Ghodssi R., Lin P., MEMS Materials and Processes Handbook

Подождите немного. Документ загружается.

724 D.R. Hines et al.

Fig. 9.45 Transfer printed ﬁlms: (a) Optical image of a Au serpentine line transfer printed onto a

Kapton substrate; (b) optical image of a Pn ﬁlm (and Au ﬁlm) printed onto a PET substrate; and

printed Au electrodes (a)and(c). Reprinted by permission from [220]. (b) Reprinted by permission

from [221]

Au to Si is low enough to allow Au to be easily printed onto a plastic substrate (such

as PET or PC). However, the adhesion of Al to Si is too high for the Al print from a

bare Si substrate.

If, however, the Si surface is coated with a TDFS SAM prior to Al deposition,

then the adhesion of the Al to the Si substrate surface can be reduced to the point

where it can be successfully printed. The same is true for PMMA ﬁlms. Also P3HT

has been shown to print better from OTS-coated Si as compared to bare Si sub-

strates. In addition to lowering the adhesion between the printable layer and the

9 MEMS Lithography and Micromachining Techniques 725

Fig. 9.46 Micrograph

showing a Si grid resonator

printed over a preprinted

cavity on PC ﬁlm

transfer substrate, the adhesion between the printable layer and the device substrate

can be increased. This can be accomplished in various ways using SAMS [223],

plasma treatments [224], increased surface roughness, and so on. The transfer print-

ing method described above has been implemented to assemble dissimilar materials

sequentially onto a single substrate for the fabrication of both electronic devices

such as thin-ﬁlm transistors and mechanical devices such as thin-ﬁlm resonators;

examples of both are shown in Fig. 9.47.

9.7 Case Studies

In a commercial semiconductor device fab or MEMS foundry, the lithography is per-

formed using large-diameter wafers (8–12 in. diameter). The wafers are mounted

in cassettes that can be loaded into each machine or solvent bath for automated

processing. In a research and development lab it is more typical to handle smaller-

diameter wafers (3–6 in.) or even pieces of a wafer such as 1 × 1cm

.The

following case studies are geared toward the processing methods in a research and

development lab.

9.7.1 Case Study 1: Substrate Cleaning-RCA Clean(s)

Werner Kern developed the basic procedure in 1965 while working for Radio

Corporation of America (RCA). The full cleaning procedure consists of three steps:

726 D.R. Hines et al.

Fig. 9.47 (a) Electronic and (b) mechanical devices assembled onto plastic substrate using transfer

printing methods (Reprinted by permission from [ 220]and[225]

1. Removal of the organic contaminants (organic clean)

2. Removal of thin oxide layer (oxide strip)

3. Removal of ionic contamination (ionic clean)

9.7.1.1 Recipe Steps

Organics Removal (RCA1)

1. Submerge substrate in a 1:1:5 solution of NH4OH + H

+ H

O at 75–80

◦

for 15 min.

9 MEMS Lithography and Micromachining Techniques 727

2. Rinse in DI water.

3. Blow dry with N

gas.

4. Put in 120

◦

C oven for 20 min.

Oxide Removal (RCA2)

1. Submerge substrate in a 1:50 solution of HF + H

O (buffered oxide etch -BOE)

at 25

◦

Cfor1min.

2. Rinse in DI water.

3. Blow dry with N

gas.

4. Put in 120

◦

C oven for 20 min.

Ionic Contamination Removal (RCA3)

1. Submerge substrate in a 1:1:6 solution of HCl + H

+ H

O at 75–80

◦

Cfor

15 min.

2. Rinse in DI water.

3. Blow dry with N

gas.

4. Put in 120

◦

C oven for 20 min.

9.7.1.2 Notes

In many cases, removal of organic contamination is sufﬁcient; however, further

cleaning can be performed for removal of oxide layer and ionic contamination.

9.7.2 Case Study 2: Substrate Cleaning, O

Plasma Clean

The O

plasma cleaning method is primarily used to remove organic contaminants.

One of the main advantages of plasma cleaning is that it does not employ wet chem-

istry. This case study is based on using a PlasmaTherm 970 Reactive Ion Etcher

(RIE).

9.7.2.1 Recipe Steps

1. Load wafer into the plasma chamber.

2. Run O

plasma for 5 min at 200 W, 100 mtorr, 20 SCCM.

3. Vent chamber and remove wafer.

9.7.2.2 Note

Depending on processing conditions, plasma treatments have the potential to affect

the roughness of the substrate surface. Usually the roughness increases.

728 D.R. Hines et al.

9.7.3 Case Study 3: Substrate Cleaning, Solvent Clean

The solvent clean is a general-purpose method that can be performed in most any

laboratory environment.

9.7.3.1 Recipe Steps

1. Submerge a substrate for 5 min in separate beakers of each of the following

solvents.

Trichloroethylene (TCE)

Acetone

Methanol

2-propanol (IPA)

2. Rinse with fresh IPA.

3. Blow dry with dry N

gas.

4. Place in 120

◦

C oven for 20 min.

9.7.3.2 Note

As the wafer is transferred from one bath to the next, be sure that the wafer surface

stays wet and that no part is allowed to dry before entering the next bath – never

let the wafer surface dry – always keep the wafer surface wet. For example, acetone

tends to leave a very thin residue on the substrate surface that is extremely difﬁcult

to remove. As an optional step, ultrasonic agitation can be added to each solvent

bath during cleaning.

9.7.4 Case Study 4: Positive Photoresist Processing: General

Processing for Shipley 1800 Series Photoresist

This case study is adapted from processing notes from the Cornell Nanofabrication

Center. Shipley 1800 series photoresist is a general-purpose, broadband (365–

436 nm) positive resist. It is best suited for use on a 5x stepper and HTG Contact

Aligner. It is not recommended for the 10:1 Stepper. The 1800 series resists include

S1805, S1813, S1818, S1818J (dyed), S1827 (0.5, 1.3, 1.8, and 2.7 µm). This case

study assumes that the substrate is a Si wafer.

9.7.4.1 Recipe Steps

1. (Optional) After wafer cleaning, perform a dehydration bake by putting the

substrate in a drying oven at 150

◦

C for 30 min.

2. Mount the substrate onto a spin coater and verify that the spin coater is set up

with the appropriate parameters.

9 MEMS Lithography and Micromachining Techniques 729

3. Apply an HMDS treatment to the substrate surface. Using a disposable pipette,

syringe, or eye-dropper, apply HMDS solution to the entire substrate surface.

Allow to sit for 10 s and then spin for 30 s at 3000–5000 rpm on a spin coater.

4. Dispense photoresist onto the middle of the substrate and spin at the desired

speed for 20–30 s (thicker ﬁlms take a longer time to reach uniformity). Substrate

surfaces containing topography may beneﬁt from slower rate rates.

5. Soft bake by removing the substrate from the spin coated and placing of a hot

plate. Bake for 1–2 min at 90–115

◦

C. Thicker ﬁlms beneﬁt from longer baking

times.

6. Mount the photoresist coated substrate and the appropriate photo mask into the

contact aligner (see contact aligner case study) and expose for the desired time.

Exposure time will vary depending on resist thickness, baking time, substrate

reﬂectivity, and so on.

7. Develop for 1 min in AZ 300MIF or CD 26. Can also use MF-321 (no dilution),

or Microposit Developer Concentrate (MDC) diluted 1:2 (MDC minimizes Al

etch rate but is not metal-ion free). Use agitation while developing.

8. (Optional) Hardbake at 115

◦

C for 1 min on the hot plate, or 20–30 min in the

oven. The hardbake serves to promote adhesion during wet etching or increase

selectivity during dry etching.

This case study is generic with some processing conditions not speciﬁcally

stated. Below is a speciﬁc case study designed for S1813. Positive photoresists can

be sensitive to humidity. After UV exposure it may be advantageous to leave the

photoresist coated substrate sit out in the cleanroom for 30 min. This allows the

photoresist layer to absorb humidity which tends to minimize the production of

bubbles in the resist layer during postbake.

9.7.5 Case Study 5: Positive Photoresist Processing: Speciﬁc

Processing for Shipley S1813

This case study is adapted from the Ph.D. thesis titled “Design of MEMS-

based Tunable Antennas, Organic Transistors, and MEMS-based Organic Control

Circuits” by Madhurima Maddela, Auburn University. The processing steps are

designed speciﬁcally for Shipley S1813 positive photoresist.

9.7.5.1 Recipe Steps

1. Clean the substrate.

2. Mount the substrate onto a spin coater and verify that the spin coater is set up

with the appropriate parameters.

3. Dispense enough S1813 photoresist to cover the center half of the substrate and

then spin coat for 30 s at 2500 rpm using a ramp rate of 500 rpm.

730 D.R. Hines et al.

4. Softbake the photoresist coated substrate by placing it on a hot plate at 110

◦

for 60 s.

5. Mount the photoresist coated substrate and the desired photo mask into an optical

projection system (i.e., a contact aligner or a stepper) and expose. The exposure

time varies from system to system and will most likely need to be determined by

trial and error. Ask a fellow cleanroom worker or technical support staff member

for a good starting point.

6. Perform a postexposure bake by placing the substrate onto a 110

◦

C hot plate for

60 s.

7. Develop the photoresist layer by submersing in CD-30 photoresist developer for

45 s. Gently agitate while in the solution. Remove the substrate and rinse for 30 s

in DI water. Remove the substrate from the DI water and dry with a N

gun.

9.7.6 Case Study 6: Positive Photoresist Processing: Speciﬁc

Processing for OiR 906-10

This case study is adapted from the Ph.D. thesis titled “Monolithic Suspended

Optical Waveguides for InP MOEMS” by Kelly, Master’s Thesis, University of

Maryland, College Park, MD, June 2005. The processing steps are designed specif-

ically for OiR 906-10 positive photoresist. In order to achieve optimum sidewall

roughness and angle and the best possible feature resolution, projection lithography

was used to pattern suspended waveguides. A 5x reduction Alphastep 1000 stepper

lithography system was used to achieve critical dimensions down to 0.5 µm. A 5 in.

chrome on quartz mask witha5cm× 5 cm mask area and a 0.1 µm spot size was

used. This 25 cm

mask area was stepped down to a 100 mm

area on the chip.

9.7.6.1 Recipe Steps

1. Clean the substrate.

2. Mount the substrate onto a spin coater and verify that the spin coater is set up

with the appropriate parameters.

3. Apply an HMDS treatment to the substrate surface. Using a disposable pipette,

syringe, or eye-dropper, apply HMDS solution to the entire substrate surface.

Allow to sit for 10 s and then spin dry for 60 s at 3000 rpm on a spin coater.

4. Dispense enough Arch OiR906-10 photoresist to cover the center half of the

substrate and then spin coat for 60 s at 3000 rpm. The target photoresist ﬁlm

thickness is 1 µm.

5. Softbake the photoresist coated substrate by placing it on a hot plate at 90

◦

Cfor

60 s.

6. Mount the photoresist coated substrate and the desired photo mask into the step-

per and exposure for the desired time. The stepper parameters and exposure time

may best be determined by ﬁrst exposing and developing a test wafer (see process

notes below).

9 MEMS Lithography and Micromachining Techniques 731

7. Perform a postexposure bake by placing the substrate onto a 110

◦

C hot plate for

60 s.

8. Develop the photoresist layer by submersing in ODP 4262 photoresist developer

for 60 s. Gently agitate while in the solution. Remove the substrate and rinse for

30 s in DI water. Remove the substrate from the DI water and dry with a N

gun.

9.7.6.2 Notes

Several alignment procedures associated with the Alphastep stepper may need to

be carefully addressed: autoleveling of wafer pieces cannot be implemented on this

stepper and the lamp intensity can ﬂuctuate over periods of only a couple of hours.

In order to level the wafer pieces, we employed a technique of shimming the chuck

using various pieces of aluminum foil and Post-It notes. A program was run to

measure the wafer height across the chip using the stepper’s laser measurement

system, then the chuck was shimmed to try to level the surface. The focal depth

of the stepper optics was 2 µm, so this process was repeated until the die-to-die

variation in height was less than 1 µm. In our case, a single 10 × 10-mm die per

chip was used so the entire chip area had to be leveled to ±1 µm to achieve optimum

resolution.

In order to provide the correct exposure time and focus depth, a focus exposure

matrix was used prior to exposing each chip. A mask with a test pattern of bars and

spaces varying from 2.0 µmdownto0.5µm in width was used for this matrix. An

exposure dose between 100 and 120 mJ/cm

with a focus depth between −1 and 1

provided the best feature resolutions and resist proﬁles. The optimum exposure time

and stepper parameters were determined from an optical inspection of the developed

test wafer(s) using a high-quality optical microscope.

9.7.7 Case Study 7: Negative Photoresist Processing: Speciﬁc

Processing for NR7-1500PY

This case study describes speciﬁc processing conditions for the Futurrex NR7-

1500PY negative photoresist. An MJB3 manual contact aligner is used for UV

exposure.

9.7.7.1 Recipe Steps

1. Perform solvent clean on a Si wafer.

2. Mount wafer on spin coater and set up spin coater with a ramp rate of 1000 rpm

and spin speed of 4000 rpm and a time of 60 s.

3. Dispense photoresist onto the surface of the wafer. Dispense enough resist to

cover at least half the surface area of the wafer. Turn on the spin coater.

4. Remove the photoresist coated substrate from the spin coated and place on a

preheated hotplate at 90

◦

Cfor1min.

732 D.R. Hines et al.

5. Mount the substrate and desired photomask onto the MJB3 contact aligned. With

the microscopy, image a feature on the mask. Then, using the z-translation bar

and ﬁne adjustment knob, raise the wafer up such that the top surface of the

photoresist is just below the bottom surface of the mask when the z-translation

bar is in the “contact” position (be sure that the slide bar is in the contact position

and not in the “soft contact” position). This is the position just before the shadow

image of the mask feature disappears. Using the z-translation ﬁne-adjustment

knob, raise the wafer into contact with the mask. This is the point where the

shadow image of the mask feature disappears. Now move the slide bar into the

soft contact position. At this point the mask and photoresist surfaces will be in

focus but not in contact. Adjust the X, Y, and Rotation of the wafer to be in

alignment with the mask. Once proper alignment has been established, move the

slide bar into contact mode. Set the exposure time to 16 s and push the expose

button.

6. Remove the wafer from the contact aligner and place onto a preheated hotplate

at 90

◦

Cfor1min.

7. Submerge the wafer into a bath of RD6 photoresist developer for 15 s. Gently

agitate either the solvent or the wafer during developing. Remove the wafer and

submerge into a bath of DI water for 30 s and apply agitation. Remove the wafer

from the DI water bath and rinse in DI water (i.e., spray down with a squirt bottle

ﬁlled with DI water). Then dry with N

gas (i.e., use a N

gas stay gun equipped

with a point-of-use ﬁlter connected to a clean dry source of N

gas).

8. Perform etch or deposition processing as required.

9. Prepare a bath of RR2 photoresist solvent that is large enough so that the wafer

can be vertically submerged (not laying ﬂat but rather standing on edge). Also

prepare separate acetone and IPA baths in which the wafer can lay ﬂat. Place the

RR2 bath on a hotplate and heat the solution to 85

◦

C. Be very careful not to let

the solution temperature rise above 100

◦

C as RR2 is ﬂammable! Place the wafer

vertically into the warm RR2 bath and gently agitate. Once liftoff is complete,

transfer the wafer to the acetone bath. Agitate gently and then transfer the wafer

to the IPA bath (as the wafer is transferred from one bath to the next, be sure that

the wafer surface stays wet and that no part is allowed to dry before entering the

next bath – never let the wafer surface dry – always keep the wafer surface wet).

Remove the wafer from the IPA bath and rinse it down with IPA from a squirt

bottle. Then dry with N

gas.

9.7.7.2 Note 1

For liftoff involving Au ﬁlms less than 100 nm thick, the metal ﬁlm tends to lift

off in small pieces and look like a ﬁne mesh as it lifts off. Au pieces that are tens

of microns in size tend to backstick to the patterned Au features and cannot be

subsequently removed. This backsticking can largely be prevented by processing

-wafer pieces in a 500 ml bath of RR2 that is stirred at 300 rpm with a 1 in. long

stir bar. The wafer piece is held with ﬂat nose metal tweezers at a position such that

9 MEMS Lithography and Micromachining Techniques 733

the top surface of the wafer is pointed slightly down from the vertical and so that

the backside is slightly turned into the solvent ﬂow. This sample position allows the

Au ﬁlm and pieces to be swept away from the wafer surface during liftoff.

9.7.7.3 Note 2

Thicker Au ﬁlms can take several to ten or so minutes to completely lift off. We have

found that the liftoff time can be dramatically reduced by performing an O

plasma

exposure after deposition but before liftoff. For example, 5 min in a PlasmaTherm

790 at 200 W, 19 SCCM O

, and 100 mtorr reduces the liftoff time from about

10 min to about 1 min for a 100 nm thick Au ﬁlm. RR2 is an aggressive solvent

and can attack the Si surface; therefore, there can be an advantage to shorter liftoff

times. Albeit with limited investigation, the plasma exposure does not appear to

adversely affect the Au or Si surfaces, such a process may not be compatible with

other materials or with previous processing steps that may have been performed on

the wafer.

9.7.8 Case Study 8: E-Beam Lithography

E-beam lithography can be performed with any number of different SEM systems,

therefore only a generic account of SEM functions is presented as part of this case

study.

1. Start with a 1 cm

Si substrate that has been cleaned (see cleaning procedures

for UV lithography).

2. Spin coat a layer of MMA at 4000 rpm for 60 s and bake for 5 min at 200

◦

3. Spin coat a layer of PMMA 950 K A7 at 4000 rpm for 60 s and bake for 5 min

at 180

◦

4. With a diamond scribe, scratch a line i nto the PMMA. Go deep enough to pro-

duce a small number of Si shards near the end of the scratch. An example is

shown in Fig. 9.48.

5. Mount the substrate on the SEM sample holder. It is usually a good idea to place

the sample in a speciﬁc orientation so that it is easier to locate and align the

Fig. 9.48 Illustration of

substrate for e-beam

lithography