Photoresist processing guidelines.

Start from a clean, dry surface

Substrate cleaning removes particles, organic films and native contamination before any resist touches the wafer, because whatever is on the surface at coat time ends up under the resist and prints as a defect. Particulate contamination seeds comets and pinholes during spin; an organic or ionic film locally changes wetting and adhesion, so the resist edge lifts or the pattern scums where the surface was dirty.

The classic wet sequence is the two-step RCA clean — an ammonia/peroxide step (SC-1) that lifts particles and organics, and a hydrochloric/peroxide step (SC-2) that strips metallic ions — often preceded by a sulphuric-acid/peroxide (piranha) strip of heavy organics. The right sequence depends on the substrate and the contamination, and aggressive cleans are not free: they grow or etch oxide and must suit the material. A final dehydration bake then drives adsorbed water off the surface, since a hydrated surface is the single biggest adhesion problem the next two steps exist to solve.

Sources: Kern (1990); MicroChemicals — Cleaning & adhesion; Cornell CNF — photolithography manual

Why HMDS priming helps

Adhesion promotion makes the resist stick to the substrate through development and every step that follows, and on silicon it almost always means priming with hexamethyldisilazane (HMDS). A freshly cleaned oxide or silicon surface is hydrophilic — covered in hydroxyl (silanol) groups and adsorbed water — and an aqueous-base developer will happily creep along that hydrophilic interface and undercut fine features.

HMDS, applied as a vapour prime, reacts with those surface hydroxyl groups and replaces them with a hydrophobic methyl-terminated layer, so developer no longer wets under the resist edge. Vapour priming on a hot surface is more uniform and reliable than a spun-on prime, which is why production tracks integrate an HMDS oven. Priming is standard for DNQ-novolak resists on oxide; it is usually unnecessary for epoxy resists such as SU-8, which adhere well on their own. What’s left after priming should be a barely-visible monolayer — over-priming a surface can itself hurt adhesion.

Sources: MicroChemicals — Cleaning & adhesion; Mack, Fundamental Principles (2007)

Spin coating, and when to laminate instead

Coating lays down a resist film of controlled, uniform thickness. Spin coating is the dominant method: resist is dispensed onto the wafer, which is then spun so centrifugal force spreads it to a thin film while solvent evaporates. Final thickness falls as spin speed rises — roughly with the inverse square root of spin speed — and also depends on the resist’s solids content and viscosity, so a thicker film comes from a more viscous resist or a lower speed. The curve of thickness against speed for a given resist is its spin curve; read a target thickness from an actually-plotted point rather than between two. The per-resist spin curves live on the recipe pages.

Spin coating always builds a thicker rim at the wafer’s edge — the edge bead — which is removed immediately after coat so it can’t flake and contaminate later steps.

Dry-film lamination

For very thick coats, coarse features, or topography a spun film would pool into, the resist can instead be applied as a preformed solid sheet — hot-roll laminated onto the substrate under heat and pressure. Lamination gives a thickness set by the sheet rather than by spin dynamics, so it is largely independent of substrate size, shape and topography, which makes it common for thick and panel-scale (PCB-style) processes — at the cost of the finest resolution a spun film reaches.

Sources: MicroChemicals — Spin-coating; Zhang et al., dry-film review (2025); Mack, Fundamental Principles (2007)

Drive off the casting solvent

The soft bake (or prebake), performed right after coating, drives the casting solvent out of the film so it is dry, dimensionally stable and adherent before exposure. It is a balance, not a maximum: underbaking leaves residual solvent that causes scumming, poor adhesion and unstable development, while overbaking can thermally decompose the photoactive compound of a DNQ resist and reduce its sensitivity. Because the right amount of bake scales with film thickness, every recipe specifies both a temperature and a time — see the per-resist soft-bake values.

Sources: MicroChemicals — Softbake; Mack, Fundamental Principles (2007)

Let a DNQ film reabsorb water

Rehydration is a deliberate wait, after soft bake and before exposure, that lets a DNQ-novolak film reabsorb some atmospheric moisture. The exposure reaction that makes DNQ soluble consumes water, so a film baked bone-dry can show reduced apparent sensitivity until ambient humidity restores it. This is why some datasheets specify a minimum delay — or a controlled humidity — between soft bake and exposure, and it matters most for thick films, which lose the most water during bake. Chemically amplified resists have the opposite sensitivity: they prefer a short, controlled post-coat delay to limit contamination uptake, so always follow the resist’s own datasheet, linked from each recipe.

Sources: MicroChemicals — Rehydration; Dammel, DNQ-based Resists (1993)

Deliver the right dose

Exposure delivers UV light to the resist in the intended pattern, changing the solubility of the illuminated regions. The governing quantity is dose — energy per unit area — and each resist has a characteristic dose-to-clear at which a positive film first clears to the substrate, from which a working dose is set. Too little dose leaves scum and footing at the resist base; too much widens or rounds features. On a maskless tool the pattern is written directly rather than projected through a photomask, and dose is set in software per layer. Working doses are resist-, wavelength- and thickness-specific and live on the recipe pages.

Inside the film, light reflected back up from the substrate interferes with the incoming beam and forms vertical intensity ripples — standing waves — which print as fine horizontal ridges on an otherwise smooth sidewall. Their severity depends on substrate reflectivity and film thickness, and a bottom anti-reflective coating or a well-chosen thickness reduces them; the post-exposure bake then smooths what remains.

Sources: MicroChemicals — Exposure; Mack, Fundamental Principles (2007)

Average the standing waves — or drive the reaction

The post-exposure bake (PEB) is a bake between exposure and development, and it does one of two distinct jobs depending on the resist chemistry. In a conventional DNQ-novolak resist, the PEB thermally diffuses the exposed photoproduct just enough to average out the standing-wave ripple left by exposure, smoothing the sidewall without significantly moving the pattern edge.

In a chemically amplified resist, the PEB is where the actual imaging chemistry happens: the acid generated by exposure catalytically deprotects (or crosslinks) many sites during the bake, so PEB temperature and time are among the most dose-sensitive parameters in the whole process and a delay before PEB invites amine contamination and T-topping. Because the PEB does such different work in the two families, its value is strictly resist-specific — see the recipe pages.

Sources: MicroChemicals — Post-exposure bake; Mack, Field Guide to Optical Lithography (2006)

Dissolve the soluble regions

Development dissolves away the soluble regions of the exposed film — the exposed area for a positive resist, the unexposed area for a negative one — leaving the pattern behind. Most conventional resists develop in an aqueous, metal-ion-free base, almost always TMAH (tetramethylammonium hydroxide), whose strength is set either as a normality (the common ready-to-use grade is 0.26 N) or as a dilution of a concentrate. Development rate, contrast and dark loss all depend strongly on that concentration and on temperature, so a recipe specifies the developer, its strength, the method and the time. Note that epoxy (SU-8-type) resists are the exception: they develop in an organic solvent, not an aqueous base.

How the developer is applied matters too. Puddle development dispenses a static pool onto a near-stationary wafer for a fixed time before a spin rinse and gives repeatable single-wafer CD control; immersion (dip) development submerges wafers in an agitated bath and favours batch throughput. Under-development leaves scum; over-development thins and rounds features and erodes the dark field. The per-resist developer and time are on the recipe pages.

Sources: MicroChemicals — Development; Mack, Fundamental Principles (2007)

Toughen the film for what comes next

The hard bake is a higher-temperature bake after development that further dries and cross-links the finished resist so it survives the etch, implant or electroplating step that follows. It raises the film’s thermal and chemical resistance and improves adhesion, but it is a trade-off: pushed too far it rounds sharp corners and can make some resists reflow, so the temperature is chosen against how much profile fidelity the following step needs. A resist meant to be stripped afterward is hard-baked gently, if at all, because an aggressive hard bake makes later removal much harder. Recommended hard-bake conditions are resist-specific — see the recipe pages.

Sources: MicroChemicals — Hardbake & reflow; Mack, Fundamental Principles (2007)

Two ways the resist leaves

Once the resist has done its job it is removed — either stripped, or used for lift-off. Stripping dissolves the resist off a substrate after an etch or implant, in a solvent or a dedicated resist remover (or, for stubborn residues, an oxygen plasma ash). A resist that was hard-baked hard, cross-linked, or implant-crusted is deliberately the hardest to strip, so the strip chemistry is matched to the resist’s history.

Lift-off

Lift-off inverts the usual order: resist is patterned first, a film (commonly a metal) is deposited over the whole wafer, and then the resist is dissolved — carrying away the film that sat on top of it and leaving film only where the resist had an opening. For this to work cleanly the deposited film must be discontinuous at the resist edge, which requires a negative or undercut (re-entrant) sidewall — wider at the base than the top — so the solvent can reach in. That undercut comes from an image-reversal resist, a dedicated lift-off resist, or a bilayer underlayer such as LOR or PMGI beneath an ordinary positive resist.

Recipes suited to lift-off are filterable in the library.

Sources: MicroChemicals — Lift-off; MicroChemicals — Photoresist removal

Reading common defects

Most process faults announce themselves as a recognisable defect, and each one points back at a step above. The glossary defines each term in full; the short table below is a field index, and the schematics show what four of them look like.

• Striations — faint radial thickness ripples from uneven solvent evaporation during spin coating.
• Comets — streaks trailing from a particle on the wafer or chuck during spin.
• Edge bead — a thickened rim of resist at the wafer edge; removed before it flakes and contaminates later steps.
• Scumming — residual resist left in a field that should have cleared — usually underexposure, an underbaked film, or spent developer.
• T-topping — a feature wider at the top than the body, most often from airborne amine neutralising photoacid in a chemically amplified resist.
• Footing — a flare of unwanted resist at a feature's base, from substrate reflectivity, under-dose at the interface, or interface chemistry.

Sources: MicroChemicals — Trouble-shooter; Mack, Field Guide to Optical Lithography (2006)

1. W. Kern. The Evolution of Silicon Wafer Cleaning Technology. Journal of The Electrochemical Society (1990). doi:10.1149/1.2086825
2. C. A. Mack. Fundamental Principles of Optical Lithography. Wiley (2007). doi:10.1002/9780470723876
3. C. A. Mack. Field Guide to Optical Lithography. SPIE Press (2006). doi:10.1117/3.665802
4. R. Dammel. Diazonaphthoquinone-based Resists. SPIE Press (1993). doi:10.1117/3.2265072
5. M. Zhang, F. Meng, X. Li, W. Zeng. Applications of Dry Film Photoresist in Micromachining: A Review. Micromachines (2025). doi:10.3390/mi16111258
6. MicroChemicals GmbH. Substrate Preparation: Cleaning and Adhesion Promotion (application note). https://www.microchemicals.com/dokumente/application_notes/substrate_cleaning_adhesion_photoresist.pdf
7. Cornell NanoScale Science & Technology Facility (CNF). Photolithography Techniques Manual. https://www.cnfusers.cornell.edu/sites/default/files/Area-Resources/PhotolithTechniquesManual20.pdf
8. MicroChemicals GmbH. Spin-coating of Photoresists (application note). https://www.microchemicals.com/dokumente/application_notes/spin_coating_photoresist.pdf
9. MicroChemicals GmbH. Softbake of Photoresists (application note). https://www.microchemicals.com/dokumente/application_notes/softbake_photoresist.pdf
10. MicroChemicals GmbH. Rehydration of Photoresists (application note). https://www.microchemicals.com/dokumente/application_notes/photoresist_rehydration.pdf
11. MicroChemicals GmbH. Exposure of Photoresists (application note). https://www.microchemicals.com/dokumente/application_notes/exposure_photoresist.pdf
12. MicroChemicals GmbH. Post Exposure Bake of Photoresists (application note). https://www.microchemicals.com/dokumente/application_notes/photoresist_post_exposure_bake_peb.pdf
13. MicroChemicals GmbH. Development of Photoresists (application note). https://www.microchemicals.com/dokumente/application_notes/development_photoresist.pdf
14. MicroChemicals GmbH. Hardbake, Reflow and DUV Hardening of Photoresists (application note). https://www.microchemicals.com/technical_information/photoresist_hardbake_reflow_uv_hardening.pdf
15. MicroChemicals GmbH. Lift-off Processes with Photoresists (application note). https://www.microchemicals.com/dokumente/application_notes/lift_off_photoresist.pdf
16. MicroChemicals GmbH. Photoresist Removal (application note). https://www.microchemicals.com/dokumente/application_notes/photoresist_removal.pdf
17. MicroChemicals GmbH. Photolithography Trouble-Shooter (application note). https://www.microchemicals.com/dokumente/application_notes/lithography_trouble_shooting.pdf

General photolithography reference material, not a specification of any particular NANYTE BEAM configuration, and not a substitute for a resist’s own datasheet. Datasheet values are starting points; optimal parameters depend on your substrate, equipment and environment. Product names and trademarks belong to their respective owners; NANYTE is not affiliated with the manufacturers mentioned. Stuck on a specific process? Talk to an engineer.