Let’s take a moment and imagine that you are designing a fantastic IoT system. You have done a great deal of work developing the concept, you know that the system works from your bench testing and you are moving to the next stage where the device is unplugged from the bench. Initially you choose a battery, but you know that the cost and effort that goes into maintaining the battery will limit the device’s use. You start to explore what is possible with energy harvesting and choose to look at solar since light is something reliably present for your application. So where do you start?
With solar cells there is not one perfect solution for all applications. As with most things in the world there are many tradeoffs when choosing one technology over another. This article will discuss some of these tradeoffs.
First, let’s look at what the efficiency of a solar cell is. Efficiency of a solar material is defined as the power output per unit area (Watts/square meter) produced compared to the optical power per unit area available from the sun. This efficiency number is stated for one specific operating environment: AM1.5 solar light spectrum, 1000 Watts/square meter light intensity, and 25C temperature. AM1.5 solar spectrum is the spectrum of light measured when light travels through 1.5 times the earth’s air mass at the equator (AM1.0). This set of criteria is unrealistic for most actual environments and as we will see, this standard efficiency is meaningless with indoor lighting.
Figure 1: (A) AM1.5 solar spectrum. (B) Spectrum of common indoor light sources.
With high efficiency indoor lighting the available light spectrum is much narrower as shown in Figure 1. Since high efficiency indoor lighting (LED, Fluorescent) are limited primarily to the visual spectrum, a solar cell that only absorbs light in the visual spectrum can be low efficiency outdoors and high efficiency indoors. This also means that a multijunction solar cell with a very high outdoor efficiency may have a very low indoor efficiency since one of the junctions in the solar cells is built to only absorb infrared light, which is not available in the indoor lighting.
If the junction designed to absorb the infrared light is not producing current, it will limit the amount current produced by the solar cells since Kirchhoff’s current law states that the amount of current at any point in a branch must be the same. The second factor that affects solar cell efficiency indoors is parasitic leakage. Each technology and manufacturing technique will yield a different amount of parasitic loss. Outdoors the light intensity is high enough that the parasitic losses are not significant compared to power production, but indoors the light intensity is low enough that the parasitic losses can become dominant and limit the ability of a solar cell to output power. Figure 2 shows how a technology like amorphous silicon (a-Si) can be low efficiency outdoors but generate the most power indoors with LED or fluorescent light sources.
Figure 2: Indoor performance of common solar technologies. Rated efficiency is the outdoor efficiency of the technology. Current density is the actual measured intensity using each light source.
Since the highest power producing technology will depend on your lighting scenario, the first thing you will want to determine before choosing a solar technology is whether your product will be used primarily indoors or outdoors. When comparing indoor specs from different companies it is also important to look at the type of lighting the indoor measurements were made with. There is no industry standard for indoor solar measurements and the same cell measured in warm white LED lighting will produce more current at the same lux level as the same cell when measured in fluorescent or cool white LED lighting.
The next question is how will solar be integrated into your product. Solar material is not compatible with a standard PCB reflow process, so integration techniques need to be considered. Panels are commercially available with a number of different connection systems. Some panels can be connected using a traditional soldering iron process by soldering tabs into a PCB through-hole/slot or onto a surface mount pad. Panels can be purchased with wire leads pre-attached. These leads can have connectors crimped onto them, can be soldered to a PCB, or can be clamped into a screw type connector. Another option is to connect the contacts on the back of a cell to a PCB using a conductive epoxy. Finally, flexible printed circuit (FPC) connectors can be added to a PCB using a standard reflow process and solar panels with tabs that fit into these connectors can be quickly and easily installed as show in Figure 3.
In addition to how to electrically connect the solar to a system, you will also need to determine if there are other physical requirements like flexibility that are required by or that will enhance your product. If you use a flexible PCB do you need a panel that can conform to your PCB or does you application require the panel to conform around a curve? Solar technologies such as amorphous silicon are able to flex and conform to a variety of surfaces as shown in Figure 4.
Figure 4: Flexible amorphous silicon panel bonded to a flexible PCB with integrated BLE radio.
Next consider the environmental requirements of your application. Will the panels be exposed to sweat, moisture, or UV light? This will guide you in choosing the encapsulation requirements for the panel. Some cells are designed to be in wet environments and some are not. Some encapsulations are UV resistant, while others will yellow and become brittle after long-term UV exposure. Will your panel be installed outdoors where self cleaning is important? Fluorinated polymer encapsulants like ETFE are known for their excellent self cleaning ability. Will your application see intermittent shading? Performance in partial shading is dependent on the panel construction and solar technology. Some panels will output proportional to the amount of panel shaded, while others will output very little even if only a few percent of the panel is shaded. If you have a very low power application and shading is a factor, you may want to consider using a panel with only one or two cells and an energy harvesting IC to harvest the power and store it in a rechargeable battery.
Finally, once you have your requirements, purchase samples and test out your use case in the real world to verify the assumptions made during the design process are valid. Depending on what you find you may opt to change the amount of solar or the battery capacity of the system to modify your factor of safety. It is important to note that increasing the battery capacity to a very large battery with a small solar panel can actually be detrimental to the system operation. Battery self discharge is proportional to battery capacity and if a battery becomes too large the self discharge could be more than your energy harvesting system can generate.
Now go ahead and jump into the world of energy harvesting and solar. I hope this article has given you a starting point.
References
1 https://www2.pvlighthouse.com.au/resources/courses/altermatt/The%20Solar%20Spectrum/The%20global%20standard%20spectrum%20(AM1-5g).aspx
Daniel Stieler is president of PowerFilm, Inc., based in Ames, Iowa. He will appear at Sensors Converge Sept. 21 in San Jose. The event includes many sessions in person and digitally from Sept. 21-23. Register here to attend.