MOSFETs revisited

I read an article today in [1] regarding the replacement of silicon chips with germanium chips in electronics. Silicon transistors are rapidly approaching miniaturization limits while germanium counterparts show two or three times better speed performance. These chips consists of billions of small transistors called MOSFETs. In this blogpost I will explain the working principle of MOSFETs (Metal Oxide Semiconductor Field Effect Transistors) widely used in electronic industries. For this, I will use the case of an nMOS (MOS where conduction is due to electrons), which is built on a p-type Si wafer as in Figure 1. MOS1

Figure 1. Image taken from [2], page 35.

The MOS consists of the transistor body, p-type in this example, and two n-type regions called source and drain.  The body is referred to  as semiconductor. On the top of n and p regions it is grown an oxide (insulator, usually SiO2 – regular glass) and then the Gate (formerly metallic, now consisting of conductive silicon).  The p-type body composed of Si is doped with acceptor atoms such as B, Ga or In, which makes available a lot of positive mobile carriers called ‘holes’. The n-type drain and source  formed of Si are doped by donor atoms such as P, As or Sb, which leads to an increased concentration of negative mobile carriers (electrons), able to move and conduct electricity. The doping of Si is accomplished through diffusion or ion implantation. As you can see in the above image (Figure 1) there is no path between the source and the drain for the carriers to flow. The area between the source and the drain is called channel and here all the magic takes place. In order to form this path a positive voltage is applied on the gate (VGS), as in Figure 2. MOS3

Figure 2. Image taken from [2], page 39.

When VGS is large enough positive charges accumulate on the gate. It is called field effect transistor (FET) because the positive voltage on the gate attracts the electrons close to the surface of the semiconductor (effect called inversion because we have an n-type region in a p-type substrate), so the transistor action is due to the field from the gate to the channel region. Now a path between the source and the drain is formed and current can flow. In order to keep this going another potential on the drain (VDS) has to be applied. The current is positive in the direction of moving positive charges so, this is why ID has an opposite direction as compared to the direction of moving electrons. If we plot the current vs. the drain-source voltage we have the following dependence: MOS2

Figure 3. Image taken from [2], page 45.

Two regions can be distinguished: a non-saturation region where the current varies approximately linearly with the applied voltage and a saturation region where the current is almost constant with the increase of the voltage. The following electron microscopy image shows how a real device looks in reality. MOS4

Figure 4. Image taken from [3], page 129.

When you have both p-type and n-type MOS on the same chip you have built a CMOS (Complementary MOS).  For a detailed tutorial regarding the math and the equations behind the MOS Transistors you should read this book [2] or watch this course [4]. References

[1] Katherine Bourzac, New Chip Points the Way Beyond Silicon.

[2] Yannis Tsividis, Operation and Modeling of The MOS Transistor, Second Edition, OXFORD UNIVERSITY PRESS.

[3] C. Auth et al., Proc 2008 Symp. VLSI Technology, p. 128-129.

[4] MOS Transistors, Coursera.

Advertisements