How to make an epoxy SU-8 mold?

The SU-8 mold fabrication process: tips and tricks

(In this review we are going to work with a silicon wafer, you can use other substrates such as glass with some little modifications of the process. If you have any question about other substrates contact us.)

1. Wafer preparation

Even if the wafer is new, it has to be prepared before receiving the SU-8 photoresist. If you are in a clean room, you can clean it with piranha solution (H2SO4+H2O2), outside a clean room you can use acetone. The cleaning can be optional if you are sure of the state of your wafer but it is strongly advised. In any case you have to heat your wafer to remove all moisture on the surface. We recommend a heating time of 15 min at 120°C in an oven for example. This heating step is really relevant because it will allow the SU-8 to stick better on the substrate.

If you have some difficulties to spread the photoresist over the surface or if during the soft bake the photoresist makes holes in the layer, it means that the photoresist doesn’t wet the substrate enough. A plasma treatment (O2 or Air plasma works fine) during 5 minutes will increase the spreading.
At last, even if, with the dehydration step, the photoresist doesn’t stick to the substrate, you can use an adhesion promoter which can be deposited by gas such as HMDS or in liquid form and spin coated first on the surface like the Omnicoat.

2. Spin coating of the negative epoxy resin SU-8 photoresist

To create the photoresist layer which is later going to be the mold, we use a spin coater. The spin coating is probably the most used technique to create a photoresist layer of a wanted thickness. We describe in more details the spin coating step in our review “how to get the best SU-8 spin coating”.
The spin coating consists in putting a puddle of SU-8 photoresist on a rotated substrate. The rotation speed, the acceleration and the SU-8 photoresist viscosity will define the thickness of the SU-8 photoresist layer.
To succeed this step some precautions have to be taken:

• Make sure your spin coater is perfectly flat.
• Put the wafer as centered as possible on the spin coater chuck to have the best spread of your layer.
• Put your photoresist as centered as possible for the same reason as the wafer.
• Coat the SU-8 photoresist in two steps, one at low speed and low acceleration during 10 to 30 seconds to spread the photoresist over the substrate and then a second step at high speed and acceleration during 30 to 60 seconds to control the layer thickness.
• Use a micropipette to dispense the photoresist to control the SU-8 photoresist quantity used.

3. Epoxy resin SU-8 soft bake (first baking of the photoresist)

The aim of the soft bake is to evaporate the solvent to make the SU-8 photoresist more solid. The evaporation will slightly change the thickness of the layer and prepare the SU-8 photoresist to be exposed to the UV. Indeed a quantity around 7% of solvent alows for good exposure. We have gathered more information about the relevant parameters to do a good SU-8 photoresist bake in our review “how to get the best SU-8 baking”.
It will depend on the thickness of the layer but keep in mind that the most problematic point will be the mechanical stress inside your SU-8 photoresist. To decrease this stress as much as possible, you have to heat and cool down slightly. Use a hot plate to bake your SU-8 photoresist, it will allow you to have heat from the bottom to the top of your wafer and thus favor solvent evaporation. You can see which hot plate we recommend here.
For heating progressively we advise you to follow a special heat pattern, with a first plateau at 65°C then a second at 95°C, the time of each plateau depending on the thickness of the SU-8 photoresist layer.
At the end of this step, the wafer can be kept in dark and on a flat surface during several weeks before continuing and finishing the process without serious consequences.

4. Epoxy resin SU-8 edge bead removal

The edge bead removal step consists in removing the photoresist edge bead around the wafer after the spin coating. The edge bead appears because of the surface tension of the photoresist. Because of it, the photoresist layer is thicker at the edge and the height namely depends on the photoresist viscosity.
Especially for very viscous photoresists, where the edge bead can rise several µm, it has to be removed. Indeed, the edge bead prevents the mask to be as close as possible of the wafer during the exposure step and this gap will lead to a resolution loss in the design.
Usually the edge bead removal is done by injecting acetone forward the edge of the substrate when it is rotated at high speed. This step can be achieved by hand or automatically thanks to bigger equipment.

5. The UV exposure process

The aim of the exposure is to initiate the cross linkage by the activation of the PAC (PhotoActiv Component) in some parts of the photoresist. This activation will change the local properties of the resin which, after baking, will be soluble or not into a solvent. Since the SU-8 is a negative photoresist, it means that the part exposed to the UV light will become hard and the other part will dissolve during the development.
Some parameters have to been chosen carefully:

• The exposure wavelength of the SU-8 is 365nm.
• The time depends on the layer thickness and the power of the lamp.
• The contact mode is important, less than a hard contact you may lose some resolution and with a vacuum contact you may stick the wafer to the mask.

6. The epoxy resin SU-8 post exposure bake (second baking of the photoresist)

The second photoresist bake is called PEB (Post Exposure Bake). It’s done as its name said, just after the UV exposure. The UV exposure activates the photoactive components in the SU-8 photoresist, but it needs energy to continue the reaction; this bake brings that energy. As for the soft bake, the most problematic point will be the mechanical stress inside your SU-8 photoresist and thus the heating and cooling down will have to be done slightly to decrease as much as possible this stress.
The heating pattern is the same as in the soft bake, a first plateau at 65°C then a second at 95°C, the time of each plateau depends on the thickness of the SU-8 photoresist layer.
Make sure your wafer is at the room temperature before continuing the process.

7. Development of the negative epoxy resin SU-8 photoresist mold

The development is the step where the non linked SU-8 photoresist is diluted in solvent. At this step the design is revealed on the substrate. The SU-8 can be developed thanks to the SU-8 developer. It’s a product from Microchem mainly composed of PGMEA (Propylene glycol monomethyl ether acetate) but you can also develop the SU-8 thanks to Ethyl-lactate or Di-acetone alcohol.

• Put your wafer in a crystallizer with 50mL of developer solution.
• The time of the SU-8 development depends on the thickness of the layer between 1 min to more than 15min.
• During the development maintain a good agitation to have the best development as possible.
• Rinse your wafer with isopropanol. If some white traces remain, develop again during 1-2 minutes. If some remain again change the solution and use 50mL of new developer for 1-2 minutes.
• Dry your wafer.

8. Epoxy resin SU-8 mold hard bake (third baking of the photoresist)

The third and last photoresist bake is called “Hard bake”, it is the last step of the process but can be optional. A lot of strength remain inside the SU-8 photoresist at the end of the process that can create cracks on the surface or even delamination of the layer… The hard bake heats the SU-8 photoresist at high temperature (more than 120°C) to suppress these strengths. Thanks to it, some cracks disappear and the SU-8 photoresist becomes harder.
Once again, it’s the same as in the soft bake and PEB, the increase and cooling down of the photoresist have to be done slightly.

9. Checking measures

At this step the SU-8 mold is already finished, but we have to check if it fits our expectations. For that you have to watch it under a microscope for a first visual checking. Then you can use optical or a mechanical profiler to measure the depth of the layer.