Energy Harvesting Update

One-on-one with Georgia Tech’s Manos Tentzeris on what’s possible today and what will unfold over the next five years.


Manos Tentzeris, professor of electrical and computer engineering at the Georgia Institute of Technology, sat down with Semiconductor Engineering to discuss energy harvesting. What follows are excerpts of that conversation.

SE: What is the state of energy harvesting and are we making progress.

Tentzeris: The latest results are systems with efficiency up to 40% to 45% utilizing ambient UHF (TV) energy. That’s currently extending to other areas such as WiFi, 3G/4G/5G telephone energy. We’re seeing the ambient UHF-harvesting systems being used for chips running up to 3.5 volts utilizing the energy just from a single TV station.

SE: How do you get that energy? What’s the process?

Tentzeris: We have developed a rectifier with a combination of charge pumps and folded dual-band antennas that are monopole- or dipole-based or even three-dimensional, lately utilizing “origami” and 3D printing approaches. These architectures are not extremely complicated and are easy to realize. The challenge is more about an accurate energy spectrum variation and efficiency model between the antenna and the rectifier. You get the energy from the base station, which is four, five or six miles away. That is in the UHF range, say 500 MHz. Typical amounts of ambient energy levels in these frequency ranges is -20 to -35 dBm. It’s small. It’s in the range of a microwatt. But there are numerous applications we can utilize this microwatt-level harvested energy. If you get to 1.8 or 2 or 3 volts DC output voltage, that’s equal to the voltages powering up most typical IC’s. Or we can utilize this energy to charge a supercapacitor, and then the supercapacitor discharges to activate the IC.

SE: That’s the duty-cycle approach, right?

Tentzeris: Yes. And we can guarantee we can generate these amounts using very small fluctuations in the ambient noise from a TV or radio. That’s one approach for energy harvesting.

SE: What will this be used for? Will it be used for a smart phone, or a medical device?

Tentzeris: You can integrate it into wearables, ‘smart skins,’ machine-to-machine (M2M) and Internet of Things (IoT) devices, and that will solve the very critical powering problem. If you’re measuring temperature or blood pressure, or in the near future, glucose, that’s going to be one of the first applications. For implantable devices, that’s a different story. A more practical approach will be some form of near-field wireless power transfer.

SE: If you’re using any inductive generation, it adds too much heat.

Tentzeris: The brute force systems suffer from three major issues. One is the high loss. The second is a dramatic loss of transfer efficiency. You can have a drop in efficiency from 80% to 40% or 50% for minor misalignments due to body motions. In order to address the two previous ones, you have to provide lots of power. Using an ambient approach, you might use one to two orders of magnitude less power for an implantable device. For the traditional transfer systems, using one or two coils, that’s an issue. For other markets , that is not an issue anymore.

SE: Are you optimistic this kind of power generation will roll out soon?

Tentzeris: For IoT and implantable devices, we’ll start seeing some systems in the next 12 to 18 months.

SE: How about for wearables? And do we potentially need different batteries?

Tentzeris: The addition of flexible supercapacitors (supercaps) will play a very significant role in the activation of wearable sensors. But to power more complicated wearable devices you will need a different kind of energy harvesting, like the near-field charge from holding a cell phone or two-way radio in your hand. We have already demonstrated that you can activate some LED dials or other systems using near-field energy harvesting.

SE: How much of this is dependent on better architectures and designs, or do we still need different approaches?

Tentzeris: If you can develop reliable flexible systems without worrying about the use of a bulky and heavy battery, that will make it much easier to design wearable devices. Typical systems now have inductors, ICs, packages, and so on. For harvesting, they are amenable to other types of manufacturing. You can use plastic or paper substrates. These can be implemented in a very small form factor that will fit into a package of ICs. Right now, you can bring these components on top of a bare IC, and you can integrate all of these sensors.

SE: Can that be done with current manufacturing techniques?

Tentzeris: Most of this can be done with additive manufacturing, such inkjet and/or 3D printing.

SE: What will this add to the overall cost?

Tentzeris: It depends on the form factor. Currently, You can duty-cycle activate systems in the range of milliwatts. For a smart phone, you need between 9 and 12 watt-hours, with an average of 4 to 5 watts. For the time being, these kinds of power levels are prohibitive for ambient energy harvesting. We are about 2 to 3 orders of magnitude below that.

SE: Is it going to be doable in the next 5 years, or will it take 20 years?

Tentzeris: There are a lot of efforts under way. The efficiency will grow slowly within the next five years, and that will be a solution for low-power devices. There is also work on kinetic and thermal energy harvesting, too. Couple that with ambient wireless energy for wearables—let’s say you have harvesting integrated with that—that could be another 150 milliwatts. So you may have a real solution in the next four to five years. But you have to combine different forms of harvesting. No one alone will be sufficient for the complete powering of practical wearable devices.

SE: What we’re hearing from the design side is that power consumption will go down with different architectures, process technologies and materials.

Tentzeris: That is true. Over the last 18 months, there have been numerous IC’s with a much lower activation voltage as well as in microwatts (‘on’-state) and nanowatts (‘off’ state), as well as better quality diodes and supercaps. In addition, additive manufacturing (inkjet and 3D printing) coupled with ‘origami’ principles have enabled novel miniaturized flexible shape-changing antennas with unprecedented efficiency characteristics for non-interrupted energy harvesting. Very soon, we will start seeing a ‘LEGO’ assembly and combination approach of all these technologies for minimum power consumption under time-varying ambient conditions.

SE: What happens with wearables like the Apple Watch, where you only get two to four hours of active use?

Tentzeris: Efficiency will keep improving, taking advantage of the technologies above, although the ever-improving intelligence on the wearable smart watch will impose more and more stringent power requirements. Still, we are going to see simple battery-assisting kinetic/ambient energy harvesting ways to push two to four hours up to six to eight hours very soon. To move to two or four days, though, will require the combination of more harvesting forms as well as the drastic improvement of energy storage capability of commonly used batteries and supercaps.