Addition of a carbon fiber brush improves anaerobic digestion compared to external voltage application

Highlights

• Large conductive surface areas were compared to the use of electrodes in AD

• Adding a large carbon brush improved performance more than current generation

• More biomass could be retained on the large carbon brush than small-size electrodes

• Methanothrix was dominant methanogens on the large-size carbon brush

• Geobacter and Methanobacterium were highly enriched on the electrodes

Abstract

Two methods were examined to improve methane production efficiency in anaerobic digestion (AD) based on adding a large amount of surface area using a single electrically conductive carbon brush, or by adding electrodes as done in microbial electrolysis cells (MECs) to form a hybrid AD-MEC. To examine the impact of surface area relative to electrodes, AD reactors were fitted with a single large brush without electrodes (FB), half a large brush with two electrodes with an applied voltage (0.8 V) and operated in closed circuit (HB-CC) or open circuit (HB-OC) mode, or only two electrodes with a closed circuit and no large brush (NB-CC) (equivalent to an MEC). The three configurations with a half or full brush all had improved performance as shown by 57-82% higher methane generation rate parameters in the Gompertz model compared to NB-CC. The retained biomass was much higher in the reactors with large brush, which likely contributed to the rapid consumption of volatile fatty acids (VFAs) and therefore improved AD performance. A different microbial community structure was formed in the large-size brushes compared to the electrodes. Methanothrix was predominant in the biofilm of large-size carbon brush, while Geobacter (anode) and Methanobacterium (cathode) were highly abundant in the electrode biofilms. These results demonstrate that adding a high surface area carbon fiber brush will be a more effective method of improving AD performance than using MEC electrodes with an applied voltage.

Introduction

Anaerobic digestion (AD) is an efficient method to degrade organic matter in wastewaters and produce methane but the process can require long hydraulic retention times for stable performance. The syntrophic relationship between fatty acid-degrading bacteria and methanogens is often considered as the rate-limiting step for soluble organic acid conversion to methane (Baek et al., 2018). Therefore, methane generation rates can be improved by reducing the background concentrations of volatile fatty acids (VFAs) produced from fermentation. Inserting electrodes like those used in a microbial electrolysis cell (MEC) into AD (i.e. AD-MEC), and applying a voltage to generate an electrical current and hydrogen gas, has been proposed to improve overall AD performance (Vu et al., 2020; Zhao et al., 2016). AD-MEC systems use exoelectrogenic bacteria on the anode to oxidize organic matter and reduce VFA concentrations, and the cathode can improve methane production either through electromethanogenesis (8H⁺ + 8e⁻ + CO₂ → CH₄ + 2H₂O) or by providing H₂ for hydrogenotrophic methanogenesis (4H₂ + CO₂ → CH₄ + 2H₂O) (Cheng et al., 2009; Yu et al., 2018). AD-MEC systems have therefore been shown to increase methane production and organic removal efficiencies (Feng et al., 2015; Zhang et al., 2013; Zhao et al., 2016).

The main function of the electrodes in an AD-MEC has not been clearly distinguished between improvements due to additional surface area for biomass retention, and benefits of electrochemical processes that can remove additional organics and produce hydrogen gas. The operation of homogenized and continuously-fed AD reactors is dependent on the retention of biomass, especially methanogens (Choong et al., 2018). Adding supporting media with a high surface area into reactors has therefore been used to ensure a high biomass concentration (Show and Tay, 1999). Most of the materials used for bioelectrodes in AD-MECs provide a relatively high surface area compared to controls without electrodes. Carbon felt or carbon cloth materials have porous structures which could benefit AD performance by retaining biomass and preventing washout, in addition to stimulating electrochemical reactions. Despite of the importance of this surface area effect, the electrode packing density (i.e. electrode surface area to reactor volume ratio) has not been well documented or specifically examined in AD-MEC studies. Electrode packing densities have varied over a wide range in AD-MEC studies, from 0.7 to 15.1 m²/m³ of reactor volume, which could have impacted AD performance even without current production (Baek et al., 2020; Cai et al., 2016; De Vrieze et al., 2014; Feng et al., 2015; Liu et al., 2016; Vu et al., 2020). Low electrode packing densities (<1 m²/m³) would not be expected to impact methane production rates by more than a small amount (0.6–1.2%) due to a small percentage of the substrate being used for current generation relative to that for overall methane production (Baek et al., 2020; Feng et al., 2015). However, AD-MEC systems with a relatively small amount of current through the circuit have shown to achieve higher or more stable AD performance than conventional AD systems (Baek et al., 2020). In addition to increasing surface area available for microorganisms, the use of electrically conductive materials can also provide a beneficial platform for direct interspecies electron transfer (DIET) between electroactive bacteria and methanogens even in the absence of current generation (Baek et al., 2015; Cruz Viggi et al., 2014; Li et al., 2015). Through DIET, the electrons from organic oxidation can be transferred directly toward methanogens without redox intermediate (i.e. hydrogen), which could make the AD process more efficient and contribute to enhanced methane generation rates (Baek et al., 2018).

In this study, we investigated the impact of surface area relative to current generation using electrodes by adding carbon fiber brushes of different sizes into AD reactors. We hypothesized that the large surface area provided by the brushes was more critical to AD performance than current generation using smaller-size MEC electrodes. Carbon brushes were chosen for these tests because they provide a very high surface area-to-volume ratio, and the open brush structure has advantages of less potential for clogging compared to other carbon-based woven materials (e.g., pieces of carbon felt or carbon cloth) (Logan et al., 2007). In addition, the potential for DIET is enhanced by using carbon brushes due to their high electrical conductivity (650 S/cm for brushes used here). This strategy to add brushes into AD provide additional advantages compared to adding particles to stimulate DIET (e.g., magnetite, carbon nanotube, activated carbon) (Baek et al., 2015; Li et al., 2015; Liu et al., 2012) because there is no possibility of material washout using the brushes compared to particles. The impact of large carbon brushes on methane generation rates and COD removal rates was examined by adding brushes that filled, or half-filled the reactor volume compared to reactors with two smaller-size brush electrodes in the presence and absence of current production. The amount of attached biofilm to these different brushes was quantified in terms of protein, with the contribution of current generation evaluated in terms of coulombic efficiency and total internal resistances based on the electrode potential slope (EPS) method (Cario et al., 2019). To obtain a more comprehensive insight into process performance, the microbial communities on the brushes and in suspension were characterized using Illumina sequencing of both active (16S rRNA) and total (16S rRNA gene) microbial populations.