Standard silicon CMOS technology can create thermoelectric micro-harvesters that could be used to power numerous IoT devices.
The development of the Internet of Things (IoT) will lead to billions of devices being installed over the next few decades and is expected to have a profound effect on society1. The massive deployment of miniaturized wireless sensor nodes could, in particular, revolutionize the way we understand complex systems and even interact with reality. But how exactly this revolution will be powered remains an open question.
Batteries are currently the only practical solution for powering autonomous embedded systems. They have, however, intrinsic limitations in term of miniaturization, as well as well-known environmental downsides if adequate disposal is not available. Alternatively, self-powered integrated systems that harvest energy from their surroundings — that is, ambient waste heat — could maintain the advantages of batteries (such as low cost, easy installation and topological flexibility) while offering additional benefits (such as having long lifetimes and being potentially biodegradable and maintenance-free). They are thus considered the ultimate solution for IoT. Writing in Nature Electronics, Mark Lee and colleagues2 now show that standard silicon complementary metal–oxide–semiconductor (CMOS) technology can be used to build effective energy harvesting devices (Fig. 1).
The researchers — who are based at the University of Texas at Dallas and Texas Instruments Inc. — cleverly integrated silicon nanostructures into on-chip thermoelectric micro-generators by taking advantage of the ability of CMOS processing to fabricate ultrahigh density devices with very small thermal and electric contact resistances. As a result, they were able to integrate an astonishing number of thermocouples — tens of thousands per square centimetre — while keeping a reasonable thermal gradient across the device and a perfect match of the internal resistance to the optimum value that maximizes the power density.
With the approach, Lee and colleagues obtain power densities as high as 29 μW cm–2 K–2 near room-temperature, which compares well with the performances of bigger generators based on exotic elements. Moreover, they provide clear design and fabrication guidelines to tune both the thermal and electrical properties of their nanostructures with conventional CMOS strategies. These strategies include using controlled doping levels to adapt the contribution of different segments to the total resistance and using packing fraction engineering to reach low levels that maximize the temperature differences between the hot and cold sides of the structures.
To date, one of the limiting factors in the development of such micro-power sources has been the inherent incompatibility of good thermoelectric materials, such as bismuth and lead tellurides, with silicon technologies. This, in turn, leads to devices with a large footprint and high thermal and electrical impedances, which dramatically reduce the power generated.
Back in 2008, proof-of-principle experiments were reported on the enhanced thermoelectric properties of silicon nanostructures3,4. Since then, the thermoelectrics community has dedicated considerable effort to understand and develop such low-dimensional structures. However, while the effect of nanostructuring on the thermal conductivity of silicon has been well explored, little attention has been paid to the integration of these nanostructures into practical devices5,6. Earlier work on the integration of silicon nanostructures has focused on nanowire-based devices (fabricated both using top-down and bottom-up approaches), showing compatibility issues with mainstream technology and a limited performance due to low thermal gradients. Thus, the sophisticated integration reported in the work of Lee and colleagues represents an important step toward creating practical devices. It also shows that a substantial specific power can be generated on-chip with this type of micro-thermoelectric generator, and at a very low cost.
In terms of the development of energy-autonomous micro-systems that can power IoT devices, the integration of this, or other harvesters in silicon technology still needs to be combined with the miniaturization, and eventually the integration, of small rechargeable storage units. Such integration can be considered the other major challenge in powering the IoT revolution. While work on the silicon integration of lithium-ion batteries has begun7, the development of these devices is still in its infancy. Looking further ahead, the coupling of harvesters and storage units in a single fully-integrated system would provide a new family of devices, which can be termed harvestorers. The low-cost batch production of these harvestorers would represent the last step in the desired IoT power revolution.
The Internet of Things (MIT Technology Review, 2014); https://go.nature.com/2J2IOYz
Hu, G., Edwards, H. & Lee, M. Nat. Electron. https://doi.org/10.1038/s41928-019-0271-9 (2019).
Boukai, A. I. et al. Nature 451, 168–171 (2008).
Hochbaum, A. I. et al. Nature 451, 163–167 (2008).
Dávila, D. et al. Nano Energy 1, 812–819 (2012).
Li, Y. et al. IEEE Electron Device Lett. 32, 674–676 (2011).
Notten, P. H. L., Roozeboom, F., Niessen, R. A. H. & Baggetto, L. Adv. Mater. 19, 4564–4567 (2017).
About this article
Cite this article
Tarancón, A. Powering the IoT revolution with heat. Nat Electron 2, 270–271 (2019). https://doi.org/10.1038/s41928-019-0276-4
Advanced Materials Technologies (2020)
Journal of the American Chemical Society (2020)
Chemical Reviews (2020)
Scaling of Power Generation With Dopant Density in Integrated Circuit Silicon Thermoelectric Generators
IEEE Electron Device Letters (2019)