This time, we'll take a closer look at photolithography. First, I'll explain the overall process sequence.
Photolithography sequence
The photolithography process can be broadly broken down into three stages: coating the photoresist (PR), exposing it to UV light through a mask, and then developing it to remove any unwanted photoresist. Let's take a closer look at each step.
Surface preparation
After wafer cleaning and before applying PR for the photolithography process, the wafer surface must be rendered hydrophobic. Because PR is hydrophobic, making the wafer hydrophobic improves the adhesion of the PR. This process can be accomplished by exposing the wafer to HMDS (HexaMethylDiSilazane) vapor.
Spin coating
PR is applied using spin coating. Initially, the wafer is spun at low speed to ensure the entire surface is covered with PR, then spun at high speed to achieve the desired thickness. It's been a year since I last used this process, but I recall spinning it at 500 rpm for 20 seconds and then at 5,000 rpm for 40 seconds.
Soft Bake
After PR coating, baking is performed to remove the organic solvent contained in the PR. This slightly hardens the PR, but this can cause the solvent to evaporate during the subsequent exposure process (UV exposure), which can negatively affect the mask or exposure equipment and reduce the sensitivity of the exposure process.
I baked it on a hot plate at 120 degrees for 2 minutes. This was because there was more baking involved, but I standardized the process recipe for convenience. In my experience, I don't recall a soft bake having a significant impact on resolution. (Of course, this could be because I'm not using a nanometer-scale process, but rather a 5㎛ pattern!)
Align & Exposure
After the soft bake, the photolithography process begins. The light used during the photolithography process is ultraviolet, as the photoresist reacts with ultraviolet light. This is why, when discussing semiconductors on TV, people often work in rooms with yellow lights. To prevent the photoresist from reacting during stages other than the photolithography process, only yellow light is used during the photolithography process (the yellow room). That's why we also attached yellow cellophane to the fluorescent lights.
Let's digress for a moment. Before proceeding with the exposure process, we first do something called alignment. They say that when making a chip, the photo process is done about 20 to 25 times. Patterning, deposition, patterning, deposition, and this process must be repeated to create a chip with the desired circuit. Then, the location of each layer must be precisely aligned each time, right?? This process is the alignment work. Most of the time, there is an alignment key at a specific location, and the location is aligned using that.
After Align is complete, proceed with UV Exposure.
Post Exposure Bake (PEB)
I mentioned earlier that, in my experience, Soft Bake doesn't seem to significantly affect the resolution of the photo process, but PEB is truly crucial. It's performed at a higher temperature (110-120 degrees) than Soft Bake and is used to create a uniform PR pattern. During the UV exposure process, acid is generated in the PR, and this acid diffuses throughout the PR, stabilizing the pattern.
At this time, the temperature had to be precisely controlled, so I checked the pattern shape one by one while changing the temperature. Since the pattern came out best at 120 degrees for 2 minutes, I set the recipe to PEB, Soft Bake 120 degrees for 2 minutes Baking.
Develop
Up to this point, the entire wafer has been coated with PR. Since the mask was used for exposure, there will be areas where the PR is illuminated and areas where it isn't. Patterning is performed using a solution called Developer. The type of PR varies depending on whether the illuminated or unilluminated areas are bleached.
Once the development is complete, the user's desired pattern will be created on the wafer, which can then be washed with DI (Deionized Water) and dried.
Hard Bake
This is the final step in the photo process. Hard baking is performed at 110-130 degrees Celsius. This removes any remaining solvent and moisture from the PR, strengthens its hardness, and improves adhesion between the PR and the substrate. (I used a 5-minute recipe at 120 degrees Celsius.)
Characteristics and types of photoresist (PR)
Photoresist's chemical structure changes when exposed to light, causing the areas exposed to and unexposed to light to melt during the subsequent development process. This characteristic can be used to selectively deposit or etch on wafers. There are two main types of photoresist: positive photoresist, which melts when exposed to light, and negative photoresist, which melts when unexposed.
Positive PR
Positive PR offers superior resolution compared to negative PR. Furthermore, incorrect baking temperature or baking time settings can lead to the formation of scum, a residual PR residue, during the development process. Positive PR, however, does not readily generate scum. Due to its excellent step coverage and thermal stability, it is primarily used in modern micropatterning.
However, it has the disadvantage of having poor adhesion to the oxide film and being expensive compared to negative PR.
Negative PR
Negative PR can be thought of as the opposite of positive PR. It has less adhesion to oxide films and is relatively inexpensive, but suffers from poor resolution. Additionally, the PR may swell during the development phase.
| Positive PR | Negative PR |
|---|---|
| Remove areas exposed to UV rays Scum x, Steop coverage is good Good resolution Poor adhesion with oxide film Expensive | Remove the areas not exposed to UV rays Not suitable for micropatterning Good adhesion to oxide films Cheap |
| The bond between the PR components surrounding the illuminated area is weakened, forming a wide profile at the top. | The bond between the PR components around the illuminated area is strengthened, forming a narrow profile at the top. |
| |
There are differences in the profile depending on the type of PR, and we used the Negative type by using the Lift-off method.
UV Exposure
Exposure method
There are three main exposure methods for basic exposure process equipment using ultraviolet light.
- Contact aligner
- Proximity aligner
- Projection aligner
| Contact aligner | Proximity aligner | Projection aligner |
|---|---|---|
| | |
Contact aligner and proximity aligner, the two early exposure methods, are distinguished by the gap between the mask and the wafer. The contact aligner method offers the advantage of being less affected by light diffraction due to the close proximity of the mask and wafer. However, the mask is contaminated with photoresist, which shortens its lifespan.
This method is only used in the lab, and in our lab, we also used the contact aligner method for exposure. So, we had to wash the mask with acetone or ethanol every time...
Proximity aligners can protect the mask, but precise patterning is impossible due to light diffraction. To address this issue, current semiconductor mass production processes use projection aligners, which place appropriate optical systems, such as lenses, between the mask and wafer to protect the mask and prevent diffraction effects.
Exposure wavelength
The light used in the exposure process is ultraviolet light. Initially, mercury was used to produce g-line (436 nm) and i-line (365 nm) light. However, with the need for more precise patterning technology, shorter wavelength ultraviolet light is now used, utilizing KrF eximers (248 nm) and ArF eximers (193 nm). For foundries performing ultra-precision patterning below 10 nm, EUV (Extreme UltraViolet, 13.5 nm) light is also used.
In our lab, we used i-line (365 nm) for patterning, and the equipment specifications allowed us to pattern up to 1 μm.
Resolution
The most important factor in photolithography is resolution. Resolution refers to the minimum pattern size that can be distinguished during patterning, and can be expressed by the following formula:
The process coefficient can be increased through PR characteristic improvements, process optimization, and mask pattern improvements (such as OPC). As the formula shows, the shorter the UV wavelength, the lower the Res value, meaning that smaller patterns can be distinguished (improved precision).
Depth of Focus (DOF)
When using an optical system (lens), an error occurs in the vertical direction, which is called DOF (Depth of Focus).
The larger the NA (numerical aperture) (larger lens) and the shorter the wavelength used, the smaller the DOF, which means it becomes more sensitive to the thickness of the PR, the height of the substrate, and the flatness.
In summary, using a lens with a large NA and exposure to short-wavelength UV light can increase resolution, but it also reduces the DOF, making the image more sensitive to vertical errors. To address this, the substrate is flattened through a CMP process or the photoresist thickness is reduced.
References: Samsung Electronics
