Low-Temperature Pretreatment (LT-PT) of Food Waste as a Strategy to Enhance Biomethane Production

Abstract

Food waste (FW) management remains a critical challenge within the circular economy framework. This study examines low-temperature pretreatment (LT-PT) of food waste and its effects on physicochemical transformations and microbial community dynamics. Artificial food waste (AFW) was subjected to LT-PT at 60 °C for 24 h, 48 h, and 72 h to assess changes in organic matter solubilization, nitrogen and phosphorus transformations, microbial composition, and biomethane potential. The results show that LT-PT promotes volatile fatty acid (VFA) accumulation, ammonification, and organic matter solubilization, thereby enhancing substrate biodegradability. The largest VFA increase was observed for acetate, whose concentration increased by approximately 0.55 g/L between 0 h and 72 h of LT-PT. Metagenomic analysis revealed a pronounced shift in microbial communities, with fermentative bacteria (Leuconostocaceae) increasing to 53.08% after 24 h of LT-PT, while Cyanobacteria decreased from 81.31% at 0 h to 19.48% at 48 h. Biochemical methane potential (BMP) tests demonstrated that longer LT-PT durations improved methane yield, with the highest production (1170 NmL CH₄) recorded after 72 h of pretreatment. Kinetic modeling using first-order and modified Gompertz equations confirmed that LT-PT enhances methane production efficiency by accelerating substrate hydrolysis. These findings indicate that LT-PT is a promising strategy for optimizing food waste valorization via anaerobic digestion, supporting sustainable waste management and renewable energy generation.

1. Introduction

Food waste (FW) and the continued reliance on fossil fuels represent two deeply interconnected global challenges with far-reaching environmental, economic, and social implications. Each year, an estimated 1.3 billion tons of food—nearly one-third of all food produced for human consumption—is lost or wasted globally. This staggering level of wastage contributes to approximately 3.3 gigatons of CO₂-equivalent emissions annually, not only due to the decomposition of organic matter but also as a result of the significant energy and resources invested in the cultivation, processing, transportation, and storage of food that is ultimately never consumed [1,2]. Moreover, due to its highly biodegradable nature, food waste decomposes rapidly, emitting unpleasant odors and creating serious risks to both public health and environmental quality [3]. In many regions, FW is typically managed through landfilling or incineration—practices that not only exacerbate environmental degradation but also represent a significant loss of potentially valuable resources. Simultaneously, the global energy system remains heavily reliant on carbon-intensive fossil fuels, a trend that continues to pose long-term ecological and climate risks. Collectively, these issues highlight an urgent need for—and growing societal and environmental pressure towards—the advancement of sustainable, renewable alternatives for both energy production and material utilization [4].

A potential solution is the bioconversion of FW into energy-rich products through anaerobic digestion (AD), where anaerobic microorganisms convert biomass into biomethane, leaving behind a nutrient-rich residue that can be utilized in other biorefinery processes [5]. AD stands out among bioenergy technologies for its capacity to handle a wide range of feedstocks, including those with high moisture content and the presence of impurities. It can be implemented in both large- and small-scale digesters across diverse geographic regions [6]. Although FW serves as a good substrate for AD, the process may experience reduced biogas yield due to the accumulation of harmful intermediate products resulting from inadequate process optimization and control [7]. Moreover, FW is composed of nitrogen, a component of proteins, which can inhibit acetoclastic methanogens and leads to the production of ammonia (NH₃) during fermentation. While hydrogenotrophic methanogens can continue biogas production, FW lacks certain trace elements essential for the accumulation of volatile fatty acids (VFA) [8]. Additionally, high solids contents in FW (∼48.6%) result in limitation of hydrolysis [3]. To address these limitations and enhance biomethane production from FW, substrate pretreatment is essential. This process breaks down complex polymer structures to facilitate more efficient digestion. Consequently, a range of pretreatment methods—including biological, chemical, mechanical, and thermal approaches—have been explored for treating FW prior to AD [9]. Among these methods, thermal pretreatment appears to be well-suited for the organic fraction of food waste in municipal waste programs (MWPs), as it typically contains a high moisture content (~81%) and a substantial number of volatile solids (~48.6%), along with significant proportions of hard-to-degrade components such as cellulose (~28.8%) and hemicellulose (~22.7%) [10]. Climent et al. [11] suggest that the optimal temperature for thermal pretreatment is in the range of 60 °C to 180 °C. However, temperatures exceeding 100 °C become progressively less advantageous, as they require a substantial energy input. Kasinath et al. [12] successfully carried out the disintegration of waste activated sludge (WAS) at temperatures below 60 °C under anoxic condition, achieving a 43% increase in methane potential. This suggests a strong potential for using low-temperature pretreatment (LT-PT) processes for the disintegration of biomass [13]. However, to the best of our knowledge, this approach has not yet been applied to food waste.

Therefore, in this study, a low-temperature pretreatment process was conducted on artificially prepared food waste designed to mimic the composition of the organic fraction from municipal waste plants (OFMWP). The objective was to identify the relationship between microorganisms and physicochemical parameters during LT-PT process of FW at 60 °C and then verify the efficiency of the disintegration process by subjecting the AFW to a biomethane potential test. The novelty of this research lies in the use of the LT-PT process for the disintegration of food waste, which has not yet been studied in this context. Previous studies focused on high-temperature disintegration of food waste (>100 °C) or low-temperature disintegration, but still at temperatures higher than 60 °C [14]. The findings of this study may contribute to the development of more energy-efficient and scalable pretreatment technologies, particularly suited for decentralized biogas systems in regions where high-temperature methods are economically or technically impractical. Furthermore, as global initiatives increasingly focus on transitioning to a circular economy and cutting greenhouse gas emissions, enhancing the efficiency of bioconversion processes—especially for abundant waste streams like food waste—has gained critical importance. Effective pretreatment strategies can contribute to more sustainable municipal waste management and help achieve renewable energy goals.

2. Materials and Methods

2.1. Food Waste Characteristic

In this study, artificial food waste (AFW) was utilized in low-temperature pretreatment (LT-PT) experiments. AFW is a standardized and carefully controlled material, making it a reliable reference for experimental studies. It is formulated to closely resemble the organic fraction of municipal waste processing (MWP) by combining specific ingredients. The AFW composition included bread (5% wt), rice (5% wt), pasta (5% wt), potato (22% wt), carrot (12% wt), cabbage (7% wt), tomato (10% wt), apple (20% wt), banana (5.5% wt), lemon (5.5% wt), coffee waste (0.25% wt), tea waste (0.25% wt), paper towels (1.5% wt), grass (0.5% wt), and leaves (0.5% wt). The pasta and rice were measured after boiling, whereas all other components were weighed in their raw, untreated forms. Hydration of AFW for LT-PT was maintained at ∼5% w/w (weight/weight percent) of total solid (TS). The weight ratios were estimated based on European consumption patterns for each product. Detailed physicochemical characteristics of AFW are presented in Table S1 (see Supplementary Materials).

2.2. Setup of Low-Temperature Pretreatment Process

Low-temperature pretreatment experiments were carried out using a laboratory-scale reactor with an active volume of 20 L. The reactor was equipped with a heat control system, mechanical stirring, and an air-pumping system to maintain precise process conditions. Detailed information on the LT-PT reactor design and its scheme (Figure S1) is provided in the Supplementary Materials. Each experiment treated 16 L of substrate under the following parameters: temperature of 60 °C; hydraulic retention times of 24, 48, or 72 h; mixing frequency of 35 Hz; dissolved oxygen (DO) concentration maintained above 0.2 mg/L. Dissolved oxygen (DO) was continuously monitored using DO probes installed directly in the LT-PT reactor. The probes were connected to a control unit regulating an air pump, which automatically supplied atmospheric air to the reactor whenever the DO concentration dropped below the set threshold. This feedback control system ensured stable DO levels throughout the process. Key parameters such as mixing frequency and DO levels were selected based on prior research and existing literature on LT-PT applications for waste-activated sludge and agricultural waste [12,13].

2.3. Physiochemical Analysis of Food Waste

The pH, electrical conductivity (EC), dissolved oxygen (DO) concentration, and oxidation-reduction potential (RedOx) were measured using an HQ40D portable multi-parameter meter (Hach, Germany). For further analysis, pretreated AFW samples were centrifuged at 10,000 RPM for 30 min, and the resulting supernatant was filtered through a 0.45 µm nitrocellulose membrane. The concentrations of key chemical parameters—including inorganic nitrogen species (N-NH₄⁺, N-NO₃⁻, N-NO₂⁻), total nitrogen (TN), total phosphate (TP), orthophosphate (P-PO₄³⁻), volatile fatty acids (VFA), and soluble chemical oxygen demand (sCOD)—were determined according to APHA standard methods [15] using a XION 500 spectrophotometer (Dr. Lange, GmbH, Hannover, Germany). Total solids (TS) and volatile solids (VS) were quantified via the gravimetric method. All analyses were conducted in triplicate to ensure reliability and reproducibility.

Short- and medium-chain carboxylic acids (C₂–C₈), as well as three common fermenting sugars (maltose, glucose, and fructose), were quantified using high-performance liquid chromatography (HPLC) on an Agilent 1200 system. The HPLC setup included a refractive index detector (RID) for sugar analysis and both RID and a diode array detector (DAD) operating in parallel for carboxylic acid detection, with RID used for quantification. Sample preparation involved adding 17 µL of 20% H₂SO₄ to 1 mL of AFW supernatant (post-centrifugation at 10,000 RPM), followed by filtration through proprietary silica-based spin filter columns (0.4 µm, 5000 RPM) developed in-house by AA Biotechnology (Poland). Analytes were separated using a Benson BP RA column (300 mm × 7.8 mm × 9.0 µm) packed with sulfonated poly(styrene-divinylbenzene). The retention times (RT) for carboxylic acids were as follows: citrate (3.47 min), lactate (5.28 min), acetate (6.31 min), propionate (7.40 min), and butyrate (9.00 min). For sugars, the RTs were: maltose (3.58 min), glucose (4.05 min), and fructose (4.30 min). Calibration was performed over a range of 0.125–10 mg/mL, with an injection volume of 10 µL.

2.4. Molecular Analysis

Bacterial community profiling was conducted based on genomic DNA extracted from artificial food waste samples using a standardized protocol. DNA was isolated from 3 mL of AFW collected at 0 h, 24 h, 48 h, and 72 h during the LT-PT process, utilizing a modified protocol with the Genomic Mini AX Bacteria+ kit (A&A Biotechnology, Gdansk, Poland). DNA concentration and purity were assessed using an NanoDrop™ ND-1000 UV-Vis spectrophotometer (Thermo Fisher Scientific Inc, Waltham, MA, USA), and integrity was verified through agarose gel electrophoresis. Extracted DNA samples were stored at 4 °C until further analysis.

The V3–V4 hypervariable region of the bacterial 16S rRNA gene was amplified using primers 341F and 785R for microbial community structure analysis. Sequencing was performed on the Illumina MiSeq™ platform (2 × 300 bp paired-end reads) (Macrogen Inc., Seoul, Republic of Korea) following the manufacturer’s standard protocol. The raw sequence data were initially processed using MiSeq Reporter (Illumina, San Diego, CA, USA), with paired-end reads joined using the FASTQ Joiner tool (Galaxy Version 2.0.1). Quality control was conducted using FASTQC (https://usegalaxy.org, accessed on 20 December 2024), applying a Phred quality score threshold of 20. Reads shorter than 120 bp were removed from downstream analyses. Operational taxonomic unit (OTU) clustering and taxonomic classification were performed using CLC Genomics Workbench v20.0 equipped with the CLC Microbial Genomics Module (QIAGEN, Hvidovre, Denmark). Taxonomic assignments were based on the Greengenes database, applying a sequence similarity threshold of 95%.

2.5. Biomethane Potential Test (BMP) and Kinetic Modeling

The biochemical methane potential (BMP) of AFW subjected to the LT-PT process was assessed using the Automatic Methane Potential Test System (AMPTSII, Bioprocess Control, Lund, Sweden). The quantities of AFW and digested sludge (inoculum) introduced into 400 mL AMPTSII reactors were calculated in accordance with the Bioprocess Control user manual. The inoculum was obtained from the Gdynia-Dębogórze wastewater treatment plant (WWTP), which serves a population equivalent (PE) of 440,000 and processes approximately 73,000 m³/day of wastewater. Prior to testing, frozen AFW was thawed and diluted with distilled water to achieve a total solids (TS) concentration of approximately 5%, a standard value recommended for AD. To establish anaerobic conditions, each reactor containing the AFW–inoculum mixture was purged with nitrogen gas. The generated biogas was directed through a CO₂ absorption unit—comprising bottles filled with 3 M NaOH solution—prior to entering the AMPTSII measurement system, where the biomethane volume was quantified via water displacement. Gas production data were continuously recorded and analyzed using Bioprocess Control™ software (version from 27 April 2020).

Based on obtained data, evaluation of methanogenesis kinetics was performed using the first-order kinetic and modified Gompertz [16] according to Equation (1) and Equation (2), respectively.

$$\mathit{V}\left(\mathit{t}\right)={\mathit{V}}_{\mathit{m}}(\mathbf{1}-{\mathit{e}}^{-\mathit{k}\mathit{t}})$$

(1)

$$\mathit{V}\left(\mathit{t}\right)={\mathit{V}}_{\mathit{m}}·\mathit{e}\mathit{x}\mathit{p}\left\{-\mathit{e}\mathit{x}\mathit{p}\left[\mathbf{1}+(\mathit{\lambda }-\mathit{t}){\displaystyle \frac{\mathit{R}\mathit{e}}{{\mathit{V}}_{\mathit{m}}}}\right]\right\}$$

(2)

where:

• V(t)—cumulative methane production at time t (NmL);
• V_m—experimental methane production potential (NmL);
• k—kinetic methanogenesis rate constant (d⁻¹);
• t—cumulative time for the methane production (d);
• e—mathematical constant (2.718282);
• λ—lag phase for methane production (d);
• R—maximum methane production rate (NmL/d),

Non-linear regression analyses were conducted to estimate the kinetic constants (k, λ, and R). A comparison between experimental and simulated methane production data, based on the first-order kinetic and modified Gompertz models, is shown in Figure S1 (see Supplementary Data).

2.6. Data Analysis

Data analysis and visualization were performed using the R Project for Statistical Computing (https://www.R-project.org, accessed on 20.12.2024), with graphical outputs generated via the ggplot2 package in R [17]. VFA and fermentable sugar production data, as well as BMPs, were analyzed and visualized using GraphPad Prism 10.

3. Results

3.1. Changes in Physiochemical Parameters of Food Waste

During the LT-PT process, complex organic compounds are degraded through the combined effects of elevated temperature and microbial activity. This decomposition is accompanied by changes in the concentrations of mineral and organic compounds, which are critical for the downstream utilization of AFW after LT-PT treatment.

Table 1. Chemical parameter changes during LT-PT of AFW at 60 °C. Soluble chemical oxygen demand (sCOD), volatile fatty acids (VFA), total nitrogen (TN), ammonium nitrogen (N-NH₄⁺), nitrite nitrogen (N-NO₂⁻), nitrate nitrogen (N-NO₃⁻), total phosphorus (TP), and phosphate phosphorus (P-PO₄³⁻).

Time of LT-PT	sCOD (g/L)	VFA (g/L)	TN (mg/L)	N-NH4+ (mg/L)	N-NO2− (mg/L)	N-NO3− (mg/L)	TP (mg/L)	P-PO43− (mg/L)
0 h	18	0.77	240	10.2	0.027	26.4	49.4	1.355
24 h	23	0.95	225	19.6	0.017	27.0	62.4	0.892
48 h	23	1.11	232	25.8	0.008	23.6	59.6	0.475
72 h	24	1.09	245	31.9	0.007	25.6	58	0.719

summarizes the variations in soluble chemical oxygen demand (sCOD), volatile fatty acids (VFAs), total nitrogen (TN), ammonium nitrogen (N-NH₄⁺), nitrite nitrogen (N-NO₂⁻), nitrate nitrogen (N-NO₃⁻), total phosphorus (TP), and phosphate phosphorus (P-PO₄³⁻) during LT-PT treatment at 0 h (untreated AFW), 24 h, 36 h, and 72 h. The results show that sCOD gradually increased from 18 g/L at 0 h to 24 g/L at 72 h, indicating slight solubilization of organic matter. Similarly, VFA concentrations increased from 0.77 g/L at 0 h to a peak of 1.11 g/L at 48 h, before slightly declining to 1.09 g/L at 72 h, suggesting both VFA accumulation and microbial consumption. Nitrogen transformations were particularly notable. TN initially decreased from 240 mg/L to 225 mg/L at 24 h, then fluctuated slightly before increasing to 245 mg/L at 72 h. N-NH₄⁺ concentrations showed a steady rise, from 10.2 mg/L at 0 h to 31.9 mg/L at 72 h, reflecting active ammonification. In contrast, N-NO₂⁻ levels declined from 0.027 mg/L to 0.007 mg/L, indicating nitrite reduction, while N-NO₃⁻ remained relatively stable (23.6–27.0 mg/L) with no significant accumulation observed. Phosphorus dynamics also varied over time. TP increased from 49.4 mg/L at 0 h to a maximum of 62.4 mg/L at 24 h, before decreasing to 58 mg/L at 72 h. Meanwhile, P-PO₄³⁻ concentrations consistently decreased from 1.355 mg/L at 0 h to a minimum of 0.475 mg/L at 48 h, then slightly increased to 0.719 mg/L at 72 h, suggesting phosphorus release followed by potential microbial uptake. Overall, these findings indicate that LT-PT treatment influences both organic matter solubilization and nitrogen/phosphorus transformations, with notable ammonification and VFA accumulation occurring over time.

Figure 1 illustrates the variations in five VFAs—acetate, propionate, butyrate, citrate, and lactate (Figure 1a)—as well as three fermented sugars (maltose, glucose, and fructose) and total sugars (Figure 1b) during LT-PT process. Acetate was undetectable in untreated AFW but increased progressively from 0.27 g/L at 48 h to 0.55 g/L at 72 h, indicating strong activity of bacteria with enzymatic pathways for acetate production. Similarly, propionate levels rose from 0.07 g/L at 0 h to 0.33 g/L at 72 h. Butyrate was absent at earlier time points but appeared at 72 h (0.3 g/L), suggesting delayed formation during LT-PT. Dahiya and Mohan [18] reported a relationship between acetate and higher-chain fatty acids, noting that in the presence of calcium ions (Ca²⁺), acetate can be converted into propionate and butyrate. As previously reported by Liu et al. [19], food waste contains a substantial amount of calcium. This may account for the observation that rising acetate concentrations are associated with increases in propionate and butyrate. Citrate concentrations showed minor fluctuations, increasing from 1.03 g/L at 0 h to 1.21 g/L at 72 h, likely reflecting the presence of lemon in the AFW composition. Lactate, initially minimal (0.01 g/L), gradually increased to 0.24 g/L at 72 h, consistent with microbial metabolic activity. Previous studies have confirmed the role of lactic acid bacteria during thermal treatment of food waste [20,21]. In addition, temperatures below 60 °C and pH values between 4 and 7 are favorable for lactic acid production [19]. Conversely, Li et al. [22] demonstrated that lactic acid production is lower under thermophilic conditions (50 °C) than under mesophilic conditions (35 °C), suggesting that the temperature used in this study (60 °C) is not optimal for lactic acid accumulation. Among sugars, maltose concentrations decreased from 3.26 g/L at 0 h to 2.47 g/L at 48 h, before slightly increasing to 2.76 g/L at 72 h, indicating partial microbial consumption followed by possible hydrolysis of complex carbohydrates. Glucose increased steadily from 5.49 g/L at 0 h to 6.06 g/L at 72 h, suggesting enzymatic release from polysaccharides. Fructose levels fluctuated, peaking at 4.09 g/L at 24 h before stabilizing around 3.65 g/L at 72 h. Total sugar content remained relatively stable, starting at 12.15 g/L and reaching a maximum of 12.86 g/L at 24 h, then decreasing to 11.95 g/L at 48 h and recovering to 12.47 g/L at 72 h. Overall, these results highlight dynamic transformations of the organic fraction of AFW during LT-PT, characterized by the accumulation of organic acids and shifts in sugar profiles, with acetate and propionate showing the most pronounced increases.

3.2. Microbial Community Structure

In systems for the biological production of bio-methane, the bacterial community is a key factor in assessing process efficiency. The performance and kinetics of methane fermentation are linked to the composition of the sludge microbiome and its metabolic pathways [23]. Similarly, bacteria play a crucial role in the LT-PT process, participating in the transformation of organic compounds, as well as the nitrogen and phosphorus cycle. Therefore, understanding the composition of the bacterial community will help to comprehend the transformations occurring in AFW during the LT-PT process.

The microbial community composition underwent significant changes over each tasted period following low-temperature pretreatment of AFW. The observed shifts indicate microbial succession patterns influenced by the thermal conditions and substrate availability. At the phylum level (Figure 2a), Cyanobacteria were the dominant phylum, accounting for 81.31% of the total microbial population in untreated AFW. Their initial prevalence suggests that they were well-adapted to the untreated food waste environment. Moreover, previous studies have confirmed that fruit-rich substrates can serve as a favorable medium for the growth of this group of microorganisms, due to the presence of fructose [24]. However, after low temperature pretreatment, their abundance declined sharply to 20.79% at 24 h, with a further reduction to 19.48% at 48 h before stabilizing at 25.62% at 72 h. This decline suggests that heat exposure negatively affected their survival or competitiveness. Cyanobacteria, typically associated with phototrophic or aquatic environments, may have been unable to efficiently utilize the organic substrates released during AFW breakdown under anoxic conditions [25]. In contrast, Firmicutes experienced a significant increase, rising from 1.99% at 0 h to 68.11% at 24 h, reaching 70.96% at 48 h, and remaining the dominant phylum at 63.78% at 72 h. This rapid expansion suggests that LT-PT created favorable conditions for Firmicutes growth, likely by releasing bioavailable carbohydrates and simple sugars from AFW. Many Firmicutes, particularly Clostridium and Bacillus species, are known for their ability to ferment complex polysaccharides and produce volatile fatty acids (VFAs) under anaerobic/anoxic conditions [26,27]. Their dominance indicates that they play a critical role in pretreatment AFW, contributing to breaking down and utilizing complex compounds such us cellulose, hemicellulose, and lignocellulosic materials [28]. Proteobacteria showed relatively minor fluctuations, maintaining a steady presence at each pretreatment time. Their abundance declined slightly from 16.09% at 0 h to 11.05% at 24 h, then further to 9.51% at 48 h, and remained relatively stable at 10.53% at 72 h. This trend suggests that Proteobacteria were not significantly impacted by thermal pretreatment and continued to perform metabolic functions necessary for AFW degradation. This aligns with previous findings that showed an abundance of Proteobacteria depending on the operating conditions of the process [29]. Moreover, several Proteobacteria, such as those belonging to the genera Pseudomonas and Enterobacter, are known for their role in organic matter decomposition and nitrogen cycling [30]. Actinobacteria remained a minor component throughout the study, representing only 0.29% in untreated AFW, followed by a sharp decline to 0.01% after 24 h and 48 h of LT-PT, before showing a slight increase to 0.02% at 72 h. Actinobacteria are commonly associated with cellulose degradation and soil environments, which might explain their low abundance in the LT-PT process in AFW [31]. Overall, the LT-PT process enhances the dominance of fermentative bacteria (Firmicutes), reduces phototrophic or non-fermentative groups (Cyanobacteria), and maintains a balanced microbial ecosystem essential for efficient organic waste bioconversion. These findings provide insights for optimizing microbial community composition in bioenergy applications, particularly in improving biogas yield and volatile fatty acid production from pretreated food waste.

At the genus level, Streptophyta_Unknown Family was the most dominant taxon (at 0 h), comprising 81.3% of the total microbial community. However, following low-temperature pretreatment, the abundance of Streptophyta_Unknown Family decreased to 20.79% at 24 h and remained low at 19.48% at 48 h, before slightly increasing to 25.62% at 72 h. Lu et al. [32] identified this bacterium as dominant during anaerobic composting of food waste; therefore, the anoxic conditions in this study may not have been favorable for its development. The most significant increase in abundance was observed in Leuconostocaceae (Unknown Genus), which increased from a negligible 0.34% in untreated AFW to become the dominant group at 53.08% at 24 h. The abundance slightly decreased but remained high at 40.91% at 48 h and 39.1% at 72 h. Similarly, the genus Leuconostoc, a well-known lactic acid bacterium [33], exhibited a substantial increase from 0.27% at 0 h to 14.68% at 24 h, further rising to 29.31% at 48 h before stabilizing at 23.85% at 72 h. These trends suggest that the LT-PT process created favorable conditions for this bacteria, likely due to the release of fermentable sugars. Leuconostocaceae (Unknown Genus) are known for their role in food fermentation, producing organic acids that lower pH and inhibit competing microbes [34]. Moreover, the primary metabolism of Leuconostoc species involves the conversion of simple sugars into lactate, acetate, and ethanol [33], which explains the observed increase in acetate and lactate production in this study (see Figure 1).

Members of the Enterobacteriaceae family, including an unknown genus and Pantoea, increased significantly following pretreatment. Enterobacteriaceae (Unknown Genus) increased from 0.87% at 0 h to 7.09% at 24 h, subsequently remaining at similar levels (6.14% at 48 h and 6.35% at 72 h). Likewise, Pantoea, a facultative anaerobe, increased from 0.16% at 0 h to 0.8% at 24 h, maintaining similar levels through 72 h. This suggests that low temperature conditions selectively enriched Enterobacteriaceae, likely due to their ability to metabolize diverse organic substrates released during FW breakdown. Moreover, Pantoea is characterized for phosphate solubilizing properties [35] and participates in the nitrogen cycle by removing nitrite [36]. Unlike the major shifts observed in other taxa, some genera maintained a consistently low abundance. Bacillus, an important fermentative and spore-forming genus, decreased from 0.96% at 0 h to 0.04% at 24 h, showing no significant resurgence. Acinetobacter also remained at low levels (0.76% at 0 h to 0.08% at 24 h, stabilizing at 0.06–0.08% thereafter). Pseudomonas, another facultative anaerobe, exhibited a minimal increase from 0.66% at 0 h to 0.73% at 72 h. These trends suggest that these genera were either unaffected by the low-temperature pretreatment or were outcompeted by faster-growing fermentative bacteria.

The results indicate that the LT-PT process significantly shifts microbial community composition, favoring lactic acid bacteria (Leuconostocaceae (Unknown Genus), Leuconostoc) and facultative anaerobic fermenters (Enterobacteriaceae, Pantoea) while reducing photorophic microbes (Streptophyta_). The dominance of lactic acid bacteria is particularly beneficial for VFA production, which is essential in anaerobic digestion and biogas generation. The reduction in non-fermentative microbes suggests that pretreatment improves microbial selectivity, creating an environment more suitable for efficient FW bioconversion. Overall, LT-PT enhances microbial succession toward fermentative and facultative anaerobic bacteria, optimizing conditions for FW valorization and bioconversion processes.

Principal Component Analysis (PCA) was applied to assess the relationships between chemical parameters and microbial community composition (Figure 3). The results show that N-NH₄⁺ and VFAs are clustered together, suggesting that these parameters are interrelated, potentially due to organic matter decomposition and nutrient cycling. Moreover, sCOD aligns closely with VFAs, supporting its role in fermentation processes that drive microbial metabolism. The relationship between bacterial communities and chemical parameters suggests that Leuconostoc may play a crucial role in fermentation and the breakdown of organic matter, likely associated with elevated N-NH₄⁺, VFA, and sCOD levels. Bacillus, Acinetobacter and Streptophyta_Unknown show moderate contributions along PC1, implying that they may be involved in nitrogen transformations (as they correlate with NO₃⁻ and NO₂⁻). On the other hand, Pseudomonas is positioned distinctly from most other microbial genera, which may indicate a different ecological role or sensitivity to physicochemical changes.

3.3. Biomethane Production and Process Kinetic

The results of the biochemical methane potential (BMP) test after LT-PT demonstrate a clear enhancement in methane production from AFW (Figure 4). The methane production increased significantly with longer pretreatment durations, with the highest cumulative methane production observed at 72 h, reaching 1170 NmL by day 12. In contrast, the untreated food waste (0 h condition) showed methane generation with a final value of 641 NmL. This confirms that LT-PT improves the anaerobic digestion process by making organic matter more accessible to microbial degradation. On the other hand, the biomethane yield obtained in this study is significantly lower than the theoretical BMP value for food waste generated in a university canteen, which was estimated by Browne and Murphy [37] at 549 NmL CH₄/g VS. In our experiment, the highest yield—achieved after 72 h of LT-PT—was 250 NmL CH₄/g VS. Although the theoretical value was not achieved, as noted by Ahire et al. [38], reaching this value is highly challenging, as it assumes that the bacteria involved in the process consume nitrogen (N), carbon (C), oxygen (O), and hydrogen (H) atoms without accounting for the presence of inhibitors. In terms of methane production kinetics, the data show a rapid increase in methane generation within the first three days, particularly in the 24 h, 48 h, and 72 h conditions. This suggests that LT-PT enhances hydrolysis efficiency, allowing for faster breakdown of food waste. This phenomenon was previously confirmed by Azizi et al. [16] and Kasinath et al. [12] in terms of LT-PT of WAS. After this initial surge, the methane yield continued to rise but at a slower rate, eventually reaching a plateau. The 48 h and 72 h conditions peaked earlier, around day 6 or 7, while the untreated and 24 h conditions took longer to stabilize. Notably, the 72 h pretreatment continued to show slight increases in methane yield even beyond day 7, indicating prolonged microbial activity and substrate availability. When comparing different pretreatment durations, the shorter 24 h pretreatment showed moderate improvements over the untreated condition but reached its peak earlier and at a lower value (539 NmL). The 48 h pretreatment resulted in a substantial increase in methane production, with values stabilizing at 509 NmL by day 5. However, it was the 72 h pretreatment that consistently outperformed all other conditions, producing the highest methane production throughout the experiment. This suggests that a longer LT-PT duration leads to greater solubilization of organic matter, making more biodegradable compounds available for microbial conversion.

To sum up, LT-PT significantly improves the biomethane potential of food waste, with the most substantial benefits observed at a pretreatment duration of 72 h. While this extended duration yields the highest methane production, further research is needed to determine the most energy-efficient approach that balances methane yield with operational costs. These findings reinforce the potential of LT-PT as a valuable pretreatment method for optimizing anaerobic digestion and improving food waste-to-energy conversion efficiency.

To evaluate methanogenesis kinetics, both the first-order kinetic model and the modified Gompertz model were applied to estimate key parameters: the methanogenesis rate constant (k), maximum methane production potential (Vm), maximum methane production rate (R_m), and lag phase duration (λ), as shown in

Table 2. Kinetic parameters for the first-order and modified Gompertz models.

LT-PT Time	First-Order Model		Modified Gompertz Model				Experimental
	k (d−1)	R2	Vm (NmL)	Rm (NmL/d)	λ (d)	R2	Vm (NmL)
0 h	147.0087	0.4425	653.9598	80.5392	−0.8084	0.9729	641
24 h	193.9367	0.6562	531.8467	178.9863	−0.0445	0.9903	539
48 h	193.2452	0.6602	507.9874	179.8553	−0.0033	0.9912	509
72 h	143.5602	0.3906	1175.8891	205.2382	0.0944	0.9964	1170

. The first-order kinetic model estimated the rate constant (k), which varied across different process times. The highest k value was obtained at 24 h (193.94 d⁻¹), followed by 48 h (193.25 d⁻¹), whereas the lowest was recorded at 72 h (143.56 d⁻¹). The coefficient of determination (R²) ranged from 0.39 to 0.66, indicating moderate predictive accuracy. In contrast, the modified Gompertz model provided deeper insights into methanogenesis dynamics. The highest methane production potential (P) was observed at 72 h (1175.89 NmL), while lower values were obtained at 48 h (507.99 NmL) and 24 h (531.85 NmL). The maximum methane production rate (Rm) showed a similar trend, peaking at 72 h (205.24 NmL/d), with slightly lower values at 48 h (179.86 NmL/d) and 24 h (178.99 NmL/d). The lag phase (λ) remained negligible across all conditions, with negative values at 0 h (−0.81), 24 h (−0.04), and 48 h (−0.0033), suggesting immediate methane production. A slight positive λ value was recorded at 72 h (0.0944 d), indicating a minor delay in methane initiation. According to Azizi et al. [16], methanogens may gradually assimilate dissolved organic compounds (inhibitory compounds) released during the LT-PT process. The modified Gompertz model exhibited consistently high R² values (0.97–0.99), demonstrating a superior fit compared to the first-order kinetic model. Overall, the results indicate that methanogenesis kinetics vary with incubation time, with longer durations (72 h) supporting higher methane production rates and cumulative yields, albeit with a slight delay in onset. These findings highlight the importance of optimizing incubation time to maximize biogas production efficiency.

4. Conclusions

The present study confirms that the LT-PT process at 60 °C significantly enhances the anaerobic biodegradability of food waste (FW), as evidenced by increased solubilization of organic matter, accumulation of volatile fatty acids (VFAs), and intensified ammonification processes. The physicochemical transformations observed during LT-PT were strongly correlated with microbial community shifts, notably the proliferation of fermentative bacteria such as Leuconostocaceae_unknown and Leuconostoc whose abundance jointly reached 72% after 48 h of LT-PT process. Biomethane potential testing revealed that extending the LT-PT duration up to 72 h led to a significant increase in methane production (1170 NmL) compared to untreated FW. Kinetic modeling using the modified Gompertz equation confirmed improved methanogenesis rates and higher methane production potential under extended pretreatment conditions, indicating enhanced hydrolysis and fermentation efficiency. These findings demonstrate the feasibility and efficacy of LT-PT as a pretreatment strategy for improving biomethane production from FW. The method may offer particular advantages in terms of energy efficiency and process stability, especially in small-scale biogas installations where high-temperature systems are impractical. Further optimization of operational parameters, such as retention time, co-substrate integration, or enzymatic supplementation, may further increase the efficiency and scalability of this approach.