Connecting the world will require a better energy solution.

The early stages of the IoT "hype curve" saw some wild predictions for the number of devices that would be deployed. They're not looking so wild now, with 15 billion devices in 2020 and 29 billion expected by 2030. About 60% of these will be consumer devices, the remainder industrial, or IIoT, devices including smart meters and sensors for monitoring automation equipment, transportation infrastructures, and buildings like offices and factories.

Knowing that IPv6's 128-bit address space would permit more than 100 IP addresses for each atom on the surface of the earth, we can see that the IoT could theoretically grow well beyond even the most ambitious predictions.

While we can solve many problems by adding more of these devices, we are creating another at the same time. Each one needs a source of energy to operate and the fact that many of them will be deployed in mobile or remote locations means a battery is the most obvious power source. Already, the US alone throws away about 3 billion batteries every year and our IoT habit could add many extra tons of hazardous waste. But there are some exciting alternatives.

None of the main battery technologies we use today is perfectly suited to IoT. Lead-acid batteries, although easily recycled and offering high performance thanks to low internal resistance, are large and heavy, typically unsuitable for use in tiny sensors. Alkaline batteries have a high self-discharge rate that limits their useful lifespan. And, while lithium-ion batteries offer greater energy density and longer lifetime, they are not easily recycled.

Arguably, waste is the greatest problem associated with using batteries to power IoT devices. Such massive deployments, and so geographically dispersed, means installing new batteries and recovering discharged units is simply impractical in most cases. Those chemicals and metals such as lithium will be left where they are deployed to decay into the environment. Even if we could recover the batteries, dealing with the waste properly will always be challenging.

Powering IoT devices from ambient energy sources can provide a perpetual supply that eliminates the replenishment and waste issues. There are issues here too, however. Photovoltaic cells have come a long way since the first solar-powered commercial wristwatches and calculators of the early 1970s. While their efficiency has increased and their cost has come down, integrating solar panels on small devices can be challenging and energy storage is needed to ensure continuity of supply in low light or poor weather conditions.

Alternatives like thermoelectric cells and kinetic generators that leverage electromagnetic induction also have shortcomings. Thermoelectric elements can generate relatively little power, while it's difficult to see kinetic generators being scaled down to a suitable size to power small IoT devices. Also, harvesting RF energy is known to provide enough energy to wake a small subsystem to retrieve and send back information stored in memory, such as authentication credentials or software version history. Again, relatively little energy is generated and a strong field is required, such as that from a reading device pointed directly at close range.

All of these are known to be effective in a suitable context and many applications are in action today. The search continues for a small, long-lasting, low/zero-waste energy source, however, to power the vast numbers of IoT devices we want to deploy in the future.

The FOXES project, by a consortium of European research institutions, has addressed some of the integration and storage issues associated with solar cells. Their Power Cube is a 3-D assembly that comprises a perovskite solar cell stacked on top of an energy-management circuit and multilayer thin-film capacitor for energy storage. With a surface area less than 1″ square, the research model can harvest about 260 millijoules per day, even in very low light conditions. To give an idea of how useful that is, the team says this is enough to power the gas-sensing IoT node they're developing, which is intended to consume less than 50µW per readout.

On the other hand, research into microbe-derived energy could provide a great way to power smart farming. Northwestern University in the US has demonstrated a fuel cell that harvests energy from microbial action. This "dirt-powered" source is capable of operating small electrical loads such as moisture sensors and touch sensors – which could be used for monitoring animal movements – and is about the size of a paperback book.

While these concepts, theoretically, can supply energy forever, they may still be too big for the most size-constrained applications. Atomic batteries that harness energy from decaying radioactive isotopes could offer a solution. They can last several years or decades, depending on the materials used. The idea is over 100 years old, although the energy available has been considered impractically small. With the minuscule demands of today's tiny IoT sensors, perhaps it's time has come; a recent Chinese project demonstrated a 0.6″ x 0.6″ x 0.2″ cell capable of generating 100µW and lasting for 50 years. Alternatively, nuclear-waste batteries made with irradiated carbon recovered from nuclear reactors, encased inside synthetic diamond that absorbs the radiation to convert into usable electrical energy, can run for thousands of years. I like the circular aspect of this concept, reusing waste materials.

Right now, each of these potential solutions has its own associated problems. I have every faith that we can engineer our way to overcoming them. Some still need to be smaller, although we can now anticipate a singularity where a small enough source can handle a frugal enough load to allow a lifetime of battery-free, waste-free operation. It could pave the way for practical deployment of not tens of billions, but hundreds of trillions of IoT devices.

Alun Morgan is technology ambassador at Ventec International Group (ventec-group.com); alun.morgan@ventec-europe.com.