Hello World ☆ 
This article is designed for beginners who want to learn about semiconductors.
I majored in physics during my university years and focused on theoretical physics.
However, during my master’s program, I unexpectedly delved into research on high-energy physics or particle physics experiments involving semiconductor detection devices.
This article compiles the knowledge I acquired about semiconductors during that time.
Given the potentially extensive content, feel free to explore based on your specific areas of interest. And most importantly, enjoy the read!
Understanding semiconductors, in other words, leads to understanding the movement of electrons.
After this, I will explain what semiconductors are. I will explain what semiconductors are and what they are not.
First, let me give you a historical background of how semiconductors developed.
The stage is set at Bell Laboratories (then known as AT&T Bell Telephone Laboratories).
This is that Bell that invented the telephone.
Before World War II, vacuum tubes were widely used.
The reason for this is that in the 1910s, telephone lines were laid across the continent from the East Coast to the West Coast of the United States, and vacuum tubes were an electrical component necessary for this communication technology. There are two main types of vacuum tubes: the dipole vacuum tube, invented by Fleming in 1904, and the triode vacuum tube, invented by de Forest in 1906.
The dipole tube has the characteristic of rectifying current in only one direction, which is an important characteristic in terms of current control.
A three-pole tube has an amplifying effect, which increases the signal.
When a call signal is sent over a telephone line, the signal becomes smaller due to various electrical losses.
Therefore, many 3-pole vacuum tubes were used to amplify this signal.
Off topic, but an old electronic calculator (as you can see in the picture, many vacuum tubes were used)
Scientists at this time were thinking of a replacement for the vacuum tube.
Specifically, they were looking at ores with the idea that vacuum tubes could be replaced by solid-state devices.
The reason for studying ores was that it was already known in the 1870s by Braun et al. in Germany that a rectifying effect was produced when a thin metal wire was brought into contact with an ore.
Scientists then analogized that, just as a rectifying two-pole vacuum tube developed into an amplifying three-pole vacuum tube, ore could be used in some way to create an amplifying effect.
By the way, why did they have to go to the trouble of finding a substitute for the vacuum tube when it has both rectifying and amplifying action?
The reason is that vacuum tubes have a number of drawbacks.
A vacuum tube has a structure similar to that of a light bulb and uses electrons emitted from a filament (thin metal wire) heated in a vacuum for amplification.
Since the filament must be red-hot (heated to bright red) to operate, it generates a lot of unnecessary heat and consumes a lot of power.
Another is that filaments have a short life span, so the AT&A Bell Telephone Labs telephone network, which used many tubes, required periodic replacement of tubes with broken filaments.
A further drawback was that because they were made of glass tubes, they were fragile, and maintenance was very expensive. In other words, it was a matter of cost.
That is why, since the beginning of World War II, Bell Labs had already been studying which ores were promising and how to make those crystals.
The reason was that they were needed for the development of radar and communications equipment.
Various ores were being used and experiments were being conducted on their rectifying action.
In particular, how high a frequency could be used was investigated.
The upper limit of the frequency of sound that can be heard by the human ear is about 20 kHz; Hz is read as hertz, meaning 1 Hz = the number of vibrations in one cycle per second.
(* When the period T(s) and frequency f(Hz) are set, the relationship f=1/T is established for these. 1 cycle per second, so substituting T=1s, naturally f=1Hz)
From this we can see that a rectifying action up to an electrical signal of 20 kHz is sufficient.
However, in order for ore rectifiers to be used in the radar of the time, they needed to be able to handle frequencies as high as MHz (M = 10 to the sixth power) and as low as GHz (G = 10 to the seventh power).
To identify airplanes and ships by radar, a spatial resolution of about 1 meter is required.
Resolution is the ability of a device or other object to measure or identify an object.
For example, most objects can be identified to the nearest 1 mm, so we can say that they have a resolution of about 1 mm.
The speed of radio waves is 300,000 km/second, so the frequency of a radio wave with a wavelength of 1 m can be calculated by dividing 300,000 km/second by 1 km to get
\(30×10^4×10^3 {\rm (m/s)}/1{\rm m}\)
\(= 3×10^8 {\rm Hz} \)
\(=(300 {\rm MHz})\)
(* Period (s)=1/frequency (Hz)).
Incidentally, in addition to the U.S. and Japan, the U.K. and Germany also focused on the development of electronic devices.
In summary, radar identification requires high performance in spatial resolution.
And to achieve high spatial resolution, a high frequency is necessary.
Ultimately, the two remaining ores that could respond to the highest frequencies in Bell Labs’ rectification experiments were two semiconductors: silicon and germanium.
Both are in the IV(14) group and are aligned above and below (see periodic table below).
Surprisingly, neither silicon nor germanium rectifiers were used and proven at the time.
However, researchers chose these two and spent eight years working on them. 1947 saw a major breakthrough.
While studying the properties of semiconductor surfaces, a researcher at Bell Laboratories discovered that amplification could occur under certain special conditions.
This was the first step toward the invention of the transistor (which, as will be explained later, replaced the vacuum tube).
Some things in the world conduct electricity and some do not.
For example, metals such as copper, silver, and iron conduct electricity. These materials that conduct electricity are called conductors or conductors.
On the other hand, glass and ceramics do not conduct electricity.
These substances that do not conduct electricity are called insulators or dielectrics.
Then, what is a semiconductor? It is a substance that conducts electricity under certain conditions and does not conduct electricity under certain conditions.
The most widely used semiconductor today is silicon.
Silicon is also called silicon in Japanese, and its element symbol is Si.
The following is an explanation of silicon.
A side note: It is often reported that silicone is implanted in the body, but it is not that silicone. What is used for cosmetic surgery is called silicone, which is a resin compound containing silicon (composed of two or more elements).
Silicon is the second most common element on earth, after oxygen, and is present in large amounts (* see Clark’s number if you are curious).
In most of nature, silicon exists as SiO2 (silicon dioxide) combined with oxygen.
Silicon dioxide is found as a major component of rocks.
The crystallized form of high-purity SiO2 is quartz.
As a man-made substance, it is found as the main component of glass and ceramics.
It is important to note that pure silicon does not exist in nature.
In other words, if silicon is desired, it must be refined and extracted by humans.
The first step is to melt SiO2, the raw material, and turn it into a sludge.
To melt SiO2, it must be heated to more than 1400°C.
For this purpose, an electric furnace is used. Electric furnaces use a large amount of electricity.
Therefore, Scandinavian companies that can use inexpensive electricity generated by hydroelectric power are the main manufacturers.
The purity of the refined silicon is very high (99.9…99% and the number of 9’s in a row of 10 to 11 digits from a quantitative aspect).
The reason why I mentioned silicon is because the characteristics of transistors and lasers change depending on the semiconductor used. And it depends on the intrinsic properties of the semiconductor.
In other words, the choice of semiconductor is very important.
To use an analogy, it is similar to tool selection. A marathon runner never runs in clogs.
The reverse is also true: a grandfather would never stroll on the asphalt wearing Mizuno soccer spikes.
In this way, we usually choose things based on who will use them and what they will be used for. Likewise with semiconductors, the properties of transistors differ greatly depending on whether silicon semiconductors or germanium semiconductors are used.
It is no exaggeration to say that all electrical appliances contain semiconductors.
Which semiconductor to research and commercialize is of great significance to researchers and companies.
Incidentally, when I was a graduate student, I researched something called a semiconductor detector using silicon semiconductors ☆.
Once pure silicon has been refined, the next step is to turn it into crystals.
So, what is a crystal?
A crystal is a structure in which atoms are arranged regularly and neatly.
The arrangement of crystals of silicon refined from SiO2 is coddled.
Crystals made of many tiny crystals are called polycrystals.
Therefore, it is necessary to make a beautiful arrangement.
A single, clean crystal without any joints is called a single crystal.
In other words, from polycrystal to single crystal.
The method introduced here is called the Choklarsky method.
The Choklarsky method was also used at Bell Laboratories.
First, highly pure polycrystalline silicon refined from SiO2 is melted again in a crucible (heat-resistant container).
At this time, the temperature of the silicon should be slightly higher than its melting point (1410°C).
Next, the melted silicon needs to be removed from the container.
How to do this is similar to fishing. Imagine a stick. Attach a tiny piece of silicon (called a seed crystal) to the end of the stick and place it in the molten silicon.
The seed crystal is slowly pulled up while rotating it. The liquid silicon attached to the seed crystal solidifies (also called crystal growth). This is what a single crystal is.
Citation URL : チョクラルスキー法 – Czochralski method – JapaneseClass.jp
The pulled-up mass of monocrystalline silicon is called an ingot (see image below).
Ingot image
Citation URL: ingotウェハー – Bing images
The pointed tip of the ingot gradually becomes larger as it is pulled up with the seed crystal.
Therefore, it becomes conical in the beginning. The speed of pulling depends on the type of material.
In the case of silicon, about 1 mm is pulled up per minute, so it takes a great deal of time.
The ingot is then cut into rings and the thin plate is called a wafer (the disk on the right side of the above image). The scene you want to see starts from about 30 seconds.
The inventor of the Choklarsky method was Jan Choklarsky (1885~1953).
At that time, ballpoint pens and pens that combined ink and pen were not yet widely used (they were invented in 1884, but the ink leaked badly, and it was not until 1943 that a usable pen appeared).
Therefore, the pen and ink were used separately.
Choklarsky’s invention was the result of a mistake.
He was studying the process of melting and solidifying tin in a crucible.
Then an idea came to him and he started taking notes. While engrossed in this, he moved his hand to the ink bottle to ink the pen, but accidentally put the pen in the crucible.
When he pulled it up, the metal nib was covered with tin as it was being stretched.
He did not miss it. His experiment to examine the solidification of tin was ruined, but he found a way to grow crystals.
From here, he explored good crystal growth methods and arrived at the Choklarsky method.
Since its application to semiconductor crystal growth at Bell Laboratories in the 1940s, it has been a very successful method for growing silicon crystals.
Transistors and integrated circuits are fabricated on silicon wafers.
Therefore, the larger the wafer, the more transistors and integrated circuits can be produced, which lowers the manufacturing cost.
Because of this, silicon wafers have become much larger over the past few decades.
As of this writing in 2021, wafers as large as 12 inches in diameter (about 30 cm) are being used, according to what I see on the Internet.
The crucibles are also quite large (I often refer to a photo of a crucible large enough to hold a person. I think the diameter is larger than a drum).
