What is Energy Harvesting (EH)?

Energy system with controllers and converters
Even after a legacy of supporting technologies, energy harvesting (the ability to scavenge energy from any source that physics affords us) has only graduated from emerging to emerged. This can be attributed to the drastic reduction in system power budgets resulting from Moore’s law and MEMS. Can you imagine an entire sensor network, processor, and wireless communication (including a display!) with a system power budget of only a few milliwatts? Now, you no longer have to.

Waste vs. Opportunity

“One man’s trash is another man’s treasure.” People in electronics spend a whole lot of engineering cycles, expense, and even energy dealing with “waste heat,” especially in application spaces such as data centers and others that utilize a lot of space and power for thermal mitigation solutions. This paper instead looks at such heat sources (along with other energy sources) as opportunities for energy capture or recycling.

The discussion around waste has always revolved around how to get rid of it, but with the perspective of any available energy as a potential power source, the outlook and the approach for handling waste completely changes. Whether in the U.S. or globally, roughly two-third of all raw energy sourced is assorted in the “waste heat” category, which actually represents a tremendous opportunity, as we will continue to explore in this discussion.

2020 U.S. energy flow chart

Fig. 1: Complete U.S. Energy Consumption with “Waste Heat” Component Highlighted [1]

Source vs. Load

In a previous blog on the Internet of Things (IoT) and Industrial IoT (IIoT), the importance of finding the right balance between available energy (e.g., system power source) and load (e.g., system power budget) was discussed, and major emphasis was put on the need to focus on reducing loads much more than simply looking for a bigger source. A quick refresher is demonstrated with the figure below:

Relationship of Available Power to System Power Budget

Think of your energy source as a denominator and your system power budget as the numerator. The viability of your system is the inflection point at which these two variables meet, and one can decrease the numerator much faster than one can increase the denominator. In other words, when the ratio becomes <1.
Energy Harvesters vs. Electronic Devices
Fig. 3: Relationship of Source to Load in EH Terms
The critical takeaway here is that making your system, and therefore your application, viable from both technical and economic perspectives exhibit far more dependency on a designer’s ability to reduce system power consumption and implement intelligent power management (IPM) techniques than on finding a bigger, denser battery and/or more efficient power converter solutions.

Battery capacity only doubles roughly every decade, whereas integrated circuits (ICs) and even microelectromechanical systems (MEMS) sensors can halve power consumption nearly every other year and still increase functionality, proving the superiority of researching power management over relying on battery improvements.

What is Energy Harvesting?

Energy harvesting (EHT) is the capture and conversion of free, ambient energies from the environment around us. For this reason, it is also known as “energy scavenging.” We can harvest every type of energy physics affords us, which includes:

  • Solar (thermal, photovoltaic or PV),
  • Kinetic (electrodynamic, vibrational),
  • Thermoelectric, piezoelectric (mechanically-induced),
  • Radio frequency or RF (near-field, far-field), and
  • Triboelectric (static electricity).

Any transducer is effectively a potential EH source, but this paper shall focus on solutions optimized for EH applications.

EH modalities tend to get grouped into categories of ac-based or dc-based according to their native outputs. The most common dc-based EH devices are PV and thermoelectric. The most common ac-based EH devices are piezoelectric, RF, and other mechanically dynamic EH sources such as vibrational, turbines, and magnetically driven electrodynamic transducers. Each of these sources has its own set of nuances and challenges, so please note that all power management integrated circuits (PMICs) are not a one-size-fits-all circuit for any EH source.

EH has come a very long way over the past couple decades in terms of technological advancement as well as penetration into practical applications. A decade or so ago, particularly before the major emergence of the IoT/IIoT, there were some common misperceptions about EH technologies. The first misconception was that EH produced a negligible amount of power for any useful applications. We shall explore this issue in more depth momentarily, but it is aligned very closely with the previous discussion about source vs. load.

Another misperception was that EH was an academic lab experiment lacking a production supply chain and a supporting ecosystem for high-volume developments. Many of the constituents of what eventually became the Power IoT ecosystem (outlined in more detail later in this discussion) have existed for many years (even decades) and come from industrial sources (from fresh startups to the biggest semiconductor companies) even before they were well known or coordinated in an ecosystem as they are today. A successful, EH-based system will require many technological contributions from EH transducers to energy storage (battery management systems [BMS]), power management integrated circuits (PMIC), and optimal loads (microcontrollers, memory, radios, sensors, displays, etc.).

Arguably, the most fundamental part of these components is the PMIC since it essentially acts as the “brain” of an EH-based system by controlling the EH power extraction and power management, and even integrating the BMS, all in one easy control IC (almost always in a sub-10 x 10 mm footprint). EH PMIC solutions have been on the market from companies like Texas Instruments and Linear Technologies (now Analog Devices) for around two decades, a little too early for the IoT boom. An example of a typical PMIC block diagram can be found below in Figure 4. You can even find off-the-shelf IoT modules today.
Energy harvesting components and storage options
Fig. 4: EH PMIC Functional Block Diagram Example [2]
Perhaps the biggest misperception around EH, and therefore the critical path inhibitor to its mainstream adoption, concerns its cost. Cost evaluation can be a bit tricky because any kind of cost-benefits or total cost of ownership (TCO) analysis is highly dependent on the system, application, and operating environment.

This complexity and application-specific analysis is the main reason why EH technologies have been regarded as not being economically viable in many use cases.

A common pitfall is to focus purely on a first-order analysis comparing bill-of-materials (BOM) cost before and after the application of EH technologies.

At first glance, replacing a coin cell battery that may cost <$1 (in high volumes) with an EH solution costing several dollars may not seem very logical, even with the benefits of sustainability, increased reliability, and self-powered deployments. But one must consider the second-order (and beyond) cost impacts to do a proper analysis.

While a detailed overview of such analyses is beyond the scope of this discussion, a simple measurement to consider is the replacement/maintenance costs. If that cheap coin cell must ever be replaced, then it will require human intervention, for example, a truck roll and/or special equipment to reach difficult spots in harsh environments, and which would mean the BOM cost savings will have been lost by many orders of magnitude.

Additional cost savings achieved by the use of EH technologies, in terms of capital and operational expenditures (CAPEX & OPEX), are discussed further down.

What is not EH?

Wireless power transfer (WPT) is common these days in low-power, untethered applications and used to wirelessly charge devices from smartphones or watches to toothbrushes. Unfortunately, WPT is also commonly associated with, or in worst case misrepresented as, EH when this is rarely the case. There are very few WPT applications designed to capture ambient, far-field RF energy for other energy source applications.

The far majority of WPT usage is really just the commutation of power from a directed (typically off-line or wall) source, where one of the “wires” just happens to be a wireless link. So, instead of RF “harvesting,” it is really a very inefficient power transfer from the wall. That said, WPT certainly has a place in the Power IoT ecosystem and can be highly complementary to many EH applications and use cases, but one should have a good understanding of the semantics and implications nonetheless.

For example, WPT shines in applications in which a viable energy source is simply not available and/or accessible. Examples of such applications include wireless sensor networks (WSN) embedded in structures and/or harsh environments (or even in the human body). Both WPT and EH share a very important objective, that is, the mitigation of primary (a.k.a., non-rechargeable) batteries.

Objectives of EH

There are numerous objectives and advantages to be achieved by the utilization of EH (and related) technologies. From a short-term perspective, the immediate need to curtail and mitigate the explosive growth in the number of primary batteries due to exponential growth of IoT/IIoT devices is a critical objective of EH. If the hundreds of billions (or possibly the trillion) devices predicted to be deployed this decade come to fruition, we are looking at disposing ~100+ M batteries a day with all their hazardous materials going mostly into landfills.

The more of these systems we can convert to use secondary (a.k.a., rechargeable) batteries (or other energy storage such as supercapacitors or ultracapacitors), the greater the benefit for sustainability. The long-term perspective of EH technologies envisions a utopian world in which many (if not most) of the devices in use are completely powered by free, ambient energy sources. Reliability is another very important benefit EH brings. This may seem counterintuitive as many folks will seize on the concept of a potentially-intermittent energy source as being in direct conflict with system reliability and uptime requirements.

While it is very true that many EH sources can be intermittent and unpredictable (i.e., light, wind, movement, etc.), many of the energy reliability concerns can be addressed with a good understanding of the operating environment and an ensuing system design that utilizes the appropriate type and amount of EH, typically combined with energy storage, and IPM to ensure an acceptable probability of success.

Hybrid EH implementations can also help address this and are explored below. Software (SW) is another area that can make a big difference since classical WSN/computing systems are not designed with an intermittent energy source in mind, but if the SW is designed to be energy aware and knows to store critical data when reaching a shutdown threshold and pick back up where it left off when crossing a turn-on threshold, then applying an EH power source can be greatly beneficial and effective.

The flip side of the reliability analysis is that a system enabled with an EH power source(s) and rechargeable/renewable energy storage can be a perpetually, self-powered system that eliminates many components that tend to be the most common causes of system failure related to quality and reliability. Eliminating as many mechanical/moveable components (i.e., connectors, battery holders, switches, wires, etc.) as possible is a big step in the right direction. Reducing overall system power to mitigate thermal stresses also facilitates increased reliability. Of course, energy storage components, such as secondary batteries, will inevitably lose capacity over time, though this can be improved with solid-state solutions and become a negligible concern with the use of capacitive energy storage.

These are worthy objectives and goals, but at the end of the day, cost objectives are the dominant factors in getting a real product to the market, and EH technology seem in conflict with the cost factor. As mentioned above, there are gaps in the payback period calculations for EH technologies because TCO has not expanded far enough beyond the first-order, BOM cost comparison. The “Supplemental Power” section below discusses further on how EH can help expand cost savings into CAPEX/OPEX beyond the direct, system cost metrics.

Common Forms of EH

The world of EH encompasses the full power spectrum from nanowatts to gigawatts. This paper is mostly focused on the lower end of the spectrum here since its discussion mainly revolves around the many WSN/IoT/IIoT applications. However, we must not forget all the PV panels and wind generators leading the charge for sustainability and the fight against climate change, generating a respectable, and increasing, portion of global energy.

This means there is quite a spectrum of types of EH solutions. PV cells and wind generators can generate megawatts of power, but there are also PV cells that fit in the palm of your hand and are optimized for indoor light sources and electrodynamic generators that fit on a bicycle wheel or even exist at the MEMS level. A full-on deep dive on EH modalities, operating principles, and overview of available solutions is a much bigger discussion for follow-up materials, but a sampling of EH transducers, supporting components (such as energy storage), and low-power applications can be found in Figure 5 below:

Electronic devices collage

Fig. 5: Assorted EH Transducers and In-situ Applications. a) An indoor-PV-powered, WSN, courtesy of Tyndall National Institute [3]; b) Cutaway view of a thermoelectric generator (TEG) [4]; c) Vibration energy harvester for low-frequency applications [5]; d) Flexible PV array for portable charging, courtesy of PowerFilm [6]; e) 1.5 mm3 complete EH-powered WSN, courtesy of University of Michigan [7]; f) Cutaway of supercap on silicon design, courtesy of Tyndall National Institute [3]; g) Flow-powered (hydrodynamic) shower temperature indicator, courtesy of Würth Elektronik [8]; h) Fully-integrated IoT power module [9]; i) EH PMIC embedded in PV-powered, high-end watch, courtesy of e-peas [10].

A nice graphical comparison to facilitate matching EH sources to common loads can be found in the figure below. Note this neither covers the full spectrum of EH nor applications, but it gives a pretty good idea of some very common applications and the sweet spot for matching EH modalities.

Even more importantly, it gives a more realistic match to the EH because it specifies the amount of EH needed to support the stated application. For instance, it is not enough to say one can use PV for an application, but rather state the approximate PV panel size along with the light source irradiance (i.e., indoor vs. outdoor, light/lux level, etc.). The same principles applies to thermoelectric generators (TEG), which are not only dependent on the size of the TEG, but also the temperature differential across it.

Diverse power consumption graph

Fig. 6: Comparison of Common Applications with Common EH Sources & Power Levels [12]

Hybrid & More Exotic Forms of EH

One great way to take advantage of EH technologies and maximize the application space coverage is via the utilization of multiple sources concurrently. This can be done to try and maximize the available energy at any given time by exploiting different EH sources that exist in the same setting.

Many EH PMICs are designed to accept at least two different EH inputs and even support multiple forms of energy storage (i.e., primary and secondary battery). For instance, a WSN mounted on a motor in a lit environment can take advantage of vibrational harvesting and PV inputs. Another reason to design for multiple sources is to mitigate energy intermittency in changing environments, particularly for untethered applications. For instance, a device worn on the body might incorporate both PV and a TEG so that it can extract energy from the sun when outdoors (it may also extract from the TEG, but how much depends on the outdoor temperature differential), then extract further from the temperature differential between indoor ambient and body temp when indoors.

Hybrid forms of EH exist as well. This is different from the use of concurrent sources as described in the last paragraph. There are EH modalities that take advantage of more than one harvesting principle in a single transducer. An example of this is a solution that does RF harvesting of heat energy (e.g., infrared RF energy) captured from a solar source. These kinds of EH solutions are far less common. On the emerging front, there are promising forms of EH that seek to address today’s challenges whether they be related to energy density, cost, or form factor. As a reminder, the value proposition of EH increases much faster over time due to the reduction of system power budgets than the increase in available energy from the EH transducer and power extraction/storage circuit combo.

This means that an EH source can effectively have very little application today, but much wider application coverage tomorrow, even if the amount of available power density for a given transducer does not change. Triboelectric nanogenerators (TENG) are a good example of this because it is hard to extract much from frictional, static electricity sources (~10s to 100s of μWs), though you can quickly generate extract enough energy to charge a capacitor and power some LEDs. These days, if you can power an LED, then you can also power a WSN with sensors, a microcontroller, and a radio.

Most emerging focus is in PV (or closely related) EH sources. The latest (third) generation of PV devices to be fully commercialized is organic PV in which cells are tuned to specific wavelengths (typically optimized for either outdoor or indoor use even if functional in both environments) and find a nice balance between cost, efficiency, and flexibility. These value propositions look for additional enhancement in the next generation of PV cells that incorporate the mineral perovskite, which are still in a pre-commercialization status [13].

Supplemental Power

When considering EH for your application, it is very important to bear in mind EH is not an all-or-nothing proposition. Designers can be quick to dismiss EH technologies for their applications because they cannot see a 1:1 relationship between their existing power source (typically a primary battery) and an appropriate EH solution.

But there can be a lot of value in supplemental and standby applications. Of course, the value assessment can be very specific to the application and operating environment, as stressed multiple times in this discussion. For instance, if EH can be used to extend battery life by 20–30% before requiring a recharge/replacement, then that could be the difference between making it through a full day without having to plug in or enabling that mobile system (i.e., drone or electrified vehicle) to stay in the air for a few more minutes or travel an extra 10–20 miles. Alternatively, EH may enable your device to send more frequent and/or bigger data transmissions to the cloud.

In standby applications that tend to have much lower power requirements than their fully-active system power profiles, EH can be used to provide the tiny bit of energy needed to power standby circuits awaiting an external signal to fully wake up. If the system’s main power supply also provides a standby rail for this purpose, then it tends to be costly in terms of size, weight, and power (a.k.a., SWaP) factors so if these supplies are disaggregated, then the added cost of the EH standby circuit may be neutralized (or even provide positive return on investment or ROI) by the savings in the main supply. Examples of large systems with small standby supplies include computer serves supporting wake-on-LAN (WoL) or IoT-controlled devices supporting wake-up radio (WuR).

In WPT applications, the wireless links for power transfer tend to have very poor commutation efficiency, as fundamentally demonstrated by the Friis equation [14], in which power losses are exponential with increasing distance between transmitter-receiver and increasing transmission frequency. Therefore, supplementing receiver power to operate at a lower receive radio sensitivity not only saves power on the receiving end but can save exponentially on the transmitting end as well. As in many EH applications, it should now become clearer how this supplemental approach can extend the savings upstream in terms of both CAPEX and OPEX. What is less obvious is how savings in the micro power realm translates to savings of kilowatts or even megawatts.

To visualize this value proposition, one must consider the full path power takes from generation to end load. For instance, mitigating a few watts at the load in a big server data center may not seem like much, but when you consider all the overhead (power for cooling, inefficient standby supplies), layers of power conversion (each with margin built in that quickly stacks up), layers of redundancy (in both power sources and system loads), energy storage for critical backup, etc., then those few watts can translate to massive savings in OPEX (cost of power) and CAPEX (reduction of overall infrastructure since sizing to a lower, max or steady-state power draw).

Enabling a Greener Future & the Power IoT Ecosystem

We have discussed the numerous benefits, value propositions, application considerations, and misperceptions associated with EH here. We have also given a brief overview on EH devices and the supporting devices/systems that combine to make up the Power IoT ecosystem. Ideally, the ability for EH to enable a more sustainable future is clear in some direct ways, such as eliminating primary batteries and reducing overall energy consumption (and therefore carbon and water) footprints.

In future discussions, we shall look into the less obvious areas of second-order and even third-order impacts of EH, such as the ability to drive improvement in the embodied energy of systems/devices, looking at the complete, comprehensive energy impact “from cradle to grave” or from raw material sourcing through manufacturing to field use and end-of-life and recycling.

The Power IoT ecosystem is constantly growing and becoming an increasingly organized group of stakeholders that are focused on collaborating to bring EH technologies, applications, and education to mainstream markets. A seminal group in defining, organizing, and perpetuating this ecosystem is the Power Sources Manufacturers Association (PSMA) Energy Harvesting Committee (EHC) [15]. The PSMA EHC put out a free, open-access white paper [16] in 2021 that provides an extensive overview of EH for a sustainably-focused IoT.

Designing with EH

For further readings on this subject, there are plenty of existing resources. While all the components of an EH-enabled system (i.e., connecting transducers, laying out PMIC circuits, working with energy storage, etc.) may seem daunting, there are a lot of solutions that help to integrate many of these functions supported by documentation, training, and very capable support teams.

Whether you are an embedded designer trying to optimize raw sensor data capture for big data analytics or only see EH (and typically power in general) as an inconvenient means to an end and just want to focus on your end application/system, there are plenty of out-of-the-box design kits and commercial-off-the-shelf (COTS) modules to facilitate the incorporation of EH technologies, even for technologists lacking the proper knowledge and experience to implement from the ground up. An EH IoT module is a component pre-made for this application space and typically combines the PMIC, BMS, radio(s), and perhaps even some sensors. This enables a designer to quickly connect to the appropriate EH transducer, choose their energy storage source, and provide regulated power rails for system loads all in one fell swoop. Take a look at an example of such a module.

If you want to go deeper and experiment with multiple EH modalities, PMICs, energy storage solutions, and even different radios/displays/sensors, then there are also more comprehensive design kits [17] that incorporate multiple options for many of these system design aspects in a development environment that provides switchable options for all and requires little to no knowledge of the circuits involved so that you can just focus on developing your application.

“Go plant a seed and harvest it!”

References

[1] "Estimated U.S. Energy Consumption in 2020," Lawrence Livermore National Laboratory, March 2021.
[2] “AEM10941," e-peas Product Overview, Viewed January 12, 2020.
[3] Roadmap – Cypress – Solar Powered BLE Sensor Beacon Reference Design Kit, Future Electronics, 2016. [Online].
Available: http://fcs.futureelectronics.com/2016/08/cypress-solar-powered-ble-sensor-beacon-reference-design-kit/.
[4] “How to Build a Homemade Thermoelectric Generator,” 2017. [Online]. Available: https://topmagneticgenerator.com/build-homemade-thermoelectric-generator/.
[5] "Vibration Energy Harvesters," Perpetuum Datasheet, Downloaded October 10, 2017.
[6] PowerFilm LightSaver Max, Accessed January 29, 2018. [Online]. Available: https://www.powerfilmlightsaver.com/lightsaver-max.
[7] D. Pasero, "IoT Sensors Powered by Solid State Batteries and Harvested Energy," Ilika Technologies, APEC 2018 Industry Session, Tampa, FL, March 6, 2018.
[8] C. Ho, "Flexible Energy Storage Considerations," Imprint Energy, 2017FLEX Short Course, Monterey, CA, June 19, 2017.
[9] V. Micelli, "Pavegen - The Future of Urban Energy," IDTechEx US Show, Santa Clara, CA, November 17, 2016.
[10] D. -H. Kim, N. Lu, R. Ma, Y. -S. Kim, R. -H. Kim, S. Wang, J. Wu, S. M. Won, H. Tao, A. Islam, K. J. Yu, T. -I. Kim, R. Chowdhury, M. Ying, L. Xu, M. Li, H. -J. Chung, H. Keum, M. McCormick, P. Liu, Y. -W. Zhang, F. G. Omenetto, Y. Huang, T. Coleman and J. A. Rogers, “Epidermal Electronics,” Science 333, 2011, 838–843.
[11] N. Dahad, “Cartier Uses e-Peas Energy Harvesting PMIC for Solar-powered Watch,” Embedded, January 20, 2022.
[12] "Research Infrastructure Position Paper, European Infrastructure Powering the Internet of Things" EU EnABLES Project, February 2021.
[13] “Champion Photovoltaic Module Efficiency Chart,” National Renewable Energy Laboratory. [Online]. Available: https://www.nrel.gov/pv/module-efficiency.html. Accessed 4/29/22.
[14] Friis Equation - (aka Friis Transmission Formula). [Online]. Available: http://www.antenna-theory.com/basics/friis.php.
[15] PSMA Energy Harvesting Forum. [Online]. Available: https://www.psma.com/index.php/technical-forums/energy-harvesting. [16] T. Becker, V. Borjesson, O. Cetinkaya, et al., "Energy Harvesting for a Green Internet of Things," Power Sources Manufacturers Association (PSMA) White Paper, October 2021.
[17] Würth Elektronik Gleanergy. [Online]. Available: http://www.we-online.com/web/en/electronic_components/produkte_pb/demoboards/gleanergy/gleanergy.php.
Applications