Graphene Facilitates Biomethane Production from Protein-Derived Glycine in Anaerobic Digestion.

Interspecies electron transfer is a fundamental factor determining the efficiency of anaerobic digestion (AD), which involves syntrophy between fermentative bacteria and methanogens. Direct interspecies electron transfer (DIET) induced by conductive materials can optimize this process offering a significant improvement over indirect electron transfer. Herein, conductive graphene was applied in the AD of protein-derived glycine to establish DIET. The electron-producing reaction via DIET is thermodynamically more favorable and exhibits a more negative Gibbs free energy value (-60.0 kJ/mol) than indirect hydrogen transfer (-33.4 kJ/mol). The Gompertz model indicated that the kinetic parameters exhibited linear correlations with graphene addition from 0.25 to 1.0 g/L, leading to the highest increase in peak biomethane production rate of 28%. Sedimentibacter (7.8% in abundance) and archaea Methanobacterium (71.1%) and Methanosarcina (11.3%) might be responsible for DIET. This research can open up DIET to a range of protein-rich substrates, such as algae.

Introduction

The European Union (EU) has committed to achieving at least 20% renewable energy share of gross energy consumption by 2020, rising to at least 27% by 2030 (European Commission, 2016). Transport is the most difficult sector to decarbonize. This is reflected in the relatively low targets set by the EU. The target for advanced biofuels, including biogas not sourced from food crops (such as wastes and algae), has been set at a minimum share of 3.6% of the total fuel consumption by 2030; the target for 2021 is just 0.5% (European Commission, 2016). Anaerobic digestion (AD) is a well-established technology for waste management and production of renewable energy in which wet organic materials are broken down and converted into biogas by a variety of microorganisms, primarily including acidogenic bacteria and methanogenic archaea. In addition to biogas production from AD, other value-added chemicals (such as medium-chain carboxylic acids), can also be produced through the anaerobic process (Xu et al., 2018). In recent years, a rapid growth in newly installed biomethane plants has been observed. The European Biogas Association reported that the biomethane plants in the EU in 2015 were capable of producing 1,230 M Nm3 of biomethane (European Biogas Association, 2016). This is readily available as transport fuel in natural gas vehicles. The market for biomethane, especially in the transport sector, will be driven by policy and renewable targets. Beyond the EU, the International Energy Agency (IEA) has suggested that biomethane for transport should rise to 3.74 EJ by 2040 from a base of 45.5 TJ in 2015 (International Energy Agency (IEA) Energy Technology Perspectives, 2017, European Biogas Association, 2016). For this remarkable rise in biomethane production to be achieved using the world's finite resources, AD efficiency must be optimized, and challenging novel abundant feedstocks (such as algae) must be assessed for biomethane production.

Many studies and reviews are available that detail technology advantages and disadvantages, diverse microbial pathways, and challenges involved in AD (FitzGerald et al., 2015, Xia et al., 2016, Batstone and Virdis, 2014, Rotaru et al., 2014b). Studies have shown that digestion of complex feedstocks is not entirely efficient. Protein-rich feedstocks (such as algae) may be difficult to handle in AD owing to the presence of unfavorable protein components and unbalanced carbon to nitrogen ratios, which together may lead to process instability, sub-optimal digestion, and, ultimately, failure. If the resource of biomethane is projected to increase as suggested by the IEA, then difficult feedstocks such as algae with a huge biomethane resource must be digested as efficiently as possible. Thus far, the efficiency of digestion of protein-rich feedstocks reported varies from 20% to 70% (Mei et al., 2016). In general, for slowly biodegradable substrates (such as polysaccharides and proteins), hydrolysis is the rate-limiting step (Ariunbaatar et al., 2014). However, when the substrates are easily degradable organics, such as monosaccharides or amino acids, the digestion process is limited by methanogenesis rather than by hydrolysis. Based on the current understanding of the microbiology of AD, the inefficiency and instability of digestion can be fundamentally optimized through the microbial interspecies electron transfer between acidogenic bacteria and methanogenic archaea.

The predominant understanding of interspecies electron transfer in AD is based on mediated interspecies electron transfer (MIET), in which hydrogen gas produced by acidogenic bacteria diffuses to hydrogenotrophic methanogens. However, from a thermodynamic perspective, the success of MIET is strongly restricted to the metabolite concentration, especially hydrogen. If the hydrogen partial pressure is high acidogenic bacteria are inhibited, whereas hydrogenotrophic methanogenic archaea by consuming hydrogen facilitate enhanced performance of the acidogenic bacteria. As an alternative, direct interspecies electron transfer (DIET) via biological pili or shuttle molecules is recognized as an improvement of this process, replacing hydrogen as the electron carrier between bacteria and archaea. Conductive materials might act as an effective electron conduit among microorganisms. Growing evidence suggests that the addition of conductive materials, such as ferroferric oxide, biochar, carbon nanotube, and graphene, can improve AD performance in terms of reducing lag phase time, increasing methane production rate, and resisting inhibitory conditions (Zhao et al., 2017, Lee et al., 2016, Yan et al., 2017, Yin et al., 2018).

Zhao et al. (Zhao et al., 2017) demonstrated that anaerobic digesters supplemented with carbon cloth had a higher capacity to resist the acidic impacts in mesophilic AD operating at 35°C. The microbial community analysis revealed that Geobacter and Methanosaeta species were closely attached to the surface of carbon cloth, suggesting that the predominant working mode for the interspecies electron exchange might shift from MIET to DIET (Zhao et al., 2017). Granular activated carbon has been reported to promote the enrichment of microbes utilizing a DIET process, thereby leading to an increase in methane production rate by 77.6% (Lee et al., 2016). Conductive materials were also shown to promote thermophilic AD (∼55°C), as evidenced by low volatile fatty acid (VFAs) accumulation, high biomethane production rate, and more robust responses against hydrogen inhibition (Yan et al., 2017). Renewable pyrogenic carbon has shown the capability of directly transferring electrons (Sun et al., 2017), suggesting a cost-effective application of DIET. To illustrate the dynamic advantages of DIET over MIET in AD, simplified models have been proposed to calculate the maximum potential of electric current (Cruz Viggi et al., 2014, Lin et al., 2017, Lin et al., 2018). Cruz Viggi et al. modeled the electron transfer flow in AD of propionate with the addition of magnetite particles (Cruz Viggi et al., 2014). Theoretical calculations revealed that DIET allows electrons to be transferred among syntrophic partners at rates that are substantially higher than those attained via MIET. Similarly, using ethanol as the model substrate, Lin et al. found that graphene-based DIET can sustain a much higher electric current flux, thereby allowing for more efficient electron transfer in a syntrophic mechanism (Lin et al., 2017).

Substrates such as ethanol, glucose, and VFAs have been investigated for establishing DIET with respect to experimental study and model development. The gap in the state of the art is the application of graphene in the promotion of digestion of protein-derived amino acids (such as glycine). This can open up DIET to a range of substrates such as algae, and potentially facilitate the growth of biomethane as a transport fuel. The degradation of glycine in AD is metabolically different from that of substrates that have previously been investigated (such as ethanol and propionate). Glycine is a type of amino acid that widely exists in protein-rich substrates (such as algae). When compared with ethanol/propionate, amino acid presents the amino group (–NH2), which results in a different metabolic pathway in AD. The degradation of glycine requires certain types of amino-acid-degrading bacteria, which can break down amino acid into acetic acid, ammonia, and CO2. Theoretically this process can lead to the generation of free electrons or hydrogen gas. However, it is unclear if these electrons can be directly used by other microorganisms through DIET. Previous research has demonstrated that various bacteria can degrade ethanol/propionate producing electrons, which can be further consumed by some methanogens. Given that glycine degradation in AD is metabolically different, such a process via DIET necessitates further investigation. Furthermore, the thermodynamic differences of glycine digestion through DIET/MIET remain unclear. The innovation in this article is that it substantially expands the potential range of substrates that could be syntrophically metabolized via DIET and demonstrates for the first time that the AD of amino acids (in this case glycine) could be facilitated by conductive graphene. This can advance the understanding on the theoretical maximum electron transfer via DIET/MIET, and the microbial and functional shift in response to graphene addition in the AD of glycine.

Results and Discussion

Effects of Graphene Addition on Biomethane Production

Conductive carbonaceous nanomaterials, such as graphene and carbon nanotube, have been assessed to improve the AD of small molecule substrates (such as glucose and ethanol) (Lin et al., 2017, Tian et al., 2017, Li et al., 2015). The metabolic pathways for microbial degradation of carbohydrates and proteins are fundamentally different. Carbohydrates and proteins are the main substrates contributing to biogas production in digestion. Carbohydrates are more favorable as a substrate for microbes than proteins in AD. Protein-derived amino acids may exhibit different metabolic characteristics and digestion performance in the presence of graphene when compared with carbohydrate-derived sugars.

The effects of graphene addition on biomethane yield and production rate of glycine are illustrated in Figures 1A and 1B, respectively. A biomethane yield of 189.3 mL/g was obtained after 16 days of digestion in the absence of graphene. There was a clear trend that illustrated that suitable additions of graphene could result in significant increases in biomethane yields. For example, the addition of 0.5 g/L graphene led to a significant increase (p < 0.05) in biomethane yield to 200.1 mL/g. However, further increasing graphene addition to 2.0 g/L adversely affected the biomethane yield, leading to a decrease to 170.0 mL/g. The higher concentration of graphene led to an inhibition effect on AD, suggesting that cytotoxicity could become a limiting factor when applying graphene nanomaterial in a microbial process. Cytotoxicity induced by several carbon nanomaterials (such as graphene and carbon nanotube) to different microorganisms has been observed in previous studies. For example, Liu et al. investigated the antibacterial effects of graphene-based nanomaterials (including graphene oxide and reduced graphene oxide) on the model bacteria Escherichia coli (Liu et al., 2011). The authors proposed a three-step antimicrobial mechanism, which includes initial cell deposition on graphene-based materials, membrane stress caused by direct contact with sharp nanosheets, and the ensuing superoxide anion-independent oxidation (Liu et al., 2011). Pasquini et al. demonstrated that surface functionalization of carbon nanotubes could lead to different antibacterial effects on Escherichia coli K12 (Pasquini et al., 2012). The size of nanomaterials is also a key factor that affects the antibacterial effects. Perreault et al. found that graphene-oxide-based surface coatings showed higher antimicrobial activity for smaller sizes (Perreault et al., 2015). It has also been reported that graphene can damage the bacterial cell membrane as a result of direct contact with the very sharp edges of the graphene nanosheets, resulting in an efflux of intracellular components such as RNA (Akhavan and Ghaderi, 2010). The plausible explanation of the antibacterial effect is possibly related to the synergistic impacts of cell membrane perturbation and oxidative stress, but the toxicological mechanisms at the molecular level remain unclear (Qu et al., 2015).

Figure 1

Effects of Graphene Addition on Biomethane Yield and Production Rate from Glycine.

(A) Biomethane yield and (B) biomethane production rate. Data are presented as mean ± standard deviation.

As shown in Figure 1B, the peak biomethane production rate in the absence of graphene was obtained at 12 days with a value of 21.8 mL/g/day. The presence of various graphene concentrations greatly reduced the time for peak production rate from 12 to 9 days. The highest peak biomethane production rate of 27.9 mL/g/day was achieved with the addition of 1.0 g/L graphene, corresponding to an enhancement of 28.0% when compared with the value without graphene addition. These results indicated that the suitable addition of conductive graphene could enable a much quicker digestion process, due to the potential establishment of DIET.

Effects of Graphene Addition on Glycine Conversion in AD

Glycine (C2H5NO2), as a typical C2 amino acid derived from protein-rich substrates (such as microalgae and seaweed), can be used as an electron-donating source for AD. The conversion of glycine in digestion requires the syntrophy of acidogenic bacteria and methanogenic archaea. Table 2 presents the metabolic pathways of glycine conversion and the associated Gibbs free energy changes (ΔG0′) under standard conditions. For the electron-producing reaction, either MIET or DIET can be employed by diverse acidogenic bacteria. In a similar way, for the electron-consuming reaction, either MIET or DIET can be employed by diverse methanogenic archaea. The connection established between bacteria and archaea contributes to an efficient electron transfer process. DIET can provide additional energy benefits to the syntrophic partners because there is no need to generate metabolites as electron carriers (such as hydrogen), and the subsequent diffusion of electron carrier is unnecessary. From the perspective of thermodynamic change, the ΔG0′ value of DIET in terms of the electron-donating process is more negative (−60.0 kJ/mol) when compared with that (−33.4 kJ/mol) of MIET, which indicates that the start-up process of glycine degradation via DIET is thermodynamically more favorable and spontaneous.

Table 2

Reactions and Changes in Gibbs Free Energy Values for Glycine Conversion to Methane with Different Pathways

Process	Reaction	ΔG0′ (kJ/mol)
Electron-producing reaction	1. MIET: C2H5NO2 + 2/3H2O → 2/3CH3COO− + 2/3H+ + NH3 + 2/3CO2 + 1/3H2	−33.4
	2. DIET: C2H5NO2 + 2/3H2O → 2/3CH3COO− + 4/3H+ + NH3 + 2/3CO2 + 2/3e-	−60.0
Electron-consuming reaction	3. MIET: 1/3H2 + 1/12CO2 → 1/12CH4 + 1/6H2O	−10.9
	4. DIET: 2/3H+ + 2/3e− + 1/12CO2 → 1/12CH4 + 1/6H2O	15.7
Acetate-consuming reaction	2/3CH3COO− + 2/3H+ → 2/3CH4 + 2/3CO2	−23.9
Overall	C2H5NO2 + 1/2H2O → 3/4CH4 + NH3 + 5/4CO2	−68.2

MIET, mediated interspecies electron transfer; DIET, direct interspecies electron transfer. Values are calculated at different temperatures under standard conditions (1 M concentration of all solutes, 25°C, 1 atm, and neutral pH). Negative value indicates that the reaction is thermodynamically favorable and proceeds spontaneously.

Graphene addition (1.0 g/L) enhanced glycine conversion during digestion (Figure 2A). As a result, the production of acetate from glycine was concomitantly accelerated (Figure 2B). The acetate concentration gradually increased with the degradation of glycine in digestion, with the peak acetate concentration of 0.89 mg/L observed at 8 days in the presence of 1.0 g/L graphene. In comparison, the peak acetate concentration of 0.93 mg/L was observed at 12 days of digestion in the absence of graphene. The results suggest that graphene was capable of promoting syntrophic reactions between acidogenic bacteria and methanogenic archaea. The enhanced syntrophy led to improved substrate degradation and utilization, namely, glycine degradation and acetate conversion. These results are in agreement with previous work, which demonstrated that the degradation of ethanol in AD was approximately 30% faster upon the addition of graphene (Lin et al., 2017).

Figure 2

Effects of Graphene Addition on Glycine and Acetate Conversion

(A) Glycine degradation and (B) acetate production and degradation. Data are presented as mean ± standard deviation.

Kinetic Analysis

To further evaluate the differences in digestion induced by graphene addition, three types of models were employed to simulate the kinetic patterns of biomethane production (first-order kinetic model, modified Gompertz model, and Cone model, see Table 1). The kinetic parameters, such as P, R, k, λ, and n were estimated by the studied models (Table 3). According to the results from the first-order kinetic model, there was a significant difference between the measured and predicted biomethane potential ranging from 34.5% to 171.5%, with relatively low correlation coefficients R ranging from 0.9138 to 0.9292. Nonetheless, the first-order kinetic model result suggested that the biomethane production rate constant increased with the addition of graphene from 0 to 1.0 g/L, and then decreased with 2.0 g/L graphene. The highest biomethane production rate constant of 0.0794 (L/d) was observed with the addition of 1.0 g/L graphene.

Table 1

Models Used to Describe Biomethane Production from Glycine With/Without Graphene Addition

Model	Equation	Parameters	N
First-order kinetic	P=P_m(1−exp(−kt))	Pm, k	2
Modified Gompertz	P=P_mexp{−exp[R_meP_m(λ−t)+1]}	Pm, Rm, λ	3
Cone	P=P_m1+(kt)^−n	Pm, k, n	3

P, maximum biomethane potential (mL/g); k, reaction rate constant (1/d); R, peak biomethane production rate (mL/g/day); λ, lag-phase time of biomethane production (d); n, shape factor; N, number of model parameter.

Table 3

Estimated Parameters from the First-Order Kinetic, Modified Gompertz, and Cone Models Simulating Biomethane Production from Glycine

Model	Parameter	Graphene Addition (g/L)
		0	0.25	0.5	1.0	2.0
First-order	Pmeasured (mL/g)	189.3	191.8	200.1	197.2	169.9
	Pm (mL/g)	513.9	332.6	303.7	265.3	256.0
	Difference (%)	171.5	73.4	51.8	34.5	50.7
	k (1/d)	0.0249	0.0482	0.0613	0.0794	0.0616
	Adjusted R2	0.9260	0.9138	0.9188	0.9185	0.9292
	RMSE (mL/g)	20.5	22.5	22.4	21.8	17.5
Cone	Pmeasured (mL/g)	189.3	191.8	200.1	197.2	169.9
	Pm (mL/g)	207.4	204.1	212.6	207.3	181.3
	Difference (%)	9.56	6.41	6.25	5.12	6.71
	k (1/d)	0.111	0.126	0.134	0.148	0.135
	n	3.44	3.69	3.49	3.47	3.25
	Adjusted R2	0.9879	0.9866	0.9870	0.9875	0.9907
	RMSE (mL/g)	8.28	8.87	8.96	8.52	6.35
Modified Gompertz	Pmeasured (mL/g)	189.3	191.8	200.1	197.2	169.9
	Pm (mL/g)	201.4	200.2	207.9	203.1	176.2
	Difference (%)	6.39	4.38	3.90	2.99	3.71
	Rm (mL/g/day)	20.3	23.8	25.1	26.6	20.4
	λ (d)	3.88	3.62	3.19	2.87	2.95
	Adjusted R2	0.9895	0.9882	0.989	0.990	0.9930
	RMSE (mL/g)	7.71	8.31	8.10	7.52	5.37

P, maximum biomethane potential (mL/g); k, reaction rate constant (1/d); R, peak biomethane production rate (mL/g/day); λ, lag-phase time of biomethane production (d); n, shape factor; N, number of model parameter; RMSE, root-mean-square prediction error.

Comparatively, both the Cone model and the modified Gompertz model exhibited a good fit with the measured data, as evidenced by the high correlation coefficients of R and low RMSE (root-mean-square prediction error) values. For the Cone model, the difference between the measured and predicted biomethane potential ranged from 5.12% to 9.56%. The model predicted that the highest biomethane yield of 207.3 mL/g (5.12% difference from the measured data) was obtained with the addition of 1.0 g/L graphene. Correspondingly, the highest obtained biomethane production rate constant was 0.148 day−1 when compared with 0.111 day−1 without graphene addition.

In agreement with the results from the Cone model, the modified Gompertz model predicted that the addition of 1.0 g/L graphene produced a biomethane potential of 203.1 mL/g (2.99% difference from the measured data). The peak biomethane production rate was accordingly increased from 20.3 mL/g/day to 26.6 mL/g/day (equivalent to a 31.0% increase). In addition, the Gompertz model gave a lag time of 2.87 days with the addition of 1.0 g/L graphene when compared with 3.88 days in the control condition. Based on the statistical indicators in terms of high correlation coefficient R (0.9895–0.9930) and low RMSE value (5.37–8.31), the modified Gompertz model demonstrated the best fit for the experimental data.

To evaluate the enhancing effects of graphene on AD, the relations of graphene additions (0–1.0 g/L) with different kinetic parameters derived from the modified Gompertz model are presented in Figure 3. A lack of correlation (R = 0.2987) was observed between graphene additions and biomethane yield potentials (Figure 3A). This can be explained by the fact that glycine was completely consumed after AD regardless of graphene addition, resulting in insignificant changes in the overall biomethane yield. Interestingly, apart from the biomethane yield potential, other parameters including the peak biomethane production rate, lag phase time, and peak time exhibited strong positive linear correlations with graphene addition. The best performance was achieved in the presence of 1.0 g/L graphene. Figure 3B shows that the peak biomethane production rate was positively (R = 0.7848) affected by increased graphene addition. Similarly, both lag-phase time (λ) and peak time (T) showed a positive linear correlation with graphene addition with R = 0.88337 and 0.96553, respectively (Figures 3C and 3D).

Figure 3

Correlations of Graphene Addition with Different Kinetic Parameters in Anaerobic Digestion

(A) Biomethane yield potential (P), (B) peak biomethane production rate (R), (C) lag phase time (λ), and (D) peak time (defined as T = P/R/e + λ).

Microbial Taxonomic Patterns in Response to Graphene Addition

In the AD of glycine, bacterial communities are responsible for converting glycine to acetic acid, carbon dioxide, and hydrogen, whereas archaeal communities are responsible for converting acetic acid or carbon dioxide to methane. To better understand the bacterial and archaeal responses to graphene addition, four groups of digestate samples, including the digestate without graphene addition and digestate with 0.25, 0.5, and 1.0 g/L graphene added were analyzed (Figures 4A and 4B).

Figure 4

Microbial Community Structures at Genus Level with/without Graphene Addition after Anaerobic Digestion of Glycine

(A) Bacterial and (B) archaeal communities. Genera with less than 1% abundances are classified as “others.”

In the samples without graphene addition, Levilinea (18.0%) and Aminobacterium (17.6%) were the dominant bacterial genera. It has been reported that Levilinea and Aminobacterium are capable of fermenting amino acids into VFAs, such as acetate and butyrate, along with the production of hydrogen gas (Yamada et al., 2006, Baena et al., 1998). Taking glycine as an example, the metabolic pathway is described as follows:C

The high abundance of amino-acid-degrading bacteria was ascribed to the adaptation effect of glycine in AD. When compared with the incubation without graphene addition, higher concentrations of graphene gradually decreased the abundance of Aminobacterium from 17.6% to 4.9%. However, the presence of 1.0 g/L graphene resulted in significant enrichment of Sedimentibacter from 4.7% to 7.8%. Interestingly, Sedimentibacter-related species can ferment glycine to acetic acid without production of hydrogen gas (Lechner, 2015), and the metabolic pathway can be described as follows:C

This suggests that the free electrons produced from glycine oxidation have to be consumed by other microorganisms, particularly the methanogens. This further implies the possibility of DIET between fermentative bacteria and methanogenic archaea. Sedimentibacter has also been identified in the core community for microbial fuel cell systems fed with swine manure, due to their exoelectrogenic role with high power generation capability (Vilajeliu-Pons et al., 2016).

The structure of the archaeal community in response to graphene addition is shown in Figure 4B. When graphene is absent, the genera of Methanobacterium and Methanosaeta are dominant in the archaeal community, accounting for 61.7% and 31.2%, respectively. Most species of Methanobacterium are recognized as hydrogen-consuming methanogens, converting carbon dioxide and hydrogen into methane. However, a previous study revealed that specific species of Methanobacterium can directly receive electrons for electro-methanogenesis, in which Methanobacterium palustre accounted for 86.7% of the total cells (Cheng et al., 2009). This finding suggests that methanogens can directly accept electrons, and builds further support for interspecies electron transfer. Methanosaeta are mainly acetate-consuming methanogens, which cleave acetic acid into methane and carbon dioxide. When increasing the graphene addition to 1.0 g/L in the digestion, the abundance of Methanosaeta decreased to 16.8%, compared with 31.2% without graphene addition. In contrast, the abundance of Methanosarcina gradually increased from 6.5% (without graphene addition) to 11.3% (1.0 g/L graphene addition). Methanosarcina are known as the most metabolically diverse methanogens. Recent studies have demonstrated that some species (such as Methanosarcina barkeri) are capable of directly receiving electrons from electroactive bacteria (such as Geobacter) for CO2 reduction (Rotaru et al., 2014a). The increase in abundance of Methanosarcina after graphene addition suggests their likely participation in the DIET process (2/3H+ + 2/3e− + 1/12CO2 → 1/12CH4 + 1/6H2O). In the present study, the enrichment of archaea Methanosarcina and Methanobacterium in response to graphene addition proposes a possibility that they may play an important role in conducting DIET.

Potential Microbial Network for Glycine Degradation in AD

AD is known to require multiple groups of microorganisms working together to convert organic substrates to methane. The complete degradation of glycine involves the cooperation of acidogenic bacteria and methanogenic archaea. Acidogenic bacteria are responsible for converting glycine to acetic acid and producing electrons (or hydrogen), whereas methanogenic archaea are capable of converting acetic acid, carbon dioxide, and electrons to methane through three major pathways, acetoclastic methanogenesis, hydrogenotrophic methanogenesis, and DIET methanogenesis.

Based on the differences in microbial community structure (Figure 4) and the metabolic functions of the bacterial genera present, a potential microbial network of the major bacteria and archaea involved in the AD of glycine is illustrated in Figure 5. In the AD of glycine in the absence of graphene, glycine is first degraded by acidogenic bacteria, which may include Levilinea (18.0%) and Aminobacterium (17.6%), as shown in Figure 5A. Glycine acidification may then lead to the production of intermediate products such as acetic acid and hydrogen (or protons and electrons). The further production of methane will then primarily depend on the acetoclastic and hydrogenotrophic pathways for methanogenesis, with acetic acid being utilized by Methanosaeta (31.2%) to produce methane and carbon dioxide. Carbon dioxide may then be reduced to methane by Methanobacterium (61.7%) and Methanosarcina (6.6%) through the hydrogenotrophic pathway.

Figure 5

Microbial Community Structures at Genus Level with/without Graphene Addition after Anaerobic Digestion of Glycine

(A) Without graphene addition and (B) with 1.0 g/L graphene addition. The numbers in brackets indicate the abundance of microorganisms.

The presence of conductive graphene in the digestion process appears to markedly change the syntrophic interactions between the acidogenic bacteria and methanogenic archaea (Figure 5B). Levilinea (17.8%) remained as the dominant acidogenic bacteria, whereas Sedimentibacter (7.8%) was enriched and is likely to have played an important role in degrading glycine. As previously mentioned, Sedimentibacter are capable of converting glycine to acetic acid and free electrons. Archaea Methanobacterium (61.7%) and Methanosarcina (11.3%) are capable of utilizing free electrons for carbon dioxide reduction. Therefore, the enrichment of these microorganisms in methanogenesis proposes a possibility that they may play an important role in conducting DIET.

Comparison of Maximum Electron Transfer for MIET and Graphene-Based DIET

Microbial electron transfer between electron-producing acidogens and electron-consuming methanogens in AD might proceed via either DIET or MIET (with hydrogen as an electron carrier). To compare the electron transfer efficiencies of DIET and MIET, previous models have been proposed based on the Nernst equation and Fick's diffusion law (Cruz Viggi et al., 2014, Lin et al., 2017, Lin et al., 2018). Using ethanol as the substrate in AD, Lin et al. found that graphene-based DIET can sustain a much higher electric current flux than hydrogen transfer, thereby allowing for more efficient electron transfer via the syntrophic mechanism (Lin et al., 2017). Substrates such as ethanol, glucose, and VFAs have been investigated for establishing DIET in AD. The present study increased the range of substrates known to be syntrophically metabolized via DIET and demonstrated for the first time that the AD of amino acid (glycine) could be facilitated by conductive graphene. The degradation of glycine in AD was metabolically different than those of the previous substrates investigated, such as ethanol and propionate. Therefore, to further reveal the difference of maximum electron transfer achieved by DIET and MIET, the simplified calculations were proposed based on the Nernst equation and Fick's diffusion law.

The maximum electron transfer for MIET in AD of glycine was calculated based on Fick's diffusion law (see Figure S2, Supplemental Information). The maximum driving force for hydrogen diffusion depends on the highest hydrogen concentration generated by acidogenic bacteria and the lowest hydrogen concentration reached by methanogenic archaea. The highest hydrogen concentration was calculated in terms of the electron-donating reaction (Table 2, C2H5NO2 + 2/3H2O → 2/3CH3COO− + 2/3H+ + NH3 + 2/3CO2 + 1/3H2, corresponding to ΔG' = 0), and the lowest hydrogen concentration was calculated in terms of the electron-consuming reaction (Table 2, 1/3H2 + 1/12CO2 → 1/12CH4 + 1/6H2O, corresponding to ΔG' = 0). To estimate the maximum hydrogen flux in DIET, the concentrations of reactants and products were set identical to those in MIET (see Figure S1, Supplemental Information). Assuming that the electrons are released from glycine degradation through the electron-donating reaction (Table 2, C2H5NO2 + 2/3H2O → 2/3CH3COO− + 4/3H+ + NH3 + 2/3CO2 + 2/3e−, ΔG0′ = −60.0 kJ/mol), the electrons are directly transferred to methanogens via graphene. Methanogens reduce CO2 to CH4 through the electron-consuming reaction (Table 2, 2/3H+ + 2/3e− + 1/12CO2 → 1/12CH4 + 1/6H2O, ΔG0′ = 15.7 kJ/mol). The maximum driving force for electron transfer is given by the redox potential (ΔE) of the overall reaction (C2H5NO2 + 1/2H2O → 1/12CH4 + 2/3CH3COO− + 2/3H+ + NH3 + 7/12CO2, ΔG0′ = −44.3 kJ/mol).

The resulting maximum electron flux via MIET and graphene-based DIET was calculated as 9.8 nA and 4.1 mA, respectively. The result demonstrated that the two electron transfer mechanisms could lead to a huge difference in maximum electron transfer flux. This is in accordance with previous studies (Cruz Viggi et al., 2014, Lin et al., 2017), in which DIET showed a clear advantage over MIET in maximum electron transfer when converting propionate and ethanol in AD.

Conclusions

This study demonstrated that graphene nanomaterial can promote AD of protein-derived glycine. The addition of 1.0 g/L graphene led to the highest peak biomethane production rate of 27.9 mL/g/day, which was 28.0% higher than that without graphene addition. It was observed that the consumption of glycine was enhanced simultaneously. Compared with the first-order kinetic model and the Cone model, the modified Gompertz model exhibited the best fit with measured data, and indeed a higher graphene addition (2.0 g/L) had a toxic effect on digestion. The analysis of kinetic parameters revealed that the peak biomethane production rate, lag-phase time, and peak time showed strong linear correlations with graphene additions from 0.25 to 1.0 g/L. The improved performance may be attributed to the potential establishment of DIET between Sedimentibacter (7.8% in abundance) and Methanosarcina (11.3%).

EU policies require a significant rise in advanced biofuels (such as biomethane) in the share of renewable energy consumption. The deployment of the proposed AD systems can help to meet this target. Although the direct use of high-cost nanomaterials in AD is unlikely to be economically feasible, the proof of concept benefits the application of DIET to promoting biomethane production from a variety of feedstock. Biochar derived from thermochemical processing of biomass may be a cost-effective alternative to graphene. Biochar produced under certain conditions can exhibit outstanding electron transfer capability. In a similar manner to graphene, high-quality biochar may enable the establishment of DIET-based AD. Biochar can act as a bridge linking biological AD and thermochemical pyrolysis. The enhanced AD system through addition of inexpensive materials can improve biogas production efficiency, thereby potentially reducing the cost of systems deployment.

Limitations of Study

One limitation of the bioenergy system using nanomaterials apparently is the high cost. Future work is required to propose an economically viable bioenergy system incorporating recyclable conductive materials (such as high-quality biochar). Effective bioreactor design improving biomethane production efficiency is also necessary to facilitate the deployment of the integrated AD system. Deep microbial analysis should be performed to further reveal the specific microorganisms involved in the DIET process.

Methods

All methods can be found in the accompanying Transparent Methods supplemental file.