Semiconducting devices are at the heart of the microelectronic revolution that ushered in the information age. Table 27-2 compares the properties of silicon-a typical semiconductor-and copper-a typical metallic conductor. We see that silicon has many fewer charge carriers, a much higher resistivity, and a temperature coefficient of resistivity that is both large and negative. Thus, although the resistivity of copper increases with temperature, that of pure silicon decreases. pure silicon has such a high resistivity that it is effectively an insulator and thus not of much direct use in microelectronic circuits. However, its resistivity can be greatly reduced in a controlled way by adding minute amounts of specific “impurity” atoms in a process called doping. Table 27-1 gives typical values of resistivity for
silicon before and after doping with two different impurities. We can roughly explain the differences in resistivity (and thus in conductivity) between semiconductors, insulators, and metallic conductors in terms of the energies
of their electrons. (We nee quantum physics to explain in more detail.) In a metallic, conductor such as copper wire, most of the electrons are firmly locked into place within the molecules; much energy would be required to free them so they could move and participate in an electric current. However, there are also some electrons that, roughly speaking, are only loosely held in place and that require only little energy to become free. Thermal energy can supply that energy, as can an electric field applied across the conductor. The field would not only free these loosely held electrons but would also propel them along the wire; thus, the field would drive a current through the conductor. In an insulator, significantly greater energy is required to free electrons so they can move through the material. Thermal energy cannot supply enough energy, and neither can any reasonable electric field applied to the insulator. Thus, no electrons are available to move through the insulator, and hence no current occurs even with
an applied electric field. A semiconductor is like an insulator except that the energy required to free some electrons is not quite so great. More important, doping can supply electrons or positive charge carriers that are very loosely held within the material and thus are ease to get moving. Moreover, by controlling the doping of a semiconductor, we can .
control the density of charge carriers that can participate in a current, and thereby can control some of its electrical properties. Most semiconducting devices, such as transistors and junction diodes, are fabricated by the selective doping of different regions of the silicon with impurity atoms of different kinds.Let us now look again at Eq. 27-20 for the resistivity of a conductor:where n is the number of charge carriers per unit volume and T is the mean time between collisions of the charge carriers. (We derived this equation for conductors, but it also applies to semiconductors.) Let us consider how the variables and T change as the temperature is increased. In a conductor, 11 is large but very nearly constant with any change in temperature. The increase of resistivity with temperature for metals is due to an increase in the collision rate of the charge carriers, which shows up in Eq. 27-24as a decrease in T, the mean time between collisions. In a semiconductor. /I is small but increases very rapidly with temperature as the increased thermal agitation makes more charge carriers available. This causes decrease of resistivity with increasing temperature, as indicated by the negative temperature coefficient of resistivity for silicon in Table 27-2. The same increase in collision rate that we noted for metals also occurs for semiconductors, but its effect is swamped by the rapid increase in the number of charge carriers.