Enlarge /. The structure of some of the protein wires used by bacteria.
There are a variety of ways that we can potentially generate all of the power that we need for tiny medical sensors or other devices with minimal power requirements. But there is often a big gap between such use cases and something that could charge your phone, for example, if you walk around in a sweater. The power generation devices either don't scale or don't start with such low power that you'll need a few tents to power a phone.
But today Nature has published a paper describing a device that the authors say should work that powers and scales medical sensors at the bottom to compete with solar modules at the top. And all the device needs to generate electricity is ambient humidity. Even better, the potential for developing the device was accidentally discovered by a student who wanted to do something completely different.
A jerk of low voltage serendipity
A surprising amount of scientific discovery is triggered by anger. The cosmic microwave background was famously discovered by people who worked on a microwave receiver and could not get rid of an annoying noise source – even after trying to remove all the pigeon guano from the hardware. In the case of the most recent work, a doctoral student named Xiaomeng Liu tried to work with some fibrous proteins made by bacteria. In many species, these submicroscopic fibers are good conductors, and a number of laboratories are studying their properties and those of the bacteria that produce them.
In this case, Liu experimentally placed a collection of the bacterial proteins between some metal plates to test their properties. But the proteins continued to generate tension that his equipment registered. Presumably, this tension disturbed everything he actually tried because he was trying to get rid of it – and at least largely failed.
The only thing that seemed to actually remove the tension was the removal of the ambient moisture. So Liu and other lab members shifted focus from trying to get rid of the stray voltage to understanding how moisture can create it.
In the end, they developed a device that was a conductive plate coated with the tiny protein fibers derived from bacteria. They place a few thin electrode strips on these fibers. The gaps between these strips allow the atmosphere to access the fibers so that moisture can penetrate the network.
A basic characterization showed that the device could generate a one-volt difference with a power density of around 40 milliwatts per square centimeter. The devices can generate half a volt even if they have shrunk to a square millimeter or if the relative humidity has dropped to only 20 percent (a value you would normally only see in the desert). The tension was at a maximum when the layer of protein fibers was 14 microns thick, so not much protein is needed to get this going.
It is crucial that the device can generate electricity for about 20 hours. During this time, the tension dropped by about 30 percent. If you stopped power production for five hours, the voltage would be fully restored, although it is not clear how many times the device could be recycled without sustained performance degradation (the authors simply say "repeated").
What is going on in the world?
All of this sounds suspiciously like free energy. So how can this possibly work? The researchers found that the device's function required a moisture gradient across the layer of the protein network – they measured a saturation of about 27 percent on the surface and only 3 percent at the base of the network. Some of the absorbed water molecules are already ionized, and the rest allow some chemical subsets of the proteins to be ionized, releasing protons into the tiny liquid pockets that form. It is these ions, the researchers suspect, that offer the ability to move charges through the electrodes.
To confirm this, the authors tried some related polymers and found that the presence of many easily ionizable groups was related to electrical performance.
This makes sense to some extent, since the gradient of the water across the device means that there is more ionized material on one side than on the other. And you can see how you give the device time to rebalance to restore the presence of some ions that were used in charge generation. However, it is not clear how this can be maintained indefinitely, as the air humidity would gradually level out over the device over time.
Nevertheless, the authors of the paper are enthusiastic about the prospect of building hardware on a large scale from this device. Since only access to the air is required, the devices can be stacked into a larger structure. The researchers expect that a cube one meter per side, in which the air flow and the devices for harvesting moisture are the same size, could generate one kilowatt of electricity. This number is cheap compared to modern solar modules, which produce about 200 watts per square meter and obviously cannot be stacked.
While it's doubtful whether you could get a lot of airflow through something with devices so close together, you could obviously sacrifice a little energy density to improve the airflow. What matters is that you can place it practically anywhere if you are exposed to humid air – and let it run at night.
The ability of these devices to avoid water saturation over time is an obvious open question, but not the only one. Proteins tend to degrade in the environment over time, and it is not clear how much the function of this device depends on the fibers maintaining their structures. The material used in this device is also harvested by shearing off the surface of bacteria in culture. This may not be a very economical way to manage mass production. Alternative polymers with a similar chemical composition could work, but have not been tested.
Finally, the researchers' model of the device suggests that these protein fibers are not the most efficient method of structuring one of these devices. They calculate that it only produces 4 percent of its theoretical maximum. Another big unanswered question is how much better we could actually be.
Nature, 2019. DOI: 10.1038 / s41586-020-2010-9 (About DOIs).