A world without semiconductor components has become unthinkable. But what are they and what makes them that special?
Semiconductor components have become extremely common in electric circuits. What's worth remembering is that their special properties are known for no longer than two centuries. It took until 1925 for Lilienfeld to patent the first concept of a transistor. In late 1947 the first actually working transistor was shown by Bardeen, Shockley and Brattain from Bell Laboratories. They were honored for this invention with the Nobel Prize in physics in 1956.
In the tutorial on thermistors we got in contact with a semiconductor component for the first time in this tutorial series, but there are plenty more to come. Diodes, transistors, LDRs, LEDs, photodiodes, all the integrated circuits (ICs) that make use of them and even solar cells are semiconductor components. Without these component the world as we know it would not exist.
Semiconductors have revolutionized the world and it is worth taking a more detailed look on them. The problem is, however, that their behavior is not easy to explain. This tutorial will be a purely theoretical one and it will drag us into areas of physics that are far away from common knowledge. Be warned, you might want to skip this tutorial and go ahead with just using semiconductor components without caring about how they actually work. Don't worry, if you do not understand parts of this article. It is not well-placed in a beginner series, but I think that some background knowledge is good, before we proceed with looking at the actual semiconductor parts.
Disclaimer: I'm not a physicist and semiconductors are an extremely complicated topic. I cannot guarantee for correctness of my explanations. For the ease of understanding some details might be left out or presented in an over simplified manner. If you think, that parts of this article need correction, contact me at email@example.com.
We already talked about the basics of conductivity in an earlier tutorial, but to understand semiconductors we need to dig a little deeper. Electric current is the movement of charges, in general the movement of electrons. Normally the electrons are attached to the nucleus and cannot move around freely. In some materials, however, it is possible for the outer electrons to disassociate from the nucleus and move through the material, if a voltage is applied. The material is conducts. We know that this is true for metals and graphite, but what about semiconductors?
Whether an electron can move around depends on how tight it is coupled to the nucleus. We need to take a closer look at the structure of atoms and especially the electrons. Under normal circumstances, however, it is not possible to observe the behavior of individual electrons. They are constantly in move and at the sizes of electrons quantum effects come into play as well. We cannot measure their position without disturbing their behavior. This is a general problem with all measurements, but in the realm of quantum physics we know that an exact measurement of their location and movement at the same time is impossible (see Heisenberg's uncertainty principle). At subatomic scale particle and wave properties cannot be easily distinguished anymore. Leaving all this aside, there are some general things that we know for certain:
All of this can be easily shown in the Bohr model. The physicist Niels Bohr postulated it together with Ernest Rutherford in 1913. Because of its simplicity, it is commonly taught in schools. The nucleus of the atom is surrounded by electrons orbiting it. As we know the attraction to the nucleus lowers with distance to the nucleus and thus the outer electrons are the ones that are more likely to be able to move around freely inside the material and make it conductive. By increasing the energy of an electron it is possible to push it into a more distant orbit.
The Bohr model has been superseded by today's nuclear shell model. The latter is a far more accurate model which is based on today's quantum physic understanding of electrons. It allows to calculate a three-dimensional probability distribution for the electron locations. The basic assumptions of Bohr and Rutherford still hold true, but the shell model additionally respects the interactions between electrons and the fact that in reality no simple and clearly limited cyclic orbits exist. The real location probability distribution is far more complex.
While this added knowledge greatly improved the understanding of chemical bindings, it is not strictly necessary to have all this detail to understand the guts of semiconductor behavior. When looking at the conductivity we normally use a simplified version of the band model. It differentiates between two bands: the conduction band and the valence band. When looking at the Bohr model the valance band corresponds to the outermost orbit where the electrons are still tied to the nucleus. The conduction band is the next higher orbit, where the electrons are less attached to the nucleus and can thus be disassociated from it.
The electrons excited into the conduction band can freely move and make the material conductive. Upon excitement, they additionally leave a vacant spot in the valence band. This spot can be filled by a different electron from the valence band. Now, a vacant spot exists at the place this electron originated from. This process carries on and it enables charge transport in the valence band. If you keep track of the hole and not the electrons, it looks like hole moves inside the valence band until it gets filled by an electron from the conduction band again. In analogy to electrons as charge carriers in the conduction band, the holes are also referred to as defect electrons: mobile positive charge carriers in the valence band.
With the knowledge from the shell model we know that there are no clearly distinguished orbits. There are regions where electrons are located with a high probability and regions where they are very unlikely located. The bands in the band model represent areas of high probability. In between them there is an area of extremely low probability. This zone is referred to as the forbidden zone or band gap. The electrons cannot simply move from the valence band into the conduction band. It is necessary to provide enough energy to the electron so that it can jump over the forbidden zone right into the conduction band.
This is the situation for most materials, but not for metals. In metals the conduction band and the valence band overlap. The electrons can easily move into the conduction band. Metals are good conductors. In isolators, on the other hand, the band gap is huge and it is highly unlikely that an electron gets excited into the conduction band, as this would require a lot of energy. The specialty of semiconductors is that they have a band gap that is small enough that some electrons can still get excited into the conduction band. The more electrons in the conduction band, the better the conductivity of the material. Let's have a look at how we can excite electrons and what relaxation and recombination mean.
There are essentially three ways to increase the amount of excited electrons inside the semiconductor material. The first one that may come to your mind is using an electricity. While it is indeed possible, it would require a really strong electric field. This method is not really relevant in normal semiconductor applications. Let's look at the other two methods. We already got in contact with the first one. We can increase the amount of excited electrons and thereby the conductivity of the semiconductor material by increasing the temperature. This is exactly what happens in an NTC thermistor. By heating it up we provide energy and increase the likelihood that electrons get into the conduction band. The resistance of the NTC thermistor decreases. The last method is using the photoelectric effect. It is possible to excite electrons using light or more specifically photons above a material specific wavelength. This effect enables photo and imaging sensor as we know them from our daily lives. A light dependent resistor is another, but much simpler sensor that makes use of this effect.
Now that we know how to excite electrons, let's talk about relaxation. If we excite an electron from the valence band into the conduction band, we end up with a free space in the valence band. Be assured, that this place will not be empty for a long time. The excited state is a suboptimal energetic state. An exited electron can jump back into the valence band. It fills the free hole. This is called recombination. So what is relaxation then? Relaxation means that the system returns to its original energetic state. Relaxation is the opposite of electron excitement. As energy is never destroyed or created, the electron has to release energy while it returns into the valence band. There are two ways for an electron to release energy. The first way is transferring energy to other particles. This causes grid vibrations or in simpler words additional heat. The second way is a lot more interesting: the energy can be released in form of a photon. As this photon now carries the energy that is released in cause of the electron transitioning from conduction to valence band, the emitted light has a distinct wavelength. If the band gap is smaller the emitted light is less energetic as well. Its wavelength is bigger (red light). If band gap is bigger the wavelength of the emitted light is smaller (blue light). This allows the production of light emitting diodes with different colors. We will look at LEDs in more detail in a later tutorial. Whether visible light is emitted during relaxation depends on the used semiconductor material and its exact band structure. Silicon for example is not suited for LEDs.
There are two types of semiconductors that need to be differentiated: intrinsic and extrinsic semiconductors. Let's have a look at both types.
Intrinsic semiconductors are pure semiconductors. They contain a very low number of foreign atoms. Like in the image below an intrinsic semiconductor out of silicon contains barely any other atom then silicon in its crystalline structure.
NTC thermistors and LDRs are both intrinsic semiconductors. Their conductivity is highly dependent on temperature and in case of the LDR the incoming light. Pure semiconductors are actually pretty bad conductors at room temperature. The conduction that can be observed in pure semiconductors is also called intrinsic conduction. When silicon is found in nature it is not pure, but it is also a lot more conductive than pure silicon. It has been found that the conductivity can be increased by introducing foreign atoms into the silicon. This brings us right into the area of extrinsic semiconductors.
Extrinsic semiconductors contain intentionally injected foreign atoms within their crystalline structure. The places were the crystalline structure is interrupted by foreign atoms are also called defects. The process of injecting the foreign atoms into the semiconductor material is called doping. Doping allows to alter the semiconductor behavior according to the needs of the individual application. The more foreign atoms one introduces into the material, the more conductive it becomes. The introduced foreign atoms enables additional energy states within the band gap. In consequence, the energy needed at once for a jump over the forbidden zone becomes smaller. In consequence electrons are more likely to get inside the conduction band even at low temperatures. The conduction at these defects in the semiconductor material is what is called extrinsic conduction. Semiconductors with a very high amount of foreign atoms possess a metal like behavior. The band gap is very small or even fully disappears leaving an overlapping valence and conduction band just like the one of metals. Such semiconductors are also called degenerated semiconductors as they do not behave like normal semiconductors anymore.
There are two types of doping that need to be differentiated: p-type and n-type doping.
For n-type doping an atom with an additional electron is added. Silicon has four valence electrons, thus a foreign atom with five valence electrons is used. A very typical choice is phosphorous. As shown in the picture below there is now an electron that is not involved in any chemical binding. The phosphor atom is also called an electron donator. The additional electron can be easily excited into the conduction band. If we look at the band model only a portion of the energy that would be needed to excite an electron in pure silicon is needed. An excitation of this electron is probable even at lower temperatures. In consequence, there are now extra electrons in the conduction band. This increases the conductivity of the semiconductor material. The conduction process primarily takes place in the conduction band. The electrons are the majority charge carriers in n-type doped semiconductors.
For p-type doping an atom with one less valence electron is added. A typical foreign atom used in this process is boron which has three valence electrons. As shown in the picture below there is now one electron missing for the fourth covalent binding. The foreign atom is an electron acceptor. The hole can be easily filled by an electron from the valence band. This leaves a vacant spot in the valence band, which allows charges to move in the valence band. P-type doping thus increases the conductivity of the semiconductor material just like n-type doping does. The conduction process primarily takes place in the valence band via the extra holes or defect electrons as they are sometimes called. They are the majority charge carriers in p-type doped semiconductors.
That was a lot of information about semiconductors. We did not even cover the basics of diodes and transistors at this point, however. I will end this tutorial with a cliff hanger: the magic happens if p-type doped and n-type doped semiconductor material comes together. I'll explain the details in a separate background tutorial on the p-n junction.