Dual role of grass clippings as buffering agent and biomass during anaerobic co-digestion with food waste

There is a dire need to replace the chemical buffers that regulate the redox environment in single-stage anaerobic digestion of food waste. Hence, the applicability of grass clippings as an eco-friendly buffering agent and biomass during the anaerobic co-digestion of food waste was explored. A focus was primarily given on the effects of grass clippings on the redox environment and acidogenesis. Concomitantly the production of volatile fatty acids, hydrogen and methane in mesophilic conditions was monitored. Organic load and substrate-to-inoculum ratio were kept constant in all the experiments, and no chemical buffer was used. The results revealed that the redox environment was regulated with 10% grass clippings by inhibiting rapid pH drop in the digester. The addition of 2, 4, and 6% grass clippings promoted acidogenesis with increased production of acetic and butyric acids, whereas 8 and 10% grass clippings promoted solventogenesis with ethyl alcohol production. Hydrogen generation from the experiments with grass clippings was in the range of 27–30% of the total biogas, which was marginally higher than the control (25%). Methane concentration was negligible in the biogas generated from all experiments. The acidification rate, VFA production/consumption rate, specific hydrogen yield, hydrogen conversion efficiency, and volatile solids removal were maximum and minimum in the reactors with 6 and 10% grass clippings, respectively. From the above results, it can be concluded that adding grass clippings to food waste would regulate the sudden pH changes and enhance the production of value-added biochemicals, making the process cost-effective.


Introduction
The United Nations Environment Program estimates that around 931 million tonnes of food waste (FW) were generated in 2019, indicating that 17% of global food production may be wasted (UNEP 2021). According to the Food and Agriculture Organization (FAO 2011), 4.4 GtCO 2 eq were generated from global food loss and waste, which is about 8% of total anthropogenic greenhouse gas emissions. In many countries, including India, FW gets disposed of along with municipal solid waste in landfills and open dump yards. This unscientific practice of FW disposal leads to a significant loss of the resource (Ferronato and Torretta 2019) that has excellent potential to generate clean energy and valueadded products via anaerobic digestion (AD) (Al-Wahaibi et al. 2020). However, single-stage AD of FW has limitations associated with a sudden drop of pH and concurrent inhibition of acidogenesis (Sarkar et al. 2019) and methanogenesis (Han et al. 2019). Drop of pH results mainly due to the accumulation of volatile fatty acids (VFA) produced during AD of carbohydrate-rich FW (Chakraborty et al. 2018a) In conventional anaerobic systems, pH regulation is achieved using chemical buffering agents, such as sodium bicarbonate, sodium acetate, phosphate buffer (Selvam et al. 2010), sodium hydroxide, (Sarkar et al. 2019), and calcium oxide (Chakraborty et al. 2018a). However, these chemicals are expensive and can adversely affect acidogenesis and methanogenesis processes (Yan 2016). Hence, there is a dire need for an organic co-substrate that could primarily act as a buffering agent to regulate the pH during single-stage AD of FW. Additionally, the co-substrate also gets digested and contributes to the production of desired products, such as VFA, ethanol and biogas. Substrates with low (C/N) ratios (Rabii et al. 2019) such as protein-rich organic waste can act as a co-substrate during anaerobic digestion of food waste and maintain high pH due to ammonification (Rahman et al. 2021). However, this buffering co-substrate should be readily available for free or at minimum cost to make the process cost-effective. While selecting the appropriate co-substrate, care has to be taken that either the co-substrate or its degradation products should not be toxic to the microorganisms and the anaerobic digestion process. One such promising co-substrate is grass clippings (GC), which are cut grasses collected in the mower when the lawn is mowed.
Since FW has a low C/N ratio, its biodegradability can be improved when it is co-digested with GC and other lignocellulosic wastes such as agricultural residues and weeds (Rabii et al. 2019) for better resource recovery (Vats et al. 2020). Anaerobic co-digestion is well-known for regulating acidogenesis and methanogenesis by diluting toxic compounds, maintaining the ideal C/N ratio, improving nutrient distribution, increasing organic loading rate, and favouring microbial syntrophy and growth (Chakraborty et al. 2018b).
There are a few investigations done on anaerobic co-digestion of various substrates, including FW and GC (Table 1). Nevertheless, these investigations focused on controlling methanogenesis (Fitamo et al. 2016), producing liquid organic fertilizer (Mostafazadeh-Fard et al. 2019), increasing biogas yield and overall performance of the AD process (Solarte et al. 2017;Darimani and Pant 2020). However, other products of economic importance, such as VFA, ethanol and hydrogen, can also be produced from the anaerobic co-digestion of FW and GC.
The VFAs are essential building block chemicals; hence there is a tremendous global demand for them (Dionisi and Silva 2016). At present, about 90% of VFA available in the market is generated from petro-based production methods, causing adverse health and environmental effects (Ramos-Suarez et al. 2021). Continuous depletion of fossil fuels has led to increased dependence on alternate energy sources, primarily alcohols. Ethanol, a second-generation fuel, has gained attention in the recent past, as it can be produced from a wide variety of lignocellulosic feedstocks, including FW (Ahorsu et al. 2018). Hydrogen formed during acidogenesis is a clean fuel, with a heat value of 142.354 kJg −1 ; hence it has a huge market potential (Xia et al. 2019).
Till date, no investigation emphasized the specific roles played by GC as a co-substrate for enhancing the VFA and hydrogen production while acting as a buffer during AD of FW. Such experiments should focus on the first four to five days, which are critical for understanding the hydrolysis and acidogenesis processes (Sarkar et al. 2021). During this phase (96-120 h) of anaerobic digestion, the pH drop and acid accumulation are maximum. Therefore, regulating the redox environment during the initial hours of the experiment is imperative to stabilize the AD process, to enhance the bioresource recovery. Hence, an investigation was done to evaluate the specific role of GC both as a buffering agent and biomass during its anaerobic co-digestion with FW. Emphasize was given to optimizing the ideal GC to FW substrate ratio; assessing the effect of GC addition on the regulation of the redox environment and enhancement of VFA, ethanol, hydrogen and methane production; and evaluating VFA kinetics and biogas production.

Substrate preparation
Slurry from an anaerobic digester (Biourja, GPS, Bangalore, India), treating FW, was collected in 5 L containers and used as the inoculum. Fresh FW was collected from the student's hostel mess at BITS Pilani, Hyderabad Campus, India. The fresh FW predominantly consisted of boiled rice, vegetable sauce (sambar), wheat bread (chapatti), egg yolks, etc. During the food waste collection, care was taken to avoid unwanted (non-biodegradable) materials. Immediately after collecting the FW, it was grounded into a slurry using a mixer grinder and used as the primary substrate throughout the experiment. Clippings of grass (Family: Poaceae; Digitaria sanguinalis) were collected from the mower used to trim the lawns in the institute. The length of the GC typically ranged between 6 and 8 mm. The GC were directly mixed as the co-substrate with FW slurry in required quantities to obtain the desired substrate ratios.

Anaerobic co-digestion of food waste and grass clippings
Experiments were conducted in anaerobic reactors (200 ml reagent bottles, Borosil) in which FW slurry weighing 80 g (VS ~ 23.91 g) was mixed with 10 ml of inoculum to make a uniform substrate-to-inoculum ratio. GC were added to this mixture in five concentrations, i.e., 2, 4, 6, 8 and 10% (w/v). Finally, 10 ml of Millipore water was added to each reactor to constant dilution. One set with all the contents except GC was used as the control. The digester bottles were sealed with a rubber septum and thoroughly mixed with a vortex mixer for about a minute. A 20 ml syringe with a needle was inserted into the rubber septum to measure the volume of biogas produced and sample the biogas to analyze its composition. Care was taken to ensure no leakage of biogas through the needles or syringes.
All the reactors were placed inside a temperature-controlled shaking incubator (RIS 24 Plus Orbital Shaking Incubator, Remi, Mumbai, India) maintained at 35 °C and set to 80 rpm, ensuring uniform mixing of the reactor contents. All the experiments were conducted for 96 h primarily to focus on the hydrolysis and acidogenesis phases, which determine the efficiency of biomass valorization and govern the production of desired value-added products. At 24 h intervals, 5 ml of the sample (mixed content) from the reactors was collected in 15 ml falcon tubes (Tarsons, India) and stored in a refrigerator for analyzing various parameters. All experiments were conducted in triplicates; two were used for sampling and one for measuring the volume of biogas.

Physicochemical characterization
Before starting the experiment, total chemical oxygen demand (tCOD), total solids (TS), and volatile solids (VS) analyses were performed for FW, inoculum, GC, and the mixed reactor contents. These analyses were repeated at the end of the experiments for the reactor contents. Analyses of pH, soluble COD (sCOD), total VFA, VFA distribution, ethanol, and biogas were carried out periodically for the samples collected from the reactors. Analyses of TS, VS, tCOD, total VFA and sCOD were done as per the procedure described in Standard Methods for the Examination of Water and Wastewater (APHA 2012). Freshly ground FW slurry was appropriately diluted for tCOD analysis. The supernatant was collected by centrifuging the FW slurry at 10,000 rpm for 10 min, filtered through 0.45 µm filter paper, diluted according to the organic load of the FW and used for sCOD analysis.
For VFA distribution analysis, samples were centrifuged at 10,000 rpm for 10 min, and the clear supernatant was filtered through a 0.45 µm cellulose acetate membrane. To 100 µl of the filtrate, 850 µl of Millipore water and 50 µl of formic acid were added to get 1 ml of sample in gas chromatograph vials to analyze VFA distribution. Finally, soluble products were analyzed using gas chromatography (HP 6890 Series, Hewlett Packard) provided with injector and flame ionization detector (FID) and operated at a temperature of 250 °C. At 20 mL/min (25 psi), nitrogen was used as the carrier gas. The oven temperature was programmed as follows: 120 °C for 5 min, increased to 180 °C at 5 °C/min, and then maintained at 180 °C for another 10 min. An Econo-Cap EC1000 (15 m × 0.53 mm × 1.20 µm) coated with 0.2 µm CP-Wax 57 CB column was used as the stationary phase in soluble VFA determination (Chakraborty et al. 2021).
Total biogas volume was measured at 24 h intervals using a 20 ml syringe inserted into the rubber septum of the reactor. The composition of the biogas was analyzed using a gas chromatograph (HP 7890 Series, Hewlett Packard) equipped with a thermal conductivity detector (TCD) and PLOT-Q column (30 m × 0.53 mm × 15 μm). Initially, the standard biogas mixture (H 2 :CH 4 : CO2::1:60:30) was run at 100 °C for 5 min, where N 2 gas (40 ml/min) was used as carrier gas at an injection temperature of 200 °C to measure the retention time and peak area of standard gases. After this run, the biogas samples collected from the experiments were run to determine the distribution of various gases. Hydrogen conversion efficiency, degree of acidification, VFA kinetics were calculated to understand the treatment efficiency according to Sarkar and Venkata Mohan (2017). Two data points of each sample were used for statistical analysis where mean value and standard deviation were estimated using the analysis of variance (ANOVA) test. Regression and correlation analyses were carried out for the VFA metabolites to check possible connections between the final VFA production variables. The level of significance of the results was checked at P ≤ 0.05.

Results and discussion pH flux and its influence on the redox environment
The physicochemical characterization of substrates and inoculum is presented in Table 2. After 12 h, in all digester bottles, the pH started falling from 6.6 to 4.6 due to rapid solubilization and concomitant acid production. The addition of GC (2 to 8%) did not significantly prevent pH decrease. After 96 h, pH was 3.2 in control and 3.8 in the other digester bottles with 2, 4, 6% of GC. Nevertheless, in the experiment where 8 and 10% GC were added, the pH was 4.2 and 4.6 after 96 h. The particulate nature of lignocellulosic materials in the GC helped to regulate the sharp decrease of pH as the difference was observed to be 1.4 units between control and the experiment with 10% GC (Solarte et al. 2017).
It is apparent that without GC addition (control), the pH drop continued after 12 h till the end of the experiment. Such sustained drop in pH was observed in AD of FW as the sole substrate (Moestedt et al. 2020). However, in the experiments with GC addition, after 12 h, no further drop in pH was observed; instead, after 24 h, the pH started increasing till the end of the experiment. Overall, results depicted that the addition of GC (10%) has regulated the sudden drop of pH and controlled the redox environment. Co-digestion of FW and GC helped to balance the C/N ratio of the substrate, thereby leading to enhanced organic matter decomposition and increased microbial biomass. The GC addition increased the activities of some hydrolytic enzymes, which helped in organic C and N mineralization with soluble phenolics  (Dias et al. 2014), which ultimately regulated the pH (Yao et al. 2009).

COD analysis
Hydrolysis and acidogenesis increase the solubilization of FW, which in turn increases the concentration of sCOD. Results depict that the sCOD production in mesophilic conditions increased until 24 h (20.65, 28.5, 30.87, 31.88, and 32.2 g/L) after which it started decreasing till 48 h (11.3, 11.86, 12.25, 15.55, and 16.38 g/L) for 2, 4, 6, 8 and 10% GC, respectively (Fig. 1a). After 48 h, the sCOD concentrations started increasing until 96 h, in the experiments with 2, 4 and 6% GC (15.2, 14.65, 13.78 g/L). Nevertheless, the sCOD concentrations continued to decrease for the experiments with 8% GC (13.5 g/L) and 10% GC (14 g/L) until 96 h. In the control experiment, sCOD production increased to 15 g/L after 24 h and kept increasing to a maximum of 18 g/L at 48 h, after which it decreased gradually to 7.65 g/L at 96 h. The addition of GC increased the sCOD production by 1.37 to 2.14 folds within 24 h, and the cumulative sCOD concentrations remained higher until the end of the experiment, compared to control. Macias-Corral et al. (2017) and Darimani and Pant (2020) reported that the addition of GC increased the substrate's solubilization, consequently increasing sCOD concentration. From
On a similar line, it was observed that the total VFA concentration was considerably higher when sewage sludge was anaerobically co-digested with FW, wastewater and GC when compared to the sludge being used as a single substrate (Fitamo et al. 2016). Likewise, in the current investigation, the cumulative VFA concentrations were 30. 18 and 32.45, 36.89, 36.43, 35.66, and 35.39 g/L in control and 2, 4, 6, 8, and 10% of GC addition, respectively at 96 h. The addition of GC increased the solubilization of substrates, which increased the production of VFAs and other soluble metabolites such as ethanol (Eq. 1), acetic acid (Eqs. 2 and 3), propionic acid (Eq. 4), butyric acid (Eq. 5). Further, the VFAs and alcohol produced during the acidogenic phase get oxidized to acetic acid, as shown in Eqs. 6-8. As could be noticed from the equations, hydrogen is generated as a by-product from all the steps of acidogenesis. (1) (2) C 6 H 12 O 6 + 2H 2 O → 2CH 3 COOH + 2CO 2 + 4H 2 (ΔG = −135.6 kJ∕mol) While ethanol production increased with time in the control, its concentration was many folds higher in the digester bottles with GC. As depicted in Fig. 2 (Supplementary Table 3), the maximum concentration of ethanol and its respective fraction out of the total soluble products were 1.71 g/L (35.80%) at 72 h in control, and 8.51 g/L (50.12%), 11.56 g/L (89.88%) and 12.31 g/L (82.05%), respectively, in digester bottles with 6, 8 and 10% of GC at 24 h. With 2 and 4% GC addition, the ethanol production was lower, 1.68 g/L (36.07%) at 48 h and 4.03 g/L (41.68%) at 72 h, respectively. The ethanol concentration decreased after 72 h in all the digester bottles and control, mainly because of the increase in pH. In the study done by Wainaina et al. (2019), lower pH (~ 3.2-5.0) led to alcohol biosynthesis and the generation of lactic and acetic acids.
Acetic acid was dominant at various points of time in the digester bottles with and without addition of GC: 2% (24 h), 4% (24 h), 6% (96 h), 8% (72 h), and 10% (96 h) and control (96 h) contributing 8.8, 3.10, 7.57, 0.92, 1.15 and 2.39 g/L, respectively. As depicted in Fig. 2, it can be concluded that with the increased addition of GC, acetic acid production got delayed and decreased, which affected hydrolysis, Venkata Mohan 2018, 2019), the rate-limiting step of the AD process (Ferdes et al. 2020). Nonetheless, the addition of 2 to 6% of GC was most favourable for the timely production of appropriate volumes of acetic acid. The propionic acid concentration was 4.05 g/L in the digester bottles with 2% GC. In the rest of the digester bottles and control, the propionic acid production was < 1 g/L (Fig. 2).
The butyric acid production got delayed and could be measured only after 72 h and the maximum production was at 96 h, in the digester bottles with GC, but it was at 24 h in control. The butyric acid concentrations were: 1.6, 2.36, 2.25, 2.51 and 2.36 g/L for digester bottles with 2, 4, 6, and 8% GC addition and control, respectively, and absent at 10% GC (Fig. 2). Eventually, butyrate gets converted to acetate and hydrogen; hence the concentrations of acetate and butyrate are inversely proportional to each other, in all the digester bottles with GC addition, except 10% (Sarkar et al. (2016) and Karthikeyan et al. (2016a). The main reason for no butyrate production at 10% GC is higher ethanol production (solventogenesis). Long-chain fatty acids (e.g., caproic acid) are generated through the β-oxidation pathway, where ethanol acts as electron donor and short-chain fatty acids (e.g., butyric acid) act as the electron acceptors (Wu et al. 2018).
In all the digester bottles with GC, valeric (C5) and caproic (C6) acids were produced, but isovaleric and isobutyric acids production was negligible. The maximum valeric acid concentrations were 1.83, 1.77, 2.02, 1.3, 2.25 and 2.34 g/L for control and the digester bottles with 2, 4, 6, 8 and 10% addition of GC, respectively (Fig. 2). Caproic acid production increased with the addition of GC, but it was less than 1 g/L. The variation in the long-chain fatty acids was due to the acidic environment and the availability of ethanol, which promoted the β-oxidation pathway to produce long-chain fatty acids (Steinbusch et al. 2011).
Higher hydrogen production in the experiments with GC could be attributed to the higher production of acetate, butyrate and propionate in these digester bottles compared to control. Additionally, from the thermodynamics of anaerobic digestion, it is evident that the acetate formation pathway from Acetyl-CoA was favoured (Eqs. 6-8), which also led to hydrogen generation (Taheri et al. 2018). However, methane production was less till 96 h, mainly due to the inhibition of methanogens by the acidic environment (Han et al. 2019). During AD of FW, methane generation begins typically after 7 to 9 days from the start of the digestion due to the slow growth rate of methanogens (Maus et al. 2020). For readily biodegradable substrates such as FW, accumulation of VFA and alcohols due to rapid hydrolysis and subsequent inhibition of acetogens and methanogens is a common phenomenon reported in several studies including Karthikeyan et al. (2016b) and Han et al. (2019). Additionally, due to its particulate nature, the addition of GC at higher concentrations will alter hydrolysis and VFA dynamics, which eventually hamper methanogenesis, as reported by Ferdes et al. (2020).
Compared to control, the VS removal and cumulative yields of ethanol and VFA were higher in all digester bottles with the addition of GC as co-substrate (Table 2). Digester bottles with 6% of GC addition achieved the highest VS removal (47.37%). The cumulative yields of ethanol (0.613 g/gVS), propionate (0.209 g/gVS), and caproate (0.064 g/gVS) were highest with 6% of GC addition. Maximum cumulative yields of acetate (0.468 g/gVS) were with 2% of GC, while at 4% of GC, butyrate (0.106 g/gVS) and valerate (0.190 g/gVS) generation were maximum. Considering the VS removal, VFA and hydrogen production, the addition of 6% of GC is recommended for anaerobic codigestion of FW in mesophilic conditions.

VFA kinetics and statistical analyses
As depicted in Fig. 3, the maximum production rate of ethanol and butyrate was observed at 24 h and 72 h, respectively. The maximum consumption of the metabolites happened at the same concentration of GC, however at different times, i.e., for ethanol -0.215 g/L.h, acetate -0.17 g/L.h, propionate −0.073 g/L.h and valerate -0.02 g/L.h at 48 h and caproate -0.0083 g/L.h at 96 h. The production and consumption rates of VFA and ethanol were higher in the digester bottles with GC addition than in the control. The higher acetate, propionate and butyrate production in digester bottles with GC addition is also correlated with hydrogen production. A similar pattern of VFA kinetics was observed in the study by Sarkar and Venkata Mohan (2017).
To confirm the above-described relationship between ethanol and VFA in their production and consumption, a statistical correlation analysis was carried out for the digester bottles with maximum concentration (10%) of GC and control. As explained in Sect. 3.3 for the digester bottles with higher GC (8% and 10%), a negative correlation was observed between the production of alcohol (increased) and acetic, butyric and propionic acids (decreased). The concentration of caproic acid was directly proportional to ethanol and inversely proportional to acetic acid. In control, acetic, butyric and caproic acids increased with reduced ethanol production, which showed a negative correlation. Propionic and butyric acids had a positive and negative correlation with acetic acid, respectively (Supplementary Table 1). It has to be noted that the dynamics of VFA production is directly dependent on the pH of the environment. In turn, the redox environment impacts the thermodynamic potential of biochemical reactions involved in acetogenesis (Eqs. 1, 2, 3, 4, 5, 6, 7 and 8)  and the syntrophic relationship between hydrolytic, acidogenic and acetogenic microflora (Owusu-Agyeman et al. 2020). ANOVA single factor analysis of all experiments and control was expressed through the sum of square, mean square, F-value, and P-value, which revealed that the experiment is significant as it has a P-value of ≤ 0.05 (Supplementary Table 2).

Prospects of producing value-added products
The VFA, ethanol, and hydrogen are utilized to produce several environment-friendly products, such as biogas, biofuel, bioplastics, and green electricity. Additionally, they serve as valuable raw materials for many industries, including pharmaceutical, food and chemical. The projected compound annual growth rate (CAGR) for hydrogen is around 6% between 2017 and 2023 and is expected to generate around USD 183.34 billion by the end of 2023. The CAGR for acetate is expected to grow at 4.27% of 18,296.90 kilotons by 2023 (Mordor Intelligence 2018). The CAGR for butyric and propionic acids will grow at 5.4% and over 3%, and their market value was estimated to reach 827.6 and 708.57 million USD during 2020-2026 (Markets and markets 2021). The CAGR for ethanol is around 6.7%, and it is expected to generate revenue of USD 115.65 billion by 2025 (Research Grand View 2021).
One of the main drawbacks of AD of FW is that the actual production efficiency of VFA, ethanol and hydrogen that could be achieved is only 80% of the theoretical production efficiency, using the pure strain of microorganisms. The productivity decreases further if a mixed microbial consortium is used (Sarkar et al. 2016). One of the approaches for improving VFA, ethanol and hydrogen productivity is to provide the ideal microenvironment/ redox environment for the microorganisms. This can be achieved by maintaining the pH of the reactor higher than the respective pKa values of the desired short-chain fatty acids. By doing so, protonation of VFA can be avoided, which will check the accumulation of anions that are otherwise responsible for reducing the intracellular pH and causing metabolic disturbances.
Anaerobic co-digestion of FW with lignocellulosic substrates, such as GC, is considered as one of the best ways to control the pH in the reactor. In this investigation, the process pH was controlled by adding GC to FW and the desired value-added products (VFA, ethanol and hydrogen) were produced. Furthermore, any chemical buffering agent was not used to make the process eco-friendly and cost-effective. Certainly, this process could be easily scaled up for mass production. With appropriate pilotscale studies, this approach of employing GC could salvage the industrial scale AD plants treating FW across the globe that are facing the major challenge of controlling the fluctuating pH and concomitant production of desired products. However, further research on adding GC to FW has to be done on a pilot scale to maintain process pH around 6, by focusing on various aspects, including the following, but not limited to optimizing: (i) mixing ratio of GC and FW, (ii) substrate-to-inoculum ratio, (iii) HRT, (iv) usage of enriched inoculum, and (v) pre-treatment of the substrates. Under such a favorable ambience in the digester, the desired syntrophic microbial diversity can be established, which will warrant the production of desired products.

Conclusions
The addition of GC to FW enhanced the rate of acidogenesis and removal of VS and COD. Consequently, the cumulative production of VFA, ethanol, and hydrogen was higher in the experiments with GC than in the control. The role of GC as a buffering agent and biomass was best at 10 and 6% addition, respectively. Furthermore, the addition of 4 and 6% GC was best for acetate, butyrate, and hydrogen production, whereas 8 and 10% of GC was best for ethanol production. Hence it could be concluded that the GC should be added in the appropriate ratio to FW to produce the desired value-added product(s). The procedure followed in this investigation to control the fluctuating pH during acidogenesis is simple yet efficient, cost-effective yet environment-friendly and could be easily replicated at the pilot and industrial scales.