Evaluating the impact of substrate composition, reactor configuration, and operating conditions on the performance of bioelectrochemical systems treating fermentation effluents.

Open Access
- Author:
- Cario, Benjamin
- Graduate Program:
- Environmental Engineering
- Degree:
- Master of Science
- Document Type:
- Master Thesis
- Date of Defense:
- January 16, 2019
- Committee Members:
- Bruce Logan, Thesis Advisor/Co-Advisor
John Michael Regan, Committee Member
Christopher A Gorski, Committee Member - Keywords:
- microbial electrolysis cells
microbial fuel cells
fermentation
dark fermentation
bioelectrochemical systems
carbon felt
fermentation effluent
anion exchange membrane
platinum/carbon catalyst
anolyte flow rate
reactor configuration
substrate composition
internal resistance
electrode potential slope - Abstract:
- Bioelectrochemical systems (BES) are technologies that convert the chemical potential energy of reduced substances into useful energy. Typically, the conversion is facilitated by exoelectrogenic bacteria which can utilize a physical electrode as a terminal electron acceptor. In microbial fuel cells (MFCs), substrate oxidation at an anode is most often coupled to oxygen reduction at a cathode to produce a positive cell voltage that can be directly utilized. Alternatively, microbial electrolysis cells (MECs) are entirely anaerobic and produce energy-dense and industrially-valuable hydrogen gas at the cathode. MECs require a small applied voltage to overcome thermodynamic limitations under standard biological conditions (>0.14 V in theory, >0.5 V in practice due to internal losses), but this voltage input is less than is needed for hydrogen production via water electrolysis (1.23 V). Although MECs were originally examined as an alternative wastewater treatment technology, they may also have industrial applications as a sustainable source of hydrogen gas. Dark fermentation pre-treatment of complex waste streams can improve BES performance by converting high-order substrates into compounds more easily utilized by exoelectrogenic bacteria. Ideally, acetate and hydrogen gas are the sole end products of dark fermentation, but real fermentation effluents often contain a mixture of volatile fatty acids, alcohols, and proteins. Past studies examining BES treatment of fermentation effluents have used many different substrates and obtained similarly-variable results. There are several factors in addition to solution composition that can impact BES performance, including electrode materials, reactor configuration, and operating conditions. The individual impact of these design variables must be better understood to optimize two-stage dark fermentation/BES processes for practical applications. With that goal in mind, three primary studies were conducted as part of this work. A cellulosic fermentation effluent sample received from the National Renewable Energy Lab (NREL) was analyzed, and a synthetic medium based on the effluent was evaluated as a substrate for BES. The real effluent sample had a total chemical oxygen demand (COD) of 6.21 ± 0.25 g/L, 82% of which was exerted by substrates specifically identified via chemical testing (15% acetate, 31% ethanol, 30% protein). Good agreement between the sample’s chemical and biochemical oxygen demands suggested that its organic content was biologically available under aerobic conditions. Three brush-anode MFCs were operated on a synthetic version of the fermentation effluent (5.07 g COD/L) for 150 days and produced 0.46 ± 0.01 V of electrical potential during typical 24-hour batch cycles (1,000 Ω external resistance, 0.29 W/m2, 7.16 W/m3). The relatively high ethanol concentration of the solution (920 mg/L) did not appear to have inhibited microbial growth, as over 95% of the ethanol was consistently removed in the MFCs. However, the power density obtained was less than in related studies with relatively higher amounts of acetate in the feed solution. A series of experiments was conducted to determine how changes in substrate composition, reactor configuration, and anolyte flow rate would impact the performance of flow-through MECs. Current production from a synthetic fermentation effluent was improved by increasing anolyte flow rate, placing the felt anode across the anolyte flow path from the membrane, and including a flow mixing device in the anolyte flow path. These improvements were likely a result of enhanced substrate and charge transfer. Performance with a medium in which acetate was the sole substrate was similarly impacted by flow rate and reactor configuration, although to a lesser extent than with effluent, possibly due to acetate’s relatively high diffusivity in water. Under optimal conditions, current production was initially greater using the synthetic fermentation effluent (424 ± 7 A/m3) than acetate (364 ± 8 A/m3), but current in the effluent-fed MEC steadily declined after one week of operation for reasons that were not clear. Despite a low cathodic efficiency (<40% recovery of hydrogen relative to current production), the peak hydrogen production rate achieved in this study (1.49 ± 0.32 L H2/L reactor/d) was the highest to date for MECs treating actual or synthetic fermentation effluents. A new method to quantify the components of BES internal resistance from polarization data was applied to acetate-fed cube MECs. Whole-cell internal resistance was calculated to be 120 ± 0 mΩ m2, due in largest part to the carbon felt anode (71 ± 5 mΩ m2, 59% of total). Substrate and ion mass transfer limitations, which are known to occur in MECs with stagnant fluid conditions and felt anodes oriented against the membrane, may explain the high anode resistance obtained here. Ohmic solution resistance (25 mΩ m2) and cathode resistance (18 ± 2 mΩ m2) were also notable sources of loss in the system, whereas the membrane resistance was calculated to be negligible. Anode activation losses appeared to be negative when the measured open circuit electrode potential (–320 ± 20 mV) was compared to the theoretical value under standard biological conditions (–280 mV). The reason for the negative value obtained was unclear, but it is possible that the theoretical half-cell potential for acetate oxidation to carbon dioxide under standard biological conditions was not an accurate estimate of the anode potential before losses in this system. The cathode was found to have notable activation losses (70 mV), consistent with abiotic half-cell results previously reported for Pt/C-catalyzed hydrogen evolution.