Effect of Co-Digestion Ratios and Temperature on Biomethane Production in Anaerobic Co-Digestion of Cheese Whey and Tomato Waste

Abstract

Tomato processing and dairy industries generate significant effluents worldwide, contributing to environmental pollution and nutrient loss. Anaerobic digestion (AD) offers a sustainable solution by treating these effluents while recovering nutrients and producing biomethane. Substrate composition and temperature play a key role in AD efficiency. This study investigates the batch co-digestion of tomato waste (TW) and cheese whey (CW) under mesophilic (37 °C) and thermophilic (55 °C) conditions over 20 days. Fresh cow manure (CM) served as the inoculum, maintaining a substrate-to-inoculum ratio of 1 (S/I = 1) across all digesters. The co-digestion ratios (CDRs), expressed as CW/TW (gVS/gVS), were set at 4.6, 1.7, 0.8 and 0.3. Co-digestion of TW with CW produced 2.5 times higher methane yield than mono-digestion of TW in both temperature conditions. Similarly, among all digesters set under both temperature conditions, digester 2 (CDR = 4.6) exhibited the highest performance, producing 44 mL/gVS-added cumulative methane under mesophilic conditions and 182.5 mL/gVS-added under thermophilic conditions. Across all CDRs, thermophilic digesters outperformed mesophilic ones, generating three times more biomethane. The modified Gompertz model effectively described the experimental data, achieving R² values between 0.97 and 1, confirming an excellent fit.

1. Introduction

Anaerobic digestion (AD) is a sustainable process for producing renewable energy by treating organic waste. In Europe alone, there are about 17,400 AD plants, with Germany (~11,000 plants) and Italy among the leading countries in this regard [1]. This process involves various archaeal communities that react with organic waste in the absence of oxygen. AD includes four main stages: hydrolysis, acidogenesis, acetogenesis and methanogenesis [2]. During hydrolysis, complex polymers are broken down into simpler monomers, which are then further converted into volatile fatty acids (VFAs) during acidogenesis [3]. In acetogenesis, these acids are converted into acetate and acetic acid [4]. Finally, in methanogenesis, the products of acetogenesis are converted to methane, carbon dioxide and a small amount of hydrogen sulfide (H₂S) [5]. When compared to alternatives like composting, incineration and landfilling, AD is both economical and environmentally friendly [6]. It reduces pollution, generates biogas and produces digestate that can be used as fertilizer [7]. AD supports United Nations Sustainability Goals such as Goal 1 (No Poverty), Goal 13 (Climate Action), Goal 7 (Clean and affordable energy) and Goal 8 (Decent Work and Economic Growth) [7]. This technology can treat various organic wastes, including agricultural and dairy processing wastes, as well as kitchen and dining food waste and many more [8,9]. AD is employed globally, from small domestic setups to large commercial facilities, especially in developed regions like Europe [10]. Many factors affect AD, including substrate and inoculum type, temperature, pH, hydraulic retention time, etc.

Among various organic wastes that can be used in AD, agricultural waste like tomato waste (TW) has unique opportunities when considered as a substrate for this process. Each year, the tomato processing industry produces 8 million tons of waste (3–5% of total tomato production) [11], consisting of peels, seeds and pulp, collectively known as tomato waste. Generally, it consists of 94–95% water content, 0.9–1.7% proteins, 0.2–0.5% fats and 3.9–5.96% carbohydrate content [12]. TW production is a challenging issue in the global tomato processing industry [13,14]. This waste comes from tomatoes unsuitable for making pastes, juices and sauces due to their maturity, color or shape. Similarly, overproduction in peak demand can also result in a surplus amount of tomato products, ultimately thrown away as waste. Similarly, surplus amounts of tomatoes overproduced due to high demand are also thrown away as waste. When thrown on bare ground, such waste can pose significant environmental risks if not properly managed [13,15].

Likewise, another organic waste from dairy effluents, cheese whey (CW), has also been considered as a promising substrate for AD. Around the world, CW is produced in the amount of 190 billion kgs per year [16]. It is a liquid that is produced during cheese production. For every 1 kg of cheese production, approximately 9L of CW is extracted. Generally, its pH is around 6, and it contains 50% of milk content, like lactose and proteins. Typically, CW consists of water (93–94%), 0.7–0.9% proteins (β-lactoglobulin and α-lactalbumin), 4–5% lactose, 0.7–0.9% minerals and a small amount of fats [17]. It has high energy content, hence, vulnerable if disposed of in the environment. It can be utilized for microbial fermentation, health-related drinks, biofuels and vinegar [18]. Several lab-scale and pilot-scale experiments have been conducted so far using CW as a feedstock [19,20]. On one hand, it can produce energy and be a way for resource recovery; however, if left untreated, it can pollute the environment [21].

Despite the potential of TW and CW as substrates in AD, they pose significant challenges when digested individually. For instance, TW has a naturally lower pH, which can hinder the activity of methanogens if the process is not properly managed [22]. Methanogens require a stable pH range of 6.5–7.5 to function efficiently. Additionally, TW has a low buffering capacity due to its low alkalinity, which limits its ability to neutralize volatile fatty acids (VFAs) produced during digestion. As a result, accumulation of these acids leads to further pH reduction that inhibits methanogenic activity and reduces methane yield. However, the accumulated VFAs, which are versatile biochemical intermediates, are becoming relevant in fields of bioplastic, solvent or biofuel production, an alternative valorization pathway highlighted in recent studies [23]. TW has a high percentage of structural cell wall components, which are difficult to degrade, causing lower biogas production in AD [24]. Similarly, CW mono-digestion can also lead to acidic pH and early acidification due to its increased biodegradability and lower alkalinity. Such conditions can easily reach high VFAs accumulations and, hence, process failure [25]. Moreover, the higher protein content and rapid degradability of CW also cause foam and scum formation, which makes it difficult for gas to escape from the substrate; this phenomenon also lowers mixing efficiency and creates dead zones [26].

Co-digesting TW with other organic wastes has proven to be a highly efficient method; for example, improvement in methane yield was observed when CM and corn stover were co-digested with TW in amounts of 12–40% of total VS added [27]. In the study of Gil et al., improved methane production was observed when TW had around a 50% share of total VS added [28]. Similarly, co-digesting CW has also shown promising results with other substrates like sludge, cattle manure and agricultural wastes. According to Hallaji et al., co-digestion can increase the efficiency of CW AD by 31–180% [29]. Additionally, the efficiency of CW AD in methane production can be increased by 30% when co-digested with sludge and fruit waste [26].

Challenges associated with AD of CW and TW, when treated separately, can be mitigated through their combined treatment. For instance, the low degradability and C/N ratio of TW can be offset by the high biodegradability and C/N ratio of CW, creating synergistic effects on methane production efficiency [30,31,32]. This study aims to evaluate anaerobic co-digestion performance of TW and CW under both different temperature conditions (mesophilic—37 °C, thermophilic—55 °C) and different co-digestion ratios (CW/TW = 4.6, 1.7, 0.8, 0.3 gVS/gVS), in order to determine their impact on methane yield, biodegradation kinetics and process stability. This is carried out under fixed substrate-to-inoculum ratio of 1.0. To our knowledge, this study is first of its kind that evaluates the influence of co-digestion ratio and temperature on CW–TW mixtures simultaneously and utilizes the modified Gompertz model to quantify kinetic parameters (P_max, R_max, λ) for each operational condition. Overall, this approach and the obtained results provide novel insights into assessing the potential of synergistic behavior and kinetic response for agro-industrial wastes during AD.

2. Materials and Methods

2.1. Substrate and Inoculum

For TW, tomatoes were purchased from a local market, crushed using a laboratory grinder (IKA MF 10 Basic, IKA Werke GmbH & Co. KG, Staufen, Germany), and passed through an ASTM E11 No. 6 sieve (Gilson Company Inc., Middleton, WI, USA) with an aperture size of 3.35 mm. CW was extracted from coagulating fresh milk using acetic acid. The whey was carefully separated from the curd manually. Both TW and CW were kept for short-term refrigeration at 4 °C, as this would preserve the substrate’s original characteristics, so realistic handling can be simulated prior to subjection to digestion. Freeze and thaw could alter matrix structure and composition. Regarding the selection of inoculum in an anaerobic system, it is important as it can affect the overall system performance in terms of stability and efficiency in biogas production. For this study, fresh cow manure (CM), collected from a cow farm in Umm-salal, Doha, Qatar, was used as inoculum. Many studies have mentioned CM to have the potential for using it as an inoculum because it is a major source of methanogens like Methanomicrobiales, Methanococcus, Methanosaetaceae, Methanosarcinaceae, etc. [33,34]. Total solids (TSs), volatile solids (VSs) and VFAs of the substrates and inoculum were measured using the standard methods [35]. Detailed properties of the substrate and inoculum are shown in

Table 1. Substrate and inoculum properties.

Parameter	CW	TW	Inoculum
pH	5.07	4.27	8.5
TS (%)	5.1	4.65	4.8
TS (g/L)	50.7	42.25	47
VS (g/L)	45	36.9	40
VS (% of TS)	88	87	85
C/N	30	14.1	13
TC (%) *	2.4	36.8	36.2
TN (%) *	0.08	2.6	2.9
Cellulose + Hemicellulose + Lignin + Ash	<1%	14.5%	-

* Percentage is based on dry weight (w/w).

, and further details on the analytical methods can be found in Section 2.2.

2.2. Analytical Methods

TS and VS were measured using an analytical balance (OHAUS, model AX224/E, Parsippany, NJ, USA), a drying oven (105 °C) and a muffle furnace (550 °C) [35]. pH was measured using a Benchtop pH Meter (Model PH-3110, Xylem Analytics Germany GmbH D-82362, Weilheim, Germany) after calibration. VFAs (mg/L) were monitored throughout the digestion process following methods followed in litrature [35]. Potentiometric titration of the digestate supernatant was carried out using 0.1 M HCl and 0.1 M NaOH solutions. The titration curve established between pH 3.5 and 7.0 was used to determine volatile acids’ concentrations. Titrations were performed using a pH Meter (Model PH-3110, Xylem Analytics Germany GmbH D-82362, Weilheim, Germany).

Methane volume and composition were measured using the Automatic Methane Potential Test System (AMPTS II Light, BPC Instruments AB, Lund, Sweden, measurement precision ±1% (CV)). Biogas was passed through a 5 M NaOH absorption unit where a thymolphthalein indicator was placed to remove CO₂. This allows only CH₄ to reach the gas-volume measuring device, which works on the liquid-displacement and buoyancy principle.

Total organic carbon (TOC) and total nitrogen (TN) were measured using a TOC/TN Analyzer (TOC-L CSH, Shimadzu, Kyoto, Japan), according to ASTM procedures [36]. Fiber fractions (cellulose, hemicellulose, lignin) were measured following Van Soest NDF/ADF/ADL procedure [37]. All the measurements were carried out in triplicate.

2.3. Experimental Program

Automatic Methane Potential Test System (Automatic Methane Potential Test System (AMPTS II Light, Bioprocess Control AB, Lund, Sweden) (Figure 1), a batch experimental setup of “Bioprocess control”, was employed for this study [38,39]. Batch digestion was chosen to identify the biodegradability and subsequent methane potential of CW and TW mixtures under strict conditions. This strategy allows genuine comparison of co-digestion ratios with temperature effects. This setup consists of four units, i.e., a sample incubation unit (Figure 1a), a CO₂ capturing unit (CCU) (Figure 1b), a gas measuring unit (Figure 1c), and a computer (Figure 1d). The sample incubation unit has six digesters, surrounded by distilled water, which keeps the digesters at a constant temperature. Each digester has a 2000 mL volume (1800 mL working and 200 mL head space). The head of each digester has two ports, one for intermittent sample collection for further analysis and the other for biogas collection from head space. The mixing mechanism is also mounted on top of each digester, having ON-OFF buttons and adjustable speed options. Biogas produced in the digester is then transferred through the pipes to the CCU, which consists of six containers connected to each digester correspondingly. Each container in CCU has an indicator solution, made of a standard solution of 5 M NaOH with a thymolphthalein indicator. This captures acidic fractions like CO₂ and H₂S and allows only biomethane to gas measuring units. When the solution reaches its capacity, the blue color indicator turns colorless. The gas measuring unit has wet flow meters, principally working on buoyancy and liquid displacement. These flow meters send digital signals through the device, and finally, data is sent to a computer where data acquisition systems record, analyze and display graphs.

Before preparing the digester solution, firstly, CW and CM were diluted with distilled water to ensure S/I = 1 (gVS/gVS) for all digesters. For digester solution preparation, substrates and inoculum were mixed in volumes as shown in

Table 2. Substrate proportions and properties in digesters.

Digester	CDRs (gVS/gVS)	CW (mL)	TW (mL)	CM (mL)	CW (VSg/L)	TW (VS g/L)	S/I	pH
D1	Control	900	0	900	37.80	0.00	1	6.8
D2	4.6	720	180	900	30.24	6.64	1	6.7
D3	1.7	540	360	900	22.68	13.28	1	6.61
D4	0.8	360	540	900	15.12	19.93	1	6.36
D5	0.3	180	720	900	7.56	26.57	1	6.04
D6	Control	0	900	900	0.00	33.21	1	5.78
Blank	Blank	0	0	900	0	0	-	8.5

, such that co-digestion ratios (CDRs), i.e., CW/TW (gVS/gVS), become 4.6, 1.7, 0.8 and 0.3, designated with D2, D3, D4 and D5, respectively. Similarly, digesters named D1 and D6 were kept as controls, having CW and TW alone, respectively. CDRs were selected such that there is a gradual transition from CW-dominant to TW-dominant digesters, i.e., D1 to D6, respectively, to systematically evaluate their synergistic and inhibitory effects. Moreover, the choice of such CDRs facilitated equal volume distribution of CW and TW across digesters for ease of handling. Three blank digesters having only inoculum were run for the same duration, for each temperature condition. The amount of inoculum was kept constant in all digestors. TS was kept around 5% for all digesters. All other properties of digester solutions are shown in Table 2. Initially, the samples were tested in mesophilic conditions (37 °C) and then in thermophilic conditions (55 °C), each for 20 days. Mixing speed (80 RPM) was kept constant for all digesters with alternate on and off cycles (5 min on and 1 min off). All the tests were carried out in duplicate trials. To compare end-products, VFA concentrations measured in mgL⁻¹ (acetic-acid equivalents) were converted to their COD equivalents by applying the stoichiometric factor 1.067 g COD g⁻¹ acetic acid [40], further converted to methane equivalents by applying AD relationship of 1 g COD = 350 mL CH₄ at STP [41], expressed as mL CH₄ g⁻¹ VS-added. It was calculated using Equation (1):

$$\mathrm{VFA}\text{–}\mathrm{C}{\mathrm{H}}_4 \mathrm{eq}. (\mathrm{mL} {\mathrm{g}}^{-1} \mathrm{VS})={\displaystyle \frac{\mathrm{VFA} (\mathrm{mg} {\mathrm{L}}^{-1})}{1000}}\times {\displaystyle \frac{350}{{\mathrm{VS}}_{\mathrm{sub}}}}$$

(1)

The percentage distribution was made through Equation (2)

$$\%\mathrm{C}{\mathrm{H}}_4={\displaystyle \frac{\mathrm{Methane}}{\mathrm{Methane}+\mathrm{VFA}\text{–}\mathrm{C}{\mathrm{H}}_4 \mathrm{eq}.}}\times 100,\%\mathrm{VFAs}=100-\%\mathrm{C}{\mathrm{H}}_4$$

(2)

2.4. Kinetic Modeling

The modified Gompertz model is generally employed to study the kinetic behavior of the cumulative methane production vs. time graph. This model has shown excellent performance as compared to other models, like first-order kinetics, in terms of fitting accuracy while showing a correlation with experimental graphs [42,43]. Such graphs usually exhibit sigmoidal trends having three stages, namely, the lag phase, the exponential phase and the plateau [44]. The corresponding three parameters are λ (lambda), P_max and R_max. In this study, Python libraries like SciPy. Optimize (version 1.11.4) was used to apply the model to the experimental curves. The general form of the Gompertz model is given as follows in Equation (3):

$$P\left(t\right)=P_{max}.exp\left\{-exp\left[{\displaystyle \frac{R_{max}.e}{P_{max}}}.(\mathsf{\lambda }-\mathrm{t})+1\right]\right\}$$

(3)

where P(t) is cumulative methane production in mL, P_max is the maximum cumulative methane value in mL, λ is the lag phase in days, t is time in days, R_max is the maximum methane rate in mL/day, and e is Euler’s constant. Differences in P_max among co-digestion ratios under each temperature regime were evaluated using two-way ANOVA based on duplicate reactors per condition, with significance set at p < 0.05.

The modified Gompertz model is the preferred kinetic evaluation model for methane production in batch AD due to its empirical fit and prediction of key curve features. Unlike the first-order model, which is unable to compare initial latency or a symmetric logistic function, the Gompertz model incorporates a lag phase (λ) followed by a sigmoidal rise and maximum production rate (R_max) [45], highly simulating typical BMP curves that follow pattern of slow start-up, rapid biogas evolution, and, eventually, a plateau. Multifarious studies reported high goodness-of-fit (high R² and low error) when compared to first-order and logistic models [46]. Gompertz-based fits often explain >99% variance in complete methane data, performing better than other models, especially when a lag phase is present [47]. Hence, it is widely used for BMP modelling due to improved efficacy and accurate interpretation (e.g., lag time, max rate) [46]. However, a mechanistic flow is present in the Gompertz equation. Its parameters do not directly explain substrate or microbial dynamics, limiting its predictive insight and making it sensitive to dataset variability [48]. Besides, it also does not describe multi-phase digestion when gas production deviates from a sigmoidal shape (e.g., biphasic feeds or toxic shocks). This necessitates either model extensions or combining multiple Gompertz components. In short, Gompertz model can capture lag and rate of methane production, often giving better fit performance than first-order or logistic models in batch tests [49].

3. Results and Discussion

3.1. Cumulative Methane Production

The critical role of various CDRs and temperature conditions on the efficiency of the AD process in producing methane with time is illustrated in Figure 2a–d. For a broader view of the entire process, Figure 2c,d display data in days. On the other hand, Figure 2a,b focus exclusively on the first 24 h, as all three kinetic phases of anaerobic digestion are clear in these graphs. Methane production during AD normally follows three kinetic phases: lag phase, exponential phase and plateau phase that reflects the overall methane production profile. The initial lag phase deals with hydrolysis of complex polymers with limited acidogenesis, which is considered a rate-limiting step as hydrolytic enzymes and microbial consortia respond to substrate. In the exponential phase, rapid methane production is achieved via readily degradable compounds that are converted into VFAs, fatty acids, acetate and hydrogen in the stages of active acidogenesis and acetogenesis. Lastly, in the plateau phase substrates are depleted and inhibitory intermediates accumulate, i.e., methanogenesis which results in stabilizing methane yields.

Methane production efficiency is highly dependent on process stability, which is linked to maintaining optimum pH and controlled VFAs accumulation [50]. During the experiments, these stability indicators were monitored during key time intervals to the time beyond which there was slight methane production observed. Graphs of pH and VFAs are shown in Figure 3a,b and Figure 4a,b, respectively.

In mesophilic conditions (see Figure 2a,c), all digesters exhibited a lag phase of 2–3 h, suggesting readily available substrate and rapid acclimatization of microbial consortia to the anaerobic environment, indicating efficient hydrolysis and metabolic adaptation. The lag phase was followed by the exponential increase in methane accumulation from 3 to 9 h. This increase can be attributed to the stable pH range (6–7) and lower VFAs accumulation (2000–3000 mg/L) at the beginning, which facilitated high microbial activity. From 9 h onwards, negligible methane production was observed (reaching a plateau), corresponding to a high accumulation of VFAs, which reached around 7000 mg/L at 10 h, thus causing a pH drop to the 4–5 range. Under acidic conditions (pH < 6), the methanogenic activity is suppressed due to prevalence of microbial community by acidogenic bacteria. Excessive production of VFAs, especially acetic acid, during acetogenesis permeates cell membrane of methanogens, forming H⁺ and CH₃COO⁻ within cytoplasm, which lowers intracellular pH and inhibits vital metabolic enzymes. Meanwhile, high substrate degradation increases internal hydrogen partial pressure that destabilizes syntrophic acetate oxidation, leading to constraining methane formation. These synergistic effects shift the process towards acid accumulation instead of methane production, as explained in the literature [51]. The system did not recover till the experiment ended, as the VFAs accumulated (nearly) 10,000 mg/L, overwhelming the system’s buffering capacity. This pattern of rapid acidification and inhibition due to uncontrolled VFA accumulation aligns with the findings in the literature [52,53,54].

Under thermophilic conditions (see Figure 2b,d), the lag phase of all digesters lasted between 6 and 8 h, almost double the duration observed in mesophilic conditions. The longer lag phase indicates that microbial consortia take longer to acclimate to high temperatures. In the next 5 h, because of the stable pH range (5–7) and low VFA accumulation of less than 3000 mg/L, all digesters produced a high amount of methane at an exponential rate. Following this phase, a plateau was achieved at the 12th hour, as VFA accumulation reached 4000 to 5000 mg/L (reducing pH below 5) (except D1 and D2). However, approaching the 24th hour, D1 and D2 recovered and started methane production exponentially till day 3 and day 4 (reaching the plateau phase). The quick recovery and prolonged methanogenic activity of these digesters correspond to the higher buffering capacity, which is attributed to the increased alkalinity of these digesters [55]. Furthermore, by 25 h, the accumulation of VFAs reached around 7000 mg/L, further dropping the pH to 4, making digester recovery further (even) difficult to recover until the experiment concluded on day 20.

Under both temperature conditions, D6 yielded the lowest methane among all digesters, i.e., 25.47 mL/gVS-added in mesophilic conditions and 66.66 mL/gVS-added in thermophilic conditions. However, when co-digested with CW, methane production increased by 2.5 times in both conditions, as observed in D2. Generally, digesters with high concentrations of CW (D1, D2 and D3) produced higher cumulative methane as compared to others [56]. Among them, D2 (CDR 4.6), characterized by maximum CW concentrations, exhibited the highest peak cumulative methane of 182.5 mL/gVS-added in thermophilic conditions and 44 mL/gVS-added in mesophilic conditions. High temperature helped in the disruption of recalcitrant cell wall components in TW, as well as the denaturing of protein content in CW, thus making a less viscous solution, having readily available organic matter for degradation.

Better performance of digesters having a high amount of CW, i.e., D2, is mainly attributed to a more balanced organic composition. TW has a higher amount of cell wall components like cellulose, hemicellulose, lignin and ash (See Table 1). These components are hard to degrade, which creates a staggered hydrolysis pattern with easily biodegradable contents in CW, i.e., lactose and soluble proteins [56]. Therefore, in the case of D2, VFAs produced by CW in hydrolysis and acidogenesis are converted more efficiently by methanogens, without causing acidification, thus producing sustained methane comparatively. Similarly, lower lignin content (compared to high TW digesters) minimizes enzymatic inhibition, while CW promotes microbial adoption and high extracellular enzymatic activity.

Similarly, enhanced performance of D2 also corresponds to balanced C/N. Such conditions ensure the optimal supply of carbon for energy and nitrogen for microbial growth, which helps in preventing inhibition of the AD process. C/N values lower than the optimum in the case of digesters having high TW concentrations, excess nitrogen might have caused ammonia accumulation, which hinders the activity of microbes, particularly methanogens. This becomes the dominant pathway, hydrogeophilic methanogens and syntrophic acetate-oxidizing [57]. Ammonia stresses cause metabolic shifts, i.e., from the acetoclastic pathway, which is effective in acetate to methane, to a hydrogenotrophic pathway, which is less efficient. In this shift, oxidizing bacteria can be found in high numbers. Furthermore, ammonia inhibition corresponds to Methanosaeta and Methanosarcina, which serve as microbial inhibition indicators [58].

A higher proportion of methane for thermophilic digesters in comparison to VFAs (55–75%), whereas mesophilic digesters showed higher VFAs accumulation (60–75%), suggesting a stronger carbon conversion capability under thermophilic conditions and inhibition-driven VFA accumulation under mesophilic conditions (see Figure 5a,b).

Methane production potential of TW, CW and similar substrates shows significant variation (12 mL CH₄/gVS-added to 440 mL CH₄/gVS-added) across the literature, as shown in

Table 3. Comparison of methane production from CW, TW and similar substrates in anaerobic digestion.

Digestion Type	Substrate(s)	Inoculum	Key Influencing Factors	Methane Yield (mL CH4/g VS-Added)	References
Co-digestion	Tomato residues + cow manure	Anaerobic digester sludge (municipal WWTP)	Mesophilic (35 °C), Batch	12–75	[65]
Mono-digestion	Tomato residues	Anaerobic digester sludge (municipal WWTP)	Mesophilic (35 °C), Batch	324	[65]
Co-digestion	Tomato waste + sewage sludge	Anaerobic digester sludge (municipal WWTP)	Mesophilic (35 °C), Batch	159	[66]
Co-digestion	Tomato residues + cow manure	Fresh cow manure	Thermophilic (55 °C), Batch	~200	[67]
Co-digestion	Tomato residues	Anaerobic digestate from dairy manure digester	Mesophilic (35 °C), Batch	30	[68]
Mono-digestion	Tomato plant waste (stems, leaves, etc.)	Anaerobic digestate from dairy manure digester	Mesophilic (40 °C), Batch	210.8	[69]
Co-digestion	Tomato leaves + cow manure	Fresh cow manure	Mesophilic (35 °C), Batch	30–35	[70]
Mono-digestion	Cheese whey	Fresh cow manure	Mesophilic (35 °C), Batch	7–400	[71]
Mono-digestion	Cheese whey (dairy effluent)	Anaerobic digester sludge (municipal WWTP)	Psychrophilic (20 °C) vs. Mesophilic (35 °C), Batch	389–436	[72]
Co-digestion	Raw cheese whey + coffee pulp	Anaerobic sludge pre-acclimated to cheese whey	Mesophilic (35 °C), Batch	71	[26]
Co-digestion	Tomato pomace + buffalo manure	Fresh buffalo manure	Mesophilic (35 °C), Thermophilic (55 °C), Batch	50–300	[36,73]
Co-digestion	Cheese whey + liquid cow manure	Lab CSTRs seeded with acclimated anaerobic consortia	Mesophilic (35 °C), Two-phase AD (acidogenic + methanogenic)	290	[74]
Mono-digestion	Cheese whey + supplementation	Digestate from farm dairy manure digester	Mesophilic (35 °C), Batch	424	[75]
Co-digestion	Food waste + cow manure	Fresh cow manure	Mesophilic (35 °C), Batch	130–170	[76]
Co-digestion	chicken manure + corn stover	Mesophilic digester effluent	Mesophilic (35 °C), Batch	160–230	[77]
Co-digestion	Food waste + cow/pig manure	Fresh cow/pig manure	Mesophilic (35 °C), Batch	14–30	[33]
Co-digestion	Tomato waste + cheese whey	Fresh cow manure	Mesophilic (35 °C), Thermophilic (55 °C), Batch	25–182.5	This Study

. For example Saranda and Nand conducted TW AD by stepwise addition for 10–12 weeks. Their results showed a good start-up condition and produced 597 mL/gVS of biogas, having 72% methane content [59]. Eslami et al. conducted a batch experiment that yielded around 280 mL/gVS methane, comprising 50% of the total biogas produced [14]. Another study shows that biogas generated from tomato processing waste was 140 mL/gVS, in which the percentage of methane was around 60% [60]. Additionally, Saev et al. conducted semicontinuous co-digestion of TW and cow manure in mesophilic conditions, resulting in 220 mL/gVS-added, for a co-digestion ratio (CDR) of 80:20 (manure:tomato waste) [61]. This variation in gas production arises from several operational and substrate-related factors, i.e., type of digesters, temperature conditions, inoculum type, pre-acclimation, substrate-to-inoculum ratio and physicochemical properties of the substrates, etc. CW methane production potential varies from 320 to 550 mL/gVS-added [62,63]. Similarly, another study finds that biogas from CW AD ranges from 171 to 280 mL/gVS-added, among which methane is around 60% [25,64].

3.2. Kinetic Modeling

To study more rigorous quantitative analysis and to apply it to large-scale biogas production, finding key kinetic parameters using a modified Gompertz model for AD is essential. Although this model has been employed in various AD studies, its application to such novel substrate co-digestion has not been explored. The study provides insights into the methane production, lag phase and degradation rates to understand the biodegradation behavior of the substrates. The modified Gompertz model fitted well (Figure 6a,b) with the experimental data (R² = 0.97–1) (

Table 4. The Gompertz model estimated parameters and R² values.

Type	Digester	Pmax (mL/gVS-Added)	Rmax (mL/h)	Lag Phase (h)	R2
Thermophilic	D1	119.3	75.8	12	0.97
	D2	182.5	97.3	6	0.98
	D3	56.29	591.1	7	0.99
	D4	58.44	1030.2	8	0.99
	D5	86.33	1161.0	8	0.99
	D6	66.63	852.9	9	0.99
Mesophilic	D1	35.11	173.5	2	0.97
	D2	43.16	268.5	2	0.99
	D3	43.15	296.4	2	0.99
	D4	39.1	352.4	3	0.99
	D5	40.41	429.4	3	0.99
	D6	25.47	334.2	3	0.99

). Results show a high impact of temperature conditions and substrate characteristics on P_max, R_max and λ values (Table 4). During thermophilic conditions, D2 showed the highest P_max of 182.5 mL/gVS-added, indicating synergetic effect of substrate composition and balanced C/N among all the other digesters, as well as high-temperature effects on substrate degradation and microbial activities. On the contrary, lower P_max values can be seen for D3 and D4, 56.29 mL/gVS-added and 58.44 mL/gVS-added, respectively, showing a suboptimal C/N ratio, hence, lower performance comparatively. Similarly, in mesophilic conditions, the digestor shows lower performance as compared to high-temperature conditions owing to decreased microbial activity; nevertheless, among all digesters, D2 performed well, having a high P_max of 43.16 mL/gVS-added, while D6 showed the lowest performance, having P_max values of 25.47 mL/gVS-added. These trends were statistically supported, as P_max differed significantly across co-digestion ratios under both mesophilic and thermophilic conditions (p < 0.05).

Production rates in thermophilic conditions were comparatively higher. Among thermophilic digesters, D5 and D6 showed the highest R_max of 1161 mL/h and 1030 mL/h, respectively. Similarly, in mesophilic digesters D4 and D5 showed the highest degradation rate of 352 mL/h and 429 mL/h, respectively. The λ values vary widely among different temperature conditions. Overall, thermophilic digesters took longer than mesophilic digesters. Among the thermophilic digesters, D2 has a lower λ (6 h) owing to high concentrations of readily degradable CW organic content. On the other hand, among mesophilic digesters, similar behavior was observed, i.e., D1–D3 had the lowest λ (2 h). These findings indicate the importance of substrate composition and C/N balance on AD kinetics. Utilizing this model, one can find the effects of different parameters on peak production, microbial adaptation (lag phase) and production rate. These parameters can be utilized to optimize and predict methane production for different experimental conditions.

4. Study Limitations and Perspective

Although the present study provides useful insights into the impact of synergistic effects of temperature and co-digestion ratio on the AD of CW and TW, certain limitations need to be addressed. The experiments conducted in batch scale only provide information on substrate biodegradability and kinetic behavior, but do not provide information regarding steady state kinetics, feed irregularities and microbial adaptation that is observed in semi-continuous or continuous digesters. Additionally, microbial community or enzymatic analyses were not included, which could have otherwise clarified the functional role of specific microbial consortia under different temperature conditions (mesophilic and thermophilic regimes). Therefore, future work should focus on validating these findings under semi-continuous operation and integrating microbial community profiling (e.g., 16S rRNA sequencing, qPCR) to simulate structure–function relationships during co-digestion.

Furthermore, a promising strategy of microbial community engineering or inoculum enrichment to enhance methane production is not practically approachable on a large-scale production as it requires techno-economic consideration. More importantly, to increase methane production, complex operation handling due to engineered or genetically optimized strains, biosafety requirements and management costs, becomes an economic burden. Therefore, practically, optimizing the operational framework (organic loading rate, feeding regime and co-substrate ratio) may require a more cost-effective approach.

5. Conclusions

A batch system was employed to co-digest cheese whey and tomato waste in different co-digestion ratios (CDRs). Cow manure was used as inoculum and digesters were kept in mesophilic and thermophilic conditions for 20 days. After comprehensive data analysis, we concluded that co-digestion of TW with CW produced 2.5 times higher methane yield than mono-digestion of TW in both temperature conditions. Thermophilic digesters outperformed threefold (300%) than mesophilic digesters in methane production. Among all digesters under both temperate conditions, D2 (having a high amount of cheese whey), owing to balanced organic composition and C/N, produced higher cumulative methane, i.e., 182 mL/gVS-added in the case of thermophilic conditions. In mesophilic conditions, the methane production stopped after the initial 5 h due to VFAs accumulation, which caused a pH drop (3–4). However, in thermophilic conditions, the digesters were relatively stable (5–7) in this period and, therefore, performed better comparatively. Modified Gompertz model accurately predicted the kinetics of the process, having R² values in the range of 0.97–1, indicating an excellent fit.