Skip to content

How does sand become chips? The raw material for the chip is sand

How does sand become chips?

How important are chips in our lives? As small as wireless headphones, smart phones, as large as cars, trains, planes and ships, and then to refrigerators, TVS, washing machines and so on are inseparable from chips. It can be said that the chip has become an indispensable electronic component in our life. But can you imagine such a sophisticated device being made of sand? This article will explain the process of turning sand into chips step by step.

Silicon extraction and purification

Chips are made of elemental silicon, but there is no natural elemental silicon in nature, mostly in the form of silicon oxides (silica, SiO2), silica happens to be the main component of sand.

The sand used in the chips is not the kind we normally see on beaches, rivers, construction sites and so on, because it contains red, yellow and orange impurities that are harder to purify.
The sand used in the chips The sand used in the chips is called silica sand, which has few other impurities and is pure silica. So how do you turn silicon dioxide into elemental silicon? Carbon is often used in industry to reduce silica at high temperatures and then to produce a coarse product containing small amounts of impurities.

Chemical formula: SiO2+2C== high temperature ==Si+2CO↑

We thoroughly mix silica and carbon, and then put them into an electric arc furnace and heat them to more than 2000℃. At high temperature, carbon will react with silica, leaving elemental silicon at the bottom of the furnace, and CO will be discharged as a gas.

The silicon is then treated with oxygen to remove impurities such as calcium and aluminum, resulting in a purity of 99 percent silicon. This is nowhere near the purity required to make a chip.

To further purify the elemental silicon, we grind it into a powder, add hydrogen chloride, put it in a fluidized bed reactor, and heat it to 300℃ so that it reacts fully to produce trichlorosilane, while removing impurities such as iron, boron, and phosphorus. The trichlorosilane is then heated to 1, 000 degrees Celsius, where it reacts with hydrogen gas to form 99.999999% pure silicon, which can be used to make chips.

But the elemental silicon at this point is only polysilicon, made up of a large number of small crystals, or microcrystals, which are joined by grain boundaries, which can confuse electrical signals, so the structure of the silicon must be changed to become monocrystalline silicon.

Monocrystalline silicon growth

The process of converting polycrystalline silicon into monocrystalline silicon is called Celeraic single crystal technology, also known as long crystal. Its technological process is as follows: melting 1 → neck shrinking 1 → shoulder growing 1 → equal diameter growing 1 → tail growing.

Melting: Simply put, polysilicon is added to the quartz pot and heated to above the melting point (1420℃). The whole project cannot contact with air, so the long crystal furnace is generally pumped into a vacuum, and then argon is added.

Shrinkable growth: tiny grains are put into the melted silicon melt. Thermal stress is generated on the contact surface due to temperature difference, resulting in dislocation. At this time, the growth of seed crystals is rapidly increased, and the diameter of the grown seed crystals is reduced to about 46mm.

Shoulder growth is to enlarge the diameter of the crystal to the desired size, which requires attention to cooling and reducing the pulling speed.

Equal diameter production, is to keep the crystal diameter unchanged, continuous elongation process.

Tail growth is the opposite of shoulder growth. When the diameter is reduced to a point, the rod will separate from the silicon melt. The whole growth process is over. Polysilicon then becomes a single crystalline silicon rod, which is often called a silicon ingot.

Currently, most silicon ingot are 150mm, 300mm, 450mm in diameter and are used to make 8-inch and 12-inch wafers.

Wafer making

Silicon rods cannot be directly etched. They need to be cut, ground, sliced, chamfered, polished and laser engraved before they become the basic material for chip manufacturing — the “wafer”.

Wafer slicing methods include traditional mechanical cutting (slicing) and laser cutting (slicing).Wafer making
Mechanical cutting is the direct action of the diamond saw blade on the surface of the wafer, creating stress and causing it to be divided. Generally, the cutting width is between 25-35μm, and the speed is 8-10mm/s. The cutting speed is slow. Different specifications require different cutting tools. In addition, mechanical cutting is easy to cause wafer edge collapse, damage and other phenomena. While the rotary grinding wheel cutting can reduce the wafer damage, but it needs deionized water cooling, resulting in high cost. A new type of laser cutting was invented.

Laser cutting will not produce mechanical stress, greatly ensure the quality of the wafer. In addition, the laser is more accurate, up to the sub-micron level, which is ideal for precision machining. Under the powerful pulse energy, the silicon material is directly vaporized to produce a uniform channel to achieve cutting.
Laser cutting
Laser cutting speed is faster, and does not need cooling water, more will not appear the problem of wear tool, can be 24 hours uninterrupted operation. Laser has better compatibility and versatility for wafer. After cutting, the wafer is coated by oxidation and chemical vapor deposition, so that a layer of SiO2 film is formed on the surface, and N-type and P-type doping are carried out in the SiO2 film.

Is it not a lot of netizens think that the wafer is like a DVD disc, in fact, it is not so, the wafer is not a standard round, generally will cut out a side, as a triangle like the “bottom”, become a “bottom circle”.

Chip manufacturing

Chip manufacturing is the most complex process, requiring more than 20 steps just to go through the process, while the actual manufacturing of a mobile phone chip requires hundreds, even thousands of steps.Chip manufacturing
Just to talk about the main process.

The main processes of chip manufacturing are: cleaning, pre-drying, gluing, pre-drying, alignment, exposure, development, vertical film, etching, and degluing.

Step 1 : cleaning

The purpose of cleaning is to remove metal ions and ensure that the wafers are free of pollution, pinholes and defects. The specific steps are as follows:

1. Use strong oxidant cleaning solution to dissolve metal ions on the wafer surface in the cleaning solution;

2. Use H+ to remove metal ions remaining on the wafer surface;

3. Clean the wafer with deionized water and completely remove the cleaning solution.

Step 2 : pre-bake

After the wafers are cleaned, they are dried to ensure that there is no residual cleaning solution, which is more conducive to coating the substrate, and the temperature of the wafers is maintained at around 80℃.

Step 3 : apply the base glue

In order to enhance the adhesion between the photoresist and the wafer, a layer of substrate is needed, and the ambient temperature of the substrate coating needs to be kept at about 100℃.

Step 4 : photoresist coatingphotoresist coating
Photoresist is a mixture of photosensitive resin, sensitizer and solvent. After being irradiated by ultraviolet light, electron beam and X-ray, the solubility will change.

Methods: The wafer was oxidized at 1000℃, and then the photoresist was evenly coated on the wafer surface by gluing machine.

Step 5 : before baking

The purpose of pre-drying is to vaporize the solvent of the photoresist, so that the photoresist adheres evenly to the wafer surface, and the dry photoresist coating is thinner.

Step 6 : Alignment

The accuracy of alignment operation is extremely high, which is definitely a major test of technology and equipment.

The alignment accuracy is required to be about 1/10 of the thinnest line width, and the accuracy varies with the process of the chip. For example, the chip alignment accuracy of 5nm process is 1nm.

Step 7 : ExposureExposure
Exposure is similar to pressing the shutter of a camera. When UV light hits the wafer coated with the photoresist, the photoresist reacts chemically.

The positive adhesive exposure part is dissolved in the developer, leaving no exposed adhesive layer. The exposed part of the negative adhesive does not dissolve in the developer, while the unexposed adhesive layer is dissolved.

Step 8 : development

When the exposed wafer is placed in the developer, the photoresist in the sensitive region dissolves. When this is done, the circuit diagram on the wafer is revealed.

Step 9 : vertical die

After the development process, the dominant pattern without photoresist and the recessive pattern with photoresist are formed. This pattern combination can be used as a template for the next step.

Step 10 : etchingetching
Etching can selectively remove specific parts of the wafer and trim or clean the edge. There are two main types of etching technology: wet etching and dry etching.

Etching is very important for the electrical properties of devices. If there is a mistake in the etching process, it will cause the hard to recover the silicon chip scrap.

Step 11 : Remove glue

After the above process is finished, the photoresist should be removed by plasma, special solvent, etc.

This is just one operation, and in fact, when you make a chip, you have to do it over and over again. Each layer is lithographed, resulting in a multi-layer, stereoscopic chip.

In order to prevent the chip scratch, and more convenient to connect with other components, but also to the chip packaging.

Package

Package
When the chips are made, they are about the size of a fingernail and so thin that they can be scratched and damaged, wiping out all that work.The chip package is to install a shell for the chip, which can effectively fix, protect, seal and connect the chip.

The metal connectors after the chip package are simply a bridge to communicate with the outside world. These Bridges connect the chip to the circuit board effectively, so that the chip can play a more secure and efficient role.The packaging process mainly includes: slicing, loading, bonding, plastic sealing, flanging, electroplating, printing, cutting molding.There are two common encapsulation methods: DIP package and BGA package.

DIP package

DIP package is double-row vertical package. The IC chip with this package looks like a black centipede at the foot of the double-row, which is impressive. This packaging method is the earliest IC packaging technology.

Pros: Low cost, suitable for small chips that don’t require too many connections.

Disadvantages: The heat dissipation effect is poor, can not meet the requirements of the current high-speed chip.

BGA package

Attach the chip with a gold wire to a metal pin,

Advantages: small size, accommodate more pins;

Disadvantages: complicated process and high cost.

After the encapsulation is completed, the functional, electrical, safety, environmental, mechanical and other aspects of the test, the job is done.

Conclusion

Turning sand into high-end chips is an extremely complex and delicate process that no single country can currently accomplish alone. The process requires not only a great deal of expertise, but also a lot of sophisticated equipment and chemicals with different properties. This is a test for every country. China started late in this field and has a thin foundation, but the development speed is very fast. We believe that in time, the efforts of scientific researchers will be able to independently build high-end chips.

Share:

Facebook
Twitter
Pinterest
LinkedIn

Product Categories

Most Popular

Contact Us

IBE News

Related Posts

Industries

INDUSTRIES WE SERVE Click on each industry to learn more about how we make custom printed circuit board solutions for each market.Our industries served include

Read More »