MOSFETs

The action of a field effect transistor depends upon the effect of the field produced by a charged region known as the gate. The term MOSFET is the acronym for metal oxide silicon field effect transistor. The diagram opposite shows why the transistor has this description. The transistor is based on a bar of p-type silicon. Two crosswise strips in the bar are doped to make them into n-type material, and metal is deposited on these to form the source and drain terminals. Between the strips, the surface of the silicon is covered with a thin layer of silicon oxide, which is a non-conductor. Metal is deposited on the silicon oxide to form the gate electrode.

From the account above we can see that the transistor consists of metal, oxide (of silicon) and silicon. It works by the effect of an electric field.

The transistor is connected into a circuit through its three terminals, source, gate and drain. Some types also have an external connection to the substrate, but in most types this is connected internally to the source as shown below. A potential is applied between the source (0 V) and the drain (positive) but there is a p-n junction between them and the p-type material, so no current can flow. It is like having two diodes connected back-to-back. The p-type material is connected to the source so it is at 0 V.

When the gate is made positive of the p-type material, it repels the holes from nearby regions of the p-type material, turning it temporarily into n-type material.

Another way of looking at this is to say that the gate attracts electrons to the region around it. These are more than enough to fill the holes in that region and the remainder are available as charge carriers. This creates a channel joining the two n-type strips. Current can then flow from the source to the drain. The greater the gate potential, the wider the channel and the larger the current.

The full name of the transistor described above is an n-channel enhancement MOSFET. The term n-channel describes the way that charge flows through the transistor, carried by electrons. The term enhancement refers to the fact that there is no channel when the gate is at 0 V, but the channel is enhanced (made wider) as the gate potential increases. This kind of MOSFET is the most widely used but there are other kinds, including the p-channel enhancement MOSFET, which is similar to the n-channel type, except that the polarities are reversed.

Power MOSFETs

MOSFETs built along the lines described, but with heavier construction, may be used for currents of a few amps. Usually such types have a metal tag to which a heat sink (p. 70) may be bolted. Several special types of MOSFET are used for higher currents. One of these types are the v-channel MOSFETs, generally referred to as VMOS. The ‘v’ in v-channel describes the shape of the channel, not its charge-carriers. In VMOS transistors, the doped layers are built up on a surface that has previously been etched to form steep-sided grooves, v-shaped in section. The action of the transistor is normal. The difference is that conduction is down the sides of the grooves. The distance from top to bottom of the grooves is short, and the channel is very wide in the direction along the grooves, so its resistance is low. This means that currents can be large without overheating the transistor. VMOS power transistors are made to carry currents of 10 A or more. They make it possible to control large currents by devices that can produce only a small current. For example, a heavy-duty relay or a 10 Ω loudspeaker can be operated directly from the output of a CMOS logic gate.

Another type of structure is the HEXFET. Like VMOS, it relies on the channel being short but wide to give very low ‘ON’ resistance and thus allow high currents without overheating.

JFETs

The full name of these is junction field effect transistors. This name comes from the fact that its action depends on what happens at a p-n junction (page 74). The structure of a JFET is shown overleaf. A bar of n-type semiconductor has metal contacts at each end. On either side of the bar is a layer of p-type material, the two layers being connected together by a fine wire. Remember that this structure measures only a millimetre or so across.

If a p.d. is applied to the ends of the bar, a current flows along the bar. Because the bar is made from n-type material, the current is carried by electrons. We say that this is an n-channel JFET.

When referring to the potentials in this diagram we take the potential of the source to be zero. Electrons enter the bar at the source terminal (the reason for its name), and leave at the positive end of the bar, the drain terminal. The layers of p-type material are known as the gate. If these are at 0 V or a slightly more positive potential, they have no effect on the flow of electrons. But, if the gate is made more negative than the source (perhaps by connecting a cell with its positive terminal to the source and its negative terminal to the gate), an important effect follows. The p-type gate and the n-type bar form a p-n junction. With the gate negative of the bar, the junction is reverse biased and a depletion region (p. 74) is formed. The channel through which they flow is made narrower. In effect, the resistance of the bar is increased.

The narrowing of the channel reduces the flow of electrons through the bar. There are two ways of using this field effect produced by the p-n junction:

• Switching: If the p.d. between the gate and source is very high, the depletion region extends across the whole width of the bar. This cuts off the flow of electrons altogether. Used in this way the JFET is a voltage-controlled switch.

• Current control: With smaller gate potentials that do not completely cut off the current, the width of the bar left free for the passage of electrons is proportional to the gate potential. As this is increased, so the width of the channel decreases, the resistance increases and the current decreases. In effect the JFET is a voltage-controlled resistor. Putting it another way, a JFET can be used to convert a change of potential to a change of current.

A JFET is always operated with the p-n junction reverse-biased, so current never flows from the gate into the bar. The current flowing into or out of the gate is needed only to change the potential of the gate. Since the gate is extremely small in volume, only a minute current (a few picoamps) is required. A small change in potential of the gate, controls the much larger current flowing through the channel. This property of the transistor can be used in the design of amplifiers.

JFETs have many applications, particularly in the amplification of potentials produced by devices such as microphones which are capable of producing only very small currents. They are also useful in potential-measuring circuits such as are found in digital test-meters (p. 148), since they draw virtually no current and therefore do not affect the potentials that they are measuring (p. 61).

A JFET similar to that described above is manufactured from a bar of p-type material with n-type gate layers. Current is conducted along the bar by holes and this is known as a p-channel JFET. In operation, the gate is made positive of the source to reverse-bias the p-n junction.

Bipolar junction transistors

Owing to practical difficulties in making FETs in the early days of semiconductors, bipolar transistors were the first to be widely used. They still are extremely popular with designers, even though the problems of making FETs have been overcome. They work on an entirely different principle to FETs.

A bipolar transistor is a three-layer device consisting either of a layer of p-type sandwiched between two n-type layers or a layer of n-type between two p-type layers. These are referred to as npn and pnp transistors respectively. The fact that conduction occurs through all three layers, which means that it involves both electrons (negative) and holes (positive) as charge carriers, is why these are sometimes called bipolar transistors. Their full name is bipolar junction transistors (or BJT) because their action depends on the properties of a pn junction (p. 74), as is explained later. The diagram shows the ‘sandwich’ structure of an npn transistor, the most commonly used type.

The three layers of the transistor are known as the collector, the base and the emitter. In effect, the transistor consists of two pn junctions (in other words, diodes, connected back-to-back). It would seem that it is impossible for current to flow from the collector to the emitter or from the emitter to the collector.

Whatever the direction of the p.d., one or other of the p-n junctions is sure to be reverse-biased. This is where the features of the base layer are important:

• It is very thin (though not shown like that in the drawing).

• It is lightly doped, so it provides very few holes.

When the transistor is connected as in the diagram, the base-emitter junction is forward-biased. Provided that the base-emitter p.d. is greater than about 0.6 V, a base current flows from base to emitter. When describing the action in this way, we are describing it in terms of conventional current, as indicated by the arrows in the drawing. What actually happens is that electrons enter the transistor by emitter terminal and flow to the base-emitter junction. Then, as in a diode (p. 77), they combine with holes that have entered the transistor at the base terminal. As there are few holes in the base region, there are few holes for the electrons to fill. Typically, there is only one hole for every 100 electrons arriving at the base-emitter junction. The remaining 99 electrons, having been accelerated toward the junction by the field between the emitter and base, are able to pass straight through the thin base layer. They also pass through the depletion layer at the base-collector junction, which is reverse-biased. Now they come under the influence of the collector-emitter p.d. The electrons flow on toward the collector terminal, attracted by the much stronger field between emitter and collector. They flow from the collector terminal, forming the collector current, and on toward the battery. In effect, the base-emitter p.d. starts the electrons off on their journey but, once they get to the base-emitter junction, most of them come under the influence of the emitter-collector p.d.

The important result of the action is that the collector current is about 100 times greater than the base current. We say that there is a current gain of 100.

If the base-emitter p.d. is less than 0.6 V, the base-emitter junction is reverse-biased too. The depletion region prevents electrons from reaching the junction. The action described above does not take place and there is no collector current. In this sense the transistor acts as a switch whereby a large (collector) current can be turned on or off by a much smaller (base) current. If the base-emitter p.d. is a little greater that 0.6 V, and a varying current is supplied to the base, a varying number of electrons arrive at the base-emitter junction. The size of the collector current varies in proportion to the variations in the base current. In this sense the transistor acts as a current amplifier. The size of a large current is controlled by the variations in the size of a much smaller current.

The structure and operation of pnp transistors is similar to that of npn transistors, but with polarities reversed. Holes flow through the emitter layer, to be filled at the base-emitter junction by electrons entering through the base.

Mass production

Very large numbers of transistors are produced at once on a single silicon wafer. A frequently used method is the planar process.

First a layer of silicon oxide is formed on the upper surface of the wafer by heating it in an atmosphere of oxygen and water vapour. The next step is similar to that used for making pcbs (p. 157), except that it is on a very much smaller scale. The silicon oxide layer is coated with a layer of photoresist. Then the wafer is exposed to uv light, usually through a mask on which the areas to be etched are left clear. The pattern of the mask is repeated over and over again to produce many transistors, which can be later separated by cutting the wafer. Another approach is to have a single mask and to step this along to produce an array of exposed areas. For finely detailed circuits, a reduced image of the mask is projected onto the photoresist layer using a lens system.

The resist is etched chemically, removing the exposed areas of the resist and the oxide beneath. Finally, the wafer is doped, as above, but with a different dopant. This time the dopant can not reach the areas covered with oxide, so only the exposed areas are doped. It is possible to control the depth to which the dopant penetrates.

When an npn transistor, such as that on page 89 is made, we begin with a wafer doped to produce n-type silicon. Areas of this are etched and doped to produce the p-type silicon of the base layer. Although the silicon in the base begins as n-type, it is converted to p-type by diffusing an excess of a hole-producing dopant, to cancel out the effects of the electron-producing dopant already present. The wafer is then exposed to a mask which causes smaller areas to be etched over the base layer, and then the wafer is doped to produce a smaller shallower n-type layer (emitter) in the base layer. Finally the collector (substrate), base and emitter contacts are added by placing the wafer in a vacuum, covering it with a mask and evaporating metal on to it. Alternatively a continuous layer of metal may be deposited and etched away to leave the required connections.

Practical transistors

A transistor is a minute object on a small chip of silicon (rarely germanium). To make it practicable to handle the transistor, it is mounted in a case or sealed into a block of plastic, with thin wires connecting the base, emitter and collector to thicker terminal wires. The title photograph of this chapter shows typical packages used for low-power transistors. These are two of the standard packages used for JFET, MOSFET and BJT transistors.

By varying the amount of doping, the method of doping, and the geometry of the regions, transistors with various characteristics can be made. Some have much higher gain than others (up to 800 times), or may be suitable for operation with high currents. In typical general-purpose transistors, the maximum collector current is only a few hundred milliamps.

Power transistors are capable of collector currents of up to 90 A. A power transistor is robustly constructed with low-resistance channels to minimize the voltage drop across it. It also has a sturdy metal tag or case for bolting to a heat sink.

Radio-frequency transistors are designed specially for operation at frequencies of several hundred megahertz, and some up to 5 GHz. At high frequencies, the capacitance between the base and emitter of a BJT may act to reduce the amplitude of the signal, so radio-frequency transistors are designed to minimize this effect. A transistor designed to operate at radio frequencies may be used in circuits other than radio transmitters and receivers. Many other kinds of device such as computers, mobile telephones, digital cameras and CD players operate at radio frequencies and high-frequency transistors are required for these too. Such devices are digital rather than analogue and the main function of the transistors is high speed switching. As explained in the next chapter, the gate or base of a transistor is biased ready for action by connecting it through a resistor to the positive supply lines. During manufacture, there is no problem (and almost no extra cost) in putting the biasing resistors on the same chip as the transistor. This simplifies the layout of the circuit board and saves the cost of separate resistors. Digital transistors with resistors included are often used for switching in digital circuits.

Most transistors are available also as surface mount transistors. The typical package, measuring only 3.0 mm × 1.5 mm is shown at top centre in the photograph on page 166.

Darlington transistors

A Darlington pair consists of two npn transistors connected as shown in the diagram. The way they work is easier to understand if it is explained in terms of conventional current. Given a small base current flowing to Q1, a much larger collector current flows into the transistor and out of the emitter. This becomes the base current of Q2, which amplifies the current still further. For small base currents, the gain of the Darlington pair equals the gain of Q1 multiplied by the gain of Q2. A typical Darlington pair has a gain of 10 000, and some have gains up to 50 000.

A Darlington pair may be assembled from two individual transistors but it is more convenient to use a Darlington transistor. This consists of two npn transistors made as one unit with internal connections and enclosed in one of the standard transistor packages. Paired FETs are available, often under the name FETlington transistors.

Thermionic valves

Although this is a chapter on transistors, which are solid state devices (that is, based on semiconduction), we should mention one of their vacuum tube equivalents. Such devices, once the only active devices available, have been almost entirely replaced by semiconductors. They still have applications in high-power circuits, such as radio transmitters, and there are audio enthusiasts who claim that the reproduction of music by a valve amplifier gives a tone far more pleasing than that from a solid-state amplifier. There are also the vintage radio experts who take a pride in collecting and restoring long defunct radio equipment and getting it to work again. Valves are still being manufactured.

The simplest thermionic valve is the diode which we have already mentioned (p. 77). This is the functional equivalent of the p-n semiconductor diode. Like all thermionic devices, it operates with a high cathode-anode voltage, usually over 100 V. The next most complex valve is the triode, which, as its name implies comprises three electrodes.

The symbol for a triode shows its basic structure but, in practice, the electrodes are differently arranged. In the centre of the tube is the heater filament, wound on an insulating former. Close around this is the cathode, a cylinder of thin metal. This may be connected directly to the heater, as in the diode diagram on page 77, or the cathode may have a separate terminal, as on the left. The next electrode is a cylinder around the cathode and concentric with it. This is made from wire mesh and is called the grid. On the outside, concentric with the other two cylinders is the anode, made from thin metal. The whole structure is enclosed in an evacuated glass tube.

In action, a current is passed through the heater, which glows red hot and heats the cathode. The heated metal atoms liberate free electrons which form a cloud around the cathode. The cathode is at 0 V, and the anode at a high voltage, 100 V or more. This causes the electrons to be attracted toward the anode. Current flows through the valve. So far, the action is the same as we have described for a diode. Electrons can pass through the meshes of the grid but, if the grid is made negative of the cathode, some of them are repelled and travel back toward the cathode. The more negative the grid, the more electrons are repelled. In other words, the flow of electrons, and hence the current through the tube, is controlled by the field around the grid and, hence, by its potential. In this way the triode is the equivalent of a field effect transistor. The fact that small changes in grid potential can cause large changes in anode-cathode current is the basis of many kinds of amplifying circuit.

Many other types of thermionic valve are made, some with four, five or more electrodes to give superior performance.