中文 English

Energy Harvesting Shows New Signs of Life

Focus is shifting to helping batteries rather than replacing them.


Energy harvesting is seeing renewed activity in select markets, years after some high-profile attempts to build this into consumer electronics stalled out.

Costs, manufacturing challenges, and market resistance kept this technology from moving forward, more than a decade after it was being touted as the best way forward for consumer electronics and devices that were hard to access. While solar, wind, and hydroelectric technology has proven successful for powering large factories and even cities, harvesting enough energy to power smaller devices largely fell flat.

“I always thought this looks promising, but it never really got traction,” said Joe Ward, e-peas’ senior director of sales and business development for North America. “But I think it’s getting critical mass now. To really appreciate energy harvesting, there has to be some pain associated with a battery. Customers need to have an issue with the longevity and the hassle of replacing it, as well as disposing of it.”

While not being adopted universally, there are specific applications where energy-harvesting uptake is underway today. Consumer applications are likely to be limited to wearables, but some industrial markets are fueling a resurgence — as long as there’s a payoff. Long-term manufacturing improvements may increase the application reach.

“The promise of energy harvesting microcontrollers, which run forever without maintenance, is still very exciting,” said James Myers, distinguished engineer at Arm. “But a number of challenges remain and have slowed general adoption outside of a few niches.”

Christian Bretthauer, principal engineer for Infineon’s PSS RF Division, pointed to where the action is likely to be. “Sensor networks that do not need to be wired and do not require battery replacements are the most promising drivers,” he said.

That’s the promise, at least. “There’s no IoT sensor out there that’s powered by energy harvesting — not today, other than typically small-area solar/photovoltaic, some thermal TEGs, and large apple-sized electromagnetic vibration harvesting to power wireless sensors in industrial IoT and on trains,” said Robert Andosca, CEO of Inviza. “There are no MEMS-based harvesters deployed in large scale by anyone.”

Defining energy harvesting
Energy harvesting refers to the ability to acquire operating energy from some aspect of the operating environment. It could come from light, temperature differentials, vibration, RF waves, or many other physical phenomena. Typically, there is enough energy only for a low-power circuit.

Fig. 1: Energy harvesting works for low-power circuits, ideally replacing the battery, but, more realistically, extending the life of the battery. Source: Renesas
Fig. 1: Energy harvesting works for low-power circuits, ideally replacing the battery, but, more realistically, extending the life of the battery. Source: Renesas

There are qualitative differences for how energy is generated. One involves true harvesting or scavenging of ambient energy sources that are free. Solar is a good example of that. The energy being captured would otherwise have been wasted.

A slight variant on that might be thought of as energy “poaching,” where energy is siphoned off from some other part of a system. For example, one might use the deformation of tires as a way to power tire pressure sensors. Here the rotation of the tires comes from the motor that has a major source of fuel or energy, so this pulls a tiny bit of that energy. In theory, that could increase the friction of the tire, increasing the load on the engine, but the expectation is that any such effect wouldn’t be noticeable. The energy isn’t strictly “free,” but it’s close enough to be considered free.

A common form of energy harvesting is large-scale photovoltaic (PV) technology. This harvests energy from the sun, but it’s a very different market from the rest of the energy harvesting world, which tends to sip small amounts of power to run small sensors and other circuits. So for practical purposes, PV for the energy grid is usually considered to be separate from energy harvesting.

A resurgence
Energy harvesting rode the same wave that MEMS and sensors rode a decade or so ago. That was partly fueled by ideas about using MEMS technology for energy harvesting. But energy harvesting can involve many technologies besides MEMS. They all generated excitement, technology ideas, and led to the creation of companies that never really got traction.

The promise of energy harvesting was — and still is — the ability to equip sensors and other devices that will be installed in remote locations with a way to get power from their surroundings. The real benefit is in eliminating the need to physically change batteries periodically. These devices would be self-powering, dramatically reducing their management cost.

One of the major hurdles, however, has been the cost of an energy harvesting implementation as compared with batteries. “For the consumer electronics space, they are talking about the sub-dollar or up to $1,” said Ken Imai, senior manager, product marketing, IoT and infrastructure business unit at Renesas. “Energy harvesting is three or five times higher.”

For consumer applications, device makers simply can’t justify the added cost on a price-sensitive device. Yes, it would be nice for that smoke alarm not to chirp for a new battery at 2 a.m., but consumers may balk at the price if energy harvesting is used. “I can do this 50-cent battery thing, or I could spend three bucks on energy,” said Ward, describing the system-maker’s dilemma. In the end, cost wins.

So the reality of costs against what customers would pay have made it look like the technology would never quite pencil out. And things mostly went on hold.

This was particularly the case with MEMS-based energy harvesting, which is usually built on silicon wafers using circuit-building techniques. As with chips, the number of yielding dies per wafer drives the economics. The typical approach to that with circuits is to continue to scale the die size to smaller dimensions. But the physics of energy harvesting requires a certain size in order to obtain good power density.

“Anything energy harvesting is area-dependent, for the most part,” said Andosca. “You can shrink them only so far. And then you hit the physics wall of power density limitations.”

That places a limit on how small an energy harvesting device can be. For expensive silicon-processing techniques, it can be hard to make the numbers work. In one example, a cantilever used to harvest vibrations saw cost capped at $10 when projecting cost reductions forward.

Use of other materials like glass for a substrate might help, while large-format processing is another option. “The flat panel display industry uses the same exact techniques to build their panels as the semiconductor industry uses, or the MEMS industry uses to make sensors,” said Andosca. “They just use different sizes of substrate and some different substrate material types.”

Eliminating batteries — or extending them
Today there appears to be a quiet renewal of interest in energy harvesting — especially in the smart-city and industrial markets. One reason is that batteries are wearing out faster than expected, and replacing them is becoming more expensive. That brings the break-even point for using energy harvesting closer.

“People are realizing that batteries don’t last as long as they say they do,” explained e-peas’ Ward. “For one reason or another, they just can’t get to that end goal of a five-year battery life. The cost of replacement is growing, since you need to roll trucks or send people out to replace the batteries. If you’re doing tens of thousands of these sensors, it’s a nightmare. Then they’ve got to dispose of the dead batteries.”

One challenge is that batteries fail at very different ages. “Some of those batteries may last four years, but the customer must start planning around the first failures, which cuts the expected product life in half (or more),” he added.

What’s inside
There are two critical pieces to an energy harvesting solution. The transducer is the real generator, and there are many different types. Some of them, like thermo-electric generators (TEGs) and PV cells, generate DC voltage. Others, like vibration harvesters, will generate AC power that will require rectification to be useful.

“In outdoor applications, PV is still the easiest, cheapest and most reliable option,” said Bretthauer. “On everything that is moving, vibration harvesters can be interesting. And on machines that produce a fair amount of heat, thermoelectric harvesters might also be a good option.”

The second piece of the solution is a power-management chip, or PMIC. E-peas refers to this as an ambient energy manager (AEM). This will manage the power generated by the transducer, although it may leave rectification and voltage shifting to the transducers, requiring a clean source at its input. There can be different versions for different transducers, since the physics of transducers determines what voltage they generate.

Fig. 2: Energy management helps to use harvested energy to power circuits as well as for directing excess energy into storage. Source: e-peas

Fig. 2: Energy management helps to use harvested energy to power circuits as well as for directing excess energy into storage. Source: e-peas

While energy harvesting mostly has been discussed for its ability to eliminate batteries, it has long been a challenge to create harvesting circuits that do better than breaking even. Those circuits themselves require power, and if they use all of the power being generated, they’re not useful. Circuits can be successful only if they can handle the power management with enough power left over to handle the operational load.

That can be tough, depending on the kind of load it’s driving. In many cases, the energy harvesting serves not to eliminate the battery entirely, but to supplement it, extending the life of the battery. It then becomes the job of the PMIC to use the harvested energy first, turning things over to the battery when necessary.

In such setups, there typically would be a second form of energy storage. Energy harvesting systems will need to store what they generate in order to decouple generation from consumption. That might require supercapacitors or rechargeable batteries (also called “secondary batteries,” in contrast to one-time use batteries, which are “primary”).

Where power requirements are higher, supercapacitors are better at grabbing and releasing the stored energy faster. Batteries have higher energy density, but their power density isn’t nearly as good. The quality of supercapacitors has dramatically improved, reducing leakage to levels far below what’s possible with standard capacitors.

While some energy harvesting devices are implemented as separate units, they can also be combined with other common circuits. Renesas has a microcontroller (MCU) with energy-harvesting management capability. It can power the MCU as well as other circuits. The VDD pin, which would normally be where power was applied as an input, can be used as an output. The company is creating an ecosystem for various different transducer types.

Fig. 3: An energy-harvesting MCU. Source: Renesas

Fig. 3: An energy-harvesting MCU. Source: Renesas

One of the costliest operations from an energy standpoint can be radio communications. But Renesas said it can power a number of low-power wide-area (LPWA) protocols. The company specifically said that customers have implemented LoRaWAN and LTE Cat M1 radios.

This isn’t likely to work for higher-power protocols – but those protocols also may not make sense for these ultra-low-power applications. “You really don’t need to send and receive a lot of data,” said Ashraf Takla, president and CEO of Mixel. “Mostly, you’re communicating slowly varying control signals, probably between a lot of different sensors and harvesting devices and things like that.”

Dynamic adaptation and data buffering may also be necessary if the power source is inconsistent. “If we have limited energy input from a harvester occasionally, we can reduce the energy discharge by, e.g., limiting the number of radio communications and keeping data in on-chip memory,” added Imai. “When we have enough energy input, we can transfer the stored data.”

Applications matter
As a result of the cost challenges, energy harvesting is more heavily focused on applications where it’s expensive to use primary batteries. There is only one consumer application area generating interest — wearables. The Casio G Shock watch is one example that uses energy harvesting, and clothing will be another.

In order for wearables to be successful, however, the technology has to be invisible – including the generation of power. “I hate charging stuff, and I’m not going to charge my pants up,” said Andosca.

The idea is to derive energy from the bending of belts, or the pressure of feet in shoes, or any of the other movements that we perform. Thin packaging may contribute to the functioning of piezoelectric harvesters, creating what’s referred to as a bender.

Agriculture presents other opportunities. One example is the use of motion sensors on the ears of cattle, which can help to alert ranchers when cows come into estrus. This is a clear example of a sensor that provides great utility but presents serious limitations for changing a battery. Similarly, stationary sensors for soil conditions would also benefit, since they would be placed in remote locations that would be expensive to service.

There are also select industrial devices that could benefit from harvesting. While most heavy industry features enormous availability of wall power, being able to install sensors or devices in hard-to-reach places with no nearby plugs can provide ease and flexibility. Imai mentioned an electronic faucet, for example, which can leverage a turbine in the water pipe or the temperature difference between hot and cold water for power.

Health care provides yet more opportunities for energy harvesting. “Everybody wants to seal the equipment and not have wires or batteries going in and out,” said Ward.

Other outdoor applications include environmental sensors, like weather stations, as well as structural-health sensors. The latter can be used to help maintain critical infrastructure like oil and gas pipelines, or bridges. Sensors likely would be installed in many hard-to-reach places, making battery replacement impractical.

Necessary battery life also varies by application. “Some agriculture products are looking for six months,” said Ward. “With the smart-building or the smart-city stuff, they’re trying to get a couple of years.”

Automotive opportunities
Automotive is another area where harvesting can help to fuel the electrification of vehicles. “When you’re looking at consumer electronics or a building, there’s a different meaning of energy harvesting, compared to when you talk about a car or a vehicle,” said Puneet Sinha, director of new mobility, mechanical analysis division at Siemens EDA.

Regenerative braking represents the best-established such approach in hybrid vehicles today. “Since 2009, Formula 1 has been using Kinetic Energy Recovery Systems (KERS or ERS-K) for kinetic energy recovery leveraging an electric motor/generator unit to turn the kinetic energy of the car during braking into electrical energy,” said Robert Schweiger, director of automotive solutions at Cadence.

Other opportunities to pull energy from various systems are being explored. “Harvesting energy and recovering it also can be done by converting the kinetic energy given off by shock absorbers,” he added. And there have been those attempts to use tire deformation to power tire pressure sensors.

But more refinement may be needed before deployment. “These applications (other than regenerative braking) are still in the university research phase,” said Sinha. “They need to be efficient, but more than that, they need to make commercial sense in terms of how much extra cost you will have to add to the system versus the benefit you will get out of it.”

Another possibility is to exploit temperature differentials, although electric vehicles generate much less heat than internal-combustion (IC) engines. “Batteries are a lot more efficient than internal-combustion,” said Sinha. “In an IC engine, maximum efficiency is somewhere around 35%, whereas batteries are almost 90% efficient. You are not playing with that much heat, so you have to remove less heat, but then you have less heat to play with.”

In another example, batteries, which don’t operate well when cold, can be heated to improve performance. “There are companies looking to put a converter in the battery pack so that they can take that heat and use it toward heating the battery,” said Sinha.

While these are all helpful possibilities, they’re not going to be the breakthroughs that push the travel distance forward. “They can make a few percent difference, but they’re not going to take the range from 250 to 500 miles,” Sinha cautioned.

In automotive perhaps more than any other application, energy-management algorithms are critical. “Think of the Tesla Model S and the Audi e-tron,” Sinha said. “They are the same vehicle size-wise and the class they are in, and both have the same amount of battery energy in them. But one vehicle gets 30% to 50% more range than the other one.”

“Going forward, more and more companies are investing to update the software on how efficiently you can extract the energy out of the battery,” he continued. “That energy was always lying there, but how effectively can you extract it out?”

Looking to the future
These developments are proceeding with little fanfare, and it’s truly still early days. As product launches and market development take place over the next year or two, we may see energy harvesting finally reaching a level of commercial viability that was unobtainable in the past.

That said, the energy density of batteries has gone up by three times in the last decade, and the costs keep coming down. That makes the economics of energy harvesting a moving target — in this case, one that gets harder over time.

All in all, Arm describes three critical challenges when looking into energy harvesting. “The first challenge is about application-specificity. You can’t use solar cells in a dark cellar or thermoelectric where there is no temperature gradient,” said Myers. “Secondly, there is power variability. With a small 1 cm2 cell you might harvest milliwatts outdoors, but only microwatts under office lighting. Thirdly, there is low power density: Harvesters have lower power output than a similar-sized battery.”

But for applications where these three challenges can be met, energy harvesting is looking promising again.


Diogene7 says:

I would think that one interesting opportunity would be to integrate at least one energy haversting technology on smartphones as a small security back-up power only for emergency calls : if the battery of your smartphone (iPhone) or smartwatch (Apple Watch) is flat, you would always be able to harvest energy for a 2mn call sending with GPS location. It should even be a mandatory feature in my humble opinion…

Leave a Reply

(Note: This name will be displayed publicly)