Particle Size Effect on Biodegradability and Kinetics During Anaerobic Digestion of Fruit and Vegetable Waste

Abstract

This study examines the anaerobic digestion (AD) of fruit and vegetable waste (FVW) and digestates to assess the effect of particle size on anaerobic biodegradability (AB) and process rate at different stages of digestion progress. Batch assays were conducted with FVW mixtures and digestates from 5, 10, and 15 days of digestion, using four particle size ranges: Ø1 < 1.8 µm, 1.8 < Ø2 < 500 µm, 500 < Ø3 < 1000 µm, and Ø4 > 1000 µm. While AB and specific methanogenic activity (SMA) showed no significant differences among FVW mixtures, particle size significantly influenced these variables. Methane yields were 298.2 and 309.8 $\mathrm{m}\mathrm{L} \mathrm{C}{\mathrm{H}}_4·{\mathrm{g}}^{-1}\mathrm{V}\mathrm{S}$ for Ø4 and Ø3 particles, exceeding the 186.7 and 161.8 $\mathrm{m}\mathrm{L} \mathrm{C}{\mathrm{H}}_4·{\mathrm{g}}^{-1}\mathrm{V}\mathrm{S}$ obtained for Ø2 and Ø1, respectively. These results indicate that particle size reduction enhanced methane production; however, reducing the particle size below 500 µm inhibits methanogenesis due to intermediate accumulation. Compared to FVW, digestates exhibited a 23% reduction in AB, a 73.9% decrease in SMA, and methane yields of 55.8–294 $\mathrm{m}\mathrm{L} \mathrm{C}{\mathrm{H}}_4·{\mathrm{g}}^{-1}\mathrm{V}\mathrm{S}$. Additionally, the surface-based kinetic constant ($K_{SBK}$) decreased from 0.4523 $\mathrm{g}·\mathrm{d}{\mathrm{m}}^{-2}·{\mathrm{d}}^{-1}$for FVW to 0.0437 $\mathrm{g}·\mathrm{d}{\mathrm{m}}^{-2}·{\mathrm{d}}^{-1}$for digestates. These differences are attributed to the rapid consumption of easily biodegradable fractions within the first 5 days of digestion.

1. Introduction

Globally, one-third of all food produced is wasted annually, and for fruits and vegetables, 60% of the production is discarded [1]. Due to the putrescible nature of this waste, improper disposal results in environmental and public health issues, such as greenhouse gas emissions, the infiltration of leachates into the soil, and the proliferation of pathogens and sanitary vectors [2]. Nevertheless, anaerobic digestion (AD) constitutes a sustainable alternative to reduce the negative impacts of waste by promoting the circular economy of these materials. During this process, organic matter is metabolized, producing methane-rich biogas and a stabilized material known as digestate, which has a high nitrogen and phosphorus content. Methane can be used to cogenerate thermal or electrical energy, while digestate can be employed as a fertilizer for soils [3].

The AD process is complex, as the degradation of organic matter occurs through four stages: hydrolysis, acidogenesis, acetogenesis, and methanogenesis, in which syntrophic relationships between the different species of microorganisms involved in each stage are evident [4]. Therefore, when implementing anaerobic technologies, it is essential to establish operational conditions that ensure the products of one stage are consumed in subsequent stages without the accumulation of intermediates to prevent inhibiting phenomena and process instability, maximizing methane production and solids removal efficiencies [5]. In this regard, the first step for the design and sizing of anaerobic reactors is the analysis of variables that determine the kinetics of the process, such as pH, temperature, and concentration, but primarily the characteristics of the substrate that determine its anaerobic biodegradability (AB) [5,6].

In the case of FVW, between 56% and 75% of the volatile solids (VSs) comprise sugars and hemicellulose, resulting in an anaerobic biodegradability (AB) greater than 72% [7]. When sugars are soluble, they are rapidly fermented by acidogenic bacteria, producing volatile fatty acids (VFAs). Therefore, when the organic loading rate applied to continuous systems exceeds 3.5 $\mathrm{g}\mathrm{V}\mathrm{S}·{\mathrm{L}}^{-1}·{\mathrm{d}}^{-1}$, VFA can accumulate, lowering the pH and inhibiting the methanogenic microorganisms [8]. However, for the readily biodegradable soluble compounds, the complex polymeric structure of FVW must be broken down. Some FVWs are characterized by up to 30% lignocellulosic fibers, which are resistant to biodegradation, thus slowing hydrolysis and limiting the overall process rate [9,10,11].

A strategy to enhance the bioaccessibility of FVW is through particle size reduction. This increases the surface area available for contact between the substrate, microorganisms, and enzymes, thereby improving mass transfer and biodegradability [12,13,14,15]. This approach directly influences the methane yields. For instance, Mozhiarasi et al. [16] reported a methane yield increase from 253.9 to 322.7 $\mathrm{m}\mathrm{L} \mathrm{C}{\mathrm{H}}_4·\mathrm{g}\mathrm{V}\mathrm{S}$ when reducing the particle size of fruit waste subjected to AD, while for FVW mixtures, the observed increase was from 132 to 237 $\mathrm{m}\mathrm{L} \mathrm{C}{\mathrm{H}}_4·\mathrm{g}\mathrm{V}\mathrm{S}$ [17]. Similarly, Izumi et al. [18] demonstrated that reducing the particle size of food waste containing FVW from 0.888 mm to 0.715 mm enhanced solubilization and methane yield by 49.8% and 28%, respectively. However, further reduction to 0.391 mm resulted in a decline in methane yield due to VFA overproduction and accumulation.

This highlights the intricate relationship between the complex physical structure and chemical composition of FVW. While particle size reduction can enhance the overall AD rate and methane production by accelerating hydrolysis through an increased surface area for enzymatic or microbial activity, it may also cause the rapid release of easily biodegradable monomers. These compounds are rapidly metabolized into VFA, whose accumulation can potentially inhibit methanogenic activity.

Therefore, to adequately characterize the AD of FVW without underestimating or overestimating the process rate and methane yields, it is necessary to consider the presence of fractions with different AB in these types of waste [19,20,21,22]. To this end, this work systematically analyzed the effect of particle size on the AB, process rate, and methane production during the discontinuous AD of FVW and digestates, aiming to elucidate the changes that occur in these variables during the process to provide valuable information for the design and optimization of AD systems. The process rate was analyzed in two ways: first, by measuring the kinetics of methane production through specific methanogenic activity (SMA), and second, by calculating and comparing the surface-based kinetic constant to determine the hydrolysis rate. The results show that particle size in the AD of FVW significantly influenced the AB and the process kinetic rate.

2. Materials and Methods

2.1. Substrates

The fresh FVW comprised two mixtures, FVW1 and FVW2, collected from a cocktail establishment in Ecatepec, State of Mexico. Digestates were sourced from a laboratory-scale up-flow anaerobic sludge blanket solid-state reactor, known as RAFAELL by its Spanish acronym (Reactor Anaerobio de Flujo Ascendente Empacado con Lecho de Lodos), designed for treating FVW to produce methane [23]. Digestate samples were taken after 5, 10, and 15 days of digestion, labeled as D1, D2, and D3, respectively.

The mixtures of fresh FVW were characterized qualitatively by determining the wet weight fraction of each type of waste. For this purpose, the waste was separated by type and chopped to obtain an average particle size of 1 $\mathrm{c}\mathrm{m}$; the mass ($\mathrm{k}\mathrm{g}$) of each waste type was divided by the total mass ($\mathrm{k}\mathrm{g}$) of the mixture. For the physicochemical characterization of FVW and digestates, pH was determined according to Fernández [24]. Moisture content, total solids (TS), fixed solids (FS), and volatile solids (VS) were estimated according to standard methods [25]. Nitrogen content was determined using the Kjeldahl method according to NMX-AA-026-SCFI-2001 [26].

The chemical oxygen demand (COD) and carbohydrates were quantified using colorimetric techniques reported by Eaton et al. [25] and Goel et al. [27], respectively. These analyses were applied to a 1 $\mathrm{g}·{\mathrm{L}}^{-1}$ suspension of dried FVW. To prepare the suspension, the FVW was dried in an oven at 105 °C for 24 h; the wastes were mechanically ground and sieved to obtain particles < 500 μm in size. The particles were then resuspended in distilled water to reach a concentration of 1 $\mathrm{g}·{\mathrm{L}}^{-1}$. The C/N ratio was estimated by multiplying the COD/N ratio by 0.35 [28].

2.2. Particle Size Reduction

Once the FVW and digestate samples were obtained, particle size reduction was performed to generate samples of each type of residue with four different particle sizes: $\varnothing 1$ > 1000 μm, 500 μm < $\mathrm{\varnothing }2$ < 1000 μm, 1.8 μm < $\mathrm{\varnothing }3$ < 500 μm, and $\mathrm{\varnothing }4$ < 1.8 μm. The ranges were established so that one ($\varnothing 3$) included the particle size of 391 μm, where methanogenic inhibition due to intermediate accumulation during the AD of food waste was observed [18], along with two larger ranges ($\varnothing 1$ and $\varnothing 2$) and one smaller ($\varnothing 4$). The smallest interval was selected to analyze the AD of particles at the threshold (0.45–2 $\mathsf{\mu }\mathrm{m}$) between suspended and dissolved solids [25].

To achieve this, 200 $\mathrm{g}$ of FVW or digestate was blended with 200 $\mathrm{m}\mathrm{L}$ of distilled water for five minutes. The blended mixture was then passed through a sieve with a 1 $\mathrm{m}\mathrm{m}$ opening to separate particles larger than 1000 $\mathsf{\mu }\mathrm{m}$ ($\mathrm{\varnothing }1$). Particles smaller than 1000 $\mathsf{\mu }\mathrm{m}$ were further sieved through a 500 $\mathsf{\mu }\mathrm{m}$ sieve, thus obtaining the second particle size fraction between 500 $\mathsf{\mu }\mathrm{m}$ and 1000 $\mathsf{\mu }\mathrm{m}$ ($\varnothing 2$).

To ensure that each fraction contained only residues with specified particle sizes, the retained material after each sieving step was washed with distilled water until the COD concentration in the wash water was below 50 $\mathrm{m}\mathrm{g}·{\mathrm{L}}^{-1}$.

Particles passing through the 500 $\mathsf{\mu }\mathrm{m}$ sieve were centrifuged at 3500 rpm for 15 min. The supernatant was then filtered through a glass microfiber filter (GF/A) to ensure the particle size was below 1.8 $\mathsf{\mu }\mathrm{m}$, obtaining the fourth particle size fraction ($\mathrm{\varnothing }4$). The resulting sediment was washed with 270 mL of distilled water, yielding the third fraction of particles with sizes between 1.8 $\mathsf{\mu }\mathrm{m}$ and 500 $\mathsf{\mu }\mathrm{m}$ ($\varnothing 3$).

2.3. Inoculum

The inoculum used was granular sludge obtained from the RAFAELL reactor. Its quality was assessed by determining its specific methanogenic activity (SMA). SMA determinations were conducted in 120 mL batch reactors, where granular sludge was inoculated into 80 mL of Revised Anaerobic Mineral Medium (RAMM) containing sodium acetate as the carbon source at a concentration of 4 $\mathrm{g} \mathrm{C}\mathrm{O}\mathrm{D}·{\mathrm{L}}^{-1}$. The inoculum concentration was set to 2.5 $\mathrm{g} \mathrm{V}\mathrm{S}\mathrm{S}·{\mathrm{L}}^{-1}$ [29]. The assays were conducted over 7 days, and methane production was quantified by measuring the displaced volume of a 1.5% KOH solution in a Mariotte-type tube. All assays were performed at a controlled temperature of 35 °C by placing the reactors in an incubator. SMA was calculated using Equation (1).

$$SMA={\displaystyle \frac{R}{CF·V·C_{VSS}}}$$

(1)

where R is the slope of the cumulative methane curve over time ($\mathrm{m}{\mathrm{L}}_{\mathrm{C}\mathrm{H}4}·{\mathrm{d}}^{-1}$), CF is the conversion factor ($418\mathrm{m}{\mathrm{L}}_{\mathrm{C}\mathrm{H}4}·{\mathrm{g}}^{-1}\mathrm{C}\mathrm{O}\mathrm{D}$ at 35 °C) [30], V is the reaction volume ($\mathrm{L}$), C_VSS is the volatile suspended solids concentration ($\mathrm{g}·{\mathrm{L}}^{-1}$), and SMA is the specific methanogenic activity $\left(\mathrm{g}\mathrm{C}\mathrm{O}{\mathrm{D}}_{\mathrm{C}\mathrm{H}4}·{\mathrm{g}}^{-1}\mathrm{V}\mathrm{S}\mathrm{S}·{\mathrm{d}}^{-1}\right)$.

The inoculum was activated by successive feedings of RAMM and sodium acetate as the carbon source until no change in SMA was detected between two consecutive feedings.

2.4. Anaerobic Assays for Determination of Biodegradability and Kinetic Parameters

The anaerobic biodegradability (AB) of FVW and digestates was determined following the methodology described by Field et al. [29]. The experiment setup consisted of preparing reactors with an effective volume of 80 $\mathrm{m}\mathrm{L}$, each containing a specific type of waste (FVW1, FVW2, D1, D2, or D3) with a defined particle size ($\varnothing 1, \varnothing 2, \varnothing 3,$or $\varnothing 4$). Each treatment, characterized by the type of substrate and particle size, was replicated three times, ensuring a final concentration of 4 $\mathrm{g} \mathrm{C}\mathrm{O}\mathrm{D}·{\mathrm{L}}^{-1}$. All reactors were inoculated with preactivated anaerobic sludge to reach a concentration of 2.5 $\mathrm{g} \mathrm{V}\mathrm{S}\mathrm{S}·{\mathrm{L}}^{-1}$. The effective volume was completed with RAMM without a carbon source. The pH of the reaction medium was adjusted to 7 with KOH, and 0.5 $\mathrm{g}$ of $NaHC{O}_3$ was added per gram of COD in each reactor to enhance the buffer capacity. The reactors were sealed, and the anaerobic process was carried out at 35 °C for 7 days. Methane production was quantified as described in the previous section. At the beginning and end of each assay, the concentrations of soluble COD [25] and VFA [31] were determined.

The anaerobic biodegradability percentage was calculated using Equation (2).

$$\%AB=\%A+\%Cells$$

(2)

where $\%A$ (Equation (3)) represents the percentage of COD transformed into $C{H}_4$ and VFA, and $\%Cells$ (Equation (4)) is the percentage of COD utilized by microorganisms for growth.

$$\%A=\%M+\%VFA$$

(3)

$$\%Cells=\left[{\displaystyle \frac{\%VFA}{1-0.196}}-\%VFA\right]+\left[{\displaystyle \frac{\%M}{1-0.028}}-\%M\right]$$

(4)

where $\%M$ is calculated as $100\left(CO{D}_{CH4}/CO{D}_0\right)$, and $\%AGV$ is $100\left(CO{D}_{VFA}/CO{D}_0\right)$. Here, $CO{D}_0$ represents the initial COD, and $CO{D}_i$ corresponds to the COD of the respective compound ($C{H}_4$ or $VFA$) at the end of the assay.

To determine the mass of FVW and digestates required to be added to each assay to achieve a concentration of 4 $\mathrm{g} \mathrm{C}\mathrm{O}\mathrm{D}·{\mathrm{L}}^{-1}$, a relationship between total suspended solids (TSS) and COD was established for each sample (Equation (5)). First, the TSS content in 10 $\mathrm{g}$ of wet FVW and digestates was determined. Subsequently, the COD associated with the TSS was measured by preparing a suspension with a concentration of 1 $\mathrm{g}\mathrm{T}\mathrm{S}\mathrm{S}·{\mathrm{L}}^{-1}$ (as described in Section 2.1).

$${\displaystyle \frac{WS}{SST}}·{\displaystyle \frac{TSS}{COD}}·{\displaystyle \frac{4 g COD}{L}}·{\displaystyle \frac{V_{medium}}{1}}=solids mass$$

(5)

where WS is the mass of wet solids of a specific particle size ($\mathrm{g}$), TSS is the total suspended solids ($\mathrm{g}$), COD is the chemical oxygen demand ($\mathrm{g}·{\mathrm{L}}^{-1}$), and $V_{medium}$ is the volume of reaction used in the assay ($\mathrm{L}$).

For all experiments, the methane yield ($Y_{exp}$) expressed in $\mathrm{m}\mathrm{L}·\mathrm{g} \mathrm{V}\mathrm{S}$ and $\mathrm{m}\mathrm{L}·\mathrm{g} \mathrm{C}\mathrm{O}\mathrm{D}$ after 7 days of digestion was calculated as the ratio of the accumulated methane during the process to the initial substrate mass (in $\mathrm{g} \mathrm{V}\mathrm{S}$ and $\mathrm{g} \mathrm{C}\mathrm{O}\mathrm{D}$).

For the global methane production rate analysis, the SMA was determined for all assays, following the method described in Section 2.3. In these experiments, the carbon source was the organic matter in FVW and digestates.

2.5. Kinetic Analysis of the Hydrolysis of Substrates

Considering that the substrate surface is colonized and hydrolyzed by extracellular enzymes, the hydrolysis rate ($r_{hyd}$ in $\mathrm{g}·{\mathrm{L}}^{-1}·{\mathrm{d}}^{-1}$) can be expressed as follows:

$$r_{hyd}=-K_{SBK}·a^{*}·S_p$$

(6)

where $S_p$ is the particulate substrate concentration ($\mathrm{g}·{\mathrm{L}}^{-1}$), $K_{SBK}$ is the surface-based kinetic constant ($\mathrm{g}·{\mathrm{d}\mathrm{m}}^{-2}·{\mathrm{d}}^{-1}$), and $a^{*}$ is the specific surface area (${\mathrm{d}\mathrm{m}}^2·{\mathrm{g}}^{-1}$). Assuming spherical particles, $a^{*}$ is calculated as $a^{*}=3\rho /R$. Where $\rho$ is the substrate density ($\mathrm{g}·{\mathrm{L}}^{-3}$), and $R$ is the particle radius [32]. Thus, the model assumes that the particle surface is uniformly degraded, leading to a reduction in particle radius over time. The rate of this surface phenomenon is represented by the kinetic constant $K_{SBK}$, which indicates the amount of substrate hydrolyzed per unit area and time [33]. Since $R$ is a function of time ($t$) and the initial radius ($R_0$), it can be described as follows [33]:

$$R=R_0-{\displaystyle \frac{K_{SBK}·t}{\rho }}$$

(7)

The constant $K_{SBK}$, which depends only on the substrate nature, was determined for anaerobic assays with $\varnothing 2$ particles. To estimate $K_{SBK}$, the hydrolysis efficiency in fractional form ($\eta$) was calculated from experimental data using the following expression:

$$\eta ={\displaystyle \frac{m_0-m_t}{m_0}}$$

(8)

where $m_0$ is the initial substrate mass ($\mathrm{g} \mathrm{C}\mathrm{O}\mathrm{D}$), and $m_t$ is the substrate mass ($\mathrm{g} \mathrm{C}\mathrm{O}\mathrm{D}$) at time $t$ ($d$). The expression $m_0-m_t$ represents the total hydrolyzed substrate mass and is equivalent to the sum of accumulated methane ($\mathrm{g} \mathrm{C}\mathrm{O}{\mathrm{D}}_{\mathrm{C}\mathrm{H}4}$) and VFA ($\mathrm{g} \mathrm{C}\mathrm{O}{\mathrm{D}}_{\mathrm{V}\mathrm{F}\mathrm{A}}$) in the reaction medium.

By substituting the mass of spherical particles ($m=4\pi R^3\rho /3$) in Equation (8) and combining it with Equation (7), Equation (9) was obtained, where $K_{SBK}$ is the only unknown parameter.

$$\eta =1-{\displaystyle \frac{{\left(R_0-\frac{K_{SBK}·t}{\rho }\right)}^3}{R_0^3}}$$

(9)

To determine the substrate density ($\rho$ in $\mathrm{g}·{\mathrm{L}}^{-1}$), a 2 $\mathrm{m}\mathrm{L}$ tube was weighed ($P_1$) and then filled with substrate particles. The tube was tapped three times against a flat surface to ensure the sample occupied the entire volume. Subsequently, the tube with the packed sample was weighed ($P_2$). The density was calculated as $P_2-P_1/V$, where $V$ is the tube volume ($\mathrm{L}$).

The kinetic constant $K_{SBK}$ was estimated by solving Equation (9) using the Newton–Raphson method, implemented with the fsolve function in Matlab R2024a.

Equation (6) was coupled with Equation (10), which describes methane generation as proportional to substrate hydrolysis, to simulate methane production.

$${\displaystyle \frac{dC{H}_4}{dt}}=Y_M·r_{hyd}$$

(10)

where $Y_M$ is the product/substrate yield coefficient ($\mathrm{m}\mathrm{L} \mathrm{C}{\mathrm{H}}_4·{\mathrm{g}}^{-1}\mathrm{C}\mathrm{O}\mathrm{D}$) calculated as $Y_M=Y_{exp}/\eta$. Here, $Y_{exp}$ represents the experimental methane yield after 7 days of digestion. The division by $\eta$ accounts for the fact that methane production in the experiments originated solely from the hydrolyzed substrate.

Finally, the system of Equations (6), (7), and (10) was integrated using the ODE45 function in Matlab R2024a software, which employs the fourth-order Runge–Kutta method. The coefficient of determination ($R^2$) was calculated to assess the model’s accuracy and reliability. A model is considered reliable when $R^2$ is close to 1 [34].

2.6. Statistical Analysis

An analysis of variance (ANOVA) was conducted to evaluate the presence of statistical differences among the values of the variables analyzed in this study (SMA, AB, and $K_{SBK}$). The factors were substrate type (FVW1, FVW2, D1, D2, and D3) and the particle size.

The statistical analysis was performed using Minitab software (version 21.1.0), with a significance level ($\alpha$) set at 0.05. To identify differences among treatments, the F-statistic was compared to the critical value $F_{\alpha ,k-1,N-k}$, where $k-1$ denotes the degrees of freedom among groups or treatments ($k$), and $N-k$ corresponds to the degrees of freedom for the error term, with $N$ as the total number of observations. The F-statistics tests the null hypothesis, which assumes no significant differences among treatment means. If $F>F_{\alpha ,k-1,N-k}$, the null hypothesis is rejected, indicating significant differences among treatment means. When such differences were detected, the Tukey pairwise comparison test was applied to identify statistically equivalent means [35].

3. Results and Discussion

3.1. Inoculum and Substrate Characteristics

The inoculum quality in each reactor was characterized by an SMA of $1.23\pm 0.36 \mathrm{g} \mathrm{C}\mathrm{O}{\mathrm{D}}_{\mathrm{C}\mathrm{H}4}·{\mathrm{g}}^{-1}\mathrm{V}\mathrm{S}\mathrm{S}·{\mathrm{d}}^{-1}$. This value is comparable to the SMA of 423 $\mathrm{m}\mathrm{L} \mathrm{C}{\mathrm{H}}_4·{\mathrm{g}}^{-1}·{\mathrm{d}}^{-1}$ (equivalent to 1.031 $\mathrm{g} \mathrm{C}\mathrm{O}{\mathrm{D}}_{\mathrm{C}\mathrm{H}4}·{\mathrm{g}}^{-1}\mathrm{V}\mathrm{S}\mathrm{S}·{\mathrm{d}}^{-1}$, calculated using $CF$ at 35 °C) reported for granular sludge from a papermaking wastewater treatment plant [36]. It is also higher than the SMA of 0.639 $\mathrm{g} \mathrm{C}\mathrm{O}{\mathrm{D}}_{\mathrm{C}\mathrm{H}4}·{\mathrm{g}}^{-1}\mathrm{V}\mathrm{S}\mathrm{S}·{\mathrm{d}}^{-1}$ reported for granular sludge from an alcohol distillery wastewater treatment plant [37].

The ANOVA analysis revealed no significant differences in the initial SMA across the experiments ($\alpha =0.05$), with an F-statistic value of 0.26, which is lower than the critical value ($F_{\mathrm{0.05,3},56}=2.78$). This indicates that the variations observed in the biodegradability tests can be attributed to the analyzed factors (particle size and substrate type) rather than differences in inoculum quality.

The composition and physicochemical characteristics of the FVW and digestates are presented in

Table 1. Composition of two mixtures of FVW.

Components	Fraction in FVW1	Fraction in FVW2
Beet	0.06	0.04
Radish leaves	0.05
Jicama	0.28	0.03
Lime	0.02
Mango		0.17
Melon		0.08
Orange		0.10
Papaya	0.38	0.21
Cucumber	0.06	0.04
Pineapple		0.16
Watermelon		0.18
Carrot	0.16

and

Table 2. Physicochemical characteristics of substrates used in anaerobic assays.

		Sample
Parameter	Unit	FVW1	FVW2	D1	D2	D3
pH	-	5	6	5.1	7	7
Moisture	%	91.5	89.4	91	88	91
Total solids	$\mathrm{g}·{\mathrm{g}}_{\mathrm{s}\mathrm{a}\mathrm{m}\mathrm{p}\mathrm{l}\mathrm{e}}^{-1}$	0.085	0.11	0.093	0.118	0.088
Volatile solids	$\mathrm{g}·{\mathrm{g}}_{\mathrm{s}\mathrm{a}\mathrm{m}\mathrm{p}\mathrm{l}\mathrm{e}}^{-1}$	0.074	0.10	0.088	0.111	0.078
COD	$\mathrm{g}·{\mathrm{g}}^{-1}\mathrm{T}\mathrm{S}$	0.813	0.890	1.18	1.15	0.83
Nitrogen	$\mathrm{m}\mathrm{g}·{\mathrm{g}}^{-1}\mathrm{T}\mathrm{S}$	19	10	27	37	38
Total carbohydrates	$\mathrm{g}·{\mathrm{g}}^{-1}\mathrm{T}\mathrm{S}$	0.435	0.627	0.231	0.520	0.280
C/N	-	14.96	31.15	15.29	10.88	7.64

- Dimensionless.

. These substrates are suitable for AD due to their high moisture content (>88%) and organic matter content (VS > 87% TS), consistent with previous studies [13]. The pH of the FVW mixtures is slightly acidic, which is typical for this type of waste. In contrast, digestates D2 and D3 exhibit neutral pH values attributed to the degradation of VFA and ammonia production after 10 and 15 days of digestion, respectively. Digestate D1 shows a pH of 5.1, indicating incomplete degradation of acidic intermediates after only 5 days of processing.

The nitrogen content is higher in the digestates due to decarbonization during methane production, with D3 showing the highest nitrogen content, corresponding to its longer digestion time (Table 2). This results in a lower C/N ratio in the digestates (7.64–15.29), comparable to the value of 12.8 reported for digestates derived from tomato residues and corn stover [38]. Conversely, the FVW mixtures exhibit differences in their C/N ratios: 14.96 for FVW1 and 31.15 for FVW2. This variation reflects their composition, as FVW1 contains more vegetables, while FVW2 has a higher proportion of fruits. The C/N ratio of FVW1 is similar to the 15.4 value reported for tomato residues [38]. This parameter is essential, as low C/N ratios may lead to ammonia accumulation and subsequent inhibition during AD [39].

3.2. Anaerobic Biodegradability of Fresh Fruit and Vegetable Waste

Figure 1 presents the AB and SMA results obtained for the FVW1 and FVW2 mixtures. AB is expressed as %AB since, as indicated in Equation (2), this parameter represents the percentage of the initial substrate converted into VFA, methane, and biomass. The ANOVA for both variables of the two mixtures revealed no significant differences between the means, as the F-statistic value was lower than the critical value in all cases (

Table 3. F-statistics values for both treatments (FVW1 and FVW2), with a critical $F_{0.5, 1, 3}=10.13.$

Particle Size	FSMA	FAB
Ø1	3.93	0.71
Ø2	0.51	-
Ø3	2.26	0.01
Ø4	6.25	5.54

). Consequently, it can be inferred that the composition of the mixtures did not influence AB and SMA and, therefore, would not affect the characteristics and AB of digestates.

The AB was above 64%, and this result agrees with the reports of other authors who classify FVW as easily biodegradable substrates due to their high sugar content, with AB values exceeding 72% [7,40]. However, it has been reported that certain FVW, such as olive mill waste or fruit waste like mangosteen, are resistant to biological degradation due to their content of phenolic, flavor, and non-biodegradable compounds [41,42]. Therefore, the FVW must always be evaluated before selecting the treatment technology.

Another ANOVA revealed that particle size has a significant effect on AB and SMA, with F-statistics of 34.03 and 17.68, respectively, both exceeding the critical value $F_{0.05, \mathrm{3,16}}=3.24$. Tukey’s test (

Table 4. Pairwise comparisons using Tukey’s test, with a 95% confidence level, for SMA at different particle sizes.

Particle Size	N	Mean $\left(\mathbf{g}\mathbf{C}\mathbf{O}{\mathbf{D}}_{\mathbf{C}\mathbf{H}4}·{\mathbf{g}}^{-1}\mathbf{V}\mathbf{S}\mathbf{S}·{\mathbf{d}}^{-1}\right)$	Grouping
Ø4	5	1.1480	A
Ø1	5	0.8280	B
Ø2	5	0.7820	B
Ø3	5	0.3780	C

Means that do not share a letter are significantly different.

) provides statistical evidence that the means of particle sizes $\varnothing 1$ and $\varnothing 2$ are the only ones that do not show a significant difference. The maximum mean of SMA was obtained for $\varnothing 4$, which can be explained by the fact that particles smaller than 1.8 $\mathsf{\mu }\mathrm{m}$ are more bioavailable, and the fermentation to VFA occurs faster than for larger particles. In contrast, for AB (

Table 5. Pairwise comparisons using Tukey’s test, with a 95% confidence level, for AB at different particle sizes.

Particle Size	N	Mean (%)	Grouping
Ø2	5	100	A
Ø1	5	96.20	A
Ø4	5	90.60	A
Ø3	5	64.40	B

Cases that do not share a letter are significantly different.

), the only statistically different mean and the lowest SMA was that corresponding to particle size $\varnothing 3$, while the others exhibit means close to 100%. The observation that the lowest SMA (0.3780 $\mathrm{g} \mathrm{C}\mathrm{O}{\mathrm{D}}_{\mathrm{C}\mathrm{H}4}·{\mathrm{g}}^{-1}\mathrm{V}\mathrm{S}\mathrm{S}·{\mathrm{d}}^{-1}$) was found for $\varnothing 3$ ($1.8-500 \mathrm{\mu }\mathrm{m}$) can be interpreted by considering that AB was the lowest (64.40%) for this particle size range, indicating that the percentage of substrate converted into VFA and methane was lower compared to the other experiments. In this context, it can be inferred that the low methane production rate, indicated by SMA, results from the limitation of direct substrate availability for methanogenesis. Thus, the reduction in particle size below 500 $\mu \mathrm{m}$ may lead to the accumulation of VFA or other intermediates, which could inhibit the hydrolytic stage of AD.

3.3. Anaerobic Biodegradability of Digestates

Figure 2 shows the SMA and AB values determined for D1, D2, and D3 digestates. Biodegradability ranges from 15 to 40% for digestates with particle size $>1.8 \mathsf{\mu }\mathrm{m}$ ($\varnothing 1$, $\varnothing 2$, and $\varnothing 3$), indicating a 23% reduction in AB compared with FVW. This can be attributed to the rapid consumption of easily biodegradable carbohydrates during the initial five days of digestion [10]. As a result, the digestates, mainly composed of fibers, have an AB below 61%, which is comparable to that of lignocellulose-rich herbaceous plants [43]. However, for the smaller particles ($Ø4<1.8 \mathsf{\mu }\mathrm{m}$), the AB ranged between 48% and 85%, as the particle size reduction enhances the substrate bioavailability.

Nevertheless, despite the increment in AB, another implication of the digestate composition and its difficult biodegradation, as shown in Figure 2b, is that the highest SMA value (0.35 $\mathrm{g} \mathrm{C}\mathrm{O}{\mathrm{D}}_{\mathrm{C}\mathrm{H}4}·{\mathrm{g}}^{-1}\mathrm{V}\mathrm{S}\mathrm{S}·{\mathrm{d}}^{-1}$), obtained for digestates with particle size $\varnothing 4$, corresponds to only 26% of the maximum value (1.345 $\mathrm{g} \mathrm{C}\mathrm{O}{\mathrm{D}}_{\mathrm{C}\mathrm{H}4}·{\mathrm{g}}^{-1}\mathrm{V}\mathrm{S}\mathrm{S}·{\mathrm{d}}^{-1}$) recorded for the AD of fresh FVW.

Table 6. F-statistics values for treatments (D1, D2, and D3), with a critical $F_{0.5, 2, 5}=5.79.$

Particle Size	FSMA	FAB
Ø1	20.49	1.92
Ø2	0.25	15.45
Ø3	5.20	1.35
Ø4	4.98	9.19

presents the statistical analysis of the means for both SMA and AB assays. Significant differences were observed between the means for the different digestates. However, this behavior is not consistent across all particle sizes. In the case of SMA, significant differences were found only among the digestates with particle size $\varnothing 1$. In contrast, for AB, significant differences were observed for particle sizes $\varnothing 1$ and $\varnothing 4$.

3.4. Methane Yield

Since Section 3.2 demonstrated that no significant differences were observed in SMA and AB between the FVW1 and FVW2 mixtures, this section focuses on presenting information regarding the FVW1 mixtures and all the digestates.

Figure 3a shows the methane yields obtained in all the experiments. The methane yield for the AD of FVW1 was in the range of 161.8 to 309.8 $\mathrm{m}\mathrm{L} \mathrm{C}{\mathrm{H}}_4·{\mathrm{g}}^{-1}\mathrm{V}\mathrm{S}$. These results are close to the reported values for the AD of FVW in previous studies (207–523 $\mathrm{m}\mathrm{L} \mathrm{C}{\mathrm{H}}_4·{\mathrm{g}}^{-1}\mathrm{V}\mathrm{S}$) [44,45].

Methane yields obtained for FVW samples with smaller particle sizes ($\varnothing 3$ and $\varnothing 4$) were the lowest. The reason for this result is evidenced in Figure 3b,c, where it can be observed that at the end of the digestion process of FVW with $\varnothing 4$ particles ($<1.8 \mathsf{\mu }\mathrm{m}$), the medium was acidified, with pH values of 5.4 and 5.2. This acidification resulted from VFA accumulation, whose concentration reached 3700 $\mathrm{m}\mathrm{g}·{\mathrm{L}}^{-1}$. This value exceeds the concentration of 2837.8 $\mathrm{m}\mathrm{g}·{\mathrm{L}}^{-1}$ reported to inhibit methanogenesis even at a pH of 6.53 [46]. Furthermore, Figure 4a shows that during the first digestion day, the slope of the accumulated methane curve for the AD of $\varnothing 4$ particles was the steepest among the different particle sizes, followed by a drastic decrease leading to a nearly horizontal curve. This suggests that the soluble $\varnothing 4$ particles were rapidly metabolized, resulting in a higher methane production rate, which was numerically expressed as the higher SMA (Section 3.2); however, the generated VFA inhibited the process, decreasing the methane yield. It agrees with the findings of Izumi et al. [18], who observed excessive particle size reduction (below 400 $\mathsf{\mu }\mathrm{m}$) of food waste containing 78.4% FVW decreased methane production due to VFA accumulation.

For the experiments with $\varnothing 3$ particles (1.8–500 $\mathsf{\mu }\mathrm{m}$), the VFA concentration was 1012 $\mathrm{m}\mathrm{g}·{\mathrm{L}}^{-1}$ with pH values of 7.2–7.4, conditions that are not typical for methanogenesis inhibition caused by VFA accumulation [46]. Moreover, its methane accumulated curve was characterized by a low slope but consistently increased throughout the digestion period (Figure 4a), which can be interpreted as a slow methane production rate rather than an interruption of the methanogenic process. This trend was quantitatively reflected as the lowest SMA (0.3780 $\mathrm{g} \mathrm{C}\mathrm{O}{\mathrm{D}}_{\mathrm{C}\mathrm{H}4}·{\mathrm{g}}^{-1}\mathrm{V}\mathrm{S}\mathrm{S}·{\mathrm{d}}^{-1}$). Therefore, this methane production pattern provides additional evidence, along with that presented in Section 3.2, that a slow hydrolysis process limited methane production. In this regard, previous works indicate that this limitation could be due to the accumulation of ammonia, VFA, or flavor compounds present in fruit waste [47,48].

Regarding the digestates, methane yields ranged between 55.8 and 294 $\mathrm{m}\mathrm{L} \mathrm{C}{\mathrm{H}}_4·{\mathrm{g}}^{-1}\mathrm{V}\mathrm{S}$, which are comparable to those obtained after 40 days of AD of fiber-rich substrates (180–370 $\mathrm{m}\mathrm{L} \mathrm{C}{\mathrm{H}}_4·{\mathrm{g}}^{-1}\mathrm{V}\mathrm{S}$) [43]. The lower yields observed are a consequence of the low AB and the insufficient digestion time (7 days). Considering the low slope of methane accumulated curves (Figure 4b–d), indicating low SMA, along with VFA concentration below 1000 $\mathrm{m}\mathrm{g}·{\mathrm{L}}^{-1}$ and pH values above 7.2 (Figure 3), it is evident that the methane production rate was limited by the hydrolysis of the substrates. This has been generalized for particulate substrates [4], especially for lignocellulosic matter with low biodegradability [38]. This idea is also supported by the fact that when comparing the results of the AD of digestates with the same particle size, the lowest methane yields were observed when using digestates collected after 15 days of digestion as substrates (Figure 3).

3.5. Surface-Based Kinetics

In

Table 7. Parameters of kinetic analysis of hydrolysis of FVW at different anaerobic digestion times.

Substrate	Digestion Time	Hydrolysis Efficiency $\mathit{\eta }$	${\mathit{K}}_{\mathit{S}\mathit{B}\mathit{K}}$ Mean $\left(\mathbf{g}·\mathbf{d}{\mathbf{m}}^{-2}·{\mathbf{d}}^{-1}\right)$	Grouping *
FVW1	0	0.9574 ± 0.04	0.4523 ± 0.143	A
D1	5	0.2373 ± 0.02	0.0427 ± 0.003	B
D2	10	0.1556 ± 0.02	0.0310 ± 0.004	B
D3	15	0.2752 ± 0.02	0.0574 ± 0.004	B

* Grouping of $K_{SBK}$ as a result of the Tukey test with 0.05 significance.

, the results of the evaluation of the hydrolysis rate through the determination of the surface-based kinetic constant ($K_{SBK}$) for FVW1, D1, D2, and D3 are presented. The ANOVA evidenced a significant effect of substrate type on the $K_{SBK}$ constant, with an F-statistic of 21.07, which is greater than the critical value $F_{\mathrm{0.05,3},7}=4.77$. However, a significant difference was found only for FVW1; thus, the hydrolysis rate of digestates can be considered equivalent.

By comparing the means of $K_{SBK}$, it is evident that the hydrolysis rate decreased tenfold after 5 days of AD, dropping from 0.4523 $\mathrm{g}·\mathrm{d}{\mathrm{m}}^{-2}·{\mathrm{d}}^{-1}$ for fresh FVW (0 days of AD) to an average of 0.0437 for the digestates. This behavior results from changes in the waste composition: at the beginning of the process, the FVW contains a high content of easily biodegradable matter, such as sugars. By the end, only fibers with low biodegradability remain. For this reason, the hydrolysis efficiency decreased from 0.9574 to 0.2752.

The calculated values of $K_{SBK}$ were able to simulate the AD of the different substrates, as shown in Figure 5. The $R^2$ values were close to 1, indicating that the model provided an adequate fit to the experimental data [34]. The magnitude of $K_{SBK}$ depends on the substrate’s nature, and the values obtained (Table 7) are consistent with those reported in the literature. In the case of FVW, its value (0.4523 $\mathrm{g}·\mathrm{d}{\mathrm{m}}^{-2}·{\mathrm{d}}^{-1}$) is in the same order of magnitude as the value reported for the hydrolysis of potato starch (0.108 $\mathrm{g}·\mathrm{d}{\mathrm{m}}^{-2}·{\mathrm{d}}^{-1}$) [33], but it is two orders of magnitude greater than the value obtained during the hydrolysis of wasted activated sludge (0.0028–0.0047 $\mathrm{g}·\mathrm{d}{\mathrm{m}}^{-2}·{\mathrm{d}}^{-1}$) [49], which is more difficult to degrade due to its content of exopolymers and cell wall material. Regarding the $K_{SBK}$ of digestates, its values are lower than those mentioned for potato starch but still higher than those mentioned for sludge.

Therefore, the determined values of $K_{SBK}$ are viable for use in simulation works, with both physical and biochemical significance. However, the model proposed in this study may overestimate methane production, as it assumes the direct conversion of the substrate into methane while neglecting the intermediate stages of the AD process. For instance, the model was not effective in globally describing the AD of $Ø3$ and $Ø4$ particles, as inhibition effects were observed, slowing or halting methane production. For a better AD process description, it is recommended to integrate hydrolysis as analyzed in this work, but in models that consider microorganism growth and inhibition phenomena, as this would allow a more comprehensive description of the AD process [50].

4. Conclusions

This paper analyzed the AB, SMA, and hydrolysis kinetics of fresh FVW and its digestates to determine and understand how these parameters change during the AD of FVW at different particle sizes. The results showed that FVW1 and FVW2 mixtures did not exhibit significant differences in AB and SMA, indicating that they may yield comparable results regardless of their composition. However, significant differences were observed in AB and SMA when FVW was digested at different particle sizes, with methane yields ranging from 161.8 to 309.8 $\mathrm{m}\mathrm{L} \mathrm{C}{\mathrm{H}}_4·{\mathrm{g}}^{-1}\mathrm{V}\mathrm{S}$. It is important to highlight that reducing the particle size below 500 $\mathsf{\mu }\mathrm{m}$ led to the inhibition of AD.

During AD, higher AD and SMA values were achieved for FVW, where easily biodegradable fractions were consumed within the first five days. AB for digestates obtained after 5, 10, and 15 days of digestion shows a 23% decrease compared to fresh FVW. The maximum SMA observed for digestates was only 26% of the maximum SMA of FVW. Furthermore, changes in substrate composition significantly reduced the hydrolysis rate, with the $K_{SBK}$ value for digestates decreasing by an order of magnitude compared to fresh FVW. These findings highlight the critical role of particle size and substrate composition in optimizing the AD of FVW and suggest that particle size reduction should be controlled to avoid inhibitory effects. However, it is also important to determine whether, in practical applications, the increase in methane production justifies the costs associated with particle size reduction. Therefore, future research could focus on conducting material and energy balances to assess the technical and economic feasibility of the process, as well as exploring and evaluating the effects of other pretreatment methods, such as ultrasound or enzymatic treatment, on particle size and methane production efficiency.