BioWire Photolithography

From FreeBio

Contents 1 General Outline 2 Design a Mask 3 Making a Master Mold 3.1 Preparation 3.2 Coating 3.2.1 Spin Coating 3.2.2 Pour Coating 3.3 Expose 3.4 Post-Expose Bake 3.5 Develop 3.6 Rinse and Dry 3.7 Verify 3.8 Storage 4 Making a PDMS mold 5 Making Stamps 5.1 Making Agarose stamps 5.2 Making PDMS stamps 6 Using the stamps

General Outline

1. Design the features you want and make a mask. Remember to decide if your features need to be transparent or opaque.
2. In the cleanroom, coat a layer of photoresist on top of a wafer. (A negative photoresist, like SU-8, will be cured by UV exposure, leaving the SU-8 wherever the light shines through. A positive photoresist, like PMMA, will be etched away by UV exposure, the cleanroom.
3. Silanize the SU-8.
4. Pour PDMS over the master in order to make a PDMS mold.
5. Silanize the PDMS mold.
6. Pour agarose (5% high gel strength) over the PDMS mold.
7. Remove agarose stamps.

Design a Mask

1. Design the patterns you want using a CAD tool. We used AutoCAD. Remember to decide if you want your features to be translucent or opaque. (This depends on what type of photoresist you will be using and if you want a positive or negative master mold.)
2. Get a mask made from that design. We went through Dr. Lu (ylu@fas).

Making a Master Mold

Preparation

• Plasma clean a Si wafer.
• Cover your hot plate with a glass plate or aluminum foil. It is best to level the surface; there are metal stages with four screws that can be placed underneath the hot plates. We recommend using the digital hot plates, so that you can ramp the temperatures up and down controllably. If you can't use them for whatever reason, cover two hot plates and preheat one to 65C and the other to 95C.
• Cover the spin coater bowl with aluminum foil. Trust us, you'll be glad you did when almost all the SU8 spins off the wafer. (This is normal for spin coating, don't worry.)
• If SU8 spills, [wait for hotplate to cool and] wipe the SU8 off with acetone.

Coating

Spin Coating

• Place SU8-2075 down on Si wafer in spin coater. Make sure the wafer is larger than the chuck. Pour more SU8 than you think you need. Most of it will spin off, but the important thing is to let the SU8 reach the edges of the wafer before starting to spin. To this end, try to pour the SU8 in a wide center circle and let it sit for 4-6 minutes. If you must, tilt the wafer to help it spread, but be aware that this might trap air bubbles. (It hasn't for us, but Andrew warned us that it might.

Remember to clean up all SU8 spills. If using tweezers, clean them with acetone and then rinse with IPA. Also remember to remove all SU8 from the bottle threads before recapping. You don't want an SU8 bond sealing the cap.

MIT recommends pouring from the bottle close to the wafer and twisting bottle while lifting to terminate pouring without getting resist on bottle threads.

• Cycle times and acceleration

• Spread cycle: Ramp to 500 rpm at 100 rpm/sec acceleration. Hold at this speed 5-10 sec to allow resist to even out over the entire surface.
• Spin cycle: Ramp to final speed at 300 rpm/sec acceleration and hold for 30 sec.
• Keeping at the final speed for a bit longer (15 seconds or so) seems to help uniformity at the price of depth. We did a full two cycles of the above recipe when aiming for 150 µms and got a very uniform surface (+- 20 µms) centered at a height of 100 µms.

RPM	Thickness
3000	75 µm
2000	110 µm
1500	150 µm	Observed: 90 µm with two consecutive spin cycles
1000	225 µm

The spin cycle recipe for 150 µm is programmed into the second floor cleanroom's SU8 spinner (right hand side) as recipe 2126.

We also have recipe 2127, but don't use it. Right now, it makes the wafer fly off and shatter, which is generally considered undesirable.

• Edge bead remover or acetone should be used if tendrils/matchsticks/cotton candy forms. Use the swabs available in the inner cleanroom and gently work around the perimeter (bottom perimeter) of the wafer after it is done spinning.

• After each spin cycle, wafers must be softbaked. If using two hot plates, use the table below, and then step down back to the 65C plate for 2 minutes. If using one of the digital hot plates, as recommended, preheat it to 65C, place the wafers down, and then ramp to 95C at 120C/hr. After leaving it at 95C for the times listed below, ramp back down to 65C at 120C/hr. Each ramping should add 15 minutes to the total time.

RPM	Pre-bake @ 65C	Soft-bake @ 95C
75 µm	3 min	9 min
110 µm	5 min	20 min
150 µm	5 min	30 min
225 µm	5 min	45 min

• If wrinkles or hills are visible after cooling, bake for longer as solvent is still present. (Besides distorting uniformity, it makes it likely that the wafer will bake onto the mask in the exposure step.
• If you want a thicker layer of SU8, spin cycles can be repeated as above. (There is no increase in baking times for subsequent layers, as the base had already been baked.) If using this approach, we recommend making one final spin cycle at high (3000 or so) RPM after the desired thickness has been reached. This will add <75 µms to the height of the wafer, but will ensure that the surface is uniform and level.

Pour Coating

• A completely alternate method was suggester by Mike Fuerstman (GMW group).

1. Plasma clean the wafers using the old plasma cleaner.
2. Weigh out ~10 g of SU8. This is non-reproducible, as SU8 is too viscous to pour easily. Completely cover wafer with resist. Tilt the wafer to spread the resist all over the wafer. You can also use a Pasteur pipette for this. Even if the layer isn't level, try to make sure that at least a little SU8 covers the entire plate.
3. Make sure that the hot plate is level. Use a four screw stage from the shelves between the inner cleanroom hoods underneath the hot plate. We recommend finding a plate of glass to put on the plate to make sure that it is level and to protect the plate without having to use aluminum foil. Place wafers on hot plate. Ramp from 40C to 110C at 90C/hr.
4. When the plate reaches 50C, pop the bubbles in the photoresist layer. Our favorite method is to blow air right above the bubbles with a Pasteur pipette. If you blow the air rapidly and right above a bubble you can pop it. Otherwise, the corner of a razor blade or the tip of a Pasteur pipette works just fine.
5. Bake for 15 hours at 110C. Set the hotplate to auto-off so that it ramps back down to room temperature after the 15 hours are up. (Times given for ~10g of SU8-2075. ~13g required 20 hours of softbake after ramping.)
6. We think that it's possible to do one spin coating at high rpm to add a thin layer to make a uniform top level. Use above protocol. This might cause the wafer to spin off the chuck and smash into smithereens. We'll have to check on that.

Expose

• If the aligner was off, turn it on and let it warm up. Then let it pulse for at least 3 seconds.
• Use the AB-M Mask Aligner to align the mask and master. Remember to use the rectangular chuck. First put on the wafer and get a subtrate vacuum, then put on the mask and get a mask vacuum, and lastly raise the wafer to [barely] contact the mask and get a contact vacuum. Use the clutch while raising the mask. The microscope on the right might also be useful for this; use the two buttons on the handle on the right to position it as needed, then raise the voltage until the imaging is clear (we needed 5V).
• Make sure the UV light source has auto expose turned off. Set the exposure time (table below). When the UV lamp is in place, make sure everyone is wearing UV goggles and press the [black] auto expose button.
• The AB-M has an intensity of 25 mW/cm^2 centered around a 365 nm peak. This accounts for the filter which removes light with wavelengths under 350 nm.

Thickness	Exposure Time
75 µm	12 seconds
110 µm	13.6 seconds
150 µm	15.2 seconds
225 µm	16 seconds
~10g	180 seconds
~13g	240 seconds

Post-Expose Bake

Place wafers on hot plate again. Again, we recommend the digital hot plates. Preheat to 65C, put the wafers down, and ramp to 95C at 360C/hr. This will take 5 minutes, so add 2-3 minutes to the 95C times listed below. (We tried ramping back to 65C at 360C/hr, but the most we got was 180C/hr, which seems to be equivalent to just shutting off the hot plate.

Two step heating applies again: 65C for 1 min and 95C according to the table below. Remember, it's important to let the wafer cool to room temperature slowly (put back on 65C) so it does not crack.

Thickness	Post-Expose Bake
75 µm	7 minutes
110 µm	10 minutes
150 µm	12 minutes
225 µm	15 minutes
~10g	70 minutes
~13g	80 minutes

• Mike's PEB protocol calls for ramping from 45C to 105C at 90C/hr. Leave at 105C for 20-30 minutes, again setting the plate to auto-off. (Once plate has been preheated, leave the timer on for 70 minutes. Allow some time for cooling down as well.)
• Remember that a long postbake will increase crosslinking in the SU8, widening features and lengthening development times.

Develop

When the plate reaches room temperature, remove the wafer and develop using PGMEA (1-methoxy-2-propanol acetate). Development times can vary widely depending on agitation rate, temperature, and other factors. Developing for too long shouldn't (in theory) matter if everything else was done perfectly, but in all probability, over-developing will lead the SU-8 to lift off from the wafer. Approximate times are in the table below. For long developments, the best approach is probably to construct a chamber like Mike's (a beaker filled with PGMEA and a raised platform to hold the wafer above a magnetic stir bar). Long developments should be done under a hood (most likely back in our lab).

Thickness	Development
75 µm	7 minutes
90 µm	Observed: 16 minutes
110 µm	10 minutes
150 µm	11 minutes
225 µm	12 minutes
~10g	25 min total, 21:30 of stirring, PGMEA change and 1 min of IPA at 15 min, 1 min of IPA at the end.
~13g	Hopefully the same as 10g.

• Development goes faster if the PGMEA is changed during the development. This removes the SU8 that has already come off to be removed from the solution.

Rinse and Dry

To check if development is finished, fill a separate chamber with IPA (isopropyl alcohol). If you see a white residue when the wafer is placed in the IPA, developing needs to proceed longer. Dry the mask with a nitrogen gun.

Verify

The features of the master should be verified using the clean room's microscope with the red filter and the machine (called a profilometer) with the needle that drags over features to measure thickness.

Storage

Store master in a petri dish covered in aluminum foil. [Andrew recommends plasma oxidizing the master.] Before making a PDMS mold, the SU8 master must be silanized. Kit has silane in his lab. He places the wafer in a vacuum desiccator overnight along wih a glass coverslip containg a small drop of silane. Mike suggested (TRIDECAFLUORO-1,1,2,2-TETRAHYDROOCTYL)TRICHLOROSILANE as an appropriate fluorinated silane to use (Gelest Inc.) and he coated the wafer for 3 hrs. Mark suggested spreading on 230 Fluid (Dow Corning).

Making a PDMS mold

1. Now we need a PDMS negative mold. Place the finished master mold feature side up in a 100 mm plastic petri dish.
2. Mix up and degas 50 grams of Sylgard 184 PDMS.
3. Pour the PDMS into the petri dish so that it covers the template competely and evenly.
4. Place the petri dish in a vacuum dessicator for at least one hour to degas the PDMS.
5. Bake the petri dish at 60 degrees C for one hour to cure the PDMS.
6. Once the PDMS is cured, use a pair of needlenose pliars to carefully break apart the petri dish around the wafer/PDMS.
7. Carefully peel the wafer off the PDMS and place it in a clean petri dish.
8. Cut away the excess PDMS from around the patterns.
9. Carefully cut the patterns into small squares. Notch the bottom edge of each stamp to keep track of which side the features are on.
10. Store the stamps feature side up.
11. This PDMS mold is the inverse of the master mold. So, if the master was a positive mold (the same as the stamp you want), the PDMS mold is a negative mold. If the master was a negative mold (the inverse of the stamp you want), the PDMS mold is a positive mold. If so, you have to cure the surface of the PDMS mold and make a second PDMS mold. This second PDMS mold will now be a negative and you can proceed.

Making Stamps

Making Agarose stamps

1. Clean the PDMS molds by placing them feature side up in a 50/50 solution of ethanol and distilled water, and sonicate them for 30 minutes, no heat.
   1. If PDMS molds have been used before, replace with autoclaving in water to remove little bits of agarose or you can just wash them with water and air dry them.
2. Treat PDMS mold with oxygen plasma and immerse in water. Pressure should be around 500 mTorr and plasma should be dark violet. This can be done in Whitesides Lab.
3. Place PDMS mold in small beaker. Pour agarose (5% high gel strength in 1X PBS, 30 mL/stamp) over the mold and let it harden. Pop bubbles on features by pipetting agarose in and out of pipet tip. This is now a positive agarose stamp.
4. Let the agarose stamp sit 20 minutes or until firm.
5. After solidification, remove the agarose stamp and PDMS mold together from the well and peel the mold away from the stamp. (In the BioLabs, using the green curvy spatulas is a very quick of getting the agarose chunk out of the beaker.)

Making PDMS stamps

1. Silanize the PDMS mold. Place a few drops of silane from Kit's Lab on glass coverslip and place next to PDMS mold inside toxic vacuum dessicator.
2. Let vacuum dessicator run for 24 hours and remove mold.
3. Follow the procedures for PDMS molds.

Using the stamps

1. To ink agarose stamp:

• If using cell culture to ink, try growing or back-diluting cells to OD600=0.3 (log phase)
• Place stamp into 6 well plate feature side up. Pour or pipette cell solution over stamp. Agarose will absorb excess liquid, leaving cells on stamp surface. Add liquid a second time, and let stand again. Remove excess liquid by blotting or by repeatedly stamping onto a petri dish.
• Alternatively, place the stamp onto a lawn of bacteria. Getting a lawn is your problem.

1. Use the perimeter stamp to put down a perimeter of cells. Let grow overnight.
2. Put the final stamp down within the perimeter. Try to center the features.