Journal of Power Sources 196 (2011) 2013–2019 Contents lists available at ScienceDirect Journal of Power Sources journal homepage: www.elsevier.com/locate/jpowsour Automatic control of load increases power and efﬁciency in a microbial fuel cell Giuliano C. Premier a,∗ , Jung Rae Kim a , Iain Michie a , Richard M. Dinsdale b , Alan J. Guwy b a Sustainable Environment Research Centre (SERC), Faculty of Advanced Technology, University of Glamorgan, Pontypridd, Mid-Glamorgan CF37 1DL, United Kingdom b Sustainable Environment Research Centre (SERC), Faculty of Health, Sport and Science, University of Glamorgan, Pontypridd, Mid-Glamorgan CF37 1DL, United Kingdom article info Article history: Received 10 August 2010 Received in revised form 15 September 2010 Accepted 21 September 2010 Available online 1 October 2010 Keywords: Microbial fuel cell MFC Control Power Coulombic efﬁciency Optimisation abstract Increasing power production and coulombic efﬁciency (CE) of microbial fuel cells (MFCs) is a common research ambition as the viability of the technology depends to some extent on these measures of perfor- mance. As MFCs are typically time varying systems, comparative studies of controlled and un-controlled external load impedance are needed to show if control affects the biocatalyst development and hence MFC performance. The application of logic based control of external load resistance is shown to increase the power generated by the MFC, when compared to an equivalent system which has a static resistive load. The controlled MFC generated 1600 ± 400 C, compared to 300 ± 10 C with an otherwise replicate ﬁxed load MFC system. The use of a parsimonious gradient based control was able to increase the CE to within the range of 15.1–22.7%, while the CE for a 200  statically loaded MFC lay in the range 3.3–3.7%. The controlled MFC improves the electrogenic anodic bioﬁlm selection for power production, indicating that greater power and substrate conversion can be achieved by controlling load impedance. Load control ensured sustainable current demand, applied microbial selection pressures and provided near-optimal impedance for power transference, compared to the un-controlled system. © 2010 Elsevier B.V. All rights reserved. 1. Introduction The performance of microbial fuel cells (MFCs) depends on cou- pled biological, chemical, electrical and physical processes, which together are time varying, nonlinear and spatially distributed. The arrangement of these processes in an MFC and their interaction determines the performance of the MFC. However, most of the research to improve the performance of MFCs has focused on reduc- ing overpotential losses, through the selection of the electrogenic bioﬁlm [1,2], substrate (fuel) [3], system design [4] and operat- ing parameters without considering dynamics and the interactions between these factors [5]. Maximizing the power produced from an MFC by selecting the load has been considered previously [6]; however this has typically involved pseudo-static loads, using ﬁxed resistances. A frequently used technique of determining the peak of the power curve (by, e.g. discharge curves or Tafel plot) [7], to establishing the current and hence load at which notional peak power is produced, will not in general allow persistent optimal power production. This is because MFC performance is time varying and often, the measurement pro- cedure itself, will affect the biocatalyst [8], implying that current operating conditions affect future performance. Altering electrical load in order to affect the power produced by an MFC has shown considerable promise and is almost ubiq- ∗ Corresponding author. Tel.: +44 1443 482333; fax: +44 1443 482169. E-mail address: gcpremier@glam.ac.uk (G.C. Premier). uitous in selecting static electrical loads for MFCs. Lyon et al. [9], were able to demonstrate that altering the external loading of MFCs would affect the anode bioﬁlm community which develops, but the different communities exhibited similar performances in terms of power production. While this does not prove that manipula- tion of the load can deliver improved performances beyond those expected from static loads, it does indicate that a relationship exists between bioﬁlm ecology and load resistance. Woodward et al. [10] employed a more sophisticated load varying technique using an optimisation algorithm based on a Multiunit Optimisation Method for maximum power point tracking (MPPT). In this work, rapid convergence on the maximum power was addressed using, two replicate MFCs to inform a MPPT algorithm [11,12]. However, the tracking rate for maximum power can only proceed at the rate which peak power varies, which in turn is dictated by the natural growth and establishment of the active electrogenic bioﬁlm pro- cesses. Pinto et al. [13], presented a two population model which they were able to parameterise and validate with data from four MFCs. Their model indicated that small deviations between internal and external impedance in MFCs could cause large power losses and so concluded that in long term operation, external load should be ‘at least periodically adjusted’ in order to avoid methanogenic activity becoming problematic and therefore, ideally MPPT should be used. The time varying nature of MFCs is evident particularly if con- sidered from start-up with inoculation of the system. Consideration of an MFC model [14] conﬁrms the nonlinearity and time varying dynamic behaviour of such systems [13]. By virtue of the bio- logically open characteristics of devices, for example applied to 0378-7753/$ – see front matter © 2010 Elsevier B.V. All rights reserved. doi:10.1016/j.jpowsour.2010.09.071