What it is
Chemically amplified resists (CAR) are photoresists in which exposure generates a small amount of photoacid that, during the post-exposure bake, catalytically drives many deprotection or crosslinking reactions per acid molecule — amplifying a modest exposure dose into a fully developed latent image.
How it patterns
- Substrate
- Resist
- Exposed
A photoacid generator, not a direct photoreaction
A chemically amplified resist carries a photoacid generator (PAG) rather than relying on light to react the resin directly. Exposure converts the PAG into a strong acid at the exposed sites, but at this point very little chemistry has actually happened — the acid is only a catalyst waiting to act.
Acid-catalyzed amplification during the post-exposure bake
The post-exposure bake (PEB) is where the amplification happens: heat mobilizes the photoacid and lets it catalyze many reactions before it is neutralized or diffuses out, so one photon-generated acid molecule can drive far more chemical change than it could alone. In a positive-tone CAR the acid deprotects an acid-labile polymer, converting the exposed resin to a form soluble in an aqueous developer. In a negative-tone CAR — such as AZ nLOF or AR-N — the same acid-catalyzed step instead crosslinks the exposed resin, making it insoluble. Either way, one exposure event is amplified into many downstream reactions, which is why a chemically amplified resist images at modest exposure doses.
Why bake time, temperature and delay all matter
Because the whole image is built during the PEB rather than at exposure, the bake's time and temperature are the parameters that most determine the result — too little PEB under-develops the amplification, too much lets the acid diffuse and blur the pattern edge. The same sensitivity to time works against the process between exposure and PEB: acid can diffuse or be quenched by airborne contaminants during that post-exposure delay, so CAR processes are typically run with a bounded, datasheet-specified delay rather than an open-ended queue time.
Not every dark-field resist is a CAR
AZ 125nXT, though sometimes grouped here, is characterized by its manufacturer as a photopolymer distinct from typical chemically amplified resists — see its recipe.
What each step does
- 01 Spin coat
Coated like any other resist, but amplification magnifies small differences. The dose and bake that work are tuned at one thickness, and a thicker or thinner film shifts both.
Adjusting it: Final speed sets thickness, read off the spin curve at the datasheet's ramp. Hold the dispense, ramp and spin time constant from run to run, so any change you see comes from the step you actually adjusted.
- 02 Soft bake
Drives out the casting solvent. In a chemically amplified resist it also sets how easily acid can move later, because leftover solvent loosens the film and lets acid spread further during the post-exposure bake.
Adjusting it: An under-baked film blurs edges, because the acid travels too far. Stay at the datasheet temperature, because some formulations start to react with heat alone if pushed hotter.
- 03 Expose
Light converts the photoacid generator into acid at the exposed sites, and the resin itself hardly changes yet. The pattern exists only as acid, and at the surface airborne bases such as amines can neutralise it.
Adjusting it: Dose sets how much acid forms. Dose and post-exposure bake trade against each other, since a hotter or longer bake needs less dose. A dose is only valid with the bake it was tuned with.
- 04 Delay before PEB
Between exposure and bake, the acid sits in the film waiting for heat. Amines in the cleanroom air can neutralise it at the surface during this time.
Adjusting it: Keep the delay short and the same every run, inside the datasheet limit. On a positive resist, a long or variable delay leaves an insoluble skin across the top of lines, called T-topping.
- 05 Post-exposure bake
The film heats and the acid can move. In a positive resist, each acid cuts protecting groups off the polymer so it becomes soluble in base; in a negative resist such as AZ nLOF, it crosslinks the resin instead. One acid drives many reactions.
Adjusting it: This is the strongest control in the process. Too little under-amplifies, leaving residue in positive resist and weak features in negative. Too much lets acid diffuse past the edge, blurring it and shifting the linewidth. How evenly the hotplate heats shows up directly in linewidth uniformity.
- 06 Develop
Aqueous TMAH removes the deprotected resin (positive) or the resin that never crosslinked (negative).
Adjusting it: Residue after a normal develop usually points back at the dose, the PEB or the delay rather than the develop time, because the solubility change was fixed in the bake.
- 07 Hard bake
Optional. Toughens a positive etch or plating mask.
Adjusting it: Skip it on a negative resist used for lift-off, because the crosslinked film only gets harder to strip. Baking a positive resist past its softening point rounds the profile.
When it goes wrong
Why won't the resist clear down to the substrate?
Cause: HMDS applied as a liquid releases ammonia into the film during the soft bake, and it modifies the resin next to the substrate so it won't develop through.
Fix: Apply HMDS only from the vapour phase onto a heated substrate, and never spin it on in a tool that also coats resist.
Source: MicroChemicals, Lithography trouble-shooter, HMDS applied correctly; contamination by other substances?
Why do thick AZ 12XT or 40XT coats bubble during the soft bake?
Cause: Put straight onto a hot plate, a thick film releases solvent faster than it can escape, and the vapour is trapped as bubbles or voids.
Fix: Ramp the soft bake: use proximity baking or step the temperature up, rather than a single contact bake.
Sources: Merck, AZ 12XT-20PL datasheet; Merck, AZ 40XT-11D datasheet, Soft bake
Why does a thin coat print differently if it waits before exposure?
Cause: A thin chemically amplified film holds little acid for its surface area, so amines in the air poison the surface while it waits between soft bake and exposure, and the profile changes.
Fix: Keep the time from soft bake to exposure short, especially for thin coats, and keep coated wafers away from amine sources such as NMP and amine-based strippers.
Source: Merck, AZ 12XT-20PL datasheet
Why does the linewidth shift after a small change in PEB temperature?
Cause: The acid-driven reaction rate follows the bake temperature, so the linewidth moves almost linearly with it. That is the chemistry, not a fault.
Fix: Control PEB temperature and hotplate uniformity as tightly as dose, and re-qualify the dose whenever the setpoint or the hotplate changes.
Source: Merck, AZ nLOF 2000 series datasheet, Post expose bake
Why do fine negative-resist lines thin or lift in the developer?
Cause: Too cool or short a PEB leaves the exposed resist lightly crosslinked, so it is still partly soluble, and the smallest features go first.
Fix: Raise the PEB into the recommended crosslinking range and check it on small test structures, rather than shortening the develop to compensate.
Source: MicroChemicals, Lithography trouble-shooter, Image reversal or negative tone resists?
Why did development slow down after I raised the PEB?
Cause: Too hot or too long a PEB crosslinks even the unexposed resist thermally, which slows development and can leave scum where the resist should clear.
Fix: Keep the PEB inside the datasheet window. If development is slow, adjust the dose or the developer rather than pushing the bake hotter.
Source: MicroChemicals, Lithography trouble-shooter, Negative or image reversal resists in image reversal mode?
Why won't AR-N 4340 develop after a hotter crosslinking bake?
Cause: Above a temperature the datasheet states, AR-N 4340 crosslinks so far that it no longer develops at all, whatever the dose.
Fix: Stay inside the manufacturer's optimum bake range, and pair the dose with the bake temperature you actually use.
How do I get more or less undercut from AZ nLOF?
Cause: The undercut of a crosslinking negative lift-off resist is set mainly by exposure dose and PEB temperature together.
Fix: Tune dose and PEB as a pair, and re-qualify them for each metal thickness you plan to lift.
Source: MicroChemicals, Lithography trouble-shooter, Undesired resist profile: undercut required?
Which developer avoids scum with AZ nLOF?
Cause: Resist that partly crosslinked where it shouldn't have leaves scum that not every TMAH developer removes.
Fix: Use the metal-ion-free TMAH developers recommended for these resists, and where partial crosslinking is a risk, the variant formulated with a scum remover.
Source: MicroChemicals, Lithography trouble-shooter, Compatible developer?
Why did dose-to-size drift when nothing else changed?
Cause: An open developer bath absorbs carbon dioxide from the air and loses strength, and dissolved resist exhausts it as wafers go through.
Fix: Track bath age and throughput and replace the developer on schedule, instead of correcting with dose.
Source: MicroChemicals, Lithography trouble-shooter, Expired developer?
Why does the resist mask fail partway through a strong wet etch?
Cause: Strongly oxidising etchants, such as those containing nitric acid, chemically attack novolak-based resist.
Fix: Hard bake after development to improve chemical stability, and check the resist's compatibility with the etch before committing wafers.
Source: MicroChemicals, Lithography trouble-shooter, Wet etching: resist attack in HNO3-containing etchants
Why are there metal fences along the edges after lift-off?
Cause: Sputtering coated the resist sidewalls because the profile wasn't undercut enough to shadow them, so the remover could only break in at random points.
Fix: Evaporate rather than sputter where possible. For sputtered metal, tune the resist for a pronounced undercut and keep the film thin.
Source: MicroChemicals, Lift-off processes with photoresist, Fences after lift-off
Why is AZ nLOF so hard to strip after deposition?
Cause: A hot PEB, or heating during deposition, pushed crosslinking so far that ordinary removers no longer dissolve it.
Fix: Keep the PEB inside its window and limit substrate heating during deposition. For resist that is already over-crosslinked, use a stripper rated for crosslinked negative resists.
Sources: MicroChemicals, Lift-off processes with photoresist, Negative resists; Merck, AZ nLOF 2000 series datasheet, Stripping
Where it fits
- For high-throughput exposure steps, chemically amplified resists need comparatively little dose, since one photoacid drives many downstream reactions rather than one.
- For lift-off patterning, negative-tone CAR resists such as AZ nLOF are formulated to hold the re-entrant or vertical sidewall profile a clean metal lift-off needs.
- For processes already running a controlled hotplate PEB step, the amplified chemistry rewards that existing bake-temperature discipline with a wide usable process window.
- For loosely controlled bake stations, CAR chemistry is a poor fit: because the image is built catalytically during the PEB, small swings in bake time or temperature shift linewidth measurably, so tight PEB control is essential.
- For long queue times between exposure and bake, the datasheet's specified post-exposure delay should be respected — extended delay lets the photoacid diffuse or become quenched, degrading the pattern.
- For a first read on tone, check the datasheet rather than assuming: CAR chemistry underlies both positive-tone (deprotection) and negative-tone (crosslinking) products.
Datasheet-cited recipes
Every value on these pages is cited to the manufacturer’s datasheet — spin curve, bake schedule, exposure dose and developer.
Further reading
- Hiroshi Ito. Chemical Amplification Resists for Microlithography. Advances in Polymer Science (2005). doi:10.1007/b97574
- H. Ito. Chemical amplification resists: History and development within IBM. IBM Journal of Research and Development (1997). doi:10.1147/rd.411.0069
- Hiroshi Ito. Chemical Amplification Resists: Laboratory Curiosity to Paradigm. Journal of Photopolymer Science and Technology (2007). doi:10.2494/photopolymer.20.319
Common questions
Why is the post-exposure bake so critical for a chemically amplified resist?
The PEB is where the actual amplification happens: heat lets the photoacid generated at exposure diffuse and catalyze many deprotection or crosslinking reactions before it is neutralized, so the bake's time and temperature — not just the exposure dose — set the final linewidth and profile.
What is chemical amplification?
It is a catalytic multiplication of a small photochemical event: exposure converts a photoacid generator (PAG) into a single acid molecule, and during the post-exposure bake that one acid molecule drives many downstream reactions rather than reacting once, so a modest exposure dose is enough to fully develop the latent image.
Are chemically amplified resists positive or negative tone?
Both exist. The same PAG-plus-PEB amplification mechanism underlies positive-tone CAR (the acid deprotects the exposed resin, making it developer-soluble) and negative-tone CAR such as AZ nLOF or AR-N (the acid instead crosslinks the exposed resin, making it insoluble) — check the specific resist's datasheet for its tone.
Expose it at 365 and 405 nm
NANYTE BEAM is a desktop maskless lithography system with software-selectable dual-wavelength exposure and 16-bit grayscale — no photomask, no mask cost, same-day iteration.
General photolithography reference material, not a specification of any particular NANYTE BEAM configuration. Product names and trademarks belong to their respective owners; NANYTE is not affiliated with the manufacturers mentioned.