Photoresist moulds for electroplating

1 · Deposit a conductive seed layer

Electroplating only deposits onto a conductive surface, so a non-conductive substrate — silicon, glass, or a wafer with an insulating oxide — first needs a thin seed layer (a plating base): usually a sputtered or evaporated metal stack, an adhesion metal such as titanium or chromium beneath the metal you intend to plate, for example a Ti/Au or Cr/Cu stack. The mould is patterned on top of this seed, and metal grows up from the exposed seed inside each open trench.

The seed must be continuous and clean everywhere plating will happen, and it doubles as the surface the resist has to adhere to — so an adhesion primer or a dehydration bake before coating earns the same care here as on any lithography (see adhesion, in the guidelines).

2 · Coat the thick resist mould

A plating mould has a harder job than an ordinary etch mask. It sits in a wet plating bath — often acidic or near-neutral, sometimes warm — for minutes to hours while metal grows, so it must resist that chemistry without swelling, softening or lifting. It has to adhere to the seed through the whole plate, hold steep, vertical sidewalls so the plated part has the width you drew, and clear cleanly to the seed at the trench floor with no footing — a residual wedge of resist at the base that would keep metal from nucleating against the seed.

Choosing a thick resist across families

Mould resists span several chemistries, and the fit depends on target thickness, aspect ratio and how you plan to remove the mould afterwards. Among thick positive DNQ-novolak resists, AZ 9260 is the founding thick-mould material — Conédéra et al. reported aspect ratios of 15–20 at 100 µm on a standard aligner — and SPR 220-7 reaches comparable thicknesses in multi-coat films used as electroplating moulds for MEMS metal parts (Kukharenka et al.). The chemically-amplified AZ 40XT and the thick positive AZ P4620 are further options in the same range.

For acrylic and epoxy negatives, the acrylic AZ 125 nXT has been demonstrated at 400, 800 and 1400 µm thick with a 20:1 aspect ratio for UV-LIGA electroplating (Staab et al.), and the epoxy KMPR reaches 180 µm moulds at an 18:1 aspect ratio (Lee et al.). Browse them all in the library filtered to electroplating / moulding.

One caveat drives the whole choice: SU-8 builds excellent thick, vertical moulds, but its fully cross-linked epoxy is hard to strip once plating is done — which is exactly why KMPR was developed as a *strippable* epoxy mould resist (Lee et al.) and why thick positive resists such as AZ 9260 and SPR 220-7 stay popular for moulds you need to remove intact. Weigh strippability before you commit, and see epoxy negative resists for the family trade-offs.

How thick to coat, at what spin speed, and whether a multi-coat build is needed to reach the target are all on each resist's recipe page — read the target off the resist's own spin curve rather than a remembered number.

Sources: Conédéra et al. (1999) — thick AZ 9260 moulds; Kukharenka et al. (2003) — SPR 220-7 electroplating moulds; Staab et al. (2011) — AZ 125 nXT microelectroplating; Lee et al. (2008) — strippable KMPR electroforming moulds

3 · Soft-bake the mould film

A thick mould film holds a lot of casting solvent, and driving it out evenly is harder than in a thin coat. Too short a soft bake leaves solvent trapped near the substrate, which bubbles or delaminates during exposure and plating; too aggressive a bake can round the sidewall profile or, in a chemically-amplified resist, disturb its sensitivity. Thick resists therefore use longer, sometimes ramped or multi-step bakes — always at the temperature and time the recipe specifies for that thickness.

4 · Expose the mould pattern

Exposure defines the trenches. On a maskless direct-write tool you write the mould pattern straight from your layout — no photomask to order — which makes iterating a mould geometry a same-day loop. Thick films need proportionally more dose than thin ones because the light has to reach the bottom of the film, and the useful exposure is bounded by depth-of-focus and by the sidewall verticality you need. A positive mould resist clears the exposed trenches on development; a negative resist such as KMPR or AZ 125 nXT cross-links the walls you keep and clears the unexposed trenches. Published dose windows are on each recipe page; the exposure section of the guidelines covers dose selection for thick films.

5 · Develop to open the plating channels

Development dissolves the trenches down to the seed. Two things matter more for a mould than for a shallow pattern: the develop has to run long enough to clear all the way to the seed with no residual footing at the trench floor, and it must do so without undercutting or rounding the sidewalls you are relying on to shape the plated metal. Thick films need correspondingly long, agitated or multi-bath develops — timings are per-recipe (see development, in the guidelines). Inspect the trench floors before plating; any thin scum the developer left on the seed is dealt with next.

6 · Descum and activate the seed

Even a clean develop can leave a monolayer of resist residue on the seed, and metal will not nucleate evenly through it. A brief descum — a short, low-power oxygen plasma — clears that residue from the trench floors. Because the same plasma lightly oxidises the exposed metal seed, plating is usually preceded by a mild activation (a short acid dip appropriate to the plating chemistry) so metal grows uniformly from a clean, conductive floor.

7 · Electroplate into the mould

With the mould open to a clean seed, the wafer goes into the plating bath and metal grows upward from the seed inside each trench, taking the negative shape of the resist. Copper, gold, nickel and their alloys are all common; AZ 125 nXT moulds, for instance, have been electroformed in acid copper at room temperature and near-neutral nickel at 50 °C (Staab et al.), and thick positive moulds have produced on-chip parts from 81 µm power micro-transformer windings (O'Donnell et al.) to millimetre-wave metal structures grown in KMPR moulds (Shin et al.). Plating is typically stopped at or just below the top of the mould; the resist confines the growth laterally, so the sidewall verticality set upstream is what the finished part inherits.

Sources: Staab et al. (2011) — AZ 125 nXT microelectroplating; O'Donnell et al. (2004) — 81 µm micro-transformer moulds; Shin et al. (2010) — ultrathick KMPR micrometallic structures

8 · Strip the mould

After plating, the resist has done its job and comes off, leaving the freestanding metal on the seed. How easily it strips is the payoff of the family choice made at coating. Thick positive resists (AZ 9260, SPR 220-7) generally clear in ordinary solvents or resist strippers; the acrylic AZ 125 nXT is designed to release even high-aspect metal, stripped in DMSO/acetone or by plasma (Staab et al.); the epoxy KMPR strips where SU-8 will not (Lee et al.). A fully cross-linked SU-8 mould, by contrast, often needs aggressive strippers, a plasma ash or a sacrificial-layer release — so plan the strip before you plate, not after.

Finally, the exposed seed between the plated features is removed by a brief blanket etch so the parts are electrically isolated. The strip & lift-off section of the guidelines covers stripping chemistry in more depth.

Sources: Staab et al. (2011) — AZ 125 nXT microelectroplating; Lee et al. (2008) — strippable KMPR electroforming moulds

Common questions

Where to next

1. Conédéra et al.. Potentialities of a new positive photoresist for the realization of thick moulds. Journal of Micromechanics and Microengineering (1999). doi:10.1088/0960-1317/9/2/317
2. O'Donnell et al.. Thin film micro-transformers for future power conversion. Nineteenth Annual IEEE Applied Power Electronics Conference and Exposition, 2004. APEC '04. (2004). doi:10.1109/APEC.2004.1295935
3. Staab et al.. Applications of Novel High-Aspect-Ratio Ultrathick UV Photoresist for Microelectroplating. Journal of Microelectromechanical Systems (2011). doi:10.1109/jmems.2011.2159098
4. Lee et al.. Fabrication of thick electroforming micro mould using a KMPR negative tone photoresist. Journal of Micromechanics and Microengineering (2008). doi:10.1088/0960-1317/18/5/055032
5. Shin et al.. UV Lithography and Molding Fabrication of Ultrathick Micrometallic Structures Using a KMPR Photoresist. Journal of Microelectromechanical Systems (2010). doi:10.1109/JMEMS.2010.2045880
6. Kukharenka et al.. Realization of electroplating molds with thick positive SPR 220-7 photoresist. Journal of Materials Science: Materials in Electronics (2003). doi:10.1023/A:1023923911921