Enhancement of Anaerobic Digestion from Food Waste via Ultrafine Wet Milling Pretreatment: Simulation, Performance, and Mechanisms
[ad_1]
1. Introduction
A third of the food produced worldwide for human consumption goes to waste every year throughout the supply chain for food, amounting to over 1.6 billion tons [1]. Inadequate management of food waste (FW) can lead to the release of greenhouse gases (GHGs) and other detrimental compounds due to unregulated decomposition. According to reports, FW is responsible for 6.7% of the total GHG emissions caused by human activities worldwide [2]. Given the abundance of natural fiber, carbon, protein, fat, and other organic components in food waste, anaerobic digestion, considered an effective treatment, has gained significant attention. The rate-limiting stage in the AD of multifaceted, hard biodegradable, organic substrates has been proven to be hydrolysis [3]. Hence, a multitude of pretreatment methodologies have been devised to dismantle the intricate composition of FW and enhance the microbial interface with the digesting substrate [4].
Pretreatment technologies aimed at enhancing the hydrolysis stage can be broadly categorized into physical and mechanical methods (e.g., mechanical pulverization, thermal, ultrasonic, microwave, etc.), chemical oxidative pretreatment, and biological pretreatment. Mechanical milling has been proven as an effective technique for reducing particle size and disrupting structure, but it has been criticized for its high energy consumption [5]. Recent research has discovered that the use of chemical pretreatment might result in the presence of reagent residues and inhibitory chemicals, which have a detrimental effect on AD [6,7]. Furthermore, the exorbitant expenses and intricate operational procedures associated with biological pretreatment pose significant challenges to its implementation in practical applications [8]. Consequently, in the case of intricate FW matrices, the pursuit of procedures that employ a reduced number of chemical additives, generate less amount of waste, and prioritize operational ease, has resulted in the revival of milling [9,10].
The three main grinding modes include hammer milling, wet disc milling, and ball milling [10,11]. Ball milling is the predominant technique for grinding FW given its elevated moisture content, which promotes particle contact and chemical reactions through impacts with medium balls and FW particles. Regarding the decrease in particle size, the majority of studies have concentrated on decreasing particle diameters from centimeters to millimeters [12,13]. These studies have examined the production of methane and the stability of the process, which have demonstrated a notable increase in methane. For example, Izumi, K et al. [14] managed to decrease the mean particle size (MPS) of FW from 0.843 to 0.391 mm by bead milling, which received a 28% boost in methane production.
The hydrolysis of FW is considerably influenced by particle size, which can affect surface properties, specific surface area, solubilization of macromolecules, etc. [15]. Researchers have conducted detailed investigations into the impact of grinding on organic matter dissolution and found that ball milling can lead to depolymerization of polysaccharide chains, disruption of ordered fibrous matrix, and degradation of lignin–carbohydrate complex bonds. As a result, the solubility of carbohydrates is enhanced without the requirement of enzyme addition [16]. Additionally, researchers have found that the grinding pressure minimally affects the growth of existing microorganisms in the original matrix, and smaller particle sizes in reactors result in an increase in the proportion of hydrolytic bacteria such as Porphyomonadaceae and Bacteoidales_UCG-001 through microbial investigations [17,18,19]. However, previous studies have not investigated the changes in particle size during continuous grinding processes or the performance of anaerobic digestion of FW at micron scales.
The UFWM process is capable of processing feeds with a reasonably large particle size (several millimeters) and reducing it to a smaller size of 10 µm [20]. Nevertheless, there is currently no known work that has successfully achieved a reduction in the D90 particle size of FW to less than 0.1 mm and its AD performance. The change in substrate morphology and surface structure is also worth studying, which can explain the effects of particle morphology due to collision during the milling process. Additionally, the potential correlation between FW particles and microbial pathways has not been clarified.
This study aimed to examine the decrease in FW particle size by UFWM and its effects on the properties of FW particles and methane generation throughout the AD process. Subsequently, three kinetic models were utilized to accurately match the empirically observed methane production and elucidate the reaction kinetics of AD under various milling times. This study sought to clarify the various ways in which UFWM might enhance AD of FW. This was achieved by examining the physical, chemical, and microbiological changes that occur in FW. The results are anticipated to explain the mechanism of particle size reduction to enhance the anaerobic digestion of food waste from multiple physical, chemical, and microbiological perspectives and offer direction for the practical treatment of FW and the recovery of energy.
2. Materials and Methods
2.1. Food Waste and Inoculum
FW used for pretreatment tests and AD tests was obtained from the refectory of Tongji University, Shanghai, China. To make the samples more representative, the mixture was collected after lunch and dinner for 5 consecutive days. Before pretreatment, large bones, eggshells, chopsticks, plastics, and other components that are difficult to biodegrade are manually picked out. Then, the collected FW is shattered by homogenizer for 30 s as the original substrate and stored at −20 °C until required for the experiments.
The inoculated sludge employed for the AD experiment was obtained from a municipal WWTP in Shanghai, China. After being left undisturbed for 12 h, the sludge was filtered using a 1 mm screen to eliminate any sizable particles or debris. Before AD, acclimation of the inoculum was carried out by adding 15 g/L (TS) fresh sludge once, two days over 30 days, incubating at 37 °C. The properties of inoculum and FW used for AD are listed in Table S1.
2.2. UFWM Pretreatment
The FW was milled to reduce the particle size (PS) to a micron level by a stirred ball–attritor (WR, JM-2A, CSWRmill, Changsha, China). Parameters of the machine were based on the preliminary investigation: mixing shaft speed (1500 rpm), material solids content (5%), ball material ratio (2), and ball size grading (1.4–1.6 mm and 2 mm). The following is the process of milling to disintegrate FW: Firstly, the FW slurry was adjusted to 5% TS using ultrapure waster Next, 800 g adjusted FW was ground by the attritor in the presence of 400 g of zirconium oxide spheres with particle size gradation. Six types of AD substrates were pretreated by varying the operation time based on hours at 1500 rpm. The slurries were collected through a 2 mm screen and stored at 4 °C (control, M-0.5h, M-1h, M-2h, M-3h, M-4h).
2.3. Anaerobic Digestion Experiment Design
A batch AD experiment was carried out in replicate gas-tight glass flasks with a working volume of 250 mL. Prior to filling flasks with substrates, each flask underwent meticulous inspection to guarantee its airtightness. The ground FW and inoculum were mixed in a 2:1 ratio (based on VS) and the remainder was supplemented with distilled water, with two additional parallels per particle size and a blank sample. Before AD, the pH of the mixed substrates was regulated to a value of 7.0 ± 0.1. Each bottle was infused with nitrogen gas for a duration of 3 min in order to establish an atmosphere devoid of oxygen and then incubated at 37 ± 1 °C and 150 rpm in a constant temperature shaker. The experiment lasted 20 days until the methane cumulation curve flattened. Methane production and proportion were measured daily. pH, VFA composition, and NH4+-N were analyzed every two days. Microbial activity was measured on the digestate after AD.
2.4. Analytical Methods
The particle size and specific surface area of the milling FW particles were determined using a laser particle sizer (Mastersizer 3000, 0.01–3000 µm, Malvern Panalytical, Stamford, Connecticut, USA). Measurement of soluble substances in micron-sized particles includes Soluble Carbon Organic Dioxide Demand (SCOD), soluble proteins (SPN), and soluble sugars (SPS). The contents of SPN and SPS were determined by the Lowry method and the anthrone sulfuric method. Processing data (TS, VS, NH4+-N) were determined by APHA 1995. The biogas concentration and composition were measured using gas chromatography (Agilent 7890B, Agilent Technologies, Santa Clara, CA, USA). The VFA composition was analyzed by gas chromatography equipped with FID (Agilent 7820, Agilent Technologies, Santa Clara, CA, USA). A sample injection volume of 1 μL was used, and nitrogen served as the carrier gas with a flow rate of 50 mL/min. The detector and injection port temperatures were maintained at 200 and 220 °C, respectively.
2.5. DNA Extraction and Metagenomic Sequencing
Specific DNA extraction methods have been described in the Supplementary Data. Sequence data associated with this project have been deposited in the NCBI Short Read Archive database (Accession Number: PRJNA1037748).
2.6. Kinetic Modeling
The efficiency of the biochemical and mass transfer processes in each AD stage affects the biogas production and composition. Kinetic modeling is a powerful tool for estimating fundamental biochemical parameters. Appropriate kinetic models can provide effective solutions for improving the selection and concentration of anaerobic digestion substrates and optimizing the performance of AD. The modified Gompertz model (Equation (1)), first-order kinetic model (Equation (2)), and logistic model (Equation (3)) were utilized to describe the relationship between microbial growth and methane production.
where P(t) is the accumulated methane production (mL CH4 g−1VS). t is anaerobic digestion time (d). Pmax is the maximum production (mL CH4 g−1VS), Rm denotes the maximum CH4 production rate (mL CH4 g−1VS d−1), λ is the lag time (d), and k is the kinetic constant in (d−1). The mean value of the experimental methane production data was used as the simulation data point.
2.7. Statistical Analysis
All the tests were conducted in triplicate. The significance of data was evaluated using the IBM SPASS Statistics 26, and p < 0.05 was significant in statistics. OriginPro 2021 was used to draw graphics and kinetic modeling.
[ad_2]
