Digifund
Digifund - Silicon Wafer Supplier VietnamDIGIFUND
Semiconductor knowledge

The Photolithography Process in Chip Manufacturing

Photolithography is the step that "prints" circuit patterns onto a silicon wafer — the heart of chip fabrication. This article breaks down each step, from photoresist coating to etching, and the role of each chemical.

1. What is photolithography?

Photolithography is a technique that uses light to transfer a circuit pattern from a photomask onto the surface of a silicon wafer coated with a light-sensitive layer. This step is repeated dozens of times to build up the transistor and interconnect layers of a chip.

2. The main photolithography steps

  1. Cleaning & surface prep: the wafer is cleaned and coated with an adhesion promoter (HMDS) so resist sticks well.
  2. Photoresist coating: resist is dispensed and spin-coated into a uniform thin film.
  3. Soft bake: evaporates solvent and stabilizes the resist film.
  4. Exposure: UV light is projected through the photomask; exposed areas change solubility.
  5. Development: a developer (TMAH) washes away the soluble resist, leaving the pattern.
  6. Etching: the pattern is transferred into the underlying layer using wet etchant or plasma (dry).
  7. Resist strip: remaining photoresist is removed with solvent (acetone) or plasma.

3. Positive vs. negative photoresist

With positive photoresist, the exposed area becomes soluble and is washed away — the pattern matches the mask. With negative photoresist, the exposed area hardens and stays — the pattern is the inverse of the mask. Positive resist offers higher resolution and dominates modern chip fabrication.

Key point: the resolution of photolithography sets the smallest transistor size — and therefore the density, performance and how advanced the chip can be.

4. Related chemicals & materials

A lithography cycle requires many high-purity supplies: the silicon wafer substrate, photoresist, TMAH developer, etchant (BOE), CMP slurry to planarize between layers, and cleaning solvents such as acetone. Digifund supplies this full range of materials and chemicals.

5. What limits resolution

The smallest feature a lithography system can print is governed by diffraction, described by the Rayleigh equation:

CD = k₁ · λ / NA

Here CD is the critical dimension, λ the exposure wavelength, NA the numerical aperture of the lens system, and k₁ a process-dependent factor. This single equation explains the whole trajectory of the industry:

  • Shrink λ. The industry moved from mercury g-line (436 nm) and i-line (365 nm) to KrF (248 nm) and ArF (193 nm) excimer lasers, and more recently EUV (13.5 nm).
  • Raise NA. A larger lens captures more diffraction orders. Immersion lithography replaces the air between lens and wafer with water, pushing NA above 1.
  • Lower k₁. This is the "process tricks" term: off-axis illumination, phase-shift masks, optical proximity correction (OPC) and multiple patterning.

An equally important second equation is depth of focus, DOF = k₂ · λ / NA². Raising NA for smaller features shrinks depth of focus quadratically — which is why wafer flatness and resist thickness uniformity become so demanding at advanced nodes.

6. Common laboratory failures and how to fix them

SymptomUsual causeFix
Resist lifting offSurface moisture or organic residueDehydration bake then HMDS priming
Striations after spin coatingUneven solvent evaporation, resist too viscousStatic dispense, higher spin speed, control room humidity
Edge beadSurface tension pulls resist to the rimEdge bead removal with solvent before exposure
Sloped rather than vertical sidewallsWrong dose or defocusRun a dose–focus matrix to find the process window
Standing waves on sidewallsLight reflected from the substrate interferes within the resistPost-exposure bake to diffuse, or an anti-reflective coating
Residual film at trench bottoms (scumming)Underexposure or incomplete developmentIncrease dose or development time, check developer age
The first thing to do with a new process: run a dose–focus matrix on Test-grade wafers. It consumes a few wafers but maps the real process window of your system, instead of guessing run by run on expensive Prime material.

Frequently asked questions

What is HMDS used for?

Hexamethyldisilazane reacts with silanol groups on silicon oxide surfaces, converting them from hydrophilic to hydrophobic. This greatly improves resist adhesion and prevents lifting during development. It is an all but mandatory step before spin coating resist.

Why is a post-exposure bake needed?

PEB serves two purposes. In conventional resists, heat diffuses the photoproducts and smooths out standing-wave ripples caused by substrate reflection. In chemically amplified resists, PEB drives the acid-catalysed reaction generated during exposure — without it, almost no pattern forms.

Should I use positive or negative resist?

Positive resist gives better resolution and sidewall profiles and is the default for most device work. Negative resist adheres better, resists chemicals more strongly and is usually cheaper, suiting coarse etch masks or thick layers. For lift-off processes, the undercut sidewall of negative resist is a distinct advantage.

Need materials & chemicals for lithography?

Digifund supplies silicon wafers, photoresist, developer, etchant, CMP slurry and semiconductor-grade solvents.

View semiconductor chemicals Request a quote