MFCs Did Not Fail — They Were Pushed Toward the Wrong Objective
Microbial fuel cells (MFCs) generated something no other electrochemical platform has: electrical signal originating entirely from biological metabolism. No applied voltage. No forced current. The electrons were transferred by the organisms themselves as a byproduct of respiration. The electrical output was not a response to a measurement. It was the measurement.
The field framed it as energy. The metric was power density. The goal was to generate enough electricity to be useful — to power remote sensors, to recover energy from wastewater, to scale toward deployment. When the power output proved too low, too variable, and too dependent on conditions, interest declined. The conclusion was that MFCs did not work.
But what failed was the power application. The signal never failed.
In mixed-community systems, current fluctuated with population dynamics — shifts in species dominance, changes in syntrophic interactions, responses to nutrient availability. In single-species systems using Geobacter or Shewanella, electron transfer rates tracked metabolic activity in ways no applied-voltage technique could replicate — because the signal was not driven by the instrument. It was driven by the biology. The variability that made MFCs poor energy devices is exactly what would make them powerful as a sensing and data platform. A current that shifts with community restructuring, that responds to metabolic transitions before bulk measurement detects them — that is not noise. That is the system reporting its own state.
Every other electrochemical technique in biology — cyclic voltammetry, impedance spectroscopy, amperometry — requires applied excitation. The instrument imposes a signal and reads the response. This captures what the system does when forced. MFCs were the one platform where the biology generated signal without being asked.
When that signal was variable and hard to reproduce, the field saw weakness. But this is the same pattern that runs through every layer of biological measurement: signal that does not fit the expected framework gets discarded. Variability gets filtered. Inconsistency gets called noise. The data that current methods remove — the transient, the irregular, the nonstationary — is the data that carries the most biological meaning. MFCs proved this at the level of electron transfer, and the field walked away because it was looking for power, not information.
The field optimized MFCs for power and discarded them when power was insufficient. What it discarded was the only electrochemical platform that captured biological signal without overwriting it.
The data from decades of MFC research — current fluctuations, variability across conditions, inconsistencies that frustrated power applications — sits in supplementary files and abandoned lab notebooks. It has never been revisited as a biological dataset. Not forced. Not filtered for stability. Just organisms generating signal on their own terms.
We did not fail to make MFCs work. We failed to recognize what they were telling us.
— Pegah Farr