A High Power-Density Mediator-Free Microfluidic Biophotovoltaic Device for Cyanobacterial Cells

Biophotovoltaics has emerged as a promising technology for generating renewable energy since it relies on living organisms as inexpensive, self-repairing and readily available catalysts to produce electricity from an abundant resource - sunlight. The efficiency of biophotovoltaic cells, however, has remained significantly lower than that achievable through synthetic materials. Here, we devise a platform to harness the large power densities afforded by miniaturised geometries. To this effect, we have developed a soft-lithography approach for the fabrication of microfluidic biophotovoltaic devices that do not require membranes or mediators. Synechocystis sp. PCC 6803 cells were injected and allowed to settle on the anode, permitting the physical proximity between cells and electrode required for mediator-free operation. We demonstrate power densities of above 100 mW/m2 for a chlorophyll concentration of 100 {\mu}M under white light, a high value for biophotovoltaic devices without extrinsic supply of additional energy.

illustrates the operation of the device. After the fabrication of the device and the electrodes, the cell medium is injected using plastic syringes, and all air is removed by applying pressure on the fluid inlet and outlet. Thereafter, the elastic tubing is cut on one side and the cells are injected through another syringe. The syringes are then left attached to prevent drying of the device. Copper wires are soldered to the electrodes to provide electric connections. Finally, the device is positioned such that the anode forms the bottom and the cells sediment on it under the influence of gravity.
A magnified version of a device filled with a Coomassie blue solution is shown in Fig. 2(a), and a true-colour microscopy image of Synechocystis sp. PC 6803 cells settled on the alloy anode is presented in Fig. 1(b).

II. DEVICE STABILITY
In order to assess the stability of our devices, a sample was loaded with cell medium and its power output was measured during more than 25 hours (see Fig. 3). After a sharp decrease in the first minutes, the device was stable for a period in excess of 24 hours.

III. ANODE MATERIAL
The low-melting point solder Indalloy 19 (Indium Corporation, Clinton NY, USA) -composed of 51% indium, 32.5% bismuth, and 16.5% tin -has been chosen due to its simplicity for generating self-aligned wall electrodes in microfluidic devices. [1][2][3] The melting point at 60 • C enables straightforward insertion into the device on a hot plate at 79 • C, with the liquid metal patterned by polydimethylsiloxane pillars due to its surface tension. Upon removing the devices from the hot plate the alloy solidifies forming a solid electrode.
The standard potentials of each of the constituent metals are 4 Sn ↔ Sn 2+ + 2e − −0.14 V.
These potentials are below the value for the oxidation of hydrogen to water at the cathode (+1.23 V), and therefore it is quite possible that oxides such as, for instance, or SnO 2 are forming on the anode. We have subtracted this oxidative current from our power estimates and did not see any significant deterioration in performance in a control over a time span of 25 hours (Fig. 3).

IV. ESTIMATES OF THE INTERNAL RESISTANCE
Measuring the voltage drop over an external resistor attached to a source yields the terminal voltage which is smaller than the actual cell voltage due to the internal resistance of the source Therefore, Since the IV -characteristics are not linear, the internal resistance of the cell, or -more likely -its output voltage, depends on the current drawn. Nevertheless, from the linear part of the polarisation curve at high currents ( Fig. 2(a)  for the abiotic control. Note that the internal resistance decreased by 1/3 with the addition of the cyanobacteria. Furthermore, since maximal power transfer to the external load is observed when the load resistance is matched to the internal resistance of the cell, we can double-check the above values by comparison to Fig. 2(b). There, the maximum power is observed for external resistances of 1.1 MΩ and 2.9 MΩ for biotic and abiotic filling, respectively. These values are in close agreement with the estimates from the polarisation curves.

V. LIGHT SOURCE
In Fig. 4 we present the measured spectrum of the lamp we have used to illuminate our biophotovoltaic cells. From these data, we can also extract the weighted average wave number to be 10 7 m −1 which corresponds to a wavelength of around 570 nm. Therefore, the average energy per photon is 3.5 × 10 −19 J, and the measured photon flux of 200 µmol/m 2 /s yields an illumination intensity at the location of the devices of 42 W/m 2 .