Photoresist Thickness for Coated Chrome Plates

Choose the thinnest coating that survives handling

Choose the thinnest coating that still covers the substrate and survives your handling, then check the aspect ratio it creates at your smallest feature. The number that decides whether your process is comfortable is t / w — coating thickness divided by the smallest feature or space — not the thickness on its own.

Most coating questions arrive as a number: "can you do 2 µm?" The useful answer needs a second number attached to it, because 2 µm of resist is a comfortable coating for a 3 µm line and a difficult one for a 0.8 µm line. The same film is either generous or marginal depending on the artwork it has to define.

This guide is written for two readers: the process engineer who will expose the plate in-house, and the engineer or buyer specifying a finished mask, who needs to understand why a feature request drives a coating decision inside someone else's process. Both are asking the same question about the same physics.

Three numbers decide the coating

Everything below follows from three quantities:

• w — the smallest line or space in the artwork, in micrometres. This is the number that constrains the coating most, and it is often smaller than the feature the customer talks about first.
• t — the coating thickness on the plate.
• α — the sidewall angle of the developed resist, measured from vertical.

On a chrome-on-glass part, the film being patterned is thin and the etch is short. The resist does not need depth to survive the etch; it needs to define an edge cleanly, and then to disappear. That single fact reorders the usual intuition: more resist is not more protection, it is more distance for the light to travel and more wall to develop.

Read the 3 µm row against the published line-width tolerance. At 3 µm of coating and a 5° wall, the slope alone accounts for 0.52 µm of width change — larger than the whole ±0.2 µm window that the finished chrome line has to fit into. A wall angle of 5° is not a defect; it is an ordinary result of a process that has not been tuned for the feature. The point is that it becomes visible only when the coating is thick relative to the feature.

The ratio, not the thickness, sets the difficulty

Divide the coating by the smallest feature and the practical picture appears. Coatings near or below the feature size behave well: the developer reaches the bottom of the pattern quickly, the walls stay steep, and the dose has a wide margin. As the ratio grows, each of those becomes a tuning problem.

• Below 1:1 the resist is thinner than the feature. Resolution is limited by the exposure tool and the resist chemistry rather than by the coating.
• 1:1 to 3:1 is the usual working range for contact, proximity and direct-write processes on chrome blanks. A well-run thin-resist process sits here and holds sub-micrometre features.
• 3:1 to 6:1 is reachable, but the dose, bake and development have to be chosen for it. Expect to iterate on dose and development time rather than to inherit a recipe.
• Above 6:1 is thick-resist territory. Resist manufacturers report aspect ratios of roughly 6:1 to 10:1 for modern thick grades under optimised conditions — optimised being the operative word, and not something to assume on a first order.

Where your thickness sits in the published range

A 0.4–20 µm coating range is not one process with a dial. It spans families of resist with different viscosities, different photoactive content and different developer behaviour, and the useful part of each family is a narrower band than the range suggests.

Two practical consequences follow. First, "0.4 µm" and "20 µm" are not interchangeable options on the same job — they are different materials with different process windows, and a request that moves between them is a different process, not a different set point. Second, the thickness you ask for should be a value the chosen grade reaches comfortably, because a viscosity pushed to the edge of its range is where coating uniformity and edge bead problems appear.

A useful side note on spin coating: film thickness falls approximately with the square root of spin speed, so a fourfold change in speed gives roughly half the thickness. That leaves less range than most people expect from one resist, which is why resist grades exist at all. Industry reference, not a TL OPTIK specification.

What the coating changes after it is applied

Thickness is not a single decision with a single consequence. It propagates through the rest of the process, and each step has to absorb part of it:

• soft bake;
• exposure dose;
• development time;
• sidewall angle;
• width at the base of the resist;
• the chrome edge after etch.

The two that surprise people most are dose and development. A thicker film absorbs more of the exposure light before it reaches the bottom, so the dose that cleared a thin coating leaves resist behind in a thick one; and the developer has to work through the full depth, which lengthens the step and costs contrast at the same time. Resist manufacturers say the same thing from the other direction: their very thick grades exist precisely because pushing a thin resist to a thick film demands high doses, produces a poor wall profile and can generate nitrogen bubbles during exposure.

For chrome work there is a further consequence worth naming. Because the resist wall is transferred into the chromium during etch, the wall angle you develop is the profile the pattern edge inherits. A coating that is generous for the feature buys margin in coverage and handling; it costs profile.

A five-step way to choose

1. Fix w first.Take the smallest line or space in the artwork, whichever is smaller. If the drawing contains both a 2 µm line and a 2 µm gap, w is 2 µm and the ratio applies to both.
2. Set a ceiling on t/w.Work at or below 3:1 unless there is a reason to go higher and time to tune for it. A 1:1 target is a good default for a first order.
3. Translate the ceiling into t.For w = 2 µm and a 1:1 target, that is about 2 µm of coating; for 1:1 at w = 1 µm, about 1 µm. Then check that the value is comfortably inside the chosen grade's range.
4. Check the process, not just the number.Confirm the resist is sensitive at your exposure wavelength, that the developer is compatible, and that the bake step fits your handling time. A coating that is right on paper and wrong for the tool is still wrong.
5. Write down how it will be measured.A target without a measurement method is not a requirement. How to fix that basis is covered under "How the coating is checked" below.

Where a drawing mixes feature sizes, the smallest one sets the coating. A 10 µm window will develop perfectly well in a resist chosen for a 1 µm line; the opposite is not true.

One product family, three supply stages

TL OPTIK supplies the same product family at three points in the chain — bare glass, chromium-coated glass, and resist-coated plate — and the coating question means something different at each.

1. You buy a resist-coated plateThe thickness becomes your process input. State the target, the tolerance, the resist grade and tone, and the coating area, and confirm the soft-bake state you want to receive.
2. You buy a finished patterned maskThe thickness becomes an internal choice inside the writing and etch route. What you specify is the smallest feature, the polarity and the width tolerance; the coating is selected to deliver them.
3. You buy a chrome-coated blank, no resistThe coating question moves to your own line entirely. The relevant TL OPTIK parameters are the film family and the substrate, not the resist.

The published process order is the same in every case: substrate preparation, chromium sputtering, resist spin coating, soft bake, pattern exposure, development, chromium etch, resist strip and cleaning, then final inspection. Which of those steps sit on our side of the shipment is the only thing that changes.

What to put on the drawing

Coating requests that take an extra round of questions almost always omit the same two or three fields. The table below separates what a quotation needs from what production needs.

How the coating is checked

Thickness is a bulk property measured at points, not a picture of the whole plate. Three things therefore have to be agreed before the first part is inspected: where the measurement points are, how they are measured, and what record accompanies the plate.

Two points are rarely enough on a large plate. Coating thickness varies between the centre and the edge, and the edge is exactly where a contact frame or a mask holder will sit. A practical basis names a grid or a set of defined locations, a stated measurement method, and the reference conditions under which the values were taken — the same discipline that applies to any other dimensional requirement on the drawing.

For the acceptance side in more detail, see Inspection and acceptance of patterned glass, which covers how to define zones, defect significance and the conditions under which two measurements can be compared.