1.2 MOS structures
Carefully designed metal–oxide–semiconductor (MOS) structures are a common building block in digital electronics, primarily intended to form part of a transistor-based switch. However, throughout the active regions of a microelectronic chip there will be secondary MOS structures that arise because metal tracks are insulated from the semiconductor substrate by a layer of oxide; equally careful design is necessary to ensure that these do not form part of a switch. The acronym is a mixture of materials classification and materials, but it is so well established that it's too late to argue that conductor–insulator–semiconductor is more generic, so MOS it is.
This text is written presuming that MOS will be pronounced ‘em-oh-ess’. The alternative pronunciation, ‘moss’, is reserved for use within the longer acronym CMOS: ‘see-moss’.
Note down three important functions that can be performed by MOS structures.
MOS structures are central to the following functions:
the detection of light;
In practice, especially for mass-market consumer products, the semiconductor is silicon; the insulator is then silicon dioxide (see Box 1: Silicon and silicon dioxide). There are various options for the conductor: for several years it was aluminium; then it became expedient to use something more refractory such as heavily doped polycrystalline silicon, deposited from silane; the interconnecting tracks typically involve an exotic intermetallic compound layer followed by copper.
Box 1: Silicon and silicon dioxide
Solid-state electronics has made much use of silicon dioxide and silicon as archetypal insulator and semiconductor. They are excellent partners for the following reasons:
Silicon can be manufactured as wafers of extremely pure, single-crystal material.
Silicon of high purity can also be grown out of the vapour phase, for instance silicon tetrahydride (silane, SiH4), at elevated temperature.
The conductivity of silicon is readily manipulated locally by the addition of controlled quantities of dopant in combination with photolithography.
The oxide of silicon – namely silicon dioxide, SiO2 – can be readily grown onto a silicon surface on exposure to oxygen or steam at elevated temperature.
Silicon dioxide can also be grown out of the vapour phase at elevated temperature, so it can be incorporated over surfaces other than silicon.
Table 1 shows some bulk physical properties for three silicon-based materials and, as a comparator, aluminium. Exercise 1 will help you to appreciate the data.
Table 1 Comparisons of silicon, silicon dioxide and aluminium at 300 K
|Si (intrinsic)||Si (heavily doped polycrystalline)||SiO2||Al|
|Density / 103 kg m−3||2.3||2.3||2.2||2.7|
|Resistivity / Ω m||2.3 × 103||1 × 10−5||1 × 1015||2.7 × 10−8|
|Breakdown strength / V m−1||3 × 107||—||1 × 109||—|
|Melting temperature / °C||1415||1415||1600||660|
|Energy gap / eV||1.1||1.1||9.0||—|
|Atom density / 1028 atoms m−3||5||5||2.2||6|
Using data in Table 1, say:
(a) how much less conductive than aluminium is heavily doped polycrystalline silicon
(b) how much less conductive than aluminium is pure silicon
(c) how much less resistive than silicon dioxide is pure silicon.
(a) We want the inverse ratio of resistivities: heavily doped polycrystalline silicon is about 400 times less conductive than aluminium.
(b) We want the inverse ratio of resistivities: pure silicon is about 1011 times less conductive than aluminium.
(c) We want the ratio of resistivities: silicon is 4 × 1011 times less resistive than silicon dioxide.
There are two other things you should know about silicon and aluminium. First, they form a eutectic alloy at 11.7% Si which melts at 577 °C, so don't go thinking that process temperatures can approach 660 °C (see Table 1) once something has been made involving these two elements in contact. In fact, 450 °C is generally recognised as the safe upper limit. Second, aluminium – like gallium, boron and indium – is a well-known dopant capable of rendering silicon a p-type semiconductor.