Wafers and Transistors

Abstract

Metal oxide semiconductor (MOS) devices and integrated circuits can only be realised with semiconductor materials. The substrate material on which they are built plays a very dominant role in their performance and reliability. Aware of this fact, Lilienfeld used copper sulphide as a semiconductor starting material in 1930. Germanium was used during the early 1950s. Until 1960, however, usable MOS transistors could not be manufactured. During the 1960s, the move from germanium to silicon was made in which the transistor channel is a thin conductive layer, which is realised electrically. Table 8.1 shows a comparison of a few germanium (Ge) and silicon (Si) material constants, which clearly shows a few reasons why silicon is used today. For example, the circuits designed in the 65 nm CMOS node use transistors with a channel length close to 60 nm while operating at 1.2 V. This causes an electric field in the channel of 20 V/μm, which is much more than the maximum breakdown field in germanium but still less than that in silicon. Another reason is the widely accepted military spec, which can be met with this material. This spec requires products to function correctly at a maximum operating temperature of 125 °C. The maximum operating temperature of germanium is only 70 °C, while that of silicon is 150 °C. A major reason for these maximum temperatures is that, for many applications, the leakage currents become unacceptably large above these temperatures. Another advantage of silicon is that it is cheap because it is widely available as (beach) sand.