The Origin of Semiconductors

Semiconductors are currently being discussed a lot, about materials, about devices, about investment. Each topic has a different focus. The materials are nothing more than the third-generation wide-bandgap semiconductor materials; the devices are mainly based on WBG materials such as MOS, HEML, etc... Today we are going to talk about the history of the semiconductor industry and understand the origin of the semiconductor industry. (More Semiconductor explained: https://www.utmel.com/blog/categories/semiconductor/what-is-a-semiconductor)

Vacuum triode

Speaking of the vacuum triode, it was invented by Lee Deforest in 1906, which made it possible for consumer electronics such as radios and televisions. It was also the brain of the world’s first electronic computer. The big guy with an area of nearly 140 square meters and a weight of 30 tons used nearly 19,000 vacuum tubes and thousands of resistors and capacitors. The vacuum triode is composed of a grid and two electrodes separated by the grid in a space sealed by glass. The inside of the sealed space is vacuum. On the one hand, it prevents parts from burning and on the other hand, it is easy for electrons to move freely.

It has two major functions: switching and amplifying. Switching refers to turning on and off current, and amplifying refers to current or small signal amplification, and maintains the original characteristic functions of the signal. But there are a series of shortcomings, such as large size, easy to loose joints, short life span, and fast aging.

This series of shortcomings led to its inevitable development or replacement. In 1949, John Bardeen, Walter Brattin, and William Shockley of Bell Labs (William Shockley) invented an electronic amplifier made of germanium, a semiconductor material, which is the first generation of transistors (the three scientists won the Nobel Prize in Physics in 1956). The transistors were originally called "transmission resistors" and later renamed transistors.

This kind of transistor includes the function of a vacuum tube, and at the same time, it is solid. It has the advantages of small size, lightweight, low power consumption, and long life. Since then we have entered the "solid-state era", which is what we have seen most.

From the development of the first transistor to our many power electronic devices, power electronic technology has been developed by leaps and bounds. These devices are generally referred to as discrete devices, that is, each chip contains only one component. And another semiconductor-related device, which we have not mentioned before, is the integrated circuit.

Integrated circuit

The integrated circuit is a kind of miniature electronic device. It uses a certain process to interconnect the transistors, resistors, capacitors, inductors, and other components and wiring required in a circuit on a semiconductor substrate, and then encapsulate them in a tube. Inside the shell, a miniature structure with the required circuit functions is formed.

The first integrated circuit was invented by Jack Kilby of Texas Instruments, but it was not in the form of integrated circuits today.

It was originally connected by separate wires. Earlier, Jean Horni of Fairchild Camera has developed a planar manufacturing process that forms electronic junctions on the chip surface to make transistors. That is the advantage that silicon is easy to form insulator silicon oxide. Then Robert Noyce applied this technology to connect discrete devices previously formed on the silicon surface, and finally formed the pattern used by all integrated circuits.

Crafts and trends

Since 1947, the semiconductor industry has begun to improve and develop in technology, and it is still developing today. Process improvement can be attributed to two categories: process and structure.

Process improvement: to manufacture devices and circuits in smaller sizes, so that they have higher density, more quantity, and higher reliability.

Structural improvements: inventions in the design of new devices make their performance better, achieve better energy consumption control, and higher reliability.

The effects of these two improvements can be well understood from the development history of several generations of IGBTs. Of course, many of the factors we mentioned before, such as some defects in the equipment process, can be attributed to the improvement of the process.

The size and quantity of devices in integrated circuits are two common signs of IC development. The size of the device is represented by the smallest size in the design, which we call the feature pattern size. The development from small-scale integrated circuits to today's millions of chips has benefited from the reduction in the feature pattern size of a single element, which has benefited from the greatly improved lithography machine patterning process and multilayer wiring technology. This point is also more discussed today, such as 22nm chips, 10nm or even 7nm, and so on. A more professional expression is the gate width, that is, we control the width of the part of the gate. Smaller and faster transistors and higher-density circuits benefit from the smaller gate width.

Speaking of this, I have to mention the term "Moore's Law" that I often see some time ago. Gordon Moore, one of Intel’s founders, predicted in 1965 that the number of transistors on a chip would be doubled. The growth rate of doubling every month is more accurate after years of actual verification, and it has become the basis for predicting the density of transistors on chips in the future. According to the number of components in the circuit, that is, the level of integration, we can divide it into several levels: small-scale integration (SSI): 2~50 pieces/chip; medium-scale integration (MSI): 50~5000 pieces/chip; large-scale Integration (LSI): 5000~100000 pcs/chip; Very Large Scale Integration (VLSI): 100000~1000000 pcs/chip; Very Large Scale Integration (ULSI):>1000000 pcs/chip.

Moore’s Law does not develop unlimitedly over time. It is mainly limited by the limitations of semiconductor materials and preparation. Therefore, you will hear the news that the semiconductor material silicon is reaching its limit. Therefore, it is necessary to develop new materials and constantly upgrade equipment and design.

Chip and wafer size

We all know that chips are made on thin wafers of silicon or other semiconductor materials called wafers.

Manufacturing rectangular chips on a round wafer resulted in some unusable areas remaining at the edge of the wafer. When the chip size is larger, these unusable areas will also be larger, so larger wafers are gradually adopted, which also "reduces" the size of the chip in disguise, and also enables a certain degree of production efficiency and yield. Upgrading is one of the reasons why wafers from 6 inches to 8 inches to 12 inches are now.

Now the chip size is getting smaller and smaller, the cost is getting lower and lower, and the performance is getting higher and higher. Thanks to the process improvement and equipment development we mentioned above, this has also led to more and more fierce competition in the semiconductor industry.

For now, there are roughly three types of manufacturers: Integrated Device Manufacturers (IDM): integrate design, manufacturing, packaging, and sales; Foundry: other chip suppliers make chips; no processing plants (Fabless ): Only responsible for chip design and sales, and most other links are outsourced.