From Deposition to the Window: How Are These Chips Made? (Plain English)

Title: From Deposition to the Window: How Are These Chips Made? (Plain English)

From Deposition to the Window: How Are These Chips Made? (Plain English)

Step 1: "Skinning" the Silicon Wafer

Take an N-type silicon wafer (resistivity 1–10Ω·cm) and load it into an LPCVD furnace at around 800°C. A layer of silicon nitride (e.g., 100nm) is grown on both the front and back sides. This layer becomes the future "window."

(Why both sides? To balance stress and prevent the wafer from warping.)

Step 2: Drawing the "Window" (Front Side)

Spin-coat photoresist onto the front nitride, then use photolithography to define the window area. The resist protects the "window," while the surrounding nitride is etched away using plasma, exposing the silicon underneath. Think of this as tracing the outline of a "skylight" on a ceiling.

Step 3: Digging the Hole (Back Side)

Flip the wafer over. Align a lithography pattern on the back side to the exact same location. Then, immerse the wafer in KOH solution at 80°C. KOH eats silicon like hot water melts sugar, but it barely touches the silicon nitride.

The silicon is etched away from the back until it is just about to reach the front nitride membrane. The result: a freestanding nitride membrane stretched across a silicon frame—like digging a hole in a thick wooden plank and leaving a sheet of paper covering the gap.

Step 4: Dicing, Cleaning, and Packing

Once the whole wafer is processed, it is diced into individual 5×5mm chips. After stripping residues and cleaning, each chip is inspected for defects in a Class 100 cleanroom, placed into Gel-Pack boxes, and shipped (10 pieces per box).