Non-catalyzed cathodic oxygen reduction at graphite granules in microbial fuel cells

Abstract

Oxygen is the most sustainable electron acceptor currently available for microbial fuel cell (MFC) cathodes. However, its high overpotential for reduction to water limits the current that can be produced. Several materials and catalysts have previously been investigated in order to facilitate oxygen reduction at the cathode surface. This study shows that significant stable currents can be delivered by using a non-catalyzed cathode made of granular graphite. Power outputs up to 21 W m⁻³ (cathode total volume) or 50 W m⁻³ (cathode liquid volume) were attained in a continuous MFC fed with acetate. These values are higher than those obtained in several other studies using catalyzed graphite in various forms. The presence of nanoscale pores on granular graphite provides a high surface area for oxygen reduction. The current generated with this cathode can sustain an anodic volume specific COD removal rate of 1.46 kg_COD m⁻³ d⁻¹, which is higher than that of a conventional aerobic process. This study demonstrates that microbial fuel cells can be operated efficiently using high surface graphite as cathode material. This implies that research on microbial fuel cell cathodes should not only focus on catalysts, but also on high surface area materials.

Introduction

Microbial fuel cells (MFCs) are an emerging technology for the treatment of reduced organic carbon in aqueous waste [1]. In a MFC, bacteria use an anode as electron acceptor for the oxidation of organic carbon to carbon dioxide (CO₂), with production of protons and electrons. The protons and/or other cations in solution are transferred to a cathode through a cation exchange membrane (CEM), while the electrons are transferred via an electrical circuit, where a resistor or power user harvests the energy liberated by the reaction. At the cathode an electron acceptor is reduced using the electrons delivered by the anode. The most sustainable electron acceptor known to date for microbial fuel cells is oxygen, due to its availability in the environment and its high redox potential. However, some limitations reduce the effectiveness of O₂ as electron acceptor. First, the low oxygen solubility in water limits the availability of the electron acceptor to the electrode. The oxygen mass transfer limitation in the liquid phase has been alleviated with the use of open air cathodes [2], but the concentration of oxygen can never exceed saturation (7.8 mg/L at 25 °C), as the electrode must be covered by a liquid film in order to allow for cation/proton transfer. Secondly, the high activation energy of O₂ reduction to water at a graphite electrode leads to high cathodic activation overpotential, i.e. drop in cathodic potential from open circuit value (E_C,OC) to closed circuit value (E_C) due to losses associated with the reaction kinetics. By reducing the MFC voltage, overpotentials strongly hinder electron transfer. Zhao et al. [3] demonstrated that graphite foil performs very poorly in a linear sweep voltammogram compared to different catalyzed cathodes. Oh et al. [4] found that during the generation of polarization curves the voltage over a MFC dropped to zero much more rapidly when O₂ was reduced on a plain carbon cathode than when catalyzed cathodes were used. Several researchers have investigated graphite-bound catalysts or soluble redox mediators as possible solutions to increase the rates of cathodic oxygen reduction. The use of hexacyanoferrate (Fe(CN)₆³⁻) as a soluble redox mediator has been unsuccessful due to the low rates of reoxidation of the compound from ferrous to ferric state [5]. Instead, the use of platinum at MFC cathodes has confirmed the lessons learned from conventional fuel cell technology, by significantly enhancing oxygen reduction rates, even at catalyst loads as low as 0.1 mg/cm² [6]. However, platinum has some severe drawbacks. First, it is subject to sulfide poisoning in wastewater applications and the cost of catalyst replacement can be prohibitive; secondly, the production of platinum is highly energy-intensive (150–250 kg of ore must be processed to produce 1 g of platinum) and thus the environmental impacts may outweigh the benefits [7]. In the search for cheaper and more sustainable catalysts, transition metal porphyrins and phthalocyanines [3] have emerged, yielding overpotentials in the same range as achieved with platinum. Another strategy has been the use of solubilised ferric iron at low pH (<2.5) as a redox mediator, which could be reoxidised by a population of (for example) Acidithiobacillus ferrooxidans [8]. This method reduces the overpotentials because Fe³⁺, not oxygen is being reduced at the cathode. The results obtained in the aforementioned studies and others are summarized in Table 1.

This study proposes a different approach to reduce the oxygen reduction overpotential. This approach is based on surface area rather than catalysis. Based on electrochemistry principles, the overpotential of oxygen reduction is dependent on the current density, as shown in Eq. (1) (Butler–Volmer equation):

$$i=i_0\left[\text{exp}\frac{\beta F{\eta }_{\text{cathode}}}{RT}-\text{exp}\frac{-(1-\beta )F{\eta }_{\text{cathode}}}{RT}\right]$$

In the equation, η_cathode is the cathodic overpotential (E_C,OC − E_C, V), R the ideal gas constant (8.31 J/mol K), T the absolute temperature (K), β the symmetry factor (a constant which determines the dependence of the activation energy on the electrode potential and for a cathodic process reflects that the activation energy increases with electrode potential), F the Faraday's constant (96,485 C/mol), i the current density (mA/m²) and i₀ is the exchange current density, a parameter that depends on the activation energy of the reduction at equilibrium conditions (in a way that higher activation energy corresponds to lower exchange current). If the overpotential is large enough (greater than 80–100 mV at 25 °C) the second term in parenthesis becomes negligible and the equation can be simplified in the form known as Tafel equation:

$$\text{ln}\left(\frac{i}{i_0}\right)=\frac{\beta F{\eta }_{\text{cathode}}}{RT}$$

The equation shows that the activation overpotential increases with the current density. Instead of using a catalyst to reduce the activation energy for oxygen reduction (i.e. increasing the parameter i₀), our approach has been to reduce the current density through use of a non-catalyzed material with a high surface area. This would still cause a decrease of the overpotential, through a reduction of the current density rather than the activation energy. The material selected was coarse and highly porous industrial grade granular graphite. Some investigators have previously observed currents generated in sediments fuel cells by oxygen reduction on plain graphite in marine environments [9], [10]. However, it was suggested that in those cases bacterial colonization of the cathode may have aided electron transfer through a biocatalytic process. In this study we eliminated every possible chemical and biological catalytic process to test the ability of plain graphite granules to carry out cathodic oxygen reduction.