Impact of Nanoparticle Addition and Ozone Pre-Treatment on Mesophilic Methanogenesis in Temperature-Phased Anaerobic Digestion

Abstract

Biodegradable organic waste offers significant opportunities for resource recovery within the frame of the circular economy. In this work, the effects of carbon-encapsulated iron nanoparticles and ozone pre-treatments in the mesophilic methanogenic stage of a temperature-phased an-aerobic digestion have been studied using biochemical methanogenic potential (BMP) tests and modeling simulation. To do that, digestates from a pre-treated thermophilic acidogenic reactor that co-digested sludge and wine vinasse were used. The addition of nanoparticles favored the removal of particulate matter, which increased by 9% and 6% in terms of total solids and volatile solids, respectively. When combined with ozone pre-treatment, these increases were 27% and 24%, respectively, demonstrating enhanced AD efficiency. The dose of iron nanoparticles encapsulated in carbon did not result in a statistically significant increase in methane production when sludge and vinasse were used as feedstock. The combination of nanoparticles with the ozone pre-treatment significantly improved the methanogenic phase of the second stage, increasing the methane production yield by 22% and reducing the lag phase from 10 days to 3 days, according to the modified Gompertz model.

1. Introduction

The abundance of wastes rich in biodegradable organic matter presents significant opportunities for resource recovery. A promising approach to valorize such wastes is through the biorefinery concept, where biomass is converted into a wide array of valuable products, including fuels, energy and chemicals. Among these processes, anaerobic digestion (AD) breaks down organic materials and produces biogas and nutrient-rich digestate useful in agriculture [1]. While mesophilic single-stage AD is the most widely implemented process [2], alternative configurations, such as Temperature-Phased Anaerobic Digestion (TPAD), have emerged to enhance the reduction of volatile solids (VS) and increase gas production compared to mesophilic single-stage anaerobic digestion [3]. TPAD involves a two-stage process: an initial thermophilic acidogenic stage (55 °C) where hydrogen is generated, followed by a mesophilic methanogenic stage (35 °C) where methane is generated. This configuration facilitates the recovery of hydrogen and volatile fatty acids (VFAs) during the thermophilic acidogenic stage.

Usually, the efficiency of conventional AD is often constrained by the nature of the substrate, particularly in the case of complex and recalcitrant wastes, such as sewage sludge [4]. To address these inconveniences, co-digestion strategies can be employed to enhance the overall digestion process. An effective approach is the co-digestion of sludge with wine vinasse. Similarly, the co-digestion of two-phase olive mill waste with cattle manure has shown promising results. By combining different wastes, the C/N ratio (20 to 30) is optimized for the development of anaerobic processes [5,6]. In addition, pre-treatment methods can be used to improve substrate solubilization and also to increase the content of biodegradable organic matter. Various physical, chemical and biological pre-treatment methods have been explored in the literature, including those using acid, alkali, Fenton reagents, calcium peroxide, H₂O₂ and ozone [7].

Acidic conditions promote the dissolution of proteins and polysaccharides into extracellular polymers and microbial cells, with effectiveness increasing as the pH decreases. In contrast, alkaline conditions facilitate optimal sludge solubilization, typically occurring in the pH range of 11 to 12.5, resulting in significant increases in soluble protein, polysaccharide and soluble chemical oxygen demand (SCOD). Pre-treatments based on H₂O₂-driven advanced oxidation processes have been shown to significantly improve methane yield and VS removal in AD. Notably, the Fenton agent, consisting of H₂O₂ and an Fe²⁺ catalyst, generates abundant hydroxyl radicals (-OH) under acidic conditions, resulting in a higher redox potential compared to H₂O₂ or ozone treatments alone [7].

During recent years, ozone pre-treatment has gained significant attention due to its high oxidative potential, which effectively degrades organic matter [8], also making it a valuable disinfectant in various fields [9]. Ozone enhances the hydrolysis of solids due to the increased surface area relative to particle volume, reducing the molecular weight of organic compounds [10]. This hydrolytic process improves the efficiency of the subsequent biogas production. For example, a 1.31-fold increase in methane production from municipal waste sludge and a 2.25-fold increase from pharmaceutical waste sludge with ozone pre-treatment were observed in Biochemical Methane Potential (BMP) tests [11]. Similarly, an increase of up to 52% in methane production from anaerobically digested sludge has been observed in BMP tests [12].

In parallel, the use of additives, including micro- and macro-nutrients, nanomaterials and biological additives, has shown promising results in enhancing microbial activity and electron transfer in AD processes, leading to improved methane production [13,14]. In particular, nanoparticles offer a high surface-to-volume ratio, enabling them to interact with many species during AD. This interaction results in higher reaction rates and electron transfer throughout the process, as well as the formation of multi-species biofilm communities within the reactor [13]. Among nanoparticles, iron nanoparticles play an important role. Iron is an essential nutrient for microorganisms, and Fe⁰ (zero-valent iron) acts as an electron donor for the AD system, promoting sludge hydrolysis–acidogenesis [14]. This enhancement of AD results in improved methane production. Numerous studies support this claim, as evidenced by BMP tests performed with sludge [13,15].

However, no studies have evaluated the effect of nanoparticles and their combination with ozone in the mesophilic methanogenic stage of the TPAD process.

In this context, the present study aims to evaluate the combined effects of carbon-encapsulated iron nanoparticles and ozone pre-treatment on BMP during the second phase of a TPAD process. Specifically, the goal is to enhance methane production and generate higher-quality digestate for agronomic use. To do that, the study utilizes digestate from the co-digestion of sludge and wine vinasse in a semi-continuous thermophilic acidogenic reactor, representing the first phase of TPAD. In addition, two kinetic models (Gompertz and Cone) are proposed as tools to describe the methanogenesis process, offering insights into the underlying dynamics of methane production.

2. Materials and Methods

In order to evaluate the effect of the nanoparticles, as well as the nanoparticles in combination with ozone pre-treatment, on the mesophilic methanogenic stage, BMP tests were carried out. The substrates used in these BMP tests were the digestate from a thermophilic acidogenic reactor—the first stage of TPAD configuration—operated under the following different conditions: without pre-treatment (C), with the addition of nanoparticles (N), and pre-treated with ozone and the addition of nanoparticles (NO).

2.1. Innoculum and Substrate Characterization

The inoculum used in this work was taken from a single-stage anaerobic digester containing sludge from a wastewater treatment plant (WWTP). This digester had a capacity of 5 L and was operated at a constant 35 °C, with a hydraulic retention time of 20 days, a pH of 7.5 and continuous stirring at 40 rpm. As previously stated, the substrates used in the BMP tests were the digestates obtained from a thermophilic acidogenic reactor operating under different pre-treatment conditions. In

Table 1. Characteristics of inoculum (I) and substrates used: effluent from reactor without pre-treatment (C), effluent from reactor with nanoparticles (N), and effluent from reactor with nanoparticles and pre-treated with ozone (NO).

Parameter	I Mean (SD 1 %)	C Mean (SD 1 %)	N Mean (SD 1 %)	NO Mean (SD 1 %)
pH	8.33 (0.22)	5.12 (0.78)	5.12 (0.97)	5.07 (1.18)
TS (g/kg)	24.05 (0.27)	19.50 (11.98)	26.50 (12.72)	22.16 (5.62)
VS (g/kg)	15.41 (0.39)	14.52 (14.49)	18.51 (0.74)	17.30 (6.77)
TCOD (mgO2/L)	32,263.65 (4.12)	51,980.32 (2.99)	51,183.69 (5.70)	46,045.10 (2.10)
SCOD (mgO2/L)	19,119.20 (12.55)	32,661.97 (15.13)	35,569.68(6.53)	36,047.66 (3.93)
DOC (mg/L)	1062.00 (0.90)	7575.03 (1.65)	9140.98 (2.66)	7457.73 (1.06)
TN (mg/L)	2332.67 (2.99)	869.00 (0.57)	1054.75 (1.71)	899.75 (0.59)
VFA (g equiv acetic/L)	0.34 (10.92)	5.54 (0.56)	5.84 (0.14)	5.62 (3.42)
Acetic acid (mg/L)	181.87 (40.88)	1333.47 (2.06)	2540.29 (1.04)	2317.66 (15.42)
Propionic acid (mg/L)	101.94 (53.56)	796.61 (3.89)	1069.07 (1.86)	844.48 (12.75)
Butyric acid (mg/L)	103.80 (31.68)	4763.22 (1.98)	3159.08 (2.46)	3407.94 (12.33)

¹ SD: standard deviation (%).

, the characterization of the inoculum and substrates used in the BMP test is presented. As can be seen in Table 1, the characteristics of the inoculum were in line with those reported in the literature using the same inoculum [16,17].

In relation to the substrates used in the BMP tests (see Table 1), it was evident that total solids (TS) and VS presented higher concentrations in the N and NO substrates due to the pre-treatments carried out. The N and NO substrates showed lower levels of total chemical oxygen demand (TCOD) but higher levels of SCOD compared to the substrate obtained without pre-treatment. This can be explained by the nanoparticle and ozone pre-treatments that were carried out. This effect can be attributed to the hydrolysis caused by the ozone [18] and due to the multiple mechanisms taking place in the surface of the nanoparticles that were dosed [13], which promote particle solubilization and improve hydrolysis.

In addition, significant improvements in dissolved organic carbon (DOC) levels were observed with the addition of nanoparticles (N), possibly because these nanoparticles released carbon during the acidogenic process (first stage of TPAD).

Regarding total acidity, no significant differences in the total acidity of the effluents were observed. However, a slightly higher concentration was observed in the effluent of the reactor operated with nanoparticles, especially for acetic and propionic acids.

2.2. Experimental Setup

In a previous stage of this study, the effect of nanoparticles and ozone pre-treatment on the thermophilic acidogenic stage (first stage of TPAD configuration) of the TPAD process was evaluated using a Continuous Stirred Tank Reactor (CSTR) (Scharlab, Barcelona) operating in an acidogenic thermophilic range (55 °C). This reactor, with a total volume of 4 L, maintained a hydraulic retention time (HRT) of 6 days. The reactor was fed with a mixture of biosolids and vinasses (BV) in a ratio of 50:50 (v/v) and subjected to pre-treatment with nanoparticles and nanoparticles combined with ozone. The effluents obtained from these experiments were used as substrates in the BMP tests carried out to mimic mesophilic methanogenic conditions (35 °C) corresponding to the second stage of TPAD configuration. The BMP tests were conducted in small, amber-colored reactors with a total volume of 250 mL and an active volume of 120 mL, operated in batch mode. Each reactor received a 50/50 (v/v) addition of inoculum and the substrate to be tested (60 mL each). In addition, a control test was performed using 60 mL of inoculum mixed with 60 mL of distilled water. The tests were carried out at pH 7.5 and in the mesophilic range (35 °C) by placing the reactors in a thermostatic stirred bath stirred at 40 rpm.

In Figure 1, a diagram of the experimental setup used in this study is presented.

In order to ensure the reproducibility of the experiments, each test was conducted in triplicate. For monitoring purposes, analyses included total chemical oxygen demand (TCOD), soluble chemical oxygen demand (SCOD), dissolved organic carbon (DOC), total nitrogen (TN), total solids (TS), volatile solids (VS), pH, total volatile fatty acids (VFAs) and individual acid concentrations. Analyses were performed both at the beginning and the end of the tests. Daily, the volume of biogas production (based on headspace pressure in the vial) and its composition were measured. The test length was defined by the evolution of methane production. The experiments were supposed to have finished once the increase in production was lower than 1% in two consecutive measurements [19].

2.3. Nanoparticles

The nanoparticles used in this work were composed of an internal nucleus of iron encapsulated in carbon material obtained by means of hydrothermal carbonization (HTC) of olive mill wastes. The encapsulation was selected to promote circular economy principles. The nanoparticles present a surface area of 190 m²/g and a size of approximately 150 ± 50 nm. Their typical composition is 2.5% Fe⁰, 42.0% total iron and 55.5% carbon. These nanoparticles are commercially available from Smallops (Don Benito, Badajoz, Spain) [20]. The optimum dosage recommended by the manufacturers, especially for AD applications, is 2 g/kgVS.

2.4. Ozonation

For ozone pre-treatment, the biosolids were ozonated at a dose of 0.018 g O₃/g TS₀ before being mixed with vinasse. This operational approach, including the specified dosage, was fine-tuned in previous research [21] with the aim of improving the performance of thermophilic dark fermentation and with a special focus on hydrogen (H₂) and VFA production. The ozonation was conducted at the Ozonation Technology Pilot Plant located at the Institute of Viticulture and Agro-Food Research (IVAGRO) of the University of Cádiz. The ozone generation process began with the pre-concentration of atmospheric oxygen using a generator (GZ20 PROY 1536, ZonoSistem, Cádiz, Spain). Subsequently, the generated ozone gas was mixed with the substrate in a column reactor by means of a porous diffuser located at the bottom of the column.

2.5. Analytical Methods

The following parameters were analyzed: pH, TS, VS, TCOD, SCOD, VFA, DOC, TOC and TN. All the analyses were performed according to the standard methods described by the APHA-AWWA-WPFC [22].

pH measurements were conducted using a Crison pH meter (Alella, Barcelona, Spain). TS and VS were determined via gravimetric technique, utilizing a self-calibrating balance. TCOD and SCOD were evaluated using the colorimetric method. For the analysis of individual VFAs, such as acetic, propionic, butyric, isobutyric, valeric, isovaleric, isocaproic, caproic and heptanoic acids, as well as the total content of total volatile fatty acids measured in mgAcHequivalent/L, gas chromatography was employed. This analysis employed a gas chromatograph (Shimadzu GC-2010, Kyoto, Japan) equipped with a flame ionization detector (FID) system and a Nukol-packed capillary column (30 m in length, 25 mm inner diameter and 25 µm film thickness). Hydrogen was employed as the carrier gas at a flow rate of 42.1 mL/min and a pressure of 75.5 kPa, yielding a linear velocity of 45 cm/s and a flow rate of 1.43 mL/min in the column. Nitrogen was used as carrier gas. The FID flame was generated with synthetic air at 400 mL/min and 50 kPa pressure, along with hydrogen at 40 mL/min and 60 kPa. The oven temperature program began at 115 °C for 0.5 min, followed by a ramp of 30 °C/min to 150 °C, then a ramp of 15 °C/min to 180 °C, and held at 180 °C for 4 min. Phenol was added as an internal standard at a perfectly known concentration (500–600 mg/L) [22,23]. The determination of TOC and TN was carried out using a total organic carbon analyzer (Shimadzu TOC-L CSH/CSN, Kyoto, Japan) in accordance with the APHA-AWWA-WPFC standard methods [22].

Throughout the BMP tests, daily gas pressure measurements were measured by means of a pressure gauge (XM 5007 I.S., Samson, Frankfurt, Germany) and used to calculate the volume of biogas generated, utilizing the ideal gas formula. The composition of the biogas was analyzed by means of the BIOGAS 5000 Geotech infrared biogas analyzer (LANDTEC North America, Colton, CA, USA), which provided measurements in percentage for CH₄, CO₂ and O₂ composition with an error range of 0.5–1.5% and in ppm for CO, H₂ and H₂S with a measurement error of 1.5–2%. The equipment is calibrated once a year by the supply company (Fonotest, Madrid, Spain) and periodically verified with a certified chromatographic gas standard.

2.6. Kinetic Analysis

Once the CH₄ production was determined, the main kinetic and stoichiometric parameters were determined by fitting the modified Gompertz and Cone models to the data set obtained. In the literature, it has been stated that the modified Gompertz Model is the most commonly used [24,25,26]. The modified Gompertz Model accurately predicts both the lag time and maximum methane production. The Cone model accurately predicts the value of the hydrolysis rate.

The Gompertz model follows Equation (1), while the Cone model follows Equation (2).

$$\mathrm{H}=\mathrm{P} \exp \left\{-\exp \left[{\displaystyle \frac{\mathrm{Rm}·\mathrm{e}}{\mathrm{P}}}\left(\mathrm{ʎ}-\mathrm{t}\right)+1\right]\right\}$$

(1)

$$\mathrm{H}={\displaystyle \frac{\mathrm{P}}{1+{\left({\mathrm{K}}_{\mathrm{hyd}}·\mathrm{t}\right)}^{-\mathrm{n}}}}$$

(2)

where H is the cumulative volume of CH₄ produced (mL), P is the CH₄ production potential (mL), R_m is the maximum CH₄ production rate (mL/h), l is the lag phase time (h), t is the incubation time (h), e is 2.718281828, K_hyd is the hydrolysis rate constant (1/h) and n is the form factor.

The errors associated with the model predictions were calculated by subtracting the experimental methane production data from the methane production estimated by the model. The root mean square error (RMSE) was calculated using Equation (3).

$$\mathrm{RMSE}=\sqrt{{\displaystyle \frac{1}{\mathrm{n}}}{{\displaystyle \sum }}_{\mathrm{i}=1}^{\mathrm{n}}{({\mathrm{y}}_{\mathrm{i}}-\widehat{{\mathrm{y}}_{\mathrm{i} }})}^2}$$

(3)

where n is the total number of observations, y_i is the experimental value of the i-th observation and ŷ_i is the predicted value for the i-th observation.

In order to determine whether there were significant differences in methane production when operating with the different strategies, a statistical analysis was carried out. The statistical tests assumed that the difference between the means of the different conditions was zero. Therefore, if the calculated p value was below the significance level (alpha), the null hypothesis was rejected, indicating statistically significant differences between the different operating conditions. These tests were performed at a 95% confidence level (alpha = 0.05). The Student’s t-test was used when the data followed a normal distribution. The Mann–Whitney–Wilcoxon U-test was used when the data did not meet the normality assumption or when equal variance could not be assumed.

In addition, the magnitude of the results was assessed using Pearson’s linear correlation coefficient “r”. This coefficient indicates the strength and direction of the relationship between variables and ranges from −1 (negative relationship) to 1 (positive relationship), with 0 indicating no linear relationship. Statistical analysis was carried out using R software (v.4.4.2) [27].

3. Results and Discussion

To investigate the influence of iron nanoparticles and their combination with ozone pre-treatment, physical and chemical parameters were evaluated at the beginning and end of the study. Particular attention was paid to the possible influence on VFAs removal and methane production.

3.1. Initial and Final Characterization of the Mixtures Tested in the Biochemical Methane Potential

The mixtures fed to the BMP tests, as well as their digestates, were characterized and compared with the characteristics of the substrates used (see Table 1).

At the beginning of the BMP, the mixture of the inoculum (with a pH of 8.33) and the effluents (with a pH around 5.5 due to their origin from an initial TPAD stage in the acidogenic reactor) resulted in a pH close to 7.5, which is optimal for BMP tests. The mixing of the inoculum with the different substrates resulted in mixtures with the characteristics presented in

Table 2. Characterization of the initial mixtures of Biochemical Methane Potential. Control (MC), nanoparticles (MN), and nanoparticles together with ozone pre-treatment (MNO).

Parameter	MC Mean (SD 1 %)	MN Mean (SD 1 %)	MNO Mean (SD 1 %)
pH	7.25 (0.80)	7.30 (0.90)	6.87 (1.25)
TS (g/kg)	23.30 (11.98)	24.72 (12.72)	23.88 (5.62)
VS (g/kg)	16.11 (0.86)	17.32 (0.33)	17.03 (0.45)
TCOD (mgO2/L)	39,074.87 (5.90)	45,288.61 (2.79)	45,129.28 (4.51)
SCOD (mgO2/L)	26,169.41 (8.99)	31,666.18(8.78)	23,779.51 (2.80)
DOC (mg/L)	4191.50 (1.84)	4144.50 (2.48)	4354.17 (1.04)
TN (mg/L)	1513.33 (1.84)	1486.67 (2.48)	1499.50 (1.04)

¹ SD: standard deviation (%).

.

In this table, it can be observed that the addition of iron nanoparticles in the acidogenic reactor, mainly in MN, favored an increase in TS and VS, together with a high SCOD. As previously mentioned, Fe⁰ facilitates the solubilization of organic matter [15]. In addition, their combination with ozone pre-treatment (MNO) resulted in a significant increase in DOC. This effect has been observed previously by Parthasarathy and Narayanan [28], where biological sludge from the municipal wastewater treatment plant was ozonated. However, the mixture (MC, MN and MNO) shows a high TCOD due to the significant addition of particulate matter with the inoculum (Table 1).

Table 3. Results obtained at the end of the Biochemical Methane Potential tests. Control (MC), nanoparticles (MN), and nanoparticles together with ozone pre-treatment (MNO).

Parameter	MC Mean (SD 1 %)	MN Mean (SD 1 %)	MNO Mean (SD 1 %)
pH	8.26 (0.73)	8.20 (0.92)	8.25 (1.23)
TS (g/kg)	18.85 (0.22)	19.88 (0.30)	18.21 (0.16)
VS (g/kg)	11.55 (0.18)	12.56 (0.61)	11.36 (0.62)
TCOD (mgO2/L)	22,584.56 (14.97)	23,381.19 (1.18)	22,823.55 (0.52)
SCOD (mgO2/L)	8882.46 (7.77)	9599.43 (8.29)	8762.97 (9.08)
DOC (mg/L)	595.93 (6.24)	1516.87 (2.70)	838.00 (1.39)
TN (mg/L)	1977.75 (5.23)	1726.00 (2.68)	1552.25 (1.30)

¹ SD: standard deviation (%).

shows the results of the physico-chemical characterization parameters of the substrates once the BMP was finished. This characterization allows us to observe the changes experienced during the BMP test, as well as the influence of iron nanoparticles and iron nanoparticles combined with ozone pre-treatment on the anaerobic co-digestion process.

A slight increase in pH has been observed at the end of the BMP tests. The pH of all the mixtures increased and reached values close to 8, a trend that has also been documented by Sillero et al. [29]. It can be seen that all the values of the parameters analyzed (Table 3) decreased with respect to the initial values (Table 2). This indicates that organic matter degradation took place during the BMP test. However, TN showed an increase in concentration during the process in all the experiments, possibly due to the fixation of inorganic nitrogen into organic forms by microorganisms.

At the end of the BMP, MN continues to show high concentrations in the digestate of TS and VS, TCOD, SCOD and DOC. This finding highlights the potential for the digestate to be used in agriculture due to its nutrient-rich composition enhanced by the presence of iron nanoparticles.

3.2. Evolution of Solids and Soluble Compounds

The BMP tests resulted in the removal of particulate matter in terms of both TS and VS, as presented in Table 2 and Table 3 and illustrated in Figure 2. The removal efficiency of VS is one of the crucial control factors in the AD process [29]. As can be seen in Figure 2A, the percentages of TS removal were 19%, 20% and 24% for the C, N and NO experiments, respectively. In the case of the VS removal, their values were 28%, 28% and 33% for the C, N and NO experiments, respectively. Based on these results, it can be stated that the iron nanoparticles, and especially ozone, modestly contribute to the improvement of AD by achieving slightly higher VS removal percentages. Ozone is an effective method for solid removal and is also used in other scenarios, such as pre-treatment to enhance water disinfection [30].

The percentages of TS and VS removal were less than 50% in all cases, similar to those obtained by Sillero et al. [31], where BMP was carried out with digestate from a BHP using sludge, vinasses and poultry manure as substrates.

Figure 2B shows the removal of oxidized compounds in terms of COD. As can be seen in this Figure, the COD removal was enhanced by the addition of iron nanoparticles and the use of ozone as a pre-treatment. A number of studies have claimed that powdered activated carbon (PAC) [32] and iron nanoparticles containing Fe⁰ [14] can improve COD removal. Similarly, the same effect has been observed with ozone [33].

In terms of TCOD, the control tests achieved 42% removal, whereas the nanoparticles tests (MN) achieved a removal of 48%, and their combination with ozone (MNO) achieved a removal of 49%. Thus, the increase in TCOD removal was 33% with the addition of the nanoparticles (iron nanoparticles encapsulated in carbon) and 35% when combined with the ozone pre-treatment. Other authors, using powdered activated carbon (PAC) at a concentration of 1.7 g/L in anaerobic membrane bioreactors (AMBRs) and treating simulated domestic wastewater at 35 °C and pH 7, achieved a 22% COD [34]. In this regard, the results obtained with ozone align with the literature. BMP tests conducted using ozone-pre-treated samples of OMW (olive mill wastewaters) achieve approximately 40% COD removal [35].

Regarding the soluble material (SCOD), its removal was improved by the addition of nanoparticles, reaching a removal percentage of 70%, while without nanoparticles, it was removed by 66%. In other studies [14], a SCOD removal ratio of 82% was achieved by applying 0.5 g Fe⁰/L during mesophilic AD in a batch process involving a mixture of dewatered sludge and secondary sludge from a WWTP. However, this dosage is higher than that used in this study (0.05 g nanoparticles/L), and it is important to note that these nanoparticles also consisted of only 2.5% Fe⁰. In the case of the combined treatment with nanoparticles and ozone, the SCOD removal only reached 63%, a value lower than that obtained in the control tests—66%. This could be explained by the previous oxidation of a fraction of the SCOD by the ozone acting negatively on the AD process performance. In general, the soluble material (SCOD) exhibits higher levels of removal than the total ones (TCOD).

The removal results obtained allow us to estimate the conversion of organic matter to methane. This allows us to assess the methane yield of different substrates and identify the most suitable pre-treatment or additive for use in anaerobic co-digestion through BMP testing.

3.3. Volatile Fatty Acid Evolution

In Figure 3A, the total acid concentration at the beginning and at the end of the tests are presented. The initial total acid concentration, measured as acetic acid equivalents, was similar for C and MNO, while MN was slightly lower (2.5, 2.4 and 2.2 g acetic acid/L, respectively).

It is important not to attempt to achieve a high total acid concentration through ozone dosing, as this does not guarantee an improvement in methane production efficiency. Previous studies suggest that the ozonation of sludge improves the hydrolysis rate and, consequently, the VFA production rate. However, a VFA production over the uptake capacity of methanogens can significantly inhibit methane production [36].

At the end of the tests, the control test showed a concentration of 0.06 g equiv acetic/L, while MN and MNO exhibited very similar values (0.09 and 0.10 g equiv acetic/L, respectively). In terms of total acidity, there was a reduction of more than 95% in all tests, indicating the proper conversion of VFAs into methane during the methanogenic stage [37]. Notably, the MNO test stands out, achieving a reduction of 98%.

Figure 3B shows the acid profile at the beginning and at the end of the test. It is noticeable that all acids were completely consumed except for acetic acid, which showed an 87% reduction in the MC and MN tests and a 95% reduction in the MNO tests.

In the literature, it has been described that intermediate doses of ozone ranging from 0.049 to 0.157 g O₃/g COD showed high acetoclastic activity with low lag phases [12]. This is consistent with the results obtained in our study, where ozone pre-treatment at doses similar to those used (0.045 g O₃/g TCOD) increased biomass viability and acetoclastic methanogenic activity, thereby reducing the lag phase. However, at higher doses, such as 0.192 g O₃/g COD [12], it was observed that ozone pre-treatment reduced biomass viability and acetoclastic methanogenic activity, resulting in an initial lag period of 10 days. Subsequently, after this lag period, ozone treatment leads to an increase in methane production.

3.4. Kinetic Analysis of Methane Production

Figure 4A shows the cumulative methane production during the BMP test, which was carried out over a period of 28 days. As can be seen in this figure, all the substrates tested showed a similar methane production trend characterized by rapid production in the early days, followed by a gradual decrease until stabilization, consistent with previously reported findings [19]. The MNO test showed the highest methane production results, with 397 mL accumulated, representing a 29% increase in methane volume when compared to the control sample. Similar results have been reported in other studies [8]. These studies used ozonated pre-treated biosolids from the municipal wastewater treatment plant. At a dose of 0.1 gO₃/gTS (higher than the dose used in this study, which was 0.018 gO₃/gTS), these biosolids produced 25% more methane than raw sludge in BMP tests.

Ozone treatment prior to the anaerobic co-digestion of sewage sludge and wine vinasse leads to a significant increase in methane production. This difference in the methane yield attributed to ozone pre-treatment is statistically significant according to the Wilcoxon test, with a p-value of 1.90 × 10⁻⁵ and an effect size (r) of 0.533 (large). This finding is consistent with previous studies reported in the literature [21], which confirmed that ozone pre-treatment during AD of sludge increases methane production. In contrast, ozone pre-treatment during the AD of wine vinasse does not lead to a significant increase in methane production. This negligible effect can be attributed to the role of ozone in reducing the content of biodegradable organic matter in the vinasse, thereby reducing both COD and DOC, as supported in the literature [38].

However, with the dosage of iron nanoparticles encapsulated in carbon used in this work, an increase in methane production was not observed for the anaerobic co-digestion of sewage sludge and wine vinasse. There were no significant differences detected according to the Wilcoxon test, with a p value of 0.45 and an effect size (r) of 0.107 (small). This finding contradicts what has been approved by the supplier for other substrates. Other studies using Fe⁰ affirm the biogas production rate [39] using higher doses. For example, a 58.99% increase in methane yield from swine manure was achieved using a dose of 10 mg Fe⁰/g VS, as reported by Palma et al. [40].

Figure 4B shows the CH₄ yield data of each test expressed in terms of added VS. It should be noted that the MNO showed the highest yield again, i.e., 195 mL CH₄/gSV added, while that for C was 160 mL CH₄/g SV. In the literature, it has been demonstrated that for mixed sludge (primary and secondary sludge) from a domestic wastewater treatment plant ozonized at 0.05 g O₃/g COD, a positive effect of the pre-treatment was observed [18]. This dosage is close to that used in this study, which was 0.018 gO₃/TS₀ = 0.045 gO₃/g TCOD. However, in the study reported in the literature [18], higher levels of performance were observed with sludge alone, at 6.8 mL CH₄/g COD·d compared to 1.1 mL CH₄/g COD·d obtained in this study.

This is because the substrate originates from a prior acidogenic process where organic matter degradation and hydrogen and carbon dioxide production have already occurred [31]. The variations in methane yields obtained in other BMP tests reported in the literature using similar substrates may be attributed to various factors, including variations in operating conditions, experimental parameters, or substrate loading. As a result, the yield results may not be directly comparable to each other.

The modified Gompertz and Cone models were fitted to the experimental data of the cumulative volume of methane over time with the aim of predicting the methane production rate during the BMP experiment. The kinetic parameters fitted to the selected models for the control test (MC), nanoparticle test (MN), and ozone pre-treated nanoparticle test (MNO) are presented in

Table 4. Kinetic analysis for methane to control (MC), nanoparticles (MN) and nanoparticles together with ozone pre-treatment (MNO).

Model	Parameter	MC	MN	MNO
	P (mL)	293.0600	277.8353	402.4725
	Difference (%)	4.8740	7.7165	1.2018
Modified	Rm (mL/h)	1.2500	1.9060	1.9691
Gompertz	ʎ (h)	168.0000	247.6787	69.9502
	R2	0.7000	0.9897	0.9977
	RMSE	30.48190	12.1915	7.3175
	P (mL)	291.7044	276.3757	411.5291
	Difference (%)	5.4830	8.2551	2.3411
	Khyd (1/h)	0.0031	0.0031	0.0058
Cone	n	8.6779	8.7669	3.2656
	R2	0.9910	0.9918	0.9962
	RMSE	11.9357	10.8882	9.15775

.

Figure 5 displays the experimental methane data points alongside the simulation data. The two kinetic models applied describe the methane production kinetics with R² values above 0.99 in all cases, except for the MC experiment in the modified Gompertz model, where R² was 0.70. Therefore, both models are applicable for all experiments except the MC.

Regarding the difference between estimated and actual methane production (difference), the MN experiment showed a larger difference of more than 7% in both models.

The RMSE values for both models were relatively low (Table 4), especially when considering the size or magnitude of the variable being predicted. This indicates that both models provide accurate predictions of the experimental data. However, the fit of the Gompertz model to the control test (MC) showed high RMSE.

Both the estimated methane production (P) and the maximum methane production rate (R_m) obtained from the Gompertz model and the hydrolysis rate constant (K_hyd) obtained from the Cone model confirm that the use of iron nanoparticles alone does not affect methane production, whereas ozone treatment does, with MNO showing higher values for these parameters.

The nanoparticle experiment produced a methane curve similar to that of the C experiment. As noted above, nanoparticles appear to have no effect on methane production, as they show a lag time of 10 days. In contrast, ozone reduces the lag phase to 3 days, as shown by the steeper slope in the early days (Figure 5). This has been confirmed by other studies, which show that similar doses of ozone used in this study on sludge also resulted in a reduction in the initial lag phase [18].

3.5. Cost-Effectiveness Study

Analyzing the cost-effectiveness of the technology, it can be stated that the energy consumption for the ozone generator used in this study was 20.2 kWh/kg O₃. Once the electricity consumption was calculated, and taking into account the recommended ozone dosage, 0.018 g O₃/g TS₀, the electrical energy consumption per gram of TS₀ was 0.36 kWh/kg TS₀. Considering the average industrial electricity prices in Spain in 2023, which is 0.142 €/kWh, the cost of ozone treatment would be approximately 0.052 €/kg TS₀. Given the total solids concentration in the sludge treated in this study, about 43.6 kg/ton, the total cost amounts to 2.25 €/ton of treated sludge. In addition to the ozone costs, it is necessary to take into account the nanoparticle treatment cost. The cost of the nanoparticles is 4 €/kg. Taking into account that the recommended dosage, according to the recommendations of the supplier, is 2 g nanoparticles/kg SV, and the concentration of the sludge and vinasses used in this study was 21.51 kg SV/ton, the nanoparticle dosage per ton was 43 g nanoparticles/ton. Consequently, the total cost for nanoparticle treatment amounts to approximately 0.17 €/ton. Because of that, the total cost for the combined treatment, nanoparticles and ozone, would be about 2.42 €/ton.

The main income is derived from the methane generated. The methane yields obtained in the different experiments were the following:

−

Without pre-treatment: 0.160 m³/kg VS₀ <> 0.110 m³/kg TS₀ <> 2.36 m³/ton

−

With ozone and nanoparticle pre-treatment: 0.195 m³/kg VS₀ <> 0.139 m³/kg TS₀ <> 3.00 m³/ton

Taking into account the lower calorific value of methane, 10.55 kWh/m³, and that the price of biomethane in September 2024 was about 0.11 €/kWh, the total revenue from methane production would be:

−

Without pre-treatment: 2.36 m³ CH₄/ton·10.55 kWh/m³·0.11 €/kWh = 2.74 €/ton

−

With ozone and nanoparticle pre-treatment: 3.00 m³ CH₄/ton·10.55 kWh/m³·0.11 €/kWh = 3.48 €/ton

Balancing the incomes and the expenses, it can be concluded that the ozone and nanoparticle pre-treatments enhance methane production by about 0.65 m³ CH₄/ton. Nowadays, the current price of biomethane, 0.11 €/m³ CH₄, is not enough to recover the expenses, about 2.42 €/ton. However, the associated advantages, such as better dewatering and VFA production, the higher quality of the biogas obtained, and the presented lower hydrogen sulfide concentration, could make the combined treatment convenient, mainly in a scenario of increasing prices of energy.

4. Conclusions

In the methanogenic phase of the TPAD, the removal of particulate matter was enhanced by the addition of nanoparticles, resulting in a 9% increase in TS removal and a 6% increase in VS removal. When combining the nanoparticles with ozone pre-treatment, these improvements were even more pronounced, with TS and VS removal reaching 27% and 24%, respectively. Additionally, the combined treatment improved total COD removal by 33% and soluble COD removal by 35%, thereby enhancing the overall efficiency of AD. Ozone pre-treatment also significantly improved VFA removal, with total acidity reduction reaching 95% in combined treatment tests, compared to 87% in control and nanoparticle-only tests.

Regarding methane production, the addition of iron nanoparticles did not yield a statistically significant increase. However, ozone pre-treatment significantly improved methane production and increased the methane production yield by 22% (mL CH₄/g VS_added). The modified Gompertz and Cone models effectively described methane production, showing that ozone pre-treatment enhanced the production potential and maximum production rate while reducing the latency period from 10 days to 3 days.

This study significantly contributes to the improvement of processes involved in TPAD by enhancing the isolated production of methane and hydrogen and optimizing the recovery of effluents rich in VFAs. Furthermore, it highlights the capability of producing highly valuable acids, such as butyric acid, which can be used as precursors of valuable chemicals, such as bioplastics. Importantly, the TPAD process can also be utilized for the treatment of other biodegradable wastes, including those with high pathogen content, such as livestock waste and sewage sludge. The implementation of a thermophilic phase ensures that the digestate obtained is free of pathogens, making it suitable for agricultural applications. In summary, these strategies enhance the production of different by-products, thereby aligning with a multiproduct biorefinery model and promoting a sustainable approach to waste management.