Standard 100 mm VLSI Process Modules

Table of Contents

(Click on MOD to go to the appropriate section.)

MOD 15 Standard Wet Oxide Etching

MOD 16 Plasma Etching in Technics-C

MOD 17 Standard Furnace Cleaning

MOD 18 Lift Off Process

MOD 19 Spin-On SiO2 (SOG) for Inter-Metal Dielectric

MOD 20 Standard Wet Nitride Etching

MOD 21 Etching Single-Crystal Silicon

MOD 22 Plasma Etching of Trenches in Single-Crystal Silicon

MOD 23 Plasma Etching of Thick PSG

MOD 24 Contact Etching

MOD 25 "Show" Wafer Process

MOD 26 Lam Monitors

MOD 27 OCG Reversal Image Process

MOD 28 Polyimide Technology

MOD 29 Standard Thick Resist Process (Shipley's SPR-220 Thick Resist)

MOD 30 Standard Aluminum Wet Etch

MOD 31 Processing Glass Wafers

MOD 32 Edge Bead Removal (EBR)

MOD 33 Pocket Wafer Fabrication for 6” Wafer

MOD 34 Anodic Bonding

MOD 35 Handle Bonding (Reversible Bonding)

MOD 36 Fusion Bonding

MOD 37 Standard Anti-Reflective Coating for DUV Process

MOD 38 ITO Deposition by Reactive Thermal Evaporation

MEMS Exchange

MOD 15

Standard Wet Oxide Etching

Purpose: To etch oxide films in buffered oxide etch

Equipment: wet process stations or fume hood

Summary:

(1) Being sure any photoresist on wafers has been hard baked, wet wafers in DI H2O to prevent bubbles sticking to film surface.

(2) Dip wafers in buffered HF (BHF) for required amount of time.

(3) Rinse/dry wafers per MOD 2.

Detailed Procedure:

(1) Wet wafers in DI H2O in tank 1 or 4.

(2) Immerse wafers in buffered HF for required amount of time based on etch rate (see below).

Etch rates (approximate/when the solution is fresh):

BHF 10/1 ~500 A/minute

BHF 5/1 ~1000 A/minute

(3) Follow with rinse/spin (MOD 2).

MOD 16

Plasma Etching in Technics-C

Purpose: Plasma etching of nitride and oxide films

Equipment: Technics-C

Note: The lam1 plasma etcher is the system dedicated to nitride etching and Lam2 is dedicated to oxide etch. Prefer to use these two systems for clean standard processes, which provide better uniformity in etching.

Summary:

(1) Carry out an O2 scourge to clean chamber.

(2) Vent the system and place wafers in chamber.

(3) Pump system down to base pressure (~35 mTorr).

(4) Introduce desired gases into chamber.

(5) Strike plasma by turning on power to desired wattage and time.

(6) Turn off power and gas.

(7) Pump down chamber to remove reacted gases.

(8) Vent chamber and remove samples.

Detailed Procedure:

(1) The status of the machine should be as follows:

Mode: Manual

SOL'N (Solenoid): Closed

Vent: Off

Power: Toggle Off, Knob Pegged Counterclockwise

Gas #1: Off

Gas #2: Off

Occasionally the solenoid , which controls the vacuum pump, is left open. If this is the case, close it before enabling the system.

(2) Carry out an O2 scourge to clean the system with O2 at 300 Watts for 10 minutes, following the outline given below for system operation (from Step (4)).

(3) Once you are ready to introduce the samples, vent the chamber by toggling the VENT switch. Be sure that the SOL'N is closed when you do this. It will take about 15 seconds for the chamber to fill. Once it is at atmospheric pressure, open it carefully -- the top is very heavy -- and place your wafers on the plate. Close the top carefully, being sure not to allow it to fall.

(4) Oxygen for photoresist descum/ashing and cleaning the system (scourge) is connected through Gas #1. The Gas #1 switch will flow (1) SF6; (2) He; and (3) O2. Check correction factors on the PD module and set points for the particular gas you are going to use.

(5) You are now ready to start the vacuum pump. Leaving the vent ON, toggle the solenoid (vacuum pump) switch open. After 2 or 3 seconds, close the vent switch to allow the pump to lower the pressure of the chamber.

(6) You can watch the pressure drop as the system comes under vacuum. When the system reaches base pressure, introduce the gases you need into the chamber by toggling the appropriate gas switches on the PD and PE modules. The pressure in the chamber will rise as gas flows in, and then stabilize. (See below for specific recipes for nitride and oxide etching.)

(7) Once gas flow into the chamber is stable and at the desired pressure strike a plasma by switching the POWER toggle on and turning the dial clockwise until desired wattage is reached. The plasma is visible through the window on the front of the chamber. Begin timing your run as required based on etch rate.

(8) Once the run is complete, turn off the power, then the gas. Always turn off the power before turning off the gas.

(9) Allow the chamber to pump down to base pressure to be sure all potentially harmful gases have been swept out of the chamber.

(10) Turn off the vacuum pump by switching the SOL'N toggle to closed position. Now you may vent the chamber. Again, remember not to vent the chamber until the SOL'N has been closed.

(11) The chamber will now come up to atmosphere and you may remove your sample.

(12) Once your sample has been removed, close the chamber and start the vacuum with the vent open. After a couple of seconds, close the vent and allow the chamber to pump down to base pressure mtorr. Close the SOL'N. Be sure that gas switches and power are off.

(13) Carry out an O2 scourge to leave the system clean. Use O₂ at 300 Watts for 15 minutes, following the outline given in steps (5) through (12).

A. Nitride Etch

Set Points: SF6 - 13.0

He - 21.0

Power: 100 Watts

Nitride Etch Rate: ~500 A/minute

Oxide Etch Rate: ~250 A/minute

Silicon Etch Rate: ~8000 A/min

Beware: This process etches single crystal silicon at about 8000 A/min!

It is suggested that you conservatively approximate the time you will need to etch through your nitride film, run your sample for half the total time, open the chamber and rotate your wafers around their central axes by 180 degrees, and then etch again. This provides better uniformity in etching.

MOD 17

Standard Furnace Cleaning

Purpose: To remove heavy metal ions from the furnace tube

Equipment: Tylan furnaces

Summary:

(1) Run STCA recipe: Standard TCA clean. The recipe has the following steps:

a) Ramp up to 1100ºC.

b) Oxygen flow: 2 minutes (0₂= 4, N₂= 0)

c) TCA clean: 5 minutes (TCA = on, 0₂= 2.0) 60 cycles of (b) and (c).

Note: The TCA flow rate can not be specified in the recipe.

1. Post-ox: 5 minutes (0₂ = 4, TCA = off)

2. Ramp down to 750ºC. This process takes about 8 hours.(The sixty cycles, loop alone, is seven hours.)

(2) Run MAIN recipe: Temperature calibration in 50 steps from 750ººC to 1100C, plus standard TCA clean. The recipe has the following steps:

(a) Temperature calibration up to 1100ºC.

(b) Oxygen flow: 2 minutes (0₂ = 4, N₂ = 0)

(d) 60 cycles of (b) and (c).

(e) Ramp down to 750ºC. This process takes about 10 hours.

Note: This calibration recipe has a tolerance of three degrees as opposed to the two degrees that the standard calibration recipe "SCAL" has. The "MAIN" program has this quick calibration that should only be used in case of a problem with the temperature is suspected. Normally, the "STCA" recipe should be chosen for cleaning.

Detailed Procedure:

(1) Put the Standard Recipes floppy disc in the disc drive.

(2) Load STCA or MAINtenance (if needed) program into required tube.

(3) When the computer asks for delay time, type in requested time such that the process ends around 8AM, or at the time you are starting your run in the morning.

(4) If the loading is completed, a GOOD LOAD response will appear on the screen.

(5) Go to the tube and press RUN on the ROP (remote operation panel), or type RUN tube#.

(6) After the cleaning is done, the alarm will sound. Press the ALARM ACK button on the ROP.

(7) For best results, use tube soon after cleaning.

Notes:

1. These programs can only be aborted in an oxygen flow step.

2. The TCA clean cannot be run in two adjacent tubes at the same time!

3. It is preferable to run a TCA clean during the night, so the temperature will not interfere with other processes.

MOD 18

Lift-Off Process

Purpose:

Lift-off is a process that enables the patterning of a thin layer metal film. It begins with the patterning of a thin layer(s) material, typically photoresist, followed by a directional metal deposition, usually via evaporation. After the metal has been deposited, the initial patterned layer is dissolved in solvent leaving behind a patterned metal film. It is critical that step coverage does not occur during the deposition of the metal. Step coverage disallows a clean “lift-off.” The process recipes described hereafter are specifically designed to achieve a sacrificial layer profile that disallows metal step coverage. This enables the likelihood of success for the final lift-off step.

Figure 18.1 - Basic Lift-Off Process Flow

(1) Sacrificial layer is patterned. Note profile disallows metal step coverage.

(2) Metal deposition, and

(3) Lift-off in solvent. Metal pattern remains.

I. G & I-Line Bi-Layer Method

This bi -layer method is a simple yet effective way to achieve a good undercut profile that disallows step coverage. The process uses standard resists, stocked by the Microlab. Automated spincoat and development tracks with standard recipes can also facilitate processing.

Figure 18.2 - G-line I-line Bilayer Method Process Flow

The acquired cross sectional profile is shown in SEM image.

Equipment:

Primeoven, Svgcoat1 or 2, Ksaligner, Svgdev,

Procedure:

1. HMDS prime wafer at Primeoven. Spincoat (svgcoat2) 1.3 micron G-line resist (5000 RPM). Soft bake hot plate 60 sec at 90ºC. Chill plate 6 sec.

2. Flood expose (ksalinger) 0.14 joule/cm^2 (non-critical, use clear field dummy mask).

3. Spincoat (svgcoat2) 1.1 micron I-line resist (4100 RPM). Soft bake hot plate 60 sec at 90ºC. Chill plate 6 sec.

4. Expose (ksaligner - soft contact mode). The optimal dose for a 5-inch chrome mask with a clear field pattern area between 5% and 10% is 0.06 joule/cm^2.

5. I-line develop. Although the automated track developer may also be used for this step, a timed tank develop allows for more precise undercut. The track develop will not necessarily yield optimal results for all applications. Optional: hotplate hardbake - 60 sec at 90ºC (see notes below).

6. Deposit metal. Keep substrate temp as low as possible so as to prevent profile degradation.

7. Lift-off in acetone.

Notes:

a. The exposure dose in STEP 4 is critical. Overexposing or underexposing the resist will ultimately effect the profile of the bilayer undercut.

b. The amount of undercut may be controlled with more precision by performing a timed tank develop.

c. GCAWS may be used with this method. Perform a focus exposure test, and a series of timed tank develops to determine the optimal exposure and develop time.

d. Depending on the feature size, an aggressive hardbake can cause critical dimension disruption due to rapid solvent release. If necessary, perform a low temp hardbake. If at all possible though eliminate the hardbake step altogether. This helps to ensure critical dimension integrity and makes the acetone liftoff step easier.

e. The amount of undercut can usually be observed with optical microscope inspection. For smaller features, a mask with a test structure array of incremental lines and spaces works well - the obliterated (100% undercut) lines vs. standing lines help pin-point the amount of undercut.

II. AZ5214

AZ5214 E is a positive resist that has the capacity for image reversal. The image reversal component is used to obtain a negative profile that prevents step coverage during metal deposition. A negative mask pattern must be used with this method.

Figure 18.3 - AZ5214 Single Layer Method Process Flow

Equipment: Primeoven, Spinner1, Sink4 or 5, Ksaligner

Procedure:

1. Spincoat (spinner1) 1.5 micron AZ5214 resist (4000 RPM). Soft bake hot plate 90 sec at 90ºC.

2. Expose (ksaligner) 0.08 joule/cm^2.

3. Image reversal bake: hot plate 2 min at 120ºC. This step is the most critical of the process.

4. Flood expose (ksaligner) 0.2 joule/cm^2.

5. Tank develop. AZ400K:H20 1:5 ratio; 30 sec.

6. Deposit metal. Keep substrate temp as low as possible so as to prevent profile degradation.

7. Lift-off in acetone.

Notes:

a. Negative resist will provide the same type profile, but can be more difficult to remove during the liftoff step. AZ5214 is a positive resist, and thus better suited for lift-off.

b. Double layer thickness of AZ5214 may be spun for thicker metal film deposition. Adjust bakes and exposure times accordingly.

c. AZ5214 may also be used as a normal positive resist.

III. LOR10B

LOR10B is a spin-on coating underlayer that selectively undercuts for lift-off. Here it is processed with Shipley 1818 resist.

Figure 18.4 - LOR10B Process Flow

SEM cross-sectional result is shown.

Equipment:  Primeoven, Spinner1, Sink4 or 5, Ksaligner

Procedure:

a. Spincoat (spinner1) 1 micron LOR10B (3000 RPM). Soft bake hot plate 5 min at 150ºC.

b. Spincoat (spinner1) 1.1 micron S1818 (4000 RPM). Soft bake hot plate 1 min at 120ºC.

c. Expose (ksaligner – soft contact mode) 0.04 joule/cm^2.

d. Tank develop MF319 60 sec.

e. Deposit metal. Keep substrate temp as low as possible so as to prevent profile degradation.

f. Lift-off in acetone.

MOD 19

Spin-On SiO₂ Glass (SOG) for Inter-Metal Dielectric

Purpose: To provide inter-metal dielectric for double metallization.

Equipment: Headway spinner, convection bake oven, and Tystar 4.

Summary:

(1) After defining the Al pattern (0.6Mm thick), bake wafers at 120ºC for thirty minutes.

(2) Coat wafer with SOG.

(3) Bake the wafers for one hour total time.

(4) Anneal the wafers per MOD 18, using recipe SOGN2 or SOGO2.

Detailed Procedure:

(1) Wafer Preparation: Dry the wafer in convection oven at 120ºC for 30 minutes. (MOD 4)

(2) SOG Coating:

(a) Enable spinner1

(b) Place wafer on the Headway spinner.

(d) Adjust speed to 3000 rpm.

(d) Set spinning time to 20 seconds.

(e) Use a dropper to dispense 3 cc of Futurex IC1-200 SiO₂ SOG on the wafer. Start spinning.

(3) Baking:

(a) Place wafer in 120ºC oven for 30 minutes. (MOD 8)

(b) Increase temperature of oven to 200ºC and bake wafer for an additional 30 minutes.

(4) Annealing: Anneal wafer in Tystar 4 using either SOGN2 or SOGO2 program (400ºC for 30 minutes.) (MOD 18)

MOD 20

Standard Wet Nitride Etching

Purpose: To etch nitride films in hot phosphoric acid

Equipment: Sink 7

Summary:

(1) Remove all photoresist.

(2) Dip wafers in oxide etch to remove any oxide from nitride film. The presence of even a very thin layer of oxide will prevent the acid from etching the nitride.

(3) Wet wafers in DI water to prevent bubbles sticking to film surface.

(4) Dip wafers in hot phosphoric acid for required amount of time.

(5) Rinse/dry wafers per MOD 2.

Detailed Procedure:

(1) Etch oxide from nitride by dipping in 10:1 HF for approximately 1 minute.

(2) Turn on hot phosphoric acid bath (right heated bath of sink7) by pushing the green TEMP CONTROL button on the right half of the sink. Set the controller to the desired temperature ( the standard is 150ºC ). Wait for temperature to stabilize. Acid will reach a rolling boil.

(3) Immerse wafers in hot phosphoric acid. The time required to etch the film will vary because small amounts of water are added to the bath as the level goes down, thus diluting the acid. For a 1000 A film, check the wafers after an hour. You should be able to tell visually when the film has been etched away.

(4) Follow with rinse/spin (MOD 2).

MOD 21

Wet Etching Single-Crystal Silicon

Purpose: To anisotropically wet etch single crystal silicon.

Equipment:

For EDP etchant: Sink 3, reflux system, left side.

For KOH etchant: Sink 3, pyrex beaker, right side.

Special Note:

Do not use wax at sink3 since it may contaminate other user's processes. The use of wax should be done elsewhere in the lab. Also, make sure all wax contaminated cassettes are completely separated from sink3 cassettes.

Summary:

(1) Clean the work area (this will take 30 minutes to 1 hour for the reflux system).

(2) Mix etchant (recipes in Chapter 1.10).

(3) Calculate etch time based on etch rate for the specific etchant.

(4) Etch silicon as required.

(5) Rinse/dry wafers per MOD 2(sink3). Clean work area.

Detailed Procedure:

A. EDP Etchant

Etch recipes and pertinent information are given in Chapter 1.10.

**** EDP IS DANGEROUS!!! IT IS IMPERATIVE THAT YOU WEAR

**** GREEN GLOVES, APRON, FACE SHIELD, AND A RESPIRATOR.

**** A RESPIRATOR FITTING SHOULD BE SCHEDULED THROUGH EH&S

**** RESPIRATORY PROTECTION AT 642-3073. ALL OF THIS SAFETY

**** EQUIPMENT SHOULD BE WORN AT ALL TIMES WHEN HANDLING THE

**** EDP APPARATUS WHICH INCLUDES CLEANING, WEIGHING AND

**** AND ADDING CHEMICALS, ADDING WAFERS, ETC.

(1) EDP does not etch silicon dioxide. Therefore, be sure to HF dip your wafer to remove native oxide from the silicon surface to be etched before proceeding. A 10 second 10:1 HF dip should be sufficient. Either silicon dioxide or silicon nitride is an acceptable masking material.

(2) Make sure the reflux system is clean. The top of the ring on the tank and the watchglass must be clean and dry when you begin to prevent noxious vapors from escaping into the room. Check the watch glass cover and the top ring of the tank for white crystalline material. If you find any, clean it off with plenty of DI water and techni-cloths. An RCA clean (Chapter 2.1 of the lab manual) of the tank is recommended prior to use.

(3) If you do not wish to carry out an RCA clean, proceed as follows to ensure that the tank is at least nominally clean before proceeding with your etch.

(4) Fill the reflux tank with DI water, and let it sit for 30 minutes to 1 hour to dissolve any chemical residues in the tank.

(5) Aspirate the water, which may be very discolored. Rinse the interior of the tank with DI water and aspirate, several times, until you are satisfied that the tank is clean.

(6) Before preparing the etchant, note that the amounts of chemical listed in the recipes of Chapter 1.10 are enough to cover 2" wafer cassettes. All volumes/weights must be doubled for use with 4" wafers! Also, see the reference listed at the end of this manual for further information. Obtain ethylenediamine, pyrazine and pyrocatechol from the Microlab office. There is a portable, battery operated scale in GL2 upon which the solid chemicals except pyrazine can be weighed. This should be done inside sink3. Plastic weighing boats are provided in the C-locker next to the sink and more can be found in the old lab. Remove the scale from the sink when you are finished using it. The pyrazine should be weighed in the enclosed scale across from sink3. The reagents are mixed in the following order:

(a) Weigh pyrocatechol and place in the reflux tank.

(b) Weigh pyrazine (if necessary - not all recipes call for it) and add it to the tank. Cover the tank with the watchglass - pyrazine is very volatile.

(d) Add the required amount of ethylenediamine.

A commonly used recipe in the Microlab is the 'F' etch with 6 g of pyrazine, which is:

320 g pyrocatechol

6 g pyrazine

320 ml DI water

1000 ml Ethylenediamine

at 110ºC

This leads to an approximate etch rate of 55 um/hr.

(7) Turn on the heater by pressing the left side Temp. Control button, and set it to the desired temperature. Wait 30 min for the bath temperature to stabilize.

(8) Place the wafers to be etched in the cassettes or holders provided in the tray labeled EDP located on the table in the center of GL2. EDP cassettes are readily identified because they are discolored. DO NOT USE THE KOH CASSETTES IN THE EDP TANK.

(9) Keep a pile of techni-cloths near to the sink so that after inserting or removing wafers from the tank and rinsing off your gloves in the glove wash, you can dry your gloves to reduce the possibility of getting EDP on your skin. Using the glass rod hook, carefully lower the cassette with wafers into the reflux tank. Start timing the etch.

(10) When the etch is complete, lift the cassette, using the glass rod, and allow the etchant to drip off the wafers and cassette. If any etchant drips on the wet process station surface, rinse it off immediately with copious amounts of water, as EDP stains the plastic. Place the cassette in the rinse tank filled with 1 cm of DI water. The water prevents EDP from collecting in the bottom of the rinse tank. Allow the wafers to cool slightly, to prevent the wafer (and membranes, where applicable) from suffering temperature shock and possible breakage.

(11) Once the wafers have cooled, fill the rinse tank and proceed with the rinse cycle. If you notice a white precipitate on your wafers, this indicates that the solution has been depleted. Making a new batch of EDP should be considered.

(12) If you are not going to use the etchant again within 24 hours, shut off the temperature controller, aspirate the etchant and fill the tank with DI water. Allow the water to sit in the tank for 30 minutes @ 80ºC. Then turn off the temperature controller, aspirate the water and disable the sink. Wipe off the outside of the tank using damp techni-cloths to remove any residue and dry using dry techni-cloths.

B. KOH Etchant

Etch recipes and pertinent information are given in Chapter 1.10. KOH etches silicon dioxide, so it is not a suitable masking material for long etches. Silicon nitride is the preferred masking material.

(1) Rinse the beaker thoroughly and aspirate to make sure it is clean before processing. Always take care not to spray water outside of the beaker since it may ruin the hot plate. Also, make certain that the spin bar and Teflon® bottom screen are inside of the beaker.

(2) With the cassette out of the beaker, add 2 liters of DI water by looking at the 2 liter mark on the beaker. For Critical runs, the DI water may be measured using the graduated cylinder located near the sink.

(3) Turn on the spinner, which is accomplished by pressing the right Megason button, and adjusting the speed using the knob under the heat lamp timer. Typically, the speed is set to 6. If the speed is too high, the spin bar will move away from the center of the beaker and stop. This can be fixed by turning the speed knob to zero, allowing the spin bar to return to the center, and then slowly increasing the speed to the appropriate setting.

(4) Add 1 kg of KOH by simply adding the contents of two 500 mg bottles of KOH (or one 1 kg bottle, if available). For critical runs, the KOH should be weighed on the portable balance inside the sink. Also, other ratios of KOH to water produce different results so see the reference at the end of this manual for more information.

(5) Turn on the hot plate by pressing the right Temp Control button. Typically, the hot plate should be set to 80 deg. The KOH will temperature stabilize at 80 deg between 30 minutes to 1 hour. During the temperature stabilization phase, a visible interface may form between the hot and cold (KOH and water) liquids. Using the cassette handle as a stirrer, this interface may be broken allowing the mixture to temperature stabilize faster.

(6) Load the wafers to be etched in the cassette and place in the beaker. Take care when removing the lid since the KOH solution condenses on it, so tap it a few times to shake off most of the drops and place on a techni-cloth. For critical runs, the cassette can be left in the beaker during the temperature stabilization period, lifted out of the KOH solution so that the wafers can be loaded, and then dropped back into the solution.

(7) Replace the lid and start timing the etch. For critical runs, since the etch rate of KOH is extremely temperature sensitive (a factor of 2 to 3 times for every 10 degrees), the wafer can be rotated half way through the etch. This time can be approximated by assuming that the etch rate of silicon is approximately 1 um/min using the suggested temperature and KOH to water ratio.

(8) Once the etch has been completed, carefully remove the lid and remove the cassette. Let the wafers cool briefly before rinsing by partially dipping the wafers in the rinse tank to avoid breaking membranes. Initiate the rinse cycle.

(9) If you are finished etching, shut off the heater by pressing the Temp. Control button and stop the spinner by pressing the Megason button. Aspirate the KOH, and rinse the tank with DI water and aspirate three times. Disassemble the beaker and plastic collar, wipe down using damp techni-cloths and dry using dry techni-cloths. Disable the sink.

For more information of anisotropic etching a good reference is:

H. Seidel, L. Csepregi, A. Heuberger, and H. Baumgartel, "Anisotropic etching of crystalline silicon in alkaline solutions, I. Orientation dependence and behavior of passivation layers, " J. Electrochem. Soc., vol. 137, no. 11, pp. 3612-3626, Nov. 1990. and

H. Seidel, L. Csepregi, A. Heuberger, and H. Baumgartel, "Anisotropic etching of crystalline silicon in alkaline solutions, II. Influence of dopants, " J. Electrochem. Soc., vol. 137, no. 11, pp. 3626-3632, Nov. 1990.

MOD 22

Plasma Etching of Trenches in Single-Crystal Silicon

Purpose: To etch deep trenches in single crystal silicon using plasma.

Note: This is to etch trenches in single crystal Si, deeper than 3 Mm; when no photoresist mask can be used. For trenches that are less than 3Mm deep, use the standard procedure for Poly-Si etch in lam4 (Chapter 7.4).

Equipment: lam4

Summary:

(1) Mask your sample with oxide.

(2) Program the recipe into the lam (see Chapter 7.4 for details).

(3) Etch the desired length of time.

Detailed Procedure: See lab manual Chapter 7.4 for lam4 operating procedures.

(1) Deposit the required thickness of oxide onto the wafers. The etch selectivity of Si to oxide, using recipe #400 in lam4 (given here), is 20:1.

(2) Coat the wafers with photoresist, expose with the desired mask and hard bake at 120ºC for thirty minutes before the oxide etch.

(3) Etch the oxide in lam2. See Chapter 7.2 for lam2 operating procedures.

(4) Strip the remaining photoresist, after the oxide etch, using standard procedures (MOD 10).

(5) Load the polysilicon etch recipe #400 into lam4.

(6) Delete the overetch step by copying steps seven to six, eight to seven, and nine to eight.

(7) Change the following parameters on the corresponding recipe steps:

RF Top W: 300 (in step 5)

He lamp t: 0 (in all steps)

Time: 60 seconds (in the new step 6)

(8) In step #5, change the "completion" parameter from [TIME & ENDPOINT] to [TIME], by moving the cursor to completion with the arrow keys and entering 1 from the available options.

(9) Program in the desired amount of time based on an etch rate of approximately 6700 A/minute.

(10) The final recipe should look like this:

Single Crystal Silicon Trench Etch

Parameter

#01

#02

#03

#04

#05

PRESSURE (mT)

RF TOP (W)

GAP (cm)

Cl2 (ccm)

O2 (ccm)

He (ccm)

HBr (ccm)

SF6 (ccm)

He clamp t

COMPL

TIME

400

1.00

100

STBL

60 sec.

400

200

1.00

100

TIME

7 sec.

425

0.80

180

400

STBL

30 sec.

425

300

0.80

180

400

TIME

ET sec.

400

1.00

400

TIME

60 sec.

(11) Program in the desired amount of time based on an etch rate of approximately 6700 A/minute.

(12) Run a test sample, patterned with some mask you are using, to check the etch rate.

Note: This procedure works fine for trenches up to 4.5 µm deep.

MOD 23

Plasma Etching of Thick PSG

Purpose: To etch thick films of PSG using plasma.

Equipment: lam2

Summary:

(1) Mask your sample with a double layer of photoresist, hard baking at 120ºC for at least 60 minutes before etching.

(2) Clean the lam chamber using the clean recipe, and one dummy wafer.

(3) Program the recipe into the lam (see Chapter 7.2 for details).

(4) Etch the desired length of time.

(5) Clean the lam chamber again using the clean recipe, and one dummy wafer.

Detailed Procedure:

See lab manual Chapter 7.2 for lam2 operating procedures.

(1) Coat your wafers with a double layer of photoresist (MOD 6), using program #4 on the svgcoat.

(2) Expose your wafer, using 1.5 times the standard exposure time.

(3) Hardbake the wafers at 120ºC for at least sixty minutes before etching (MOD 8)

(4) Load the CLEAN recipe into lam2 and run the recipe using a dummy wafer.

(5) Load the standard oxide etch recipe into lam2.

(6) The following should be the parameters in the etch step of the standard recipe:

Pressure: 2.8 Torr

RF Top: 850 Watts

Gap: 0.38 cm

He: 120 sccm

CHF3: 30 sccm

CF8 = 90 sccm

Etch Time: 01:00

(7) After each etch step, a cool down period is required to minimize resist burning and erosion. Program the step subsequent to each etch step as follows:

Pressure: 2.8 Torr

RF Top: 0 Watts

Gap: 0.38 cm

He: 120 sccm

CHF3: 30 sccm

CF4: 90 sccm

Time: 01:00

(8) Program in the required number of etch/rest cycles necessary to etch your film, based on an etch rate of approximately 6000 A/min (for densified PSG). It is recommended that you etch a test wafer for one minute, with the same pattern, to measure the etch rate for your own process.

(9) Following the final etch step, program the subsequent step as follows (to bring the gap to the nominal value of 1.35 cm):

Pressure: 0.0 Torr

RF Top: 0 Watts

Gap: 1.35 cm

He: 0 sccm

CHF3: 0 sccm

CF4: 0 sccm

Time: 00:10

The final recipe should look like this

Thick PSG Etch

Parameter

#01

#02

#03

#04

#05

PRESSURE (TORR)

RF TOP (W)

GAP (cm)

C2F6 (ccm)

O2 (ccm)

He (ccm)

CHF3 (ccm)

CF4 (ccm)

COMPL

MAX

2.8

0.38

120

stbl or time

30 sec.

2.8

850

0.38

120

time

60 sec.

2.8

0.38

120

time

60 sec.

2.8

850

0.38

120

time

1.35

120

time

10 sec.

Notes:

(1) Steps 2 and 3 should be copied for as many times as necessary to etch your particular film.

(2) If you need to end point your process at the final step, set the "End Point Menu" accordingly. See Chapter 7.2 of the Microlab manual for details.

(10) Run the wafer.

MOD 24

Contact Etching

Purpose: To etch contact openings prior to metallization.

Equipment: lam2, sink8

Summary:

(1) Define the contact openings using standard photolithography processes. Be sure to descum and hard bake at least 30 minutes before etching.

(2) Measure the initial oxide thickness in the contact openings using the nanospec (you'll probably have to use the 100X lens for this measurement - see Chapter 8.33 for details). Add 2000A to this value to compensate for the extra oxide thickness over the polysilicon gate region.

(3) Clean the lam chamber using the clean recipe.

(4) Program the recipe into the lam (see Chapter 7.2 for details).

(5) Etch the desired length of time.

(6) Wet etch to remove remaining oxide.

(7) Clean the lam chamber using the clean recipe.

Detailed Procedure:

See lab manual Chapter 7.2 for lam2 operating procedures.

(1) Load the CLEAN recipe into lam2 and run it using one of the clean dummy wafers for lam2.

Note: DO NOT have the loading cassette ready when loading the recipe, if you are going to modify the recipe. The etch process will start as soon as the recipe is loaded, if the loading cassette is ready.

(2) Load the standard oxide etch recipe into lam2.

(3) Calculate the etch time required to etch through your film based on an average etch rate value of the last process monitor test posted on the lam2 comments. If your film thickness exceeds 1 Mm, see MOD 23 for instructions.

(4) Enter the calculated etch time into the L/S (low selectivity) etch step of the standard recipe:

Pressure: 2.8 Torr

RF Top: 850 Watts

Gap: 0.38 cm

He: 120 sccm

CHF : 30 sccm

CF 3= 90 sccm

Etch Time: Variable

(5) After the L/S etch step, a 50% H/S (high selectivity) overetch is required to guarantee that the contacts are clear. Program this etch step as follows:

Pressure: 3.0 Torr

RF Top: 700 Watts

Gap: 0.40 cm

He: 110 sccm

CHF : 35 sccm

CF : 30 sccm

Time: 50% Overetch

(6) Run the wafer.

(7) Measure remaining oxide thickness in contact openings. It should be less than 200A. If it is more, plasma etch again for time required to clear remaining oxide. The color in the contact area will be highly reflective (white) if you've cleared the oxide. Other colors indicate the presence of oxide or other films.

(8) Without removing the photoresist, wet etch the wafers in 5:1 BOE (sink8) until contact openings enlarge to within 1Mm of the edge.

MOD 25

"Show" Wafer Process

Purpose: To make 4-inch or 6-inch "show wafers” for special occasions.

Equipment: sink6, tystar2, 3, or 4, primeoven, svgcoat6, svgcoat1 or 2, ksaligner, svgdev, svgdev6, oven-vwr, sink8, sink5, matrix, technics-c, cpa

Summary:

4” wafers:

(1) Grow thermal oxide on bare Si wafer.

(2) Make a transparency to use as a mask.

(3) Prime wafers in primeoven.

(4) Spin coat wafers with 1 - 2 microns photoresist using standard coat programs at svgcoat1 or svgcoat2.

(5) Expose wafers at the ksaligner.

(6) Do a post-exposure bake on the svgdev.

(7) Develop wafers on the svgdev.

(8) HF etch in sink8.

(9) Strip off photoresist.

6” Wafers:

(1) Grow thermal oxide on bare Si wafer.

(2) Make a transparency to use as a mask.

(3) Prime and spin coat wafers with 12000A I-line resist on the svgcoat6.

(4) Expose wafers at the ksaligner.

(5) At svgdev6, do a post-exposure bake and develop the wafers.

(6) HF etch in sink8.

(7) Strip off photoresist.

Detailed Procedure:

(1) Decide what color you would like for the background. Royal blue is nice and you can use the 1000 A oxide wafers as the substrate. Purple is also nice-this can be done with ~5000A of thermal oxide. You can check the oxide color chart in the Tylan notebook (VLSI area) for other color/thickness ideas.

(2) Perform a standard pre-furnace clean of the wafers (MOD 1)

(3) Grow thermal oxide on the bare silicon wafers in one of the tystar wet oxidation furnaces (tystar2, 3, or 4). Use a wet oxidation recipe, as it is quicker. To find time required for desired oxide thickness, see an oxide growth chart (in Tylan notebook, VLSI area) or use an oxide growth calculation program (several are available through the internet).

(4) Use computer software such as Word, Powerpoint, or Photoshop to create a document with the desired words and/or images. If possible, images should be converted to black-and-white. Half-tone conversion is best if available, but good results have also been obtained with grayscale.

(5) Print the document onto a transparency, which will be used as a mask. For this standard process module, with positive resist over dark blue 1000A oxide on lighter silver-gray silicon, your transparency should be a positive mask, not a negative one. You can use a laser printer to print directly on the transparency, but be sure that the printer and transparency are compatible or the transparency could melt. An inkjet printer has also been used successfully.

(6) Dehydrate and prime the wafers. For 4” wafers, use primeoven. 6” wafers are primed as part of the svgcoat6 track.

(7) Coat 4” wafers with G-line resist (OCG 825, program #2, svgcoat1 or 2). 6” wafers on svgcoat6 can be primed and then coated with 12000A I-line (OiR 897 10i, program #3). Include a 90 C soft bake after coating.

(8) Obtain two clear glass mask plates larger than the wafer size. (The idea is that you will create a “sandwich,” with the transparency resting directly on top of the coated wafer and with both these two in between the two glass mask plates, in other words: plate--wafer--transparency--plate.)

(9) Cut the transparency to fit the mask plate, and tape it tightly to the top plate at the corners. The transparency should be affixed such that it will not print backwards on the wafer.

(10) At the ksaligner, install the mask holder frame appropriate to the mask plate size you are using. Leave a dummy mask in the ksaligner mask frame, so that it doesn’t get vacuum errors. Do not load your masks in the frame, and do not load any wafers into the ksaligner.

(11) Take the mask plate which does not have the transparency affixed, and place it carefully into the recessed area on top of the mask frame. The vacuum leaking from below will probably hold the plate somewhat firmly to the frame, this is normal.

(12) Place a resist-coated wafer on top of this plate.

(13) Place the mask with the transparency on top of the wafer, transparency side down.

(14) Center the mask plate over the wafer so that the image and/or text is centered over the wafer (make sure that the transparency side is in contact with the wafer). Be careful not to scratch the resist while centering the transparency, try to keep it lifted it a bit while moving it.

(15) Before exposing, take a moment to make absolutely certain that there is enough clearance for the exposure assembly to slide out without hitting anything on the mask frame. If it crashes, it could damage the ksaligner.

(16) Take note of the seconds on a watch or clock, and press the button “lamp test”. This causes the exposure assembly to slide out and flood the mask frame with UV light, but without moving the wafer chuck.

(17) Exposure should be about 10-20 seconds. The ksaligner does not run the timer during a lamp test, so you will need to count the seconds yourself with a watch or wall clock. If you are printing a grayscale image, you may need to do a test to find the optimal exposure time for bringing out the grayscale.

(18) Press “lamp test” again to stop UV exposure. The exposure assembly will slide back in.

(19) Do a post-exposure bake and then develop. For 4” wafers use the svgdev, for 6” wafers use svgdev6.

(20) Optional: Hard bake the wafers in a 120ºC oven (oven-vwr in Y2) for 30 minutes.

(21) Etch the wafers in the buffered HF bath at sink8, until the solution readily beads off of any areas not coated with resist (about a minute or less). Rinse in QDR, and dry in spindryer.

(22) Strip the photoresist in sink5 PRS-3000, matrix, or technics-c (MOD 13). Voilà! You now have a beautiful work of art!

Optional: This process can be modified for other films, for example 1000A – 2000A Aluminum sputtered on top of 1000A oxide in cpa can be etched in aluminum wet etch at sink8 without attacking the oxide, yielding bright white-silver positive areas on dark blue negative areas (this requires a negative mask if any images are present.)

MOD 26

LAM Monitors

General Operation and Procedures

Measurements:

► Measure in following positions: T(1,4); C(4,4): F(8.4); L(4,1); R(8,1). (Pick position in the die closest to the center of wafer.)

► Measure in the same spot before and after etching.

Calculations:

► For Lams 1, 2 and 4:

Etch Rate: (PreEtch - PostEtch) / 0.5 A/min

For Lam3:

Etch rate = Step Height / 0.5

► %Unif: (Max - Min) / (Max - Min)

PC Screen resetting on Lam 1 and 2:

► ctrl-alt-delete (may need to do it twice)

► F1

► password: lam1-pc

Y → enter

choose Lam1 in small window → enter

► esc (proper screen should appear)

Purpose: Nitride etch Monitor Test

Equipment: LAM 1

Time of Execution: 30 seconds per wafer

Detailed Procedure:

(1) Measure nitride thickness on Nanospec, Program#6, Lens:10X, Tox:1000A

(2) On PC key board hit

- ctrl-esc → choose Lam1 → esc

(3) Load: O2clean.RCP (program4) and load 1 bare Si wafer. → Press START.

(4) Load NITSTN1.RCP (recipe #3) and edit as follows:

NITSTN1.RCP (recipe #3)

Recipe

Step #1

Step #2

PRESSURE

RD POWER (W)

GAP

He (sccn)

SF6 (sccm)

TIME

COMPL

375

1.35

175

20 sec.

TIME

375

150

1.35

175

30 sec.

TIME & ENDPT

(a) Eliminate overetch step.

Copy step #5 → #3 go to copy press field select.

Copy step #6 → #4 go to copy press field select.

(b) Go to Parameters: [MACHINE] should be flashing and press field select.

Change endpoint #1. Set normalize to 10 sec. and trigger to 25%.

(5) Press STATUS --> go to monitor hit esc --> enter to see the plot come up on the screen.

(6) Load 3 dummies + 3 test wafers with CMOS active pattern on nitride (tylan9) over 1000A of oxide.

(7) Press START (on first 3 dummy wafers press field select once power becomes stable - make sure manual endpoint is flashing).

Watch RF Power < 10.

Then run 3 work wafers.

Specification:

Etch rate: 1000 A/min+/-10%

w/in W Unif: 10%

W-W Unif: 10%

Note: Nitride with lower stress (tylan 18) will have lower etch rate.

Purpose: Oxide etch Monitor Test

Equipment: LAM 2

Time of Execution: 30 seconds per wafer

Detailed Procedure:

(1) Measure oxide thickness on Nanospec, Prgram#1, Lens:10 x

(2) On PC key board hit

- ctrl-esc → choose Lam2 → esc

(3) Load: Clean recipe and load 1 wafer with oxide. → Press START

(4) Load Recipe #b (SIO2ET) and edit as follows:

SIO2ET (Recipe b)

Recipe

Step #2

PRESSURE (TORR)

RD POWER (W)

GAP

He (sccn)

CHF3 (sccm)

CF4

TIME

COMPL

2.8

850

0.38

120

30 sec.

TIME & ENDPT

(a) Step#1 change [COMPL] to [time only] time = 00:20 min:sec

(b) Edit recipe to eliminate overetch step.

Copy step #5 → #3 go to copy press field select

Copy step #6 → #4 go to copy press field select.

Notify endpoint #1. Set normalize to 10 sec and trigger to 25%.

(5) Press STATUS → go to monitor hit esc → enter to see plot come up on the screen.

(6) Load 3 dummies + 3 test wafers with resolution mask pattern on 6000 A of thermal oxide.

(7) Press START (on first 3 dummy wafers press field select once power becomes stable - make sure manual endpoint is flashing).

Check Pressure: 2.8 mT

Temp < 20

Specification:

Etch rate: 5800 A/min ± 10%

w/in W Unif: 20%

W-W Unif: 20%

Note: For better uniformity use Press: 2.9T, Gap: 0.42, RF: 750, He: 110, CF4: 95 sccm, CH3: 30

Purpose: Aluminum etch Monitor Test

Equipment: Lam 3

Time of Execution: 30 seconds per wafer

Detailed Procedure:

(1) Make sure Lam4 is not in use when using recipe w/ CL2.

(2) Load 3 dummies + 3 test wafers with resolution mask pattern on 7000 A of Al over 1000 A of thermal oxide.

(3) Put module in → Press load

(4) Modify recipe:

Parameters

Step #1

Step #2

Step #3

PRESSURE (MTORR)

RF POWER (W)

BCl3 (sccm)

Cl2

CHF3 (sccm)

CF3 (sccm)

COMPL

MAX

250

time

20 sec.

250

time

30 sec..

100

Recipe

10 sec.

(a ) Step#1 change [COMPL] to [time only].

time = 00:20 min:sec

(b) Step #2 change time to 30 sec.

Copy step #4 → #3 go to copy press field select

Copy step #5 → #4 go to copy press field select.

(5) Press STATUS

(6) Press START (on first 3 dummy wafers press field select once power becomes stable - make sure manual endpoint is flashing).

(7) To go to [Air Lock] go to recipe press field select when [Reactor] is flashing.

Specification:

Etch rate: 5000 A/min ± 10%

w/in W Unif: 20%

W-W Unif: 20%

Measurements:

► Strip PR in PRS2000 and measure step height in AS200

Purpose: Polysilicon etch Monitor Test

Equipment: LAM 4

Time of Execution: 30 seconds per wafer

Detailed Procedure:

(1) Measure poly thickness on Nanospec (Program#4), RI = 3.7, Tox:1000 A and Lens:10 x.

(2) Make sure Lam3 is not in use when using recipe w/CL2.

(3) Load 3 dummies + 3 test wafers with CMOS-Poly mask pattern on Polysilicon over 1000A of thermal oxide.

(4) Status - Check for idle condition.

(5) Check alarm and clear them.

(6) Load Recipe: Select 1) Recipe then press LDNEW → 400 → load

(7) Modify recipe as follows:

Select recipe and press copy to insert step between 4 and 5 by going to

COPY 5 to 6 → ENTER

COPY 4 to 5 → ENTER

Modified Polysilicon Etch Recipe 400

Parameter

Step #5

Step #6

PRESSURE (MTORR)

RF TOP (W)

GAP (cm)

Cl2 (sccm)

He (sccm)

SF6 (sccm)

HE CLAMPT

COMPL

TIME

CHANNEL

DELAY

NORM SEC

NORM VALUE

TRIGGER %

425

0.80

400

8.0

[STABL]

15 sec.

425

0.80

180

400

8.0

[ENDPT]

30 sec.

5000

50%

(8) Go back to Menu, press STATUS

To see graphics press PLOT → Trend

Specifications:

Etch rate: 4900 A/min ± 10%

w/in W Unif: 5%

W-W Unif: 10%

Note: When using endpoint copy start by 11 → 12, 10 → 11, 4 → 5

Change step #5 Cl2 = 0

Step #6 Trigger = 90%

Time = 90 sec It will endpoint by itself.

MOD 27

OCG 825 Reversal Image Process

(1) Standard clean wafers.

(2) Spin dry.

(3) Standard dehydration bake wafers in furnace tube for 10 minutes.