When it comes to outdoor Ambient IoT deployments, solar energy is often the first go-to option due to its abundance and the unmatched efficiency of solar cells. However, there are many situations where solar energy may not be readily available, necessitating the exploration of alternative energy sources. We encountered such a challenge while working on the FutureArctic project. Our task was to deploy a large network of low-power outdoor sensing devices in a remote site (ForHot) in Iceland to facilitate seamless data collection for ecologists and biologists studying future carbon release from the Arctic ecosystem. In Iceland, harsh winter conditions result in limited solar radiation for nearly six months. Moreover, heavy snowfall can accumulate on solar panels, drastically reducing their power output to almost zero. However, this site has a very peculiar characteristic; it was geothermally warmed. So, the soil there was at a higher temperature, even up to 100°C. We could use the temperature difference between geothermally warmed soil and air (<25°C)  to generate energy using a thermoelectric energy generator (TEG). 

ForHot is a special case with geothermally warmed soil. But is it possible to extend the use of this specific energy source to other locations? If we could, it would unlock a unique landscape of energy opportunities. In many situations where solar power is not ideal, we could instead rely on soil-air thermal energy. Intrigued by this possibility, I undertook a thorough investigation of soil-air thermal energy during my PhD, which formed a substantial part of my research.

Intuitively,  a temperature difference between soil and air is possible anywhere because these mediums have distinct thermal properties. While both receive solar radiation during the day, the soil warms up relatively slower than the air. Therefore, the soil temperature always lags behind the air temperature. To verify this, I conducted a series of data measurement campaigns. Soil and air temperature at three different locations in Antwerp, Belgium, were measured continuously for 6 months. The air temperature and soil temperature at 15 cm were measured every 5 minutes. The experiments were conducted during winter months (Oct-May) so that the influence of solar radiation is minimal. From the collected data, a significant difference between the soil and air temperature became evident throughout the measurement. For the three locations, the average temperature difference was more than 2.5°C. 

Motivated by the positive outcomes from the measurement campaign, I decided to build an energy harvester that could harness energy from soil-air temperature differences.  For sure, this requires a TEG. In addition, a set of components is necessary to apply the temperature difference between the soil and the air across the TEG. This involves copper rod(s) on the soil side and a heat dissipator (absorber) on the air side. We chose copper rods due to their very high thermal conductivity. The heat dissipator could be something like a heat sink which relies on natural convection or a black radiator which dissipates heat through radiation. We used a radiator on the air side as it can further enhance energy generation through solar absorption during the day.

I tested this generator  (I call it SoTEG- Soil-air Thermal Energy Generator) in an outdoor environment for two weeks during October and November. With a 3°C temperature difference between the soil and the air, the SoTEG could generate around 0.1 mW of power and when combined with a power management unit (PMU), LTC3109, the actual energy available to the load is around 0.03 mW at 4.2 V. This power is sufficient to trickle charge many low power outdoor sensing device, for example in an agriculture environment. By improving the harvester design and using a PMU with higher efficiency, such as MCRY-12, which offers > 70% efficiency, we can enhance the power generation even more. In addition, considering that empirical evaluations may not be feasible all the time, I developed a simulation model of the SoTEG to estimate the feasible power output if other ambient parameters are available. You can find the model here.  Using this model and a massive dataset from CuriuezeNeuzen, we conducted a large-scale analysis of soil-thermal energy potential of the whole Flanders region in Belgium. The results of this study will be made public soon.

We plan to further investigate the possibility of using this unique energy source in real-world applications through the recently awarded IOF Proof of Concept (PoC) project, Sustainable and Energy Neutral Soil Sensing (SENSS).  The idea behind SENSS is to build an energy-autonomous and batteryless system that can sense different soil parameters including soil temperature, nutrients and soil CO2 emissions to assist farmers in selecting optimal fertilization strategies. This approach will drastically reduce the expenses incurred by farmers while simultaneously helping to control carbon emissions from agricultural fields.

References

1. Puluckul, P.P., 2024. Building Reliable and Sustainable Internet of Batteryless Things, Doctoral dissertation, University of Antwerp
2. Puluckul, P.P. and Weyn, M., 2024. Harvesting Energy from Soil-Air Temperature Differences for Batteryless IoT Devices: A Case Study. IEEE Access.