en English

From sand to Wafer, what has gone through?

Why can sand be used as a raw material to make wafers?

The main reason for choosing sand is that SiO2 is the main component of sand, and silicon (Si) is the raw material of semiconductors, so it is directly extracted from sand. First, sand (SiO2) reacts with carbon at high temperatures (2000C) to produce silicon and CO2. At this time, the Silicon is metallurgical Grade (MGS: Metallic Grade Silicon), which is crude polysilicon. After purification of polysilicon, the purified TCS reacts with H2 at 1100C to generate electronic grade (EGS: Electric Grade Silicon) polysilicon and HCl, and then the production of monocrystalline silicon.

How to get from ingots to individual wafers?

  • Trimming/groove-cutting: Wafer has a Notch (6 inches is flat), which should be cut when the ingot is finished. Must be along
  • Cutting silicon ingot into wafer: the line cutting (one can cut many slices) is adopted. The knife rotates but the silicon ingot /ingot only translates and does not rotate. The thickness of each wafer is determined by the distance between the two blades. Generally, the thickness of the wafer is 525mm for 4 inches, 625mm for 5 inches, 675mm for 6 inches, 725mm for 8 inches, and 775mm for 12 inches.
  • Edge Stitching: A well-cut wafer must have a sharp cylinder at its Edge. It needs to be sharpened into a circle to reduce stress. Grind the wafer in a fixed slot, like a knife.
  • Lapping: Because the wafer that has just been cut must have a lot of damage and a rough surface, this step is similar to CMP and can be ground flat with slurry. Therefore, the wafer is sometimes called a Polish wafer.
  • Wet etch: because the polishing is still mechanical, it still cannot completely remove the damage, so it requires a chemical reaction to remove the surface defect. HNO3+HF+CH3COOH(acetic acid) was mainly used at 4:1:3. In Nitric acid oxidation, HF eats SiO2.
  • Anneal: For fear of lattice damage, annealing can remove lattice damage. Ar gas is generally used, so we sometimes see our flow with the Ar Anneal wafer. The reason we use Ar is that Ar is an inert gas that doesn’t react, and the reason we don’t use H2 was previously said to cause the surface concentration to change.

Here the silicon wafers are done. At present, the global silicon wafers have been concentrated in the hands of the top five suppliers, including Shin-Etsu Semiconductor, Shenggao, Taiwan’s Global Crystal, Germany’s Silitronic, and South Korea LG, etc., with a global market share of 90%.

What is the Wafer?

Wafers are composed of pure silicon (Si). It is generally divided into 6-inch, 8-inch, and 12-inch specifications ranging. Samsung, Toshiba, Western Digital, Micron, etc. purchase silicon wafers to make into NAND Flash and DRAM wafers. Because of the different nanotechnology used by each company, the produced NAND Flash chips differ in performance, cost, and other aspects.

After making it into a NAND Flash wafer, the Wafer will be cut into one by one wafer, the professional term is called Die, the capacity unit of a single Die is Gbit, different from the unit of chip Gbyte, because 1 Byte=8 bit.

After cutting, then after testing, Samsung, Toshiba, Western Digital, Micron, etc. will be divided into qualified and unqualified Flash Die, test qualified Flash Die for packaging, unqualified Downgrade Wafer is some residual edges, the original factory will do scrap processing, there are also many in the form of blue film Wafer flow to the downstream packaging plant for processing, recycling Utilization, known as the black film Ink Die, and finally used by some small manufacturers for cheap bundled SD cards and USB flash drives and other products.

Good Die is stacked and packaged into different NAND Flash chips, and then packaged with control chips to produce eMMC products. eMCP is a multi-chip package of eMMC and LPDDR and then used in cell phones, tablets, smart boxes, and other end products. SSD is a PCBA board with one or several NAND Flash chips, and control chips or DRAM chips. SSD is made of one or more NAND Flash chips and control chips or DRAM chips on a PCBA board and is mainly used in the field of computers and servers.

Table of Content

Leave a Comment

Your email address will not be published.

two − 1 =

Let's Have A Chat

Learn How We Served 100+ Global Device Brands with our Products & Get Free Sample!!!

Email Popup Background 2