Semi-Conductors In the metal materials considered earlier, the coupling of the atoms together to form the material decouples an electron from each atom setting it free to roam around inside the material. These free electrons can easily be made to move by an applied electric field enabling electric currents to flow through the material. It is the large number of these electrons in metals that make them good conductors; without them electrons would flow much less easily through them. This can be seen in 2 other groups of materials: insulators: have very few free electrons in the material, because electrons are strongly bound to the atoms of the material. Pulling electrons out of the material or pushing them in, is extremely difficult. Electrons do not flow in larger numbers through the material in response to an electric field. The resistor of the material is effectively infinite, R. [The quantum explanation is rather more complex, but the result is the same: insulators are extremely poor conductors.] semi-conductors: are neither insulators nor good conductors. The outer electron of each atom in the semi-conductor is weakly bound to the atoms; some outer electrons are set free to move about the material, but most are bound to atoms. Thus, there are some free electrons in the material but only enough to make it weakly conducting. We shall not examine insulators further, but semi-conductors are of great importance because their conductivity can be improved to produce 2 different types of conducting material. These forms can be used together to produce diodes and transistors: electrical devices of particular, importance and interest. The conductivity of semi-conductors is enhanced by embedding atoms of another material into the semi-conductor. This is called `doping. There are 2 types of dopant: donor and acceptor materials. The best known semi-conductor is silicon, but there are others e.g. germanium, diamond, siliconcarbide. N-type Semi-Conductors are produced when donor atoms are embedded in a semi-conductor, outer electrons from donor atoms are set free to move through the material. This increases the conductivity of the material. Because the number of free electrons in the semi-conductor is quite low, it does not need much donor material to increase the conductivity significantly (see box). N-type materials are like metals in that there are lots of free electrons. Phosphorus is a donor material that is used with silicon. Let s assume that there are 10 28 atoms /m 3 in the semi-conductor material, and that 1 in 10 12 atoms release an electron into the material. Then there are 10 12 electrons /m 3. Adding 1 donor atom for every 10 6 atoms (0.0001%), gives 10 22 donor atoms/m 3 and 10 22 electrons/m 3, assuming each donor atom gives up an electron: the number of free electrons, and thus the conductivity, has increased by 10 10 times. In practice only a fraction of the donor atoms give up an electron, and the conductivity increases much more slowly. In silicon, adding 1 donor atom for every 10 5 silicon atoms increases conductivity by 10 3. P-type Semi-Conductors are produced when acceptor atoms are embedded in a semi-conductor. An acceptor atom easily captures an extra electron and binds it firmly, becoming negatively charged. When acceptors atoms are embedded in a semi-conductor, some of them attract and bind an extra electron. Once all the free electrons in the material have been bound, electrons are pulled off their surrounding semi-conductor atoms: this is quite easy as their outer electron is only weakly bound. This process leaves a large number of semi-conductor atoms with a positive charge (roughly equal to the number of acceptor atoms with a negative charge). It is quite easy for such a positively charged atom to pull the outer electron off a neighbouring neutral semi-conductor atom, swapping the charge. It is usual to label the positively charged atom a hole (a place where an electron can go), and to speak of the hole moving. In a P-type semi-conductor, this movement of an electron from an atom to a hole occurs frequently and the holes move around rapidly: in silicon, the hole mobility in p-type material is about 1/2 that of the electrons in n-type material. 12
If only sufficient acceptor atoms are embedded to remove the free electrons, the material becomes an insulator. Adding further acceptor atoms creates a hole for each embedded atom, and the holes created can be used to carry an electric current. Applying a voltage across a p-type semi-conductor leads to electrons being pulled out at one end of the material and being pushed in at the other; electrons move through the material via the holes. However, it is more convenient to think of the positively charged holes moving through the material from the higher voltage to the lower: holes are ejected into the material at the high voltage end (an electron is pulled off an atom creating a hole); at the low voltage end holes are pulled out of the material (electrons are injected into the material and coalesce with the arriving holes). A simple analogy for this flow of holes is a queue of people waiting on a row of chairs outside an office. When the person at the head of the queue goes into the office, an empty seat, a hole, is created at the head of the queue. When the next person in the queue moves chair into the vacant chair, the hole moves in the opposite direction. This continues as the remainder of the queue shifts along, until the hole reaches the end of the queue, when it is disappears as another person arrives and sits down. The hole is injected at the head of the queue, flows to the tail of the queue and is removed. The greater the hole density the higher the conductivity of the material, so that the quantity of embedded acceptor atoms controls the conductivity, and a small percentage of acceptor material causes a very much larger increase in conductivity. For silicon, Boron is the usual acceptor element. Semi-Conductor Diode The figure shows a semi-conductor diode. The material has abutting n-type and p-type material and a narrow region where the 2 materials merge: this is the p-n junction. This has unique electrical properties. Electrons can be made to flow easily from the n-type region into the p-type region, and for holes to flow in the other direction. It is very much more difficult to get the opposite flows. For low voltage drops across the junction, the diode has very low resistance, R 0, in one direction, but very high resistance, R, in the other direction. A full explanation of this is too complex for here. It can be considered that it is easy to move electrons from the n-type into the p-type region because there are lots of free electrons in the n-type region; it is much more difficult to remove an electron from the p-type region because there are no free electrons. A similar argument may be made for the holes. When the p-region connects to a higher voltage source and the n-region to lower voltage source, a current flows and the p-n junction is said to have a forward-bias or to be forward-biased. When the voltage connections are reversed (n-type more positive than the p-type region), no current flows and the junction is said to have a reverse-biasi or to be reverse-biased. The width of the transition region between n-type and p-type can be very narrow, much less than 1 micron (10-6 m). Diodes are useful devices in their own right, e.g. in turning alternating current into direct current. The p-n junction is fundamental to the operation of the transistor. Transistor The transistor like the diode is made from abutting regions of n-type and p-type semi-conductor material, but the transistor has one material type sandwiched between the other type so that there are 2 p-n junctions. There are 2 possibilities: the p-n-p transistor with n-type material sandwiched between p-type material; the n-p-n transistor with p-type material sandwiched between n-type. There are a 13
variety of transistors structures and modes of control/operation. We examine one type, the so-called MOSFET (Metal-Oxide-Silicon Field Effect Transistor), in both n-p-n and p-n-p forms and the operation of these as voltage-controlled switches. n-p-n transistor (MOSFET) The diagram shows a silicon n-p-n field effect transistor. It is built on the surface of a wafer of p-type material about 1 mm thick: this is called the substrate. The transistor is formed by producing 2 n-type regions in the surface layer. These n-regions with the p- region in-between form the n- p-n sandwich. Above the middle of the sandwich, a region of silicon-oxide, an insulator, is formed, and above this is placed a layer of conducting material, called the Gate. The remaining parts of the diagram are metal regions attached to the n-type regions as connecting wires, and a connection of the substrate to the low voltage terminal of the power supply. The Gate used to be made of metal, thus giving the phrase Metal-Oxide-Silicon from the arrangement of Gate, insulator and the substrate. Nowadays, the Gate is made from a form of silicon that is a good conductor. This is poly-crystalline silicon, which has very many different crystalline regions with many boundaries between crystals. At these boundaries many free electrons are released by silicon atoms to give a large enough free electron density in the material for a reasonable conductivity. [The substrate is itself cut from a single crystal of silicon, i.e. it has a uniform crystal structure with no crystal boundaries]. With modern manufacturing techniques, the n-type regions are of the order of 0.25 microns wide and deep, and a similar distance apart. These dimensions are continuously being reduced as technology improves, and dimensions of 0.15 microns and smaller can be produced by a few manufacturers. The explanation of the operation of this transistor as a simple switch requires its attachment to a power supply. For further examples we assume a 5V power supply since this is the traditional voltage used in digital circuits (many circuits now use 3.3V or a mixture of 3.3V and 5V supplies). To show the operation of the n-p-n transistor, the low voltage power supply terminal, 0 Volt, is directly connected to one of the n-type regions. Whichever region is chosen is labelled the Source: the name reflects the fact this region will be the source of electrons when a current flows through the transistor. The other n-type region will be connected indirectly to the positive, 5 Volt, terminal of the power supply though other devices that will act to reduce the current flow when the transistor is conducting, as otherwise a short-circuit condition will exist. This n-type region is labelled the Drain, as it is the terminal into which electrons flow when the transistor conducts. The voltage on this region will be controlled by the operation of the transistor and may be 0V or 5V, i.e. it is greater than or equal to the substrate voltage. The switching action between the Source and Drain regions is under the control of the Gate: this is the switch control input of the transistor with the voltage on the Gate, either 0 or 5 volts, deciding the state of the switch as either conducting or non-conducting. 14
Gate at 0 Volts - switch non-conducting: Assuming that there is no current between the n-regions then all the voltage from the power supply will appear across the p-region of the sandwich, and the Source will be at 0V and the drain at 5V: this is the result that we saw in Circuit 2 of page 9 of the Basic Electricity notes. Examining the Source-Substrate junction, both the source and the substrate are at 0 Volts, so that there is no voltage drop to drive a current between Source and Substrate. Turning to the Drain-Substrate junction, this is a p-n junction with the p-region, the Substrate at 0V, and the n-region, the Drain is at 5V, so that the junction is reverse-biased and no current can flow through the junction. Since no current can flow into or out of the Gate, because of the insulator, there are no current flows anywhere within the transistor, confirming the assumption with which we started. Gate at 5 Volts - switch conducting: What difference does changing the voltage on the Gate to 5V? This introduces an electric field across the insulator: with 5V on the Gate, 0V on the substrate, and an insulator thickness of 1 micron, the electric field strength is 5 million Volts/metre - enough to make your hair stand on end. This electric field reaches down into the substrate below and pulls electrons towards the surface of the substrate into the channel region below the insulator: the figures shows the channel growing. Electrons move into the channel region until there are sufficient excess electrons to screen the substrate below the channel from the electric field. The channel region is now rich in free electrons and is now no longer a p-type region, but is an n-type region. There is now no reverse-biased p-n junction to stop current flowing nor is there an n-p-n sandwich: just one n-region through which current can flow. In the figure right the Drain voltage has changed since with the current flow through the switch the voltage drop from the power supply will appear across some more resistive part of the circuit. An alternative way to look at this operation is that at the moment the electric field appears across the insulator, the voltage of the region below the insulator instantaneously becomes 5V; this voltage forward-biases the Channel-Source p-n junction and pulls electrons out of the Source into the channel; these electrons move across the channel to the Drain, filling the channel and reducing the voltage on the channel to 0V. While the Gate stays at 5V the channel remains in place. Switching Gate voltage from 5V to 0V: When the Gate is set back to 0V, the electric field disappears and there is nothing to hold the electrons in the n-region. They are rapidly absorbed into the Drain and Source regions. Another way of looking at this is that at the change of voltage of the gate, the voltage on the channel also immediately drops by 5 Volts to -5V (the channel has an excess of negative charge with respect to the surrounding 0V regions, i.e. is more negative) and the surrounding regions at 0V suck the electrons from the channel. The n-p-n transistor is commonly called an n-type transistor, because in its conducting phase it becomes all n-type material. Its standard logic symbol is shown right in 2 forms: one form shows the connection from the substrate to the 0V power supply connection, while the other leaves this out. The symbols highlight the isolation of the gate input from the drain and source terminals through the gap between the gate input and the drain-source line. 15
p-n-p MOSFET transistor The p-n-p transistor is very similar in operation to the n-p-n transistor, but there are important differences; the major one being that the Gate voltage for conduction is 0V opposite to that for the n-p-n. The figure shows the layout of a p-n-p transistor. It can be seen that the n-type and p-type regions are reversed from the n-p-n transistor: there is now an n-type substrate with p-type Source and Drain regions. The substrate is attached to the 5V power supply terminal. In order to illustrate the operation of the p-n-p transistor, different power supply connections are used: the Drain region is directly connected to the 5V power supply terminal, while the Source is connected indirectly to the 0V supply terminal via other circuit elements that restrict current flow when the transistor is conducting. The voltage of the source is controlled by the switching action of the transistor: it is 0V or 5V. Gate at 5 Volts - switch non-conducting: Assuming that there is no current between the p- regions then all the voltage from the power supply will appear across the n-region of the sandwich, and the Source will be at 0V and the drain at 5V. Examining the Drain-Substrate junction, both the substrate and the drain are at 5 Volts, so that there is no voltage drop to drive a current between them. Turning to the Source-Substrate junction, this is a p-n junction with the n-region, the Substrate at 5V, and the p-region, the Source is at 0V, so that the junction is reverse-biased and no current can flow through the junction. Since no current can flow into or out of the Gate, because of the insulator, there are no current flows anywhere within the transistor, confirming the initial assumption. Gate at 0 Volts - switch conducting: There is now an electric field across the insulator: with 0V on the Gate, 5V on the substrate. This electric field reaches down into the substrate below, pushing electrons away and creating holes in the channel region: the figures shows the channel growing. Holes move into the channel region until there are sufficient excess holes to screen the substrate below the channel from the electric field. The channel region is now rich in holes and is now no longer a n-type region, but is a p-type region. There is now no reverse-biased p-n junction to stop current flowing nor a p-n-p sandwich: just one p-region though which current can flow. In the figure to the right the source voltage has changed since with the current flow through the switch the voltage drop from the power supply will appear across some more resistive part of the circuit. An alternative way to look at this is that at the instant the electric field appears across the insulator, the channel region voltage drops to 0V; this voltage forward-biases the Drain-Channel p-n junction and pulls holes from the Drain into the channel; these holes move across the channel towards the Source. While the Gate stays at 0V the channel remains. The p-n-p transistor is commonly called a p-type transistor, because in its conducting phase it becomes all p-type material. Its standard logic symbols are shown to the right. 16