Dalla sabbia ai chip - La produzione di Silicio

Hi guys, I'm Marco for amici il Cuso and welcome to this second video recap of the posts published this summer on Facebook and Instagram. If you still don't follow us on these social networks, you can do it using the link in the description. Today we will gather two posts because we will see together how to purify and crystallize the silicon that is used, among other things, to produce photovoltaic panels and microprocessors of mobile computers and all the rest, starting from the raw materials. Silicone is the second most abundant element of the Earth's crust, after oxygen, and with the latter forms the most abundant mineral class of all, silicates. Silicone has numerous applications in the electronic field, microchip, metallurgical, it is used in the aluminum forge, and energy and photovoltaic panels. However, silica in nature is practically never found in pure form, only in 1823 Johannes Verzelius managed to purify it for the first time. Different applications then require different levels of purification, but in general it is always necessary to start from a common point, separate the silica from the oxygen, and first still choose the right raw material. We know, for example, that sand is mainly made up of crumbled quartz, which is a form in which silica can be presented. Unfortunately, however, the sand of our beaches is too contaminated by other elements to be used as a source for the purification of silica. The choice then falls on pure quartz crystals, extracted from the mine, in which the contaminations are much lower. Ok, we have chosen the source of silica. And now? The process of purification currently used consists of a reaction between silica and carbon, which occurs in an arc electrode. At temperatures ranging from 1500 to 2000 ° C, silica and carbon turn into silica and carbon dioxide. But then, the starting process to produce silica for photovoltaic cells that should lead to decarbonization, produces CO2? Yes, but obviously everything must be framed in the general balance. This process, being a first step, produces a fairly low quality silica. An average purity of 98% is estimated, which is very good for metallurgical applications. This is why it is called metallurgical grade silica, while more advanced applications such as electronics require higher purities. Metallurgical grade silica is the starting point for applications such as solar cells and electronics, but only about 20% of the silica is purified at least to this degree. The part of the lion, that is about 80% of all the silica produced on a global scale, is made by the so-called ferro-silica. As the name suggests, this is a very variable alloy of iron and silica. Consider that silica can go from 8% to 95% in weight, and commonly the alloy is called FeZXX, where instead of XX the approximate percentage of silica present is indicated. Ferro-silica is used in the production of steel and in general to reduce the metal bones to their elemental form, so you can understand that it is highly requested. The production process is quite similar to that of metallurgical silica, but in the furnace, in addition to carbon, obviously also iron is introduced, typically derived from ferrous rods. It is not a particularly pure alloy in terms of other components, but for the purposes required it is more than good. However, if we want electronic or solar grade silica, we must eliminate these foreign components. The presence of impurities, mainly made up of transition elements, creates levels in the gap of silica that should not be there, because they alter their electronic properties. Even if the silica should then be drugged for the particular application, the level and quality of the drug must be checked with care. The purification starts with a reaction with hydrochloric acid at 300 ° C, which produces trichlorosilane and molecular hydrogen. Many impurities react with the chlorine and form their own halogens, therefore being removed from the boxes. Trichlorosilane is a very volatile liquid, with a very low boiling point, not even 32 ° C, and therefore can be separated for distillation, as if we wanted to obtain a silica clump. To return to elementary silica, trichlorosilane is made to react with hydrogen at high temperature, above 1100 ° C, and for long periods of time, even longer than a dozen days, this to reverse the previous process. The products are in fact silica and hydrochloric acid. This time, however, silica has a much higher degree of purity, because most of the impurities are present in concentrations less than one part per billion, and the general purity is up to, well, this figure here. To promote the growth of silica during this process, the reaction is carried out in an empty room, where a small bar of polycrystalline silica is present. This bar acts as a seed or crystallization germ, and the silica that derives from the reaction grows literally on top of it, enlarging it. This process was developed by Siemens, and is in fact called the Siemens process. At this point, the beautiful pure silica can be used in solar cells, for which polycrystalline silica can be enough, or further worked to make it monocrystalline, which is instead essential for the production of microchips. Here, now we have to understand the difference between polycrystalline silica and monocrystalline. A crystal is, to simplify a bit, a solid in which the molecules, or as in the case of silica, the atoms, are arranged in an order that extends in the medium-long range. This means that we can identify a base structure, an elementary cell, made up of a certain number of atoms, which then repeats itself many times in space. In the case of silica, the atomic bonds have a pyramidal structure, with an atom in the center and another four at the top of a pyramid. It is the same structure with which the carbon atoms are arranged in the diamond. The elementary cell that repeats itself has a slightly more complicated shape than that, because behind this fact there is the need to have a structure that actually repeats itself by translation, long well-defined directions, but this is of little interest to us now. What interests us is that any solid crystal is, in general, given by the combination of many different crystals joined together, and for this reason it is called a polycrystal. In practice, when the substance solidifies during cooling by the state fused, it begins to form here and there, by chance, tiny crystalline germs. These germs, if the conditions are appropriate, and they are not always, so much so that some solids are in fact completely amorphous, they grow in size. Each of them has the characteristic regularity of a crystal, but it is quite obvious that the orientation of the base cell for each will be completely random. Germs will grow into grains, and at a certain point the grains will bind together forming the final solid. But they will all be oriented in a random direction, and therefore the solid will in fact be an aggregate of many grains. This is a very common condition in solids that tend to crystallize easily, like metals. A solid polycrystalline is considered a crystal from certain points of view, for example it has a well-determined melting temperature, since each grain melts at the same temperature. Under certain others, the presence of grain edges, that is, the surface that separates two adjacent grains, has macroscopic effects, for example on the properties of electrical and thermal transport. A polycrystalline is of isotropic norm, that is, it behaves in the same way regardless of the direction in which it is observed, or tested. This should not surprise us, given the chance with which the grains are oriented, taken in any direction, the solid will always present itself in the same way. Another property of polycrystalline is that it has no precise cutting directions, but we will come back to this in a moment, because this is a further problem for the electronics industry. Although it is not at all simple, it is still possible to make a solid of polycrystalline silicon transform into a monocrystalline, or it can be made directly crystallized into a monocrystalline. By monocrystalline we mean a solid of arbitrary dimensions, even if limited by the available technology, made up of a single grain. Taken from an elementary cell, I can move it without ever rotating it from a theoretically arbitrary distance and always remain in the regular structure of the solid. It is a difficult operation, which can be obtained in various ways. A classic is the Czochralski method, I hope I pronounced it correctly, and its variations. A small monocrystalline silicon, appropriately oriented, is immersed in the molten silicon at a very precise temperature, 1425 degrees. The silicon, which is very few degrees above its melting point, uses the small crystal as a germ and begins to grow around it. However, unlike what happens in the normal crystallization process, this time the grain is one and can theoretically grow to size. Other precautions can be taken during the process, for example, it is possible to dissolve drug-containing substances in the fusion to obtain a drug-containing liqueur directly, or localized heating allows a partial fusion and slow recrystallization of the liqueur to eliminate defects in the crystalline reticle. Because, as much as the system is growing like a single grain, this is not exempt from crystalline defects that thermodynamics, subduedly, always tries to end. In fact, there are technical difficulties in obtaining giant monocrystals, although currently we are able to obtain cylinder caps up to 300 mm in diameter and we are trying to rise in size up to 450 mm. But why give our souls to build such a complicated object? Because when a crystal is a monocrystal, it has very different properties from the polycrystal. Remember when I told you that a polycrystal is all isotropic? Well, a monocrystal, no. Its electrical, thermal, mechanical properties also vary significantly from direction to direction. In addition, a monocrystal can be cut with incredible precision, literally atomic, along some precise directions. The diamond cutters know something about it, natural monocrystals. And this thing has been confirmed to me by those who know the gems well. Cristian from the channel Gemmologia per Tutti. It's worth taking a walk. And the cut obviously allows us to have very precise slices of thickness. Looking around, you might come across indications like Wafer 1.0.0. This indication, which reads 1.0.0 and not 100, is a Miller index, a way to indicate the direction inside the monocrystal, which is a fascinating but complicated topic. In short, working the silicon is a long, expensive and complex process. But the modern industry, especially that of chips, requires increasingly controlled silicon, with the possibility of obtaining structures of nanometric dimensions. The cell phone or computer you are using to watch this video requires a technology that would have been unthinkable at the dawn of computer science and also requires a physics that is no longer that of Ohm and classic resistances. Well guys, this post-summer video ends here. If you liked it, thumbs up and share it with your friends. Follow us on Facebook and Instagram and I also remind you that on Docity you can find my introductory course on quantum mechanics. I thank you for watching. I'll see you in the next video and be sure to be curious. Physics is where you don't expect it.