字幕表 動画を再生する 英語字幕をプリント This episode of Real Engineering is brought to you by Brilliant, the problem solving website that teaches you to think like an engineer. Installed global capacity of solar cells has increased year on year for the past decade, fueled by the plummeting prices and rising efficiency of solar cells. [1] Forcing fossil fuel producers out of the market through technological advance. At the end of 2019, the total installed capacity of photovoltaic cells exceeded 630 thousand Megawatts, an astounding figure that is going to continue to rise in the coming decades. However, in the 40 years we have been using solar cells, there has been a mysterious flaw that has been sapping away potential electric current from the photovoltaic cells. Upon testing in the laboratory, newly manufactured solar cells display an efficiency of about 20%. Meaning they could convert 20% of the incoming energy from sunlight into electric current. However, within hours of operation, that efficiency would drop to 18%. [2] A 10% drop in total electric generation. Losing 10% of 630,000 Megawatts of power is no small problem. That's equivalent to about 30 nuclear power plants worth of power capacity, if the solar panels could operate all day, which they can't, but you get the point. There's a lot of potential electricity being lost. It's no wonder that scientists and engineers have been hunting down the cause of this problem, termed light induced degradation, for 40 years. And last year, we may have finally cracked the problem and found the cause behind this mysterious loss in power. To understand it we first have to understand how photovoltaic cells work. Photovoltaic cells use the photovoltaic effect to generate a current. An effect where photons of a particular threshold frequency striking a material can cause electrons to gain enough energy to free them from their atomic orbits and move freely in the material. [3] This is best achieved with semiconductors, whose unique properties lying between conductors and insulators allows them to most easily elevate electrons from atomic orbit to moving freely among their atoms. Some of the first solar cells were created using selenium, like this one, created by Charles Fritts, sitting atop a New York in 1884 [4] A revolutionary device that produced a consistent current of electricity, but it was achieving an efficiency of just 1%. Converting 1% of the energy striking it in the form of light, into electricity. This, in combination with the high cost of selenium, made it a unviable source of electricity. To succeed these devices needed to compete with fossil fuel power sources. Before the photovoltaic effect could power the world, scientists and engineers would need to figure out how to increase that efficiency percentage and do it with cheaper materials. Enter Silicon. A common semiconductor material that has formed the bedrock of the electronic age. This is going to be our starting material for our solar cell. Let's build a solar cell from scratch and see how efficiencies were gradually increased over time. Let's first look at what happens when light interacts with a pure silicon crystal like this. Incoming light can do one of three things. It can be reflected, absorbed, or simply pass right through it. If the light is reflected or passes through, it cannot produce the photovoltaic effect. Step one to improving our efficiency is to minimize the amount of light that gets reflected off the material. This is wasted energy that affects our efficiency level. In fact, 30% of light that strikes untreated silicon is reflected. So before we even start, our maximum efficiency drops to 70%. [5] For this reason, Silicon is often treated with a layer of silicon monoxide which can reduce the light reflected to just 10%, while a second layer with a secondary material, like Titanium Dioxide can reduce it as low as 3% [6] Texturing the surface of the material can further increase the probability of the light being absorbed. If it is textured like this, light that is initially reflected has another chance to strike the material and be absorbed. Only light that is absorbed can potentially cause the photovoltaic effect, but not all light will. We need photons above a threshold energy to increase an electron's energy enough to allow it to move freely in the material. A photon's energy is defined by multiplying planck's constant by its frequency. Silicon requires photons with 1.1 electron volts to produce the photovoltaic effect, which corresponds to a wavelength of 1,110 nanometres [7] This lies around here in the light spectrum and any lower energy light from here down cannot cause the photovoltaic effect. This light will simply cause the atom to vibrate and create heat. This graph shows the total solar energy being emitted by the sun, however a good deal of this does not reach the Earth's surface as it is absorbed in the atmosphere. This is a more realistic graph. About 4% of the energy reach earth's surface is in ultraviolet, as the sun emits relatively little ultraviolet photons. 44% is in the visible spectrum, and 52% is in the infrared spectrum. This may sound surprising, as infrared light is lower energy, but it covers a wider range of the spectrum and thus accounts for more energy. Because silicon cannot make use of light with a wavelength greater than 1,110 nanometres, everything from here up is energy we cannot convert to electricity. This represents about 19% of the total energy reaching earth. Another thing to note is that light with higher energy does not release more electrons, it simply produces higher energy electrons. For example, blue light has roughly twice the energy of red light, but the electrons that blue light release simply lose their extra energy in the form of heat. Producing no extra electricity. This energy loss results in about 33% of sunlights energy being lost.[8] So these spectrum losses alone cause a 52% loss in efficiency. This is a lot of energy to lose, but silicon sits near the ideal threshold frequency that balances these two energy losses. Capturing enough of the lower energy wavelengths, while not losing too much efficiency as a result of the material heating up. [9] The reason's solar panels lose efficiency as they get hotter is quite complicated and outside the scope of this video, but for now all you need to know is that silicon balances these factors best for terrestrial purposes. This is such a large loss in power that in some climates active cooling, which takes some of the electricity the panels create to cool the panels, actually results in more electricity being generated. Onto the next problem. Knocking an electron free by itself does not create an electrical current in our circuit. It just frees an electron to float freely about the material. To create a useful current we have to force this electron around an external circuit where it can do work. Freeing an electron also creates a positively charged “hole” in its place, that is also free to move about the material. If an electron meets a hole, it simply fills it and our energy is wasted. The next trick to maximise efficiency is to limit the chances electrons have to fill these holes and to force them into our circuit as quickly as possible. To do this we use the unique properties of silicon. Silicon has 4 electrons in its outer shell, and thus readily forms a crystal structure with 4 neighbouring atoms using covalent bonds, a bond where neighbouring atoms share an electron pair. We can manipulate this behaviour and tailor the crystals material properties by adding impurities, called dopants. Say we add boron atoms to the silicon crystal wafer. These boron atoms have 3 electrons available for bonding with the silicon crystal, but silicon wants 4. So this creates a “hole” in the crystal that wants to be filled with an electron. We call this a p-type as it has positive charge carriers Now, let's say we create another wafer of silicon, but this time we add atoms with 5 electrons available for bonding, like phosphorus. Again the phosphorus bonds with the silicon, but this time we have an excess electron that can float freely about the material. We call this an n-type, because it has negative charge carriers. Now, let's sandwich these two materials together and see what happens. The positive holes and negative electrons migrate towards each other. The electrons will jump into the p-type and the holes jump into the n-type. This causes an imbalance of charge, because now the p-type side has more negative charges, and the n-type has more positive charges. We have just formed an electromagnetic valve that allows electrons to pass in one direction. Let's see how this works. Suppose a photon with sufficient energy enters the p-type side of the solar cell and knocks an electron free. The electron starts bouncing around the material and one of two things can happen. It can recombine with a hole, resulting in no current, or it can come into the electromagnetic field at the junction of the material. Here the electromagnetic field actually accelerates the electron across the junction into the n-type side [10], where there are very few holes for it to fill, and to boot, the junction's electromagnetic field actually prevents the electron from passing back to the other side. A similar thing happens on the n-type side, where holes are selectively transported across the junction before they can recombine. This means one side of the junction becomes negatively charged, while the other side becomes positively charged, we have created a potential difference or in other words a voltage. If we add some metal contacts and an external load circuit, these electrons will pass along the circuit to recombine with the holes on the other side. We have just created a solar cell. You may notice a problem here though, by adding metal contacts to the upper surface of the solar cell, we have just blocked light from entering the cell, and thus reduced it's efficiency. This is yet another problem engineers have had to think carefully about in their quest to optimize solar cell efficiency. Over the years engineers have optimized both the shape and manufacturing techniques to minimize the area covered by the metal electrodes, while also minimizing the resistance the electrons will face in entering the external circuit. One research paper used topology optimization to design these electric contacts. [11] Topology optimization uses algorithms to optimize the design of objects using constraints the engineer inputs. Using this method for the electric contacts produced something remarkably like the vasculature of a leaf, and that shouldn't really surprise you. Footage: Vasculature tissue on a leaf does not perform photosynthesis. It instead brings the water that is essential for photosynthesis to the leaf and extracts the useful products,, serving a similar purpose as our electric contacts, so of course plants have developed the perfect shape to optimize the energy they can absorb from the sun. Plants have had millions of years to evolve this shape[10] However, most solar cells use a simple grid shape, as it is cheap to manufacture. This typically results in an efficiency loss of about 8%. All told, these effects result in typical modern silicon solar cells having a laboratory tested efficiency of 20%. So, what was happening to cause that drop to 18% after a couple of hours of operation? This problem was the focus of hundreds of scientific papers and many found clues to the problem. [12] Many noted that the efficiency drop was correlated to the concentration of boron and oxygen in the silicon and noted that the drop did not occur when boron was substituted for gallium. Thus, it was known a boron oxygen defect was causing the issue. Others found that the defects could be reversed by heating the silicon in the dark at 200 degrees for 30 minutes, but it would return once again upon exposure to the light. Efforts in reducing the problem have primarily focused on reducing the concentration of oxygen impurities in the silicon wafers, which occur as a result of the czochralski silicon wafer manufacturing technique that is the source of the 95% of silicon solar cells. These manufacturing techniques are still a point of research [14] and the engineers and scientists were working blindly. Little was known about the actual defect creation process and how exactly it was causing such a large drop in efficiencies, leaving engineers with less information to solve the problem with. This paper used a special imaging technique and observed these boron oxygen molecules converting into something the paper refers to as “shallow acceptors” when exposed to light. [13] In essence, they observed the defects transforming into little electron traps that acted as recombination sites, and thus reduced the time and probability of electrons entering the circuit to do work. With this knowledge, engineers can now develop better techniques for preventing this phenomenon and hopefully help increase our renewable energy capacity in the coming years. It's easy to think that technology has reached a point of being so advanced that knowing where things can be improved is practically impossible for the average person, but that simply isn't true. A little bit of research into any area will reveal countless problems humans are still grappling with fixing. When I started researching this video, I knew little about solar panels beyond the basics. In order to make this video, I took a week to deep dive into some college textbooks using my knowledge of material science and electronics to guide my research, but I had some gaps in understanding that the college textbooks just assumed I had preexisting knowledge of. Terms like “band gap” and “fermi levels” kept appearing, and without understanding these terms, I couldn't make complete sense of the explanations. These were like canyons in my journey for knowledge. I couldn't advance until I filled them in. So, half way through the research and writing process, I decided to stop what I was doing. I changed my tactics. I decided to take the Brilliant course on solar energy, because I knew Brilliant would guide me through the very basics of the subject right through to the more complicated concepts. It worked a treat. All the little gaps in my understanding were filled in and I could now read scientific papers and college textbooks without feeling like I was reading a foreign language. This is the magic of Brilliant. Brilliant's courses are curriated incredibly well and allow you to go from knowing nothing to being an expert. This is just one of many courses on Brilliant. Brilliant's thought-provoking math, science and computer science courses help guide you to mastery by taking complex concepts and breaking them up into bite-sized understandable chunks. You'll start by having fun with their interactive explorations and over time you'll be amazed at what you can accomplish. If you are naturally curious, want to build your problem-solving skills, want to develop confidence in your analytical abilities, or like me, find yourself struggling to parse the language of advanced college textbooks because you are missing some fundamental knowledge, then get Brilliant Premium to learn something new every day. As always, thanks for watching and thank you to all my patreon supporters. If you would like to see more from me, the links to my twitter, instagram, discord server and subreddit are below.