As IoT solutions continue to thrive, so too does the demand for low-energy devices with the ability to operate on natural energy. Questions now arise: What challenges do energy converters face when harvesting energy? What energy options exist for devices? What semiconductor solutions are available to engineers?
What Is Energy Harvesting?
Energy harvesting, as the name suggests, is the act of harvesting energy to power a device. Unlike battery- or mains-powered devices, an energy harvester can extract energy from its surrounding environment and thus removes the need for a power connection. While energy-harvesting solutions are becoming more common, the idea of energy-harvesting circuits has been around for decades.
One such example is a calculator that incorporates small solar cells that can work under ambient light. Such calculators have been around since the 1990s and were only possible thanks to the low-energy consumption of basic calculator chips and LCD displays with no backlights.
For devices that require more power, energy harvesters will often trickle-charge a small capacitor or battery over time, and when a specific charge or voltage has been achieved, the energy is quickly released. While this only allows a device to be operational for brief periods of time, it does allow for that device to be entirely independent of any external power sources (i.e., it can operate in the most remote of places).
As energy sources used by energy harvesters are often limited, devices using energy harvesting have historically been limited to very basic functions. But the introduction of low-powered SoC (Systems-on-Chips) and advanced-power-management IC (integrated circuits) is starting to open the world of electronics to energy harvesters as a viable power source. So, what challenges do energy harvesters face? What power sources exist? And what solutions currently exist for engineers?
Where Can Energy Come From?
When it comes to energy harvesters, just about any energy source that naturally occurs in the environment has the potential to provide power. These energy sources include light, wind, mechanical action, temperature, sound, and radio.
Light is one of the more common sources of energy as it is widely abundant, but it should not be mistaken with solar. Solar panels and solar cells are used to generate electricity when exposed to sunlight (as is the same with an energy harvester); unlike energy harvesters, though, they effectively cease to function when shaded. A light energy harvester, however, is able to operate even in low light levels and are commonly found on calculators and small devices. This gives light-based energy harvesters the opportunity to operate even at night under artificial light (e.g., streetlights and building signs). Additionally, the small size of solar cells used in energy harvesters makes them easy to integrate into a design, and the lack of moving parts helps to improve reliability when used in a remote location.
Wind is another example of a natural source of energy that can be harvested. Like solar panels, there is a clear difference between a wind turbine used to generate sizeable amounts of energy and one that would be miniaturized for use with a single device. An energy harvester using wind would be configured to operate in low breezes caused by air convection, air flow through buildings, and the movement of a device through air. However, even a miniature wind harvester would be sizeable, which is why they are seldom used (if wind energy is desired, a turbine will likely be used).
Mechanical action can also be used as a source of energy for small devices with examples including compression, tension, bending, and acceleration. Typically, a piezoelectric element is used to convert a mechanical force into electricity (e.g., doorbells and pavement). Wearable devices can generate energy from the movement of arms via magnet suspended by springs surrounding a coil. Such energy sources are ideal if a device expects to experience frequent mechanical forces, but the use of moving parts can see wear and tear over time—and inconsistent mechanical energy can make it difficult to use as a reliable energy source.
Energy harvesters can also extract thermal energy, so long as a thermal gradient exists (i.e., both a hot side and cold side). This is usually exploited with the use of thermocouples that generate a voltage when exposed to a thermal gradient. If the temperature gradient is consistent, then it can make for a very reliable energy source, but the low efficiency of thermocouples makes them impractical to implement. As thermocouples are a solid state, there are no moving parts, which makes them reliable for long-term usage. However, the use of a dedicated TEG (ThermoElectric Generator) can help to boost the energy harvested from a thermal gradient.
Radio emissions from human activity also present energy harvesters with a viable energy option. Radio waves from Wi-Fi routers, cellular networks, and radio stations can all be harvested to power devices. In fact, crystal radios located near a radio station can operate without the need for any additional power. However, most sources of radio energy are extremely small, and this can make it difficult for energy harvesters to operate.
What Challenges Do Power Converters Face?
While energy naturally exists in the environment around us, tapping that energy presents major challenges. The first major hurdle that energy harvesters must overcome stems from the fact that naturally occurring energy sources are obscenely small. While solar cells can generate sizable voltages, the energy produced by mechanical, vibration, radio, and sound can be in the nanowatts. Considering that voltage and current are both proportional to power, the generated power will produce either a miniscule voltage or current.
Thus, we arrive at the first major hurdle faced by energy harvesters: energy extraction. An energy-harvesting circuit whose generated voltage is too low will not be able to activate semiconductor components due to the need for forward voltage bias (e.g., silicon diodes will not conduct until a forward voltage drop of 0.6V is observed).
Increasing the voltage can be done by changing the nature of the harvester (e.g., adding more loops to an inductor or increasing the input impedance of the energy harvester), but this results in a reduced current. A reduced current means that more time is needed to charge a capacitor or battery, and this results in a smaller operating-charging ratio.
If an energy harvester can overcome this challenge, the second challenge is reliably storing that energy. Capacitors are small, can be used to store charge, and are able to react quickly to large current demand; they can be leaky depending on the technology used, though. Additionally, the charging circuit will also have some level of inherent leakage caused by non-ideal components. Thus, an energy harvester must ensure that it can charge faster than the rate of discharge from natural leakage.
The third challenge faced by energy harvesters is determining when enough energy has been stored so that it can be used to power a device. One of the simplest ways to do this is to use a capacitor of a known size and then wait for the voltage to reach a certain level (this will directly correspond to the energy stored) before feeding it into a power converter. However, the very act of measuring the voltage requires energy (assuming the voltage detection circuitry is based on active components), and this impedes the speed at which the system can charge.
Once the correct amount of energy has been stored, the final challenge is to release the energy efficiently to a device at a suitable voltage. Linear regulators are very good at providing noise-free supply voltages, but they are extremely wasteful; therefore, energy harvesters use switching regulators. However, these can introduce noise into a circuit—meaning that great care must be taken during the design stage.
In summary, energy sources are often miniscule in size, meaning that impedances between the energy source and the harvester must be correctly sized; current leakage means that component selection must be done carefully; the low voltages associated with energy harvesters means that active components with low forward voltage drops must be chosen; and power converters must be as efficient as possible.
What Options Do Some Devices Have?
The challenge of needing power to drive active components that monitor the current stored energy can be met with the use of small-coin cell batteries. While these have extremely limited capacity, it can be significant enough to run the monitor components for an extremely long period of time (potentially years).
The only purpose of such a battery is to maintain a bias voltage that allows the power monitor circuitry to operate, and, if MOS technology is used, then the entire design consumes an insignificant amount of current, making the coin cell a viable option. However, the use of a battery defeats the purpose of energy harvesting, and the fact that the battery will eventually run out does not bode well for remote apps.
With regards to energy storage, capacitors are ideal due to their low resistance, fast charge and discharging capabilities, and ease of integration. However, super capacitors allow for much greater storage capabilities at lower voltages, and their low-voltage operation also makes them ideal for use with low-voltage energy sources.
As most energy sources are extremely small, it can also be advantageous for engineers to combine multiple energy sources into a single device. For example, a small solar cell in combination with a thermocouple and piezoelectric generator can be used to gather energy from light, sound, and heat simultaneously. Such an arrangement provides more energy options to the device while increasing the charge speed. In fact, some energy-harvesting IC have dedicated inputs for connecting to multiple energy sources.
The MAX17710 is one example of an IC designed with energy-harvesting in mind. The IC integrates two input power options for different energy sources (one high-voltage source and one low-voltage source), a boost regulator, and an internal state machine for connection to a microcontroller. Additionally, the MAX17710 is also designed to work with THINERGY MEC101 batteries that are flat, solid-state batteries. The thinness of these batteries combined with their solid-state nature means that they are compact, inherently safe, and ideal for use in slim remote devices.
The STMicroelectronics SPV1040 is an example of a solar battery charger that is specifically designed to efficiently harvest energy with the use of MPPT (Maximum Power Point Tracker). In essence, an MPPT allows for setting the load point of a solar cell to be at its optimum for maximum power transfer, and the SPV1040 utilizes this to achieve up to 95% efficiency. Furthermore, the SPV1040 can operate down to 0.3V (i.e., very small voltages).
The e-peas AEM30940 is a dedicated energy-harvesting IC that can self-start from an input voltage of 380 mV with an input power of 3 µW. Extracting energy from stray RF, the IC integrates an MPPT for maximizing power transfer from the radio input and energy converter; can operate with all storage options including batteries, capacitors, and super capacitors; and can be connected to a primary battery for fail-safe operation. Additionally, the AEM30940 integrates multiple LDO regulators for producing various output voltages needed by many modern microcontrollers (e.g., 1.8V core and 3.3V I/O).
Conclusion
When trying to extract energy from the surrounding environment, energy harvesters face a multitude of challenges—the availability of energy, the need for active components to sense charge states, current leakage from components, and the high efficiencies needed by power converters. The fundamental limits of semiconductor materials prevent ultralow voltage operation, and the non-consistent nature of naturally occurring energy sources makes reliable operation virtually impossible.
However, great strides in energy harvesting have been made, and options do exist for engineers, such as the e-peas AEM30940 and MAX17710. But regardless of the technology being used, the best moves an engineer can make is to choose microcontrollers with extremely small energy requirements as well as those that have deep sleep modes. At the end of the day, there is no need to worry about storing large amounts of energy if a design requires little energy in the first place.