Microbial biofilms for electricity generation from water evaporation and power to wearables

Abstract

Employing renewable materials for fabricating clean energy harvesting devices can further improve sustainability. Microorganisms can be mass produced with renewable feedstocks. Here, we demonstrate that it is possible to engineer microbial biofilms as a cohesive, flexible material for long-term continuous electricity production from evaporating water. Single biofilm sheet (~40 µm thick) serving as the functional component in an electronic device continuously produces power density (~1 μW/cm²) higher than that achieved with thicker engineered materials. The energy output is comparable to that achieved with similar sized biofilms catalyzing current production in microbial fuel cells, without the need for an organic feedstock or maintaining cell viability. The biofilm can be sandwiched between a pair of mesh electrodes for scalable device integration and current production. The devices maintain the energy production in ionic solutions and can be used as skin-patch devices to harvest electricity from sweat and moisture on skin to continuously power wearable devices. Biofilms made from different microbial species show generic current production from water evaporation. These results suggest that we can harness the ubiquity of biofilms in nature as additional sources of biomaterial for evaporation-based electricity generation in diverse aqueous environments.

Fig. 1. Electric outputs from G. sulfurreducens biofilms.

a A harvested biofilm sheet of G. sulfurreducens strain CL-1 (schematic inset), floating on water. Scale bar, 2 cm. b Schematic of using laser-patterned biofilms to construct (i) single device and (ii) interconnected device array on a PDMS substrate (gray), with a portion of the biofilm at one electrode immersed in water. A tissue paper was used to support the biofilm in the single device, which may assist water evaporation but does not contribute to electric signals (supplementary Fig. S4). c Cross-sectional TEM of a G. sulfurreducens strain CL-1 biofilm. Scale bar, 1 µm. d A continuous recording of the short-circuit current (I_sc) from a device for one month. The device had the structure shown in (b-i), with an electrode spacing 2 mm and lateral width 1 cm, yielding an estimated energy density V_o·I_sc/4 of ~1 µW/cm² in the active biofilm region. The test was performed in the ambient environment with relative humidity (RH) fluctuating between 30–55%. e Open-circuit voltage (V_o) from an integrated device array (bii) floating on a water surface, which was used to power up an LCD (inset). The actual device is shown in supplementary Fig. S8.

Fig. 2. Integrated biofilm devices using mesh electrodes.

a Schematic and actual photo (bottom) of a biofilm device. Scale bar, 0.5 mm. b Representative open-circuit voltage V_o (top) and short-circuit current I_sc (bottom) from a device placed on a water surface (schematics). The device size was 5 × 5 mm². c V_o (black) and I_sc (red) from devices with respect to different porosities in the mesh electrodes. The pore size was kept 100 µm and the device sizes were 25 mm². d V_o (black) and I_sc (red) from devices with fixed porosity of 0.4 but varying pore sizes (from 1000 to 20 µm) in the mesh electrodes. The device sizes were 25 mm². At the pore size of 20 µm, the estimated optimal energy density (V_o·I_sc/4) was ~0.84 µW/cm² in the active biofilm region. The previous device structure (Fig. 1bi) yielded higher density, probably because the biofilm-tissue paper interface improves water transport. e V_o (black) and I_sc (red) from devices of different areas, with the porosity and pore size in the mesh electrodes, kept 0.4 and 100 µm. f Measured V_o from devices connected in series (upper inset) using a “buckle” design in electrodes (bottom schematic). All the tests were performed in the ambient environment (RH~50%). All the error bars are standard deviations.

Fig. 3. Wearable powering.

a Open-circuit voltage V_o (gray) and short-circuit current I_sc (red) from biofilm devices placed on deionized water, 0.5 M NaCl, 0.5 M KCl, and artificial seawater (475.7 mM NaCl, 10.8 mM CaCl₂, 25.6 mM MgCl₂‧6H₂O, 28.2 mM MgSO₄) solutions. b A biofilm device was patched on the skin (top) and removed 18 h later (bottom). Scale bars, 0.5 cm. c V_o (gray) and I_sc (red) from biofilm devices patched on sweating skin (top) and dry skin (bottom), before (left) and after (right) 18 h. d (Left) schematic of connecting biofilm devices to wearable sensors for wearable powering. (Right) Actual photos of powering a skin-wearable strain sensor with one biofilm device (top) and an electrochemical glucose sensor with three biofilm devices (bottom). Scale bars, 1 cm. e Measured pulse signal (left) from the wrist and respiration signal (right) from the chest using the biofilm-powered strain sensor. f Amperometric responses from a biofilm-powered glucose sensor placed in solutions having glucose concentrations (C) of 0, 100, 200, and 300 µM, respectively. The inset shows the calibrated response curve. g (Top) A continuous measurement of current from a biofilm-powered glucose sensor during exercise. (Bottom) Calibrated glucose levels from collected measurements before (blue) and after (orange) a meal. All the error bars are standard deviations.

Fig. 4. Devices made from filtered biofilm-mats.

a Schematic of harvesting biofilm-mats by filtering microbial solutions. The bottom photos show filtered (left) G. sulfurrenducens and (right) E. coli mats, respectively. Scale bars, 1 cm. b Average V_o (gray) and I_sc (red) measured from devices fabricated with biofilm-mats of G. sulfurreducens, genetically modified G. sulfurrenducens Aro-5 strain, E. coli, and genetically modified E. coli strain. The devices had the same size of 5 × 5 mm², with a porosity of 0.4 and pore size of 100 µm in the mesh electrodes. c Cross-sectional TEM image of a biofilm-mat assembled by filtered E. coli. Scale bar, 1 µm. d Scanning electron microscope images of (left) a tissue paper and (right) a tissue paper infiltrated with E. coli. Scale bars, (left) 100 µm, (right) 10 µm. e Average V_o (gray) and I_sc (red) measured from devices fabricated with tissue paper infiltrated with Geobacter and E. coli. The devices had the same size of 5 × 5 mm², with a porosity of 0.4 and pore size of 100 µm in the mesh electrodes. All the error bars are standard deviations.