Every chip manufactured today, from the processor in a smartphone to the memory in a data centre server, is defined by patterns etched into silicon with extraordinary precision. The material that makes this precision possible is the photoresist. Understanding what photoresists do, how they behave under light, and why their chemistry must evolve as transistors shrink is fundamental knowledge for anyone working in semiconductor process integration or materials engineering. This article builds that understanding systematically, starting with the material itself and working through to how resist selection shapes the performance of finished devices.
The concepts covered here progress from foundational to applied: what a photoresist is and what it contains, how it responds to exposure, the complete lithography sequence, why advanced nodes demand different chemistry, how performance is quantified, and finally how resist choice influences the device that comes out of the fab.
What is a photoresist and what is it made of?
A photoresist is a light-sensitive polymer film applied to a wafer surface that can be selectively removed after exposure to radiation, leaving behind a precise pattern that guides subsequent etch or implant steps. The word “resist” captures its function: the material resists chemical attack in areas where it is meant to remain, and surrenders in areas where it is not.
Most photoresists share a common formulation architecture built from four functional components:
- Polymer matrix: the structural backbone, typically an acrylic or cyclic olefin copolymer, that gives the film its mechanical and etch-resistance properties.
- Photo-active compound (PAC) or photoacid generator (PAG): the component that absorbs incoming photons and triggers a chemical change in the film.
- Solvent: carries the solid components in solution for spin-coating and evaporates during the post-apply bake.
- Additives: quenchers, surfactants and adhesion promoters that fine-tune contrast, defectivity and film uniformity.
For EUV photoresists, the conventional organic polymer architecture is increasingly replaced by inorganic or metal-containing formulations. Metal oxide resists, for example, absorb EUV photons far more efficiently than organic polymers because metals such as tin or hafnium have much higher extreme-ultraviolet absorption cross-sections. This difference in composition is not cosmetic; it has direct consequences for sensitivity, resolution and the process window available to the engineer.
How photoresists respond to light exposure
The exposure step converts a latent aerial image from the scanner’s optics into a chemical image inside the resist film. The mechanism depends on whether the resist is a positive-tone or negative-tone formulation, and this distinction is worth establishing clearly.
In a positive-tone resist, exposure makes the illuminated regions soluble in developer. The photoacid generator releases acid upon absorbing photons; during the post-exposure bake (PEB), that acid catalytically cleaves protecting groups from the polymer, converting hydrophobic chains into hydrophilic ones that dissolve readily in aqueous alkaline developer. The pattern that remains on the wafer is therefore the inverse of the bright regions in the mask.
In a negative-tone resist, exposure crosslinks or insolubilises the illuminated regions. The unexposed areas dissolve, leaving behind a pattern that mirrors the bright regions of the mask. Negative-tone developers, often organic solvents, are used with certain chemically amplified resists to invert tone without changing the underlying chemistry.
The amplification mechanism in chemically amplified resists (CARs) is worth understanding in its own right. A single absorbed photon generates one acid molecule, which then catalyses dozens or hundreds of deprotection reactions during the PEB. This chain reaction multiplies the chemical signal from a weak photon dose, making CARs far more sensitive than earlier single-photon-per-reaction systems. The trade-off is that acid diffusion during the bake can blur the chemical image, which is one reason PEB temperature and time are tightly controlled.
The full lithography process step by step
Lithography is not a single action but a sequence of tightly coupled steps. Each step conditions the wafer for the next, and a variation at any point propagates through to the final pattern.
- Surface preparation: the wafer is cleaned and primed, typically with hexamethyldisilazane (HMDS), to promote resist adhesion by making the oxide surface hydrophobic.
- Underlayer and anti-reflective coating deposition: one or more spin-on layers are applied beneath the resist. A silicon-based anti-reflective coating (SiARC) suppresses standing waves caused by back-reflected light and acts as a hard mask during etch. These underlayers are critical to pattern fidelity, particularly at tight pitches.
- Resist spin-coating: the resist solution is dispensed onto the spinning wafer; centrifugal force spreads it into a uniform film, typically 20 to 100 nm thick at advanced nodes.
- Post-apply bake (PAB): a low-temperature bake drives off residual solvent and densifies the film.
- Exposure: the scanner projects the mask pattern through its optical column onto the resist. At advanced nodes this uses deep ultraviolet (DUV) at 193 nm with immersion water, or extreme ultraviolet (EUV) at 13.5 nm.
- Post-exposure bake (PEB): heating drives the acid-catalysed deprotection reaction, converting the latent chemical image into a solubility contrast image.
- Development: aqueous or solvent developer dissolves the soluble regions, revealing the patterned resist profile.
- Inspection and metrology: CD-SEM and overlay tools verify that the printed pattern meets specifications before the wafer proceeds to etch.
The underlayer stack beneath the resist deserves particular attention. Without effective reflection control, standing waves in the resist film create periodic CD variations that no amount of exposure tuning can fully correct. This is why silicon-based anti-reflective coatings and underlayers are an integral part of the patterning stack rather than optional additions.
Why resist chemistry changes at advanced nodes
Building on the exposure mechanism described earlier, it becomes clear why the resist chemistry that worked at 65 nm nodes cannot simply be reused at 3 nm. Three physical constraints tighten simultaneously as pitch shrinks: resolution, line-edge roughness (LER), and sensitivity. These three parameters are coupled in a fundamental trade-off sometimes called the RLS triangle, and optimising all three at once is one of the central challenges of advanced lithography materials development.
At EUV wavelengths, organic polymer resists face an absorption problem. The 13.5 nm photons carry enough energy to ionise molecules directly, but the low absorption of conventional polymers means that a significant fraction of photons pass through the film without contributing to the chemical image. Stochastic effects, the random variation in where photons are absorbed and where secondary electrons travel, become a dominant source of LER and local CD variation at sub-20 nm half-pitch.
Inorganic and metal oxide resists address this directly. Their higher EUV absorption efficiency means more photons contribute usefully to the exposure event per unit dose, improving sensitivity. Their inorganic backbone also offers superior etch resistance compared to organic polymers, which matters because the resist film must survive long enough to transfer the pattern into the underlayer and substrate below. The development of next-generation EUV and high-NA EUV resist materials, including metal-containing formulations, is an active area of research that will define process capability at the most advanced logic and memory nodes through the remainder of this decade.
Key performance parameters and how they are measured
Evaluating a photoresist requires quantifying several interdependent properties. Understanding what each parameter means, and how it is measured, allows process engineers to make informed comparisons between candidate materials.
Resolution
Resolution describes the smallest half-pitch or feature size the resist can faithfully reproduce. It is measured by printing a series of decreasing pitches on a test wafer and examining the resulting profiles with a CD-SEM. Resolution is limited by both the optics of the scanner and the diffusion length of the reactive species within the resist film.
Line-edge roughness
LER quantifies the deviation of a feature edge from a perfectly straight line, measured in nanometres (3-sigma). High LER translates directly into variation in transistor gate length, which degrades device performance and yield. LER is inherently stochastic and tends to worsen as exposure dose decreases, creating tension with sensitivity targets.
Sensitivity
Sensitivity, or dose-to-clear (Eop), is the minimum exposure energy in millijoules per square centimetre needed to fully clear the resist in exposed areas. Higher sensitivity reduces scanner throughput requirements but, as noted above, often increases LER. For EUV processes, sensitivity is a throughput-critical parameter because EUV source power remains a limiting factor in high-volume manufacturing.
Contrast and process window
Contrast describes how sharply the resist transitions between exposed and unexposed states. A high-contrast resist produces nearly vertical sidewalls and tolerates a wider range of exposure and focus conditions, which is the process window. A narrow process window means more wafers fall out of specification when scanner conditions drift, directly affecting yield.
How photoresist selection shapes device performance
The parameters described above are not abstract quality metrics; they have direct consequences for the transistors, interconnects and memory cells that the lithography step is defining. Resist selection is therefore a materials decision with device-level implications.
Gate critical dimension (CD) uniformity across a wafer is one of the clearest examples. A resist with poor contrast or high LER introduces variation in the printed gate length. Since transistor switching speed and leakage current are both sensitive functions of gate length, CD variation translates into a distribution of electrical performance across the die. At advanced logic nodes, where gate lengths are measured in single-digit nanometres, even sub-nanometre CD variation is significant.
In NAND flash memory, the vertical channel hole diameter must be tightly controlled across billions of cells on a single die. Resist performance at the hole-patterning step directly determines the distribution of cell threshold voltages and, ultimately, the bit error rate of the finished device. Similarly, in DRAM, the capacitor contact pitch is defined lithographically, and resist-driven CD errors at that step affect storage capacitance and refresh performance.
The choice of underlayer materials works in concert with the resist. A well-matched SiARC suppresses reflection, tightens the CD distribution and improves the process window available to the resist above it. The full patterning stack, from the silicon underlayer through the anti-reflective coating to the resist itself, must be engineered as a system rather than as independent layers. Explore PiBond’s advanced materials and applications to see how silicon-based anti-reflective coatings, underlayers and spin-on hard masks are developed and supplied as part of an integrated patterning solution.
For engineers specifying materials at advanced nodes, the practical implication is that resist selection cannot be separated from the broader process integration context. The scanner wavelength, the underlayer stack, the etch chemistry downstream, and the device performance targets all constrain which resist formulations are viable. As nodes advance toward high-NA EUV and beyond, the materials requirements will continue to tighten, making the chemistry of the resist and its surrounding stack one of the most consequential decisions in the patterning process flow. To discuss specific material requirements or development opportunities, contact PiBond directly.
Related Articles
- What refractive index control means for high-density semiconductor optical systems
- How cleaning chemistry compatibility shapes process integration decisions
- Semiconductor dielectric materials supplier selection criteria for fab procurement teams
This content was generated with the help of AI and it may contain mistakes