The Chemistry Nobody Talks About
Ask someone about chipmaking and they'll mention transistors, lithography, maybe EUV. Hardly anyone mentions photoresist — the light-sensitive polymer that makes the entire patterning process possible. Yet without advances in resist chemistry, every lithography improvement would be useless. You can have the most expensive scanner in the world, but if the resist can't resolve the pattern cleanly, the chip doesn't get made.
How Photoresist Works
A photoresist is a polymer film that changes its solubility when exposed to light. In a positive-tone resist (the dominant type in semiconductor manufacturing), exposed regions become soluble in developer solution and wash away. In a negative-tone resist, exposed regions crosslink and become insoluble — the unexposed parts wash away instead.
Chemically Amplified Resists
Modern resists for deep-UV (DUV) lithography at 193nm and 248nm are chemically amplified (CAR). When photons hit the resist, they generate photoacid molecules. During a post-exposure bake (PEB), these acid molecules catalytically deprotect many polymer chains — one photon can trigger hundreds of chemical reactions. This amplification makes the resist much more sensitive, which matters enormously for throughput.
The standard DUV resist stack includes three key components: a base polymer (typically a polyhydroxystyrene derivative for 248nm or an acrylate/methacrylate copolymer for 193nm), a photoacid generator (PAG), and a base quencher that limits acid diffusion to maintain resolution. The film thickness is usually 50-200nm, deposited by spin-coating from a solvent solution.
We covered a related topic in Persistent Memory: Intel Optane Legacy, CXL-Attached PM, and.
The EUV Resist Challenge
EUV lithography at 13.5nm brought a completely new set of problems for resist chemists. The photon energy is 92 eV — roughly 14 times higher than 193nm DUV. This changes the fundamental interaction between light and resist.
Stochastic Effects
At EUV doses, each pixel of the pattern receives only a handful of photons. For a 20nm feature at a typical dose of 30 mJ/cm², you're talking about maybe 30-50 photons per pixel area. Statistical variation in how those photons land creates roughness at the feature edges — this is the infamous line-edge roughness (LER) problem that has plagued EUV since its introduction.
There's a three-way tradeoff that resist engineers call the "RLS triangle": Resolution, LER, and Sensitivity. Improve any two and the third gets worse. A more sensitive resist needs fewer photons (better throughput) but has worse stochastic noise (worse LER). A less sensitive resist gets smoother edges but tanks your scanner productivity.
See also: Backside Power Delivery: Why Routing Power Under the Transis.
Metal Oxide Resists
One of the most significant developments in EUV resist chemistry has been the move toward metal oxide-based resists. Companies like Inpria (acquired by JSR in 2021) developed tin-oxide based resists that have much higher EUV absorption than organic CARs. Higher absorption means more of the incoming photons actually do useful chemistry, improving the sensitivity-roughness tradeoff.
These metal-containing resists also offer better etch resistance than organic resists, which is important because EUV resists need to be very thin (30-40nm) to maintain resolution — and a thin resist film has less material to protect the underlying layers during etching.
The Japanese Supplier Dominance
The photoresist market is overwhelmingly Japanese. Tokyo Ohka Kogyo (TOK), JSR, Shin-Etsu Chemical, Fujifilm, and Sumitomo Chemical supply roughly 90% of all advanced photoresists globally. This concentration became front-page news in 2019 when Japan restricted exports of hydrogen fluoride, photoresists, and fluorinated polyimide to South Korea amid a diplomatic dispute. Samsung and SK Hynix were suddenly scrambling for alternative resist supplies.
We covered a related topic in Interconnect Scaling: Why On-Chip Wiring Is the New Bottlene.
The reason for Japanese dominance isn't mysterious — it's decades of accumulated process knowledge and customer relationships with the major fabs. Resist development requires extremely close collaboration with the foundry, tuning formulations to specific scanner configurations and process conditions. Switching resist suppliers isn't like switching commodity chemicals; it requires months of qualification and can require revalidating the entire lithography process.
What's Next
For High-NA EUV, the resist challenge intensifies. The higher resolution means even thinner films, tighter LER requirements, and the need for new development and rinse chemistries. Dry development processes — using plasma instead of liquid developer — are being explored as a way to pattern thin resists without mechanical damage from liquid flow.
Directed self-assembly (DSA), where block copolymers spontaneously form regular patterns guided by a lithographic template, remains a research topic. It could potentially improve resolution and roughness beyond what conventional resists can achieve, but it's still years from production readiness. For now, the semiconductor industry's most advanced technology remains critically dependent on advances in polymer chemistry — a fact that still surprises most people outside the field.