Storage Alternatives for Energy Harvesting Applications

Alternative energy sources deliver small amounts of power intermittently, and at times power levels may not match the needs of the applications that depend on them. Some form of energy storage is needed, though the solution will vary with the demands of the application. This article will examine various energy buffering solutions, including small form-factor batteries, thin-film batteries, and supercapacitors, highlighting both their specifications and the applications to which they are best suited.

The supply side

Just as semiconductor buffers smooth the flow of data between a processor and memory, some form of buffering is needed between energy harvesting sources and their associated applications. The amount of buffering needed increases in proportion to the mismatch between supply and demand.

On the supply side there is a great deal of variability in the amount of power available from different energy harvesting sources (see Table 1). With the exception of the ambient RF, typical energy harvesting devices can supply anywhere from 10 µW to 1 mW, though rarely on a steady basis. Solar power is obviously only available when the sun is out; vibration/motion sources only work when the motor is turned on or the person being monitored is moving; and thermal sources only work when there is a reliable source of heat and a heat sink to generate the temperature gradient required by thermal electric generators (TEG's).

Table 1: Power available from energy harvesting sources.

Another limiting factor is the inefficiency of most energy scavenging devices. Looking again at Table 1, there is often a few orders of magnitude difference between the theoretically available source power and the actual power you can harvest using currently available devices. This problem can be addressed by adding more solar cells or TEGs in parallel, for example, if your application will allow it. Still, you are dealing with micropower sources and there is a limit to what you can do with them.

In many cases, you can increase the amount of energy harvested by utilizing a variety of sources. For example, a remote wireless strain sensor on a building might combine a small solar cell with a TEG; on a bridge roadbed, a similar device might combine a TEG with a vibration sensor. If your application can combine light, vibration, and thermal energy sources, it may be able to operate indefinitely by just harvesting these energy sources.

The demand side

The obvious response to limited supply is to limit demand. If over time, the energy requirement of your device is equal to or less than what your energy harvesting sources can supply, then the device should be able to operate independently for an indefinite period of time. If supply falls short of demand, then you have to consider whether the time your device can operate before needing attention is adequate under the circumstances; if not, then energy harvesting sources may not be part of the solution.

Wireless remote sensor nodes rely heavily on energy harvesting. To minimize power consumption, they typically power down the microcontroller as far and as frequently as possible, only waking up to check for beacons and to quickly burst out data for a few microseconds every second. In addition, they utilize small, low-power software stacks such as ZigBee® and Bluetooth® low energy. Some companies have gone so far as to create their own smaller, proprietary RF protocols, such as Texas Instruments’ SimpliciTI™ and Microchip’s MiWi™.

Wireless sensor nodes highlight the need for energy buffering, since the current demand during transmit is often 10 to 100 times that required while the MCU is in a sleep state. In these applications, capacitors or even supercapacitors are often placed in the supply line to be able to handle the sudden surge in current. Thin-film batteries could be permanently damaged without this sort of protection.

High-current batteries

If your application has a particularly high but infrequent current requirement – for example, a small motor that cycles periodically – a miniature lead-acid battery may fill the bill. The EnerSys Cyclon 0810-0004 2 V thin-plate, sealed-lead battery, delivers 2.5 Ah at 260 mA. Just 1.35 inches in diameter by 2.41 inches high, the absorbed glass mat (AGM) Cyclon battery can fit into the smallest embedded applications.

Lead-acid batteries have their own unique charging requirements. Overcharging or repeatedly fully discharging these batteries can markedly shorten their life expectancy. The cycle life of a lead-acid cell is a function of depth of discharge (DOD), temperature, and charging rate. According to EnerSys, the life of a Cyclon cell can vary from 300 to more than 2,000 charge/discharge cycles. If a proper state of charge is maintained, Cyclon cells can last up to ten years.

Since the output voltage of a lead-acid battery varies with the state of charge (see Figure 1), both charge control circuitry and voltage regulation between the battery and the application are necessary.

Figure 1: Cyclon battery medium rate discharge voltage profile (Courtesy of EnerSys).

Lead-acid batteries are inexpensive and well-suited to outdoor applications that may be subject to extreme temperature.

Coin cells

Small lithium-ion coin cells are not rechargeable, but they still have a place in low-power devices that harvest ambient energy sources. A low-power sensor application could be powered by energy harvesting sources when they are available – with their energy stored in a capacitor or thin-film battery – switching over to the coin cell when those sources are not available. Such an arrangement could greatly extend the life of the coin cell battery.

In contrast to lead-acid batteries, lithium-ion coin cells have an extremely flat discharge curve (see Figure 2) thanks to a very low self-discharge rate. The ubiquitous CR2032, for example, is rated at 3.0 V at 225 mAh with a recommended continuous standard load of 0.2 mA. Figure 2 shows the cell voltage staying relatively flat while delivering 190 μA into a 15 kΩ load for over 1,000 hours, after which it quickly fails.

Figure 2: Discharge characteristics for CR2032 coin cell (Courtesy of Panasonic).

Manganese-lithium batteries – unlike their Li-Ion counterparts – are rechargeable, which makes them a natural fit for energy harvesting applications. The Panasonic ML-1220/F1AN is a rechargeable SMD coin cell rated at 3.0 V at 17 mAh while drawing 30 μA. While its discharge characteristics (see Figure 3) are similar to those of the CR2032, the ML-1220 can endure 1,000 charge/discharge cycles between ten percent state of charge and nominal capacity. The device requires a fixed-voltage charge of 2.8 to 3.2V at 1.2 mA or below. With a voltage regulator between the energy scavenging sources and the battery, that should be a fairly easy target for such micropower sources to hit.

Figure 3: Discharge characteristics of the Panasonic ML-1220 manganese-lithium battery (Courtesy of Panasonic).

Coin cell batteries, whether rechargeable or not, are a natural fit for many, if not most energy harvesting-based applications. Since even the CR2032 can last up to ten years if your application is sufficiently low powered, the ability to recharge may not be an issue. For example, the coin cell that is maintaining the BIOS configuration data in the computer on which you are reading this article should be good for up to ten years. The chance of the average person not replacing their computer within ten years is close to zero.

When an application requires more power than a standard Li-Ion coin cell can deliver for sufficient amount of time, designers should consider rechargeable manganese lithium coin cells such as the ML-1220 in conjunction with energy harvesting sources of power.

Thin-film batteries

For ultra-low-power devices that need to operate indefinitely, thin-film batteries are a natural storage choice to use with energy harvesting sources. Cymbet’s solid-state, surface-mount, thin-film EnerChip™ batteries are increasingly paired with small PV cells in energy harvesting applications. The Cymbet EnerChip CC CBC3112-D7C-TR1 is rated at 3.3 volts with a capacity of 12 μAh at 50 μA. These are obviously quite small batteries, so they are targeted at ultra-low-power applications such as wireless sensors, RFID tags, and standby supply for NV-SRAM in real-time clocks. The CBC3112 comes in a 20-pin, 7 mm x 7 mm square dual flat no-lead (DFN) package that is only 0.9 mm thick.

EnerChip CC devices contain a built-in power manager, temperature compensated charge control, and built-in energy storage protection. During normal operation, the chip charges itself with a controlled voltage using an internal charge pump that operates from 2.5 to 5.5 volts. When the primary power supply dips below a user-defined threshold voltage, the EnerChip CC signals this event and routes the EnerChip voltage to VOUT in order to maintain continuous power to the MCU or other circuitry. A single EnerChip CC can charge up to ten additional EnerChips connected in parallel.

The EnerChip CC’s discharge characteristics (see Figure 4) closely resemble those for Li-Ion coin cells, with the output voltage remaining largely flat before eventually falling off a cliff. However, CBC 3112 can withstand over 5000 charge/discharge cycles, recharging from 10 percent to 80 percent state of charge in ten minutes.

Figure 4: EnerChip discharge characteristics (Courtesy of Cymbet).

 For those wishing to look further into this technology, Cymbet offers both the Cymbet Eval Board for CBC3112 and the Cymbet Enerchip CC Eval Kit.

Infinite Power Solutions (IPS) offers a line of rechargeable THINERGY™ micro-energy cells (MECs), solid-state batteries targeting energy harvesting applications. The THINERGY MEC225-1S provides a nominal 4.1 volts with a capacity of 130 µAh at 65 µA. At just 12.7 mm x 12.7 mm x 0.2 mm, the devices definitely deserve to be called thin.

THINERGY MECs utilize a lithium cobalt oxide (LiCoO2) cathode and a Li-metal anode with a solid-state electrolyte called LiPON (Lithium Phosphorus Oxynitride), which gives them a discharge curve similar to other lithium-based batteries (see Figure 5), though with some important differences.

Figure 5: Typical discharge curves at 250C for THINERGY MEC225 (Courtesy of Infinite Power Solutions).

THINERGY MECs have a number of advantages in energy harvesting applications. They have a rated cycle life of 100,000 charge discharge cycles from a ten percent depth of discharge with a typical application load; they can be recharged to 90 percent state of charge in 15 to 20 minutes. The smallest device in the THINERGY MEC family of products, the MEC-225, can accommodate a discharge rate of up to 7 mA and an ultra-low self-discharge rate of one percent per year at 25°C. Of particular interest is their ability to accept charge currents along 1 µA, well within the range of all but the smallest micro-power source.

IPS makes the Infinite Power Solutions Energy Harvesting Evaluation Board that includes a four volt 0.7 mAh THINERGY MEC, a small solar cell, and associated circuitry to enable you to experiment with a solar powered energy harvesting application.

Supercapacitors

To be able to handle sudden current surges – such as when the quiescent wireless sensor node suddenly transmits a burst of data – thin-film batteries employ a large capacitor or often a supercapacitor across their output.

Electric double-layer capacitors (EDLCs) – also known as supercapacitors – have an energy density hundreds of times greater than electrolytics. Their high density is a direct result of the extremely close proximity of the conductive layers, which also results in a very low breakdown voltage. Supercapacitors have a lower energy density than batteries, but a far higher power density, since, unlike batteries, they can be discharged almost instantaneously.

The Taiyo Yuden PAS414HR-VG1 is a 3.3 volt, 60 mF surface-mount coin cell EDLC that measures only 4.8 mm diameter by 1.4 mm high. The PAS414HR-VG1 has a capacity of 20 µAh measuring from 3.3 volts to 2.0 volts and approximately 40 µAh when discharging down to 1.0 volt.

Figure 6 shows the charge/discharge characteristics for the PAS414HR-VG1 under different load conditions. The curves highlight some of the advantages as well as disadvantages of super caps in energy harvesting applications. On the plus side, they have very fast charge and discharge rates, the latter being a key advantage in applications that occasionally require high peak power delivery. On the downside, their output voltage drops quickly as they discharge, generally requiring a buck boost voltage regulator to assure constant output from the supply to which they are attached.

Figure 6: Charge/discharge characteristics for the PAS414HR-VG1 supercapacitors (Courtesy of Taiyo Yuden).

Supercapacitors tend to suffer from internal leakage, resulting in a much faster self discharge rate than batteries or capacitors. As a result, they are usually used in conjunction with batteries to buffer the load for the power source and provide peak power when needed. They augment batteries in energy harvesting solutions, they do not replace them.

Which storage medium is best?

While acknowledging the advantages of coin cell batteries (rechargeable and not) and supercapacitors in various applications, Infinite Power Solutions stresses the advantages of thin-film batteries in energy harvesting applications. Table 2 gives a quick overview of key parameters for each of the primary technologies.

Table 2: Storage choices for energy harvesting applications (Courtesy of Infinite Energy Systems).

Traditional rechargeable batteries are a good choice when AC power is available; lead-acid batteries are the least expensive and most durable, though they have a higher self discharge rate, lower energy density, and shorter cycle life than lithium-based rechargeable cells.

Supercapacitors excel at handling energy surges, but compared to rechargeable coin cell and thin-film batteries, they have a number of disadvantages, including:

• High internal leakage currents resulting in relatively rapid cell discharge;
• Voltage decreases linearly during discharge; decreases linearly during discharge;
• Degradation and short lifespan at high temperatures;
• Large size and poor energy density;
• Charge balancing required when connected in series for higher voltage;
• More frequent replacement due to degradation and lifespan issues.

Thin-film batteries have numerous advantages in small battery-powered devices, including high energy density, small size, minimal charging requirements, voltage stability during discharge, and low self-discharge rate. However, they also have a number of disadvantages, including high cost, limited storage capacity, limited surge capability, and some fragility regarding the minimum and maximum voltage levels they can tolerate – though the latter is generally handled by an accompanying power management IC (PMIC).

By and large for low-power applications that rely on energy harvesting – particularly in those cases where size is a more important consideration than cost – thin-film batteries will be the storage medium of choice.

Summary

There are numerous storage devices available for energy harvesting applications, including miniature lead-acid batteries, lithium-ion coin cells, supercapacitors, and thin-film batteries. Each has advantages and disadvantages, and the choice will vary with the application.