< Back to previous page

Publication

Microelectronics for Microorganisms: Measurement Instrumentation and Optimization Algorithms for Electroactive Biofilms

Book - Dissertation

Bioelectrochemical systems (BESs) are a novel type of bioreactors where electroactive biofilms (EABs) drive electrochemical reactions as an environmental friendly alternative to many classical chemical pathways. The research community originally focused on microbial fuel cells which directly produce electrical energy. More recently, the focus has shifted towards microbial electrolysis cells, which produce valuable chemicals from organic waste streams, and towards the development of sensors with self-regenerative properties for long-term applications. The catalytic current density of these bioreactors is, however, still too low to compete with established technologies, except for niche applications. Despite being an active research field, research progression is slow and experimental-dependent. These experiments are challenged by a long experiment duration, in the order of several days to weeks, together with an inherent stochastic variation within similar experiments. Dedicated instrumentation is essential to overcome these hurdles. While demonstrating top-notch performance, existing state-of-the-art commercial measurement instrumentation, i.e. potentiostats, is extremely expensive with a price-per-channel in the order of $1-10k (€1-10k). This makes experimentation beyond a few individual channels impractical. Furthermore, only "passive" measurement techniques are supported, where the measurement points are all predefined. "Active" measurements, where the measurement values are determined at run-time based on the measurement results of the previous ones, are not supported. State-of-the-art low-cost research instrumentation contains an analog feedback architecture that becomes unstable for large capacitive loads, for which BESs are known for. Furthermore, these devices are constrained to a single channel only. State-of-the-art multichannel instrumentation contains multiple (up to 1024) sensing channels but only a few (typical only one) stimulation channels. Parallel read-out is thus supported but individual parallel stimulation is not possible, making these instruments not suitable for high-throughput experiments. This work aims at increasing the research progression speed by: (1) increasing the experiment throughput using parallelism, and (2) reducing the experiment measurement time using search & optimization algorithms, which optimally trade measurement time for measurement accuracy. To achieve these two goals, this work proposes two innovations for the instrumentation hardware. For the first hardware innovation, this work introduces an inherently BES stable, digital feedback loop for the potential control. This digital potential control loop replaces the unstable analog potential feedback loop of the lowcost state-of-the-art alternatives. This potentiostat channel architecture has been implemented in a 6-channel low-cost, $50-per-channel (e45-per-channel), printed circuit board (PCB) prototype. The digital-oriented architecture naturally supports the customized search & optimization algorithms. For the second hardware innovation, this work introduces a scalable, affordable multichannel potentiostat channel hardware architecture using time division multiplexing. This potentiostat channel architecture benefits from the introduced digital potential control loop and is implemented on PCB. The prototype contains 128 individual channels, 16× more than the state of the art. The cost-per-channel is only $5 (€4.5), 4× lower than the state of the art, the area-per-channel only 100 mm2, 5× lower than the state of the art. Next, this work introduces two novel measurement algorithms that optimally exploit the advantages of the 6-channel potentiostat (support for "active" measurements) and the 128-channel potentiostat (support for 128 simultaneous experiments). The first algorithm, fast polarization curve characterization, measures the BES polarization curve. The measurement time of this algorithm is logarithmic dependent on the measurement resolution instead of linear for the existing techniques, resulting in a substantial reduction of the measurement time for equal accuracy. A practical speedup of 2-12× is demonstrated in a case study that evaluates the impact of pH on a BES. The second algorithm, parallel cyclic voltammetry, measures instantaneously the polarization curve by exploiting the 128 parallel channels. This algorithm is about 72× faster than a classical 1 mV/s cyclic voltammetry experiment. Besides those two algorithms, this work also introduces the high-throughput experiment paradigm, made feasible by the 128-channel potentiostat. The advantages of using this paradigm are: BES studies with many different data points, and at the same time, BES studies with sufficient replicates to obtain statistically sound results. These advantages are demonstrated in a case study investigating the impact of the anode potential on the current density of EABs.
Accessibility:Closed