Includes information on
- Wafer preparation and spincoating with various photoresists
- o Shipley 1818 (2um layer)
- o Shipley SJR5740 (~10um layer) -- discontinued product
- o Shipley SPR220-7 (~7-10um layer) -- replaces SJR5740
- o AZ-4620 (~10um layer) -- replaces Shipley SPR220-7
- o Shipley 1805 (500nm layer)
- o SU-8 25 (~25um layer)
- Mask alignment & exposure with the Karl Suss Mask Aligner and Quintel Mask Aligner
- Developing exposed wafers
Preparation for spin coating
1. Inspect the wafer for dust, smudges, or scratches. Blow particles and dust off the wafer with the N2 gun.
- Scratches or defects on the glass or in the Si film can affect spincoating and etching, so choose a different wafer if these defects might overlap with your feature area.
2. Dehydration bake in oven-vwr at 120°C, 30 min
- Transfer wafers into a teflon cassette (located in drawer below oven). Label your cassette with a label-ring (also in drawer, fits into a wafer slot in the cassette).
3. HMDS priming, 5 min (for S1818, S1805, SJR5740, AZ4620 photoresists)
- Using the red rubber gripper, transfer wafer cassette to the HDMS coating tank in sink4. Be careful not to touch the wafers.
- Make sure the sink is enabled, then set the timer for 5 min. Check to make sure that bubbles are visible in the HMDS reservoir.
- Once the timer/bubbler stops, remove the wafer cassette (again with the red gripper). Move wafers back into your box for transfer to the spinner.
The most commonly used photoresist for our fab is Shipley 1818. The large common bottle of S1818 is in the left C-locker in Y1; smaller dark-glass dispensing bottles are available in the microlab. Spincoating is done using the Headway spinner, also in Y1.
- The start/stop control is located on the floor pedal directly beneath the spinner.
Right = START, left = STOP.
CAUTION: make sure to turn the chuck speed to zero before you press the STOP pedal, as this both stops the chuck and cuts the vacuum. If the chuck is still spinning, there’s a good chance the wafer will fly off.
Spin coating protocol
General protocol for using spinner1:
1. Enable spinner1 and make sure the vacuum reading is in the correct range.
2. Select a chuck, fit it onto the spindle, and press down until you hear it click into place. The chucks for both 4" and 6" wafers are in the drawer below the spinner control box.
3. On the spinner control box: the display indicates the speed (in krpm), the left knob controls speed, and the right knob controls the time.
- One half-turn on the speed knob ≈ 500 rpm.
- There is a 30-second mark beside the time knob; for more accurate timing, use a separate timer and manually turn the speed knob to zero at the end of the spin time. Make sure that the speed knob is set to zero before pressing the START pedal, and remember to turn the speed to zero again after spincoating, before pressing the STOP pedal.
4. Place wafer on chuck, centering it as accurately as possible. Test-spin the wafer to check that it’s aligned properly; if the wafer is off-center, adjust it appropriately.
5. Follow the spincoating parameters listed belpw for S1818, S1805, SJR5740, or SU-8 25. For other photoresists, consult the manufacturer's instructions for spincoating a desired film thickness.
1. Shipley 1818 positive PR:
- 500 rpm to spread, 2500 rpm to spin, 30 sec total spin time.
- Approximate photoresist thickness = 2µm.
2. Shipley SJR5740 positive PR:
- 500 rpm to spread, 2500 rpm to spin, 30 sec total spin time.
- Approximate photoresist thickness = 10µm.
3. Shipley SPR220-7 positive PR:
- 500 rpm to spread, hold for 10s, ramp 10s to 1800 rpm, spin at 1800 rpm for 30s.
- Approximate photoresist thickness = ~10um.
4. AZ4620 positive PR:
- Pour at 0 rpm, ramp to 500 rpm @100 rpm/s to spread, ramp to 2500 rpm @ 300 rpm/s, hold for 30 sec.
- Approximate photoresist thickness = 10µm.
5. Shipley 1805 positive PR:
- 500 rpm to spread, 5000 rpm to spin, 30 sec total spin time.
- Approximate photoresist thickness = 500nm.
6. SU-8 25 negative PR:
- Pour at 0 rpm, ramp to 500 rpm @100 rpm/s to spread, ramp to 2000 rpm @ 300 rpm/s, hold for 30 sec.
- Approximate photoresist thickness = 25µm.
- Clean the mouth & cap of your photoresist bottle after each pour with a technicloth/acetone and keep the bottle closed. Dried photoresist residue can produce defects in the film.
6. After the wafer has stopped spinning, let it sit for 15-20 sec for the photoresist to firm up, then transfer the wafer to a hotplate or oven for softbaking.
Softbaking protocol for S1818, SJR5740, SPR220-7 and AZ4620
1. Shipley 1818
- Transfer wafer to 120 °C hot plate for 2.5 min. If you are processing a number of wafers, it may be more efficient to put them in the 90 °C oven for 30 min.
- Cool wafers for 10-15 sec on cooling block.
2. SJR 5740
- Bake on 70 °C hotplate for 6 min.
- Transfer to 90 °C hotplate, bake for 7 min.
- Cool for 10-15 sec on cooling block.
from Microchem rep:
- Bake on 90°C hotplate for 6 min
- Transfer to 115-120°C hotplate for 6 min
- edit: DO NOT COOL ON COOLING BLOCK. set on a technicloth to cool slowly; thermal shock will crack photoresist.
Microlab SPR-220 protocol:
- Bake on 115°C hotplate for 5 min
- Cool slowly, do not use cooling block.
- After exposure, wait 30 min, then do post-exposure bake at 115°C for 6.5 min. (Again, cool slowly.)
- Hardbake parameters: 80ºC hotplate for 15 minutes, cool slowly to room temp.
- Softbake @ 110°C hotplate for 2 min.
- Cool slowly, do not use cooling block.
- Set on technicloth to cool slowly.
1. Clean the opening of the photoresist bottle(s) with a technicloth/acetone, close tightly, and store in the appropriate locker/drawer. When cleaning the bottle-opening, take care to not allow any acetone or particles to fall into the photoresist or remain on the bottleneck.
2. Clean the vacuum chuck, also with a technicloth/acetone.
3. Change the Al foil in the spinner.
- Carefully remove the foil from the spinner, dump as much of the photoresist as possible into the waste bottle in the sink/hood, then throw away the used foil.
- NOTE: If the waste bottle is full to the shoulder of the bottle, do not pour in more photoresist/acetone. First remove the full bottle, cap and label it properly, and put it in the waste disposal cabinet in room 432B. Obtain a new organic-waste bottle from the blue bin behind 432B (old H2O2 with vented caps only), attach the funnel, and set the new waste bottle in the vented container in Y1. Overfilling an organic waste bottle makes a nasty mess and will bring down the wrath of Marilyn, so don't do it.
- Get a new piece of Al foil & press it over the spinner area. Take care to leave an opening for the spindle and vacuum openings, so that the spinner ventilation can still work.
Shipley 1818, SJR 5740, SPR220-7, AZ4620 (positive photoresists)
These are the most commonly-used photoresists for the Mathies lab.
Shipley 1818 has a film thickness of ~2µm, SJR5740 has a film thickness of ~10µm. (Note that SJR5740 has been replaced by SPR220-7.)
- Dehydration bake in 120°C oven, 30 min
- HMDS prime, 5-10 min
- Spincoat: 500 rpm to spread, 2500 rpm to spin, 30 sec total spin time.
o S1818: 2.5 min on 120°C hotplate or 30 min in 90°C oven. Cool for 10-15 sec on cooling block. o SJR5740: 6 min on 70°C hotplate, then 7 min on 90°C hotplate. Cool for 10-15 sec.
- Expose & develop
o S1818: minimum exposure energy = 80 mJ/cm2 (I-line UV), develop for 1 - 1.5 min in 1:1 MPDev:H20 o SJR5740: minimum exposure energy = 300 mJ/cm2 (I-line UV), develop for 1 - 1.5 min in straight MPDev concentrate
- RIE ashing: The O2 ashing step etches the photoresist layer, so it's important to limit the power and time settings.
02 ashing for 1 min at 150W and SF6 etch for 1.5 min at 200W are roughly the maximum values for those parameters that should be used.
Notes on Shipley SPR220-7: a ~10um thick resist to replace SJR5740, since Microchem has discontinued SJR5740.
- Dehydration bake & HMDS prime, same as above
- Spincoat: 500 rpm to spread, hold 10s, ramp 10s to 1800rpm. hold 30s at 1800rpm.
- Softbake (protocol from Microchem rep, see Microlab manual for other SPR220 protocol): 6 min at 90°C, then 6 min at 115°C. Cool slowly to room temp. DO NOT USE COOLING BLOCK -- thermal shock causes cracking in the PR film.
- Exposure energy = 320 mJ/cm2 minimum
- Develop in straight MPDev for 30sec - 1 min
Notes on AZ4620: a ~10um thick resist to replace SPR220 on particular photolithography.
- Dehydration bake & HMDS prime, same as above
- Spincoat: pour at 0rpm, ramp to 500rpm @100rpm/s to spread, ramp to 2500rpm, hold for 30sec.
- Softbake: 2 min at 110°C. Cool slowly to room temp. DO NOT USE COOLING BLOCK.
- Exposure energy = 400 mJ/cm2 minimum (calculated by g-line)
- Develop in 1:3 AZ400K:water for 1.5min
Shipley 1805 (positive photoresist)
Shipley 1805 photoresist gives an approx. film thickness = 500nm, and is best suited for features that are 3um wide or less in the mask.
- HMDS prime the wafer for 5~10mins after dehydrating the wafers in VWR (@120C) for 10~15mins.
- Spincoat: 500 rpm to spread, 5000 rpm to spin, 30 sec total spin time.
- Soft bake @ 120°C hotplate for 1.5min.
- Expose & develop: minimum exposure energy = 100 mJ/cm2 (I-line UV), develop for 30 sec in 1:1 MPDev:H20
(100mJ/cm2 = ~5-6 sec @ 20mW/cm2 in the KS Aligner)
- RIE in ptherm for transferring patterns onto the A-Si layer:
o O2 plasma for 15sec @ 150W, SF6 plasma for 30 sec @ 200W
Since this PR is recommended for features 3um or less wide, KS Aligner or Quintel (preferably KS Aligner) should be used in one of the contact modes. Even though 1.5mins @ 125C hotplate does dry up the solvent reasonably in the PR film and it doesn’t seem to stick to the mask during exposure, it’s recommended to check the mask between exposures and increase soft bake time to 2mins if the PR film is found to stick to mask.
O2 and SF6 plasma tend to etch the PR film during pattern transfer to A-Si process. Since S1805 gives a PR thickness of only 500nm, care should be taken not to exceed the above mentioned etch times in ptherm.
SU-8 for Polymer Molding
SU-8 2002 negative photoresist
This recipe gives an approximately 3µm thick PR.
- Surface preparation. Silicon wafers or A-Si coated glass wafers do not require HDMS priming with SU-8. After dehydrating the wafers in an oven 120C for 10-15mins, SU-8 can be spin-coated directly at the spinner.
- Spin Coating. Within a few minutes of the above described surface treatment step, dispense PR directly from the dropper bottle onto the slow-spinning (~10-15rpm) wafer. When PR spreads over the entire wafer, ramp to 640rpm hold for 30secs.
Note on other thicknesses: 500rpm for 30sec -> 3.3µm; 800rpm for 30sec -> 2.7µm, 1000rpm for 2.5µm.
- Soft bake. Since this is thicker than proposed 2µm film PR, it’s necessary to perform soft bake in the hot plate ovens to achieve more uniform densification of the film. Putting the wafer directly on hot plates would create stress thus induce peeling. Pre-bake @ 65C for 3mins followed by soft-bake @ 95C for 4mins. Closely monitor the temperatures on the hot plate ovens because fluctuations also induces stress in film.
- Exposure. UV in the I-line range (365nm) initiates the cross-linking/curing process over the exposed areas. A little over exposure (200mJ/cm2) is recommended to obtain better adhesion and vertical side walls on the features after development. If using KS Aligner, use in the proximity mode with exposure gap = 10~25um depending on how critical the final feature size is.
- Post Exposure Bake (PEB). This is a very very critical step and if not performed properly, leads to defects in the final features. It is performed in hot plate ovens @ 65C for 3mins followed by 4mins @ 95C. Extended stay in the ovens may dry up the PR film and create additional cross-linking.
- Develop. 1.5mins in SU-8 developer should be sufficient to develop the film. If this doesn’t seem sufficient on visual inspection, gently shake the wafer in the developer for 1~2 more minutes. Quickly rinse off the developer with sufficient IPA.
- Hard bake. This step is recommended, if the features are going to be a permanent part of the wafer, cure the wafer @ 150-200C for 15-30mins. Take care to ramp the temperature to 200C and ramp it down so that thermal stress induced in the film is minimal. Too much stress can peel the film off, especially when the film is thicker than optimal.
SU-8 2025 negative photoresist
SU-8 2025 gives an approximate film thickness of 25µm.
- Surface preparation. After dehydrating the A-Si coated glass wafers in VWR (@120C) for 10~15mins, they can be directly spin coated with SU-8. But if SU-8 is going to be a structural feature on the wafer or it’s going to be spin-coated directly over glass or good adhesion of SU-8 to the surface is critical, application of AP300 adhesion promoter is recommended. Pour 2-6ml of AP300 over the stationary wafer at spinner1. Ramp to 4000rpm and hold it at that speed for 30secs. Note- Opened AP300 bottles should be stored under N2 atmosphere or replaced after 3months of opening the bottle.
- Spin Coating. Within 1min of the above described surface treatment step, dispense PR directly from the stock bottle (since, viscosity = 2500cStokes, high) onto the stationary wafer. Ramp from 0 to 500rpm @ 100rpm/sec (=> 5sec ramp time) for spreading the PR film over the entire wafer. Following this, ramp from 500rpm to 2000rpm @ 300rpm/sec (=> 5sec ramp time) and hold @ 2000rpm for 30secs.
- Soft bake. Since this is a thick film PR, it’s recommended to perform soft bake in the hot plate ovens (to the right of hot plates in Y1) to achieve more uniform densification of the film. Pre-bake @ 65C for 4.5mins followed by soft-bake @ 95C for 8.5mins. The thermocouples attached to the hotplate ovens are usually off and hence, use thermocouple from the lab. (310 Lewis) to get a better idea of temperature inside the ovens.
- Exposure. UV in the I-line range (365nm) initiates the cross-linking/curing process over the exposed areas. A little over exposure (300mJ/cm2) is recommended to obtain better adhesion and vertical side walls on the features after development. If using KS Aligner, use in the proximity mode with exposure gap = 10~25um depending on how critical the final feature size is.
- Post Exposure Bake (PEB). This is a very critical step and if not performed properly, leads to defects in the final features. It is performed in hot plate ovens @ 65C for 2.5mins followed by 4.5mins @ 95C.
- Develop. 4mins in SU-8 developer should be sufficient to develop the film. If this doesn’t seem sufficient on visual inspection, gently shake the wafer in the developer for 1~2 more minutes.
- Hard bake. Though this step is not necessary, if the features are going to be a permanent part of the wafer, cure the wafer @ 200C for 15mins. Take care to ramp the temperature to 200 and ramp it down so that thermal stress induced in the film is minimal. Too much stress can peel the film off, especially when the film is thick.
Alignment and Exposure: General
After the wafers have been coated with photoresist and soft baked, they may be taken to either quintel (Quintel Mask Aligner) or ksaligner (Karl Suss Aligner) for mask alignment and pattern exposure. The quintel can only accommodate 4” wafers; the ksaligner can handle 4" or 6" wafers. The ksaligner is also generally better for doing back-side alignment.
1. Shipley 1818 requires a minimum of 80 mJ/cm2 (i-line) exposure, SJR 5740 requires ~300 mJ/cm2 (i-line), SPR220-7 requires a minimum of 320 mJ/cm2 (i-line), and AZ4620 requires a minimum of 400mJ/cm2(g-line). Although AZ4620 is a broad band photoresist, exposure power is calculated by g-line. Check the intensity of the lamp, then calculate the necessary exposure time.
2. To measure the lamp intensity, use the Karl Suss UV intensity meter and follow the instructions given in the quintel or ksaligner lab manual.
- For quintel, the intensity is often marked on a piece of tape on the front of the machine.
- For ksaligner, the intensity can also be read (during a test exposure) on the display of the box underneath the ksaligner table.
3. Check the lab manual for specific instructions on how to set the program, exposure time, and other paramenters.
4. Before loading your mask in the aligner, make sure that it is clean, with no particles or photoresist residue on the feature areas. You may be able to just blow the particles off the mask with the N2 gun; if not, use acetone and a lollipop swab to clean off the mask, then rinse with IPA and blow dry.
1. Remove the dummy mask. Place your mask chrome (brown) side down on the mask vacuum chuck, with the letters facing to the right. Center mask over the opening and press MASK VACUUM to lock the mask in place.
2. To align the first pattern to a wafer:
- Load the wafer and wait for the display to read "Separation".
- Use the right-hand joystick to align the wafer to the mask. Center the edge of the wafer in each of the four alignment windows. (If there are no alignment windows, alignment is much more difficult, so make sure to include them in your mask design.)
o Coarse align is 3:1; fine align is 150:1. Press & hold the COARSE button on the left front panel while using the right-hand joystick to switch to coarse align.
3. To align the mask pattern to a pattern already on your wafer:
- Turn on the split-field screen and adjust the illumination such that you can see your mask pattern. The illumination control knob should be set to MIC, and illumination intensity set so that each screen shows the features clearly.
- Find the desired alignment marks with each microscope objective, using the control knobs on the microscope assembly.
o Magnification is controlled independently for each objective; it's usually easier if both objectives are set to the same magnification.
- Using the split-field screen, align the mask and wafer. The left-hand joystick moves the wafer and mask assembly together (scan-align mode); the right-hand joystick controls alignment between the wafer and the mask. The micrometer on the left side of the stage to rotate the wafer relative to the mask.
4. When the alignment is set, press the SEPARATE/CONTACT button to bring the wafer back into contact with the mask. The display should read "Contact". Do not move the wafer when it is in contact with the mask, as this could damage the wafer and/or the mask.
5. Press EXPOSE. Don’t look at the light (UV hazard). When the exposure is finished, the wafer should be automatically lowered and ejected.
6. If the alignment is good on the first wafer, the following wafers should be automatically close to good alignment. Some fine-tuning (using the procedures described above) may still be necessary.
Developing Exposed Photoresist
- For Shipley 1818 photoresist, develop each wafer for 1 - 1.5 minutes in a 1:1 solution of Microposit Developer (MP DEV) concentrate and water.
- For SJR 5740 photoresist, develop each wafer for 1 - 1.5 minutes in straight MP DEV concentrate.
- For SPR220-7 photoresist, develop each wafer for 0.5 - 1 minute in straight MP DEV concentrate.
- For AZ4620 photoresist, develop each wafer for 1.5 minute in 1:3 solution of AZ300K developer and water.
Agitate the wafer in the developer, then rinse with water, blow dry, and inspect under the microscope. The feature walls should be even and clearly defined, and the exposed areas should be clear. Make sure to check areas with fine or narrow features (hyperturns, etc), as those are the most likely to be underdeveloped.
- Note: the developer etches away exposed photoresist much faster than unexposed PR, but the unexposed areas do also become thinner. If you develop for too long, you can etch the unexposed areas too much and render the photoresist mask layer useless.
- If you see photoresist residue in the exposed areas, the wafer may be underexposed or underdeveloped. First try developing the wafer longer, for 10-20 seconds at a time, and inspect the wafer each time. You can increase the developing time by ~1 minute, but if there is still residue remaining in the exposed areas after that, increase your exposure time on the next wafer.
- If there are one or two areas where the photoresist did not expose properly (because of particles on the mask, etc), you can use a swab to put developer just on those areas. Rub the area gently with the swab for a few seconds, then rinse and inspect. Again, be careful not to overdevelop.
- Finally, make sure that your chrome mask is clean. Photoresist can sometimes come off the wafer and stick to the mask, or there may be dust or other particles blocking the path of the UV light through the mask. If the mask needs cleaning, remove it from the aligner and follow the cleaning steps described above.
- Eric Chu 17:15, 21 May 2009 (PDT):
or instead, discuss this protocol.