In this video, we will discuss photodetector. So the most common form of photodetector based on semiconductor is photodiode. And photodiode is essentially a p-n junction on the reverse bias. So as shown in this diagram on the right, there is a p-n junction here near the surface and the light comes in. And the light is absorbed by the semiconductor, produces electron hole pair. If that electron and hole pair production occurs within the depletion region of the p-n junction, then there is a built-in electric field. And therefore the holes and electrons are separated by the built-in electric field. And they can be collected by the electrodes that attached on the p-n inside and register a current. And that current will be proportional to the number of photons that is absorbed. So this is how you detect light using semiconductor p-n junction. It's a very compact solid-state detector. And here is the photograph of a packaged p-n diode sold by a company and it's very cheap and very handy. The quantum efficiency tends to be very high. Meaning that you almost always get one electron hole pair per photon coming in and the speed tends to be pretty high. But there are factors that determine the speed of detection and that is the carrier dynamics, but also the junction capacitance. So the p-n diode and p-i-n diodes are typically used for photodiode. And in a p-i-n diode, p-type and intrinsic and n-type diode, is essentially the same as p-n diode. Except that you have an intrinsic region in between and that artificially increases the depletion region. So in a p-i-n diode, tour depletion region width is essentially determined by the intrinsic region. Which generally are designed to be greater than the typical p-n junction. So in any case, the operating mechanism is quite similar and in a p-n diode under four bias you inject carriers. And these injected carriers then diffuse away from the junction. And that's essentially the for bias current. So in a PN Junction under for bias, the current that you observe is essentially diffusion current. On the reverse bias however, there will be very, very little diffusion current because the energy barriers are increased beyond that already establishing at equilibrium condition. So diffusion is further suppressed and so therefore the current under reverse-bias condition is really dominated by generation and recombination current. Typically generating current because the p-n product, and the hole and electron concentration is reduced below the equilibrium value under reverse-bias condition. And therefore the semiconductor responds by generating carriers. So the generation current is the dominant current mechanism under reverse-bias condition. So, when the semiconductor is exposed to light, then the photo generation, it can be a dominant mechanism of the current. And this is the mechanism that is used in a photodiode. The reverse bias also helps in the photodetector in another way. That is, that the depletion region is increased under reverse bias. And the depletion region is really the region where you have a built-in electric field. And the photogenerated carriers are separated by that electric field and your registered current. So the photodiode is typically operating under reverse-bias condition. To look at the characteristics of these photodiodes a little in more detail. One of the key characteristics of characterizing the performance of any photodetector is the quantum efficiency and bandwidth speed. Now, the quantum efficiency is essentially, can you collect all electrons and holes generated by photons? So, does the incoming photon produce electron hole pair always? And also, once the electron hole pairs are generated, can you collect all of those carriers without losing them into defects. And other recombination mechanism that might be present in your semiconductor? So, the modern semiconductor technology can produce very, very high-quality. Especially silicon-based devices are very, very pure high-quality material. So the quantum efficiency can be very high. The bandwidth depends on three things, carrier diffusion towards the depletion region edge. If the carriers are generated outside of depletion region, then these carriers must enter the depletion region where there is a built-in electric field. And then, the carriers are separated, electrons and holes are separated and then you can measure a current. So the carrier diffusion into the depletion region edge, that's one component. And then the other is once the carriers are in the depletion region, whether they are generated outside the depletion region and diffuse into the depletion region. Or they are generated within the depletion region to begin with. In either case, they have to go across, traverse the depletion region and reach the neutral region and then get collected by the electrode. So that drift time within the depletion region is another component. And then of course, like on any other electrical circuit, there is a finite time response due to the capacitance effect. So RC time constant is another factor that determines the speed of your photodetector. So as I said, if the carriers are generated outside the depletion region, they need to first diffuse to reach the depletion region. And once they reach the depletion region, and then they are swept away. They are drifted by the built-in electric field. The drift time, so the diffusion here the diffusion process is governed primarily by the diffusion coefficient of the semiconductor. Now, once the carriers are in the depletion region, then they need to be drifted and go travel across the depletion region. And reach the electrode in the end and there are two trade offs that we need to consider. So first, you want to have a large absorbing volume so that you can increase light absorption. You don't want to lose both towns because the absorbing layer is too thin, than the total absorption. There is a high transmission of light, so the absorption is limited. That's going to decrease your quantum efficiency. So you want to have a large absorbing region and in that sense depletion region width should be high. However, if you're depletion region was high then it is going to take that much longer for carriers to travel across the depletion region and so it slows down. So you can see there's a trade-off between quantum efficiency and the time response. And of course as I said, the p-n junction on the reverse bias is a capacitive device. So there is a finite junction capacitance and the junction capacitance gives rise to an RC time constant. And that is going to be another major factor that determines the time response or the bandwidth of your photodetector. Now, if you increase the reverse bias quite high, then you can have an avalanche breakdown. An avalanche breakdown occurs through this impact on the ionization process that's schematically shown in the figure here. So the electron here gets accelerated by the built-in electric field and when the electric field is very, very high. This electron will attain very high energy. And so let me use a pen here, so this electron is going to get accelerated very high. So the kinetic energy of the electron becomes very high. And this electron has enough energy to collide with one of these electrons that was originally here in the valence band and knock that one off. Creating one hole and one additional electron and this electron is going to lose the energy and reach the bottom of the conduction band. So this process is called the impact ionization and you can see that it is a carrier multiplication process. So if your photodetector operates in this regime, then you can imagine that one photon that came in produces one electron hole pair. And this whole is going to go out this way. This electron is injected in to the depletion region. Normally, if you don't have this impact ionization process, this electron will just go through the depletion region. And get collected by the electrode on the inside and register current. So that's the normal operation of p-n-junction-based photodiode. You get one electron in one hole per one incoming photon. But if you have impact ionization process, then one electron injected will produce many electrons and many holes. And so this mechanism gives you gain and it becomes possible to effectively detect very, very low level light signal. And this type of photodiode is called the Avalanche photo diode. And if you recall the impact ionization process is terrorized by this coefficient, impact ionization coefficient. And when both electron and holes Alpha sub N and Alpha sub P are the impact ionization coefficient for electron and hole when they are the same. Meaning that the electrons and holes are equally likely to impact ionize and produce current. Then, you have a breakdown. In other words, your gain M here is the ratio of the final density of electron to the initial density of electron. And this ratio can diverge and what that means is that you have an indefinite increase in current which is not good. This is a breakdown and you don't really have a control over the current. So you can't really have a normal operation of a device in this condition. So, breakdown is something to be avoided in the photodetection. So if your impact ionization coefficient for holes is zero, meaning that only electrons impact ionized. Holes don't. In that case, the expression for the gain coefficient here M has no singularity. So you have always a finite gain which can be very high. And there is no condition that leads to indefinite uncontrolled increasing current. So this is the ideal mode of operation and this will be the condition that you would like to produce for avalanche photodiode. But in natural semiconductors, electrons and holes always impact ionize both of them. So how do you produce this condition where only one carrier species impact ionizes? You can use a superlattice structure as shown in the figure in the right. Superlattice structure is composed of repeating layers of wide bandgap material and narrow bandgap material. And so when an electron enters the superlattice structure. Then every time the electron crosses the interface from the wide bandgap material to the narrow bandgap material the electron gets a boost in its kinetic energy due to the conduction band discontinuity. And therefore, even when the overall electric field is small, you can still achieve a very high kinetic energy. And initiate impact-ionization events. Now, holes experience the same thing. So hole gets a boost in its kinetic energy every time it crosses the interface from wide bandgap material to the narrow bandgap material. However, you can always choose the materials in such a way that the conduction band offset is large and valence band offset is small. This way, you can create a structure in which only electrons impact ionize and holes don't. And this is a structure that is commonly used in avalanche photodiode.
