Photoresist Process
The photoresist process is the sacrificial imaging step that every other film-forming process in this cluster eventually depends on — coat, soft bake, expose, post-exposure bake, develop, hard bake, and later strip — and it is the one "process" row here that never survives to become part of the finished chip, unlike the permanent BARC, DARC, APF, and spin-on carbon films it sits on top of in a trilayer stack. Resist suppliers JSR, Tokyo Ohka Kogyo, Shin-Etsu, and Fujifilm sell this chemistry by the drum, and Tokyo Electron's Lithius coat-track systems spin it onto 100,000+ wafers a month at a 300mm fab, holding a film only 40-150nm thick (EUV) to single-nanometer thickness uniformity before anything is even exposed.
Every step in that sequence exists to control one variable: how precisely a line between "exposed" and "unexposed" resist can be placed, and chemically amplified resist (CAR) chemistry is what makes that line sharp enough to matter at modern dimensions. A single EUV photon carries roughly 92eV, enough energy to break bonds but nowhere near enough photons per unit area to directly expose every molecule in a 40nm-thick film at a usable dose — so instead, each absorbed photon triggers a photoacid generator (PAG) to release one acid molecule, and during the post-exposure bake that one acid molecule diffuses through the polymer matrix catalytically, deprotecting thousands of sites along the way before it's finally quenched. That amplification is the only reason resist exposure doses can stay in the tens of millijoules-per-square-centimeter range instead of requiring orders of magnitude more photons, but it comes with a direct tradeoff: the same acid diffusion length that amplifies sensitivity also blurs the boundary between exposed and unexposed regions, which is the dominant source of line-edge roughness in EUV resist today.
That dose-contrast curve on the right is also exactly how a fab specifies resist performance to JSR, TOK, or Shin-Etsu before ever committing it to a production line — the steepness of the transition, called the resist's contrast or gamma, predicts how much a dose error at the wafer translates into a critical-dimension error on the finished line. A resist with low contrast needs a large, well-behaved dose window to stay in spec, which is forgiving of scanner dose variation but blurs edges; a resist engineered for high contrast snaps from "fully retained" to "fully cleared" over a narrower dose range, tightening CD control at the cost of making the whole process less tolerant of any dose or focus error the exposure tool introduces upstream. EUV resists push this tradeoff hardest: at a typical 30-40mJ/cm² exposure dose, stochastic photon shot noise alone — not film chemistry imperfection — sets a hard floor on how sharp that curve's transition can get, which is why EUV resist line-edge roughness has become as much a photon-statistics problem as a chemistry problem.
The coat, bake, and strip steps bracketing exposure matter just as much as the chemistry, because photoresist is the only layer in this materials cluster designed to be removed completely rather than integrated into the final device. Soft bake at 90-130C has to drive off the PGMEA casting solvent without pre-reacting the PAG; post-exposure bake temperature and time are controlled to tenths of a degree and seconds because that's the step actually doing the acid-catalyzed chemistry that developer later reads out; and hard bake densifies the surviving resist pattern so it can withstand the ion bombardment of a subsequent dry etch without reflowing or cracking. Once that etch is done, the resist's entire job is finished — it gets removed by O2 plasma ashing or a wet strip, a one-way trip that distinguishes photoresist from every other row in this cluster, where TEOS, APF, BARC, and DARC films are chosen specifically because some or all of them survive into the finished stack.
Read photoresist as the temporary, disposable twin of the permanent process rows around it: identical attention to film thickness, bake temperature, and composition control, but aimed entirely at creating a precise, removable pattern rather than a permanent functional layer, which is exactly why its process window is judged by a dose-contrast curve instead of the resistivity, stress, or step-coverage numbers that define success for the films it sits on top of.