The detailed manufacturing process of chips!

█ Oxidation

First, an oxide layer is applied to the diced and polished wafer. The purpose of oxidation is to form a protective film (oxide layer) on the fragile wafer surface. This oxide layer prevents the wafer from being affected by chemical impurities, leakage current, and etching.

Dry oxidation involves introducing pure oxygen, which flows across the wafer surface and reacts with silicon to form a silicon dioxide layer. Wet oxidation uses both oxygen and highly soluble water vapor.

█ Photolithography (Coating, Pre-baking, Exposure, Post-baking, Development)

Photolithography, simply put, is like a printer "etching" the chip circuit pattern onto the wafer.

Photolithography can be divided into three main steps: coating, exposure, and development. Let's look at each one in turn.

First is coating. This adhesive is called photoresist, sometimes also called photoresist, and is a photosensitive material.

There are two types of photoresist: positive and negative.

Positive photoresist, when exposed to a specific light beam (exposure), undergoes a change in its molecular structure, becoming more easily dissolved. Negative photoresist, on the other hand, becomes difficult to dissolve after exposure. Positive photoresist is used in most cases.

Coating

A photomask is a glass or quartz plate with a patterned layer of an opaque material (such as chromium). This pattern is essentially the blueprint for the chip, or the integrated circuit layout.

Photomask

In a lithography machine, the wafer and the photomask are precisely fixed. Then, a special light source (mercury vapor lamp or excimer laser) emits a beam of light (ultraviolet light). This beam passes through cutouts in the photomask and multiple lenses (to focus the light), ultimately projecting it onto a small area of ​​the wafer.

The intricate circuit pattern is thus "projected" onto the wafer.

Using positive photoresist as an example, the photoresist in the exposed area becomes more easily dissolved. The unexposed photoresist remains undamaged.

The mechanical fixtures holding the wafer and photomask constantly move, and the light beam continuously illuminates the wafer. Ultimately, the circuitry for dozens to hundreds of chips is "drawn" on the entire wafer.

The Lithography Process

After the silicon wafer exits the lithography machine, it undergoes a heating and baking process (baking for 20 minutes at 120-180℃), known as post-bake.

The purpose of post-bake is to ensure the photochemical reaction in the photoresist is fully completed, compensating for insufficient exposure intensity. Post-bake also reduces the ring-like patterns that appear after photoresist development due to the standing wave effect.

Next comes development. After exposure, the wafer is immersed in a developing solution. The developing solution removes the exposed photoresist (positive photoresist), revealing the pattern.

█ Etching

Okay, let's continue discussing the chip manufacturing process. Now, although the pattern is visible, we've only removed a portion of the photoresist. What we really need to remove is the underlying oxide layer (the part not protected by the photoresist).

In other words, we need to continue "digging" deeper. The process used at this stage is etching. Etching processes are divided into two types: wet etching and dry etching.

Wet etching involves immersing the wafer in a liquid solution containing specific chemicals, using a chemical reaction to dissolve the semiconductor structure (oxide film) not protected by the photoresist.

Dry etching uses plasma or ion beams to bombard the wafer, removing the unprotected semiconductor structure.

Two concepts are important in etching processes: isotropy (or anisotropy) and selectivity.

Wet etching etches in all directions, hence the term "isotropy." Dry etching etches only in the perpendicular direction, hence the term "anisotropy." The latter is clearly better.

During etching, both the oxide layer and the photoresist are etched. Under the same etching conditions, the ratio of the photoresist etching rate to the etching rate of the material being etched (oxide layer) is the selectivity. Clearly, we need to etch as little photoresist as possible and etch as much oxide layer as possible.

█ Doping (Ion Implantation)

Okay, that concludes the "hole-making" process. At this point, the wafer surface has been etched with various trenches and patterns. Next, let's look at the doping process.

When introducing basic chip knowledge (How exactly do semiconductor chips work?), I mentioned that transistors are the basic building blocks of chips. Each transistor is based on a PN junction. As shown in the diagram below (MOSFET transistor, NPN), it includes P-wells, N-wells, channels, gates, etc.

The previous photolithography and etching only created the holes. Next, we need to construct P-wells and N-wells based on these holes. Pure silicon itself is non-conductive. To make non-conductive pure silicon a semiconductor, we must introduce impurities (called dopants) into the silicon to change its electrical properties.

For example, doping silicon with phosphorus, antimony, and arsenic can produce N-wells. Doping with boron, aluminum, gallium, and indium creates a P-well.

N atoms have free electrons. P atoms have many holes and a small number of free electrons. By applying a gate and voltage to the channel, electrons in the P atoms are attracted, forming an electron channel (channel). Applying voltage between the two N atoms creates a current between the NPN junctions.

As shown in the diagram below:

In the diagram, the bottom is the P-well substrate. The two holes are N-wells. That is, when making this NPN transistor, ion implantation is used before the initial oxidation process. The substrate is first doped with boron (containing a small amount of phosphorus), becoming a P-well substrate. (For ease of reading, I didn't explain this step earlier.) Now, the hole-making part can be doped with phosphorus to become an N-well. Do you understand? The purpose of doping is to create a PN junction, creating a transistor.

Doping includes two processes: thermal diffusion and ion implantation. Because thermal diffusion is difficult to achieve selective diffusion, ion implantation is currently used for most applications except for specific needs.

Ion implantation is the process of using a high-energy particle beam to directly inject impurities into a silicon wafer.