For the batch lab scale anaerobic co-digestion experiments, anaerobically digested sewage sludge was used as the inoculum, collected from the primary anaerobic digester from a local conventional anaerobic digestion operation at a publicly owned wastewater treatment plant. Fats, oils, and grease (FOG) were collected from the surface of the primary clarifier, and thickened waste activated sludge (TWAS) was collected from the bottom hopper of the secondary clarifier. These feedstocks were obtained from a medium-sized wastewater treatment facility (10 - 30 MGD capacity) (Boca Raton Glades Road Wastewater Treatment Plant). Even though FOG collected as floating scum has lower methane potential compared to FOG sourced directly from food service establishments, the ease of accessibility was a main consideration. This approach underscores the practicality of implementing anaerobic digestion by utilizing waste materials that are readily available and easier to source locally. Different fruit juicing waste residuals were obtained from a local small processing facility (Raw Juce Company) to characterize the fruit juicing waste residuals and select suitable candidate food waste feedstocks to undergo anaerobic co-digestion treatment. Prior to analysis, the samples were refrigerated.

Key water quality parameters were analyzed prior to incubation and periodically throughout the batch anaerobic co-digestion experiments. These included pH, alkalinity (mg/L as CaCO₃), total volatile acids (mg/L as CH₃COOH), ammonia (mg/L as NH₃-N), chemical oxygen demand (COD; mg/L as O₂), total nitrogen (TN; mg/L as N), total solids (% wet weight), and volatile solids (% total solids). The water quality parameters were measured following the Standard Methods for the Examination of Water and Wastewater [51] .

Fruit waste characterization was performed to identify the most suitable fruit juicing waste residuals to be used as a substrate feedstock for anaerobic co-digestion. Waste products from the processing of green apples, lemons, pears, red apples, cucumbers, and kale were collected from Raw Juce Company in sterile Whirl-pak bags. About 30 grams of each material was measured and then combined with 100 mL of distilled water in a commercial blender to form a homogeneous slurry that was subsequently analyzed for each of the water quality parameters of interest. Initial characterization was also performed on the individual feedstock of anaerobically co-digested sewage sludge, thickened waste activated sludge (TWAS), and FOG. The C:N ratio of the feedstocks was calculated by converting the chemical oxygen demand (COD), expressed in mg/L as O₂, to mg/L as C, and then dividing this value by the total nitrogen (TN), expressed in mg/L as N.

To determine the biodegradability and technical feasibility of a range of substrates and feedstocks for anaerobic digestion, batch anaerobic co-digestion bioassay tests were conducted in 500 mL serum bottles. Specific substrates were selected based on factors discussed in the introduction. Tests were conducted at mesophilic temperatures (35˚C ± 0.5˚C) for SRT = 63 days and at adequate ratios of substrate to inoculum on a volume basis, as shown in Table 1 . The substrate ratios for this study were selected based on guidance from the USEPA Quality Assurance Protocol Plan (QAPP) and findings from prior research. Two controls—100% anaerobic digestate and 100% TWAS—were included for comparison purposes. Anaerobic digestate and TWAS was mixed with only one type of substrate at a time, followed by a combination of both substrates. This approach provided valuable insights into the effects of each individual substrate and in combination on biogas production. The feedstock combination ratios were selected primarily based on previous research indicating that adding FOG at a rate of 10% - 30% of the total digester feed to two full-scale wastewater anaerobic digesters increased biogas production by 30% - 80% [41] [52] . Furthermore, a 50% inoculum ratio was maintained across all mixtures within the working volume for simulating a realistic scenario where a full-scale anaerobic digester at a wastewater treatment plant typically operates at 50% - 80% capacity. At this inoculum level, adequate microbial biomass will be present to degrade the substrates while maintaining a headspace of 400 mL in all the 500 mL serum bottles for safe and efficient biogas collection.

Table 1. List of specific mixtures tested.

^aS: anaerobic digestate; ^bTWAS: thickened waste activated sludge; ^cFOG: fats, oils, and grease; ^dHSW: high-strength waste; ^eRA: red apple juicing waste residuals.

Mixtures were prepared in duplicate in 300 mL beakers and then transferred to a 500 mL serum bottle with a working volume of 100 mL and headspace of 400 mL. Therefore, two serum bottles were used for each ratio for reproducibility of the results in the batch bioassay tests. Prior to incubation, alkalinity, pH, and volatile acids of all the mixtures were measured. The volatile acids/alkalinity ratio of all mixtures was found to be greater than 0.3. Therefore, approximately 2 grams of sodium bicarbonate powder were added in all mixtures before incubation, and the ratio of volatile acids/alkalinity was brought to within 0.1 - 0.3. Subsequently, N₂ gas was used to purge the headspace of each serum bottle for at least 1 minute to create anaerobic conditions. After purging, an aluminum cap with rubber septum was immediately secured over the serum bottle using a specialized crimping tool. Serum bottles were then partly submerged in a water bath and incubated at 35˚C ± 0.5˚C ( Figure S1(a) (supporting information)). Bottles were manually inverted ten times per day during the incubation period to ensure adequate contact between microorganisms and the substrate, facilitate uniform distribution of temperature, and prevent scum formation, solids deposition, and accumulation of toxic compounds.

Biogas sample duplicates were collected periodically using an 18-gauge, 1.5-inch needle attached to a 550 mL plastic syringe ( Figure S1(b) (supporting information)). The 550 mL capacity plastic syringe was selected with respect to the 500 mL capacity serum bottle to accommodate for the full biogas volume expected to be produced in each measurement interval minimizing the need for repeated withdrawals of the biogas from the headspace of each serum bottle. The syringe was tightly fitted to a 2 micron syringe filter via an adapter to trap any water vapor in the biogas to ensure accuracy in biogas volume and composition measurements. The filter assembly was then connected to the needle. Afterwards, the needle was used to pierce through the rubber septum and the plunger was allowed to move freely until the internal pressure was neutralized and the plunger stopped moving, indicating that equilibrium was reached. The volume of biogas collected was then recorded from the graduated marks of the syringe. Syringe collection of biogas is flexible, economical, convenient and easy to set up with minimal biogas loss. Biogas samples were then injected into a portable biogas analyzer (LandGEM 5000+), and the biogas composition in percent by volume (CH₄, CO₂, and O₂) was recorded. The instrument was purged with ambient air and then calibrated daily using specialty calibration gas with composition of 50% CH₄, 35% CO₂, and 15% N₂.

Siloxanes were analyzed by a certified laboratory (Atmospheric Analysis & Consulting, Inc.) using GC/MS in accordance with EPA Method TO-15M, which defines the performance criteria of sampling and analysis of volatile compounds in air contained in sample canisters. Once per run, biogas samples were withdrawn from the digesters as described in Section 2.2 and then transferred to sample collection canisters via suction using tubing connected to a filter to trap any water vapor present and reduce interference from any trace constituents.

The co-digestion economic analysis tool (CoEAT) is an Excel-based framework designed by USEPA (2010) to assess the preliminary economic feasibility of proposed anaerobic co-digestion. The CoEAT model calculates fixed and recurring costs, solid waste diversion savings, capital investment, and expected biogas production and energy value [53] based on user-provided input. The model considers feedstock availability, capital investment, food waste collection capacity, transportation distances, number of personnel required, average labor cost in the service area, and feedstock moisture content. Additionally, the model accounts for digester sizing, solids retention times, digester capital costs, operation and maintenance requirements, and financial data.

In this research study, the CoEAT model was run for two scenarios: 1) a medium-sized wastewater treatment plant (10 - 30 MGD capacity) receiving 10 tons/day of supplemental food waste but without co-generation capability and 2) a larger-sized wastewater treatment plant (>30 MGD capacity) receiving 20 tons/day of supplemental food waste with an existing co-generation plant. Additionally, the cost of co-digestion on a per ton basis was determined manually by considering the total capital cost, total operating cost, revenue earned from replacing grid-supplied electricity with biogas electricity generation and avoided landfill tipping fees, as well as revenue earned from implementing a food collection fee in the service area. Table 2 shows the input parameters considered to perform the initial economic feasibility assessment using CoEAT model for both the medium-sized and larger-sized wastewater treatment plants.

Table 2. Input parameters provided to the CoEAT model for both medium-sized and larger-sized wastewater treatment plants

The average number of roundtrip miles per collection truck to complete a food waste pickup and delivery was determined using Equation (1). The average number of roundtrip miles to dispose of the biosolids (landfilled or land applied) was calculated using Equation (2), as follows.

${\text{D}}_{\text{food}}=\left({\text{R}}_{\text{food}}+{\text{D}}_{\text{plant-to-station}}\right)\times 2$ (1)

${\text{D}}_{\text{biosolid}}=2\times {\text{D}}_{\text{plant-to-biosolid}}$ (2)

D_food represents average roundtrip distance for collecting and transporting food waste to the digester (in miles), R_food represents the average radius of the service area, D_{plant-to-station} denotes the distance between the wastewater treatment plant and the transfer station (in miles). D_biosolid denotes the average roundtrip distance for biosolids disposal (in miles) and D_{plant-to-biosolid} is the distance between the wastewater treatment plant and the biosolids processing or disposal site (in miles)

The food waste preprocessing costs at the digester location and transfer station were estimated to be $91 per ton per day [55] , considering the costs associated with preprocessing, which includes receiving stations, pumps, heaters, odor control, piping, and grinders, etc. Lastly, default values were used for the ancillary equipment costs incorporated during the modeling scenarios.

Table 3 summarizes the physical and chemical characteristics of each feedstock tested. The FOG and fruit juicing waste residuals were acidic with low to no alkalinity. The inoculum, thickened waste activated sludge, and FOG had low C:N ratios (20-30:1 is ideal for anaerobic co-digestion [27] ). For efficient anaerobic co-digestion to occur, a substrate with a low C:N ratio, such as FOG, should be co-digested with a feedstock that has a high C:N ratio [46] . Red apple juicing waste residuals had the highest C:N ratio (135) among the high-strength organic waste feedstocks tested. Even though lemons had C:N ratio of 30, the presence of citrus acid and D-limonene, an antimicrobial agent present on the lemon peels are toxic to microbial cells during AD, resulting in catastrophic failure of the digester [56] . On the other hand, kale and cucumber had C:N ratios below 10 indicating unsuitability of these feedstocks as an AD additive. Therefore, selecting a mixture of fruit waste might have reduced the methane yield. The selection of abundant, locally available red apple waste as a sole fruit waste residual reduced the variability in the AD process, ensuring consistency and reproducibility in the methane yield results and the influence on pH and VFAs, such that the measured methane production rate can be attributed to red apple waste’s specific characteristics. Besides, preliminary studies were conducted on red apple, pear, and lemon juicing waste residuals, which indicated that red apple waste had the highest methane production potential (523 mL/gVS) compared to pear waste (325 mL/gVS) and lemon waste (30 mL/gVS) [55] . Therefore, the selection of red apple waste for the batch bioassay tests was based on prior experimental evidence demonstrating superior performance and its distinctive properties (C:N ratio of 135).

Table 3. Characteristics of anaerobically digested sewage sludge (inoculum), TWAS, FOG, and different fruit waste juicing residuals.

While we acknowledge that a mixed fruit waste approach could provide a broader spectrum of analysis and a more realistic scenario, our goal was to enhance methane yield while maintaining uniformity in substrate composition. Moreover, this study provides a comprehensive analysis of the behavior of red apple waste during anaerobic digestion, disseminating useful baseline data for future studies when mixed fruit waste will be employed as a comparative feedstock. Additionally, none of the feedstocks had toxic levels of ammonia (>3000 mg/L as NH₃-N; [26] ).

Figure 2 shows the variation in COD levels of different mixtures over time. On day 7, the COD values of all mixtures increased, including controls attributed to hydrolysis and solubilization of particulate organics that increased the bioavailable organic content and stimulated microbial growth. The COD of the 100% inoculum fluctuated throughout the incubation period but tended towards stabilization. Beyond 42 days, the mixtures generally experienced a decline in the rate of organic consumption, indicating evidence of stabilization.

Figure 3 shows the variation in ammonia over time. In general, ammonia concentration increased due to the hydrolysis of nitrogen-bound organic molecules, resulting in ammonia release. Proteins are large organic macromolecules created from building blocks of amino acids linked by peptide bonds. During the hydrolysis stage of anaerobic digestion, larger protein molecules are decomposed into polypeptides and simpler amino acids that undergo further degradation to form volatile fatty acids, carbon dioxide, hydrogen, and ammonia [36] . Nevertheless, throughout the incubation period, ammonia levels in the control and all other tested mixtures remained below the toxic threshold limit (~3000 mg/L as NH₃-N; [26] ). A full-scale anaerobic digester typically operates with an SRT of 30 days, and Figure 3 suggests that the ammonia concentrations in all mixtures tested were well below the inhibitory level after an incubation period of 28 days. The observed increase in ammonia concentrations at SRT > 42 days could be attributed to cell lysis, releasing cellular proteins and nucleic acids that served as an additional nitrogen source similar to the phenomenon observed by Chuchat and Skolpap (2015) [57] during anaerobic digestion of waste activated sludge.

Figure 4 shows the variation in total nitrogen measurements over the incubation period carried out on the slurry and includes the particulate organic matter. Unlike the carbon content that decreases after the digestion due to the conversion of organics into methane, the total nitrogen content remains relatively constant as the speciation may change. As expected, the total nitrogen content remained relatively stable across all mixtures and controls, with fluctuations observed specifically in the TWAS control. Figure 4 shows that the TWAS control had higher TN values compared to the feedstocks that contained high-strength organic wastes.

Figure 5 shows the variation in alkalinity over time. For mixtures with 10% FOG, alkalinity decreased until day 21. On day 7, the alkalinity fell below the minimum recommended alkalinity level (5000 mg/L as CaCO₃) inside the digester in one of the duplicate samples containing 10% FOG and 10% red apple juicing waste residuals. The decrease might be attributed to the rapid degradation of soluble organic sugars into volatile acids. However, the pH and alkalinity values of the mixture later recovered and remained stable, possibly because ammonia was released as digestion continued.

Figure 6 summarizes the variation in volatile acids monitored over 63 days. The concentration of volatile acids in all mixtures increased and peaked on day 21, indicating that the consumption rate of volatile acids was lower than the generation rate until day 21. The inoculum, already in a nearly stabilized condition, had a relatively constant concentration of volatile acids, indicating negligible methane production over the incubation period. As seen with COD, an extended lag phase was observed with VFAs for the mixtures containing 10% FOG. This delay was likely due to the time required for the digested sludge to acclimate to the complex 10% FOG added to the digesters, rather than to methanogenesis inhibition caused by souring from a rapid buildup of volatile acids, as the digestion process eventually recovered.

Figure 7 describes the variation in the volatile acids/alkalinity ratio over time. Initially, the value climbed until day 28, when nearly all mixtures exceeded the optimal range of 0.1 - 0.3 [25] , presumably because hydrolysis and acidogenesis were primarily taking place. On day 42, the ratio of almost all mixtures, including 100% TWAS, dropped below 0.3, indicating a delayed onset of methanogenesis until day 28. Consequently, the elevated ratios observed earlier in the process were likely due to the acclimation time of the microbes to the substrates, with no impending digestion failure expected.

The pH is a primary factor that can indicate the stability of the anaerobic co-digestion process. Figure 8 illustrates the variation in pH over time. The pH remained stable in the range of 7.5 - 8.5 for the duration of the incubation period. No attempt was made to optimize the bicarbonate addition, but the level can likely be reduced since no signs of souring were observed in any of the experiments.

Figure 9 summarizes the reduction of volatile solids (%) in the mixtures monitored over time. Unlike other parameters, volatile solids were measured only on days 14, 42, and 63. Despite this limitation, the selected time intervals effectively captured the trend in volatile solids reduction, allowing for meaningful interpretation of the results. The volatile solids decreased by a maximum percentage (~85%) for the mixture containing 10% FOG and 10% red apple juicing waste residuals within 63 days of incubation. After day 42, the reduction of volatile solids in the mixtures leveled off, indicating stabilization of the digestion process. Meanwhile, the mixture containing 10% FOG as the single substrate reduced volatile solids by 80.5%. Furthermore, except for the mixture containing 10% FOG combined with 10% red apple juicing waste residuals, slight differences were observed in the volatile solids reduction between 42 and 63 days for the other mixtures. The discrepancy may be attributed to differences in microbial density between the respective mixtures on day 42 and day 63.

Figure 10 summarizes the variation in cumulative biogas yield over 63 days incubation. Until day 21, the mixture containing red apple juicing waste residuals produced more biogas compared to controls and those containing FOG. This indicates the presence of more readily biodegradable soluble carbohydrates. After 21 days, a linear increase in biogas production was observed for the mixture containing 10% FOG and 10% red apples, suggesting a lag phase of approximately 21 days for biogas production for mixtures with FOG. The observed lag phase may be attributed to the inhibitory effects of LCFAs produced during the degradation of lipids in FOG. It took time for the anaerobic microbial community to acclimate to the complex FOG substrate and its hydrolysis products.

The observed lag phase aligns with findings from Usman et al. (2020) [58] , who reported a 25-day lag phase during anaerobic digestion of FOG, attributed to the initial presence of LCFAs. In the current study, the lag phase was likely caused by physical transport limitations of LCFAs rather than metabolic inhibition. Additionally, insufficient nitrogen purging might have resulted in extending the lag phase by leaving residual oxygen in the headspace, potentially affecting initial microbial activity. Other factors, such as LCFA concentration, inoculum characteristics, and temperature fluctuations, could also have influenced the duration of the lag phase. To mitigate this delay, future studies could explore strategies such as inoculum pre-acclimation, improved nitrogen purging, and bioaugmentation with LCFA-degrading microbes.

Despite the lag phase, Figure 10 shows that the mixture containing 10% FOG had a biogas yield of 453 mL/gVS after 63 days, while the mixture containing 10% FOG along with 10% red apple juicing waste residuals demonstrated an even higher biogas yield of 492 mL/gVS. This yield was 2 - 3 times higher than the controls (e.g., no FOG or fruit waste) and also surpassed the mixture containing only 10% red apple waste (369 mL/gVS). These findings align with Abdallah et al. (2022) [59] , who noted that lipid-rich substrates can achieve higher methane yields but often exhibit slower hydrolysis kinetics. Similarly, Xue et al. (2019) [60] highlighted a positive correlation between biogas production and lipid content in the digester medium.

Although biogas yield trends indicate a lag phase and microbial acclimation, a detailed analysis of the microbial interactions responsible for biogas production and process stabilization was beyond the scope of this study. However, the performance and stability of anaerobic digestion are directly related to the microbial composition involved in the process [61] . Future studies incorporating metagenomic analysis and microbial sequencing could elucidate microbial diversity, community composition, and functional interactions, particularly those involved in LCFA degradation and syntrophic relationships between fermentative bacteria and methanogens. This would provide deeper insights into their roles in anaerobic co-digestion and enhance our understanding of biogas production dynamics during co-digestion.

Figure 11 shows the variation in methane yield over time and demonstrates that the feedstock composition plays a vital role in anaerobic co-digestion performance. The highest recorded methane yield (280 mL/gVS with 55% methane) was obtained for the mixture containing 10% FOG and 10% red apple juicing waste residuals on day 63, after exhibiting a lag phase of 21 days ( Figure 11 ). The methane yield of the inoculum stayed below 50 mL/gVS throughout the incubation period. The methane yield of the mixture containing 10% FOG as the sole substrate started to increase after day 21 and attained stabilization within 63 days with a methane yield of 243 mL/gVS with 53% methane compared to the highest methane yield of 201 mL/gVS obtained by Mahat et al. (2020) [36] after SRT = 54 days by mixing substrate rich in lipid content with anaerobic digestate in equal amounts.

It is anticipated that increasing the percentage of FOG (>10%), regardless of whether fruit juicing waste residuals were added, will extend the observed lag phase (>21 days) as the microorganisms need to adapt and degrade even higher concentration of complex LCFAs produced from the hydrolysis of FOG [55] [58] [62] . LCFA intermediates are known inhibitors of nutrient transport and enzyme activity and can damage cell membranes and lead to biomass flotation due to coagulation [11] [55] [63] . While this study focused on a 10% FOG loading rate to balance substrate availability and process stability, further research at higher FOG loadings (>10%) or on optimizing the FOG-to-fruit waste ratio is recommended to better evaluate microbial adaptation and system performance during process scale-up. Specifically, future studies could increase FOG loading rates from 10% to 30%, while varying fruit waste ratios to observe their influence on methane yield without compromising the process stability. Such research would provide valuable insights for scaling up anaerobic digestion systems to handle high-strength organic waste streams in industrial settings. Despite this, FOG addition increased methane yield more than simply adding red apple juicing waste residuals alone since FOG lipids typically have a higher energy density (700 - 1430 mL CH₄/gVS; [64] ) than carbohydrates from fruit waste (160 - 350 mL CH₄/gVS; [65] ). Figure 11 shows that replacing 10% of red apple juicing waste residuals with 10% of FOG increased the methane yield from 158 mL/gVS to 243 mL/gVS.

Conversely, Figure 11 also shows that using red apple juicing waste residuals as the sole substrate resulted in a higher methane yield of 158 mL/gVS compared to the TWAS control (133 mL/gVS). This difference can be attributed to the presence of complex extracellular polymeric substances with low biodegradability in the TWAS control, making it resistant to anaerobic co-digestion, resulting in reduced methane production [66] . In contrast, the presence of 10% red apple juicing waste residuals in the waste slurry, with its aqueous composition and simpler structure, enhanced the biodegradability and made it easier for the microbial fauna to access the soluble sugars (fructose, sucrose, glucose, and sorbitol) and increase methane yield. However, the rapid exhaustion of these readily accessible carbon sources might have resulted in the marginal observed increase in the methane yield compared to the TWAS control.

An interesting observation was that the combination of red apple juicing waste residuals and FOG resulted in the highest methane yield measured (280 mL/gVS), twice as much as the TWAS control. The mix of nitrogen-rich TWAS with lipid-rich FOG and carbohydrate-rich fruit juicing waste residuals of the red apple variety were compatible and increased the methane yield. The simple sugars present in the fruit juicing waste residuals may have provided more bioavailable carbon to increase the microbial growth rates supported by the presence of high energy density lipids from the FOG. The red apple waste offered rapidly digestible carbohydrates as an energy source, thereby increasing the biogas production rate and improving the reaction kinetics. Meanwhile, FOG’s lipid content provided a slower released, more sustained source of carbon. Together, these two substrates created a more stable process and continuous methane production. Additionally, adding 10% red apple waste slurry might have aided in physically dispersing the FOG more evenly inside the digester, thereby increasing the bio accessibility of FOG for degradation. In addition, the presence of red apple waste likely enhanced the microbial diversity and density, increasing the robustness of the digestion process. Therefore, the combined recipe created an environment favorable for methanogenic growth by enhancing buffering capacity, diluting potential toxins, providing essential nutrient supplementation, and establishing more balanced conditions for anaerobic co-digestion. Figure S2 shows that the combined mixture (50S:30TWAS:10RA:10FOG) produced more methane compared to a similar study conducted by Mahat et al. (2020) [36] , who used feedstock containing food processing wastewater (FPW) and anaerobic sludge mixed at a ratio of 1:1 (50 mL:50 mL) with a value of 201 mL/gVS.

High-strength organic substrates, such as fruit juicing waste residuals, are generally not expected to generate siloxane compounds, precursors, or derivatives during anaerobic co-digestion. This inference is based on the fact that siloxanes are typically used in personal care products and industrial applications, such as fuel additives and antifoaming agents [67] , which are unlikely to be present in fruit waste streams. However, there are various pathways through which siloxane compounds can contaminate feedstock—for example, siloxane-containing pesticides or siloxane-based cleaning agents. Figure 12 summarizes the total siloxane concentration and the concentration of octamethylcyclotetrasiloxane (D4) in the biogas of several mixtures after 30 days of co-digestion under mesophilic conditions. D4 and D5 are the most prevalent form of siloxanes in biogas, which typically comprise 90% of the total siloxanes in biogas [68] . D4 comprised approximately 60% of the total siloxanes in this study.

Figure 12 shows that all siloxane levels measured in this study were below 2 mg/m³, which did not exceed the emission limit for internal combustion engines (15 mg/m³) or thermoelectric units in Germany (6 mg/m³). However, all biogas samples analyzed had siloxane concentrations higher than 0.1 mg/m³, the recommended limit for biogas intended to be used in turbines and fuel cells. A more stringent limit is established for turbines owing to their greater susceptibility to siloxane damage [69] . With respect to the emission limit for internal combustion engines (15 mg/m³), there was barely any variation in siloxane levels across the control and the mixtures, as the concentrations remained consistently below 2 mg/m³, ranging from 0.8 - 2 mg/m³. D4 concentrations in all biogas samples were below the permissible limit of 11 mg/m³ for energy recovery systems [70] . Thus, addition of high-strength organic wastes for co-digestion with TWAS did not increase the siloxane concentrations such that no additional costs would be expected for siloxane removal from co-digested derived biogas.

For the CoEAT model input in this study, the biogas production potential value was determined from the batch biodegradability tests to be 15 ft³/lb (0.94 m³/kg) of volatile solids destroyed obtained for the 10% FOG and 10% red apple juicing residuals waste slurry. The financial outputs for the two scenarios from the CoEAT model are summarized in Table S2 . The net benefit will be positive after 15 years, suggesting the implementation of co-digestion at an existing wastewater treatment plant as a long-term investment expected to have a net positive benefit with $39,472 and $52,488 for scenarios 1 and 2, respectively. The analysis assumes that 10 - 18 tons/day of food waste will be diverted from landfills. It has been reported that a 10% food waste addition to the digester feed marginally increases sludge volume production due to the high biodegradability of food waste (volatile solids reduction > 85%), while a 50% addition could lead to a 24% increase in dried sludge mass, potentially raising operational costs and necessitating further analysis of increased sludge handling and disposal costs [71] . Using a 5% interest rate and a 15-year planning horizon, the cost per ton of food waste was determined to be $220/ton for scenario 1 and $114/ton for scenario 2.

We also investigated the cost of an annual subscription service based on 50 clients providing 50 tpy of waste. For scenario 1, a $160/ton subscription yields +$81,000 net benefit; scenario 2, a $120/ton subscription provides +$240,000 net benefit. We did not account for avoided landfill air space savings (estimated broadly at $100 - $400/ton but varies by location, available space, and landfill age), flow control issues, or costs associated with an additional 20% - 25% more biosolids disposal volume.