From Sand to Polished Wafer: How Silicon Ingots Are Made

The semiconductor industry begins with one of the most common materials on earth: silica sand. Turning sand into the base material for integrated circuits requires precision, scale, and control. Every polished wafer you see in a fab starts from this process.

Step 1: Mining and Preparing Silica

You start with quartz-rich sand. This sand is mined from deposits with high purity, often more than 98 percent silicon dioxide. The sand is washed and processed to remove impurities such as iron and aluminum. The goal is to reach metallurgical-grade silicon with a purity of about 98 to 99 percent.

Step 2: Producing Metallurgical-Grade Silicon

The prepared sand is fed into electric arc furnaces at temperatures above 2000°C. The oxygen separates from the silicon dioxide, leaving molten silicon. This product is metallurgical-grade silicon, used in steelmaking and solar industries. For semiconductors, you need far higher purity.

Step 3: Chemical Refining to Polysilicon

To reach electronic grade purity, producers refine metallurgical silicon through chemical reactions. One common process is the Siemens method. In this process, silicon reacts with hydrogen chloride gas to form trichlorosilane. This gas is distilled to remove trace metals and then decomposed back into pure silicon. The result is polycrystalline silicon rods with a purity of 99.9999999 percent. At this point, the material is suitable for integrated circuits.

Step 4: Growing a Monocrystalline Ingot

Your next step is to transform polycrystalline silicon into a single crystal. The most widely used process is the Czochralski method.

• A quartz crucible is filled with pure polysilicon chunks and melted at about 1500°C.

• A small seed crystal is dipped into the melt and slowly pulled upward while rotating.

• The molten silicon solidifies on the seed, forming a large cylindrical single crystal.

This process produces ingots up to 300 millimeters in diameter and weighing several hundred kilograms. TSMC, GlobalFoundries, and Analog Devices all rely on such wafers as their starting point.

Step 5: Shaping the Ingot

Once the ingot is complete, it is ground to achieve a uniform diameter. A flat or notch is added along the side to indicate crystal orientation. The ends are cut off to remove imperfections from the seed and tail.

Step 6: Slicing the Wafers

Diamond wire saws slice the ingot into thin discs. A single ingot yields thousands of wafers, each typically between 725 and 775 micrometers thick for 300 millimeter wafers. During slicing, manufacturers aim to minimize kerf loss, the wasted silicon lost as dust.

Step 7: Surface Treatment

Freshly sliced wafers have saw marks and surface damage. They go through lapping, etching, and chemical cleaning to remove these defects. Lapping flattens the surface. Etching removes damage from the slicing process. Chemical baths prepare the wafer for polishing.

Step 8: Polishing

The final step is chemical mechanical polishing. Wafers are pressed against polishing pads with a slurry containing abrasive particles and reactive chemicals. This produces a surface with atomic-level flatness and mirror-like quality. At this stage, the wafer is ready for semiconductor device fabrication.

Why Polished Wafers Are Expensive

A 300 millimeter polished wafer today costs up to 25,000 dollars once processed into a finished IC wafer in advanced fabs. Several factors drive this cost:

• Purity levels of 99.9999999 percent require complex refining.

• Growing and handling large single crystals takes weeks of continuous operation.

• Each wafer passes through hundreds of process steps before reaching final IC form.

• Advanced nodes like 5 nanometer and below require extreme precision and expensive equipment.

When you hold a polished wafer, you are looking at the result of thousands of steps, billions of dollars of infrastructure, and decades of process development.

© 2025 Silicon Masters LLC - All Rights Reserved