Impact of Organic Load on Methane Yields and Kinetics during Anaerobic Digestion of Sugarcane Bagasse: Optimal Feed-to-Inoculum Ratio and Total Solids of Reactor Working Volume

Abstract

The effect of increasing organic load on the specific methane yields (SMYs) and kinetics of methane production during the anaerobic digestion (AD) of sugarcane bagasse (SB) was investigated in batch experiments at 37 °C. The organic load of the batch AD system was increased based on an increase in the feed-to-inoculum (F/I) ratio (T1–T5) and increase in the Total Solids (TS)% of the working volume (T6–T10). The results show that in both the treatment sets, an increase in organic load led to a decrease in SMY. Higher organic loads in terms of F/I ratio (T4 and T5) were inhibited due to Volatile Fatty Acid (VFA) accumulation. On the other hand, higher organic loads (T8, T9 and T10) in terms of the higher TS% of the working volume was inhibited by the accumulation of NH₄-N. Thus, an organic load of 50 gVS/L at an F/I ratio = 1.0 and TS = 10% (T3) was found to be the highest organic load that had no significant inhibitions among the tested treatments. The results from the kinetic studies show that the first-order kinetic model is the best fit for the SMY data, with average differences% of 2.32% and 3.13% for treatments T1–T5 and T6–T10, respectively.

1. Introduction

Sugarcane (Saccharum officinarum) is one of the most important crops cultivated in the tropical and subtropical regions that contributes to more than 70% of the world’s sugar demand [1]. A large quantity of agricultural residue known as the sugarcane bagasse (SB) is generated during the processing of these canes to produce sugar. Around 526 million tonnes of SB are generated as a result of sugarcane processing each year [2]. Nearly one-fourth of the weight of a sugarcane plant is attributed to be SB, which is the dry fibre pulp left behind after the extraction of sugar juice [3]. It is estimated that nearly 260–280 kg/tonne of SB is generated from sugar mills as a result of sugar production [4,5]. SBs produced from these sugar mills are largely used in cogeneration applications to produce heat and electricity [5,6,7]. However, incomplete combustion of these SBs can result in the release of pollutants such as nitrogen oxides, carbon monoxide and other respirable particulate matter that has a hazardous impact on the environment and human health [8]. Instead of burning these residues as a fuel for boilers, such by-products can be converted to more valuable products using processing technologies such as liquefaction, gasification, pyrolysis and anaerobic digestion [1]. For instance, Cai et al. [9] reported the advantages of using the sludge incineration ash formed as a result of the incineration of waste-activated sludge for the synthesis of functional environmental materials. Similarly, Arelli et al. reported that anaerobic digestion (AD) is one of the most viable processes for converting lignocellulosic organic wastes such as sugarcane bagasse to biomethane [10].

AD is considered to be one of the most effective and sustainable environmental technologies for treating wastes such as the agro-wastes and agricultural residues generated as a result of cultivation and converting them to biogas for renewable energy generation [11,12]. These agro-wastes, which mainly consist of cellulose, hemicellulose and lignin [13,14], are considered to be valuable renewable resources with high potential for biogas generation [15]. Thus, using agro-waste for recovering energy through AD is considered to be a more sustainable waste management practice [16]. However, the resistant and recalcitrant structure of the lignocellulosic biomass due to the presence of a complex lignin structure results in a lower availability of cellulose and hemicellulose for biochemical conversion [5,13]. Thus, the hydrolysis of lignocellulosic biomass is often considered to be a rate-limiting step during AD [17]. Even though the complex lignocellulosic structure hinders the process, the AD of lignocellulosic substrates is widely accepted, as it is more efficient compared to other thermochemical and biological processes [18]. The higher energy output-to-input ratio (28:1) for AD makes it more acceptable compared to other waste management techniques [19]. For instance, Paulose et al. reported a specific methane yield of 187.9 mL/gVS_added when untreated SB was subjected to AD at mesophilic temperature at an F/I ratio of 0.5 [3]. Vivekanand et al., 2014, also reported similar results for untreated SB subjected to AD [20].

However, conventional batch AD processes are performed at lower organic loads, resulting in higher reactor volume and thus higher investments [21]. The amount of waste that can be digested in an AD reactor working at low organic load will be minimal and results in lesser total production [22]. On the other hand, AD at higher organic loads has a higher chance of process failures due to unfavourable conditions [23]. Even though these high-organic-load processes can take up higher quantities of organic wastes, the process stability is largely affected by these unfavourable conditions [24]. Thus, it is essential to determine the maximum organic load of sugarcane bagasse (SB) that can be fed to a batch AD system without inhibiting the AD process. Increasing organic load in terms of increasing the TS content of the reactor working volume has several other advantages over the conventional AD system, such as lower water and energy requirements for degrading feedstocks with high solid content, especially agricultural crop residues, which have solid contents >50% [25]. Further, the use of smaller reactor volumes and higher organic loads may improve the volumetric efficiency of the process [26,27].

Even though a higher organic load helps to attain higher volumetric production, the AD of lignocellulosic substrates at higher organic loads is subjected to severe stress conditions compared to that of the lower organic loads [28,29]. The presence of a complex lignin structure in the lignocellulosic substrates makes the hydrolysis of cellulose and hemicellulose more difficult, especially under a reduced-moisture condition [30]. Many researchers have reported lower cumulative and specific methane yields for different substrates subjected to AD at higher organic loads [26,31,32]. Several factors affect the lower methane yields from these substrates for a low-moisture environment. For instance, Wang et al. [33] reported a 10% reduction in methane yield when the organic load was increased by increasing the TS% from 15% to 20%, while codigesting pig manure with food waste at mesophilic temperature. Similar results were also reported by Motte et al. when wheat straw was subjected to AD at different TS levels (15–25%) to increase the reactor organic load [34]. Process failures due to the accumulation of metabolite intermediates was reported by many researchers under higher organic loads based on the increased TS content of the reactor [32,35,36,37]. The possible reason for such an accumulation of metabolite intermediates at higher organic loads has been attributed to the lesser mobility of microbes to access the substrates due to low-moisture conditions and a lack of available free water in the reactors [24,38,39].

The ratio of feed to inoculum (F/I) in a reactor also plays a vital role in the AD process dynamics [40]. A lower F/I ratio reduces the chances of reactor failure [36]. However, operating a reactor at a very low F/I ratio is not feasible from an operational point of view [41]. Determining the optimum organic load in terms of the F/I ratio for a substrate above which the process inhibits is essential to determine the optimum reactor operational condition [41,42]. A few literature studies have been found, investigating the effect of the F/I ratio on methane production and process dynamics of various substrates subjected to AD. For instance, Jiang et al. [43] reported a 30% improvement in methane yields at an F/I ratio of 1.0 compared to an F/I ratio of 3.0 during the AD of pig manure codigested with food waste in higher TS conditions compared the conventional AD process. Similar results were also reported when an organic fraction of municipal solid waste was subjected to different organic loads based on increases in F/I ratios, such as 1.0, 2.0 and 3.0, under reduced-moisture conditions, where an F/I ratio of 1.0 produced 156% more methane yield than an F/I ratio of 3.0 [21]. Although many researchers have tried to formulate an optimal F/I ratio or organic load at higher TS levels, Wang et al. [36] have reported that there is no optimal F/I ratio, as it depends on a number of factors such as the operational conditions and substrate characteristics. Nevertheless, no studies were found on the effect of increasing the organic load of sugarcane bagasse either by increasing the F/I ratio of the reactor or by increasing the total solid contents in the reactor working volume. Thus, this study aims to investigate the effect of increasing organic load on SMY and process kinetics when SB is subjected to AD.

2. Materials and Methods

The SB was collected from Racecourse Sugar Mill, Mackay, QLD, Australia. Upon arrival at the laboratory, SB was milled to a particle size of 2 mm using a Retsch SM100 ball mill (Retsch GmbH, Haan, Germany). The milled samples were then secured in air-tight bags and stored in a walk-in cold room until further use. Solid inoculum (SI) and liquid inoculum (LI) were collected from a full-scale biogas plant treating primary and waste-activated sludge at 37 °C (Queensland Urban Utilities, Luggage point, Brisbane, QLD, Australia). SI is the residue left over after the removal of the supernatant following the centrifugation of LI and hence has a higher TS% than the LI. Prior to the experiment, inoculum was degassed at 37 ± 1 °C for 5 days to eliminate the residual methane.

2.1. Experiment Design

Two sets of experiments (T1–T5 and T6–T10) were designed to increase the organic load in batch AD of SB. In treatments T1 to T5, the F/I ratio was increased from 0.3 to 3.0 at a constant 10%TS of working volume, whereas in treatments T6 to T10, the working volume TS was increased from 3 to 16% at a constant F/I ratio of 1.0 (

Table 1. Batch assay experimental design.

Treatments	F/I Ratio	TS (%)	Organic Load (gVS/L)
T1	0.3	10	19.9
T2	0.5	10	28.5
T3	1.0	10	50.0
T4	2.0	10	80.3
T5	3.0	10	106.1
T6	1.0	3	13.6
T7	1.0	8	40.3
T8	1.0	12	65.2
T9	1.0	14	88.4
T10	1.0	16	110.0

).

In T1–T5, inoculum with higher TS% than the normal wet AD was used to perform the experiment at a higher organic load. Inoculum with higher TS content was prepared by mixing SI and LI in such a way that the overall TS of the reactor was regulated at 10%. In T6–T10, the TS of the reactor working volume was varied by altering the TS of the inoculum used in each treatment. In each of these treatments, calculated quantities of SI and LI were mixed together to attain the desired overall TS of the reactor working volume. Background methane production for each of these treatments were subtracted using their respective blank assays comprising SI and LI in the same proportion as that of these treatments.

2.2. Biochemical Methane Potential Test

Biochemical Methane Potential (BMP) tests were performed on all treatments (T1–T5 and T6–T10) to investigate the effects of increasing organic load on methane yields during the anaerobic digestion of SB. Three replicates of each treatment were performed and their average value was noted to reduce error. Batch assays were prepared in 250 mL GL45 Schott glass bottles. The desired TS levels of the reactor working volume in these assays were regulated with SI, LI and deionised water. The F/I ratio (VS basis) of these assays was maintained by mixing a calculated amount of SB with the prepared inoculum mixture. Assays were then flushed with 100% N₂ for 2 min to maintain anaerobic conditions and capped with butyl rubber stoppers, after which they were statically incubated at the mesophilic temperature of 37 ± 1 °C. Triplicates of blank assays at the same TS level of the inoculum in the treatment assays (T1–T10) were also incubated at mesophilic temperature to subtract the biogas production from the assays in the absence of any substrates. The overhead pressure was measured on a daily basis and the biogas production was calculated using the ideal gas equation [3]. Biogas composition was categorised using Shimadzu GC-2014 ATF, Rydalmere, NSW, Australia. The experiments were terminated at a stage where the daily biogas production depleted below 2% of the previous day.

2.3. Kinetic Modelling

The first-order kinetic equation (Equation (1)) and the modified Gompertz equation (Equation (2)) were used to determine the different modelling parameters in T1 to T10. Different model parameters were estimated using Microsoft Excel’s^® (Version 2407 Build 16.0.17830.20210) solver function.

$$B\left(t\right)=B_0 [ 1-e^{\left(-k_{hyd} . T_{delay}\right)}]$$

(1)

$$B\left(t\right)=B_0 . ex{p}^{[-ex{p}^{\left({\displaystyle \frac{R_{max} . e}{B_0}} . \left(\lambda -t\right)+1\right)}]}$$

(2)

where B(t) is the specific methane yield (NmLCH₄/gVS_added) on the t^th day, B₀ is the expected maximum methane yield (NmLCH₄/gVS_added), k_hyd is the methane production rate constant (day⁻¹), T_delay, λ is the lag phase (days), e = exp(1) = 2.718282 [44,45], and R_max is the maximum methane production rate (NmLCH₄/gVS_added/day).

2.4. Theoretical Methane Yield and Experimental Biodegradability

Theoretical methane yield of sugarcane bagasse was calculated by using Buswell and Mueller equation which was later modified by Boyle to include nitrogen and sulphur, thereby obtaining ammonia and H₂S in the produced gas (Equation (3)) [3]. Theoretical methane yield (BMP_th) was calculated based on the carbon (C), hydrogen (H), nitrogen (N), oxygen (O) and sulphur (S) contents in the TS of the substrate using (Equation (4)) [46].

$$C_n{H}_a{O}_b{N}_c{S}_d+\left(n-{\displaystyle \frac{a}{4}}-{\displaystyle \frac{b}{2}}+{\displaystyle \frac{3c}{4}}+{\displaystyle \frac{d}{2}}\right) H_{2}O \to \left({\displaystyle \frac{n}{2}}+{\displaystyle \frac{a}{8}}-{\displaystyle \frac{b}{4}}-{\displaystyle \frac{c}{8}}-{\displaystyle \frac{d}{4}}\right)C{H}_4+\left({\displaystyle \frac{n}{2}}-{\displaystyle \frac{a}{8}}+{\displaystyle \frac{b}{4}}+{\displaystyle \frac{c}{8}}+{\displaystyle \frac{d}{4}}\right)C{O}_2+N{H}_3+d{H}_{2}S$$

(3)

The subscripts of the elements were determined by dividing the elemental composition in the TS of the substrate with the corresponding atomic weight of the element [3].

$$BM{P}_{th}={\displaystyle \frac{22400 \left(\frac{n}{2}+\frac{a}{8}-\frac{b}{4}-\frac{c}{8}-\frac{d}{4}\right)}{12.017n+1.0079a+15.999b+14.0067c+32.065d}}$$

(4)

The ratio of the SMY to the BMP_th expressed in percentage is known as the biodegradability index (BDI) and can be calculated using Equation (5) [47].

$$BDI \left(\%\right)=\left({\displaystyle \frac{SMY}{BM{P}_{th}}} \right)\times 100$$

(5)

The experimental biodegradability or VS_R % is the percentage of VS degraded in the assay as a result of the AD process. Thus, the VS_R % can be determined using Equation (6) as given below [46].

$$V{S}_R \% \left(\%\right)=\left({\displaystyle \frac{V{S}_i-V{S}_f}{V{S}_i}} \right)\times 100$$

(6)

where VS_i and VS_f are the initial and final amount of VS in the BMP reactor before and after the incubation process.

2.5. Analytical Methods

The different physio-chemical properties of the substrate, liquid inoculum, solid inoculum and the digestates were determined according to the standard methods and protocols described elsewhere [48]. TS and VS were analysed according to the standard methods described elsewhere [49]. A Lachat Instruments USA, Quick Chem 8000 Flow Injection Analyser (FIA) was used to analyse phosphate (PO₄-P) and ammonium nitrogen (NH₄-N) as described elsewhere [50]. Total Kjeldahl Phosphorus (TKP) and Total Kjeldahl Nitrogen (TKN) were analysed using Inductivity Coupled Plasma-Optical Emission Spectroscopy (ICP-OES) installed with WinLab32 for ICP software version 2.2 (Perkin Elmer, Waltham, MA, USA Optima 7300 DV) according to the protocol published elsewhere [51]. A gas chromatograph (Agilent technologies 7890A, Santa Clara, CA, USA) equipped with a Flame Ionisation Detector (FID) and an Agilent DB-FFAP column was used to analyse VFAs [52]. Elemental composition (C, H, N, O, S) were determined as per the protocols published elsewhere [53]. The critical water content of SB was determined according to the drying test for the Water Absorption Index (WAI), determined according to the protocol published elsewhere [54]. Briefly, 1.25 g of SB was transferred to a 50 mL centrifugation tube and 15 mL of distilled water was added. The sample was mixed for 10 min at room temperature and centrifuged at 3000 rpm for 10 min. Upon centrifugation, the supernatant was discarded, and the sample was weighed and reported in g/gTS [55,56].

The over pressure in the headspace was measured by using a Keller Leo 2Ei (−1 to 30 bar) pressure transducer (Keller Druckmesstechnik AG, Winterthur, Switzerland). The composition of biogas was analysed by using a Gas Chromatograph, Shimadzu GC-2014 ATF, Japan with a 1 mL sample loop (Valco GC valve) for sample injection. The injector and oven temperature were set at 75 °C and 110 °C, respectively, during the process. However, a temperature of 120 °C was set for the GC gas detector. Argon gas of ultra-high purity (UHP) at a flow rate of 27.5 mL/min and 723.8 kPa was used as the carrier gas. External gas samples obtained from British Oxygen Company (BOC) (North Ryde NSW, Australia) were used to calibrate the GC at three different concentration levels.

Specific biogas yield (SBY) and SMY were calculated according to Equation (7) and Equation (8), respectively. Biogas/methane production from blank (mL) was subtracted from the biogas/methane production (mL) from assays in STP conditions to obtain the net biogas/methane produced from the substrate alone [17,57].

$$SBY \left({\displaystyle \frac{Nml}{gV{S}_{added}}}\right)={\displaystyle \frac{mL bioga{s}_{sample assay}-mL bioga{s}_{blank}}{V{S}_{added}}}$$

(7)

$$SMY \left({\displaystyle \frac{NmlC{H}_4}{gV{S}_{added}}}\right)={\displaystyle \frac{mL methan{e}_{sample assay}-mL methan{e}_{blank}}{V{S}_{added}}}$$

(8)

where VS_added is the substrate VS in the assay.

2.6. Statistical Analysis

The results of each treatment considered were the average values from the three replicates for each treatment ± standard error from their mean. SMY for T1–T5 and T6–T10 were assessed for significance using one-way analysis of variance (ANOVA) or Tukey’s post hoc test (t-test) at a 0.05 significance level using IBM SPSS Statistics 29^® software.

3. Results and Discussion

3.1. Chemical Composition of Substrate and Inoculum

The chemical compositions of SB, LI and SI are presented in

Table 2. Chemical composition of sugarcane bagasse (SB), liquid faction of inoculum (LI) and solid faction of inoculum (SI).

Parameter	SB	LI	SI
TS (%ww)	80.9	2.8	20.4
VS (%ww)	72.3	2.0	14.6
VS/TS	0.9	0.7	0.7
Density (g/mL)	0.2	1.0	1.0
TVFA (g/L)	NA	0.2	0.6
WAI (g/gTS)	10.5	NA	NA
TKP (gP/kgTS)	0.2	27.7	24.7
TKN (gN/kgTS)	2.1	60.4	53.5

NA: Not analysed.

. The results show that the TS contents of SB, LI and SI were 80.9%, 2.8% and 20.4% wet weight (w/w), respectively. The corresponding VS values were 72.3%, 2.0% and 14.6% w/w, respectively. The higher TS content and VS/TS ratio of SB suggests that it is an ideal substrate to study the effects of organic load in a batch AD process [17]. Moreover, the higher TS content of the SI helps in regulating the overall TS of the reactor working volume. The density of SB was found to be 0.2 g/mL and was accounted for while calculating the headspace of assays. A higher WAI of 10.5 g/gTS was obtained for SB, indicating the water absorbing capability of SB. The concentrations of Total Kjeldahl Nitrogen and Phosphorous for SB, LI and SI were 2.1, 60.4 and 53.5 and 0.2, 27.7 and 24.7 g/kgTS, respectively. The elemental composition of SB is as given in

Table 3. Elemental composition of sugarcane bagasse (SB).

Parameter	SB
Carbon (%TS)	44.5
Hydrogen (%TS)	5.7
Nitrogen (%TS)	0.5
Oxygen (%TS)	44.4
Sulphur (%TS)	0.1

and was used to calculate the BMP_th of SB.

3.2. Effect of Organic Load on SMY of SB

The effect of organic load on SMY during the mesophilic batch AD of SB is presented in Figure 1 and Figure 2. In both the experiments, a decline in SMY was observed with the increase in organic load. In experiment 1 (T1–T5), where the organic load was increased from 19.9 gVS/L to 106.1 gVS/L, an SMY of 59.27 NmLCH₄/gVS_added (p < 0.05) was obtained for T1 (F/I = 0.3, organic load = 19.9 gVS/L). However, for T2 (F/I = 0.5, organic load = 28.5 gVS/L) and T3 (F/I = 1, organic load = 50.0 gVS/L), decreases of 29% and 59% were observed in SMY (p < 0.05). Treatments T4 (F/I = 2, organic load = 80.3 gVS/L) and T5 (F/I = 3, organic load = 106.1 gVS/L) were completely inhibited and their methane production was lower than that of the blank.

Similar results were also reported by many researchers at higher F/I ratios tested [36]. For instance, Jiang et al. reported the complete inhibition of 75:25 and 100:0 feed ratios at F/I = 3 (organic load = 155.1 and 163.0 gVS/L) and 100:0 feed ratio at F/I = 1 (organic load = 99.5 gVS/L) during the dry codigestion (TS% of working volume = 20%) of food waste with pig manure [43]. In this study, the inhibition in assays at F/I > 1 could be due to the unfavourable conditions existing in these assays, such as an imbalance between the rates of acidogenesis and methanogenesis, acidification by substrate overload, and low free water content leading to reduced mass transfer in the assays [58,59]. Further, the high TS content (high organic load) coupled with the lower amount of available free water prevented the dilution of potential inhibitors such as VFAs [58,60]. Finally, the rates of hydrolysis were low as the amount of substrate VS added to the assays were higher than that of the inoculum VS [17,61]. Thus, batch AD systems with higher organic load require larger quantities of inoculum during start-up in order to overcome inhibition due to metabolic intermediates [62]. Higher F/I ratios such as F/I = 2.0 and F/I = 3.0 resulted in larger amounts of substrates in the reactor than the inoculum. This relative abundance of feed compared to the microorganisms resulted in process imbalance and thereby the accumulation of intermediate VFAs [63,64]. Conversely, methane production started immediately without any significant lag phase at T1 and T2, and after a lag phase of 7 days for T3 (Figure 1). However, higher methane production rates were noticed for T1 and T2 compared to T3. These results were similar to the previous studies reported for the organic fraction of municipal solid wastes (OFMSW) by Dastyar et al., where OFMSW was subjected to mesophilic high-solid AD with leachate recirculation at different F/I ratios such as 1.0, 2.0 and 3.0 (organic loads of 84, 122 and 144 gVS/L, respectively), of which F/I = 1.0 had 30% and 156% more biomethane yield than F/I = 2.0 and 3.0, respectively [21]. Therefore, a maximum F/I ratio of 1.0 should be maintained during the AD of lignocellulosic substrates such as the SB in higher TS levels of the reactor working volume to prevent process inhibitions.

In experiment 2 (T6–T10), the organic load was increased from 13.6 gVS/L to 110.0 gVS/L. The highest SMY of 57.5 NmlCH₄/gVS_added (p < 0.05) was observed for T6 (TS% = 3%, organic load = 13.58 gVS/L). With the increase in organic load in T7 (TS% = 8%, organic load = 40.3 gVS/L), T8 (TS% = 12%, organic load = 65.2 gVS/L), T9 (TS% = 14%, organic load = 88.4 gVS/L) and T10 (TS% = 16%, organic load = 110.0 gVS/L), the SMY of the assays was decreased by 52%, 74%, 85% and 88% (p < 0.05), respectively.

A decrease in SMY with an increase in organic load in terms of the increased TS% of the reactor working volume was also reported by Abid et al. [23] in their studies, where cow manure was subjected to AD at different TS levels from 5% to 20%. The highest SMY of 422.95 mL/gVS was obtained for an assay with 5%TS. The SMY reduced by 17%, 30% and 73% when the working volume TS was increased by 5%, 10% and 15% [23]. In a similar study by Yan et al. [31], where a codigestion mixture of maize stover and dairy manure was subjected to AD at different TS levels, a 347% reduction in SMY was reported when the TS of the reactor was increased from 6% to 16% [31]. The decrease in SMY at higher organic loads in terms of the TS% of the reactor working volume could be due to the process imbalance and formation of metabolic intermediates due to the reduced moisture content in those assays [23,31].

Free water availability (FWA) in each treatment was calculated using the WAI for SB (Figure 3). It was observed that both SBY and SMY decreased as the FWA in the treatments decreased. A negative value for FWA denotes the deficit of free water in those assays. Interestingly, it was noted that for all those treatments with negative FWA, the SMY was either zero or negligible due to possible inhibition in those assays. An increase in organic load results in a higher quantity of SB, which in turn will absorb more water from the reactor working volume, resulting in the reduced availability of moisture for metabolic activity. The experimental results show that those treatments with a negative value for FWA were inhibited, possibly due to mass transfer limitations and the accumulation of inhibitory compounds [65,66]. Similar findings were reported by Bollon et al., where the diffusion coefficient was decreased by factors of 50 and 185 when the TS% of the working volume was increased to 8% and 25% during the dry mesophilic AD of MSW [59]. In another study by Motte et al., where wheat straw was subjected to AD at different TS levels (10%TS to 33%TS), he observed that the lack of free water available at higher TS levels (higher organic load) affects microbial activity and microbial metabolism [67].

3.3. Effect of Organic Load on Chemical Composition of Digestates

Table 4. Chemical compositions of digestate among various tested treatments.

Parameter	T1	T2	T3	T4	T5	T6	T7	T8	T9	T10
TS (%w/w)	8.93	9.18	8.13	10.06	10.49	3.02	8.83	16.45	21.69	26.97
VS (%w/w)	6.02	6.31	5.74	8.22	8.97	2.30	6.57	12.53	17.13	21.36
VS/TS	0.67	0.69	0.71	0.82	0.85	0.76	0.74	0.76	0.79	0.79
pH	7.95	7.92	7.81	5.19	5.25	7.66	7.79	8.22	8.40	8.47
NO2-N (mg/L)	0.00	0.00	0.00	0.00	0.00	10.10	5.06	11.80	9.94	5.03
NO3-N (mg/L)	253.05	249.58	340.47	327.39	287.91	372.12	166.14	440.43	299.78	430.74
NH4-N (mg/L)	2464.08	2387.20	2231.95	1882.52	1372.48	1009.74	1961.03	3201.83	3578.28	4071.59
PO4-P (mg/L)	158.16	172.25	127.68	854.86	705.96	148.32	128.92	105.77	120.27	114.47
TKP (gP/kgTS)	34.00	31.95	27.75	18.80	14.05	17.42	15.33	14.61	13.73	12.37
TKN (gN/kgTS)	55.00	51.00	46.75	32.10	23.40	28.83	26.56	24.67	23.58	21.85
Total VFA (g/L)	0.12	0.17	0.20	21.10	17.69	0.17	0.08	0.22	0.27	0.30
Acetic acid (mg/L)	36.89	49.07	93.17	8571.13	5396.51	91.09	40.22	118.29	165.12	196.64
Propionic acid (mg/L)	11.48	21.58	13.41	931.04	956.30	14.77	6.99	16.06	15.23	14.73
Iso-Butyric acid (mg/L)	0.00	0.00	0.00	334.96	202.66	0.00	0.00	0.00	0.00	0.00
n-Butyric acid (mg/L)	13.92	28.19	15.19	7914.72	7067.40	17.48	8.17	21.62	20.69	20.49
Iso-Valeric acid (mg/L)	9.40	9.75	3.43	512.34	375.65	0.00	2.01	4.79	4.41	4.24
n-Valeric acid (mg/L)	14.96	13.57	10.73	234.56	395.56	13.27	5.45	16.60	14.20	12.21
Iso-Caproic acid (mg/L)	10.79	13.57	19.20	23.41	0.00	29.46	12.91	35.01	36.73	37.39
n-Caproic acid (mg/L)	19.83	34.80	40.49	2573.41	3300.39	7.34	4.14	11.33	10.87	9.60
Total Alcohol (mg/L)	40.71	46.64	69.76	58.98	374.09	0.00	11.62	13.65	2.73	0.00
Ethanol (mg/L)	40.71	44.90	69.76	58.98	297.55	0.00	11.62	13.65	2.73	0.00
Propanol (mg/L)	0.00	0.00	0.00	0.00	0.00	0.00	0.00	0.00	0.00	0.00
Butanol (mg/L)	0.00	0.00	0.00	0.00	58.18	0.00	0.00	0.00	0.00	0.00
1-Hexanol (mg/L)	0.00	1.74	0.00	0.00	18.35	0.00	0.00	0.00	0.00	0.00

presents the chemical composition of digestates at the end of the experimental run. An increase in VS/TS ratios when the organic load increased was observed in both the experiments, indicating the higher degradation of organic matter at lower organic loads [17]. Both T4 (F/I = 2) and T5 (F/I = 3) digestates have an acidic pH of 5.25 and 5.19, indicating VFA accumulation and thereby process inhibition in these assays (Figure 4). For instance, Wang et al., 2023, reported that a pH value less than 6.5 during hydrolysis can negatively impact the methanogenic archaeal activity or even inhibit the anaerobic digestion process completely [36] in their review of the dry anaerobic digestion of organic waste where higher organic loads of the reactors were tested. Also, Rocamora et al. reported that there are higher chances of process inhibition when the rate of VFA formation as a result of hydrolysis is faster than the rate of methanogenesis, which can lead to pH drops and thereby an inhibition of the methanogenic archaea [41]. Among the tested treatments, a decrease in TKN values was observed with the increase in organic load for both the experiments. It decreased from 55 gN/kgTS (F/I = 0.3) to 23.4 gN/kgTS (F/I = 3.0) for the first experiment, whereas for the second experiment it decreased from 28.83 gN/kgTS (TS = 3%) to 21.85 gN/kgTS (TS% = 16%). On the other hand, the NH₄-N values decreased from 2664.08 mg/L (F/I = 0.3) to 1372.48 mg/L (F/I = 3.0) for the first experiment (T1–T5) but increased from 1009.74 mg/L (TS = 3%) to 4071.59 mg/L (TS = 16%) for the second experiment (T6–T10). Ammonium inhibition was evident in T8 (TS = 12%), T9 (TS = 14%) and T10 (TS = 16%), where higher NH₄-N values of 3201.83, 3578.28 and 4071.59 mg/L were observed [68,69,70]. Benabdallah et al. also reported ammonia inhibition at higher organic loads based on TS levels in their study on the organic fractions of municipal solid waste [70]. Yan et al. reported that higher NH₄-N concentrations (>2.5 g/L) can result in the inhibition of the AD process as methanogens are the least tolerant to ammonia inhibitions [31]. Moreover, reduced moisture content in assays with a higher TS% of reactor working volume limits the dilution of inhibitory compounds such as NH₄-N, resulting in localised accumulation and thereby process inhibition [59].

Total VFA and individual VFA concentrations are presented in Figure 4. The results show that VFA production profiles varied and were dependent on organic load, pH, hydrolysis rate and subsequent production and conversion rates of individual VFAs. Among the tested treatments, T4 and T5 had the highest TVFA concentration and also showed a similar VFA production profile. The TVFA concentrations in T4 (F/I = 2.0) and T5 (F/I = 3.0) were 21.10 and 17.69 g/L, respectively, with 70–80% of total VFA accounting for butyric acid and acetic acid. Similar VFA concentration levels and profiles were also reported for an F/I ratio of 3.0 during the AD of food waste at higher organic loads [35]. The possible reason for the high TVFA noticed at both F/I = 2.0 and 3.0 was attributed to the larger amount of organic load from the substrate in the assays than that of the inoculum. Thus, higher organic loads along with lower moisture content resulted in VFA accumulation and hence led to process inhibition [32]. Xu et al. also reported similar VFA inhibitions in their study where different F/I ratios (2.0, 4.0 and 6.0) were tested during the codigestion of expired dog food with corn stover at a higher organic load setting (TS% = 22%) [27]. The high levels of butyric and acetic acids along with the high concentrations of alcohols noticed in T4 and T5 (Table 4) indicate the imbalance between the acidogenesis and methanogens. Further, the increased TS content of the reactor working volume reduced the mass transfer due to lower free water in the system, thereby preventing the dilution of potential inhibitors (i.e., NH₄–N or VFAs) compared to in wet AD [58]. Moreover, the generation of butyric acid also played an important role in the accumulation of caproic acid (Figure 4).

3.4. Effect of Organic Load on Biodegradability Index and VS Removal

BMP_th, BDI and VS_R were calculated using Equations (4)–(6). The biodegradability index (BDI) decreased with the increase in organic load in both the experiments owing to the reduction in methane production with the increase in F/I ratio and TS% of the working volume (

Table 5. BDI and VS_R for various treatments.

Treatments	BMPth	SMY	BDI	VSi	VSf	VSR %
	(NmLCH4/gVSadded)	(NmLCH4/gVSadded)	(%)	(gVS)	(gVS)	(%)
T1	437.64	59.27	13.54	8.81	6.67	24.32
T2	437.64	41.95	9.59	9.92	7.33	26.12
T3	437.64	24.05	5.49	13.22	7.59	42.57
T4	437.64	0.00	0.00	9.92	6.76	31.82
T5	437.64	0.00	0.00	14.33	9.31	35.03
T6	437.64	57.47	13.13	2.77	2.34	15.31
T7	437.64	27.87	6.37	8.54	6.96	18.55
T8	437.64	15.09	3.45	14.34	13.77	3.98
T9	437.64	8.73	1.99	20.14	19.52	3.11
T10	437.64	7.06	1.61	25.94	25.19	2.89

). In experiment 1, the maximum BDI observed was for T1 (F/I = 0.3) and was only 13.54%, whereas in experiment 2, a maximum BDI of only 13.13% was achieved (T6; TS = 3%). The lower BDI among all these treatments may be due to the increased organic load coupled with the reduced moisture content in these assays, resulting in slower hydrolysis and lesser direct interspecies electron transfer than in the conventional wet AD [44,71,72,73].

The percentage of VS removed (VS_R %) is the percentage of VS destroyed or removed as a result of the AD process [46]. Table 5 presents the VS_R calculated based on the initial VS of the assays (VS_i) and the VS of the digestate (VS_f). In experiment 1, VS_R increased with an increase in F/I ratio until T3 (F/I = 1). T4 and T5, which were inhibited due to VFA accumulation, had VS_R values of 31.82% and 35.03%, respectively, as the hydrolysis phase in these treatments successfully converted a portion of the available VS to VFAs [40]. A similar trend in VS_R with an increase in F/I ratio was also reported by Jiang et al., where different feed ratios of food waste and pig manure were subjected to dry codigestion at F/I ratios of 1.0 and 3.0 [43]. The VS removal of a 25:75 feed ratio of food waste to pig manure showed an increase from 24% to 40.2% when the F/I ratio increased from 1.0 to 3.0 in their studies [43]. In experiment 2, VS_R decreased from T6 to T10, indicating lower VS reduction with the increase in organic load based on the TS% of the working volume. Similar results were reported by Yan et al. [31], where maize stover and diary manure were codigested at different organic loads (TS levels). In this study, VS_R reduced from 77.9% to 17.1% when the TS of the working volume was increased from 6% to 16% [31].

3.5. Effect of Organic Load on Methane Production Kinetics

The first-order kinetic model and modified Gompertz model were used to predict the methane yields and to evaluate the methane production kinetics of the different treatments tested in this study. Treatments with no methane production (T4 and T5) were not analysed. Equations (1) and (2) were used to predict the methane yield on the t^th day for the first-order kinetic model and the modified Gompertz model, respectively. The diff % in

Table 6. Modelling parameters for the first-order kinetic model and the modified Gompertz model for the different treatments tested.

First-Order Kinetic Model
	SMY	Bo	Diff %	khyd	Tdelay	T90	Tef	rRMSE	R2
	(NmLCH4/gVSadded)	(NmLCH4/gVSadded)	(%)	(d−1)	(d)	(d)	(d)	(%)
T1	59.27	56.69	−4.37	0.09	1.02	39.44	38.41	5.23	0.9958
T2	41.95	40.94	−2.48	0.08	1.67	40.24	38.57	2.67	0.9981
T3	24.05	24.36	−0.12	0.05	5.55	49.16	43.61	1.00	0.9992
T4	0.00	NA
T5	0.00
T6	57.47	57.82	−0.75	0.05	2.13	51.35	49.22	3.63	0.9980
T7	27.87	27.09	−4.48	0.05	3.77	55.69	51.92	1.36	0.9959
T8	15.09	14.71	−3.60	0.05	2.37	60.13	57.76	3.26	0.9926
T9	8.73	8.55	−3.53	0.05	3.24	60.22	56.78	0.61	0.9975
T10	7.06	7.17	−3.31	0.03	1.97	63.37	61.40	0.90	0.9907
Modified Gompertz Model
	SMY	Bo	Diff %	Rmax			λ	rRMSE	R2
	(NmLCH4/gVSadded)	(NmLCH4/gVSadded)	(%)	(NmLCH4/gVSadded.day)			(d)	(%)
T1	59.27	54.98	7.22	3.41			0.51	10.51	0.9829
T2	41.95	39.73	5.31	2.20			2.91	5.26	0.9928
T3	24.05	22.99	4.47	0.82			4.95	3.43	0.9908
T4	0.00	NA
T5	0.00
T6	57.47	54.08	5.93	1.97			1.69	8.96	0.9875
T7	27.87	24.92	10.61	0.90			2.06	6.10	0.9739
T8	15.09	13.77	−3.60	0.54			3.26	2.64	0.9846
T9	8.73	7.95	8.93	0.29			2.82	0.61	0.9838
T10	7.06	6.43	9.34	0.17			1.34	1.69	0.9669

NA: Not available; T₉₀: time taken for producing 90% of the maximum CH₄ yield; T_ef: effective CH₄ production time (T₉₀–T_delay); Diff%: percentage difference in the predicted yield from SMY at the end of the experimental run.

denotes the percentage difference in final experimental SMY to the final predicted yields at the end of the experimental run. Based on the diff %, the first-order kinetic model was found to be the best fit for the experimental data in both the experiments (T1–T5 and T6–T10). The modelling parameters along with the SMY obtained for different treatments tested are as shown in Table 6.

Figure 5 represents the curve fitting of the predicted models to the experimental SMY among the treatments. As the experiments were conducted in triplicate, the minimum and maximum deviation deviations from the mean values are reported as vertical error bars. The higher R² values and lower rRMSE values obtained for the first-order kinetic model proved that the first-order kinetic model was the best fit for experimental methane yields when compared with the modified Gompertz model.

The hydrolysis constant (k_hyd) refers to the rate of feed digestion by the microbes before they become inactive [48]. The k_hyd values for the first experiment according to the first-order kinetic equation were estimated to be 0.09, 0.08 and 0.05 d⁻¹, respectively, for T1, T2 and T3. On the other hand, the k_hyd values of the second experiment were constant (0.05 d⁻¹) except for T10, where it slightly reduced to 0.03 d⁻¹. The T₉₀ (time taken to obtain 90% of the total methane yield) of the different treatments according to the first-order kinetic equation was found to increase with the increase in organic load for both the experiments. This might be due to the VFA or ammonia inhibitions experienced among the higher organic loads of these tested treatments [31,40]. Table 5 also shows the effective methane production time (T_ef) calculated by subtracting T_delay from T₉₀. Similar to the T₉₀, T_ef also showed a similar trend with the increase in organic load. It increased from 38.41 days (T1, F/I = 0.3) to 43.61 days (T3, F/I = 1.0) for experiment 1 and 49.22 days (T6, TS = 3%) to 61.40 days (T10, TS = 16%) for experiment 2. An imbalance in process parameters as a result of increased organic load leading to the formation of metabolic intermediates resulted in an increased effective time for high organic loads [31,40]. For instance, Li et al. [74] also reported similar VFA accumulation and inhibition when the organic load was increased as a result of an increase in F/I ratio during the AD of food waste with waste cooking oil [74]. Similarly, Yi et al. [75] also reported higher amounts of metabolite intermediate formation at higher organic loads in terms of increased TS content when food waste was subjected to AD at different TS levels [75].

For experiment 1, the modified Gompertz model predicted a longer lag phase for higher organic loads in terms of F/I ratio. A lag phase of 4.95 days was predicted for T3 (F/I = 1.0), whereas lag phases of only 0.51 days and 2.91 days were predicted for T1 (F/I = 0.3) and T2 (F/I = 0.5), respectively. A similar trend was observed for the studies by Jiang et al., where assays with an F/I ratio of 1.0 had a shorter lag phase of 9 days compared to the 36-day lag phase for assays with an F/I ratio of 3.0 in the 25:75 feed ratio during the dry codigestion of food waste with pig manure [43]. The R_max values of T1 (3.41 NmLCH₄/gVS_added.day) and T2 (2.20 NmLCH₄/gVS_added.day) were higher compared to that of T3 (0.82 NmLCH₄/gVS_added.day) in the modified Gompertz model, which was justified by a higher methane yield in T1 and T2 compared to that of T3. Jiang et al. also reported higher R_max values for lower F/I ratios in their studies on the dry codigestion of food waste with pig manure [43]. For experiment 2, the predicted lag phase according to the modified Gompertz model increased with the increase in organic load for the uninhibited assays. Lag phases of 2.06 days and 1.69 days were predicted for T7 (TS = 8%) and T6 (TS = 3%), respectively. The shorter lag phase in assays with lesser organic load in terms of TS% implies a faster degradation of complex organics in these assays [23]. A similar trend was observed among the R_max values of these assays as the R_max values decreased with the increase in organic load. It reduced from 1.97 NmlCH₄/gVS_added.day for T6 (TS = 3%) to 0.17 NmlCH₄/gVS_added.day for T10 (TS = 16%). This was mainly due to the slower rate of production in assays with higher organic loads in terms of TS content, as the reduced moisture level would have resulted in lesser interspecies electron transfer in these assays [73]. A similar trend in R_max values was reported by Yan et al. [31] in their studies where maize stover and dairy manure were codigested at different organic loads in terms of the TS% of the reactor. In their studies, the R_max value varied from 17.95 to 2.81 NmlCH₄/gVS_added.day when the TS of the reactor was increased from 6% to 16% [31].

Thus, during the AD of SB, operating the reactor at a higher organic load of 50 gVS/L is found to be more beneficial as it can digest more SB than the conventional low-organic-load AD of SB using the same reactor working volume without decreasing the total methane production. Digesting more SB using the same reactor volume will significantly improve the process economics, as reactor sizing has a significant impact on the capital cost of an AD plant. It will also result in a better carbon footprint of the sugar processing industry as extracting energy through bioprocess technology such as AD and using it for process requirements will reduce the emissions associated with burning these agro-wastes as fuel for boilers to produce process steam. As yield decreases at higher organic loads, several yield-improving techniques such as the pretreatment of substrates or codigestion with animal manure can be tested at this higher organic load to improve the yield and thereby the volumetric productivity.

4. Conclusions

The results from this study prove that, for SB, increasing the organic load of a batch AD reactor based on the F/I ratio or TS content of the reactor working volume has profound effects on its SMYs and methane production rates. A decrease in SMY and FWA with an increase in the organic load of the reactor was evident in both the experiments. A lower FWA for SB at higher organic loads adversely affected the process parameters. The results from the first experiment (T1–T5) showed that higher organic loads tested, such as F/I = 2.0 and F/I = 3.0 with organic loads of 80.3 and 106.1 gVS/L, respectively, were completely inhibited due to VFA accumulation. The increase in organic load based on an increase in the TS content of the reactor at a constant F/I ratio also showed a significant decrease in SMY. The results show that at an F/I ratio of 1.0, higher organic loads such as 65.2 (T8; TS = 12%), 88.4 (T9; TS = 14%) and 110.0 gVS/L (T10; TS = 16%) were inhibited due to higher NH₄-N concentrations in these assays. Thus, the results from these studies suggest that the maximum possible organic load for an AD reactor digesting SB at mesophilic temperature to function without process inhibition is 50 gVS/L (T3; F/I = 1.0; TS = 10%). The results from the kinetic study suggested that, for both the experiments (T1–T5 and T6–T10), the first-order kinetic model is the best fit for the experimental SMY when compared with the modified Gompertz model.