Autonomous robots using microbial fuel cells technology

13 11 2010

Powering any moving machine or appliance requires a  portable source of electricity. The normal source used is a battery, that steadily converts the chemically reactive substances in its interior into some other products. The difference of free energy released in the reaction is taken by electrons that travel in the outer circuit so that the electrical power is obtained via the electrodes.  Fuell cells make a similar function, but the operation is not limied by the amount of chemical reactives in its interiors. Instead, a chemical fuel can be continuously fed that drives the chemical reactions from which the electrical power is obtained. The fuel cell runs for so long as the fuel is available.

Microbial fuel cells (MFC) are special in that the supply fuel can be organic matter or waste watter. The MFC contains bacteria in contact with electrodes. The bacteria are very small microoganisms (size appr. 1 µm) which can convert many organic compounds  into CO2, water and energy. The bacteria extract electrons (oxidize) some organic substrate such as glucose, and simultaneously produce protons. The micro-organsisms carry out this process in order to use the produced energy to grow and to maintain their metabolism, but we can intercept the process to extract a part of the energy as electricity. Electrons can be transfered from the bacteria to the negative potential electrode (the anode) as in a battery. The converse process, electron injection, occurs at the positive electrode, at which other micoorganisms can take advantage of the cathode as a source of electrons before transfer to a final oxidant.  The circuit is closed by proton transfer across a selective membrane between the electrodes.

The origin of energy output of this process in terms of electricity is explained in a recent review article (Falk Harnisch and Uwe Schröder Chem. Soc. Rev., 2010, 39, 4433-4448). The chemical energy (free energy) of the fuel Efuel can be converted to electricity down to the energy of the substrate oxidation, Eox, except for some energy losses. The losses at the anode are related to biological energy dissipation such as anabolic cell processes. Losses at a microbial cathode can have a similar origin.

One exciting application of MFC is to create autonomous artificial systems that can obtain energy from organic matter in the surroundings (like predators). The Bristol Robotics Laboratory has a history of creating robots that can run off the decomposition of organic matter. Their latest is the Ecobot-III Bio-Regulation and Energy Autonomy with Digeston or simply the BREADbot. This robot has an ‘artificial gut’ that holds a host of sludge bacteria to decompose all kinds of waste. 48 MFC harness the decomposition process for small amounts of electricity. BREADbot ‘eats’ wastewater for fuel and freshwater to replace moisture lost to evaporation. The robot travels back and forth between supplies for each along a stainless steel track.

Ecobot artificial stomach robot

And of course it has a defecation system…


ISTest: a new tool to treat impedance spectroscopy data of dye solar cells.

9 11 2010

Research in dye-sensitized solar cells (DSC) continues to expand across the world, and important industrial ventures are rising. For many years, research in the DSC was mainly a scientific quest to understand the mechanisms, and several types of measurements, for example electron lifetime, have been widely used. At this point, while the scientific side is still important and can provide new discoveries, especially in innovative configurations of the device, there is a priority to improve the efficiency and robustness of the DSC for specific applications and masive energy production from sunlight.

A wide variety of material and nanostructures is being investigated and proposed in publication. The problem is that changing something in the DSC usually modifies the whole behaviour of the device, and good characterization is critical, especially when specific effects have to be detected in the performance of the device.

Impedance Spectroscopy has been recognized as an esential method to test the properties of DSC. This is because the IS measurement provides integral information on the main aspects of the device performance, at given conditions of illumination and voltage: The recombination rate, series resistance, energetics, etc. The equivalent circuit parameters can be followed along the current-potential curve and explains the reason for the current-voltage characteristics in terms of the internal cell processes.  However, to obtain this information in a useful way a proper experimental procedure and interpetation of the results is required.

For example, many papers when comparing two DSC with a given modification,  report the impedance spectra at Voc of the two cells and compare the size of some arcs. This procedure is basically useless for the interpretation, because when the cell is modified (the dye changed, the titania particles in another structure, whatever) the cells, even if measured at the same voltage, are very likely to be in different internal states. So the parameters cannot be directly compared: they change simply because the number of electrons is different in each case, and the significant differences, for example the charge transfer kinetic constants, cannot be seen in this way.

The proper way to treat IS data of DSC, widely used by the key experts in this topic, is to find the  resistances and capacitances (given by the standard modelling circuit) at different potentials in a given illumination level. Next, it is important to corrrect the IR drop in the potential scale. Some labs make very high efficiency DSC (notably M. Grätzel and P. Wang), and for this the series resistance in their cells is very tiny. But obtaining such top cells requires sophisticate care in preparation that is not actually followed in most labs. If the current is sufficiently large (as expected for publishable cells!) the series resistance effect distorts the potential axis and makes the interpretation very difficult, or impossible.

The method to correct this is to measure the current-potential curve at the same time as the IS data, so that IR drop can be removed. Once this is done, it is also important to shift all the voltages to a scale in which the conduction band of titania would be at the same equivalent level. So the first potential is termed Fermi-level potential (VF) because is reflects the rise of Fermi level in TiO2, and the second one is termed common equivalent conduction band potential (Vecb). This notation is followed in our latest publications.

Since this procedure has to be applied in every case and it is sometimes cumbersome, the company ISTest has developed a method to treat the data and provide the required results. The company is a spin-off at UJI that is currently advised by the Bisquert group, notably by Fran Fabregat, who has done in the past critical breakthroughs in the interpretation of Impedance Spectroscopy of dye-sensitized solar cells. The registration and access to the program is free of charge, although in the future the company will run the program independently and will charge some fee for the use. The method works online, you insert the jV and IS parameters and get back the solar cell characteristics and the parameter plots ready for interpretation.

So for example the measured jV curves of two DSC have the following aspect

You can see that there is some significant series resistance here, because the drop of the current is tilted, when approaching Voc. However this could be due also to the recombination rate variations… To distinguish this simply in the jV curve is not possible a priori, because the series resistance contains transport and electrochemical components, and the recombination resistance is unknown a priory and depends heavily on the conditions at the interface. Either factor could cause variation in the performance, and we need a method to check this.

That is the aim of the IS measurement and interpretation.  To the above cells we apply the IS measurements and derive the recombination resistance, which shows directly the recombination characteristic at each potential (free of other effects). The recombination resistance measured from IS is

However, while the vertical scale is the recombination resistance separated for other factors in the spectra, the horizontal scale is the total potential in the cell and is not related only to the electron Fermi level position (which is what we need). In fact the aspect of the resistance is very odd, as it is flattened at the beginning and end potentials.

Now if we run the program we obtain the following results

Sample name Sample1 Sample2
Voc(V) 0.677 0.798
Jsc(mA/cm2) 17.76 8.21
FF 0.639 0.735
Eff(%) 7.688 4.812
β 0.852 0.695
Jo(mA/cm2) 3.583 E-09 4.028 E-09
ΔEc vs ref (mV) Ref 95.245
Rseries Average(Ω) 29.89 21.11
FF(internal) 0.834 0.826
Eff(internal, %) 9.888 5.352

After the fit we can use the VF scale (in which IR is removed in the potential scale) and the recombination resistance is

It is quite straight! This is useful because from this exponential dependence of Rrec on VF we can obtain very reliably the beta-parameter of recombination, and the prefactor of the resistance, associated to the fundamental recombination models. From these parameters we achieve excellent knowledge of the recombination in these DSCs that could not be obtained from the above original curves.In addition, the chemical capacitance will tell us the relative position of the conduction band in each cell. So we can make a final plot with respect to a potential in which the capacitances overlap

Here we see that the cell 1 has  a lower recombination resistance than 2. Since the Fermi level is plotted in a scale where it represents the same density of electrons, now the lower resistance is associated to a higher charge transfer rate.

Indeed we can see the jV curve with respect to VF:

This is much better than before! See in the Table that the cell with the larger current passes from efficiency 7.7 to 9.9 when we remove the series resistance effect. This is quite useful to asses the materials under investigation, as we want to know what efficiency they are capable of producing is a variety of scenarios of cell construction method. As we observed in the Rrec-Vecb plot, cell 1 has a little more recombination than cell 2. However the recombination is the two cells is in practical terms very similar: the difference in Voc corresponds roughly to the difference in the position of the conduction band that is calculated from the chemical capacitance. Thus the main difference is that cell 1 has much larger injection, which gives the high photocurrent. This explains the subtantially higher efficiency of cell 1.

We hope that this tool with facilitate the application of IS characterization of DSC in full force. With a unified method of interpetation of the parameters, the reading of results will be much easier for everyone. This will improve the understanding of the innovations and the development of better DSC.

Try it at