How a slice of silicon often thinner than a human hair can harvest sunlight to make electricity may seem like magic. But what may appear as a bit of sorcery actually boils down to uniting science and engineering wizardry with some of Earth’s most abundant resources—sunshine and silicon.
Photovoltaic (PV) cells are made of a special class of materials called semiconductors. Of all the semiconductor materials, silicon is most commonly used because of its availability (it’s the second-most abundant element in Earth’s crust) and its special chemical properties.
An atom of silicon has fourteen electrons arranged in three different levels, or shells. The first two shells, those closest to the center, are completely full. The outer shell, with four electrons, is only half full. A silicon atom will always look for ways to fill up its last shell (which would like to have eight electrons). To do this, it will share electrons with four of its neighboring silicon atoms. It’s like every atom holds hands with its neighbors, except that in this case, each atom has four hands joined to four neighbors. That’s what forms the crystalline structure, and that arrangement turns out to be important to the function of a PV cell.
Energy added to pure silicon can cause a few electrons to break free of their bonds and leave their atoms, leaving a “hole” (an unfilled bond) behind. These “free carrier” electrons wander randomly around the crystalline lattice structure, eventually falling into another hole. But there are so few free carriers available in pure silicon that they aren’t very useful. Scientists found they could improve silicon’s electron carrier ability (conductivity) by adding other atoms in a process know as “doping.”
Silicon doped with an atom of phosphorous here and there (maybe one for every million silicon atoms), will still bond with its silicon neighbor atoms. But phosphorous, which has five electrons in its outer shell, has one electron that doesn’t have anyone to hold hands with, so it takes a lot less energy to knock it loose. As a result, most of these electrons do break free, resulting in more free carriers. Phosphorous-doped silicon is called N-type (“n” for “negative”) because of the prevalence of free electrons.
But only one part of our solar cell can be N-type. The other part is typically doped with boron, which has three electrons in its outer shell. Instead of having free electrons, P-type (“p” for “positive”) has free holes.
The interesting part starts when you put N-type silicon next to P-type silicon—a silicon sandwich of sorts. When the electrons and holes mix at the junction between N-type and P-type silicon, silicon’s neutrality is disrupted and the free electrons mix to form a barrier, making it harder and harder for electrons on the N side to cross to the P side. Eventually, equilibrium is reached, and an electric field separates the two sides. The electric field allows (and even pushes) electrons to flow from the P side to the N side, but not the other way around. A P–N junction is commonly known as a diode—an electrical one-way valve for electricity. The special thing about PV cells is that they are diodes designed to absorb energy from sunlight.
When a photon—the electromagnetic energy of sunlight—with enough energy hits the N-layer, it knocks an electron free. These electrons stay in the N-layer. When a photon of light hits an atom in the P-layer, it knocks an electron free that can easily cross into the N-layer. The result is that extra electrons accumulate in the N-layer. A series of metal wires (traces) attached to the N-layer gives the electrons someplace to go, and they enter a DC circuit, flowing from the negative side of the cell and re-entering the cell through the positive side.
PV modules are made by connecting numerous cells in series, parallel, or series/parallel to achieve useful levels of voltage and current. These cell networks include positive and negative wiring terminals so we can channel the electricity generated to our uses. As long as sunlight is coming in, the electrons will keep flowing and can deliver electrical energy to a load that’s connected to the circuit.
One way to think about the process of electron movement is to imagine that the P-layer is a pool filled with electrons and your deck is the N-layer. If a sufficiently strong photon hits one of the electrons in the pool (P-layer), it can kick it up onto the deck (N-layer) where you can catch it and put it to useful work. Ideally, every photon coming into the pool would bump an electron up onto the deck that you could collect and put to use. However, silicon’s limitations, along with design challenges, prevent PV cells from being 100% efficient. In reality, most commercially available cells are between 4% and 22% efficient at converting the energy in the photons to useful electricity. Here are several reasons why:
Too Little or Too Much Energy. The light that hits a cell contains photons with a wide range of energies, but a PV cell will only respond to certain energies, or wavelengths. The required level of photon energy to activate an electron is referred to as the band gap. Different types of photovoltaic materials have different band gaps—higher and lower decks, so to speak. Some photons don’t have enough energy, and although they bump electrons, they don’t give them enough energy to get them up on the “deck.” This energy is wasted as heat. The lower the deck (lower band gap), the lower the minimum energy required.
So why can’t we choose a material with a really low band gap, so we can use more of the photons? Unfortunately, the band gap also determines the voltage of our solar cell. If it’s too low, what we make up in extra current (by absorbing more photons) we lose by having a small voltage (remember that power is voltage times current). If the incoming photon is too strong, it bumps the electron up higher than the deck, before it falls back down. In a PV cell, this energy expenditure is also wasted in the form of heat.
To capitalize on the higher energies of some photons, some exotic PV materials have two levels of decks. If a photon has enough energy, it can bump the electron all the way up to a higher deck where it can be collected. Some amorphous PV modules have two or three levels of decks, so if an electron isn’t excited enough to get on the highest deck, it might at least end up on a lower one and be used there.
These two effects alone—too little energy and too much energy in incoming photons—account for the loss of about 70% of the radiation energy incident on our cell.
Imperfect Junctions. A second source of inefficiency is that a lot of electrons just roll through slots between the deck boards before you can collect them. A perfect crystal doesn’t have any holes—every electron that is collected stays on the deck until it can be collected. However, polycrystalline solar cells have joints between crystals, resulting in an imperfection in the P–N junction—holes in the deck, so to speak, that allow electrons to slip back into the pool before they can be collected.
Even in a single-crystal solar cell, you still can’t collect all the electrons. The metal traces that collect electrons in a PV cell are spaced apart, and an electron that ends up too far from it may be lost before it can travel to the nearest trace and be collected.
Amorphous silicon has a similar problem called hydrogen diffusion. Instead of being a solid silicon crystal, it has all kinds of loose hydrogen atoms, which function like a deck full of gaps. Also, electrons in a position to be bumped by photons are fewer and farther between because the hydrogen leaves less silicon to hit. The hydrogen atoms are the reason that amorphous silicon decreases in efficiency over the first few months before stabilizing: Hydrogen in the atmosphere slowly diffuses into the module.
Reflection, Obstruction & Temperature. Silicon is very reflective, which makes harvesting sunlight challenging, since a cell can’t use photons that are reflected. For that reason, an antireflective coating (typically titanium dioxide or silicon nitride) is applied to the top of the cell to reduce reflection losses to less than 5%. This coating is what gives solar cells their blue appearance, instead of gray, as raw silicon would appear. The antireflective coating can be modified to get different colors, such as red, yellow, green, or gray, but these colors are less efficient than dark blue, so you very rarely see PV modules in these other colors. The glass on a module also has a special textured surface to minimize the reflection of sunlight.
Because silicon is a semiconductor, it’s not nearly as good as a metal for transporting electrical energy. Its internal resistance is fairly high, and high resistance means high losses. To minimize these losses, a cell is covered by a metallic contact grid that shortens the distance that electrons have to travel from one side of the cell to the other while covering only a small part of the cell surface. We could cover the bottom with a metal, allowing for good conduction, but if we completely cover the top too, photons can’t get through the opaque conductor and we lose all of our energy. If we put our contacts only at the sides of our cell, the electrons have to travel an extremely long distance (for an electron) to reach the contacts.
Various solutions to this obstruction have been considered, from BP Solar’s laser-grooved buried-grid modules that put the collection grid in trenches instead of using flat ribbons on the surface, to placing the metal contacts on the back surface of the cell (as on SunPower modules), to transparent conducting layers that are being used for some amorphous and organic PV materials.
Temperature also affects a cell’s efficiency. Typically, for each degree centigrade increase in operating temperature over its rated temperature, a PV cell loses about 0.5% of its specified power. For example, a PV module that experiences temperatures 50°C higher than its rated temperature (which is quite common for rooftop modules) may produce 25% less than its rated power. This happens because the thermal energy is distributed unevenly, with some electrons having enough energy to “go the wrong way”—back across the barrier, where they fall into holes we don’t want them to.
After all this talk about efficiency, you might be surprised to discover that buying the most efficient module on the market shouldn’t be your only goal. When you’re talking about energy production, it’s watts that we’re really after. If a less efficient PV module allows us to get those same watts for less cost, it may be a more cost-efficient choice than a more efficient, but more expensive, module.
If you have limited space on your roof or a small solar window, using more efficient modules can often make sense. But if you have acres of warehouse roof, for example, it may not. It all depends on your particular situation. To optimize your investment, prioritize cost per installed kilowatt-hour, longevity, and efficiency, in that order, if space is not a consideration.
Zeke Yewdall is chief engineer at Sunflower Solar, a PV design/install company in Boulder, Colorado.
Sam Ley is a physicist who works at Sunflower Solar, and has extensive experience in science education at museums.
Portions of this article were adapted from Scott Aldous’s article, “How Solar Cells Work,” courtesy ©2007 HowStuffWorks.com.