45.2 NmL H₂/L-h, respectively). Increasing pretreatment severity led to a decline in VHPRmax, reaching 12.0–15.0 NmL H₂/L-h in the fractions obtained at 200–250 °C. In comparison, the positive control with glucose achieved a VHPRmax of 132.5 NmL H₂/L-h.

The diauxic pattern observed can be attributed to the sequential utilization of different carbohydrate fractions rather than to metabolic shifts within the microbial community, as the profiles of soluble metabolites remained qualitatively similar across all treatments (Figure 3) [37]. In agreement, the main metabolites detected at the end of fermentation for all conditions were acetate, butyrate, and propionate, which together accounted for over 90% of the total soluble products. Butyrate and acetate were consistently the predominant metabolites regardless of the tested condition, indicating a butyrate- and acetate-type fermentation pathway, whereas propionate appeared only in minor amounts (below 0.35 g/L) [46]. It should be noted that these organic acids represent only the final stage of batch fermentation and should therefore be interpreted with caution. Future studies should track the kinetics of metabolic intermediates to better resolve the H₂-producing pathways involved. This consistent metabolic pattern suggests that pretreatment severity primarily affected substrate availability and degradation rate rather than altering the fermentation routes. Consequently, the differences observed in the duration or intensity of the two hydrogen-producing phases are more likely linked to substrate accessibility and consumption dynamics, while the precise cause of the delayed second phase remains uncertain and may depend on substrate composition and microbial adaptation. In this context, the first, small hydrogen peak likely reflects readily accessible sugars, whereas the second peak arises from the gradual availability of structurally bound carbohydrates, which also explains why the diauxic pattern persisted after hydrothermal pretreatment. Unlike the batch BHP tests, which involve a single exposure of the substrate to the active microbial community, continuous systems are expected to minimize substrate accessibility limitations, as fresh material is constantly supplied and the microbial population can progressively enrich and stabilize its hydrogen-producing activity over the medium and long term.

To gain a more comprehensive understanding of the dark fermentation process, the gas composition was monitored during the BHP tests, while COD and total sugar concentrations were determined in the effluents at the end of the assays (

Table 5. Gas composition and effluent characteristics from biochemical hydrogen potential (BHP) tests.

Condition	%H2 (v/v)	Final pH	COD (g O2/L)	Residual Total Sugars (g/L)
Positive control	41.20 ± 4.03	5.84 ± 0.01	7.72 ± 0.72	0.74 ± 0.05
Negative control	24.58 ± 2.16	6.32 ± 0.02	7.24 ± 0.56	0.51 ± 0.15
Pulp	56.80 ± 3.76	4.78 ± 0.02	9.95 ± 0.66	1.13 ± 0.10
P-S150	52.92 ± 3.89	4.90 ± 0.01	10.01 ± 1.16	1.22 ± 0.10
P-S200	47.90 ± 2.34	5.27 ± 0.01	11.05 ± 1.23	1.51 ± 0.26
P-S250	40.91 ± 1.65	5.35 ± 0.01	11.11 ± 0.99	1.72 ± 0.10
P-L150	50.24 ± 1.76	5.03 ± 0.02	7.96 ± 0.91	0.89 ± 0.01
P-L200	49.94 ± 2.76	5.20 ± 0.02	8.93 ± 0.79	0.71 ± 0.01
P-L250	44.94 ± 2.54	5.18 ± 0.02	9.15 ± 0.68	0.64 ± 0.01

). In all cases, the off-gas produced contained only hydrogen and carbon dioxide, with no detectable methane under any of the tested conditions. The hydrogen content measured in the headspace during the accelerated production phase varied notably among treatments (Table 5). The untreated pulp exhibited the highest hydrogen fraction (56.8% v/v), followed by the solid (52.9%) and liquid (50.2%) fractions pretreated at 150 °C, in agreement with their higher cumulative hydrogen yields. Increasing pretreatment severity led to a progressive decline in gas-phase hydrogen concentration, reaching 40.9% and 44.9% for the solid and liquid fractions pretreated at 250 °C, respectively. The positive control reached, on average, 41.2% H₂, whereas the negative control exhibited the lowest value (24.6% H₂), showing relatively little hydrogen-producing activity. On the other hand, final pH values ranged from 4.8 to 5.4 for all A. aculeata assays, while the positive and negative controls stabilized at 5.8 and 6.3, respectively. The lowest pH occurred for the untreated pulp, consistent with higher activity. Although the pH of some assays approached the lower limit for dark fermentation, no apparent inhibition was observed in this study, suggesting that buffering could further enhance hydrogen yields. Regarding COD removal, after correcting for the residual COD of the blank controls (7.24 g O₂/L), the net COD remaining at the end of fermentation ranged from 0.7 to 3.9 g O₂/L (Table 5). The solid fractions showed higher residual COD values (2.7–3.9 g O₂/L) than the liquid hydrolysates (0.7–1.9 g O₂/L), indicating a lower degree of substrate conversion in the solid phase. Among the liquid fractions, P-L150 exhibited the lowest net COD. The increase in final COD at higher pretreatment temperatures suggests the accumulation of non-fermentable organic compounds. As for total sugars, after subtracting the residual sugars of the blank controls (0.51 g/L), the net sugar concentrations ranged from 0.1 to 1.2 g/L (Table 5). The liquid hydrolysates showed the lowest residual sugars (0.1–0.4 g/L), indicating efficient carbohydrate utilization, whereas the solid fractions retained higher amounts (0.6–1.2 g/L), suggesting the presence of less accessible carbohydrates in the solid matrix.

3.5. Mass and Energy Analysis

The mass–energy balance revealed that the untreated pulp (scenario 1) achieved the highest hydrogen production (104.8 m³ H₂/ton) and energy recovery (15,026 MJ/ha-year), confirming that the native pulp structure is highly suitable for dark fermentation and that no hydrothermal pretreatment is required to ensure efficient hydrogen generation (

Table 6. Hydrogen production and energy performance in the two scenarios evaluated.

Scenarios	Reaction Conditions	H2 Production Yield (m3/ton)	H2 Energy Yield (MJ/ha-year)	Total Energy Yield of H2 (MJ/ha-year)
Scenario 1: Raw material
Pulp	No reaction conditions applied	104.84 ± 12.15	15,026.19 ± 335	15,026.19 ± 335
Scenario 2: Subcritical water hydrolysis
P-S150	P = 170 bar; T = 150 °C; τ = 3.2 s.	35.56 ± 2.71	5096.64 ± 197	14,513.09 ± 457
P-L150		65.7 ± 3.87	9416.45 ± 412
P-S200	P = 170 bar; T = 200 °C; τ = 3.2 s.	26.43 ± 1.98	3788.08 ± 385	10,016.98 ± 632
P-L200		43.46 ± 3.86	6228.90 ± 501
P-S250	P = 170 bar; T = 250 °C; τ = 2.7 s	7.86 ± 1.35	1126.53 ± 156	9136.97 ± 436
P-L250		55.89 ± 4.78	8010.43 ± 407

). When subcritical water hydrolysis was applied (scenario 2), the mildest condition (P-S150) provided the highest total hydrogen energy yield (14,513 MJ/ha-year), due to the combined contribution of the solid and liquid fractions. When analyzing the energy distribution in P-S150, about 65% of the total hydrogen energy yield originated from the liquid hydrolysate, while the remaining 35% was contributed by the solid residue (Table 6). This indicates that most of the fermentable potential was transferred to the liquid phase during hydrolysis. However, in practical terms, if pretreatment were required, the separation of solid and liquid fractions would not be necessary, since both can be co-fermented as a single substrate stream in subsequent dark fermentation steps. At 200 °C, both fractions decreased in energy potential, maintaining a similar ratio but reflecting a loss of overall fermentable matter. Under the most severe condition (250 °C), nearly 88% of the total hydrogen energy originated from the liquid phase, confirming extensive solubilization of organics and a concomitant depletion of the solid matrix. However, this increased solubilization did not translate into higher total energy recovery. These results show that pretreatment reduces total energy recovery and adds thermal and operational costs, confirming that untreated pulp is the most energy-efficient option.

In Brazil, Acrocomia aculeata is still at an early stage of domestication and development as a commercial crop. Between 2018 and 2021, approximately 1750 hectares were established, increasing to around 2000 hectares in 2022, mainly under silvopastoral systems. Nevertheless, extractive harvesting remains the predominant practice within the production chain [10]. As an illustrative example, considering an energy yield of 15,026 MJ/ha-year obtained for the untreated pulp, the current cultivated area in Brazil would correspond to a potential hydrogen energy output of about 30,000 GJ/year. Although modest at the national scale, A. aculeata offers a realistic opportunity to foster decentralized biohydrogen production in Brazil’s emerging bioeconomy, supported by its cultivation model involving smallholders, extractive collection, and silvopastoral systems.

Beyond its potential as a single energy source, A. aculeata represents a versatile feedstock for complementary biofuel production routes within integrated biorefinery schemes. A. aculeata pulp can contain more than 28% (w/w, dry basis) oil [47]. Currently, oil extraction from A. aculeata pulp is mainly performed by mechanical pressing, leaving a residual lipid content of approximately 15% (w/w, dry basis) in the resulting press cake [48]. In the present study, ethanol defatting was applied to further reduce the lipid fraction, yielding a cake with about 10.5% (w/w, dry basis) residual fat content. In this sense, the defatted pulp cake can subsequently be valorized through dark fermentation to produce biohydrogen, providing a complementary route to biodiesel production within an integrated and sustainable biorefinery framework. Future studies should evaluate the fermentative hydrogen production potential of the residual pulp cake derived from the oil extraction process for biodiesel production. Continuous-operation studies are also needed, which will require a detailed assessment of the operating conditions and the microbial communities involved.

4. Conclusions

The present study provides the first systematic evaluation of Acrocomia aculeata pulp waste as a substrate for fermentative hydrogen production. The untreated pulp exhibited a favorable physicochemical composition, with high carbohydrate content and a low mineral fraction, confirming its suitability for biological conversion. Subcritical water hydrolysis progressively increased the solubilization of organic matter, nitrogen, and sugars, while also generating moderate amounts of organic acids and negligible levels of 5-HMF. BHP assays revealed that the untreated pulp achieved the highest hydrogen yield, while both solid and liquid fractions obtained at 150 °C maintained significant fermentative activity. Increasing pretreatment severity to 200–250 °C led to a marked decline in hydrogen yield, likely due to the partial degradation of fermentable carbohydrates. The metabolic profiles were consistently dominated by butyrate- and acetate-type pathways. The mass-energy balance demonstrated that the untreated pulp achieved the highest total energy recovery, whereas mild pretreatment resulted in a comparable yield with most of the hydrogen energy recovered from the liquid phase. In contrast, higher severities reduced the total energy recovery by up to 40%. Overall, these findings demonstrate that A. aculeata pulp waste inherently possesses a high fermentative hydrogen potential without requiring hydrothermal pretreatment, highlighting its relevance as a cost-effective and sustainable feedstock for renewable hydrogen production and its potential integration into circular agro-industrial biorefineries.