Metalica Sputter System
Table of Contents
1. Background Information
2. Equipment Description
3. Process Capabilities
4. Operating Procedures
5. Process Monitoring and Machine Qualification
6. How To Become an Metalica User
The Metallica sputtering system
Location of Metalica in the SNF
1. Background Information
1.1. Metal Sputtering
Sputtering is a process whereby atoms are ejected from a solid target material due to bombardment of the target by energetic ions. In the Metalica system, a bias is placed between the sputter target and the substrate. This causes an ionized plasma of argon to form in the space between the two surfaces. The bias accelerates the ionized (and hence charged) argon atoms towards the sputter target, directed partly by the magnetic field of a magnet located beneath the target. When the argon atoms strike the target surface, they can eject small clumps of target atoms which fly through the plasma and deposit on the substrate. Because the distance between the target and the substrate is small relative to the target area, and material can be deflected by collisions in the plasma, these clumps arrive at a variety of different incidence angles and can coat the sidewalls of substrate features. Once they impact on the surface, target clumps stick to the surface in a relatively uniform, pinhole free layer.
1.2. Sputtering and Vapor Pressure
Chamber pressure during sputtering can influence the quality, uniformity, con-formality, and stress of the resultant film. However there is a limited process window available – if the pressure is too high, the electrodes may arc, eliminating the potential needed to maintain a sputtering plasma. If the process pressure is too low, a plasma may not form at all, or the deposition rate can be significantly reduced. The recommended pressure during sputtering is 5e-3 Torr.
In general, the adhesion of sputtered films is better than that of evaporated films using the same film/substrate combination. Adhesion depends on the film material, the substrate material, and the stress in the film. Certain applications require films such as Au and Ag may need a thin “adhesion layer” to be added first. Cr is the most commonly used adhesion layer when depositing Au or Ag on Si or quartz substrates.
2. Equipment Description
Metalica is a metal sputter deposition system that was built in-house to provide flexible processing options for non-CMOS compatible materials. The Metalica consists of four primary sub-systems: the bell jar process chamber to hold the targets and substrates, two independent magnetron S-gun sputter sources with their associated DC power supply, a vacuum system to evacuate the process chamber, and an argon mass-flow controller to regulate the process chamber environment.
2.1. Sputter Targets
There are two S-gun magnetron sputtering sources in the chamber. Each can hold a different 1" DIA X 0.125" thick target. Silver heat sink compound is applied on the back of targets (between S- gun) for keeping the targets cool and electrical contact. No backing plate is required as long as targets are of these dimensions.
Current materials available include Ag, Al, Au, Cr, Cu, Mo, NiCr, Pd, Pt, Ti, W and TiW 90/10wt%. Deposition rates are relatively slow. Au sputters at approximately 600 A/min, Ti and Cr at approximately 40 A/min, Cu at 200 A/min (one wafer at a time) using approximately .250 A sputter current
Most conductive metals can be sputtered with the 600W DC power supply. Insulating materials cannot be sputtered with the current (DC) power supply. Ferromagnetic materials (Co, Ni, Fe) cannot easily be sputtered. Ferromagnetic materials have been sputtered by bonding a thin sample of the desired material to an aluminum backing plate, however very short target life and a difficult-to-sustain plasma make this process difficult.
Metalica has no substrate temperature control. It can not reverse bias the substrate to do a pre-deposition back-sputtering clean, or back-sputtering during deposition to improve film uniformity or adhesion.
2.2. Vacuum System
A medium to high vacuum system capable of base pressures in the 1e-7 Torr range. Some seals are copper crush, but bell-jar and pump flanges are o-ring type seals. Roughing is via an Edwards XDS35I dry pump next to the system. High-vacuum is achieved using CT 8F Cryogenic on-board pump. The chamber is rough-pumped above the gate valve to approximately 150 mTorr via the rough pump. The valve to the rough pump is manual, allowing slow, controlled opening. Rough pumping takes approximately 3-5 minutes. After roughing the chamber, the roughing valve is closed. The gate valve should now be opened, exposing the chamber to the high vacuum pump. Roughing pressure is measured by a t/c vacuum gauge (10-1000 mTorr) type digital gauge. High vacuum pressure is measured by a digital ionization gauge ranging from 1e-8 Torr up to 5 mTorr. This gauge is also used to measure the chamber pressure during deposition (typically 5mTorr). The chamber is rather large (100L), and therefore requires long pump down time (~1 hour) for reasonable base pressures (~1e-6 Torr range). An unused ion pump still exists below the gate valve but was disconnected after the system was converted from a flash evaporator to a sputter system. Cryo-pump regeneration is required every few months (depending on usage). Regeneration takes approximately fours to complete. The Cryo-pump is regenerated manually, when the cryo head temperature rises above 15 K.
2.3. Process Gas Flow
Argon gas flow is measured and set by a Tylan MFC system. Ar flow is approximately 30 sccm when using the CT 8F on board cryo-pump. The gate valve is left wide open for CT 8F on board sputtering, and pressure is controlled by altering the MFC setting. For cryo-pump sputtering, the gate valve is throttled to achieve the proper chamber pressure.
2.4. Metalica Interlocks
This system has very few interlocks and is therefore easily susceptible to severe damage resulting from improper operation, and also has the potential for serious injury to the operator.
2.4.1. The power supply is interlocked to the lab computer (Coral) so that it cannot be turned on by unauthorized personnel.
2.4.2. The bell jar hoist is interlocked to a vacuum switch so that the hoist motor may not be activated while the chamber is under vacuum to prevent damage to the motor and bell jar.
2.4.3. The power supply is internally interlocked so that it will turn off if a plasma fails to strike or the set point cannot be reached. The ion gauge will automatically turn off if the pressure is too high.
3. Process Capabilities
3.1. Cleanliness Standard and Allowed Materials
The Metalica is in the GOLD contaminated equipment group. Substrates can come from any piece of equipment in the CLEAN, SEMICLEAN, or GOLD groups. Most vacuum-compatible materials are allowed in this system. Photo-resist is acceptable on wafers as long as it has been hard baked. Kapton tape is allowed for holding smaller samples to holder-wafers. Sodium and potassium-containing glasses are allowed. Most semiconductor materials, and high-vacuum compatible plastics are also acceptable. After deposition of ANY material in the Metalica (including aluminum), substrates are considered contaminated with gold and non-standard metals and can only go into other GOLD contaminated equipment.
3.2. Available Materials
Targets are metal cylinders 1" in diameter by 1/8" thick. The following sources are available in the Metalica:
Ag, Al, Au, Cr, Cu, Mo, NiCr, Pd, Pt, Ti, W
Metalica can not sputter semiconductors (silicon, germanium, etc) or dielectrics. Please contact firstname.lastname@example.org about introducing a new source for the Metalica.. NOTE: it is recommended you discuss you process with a member of the quality circle prior to contacting specmat.
3.3. Deposition Thickness and Uniformity
Metalica can provide thin films (<1μm thick) of various metals with NO precision measurement of film thickness. All deposition thicknesses are estimated by multiplying an approximate deposition rate (depending on the material and supply power) by the deposition time (which is controlled manually). Metalica should not be used for depositing thick films (>1μm) due to time, wear on the system, and adhesion issues. The maximum continuous sputter time is 10 minutes on a single source gun. After this, the source gun must be allowed to cool down for 5 minutes before sputtering again with the same gun. Deposition thickness of precious metal targets (Au, Pt, Pd) should not exceed 2000 A due to material cost.
The deposited film is reasonably conformal and unsuitable for lift-off metal patterning or shadow masking. Uniformity values are highly dependent on deposition time, pump down pressure, argon flow, and target wear. Cross-wafer uniformity is roughly 10% and wafer-to-wafer uniformity is roughly 10%. Measuring step height on the deposited wafer (by using kapton tape as a temporary mask, or subsequent etching of the metal film) is the only reliable way to obtain film thickness data. Using a piece of kapton tape on a blank wafer and measuring the step height under identical deposition conditions is another less accurate option.
3.4. Equipment Capacity
The system can hold up to four 4" wafers. Usually one of these wafers is a dummy wafer for pre-cleaning targets and bring up the power prior to actual deposition. Smaller samples may be mounted to a 4" wafer using kapton tape or hard-baked resist. Loading, operation, and unloading is manual. Samples larger than 4” are not possible. Wafers are sputtered one at a time, and rotated into position by a manual control. It is therefore possible to deposit different thicknesses (or different materials, in a different order, etc.) onto the four wafers in the same run. This is a sputter up configuration, so wafers are held above the source, facing down. A small (1/8" to 1/4") annulus on the perimeter of the wafer will not be sputtered due to the tooling that holds the wafer.
4. Operating Procedures
4.1. Safety Warnings
While relatively safe to operate, this machine can cause injury if not used in a cautious manner.
4.1.1. The chamber can crush hands if lowered onto them.
4.1.2. The operator can inhale particles while handling chamber parts and is advised to cover the mouth.
4.1.3. While the amount of x-rays escaping the machine should be safe, we do not have current measurements of the level of exposure. Do not operate this machine if you could be pregnant.
4.1.4. DO NOT ATTEMPT TO REPAIR THIS SYSTEM YOURSELF. Please contact maintenance. For certain issues (link to trouble shooting lis), you may contact the super-user outside of staff hours.
4.2. Before Operation
Check that the desired deposition can be done within normal power and sputtering times. Metalica is not designed to sputter thick films or large numbers of wafers. It is a research tool not intended for production.
Power ranges from 50 – 100 watts; maximum sputter time is 10 minutes on a source gun, when >50 watts of power used. Cool down for 5 minutes before sputtering again with the same S-GUN. Limit by poor cooling. The 5 minute cooling is good for a couple of 100 nm, but if thicker use 10 minute cool down before starting deposition with the same source gun.
4.3. Initial system checks
4.3.1. Verify that the system is not enable on Coral or does not have a tag indicating it is “IN USE”.
4.3.2. Verify on the panel “Metalica Target Tester” that the toggle switch is in the up position (Test Mode).
4.3.3. Verify that the gate valve to the cryo-pump is fully and snugly closed (clockwise.)
4.3.4. Verify that the roughing pump valve is snugly closed.
4.3.5. Verify that the cryo-pump temperature on the control box is between 11 and 14K.
4.3.6 Verify that the toggles to Roughing valve and Cryo Regen switches are in the down position (off).
4.3.7. Verify that target materials and supplies are available.
4.3.8. If the above conditions are met, enable Metalica on Coral and proceed.
4.4. Venting and Loading Samples
CRITICAL: The gate valve and the roughing valve must NEVER be open at the same time or severe contamination will result when roughing oil is sucked into the chamber and cryo-pump.
4.4.1. Turn the tag to “IN USE” on the front panel of the machine.
4.4.2. Open the N2 vent line fully and wait until the hoist “READY” light is lit (~ 1 min.), plus 1 additional minute. Verify that it appears fully vented by looking for decompression of the seal ring.
4.4.3. Lift the bell jar by tapping on the hoist up button initially (the seal ring may come off if the bell jar is lifted too quickly). Use one hand to steady the bell jar in order to avoid jolting the inner assembly. Guide and adjust the seal ring if necessary.
4.4.4. Turn the N2 purge valve clockwise to close.
4.4.5 Remove Platen, Chimneys, Cathodes, and Anodes (two pieces) and place on clean wipes.
4.4.6. Inspect the parts, targets and surrounding areas for target wear, flaking, or other problems.
4.4.7. Using a small amount of silver paste, apply with a swab stick to the back of the 1-inch target. All silver color metal targets have the symbol scribed on the backs (Ti, Pt, Cr, etc.).
4.4.8. Load target(s), switching toggle to desired target (Target A or Target B) on the panel “Metalica Target Tester”. An amber light will appear on the selected Target A or B.
4.4.9. Keep an eye on the “Target Status”, while assembling target(s). If green LED is lit, it is OK to continue. If the red LED is lit, the resistance is below 200K Ohms and must be corrected before continuing.
4.4.10. Load samples face down (maximum 1 dummy wafer plus 3 real wafers) using tweezers. Be sure nothing hangs below the level of the target shields, and the alignment mark is aligned for viewing through the porthole. NOTE: A dummy wafer may be loaded to receive the first “dirty” layer sputtered from the target and to bring up the target power. Test wafers may also be loaded with Kapton tape or lift off (TW with 3612) for measuring step height on the Contaminated Gold Alpha step 500 profilometer.
4.4.11. Rotate the wafer pallet to the first desired position (numbered 1-4.
4.4.12. Lower the bell jar with the hoist down toggle button.
4.4.13. Inspect through view port with flashlight that you can see alignment marks to assure you can see when system is pumped down for deposition.
4.4.14. Switch roughing pump toggle up (red and green light will come on) before opening rouging valve.
4.4.15. Open roughing valve, pump will be nosy at first and taper off in one minute.
4.4.16. When the base pressure (as monitored on the t/c vacuum gauge on the front panel) reaches 150 mT or less, close the roughing valve.
4.4.17. Switch roughing pump toggle down (green light will go off, red light will take up to one minute longer to go off).
4.4.18. Slowly open the gate valve fully and verify that the pressure drops to the right dramatically.
4.4.19. Press the ON button on the ion gauge, than press IG1. Be sure it is never exposed to a pressure above cry-pump compatible levels. It should automatically shut itself off (Ion gauge).
4.4.20. Let the chamber pump until desired base pressure is reached, about 1.0e-6 Torr or below (1 hour depending on the amount of time the chamber was open to atmosphere while loading targets and substrates)
4.5.a Record pressure and time on the log sheet.
4.5.1. On the “Metalica Target Tester” panel switch toggle to the down position “HV ENABLE”.
4.5.2. On the “Metalica Target Tester” panel switch toggle to desired Target (A or B).
4.5.3. Close the gate valve all the way, then re-open approximately ~2 1/4 full turns +/-.
4.5.4. Turn the Argon MFC controller to position 2.
4.5.5. Turn the gas inlet T valve toward Argon (quarter turn counter clockwise).
4.5.6. Let the pressure stabilize (IG1 and Ar MFC ~30 sccm) and then adjust pressure to about 5e -3 Torr by adjusting the gate valve slightly.
4.5.7. Turn the knob on the right side of the Bell Jar to rotate either to a “dummy” wafer, or a “real” wafer over the correct source gun. Using a flashlight and looking in the front view port will help with proper wafer alignment.
4.5.8. Turn on the power supply unit by pressing the POWER button.
4.5.9. Select power set-point by toggling the set-point to watts (50 – 100 watts) and turning the large dial below the set-point to the desired power.
4.5.10. Double check that the correct source gun (A or B) is selected and that your wafer is over the selected source gun. Never sputter without a wafer or sample overhead. Calculate deposition rate and set manual timer for the desired deposition time. The maximum continuous deposition is 10 minutes, after which a 5 minute cool down period is required before sputtering with the same source gun.
4.5.11. Press the OUTPUT ON button on the power supply to start sputtering. Adjust power if needed to the desired number. Plasma should strike within a few seconds. Note: if the power is to be turned off for any reason, one can simply leave the power set-point as is. The controller will safely ramp up to the setting safely.
4.5.12 Check for arcing or other probes during sputtering.
4.5.13. When deposition is complete, press OUTPUT OFF on the power supply.
4.5.14. Record deposition material, power, current, and time values in logbook, checking that power, voltage, and current readings are similar to previous runs. If not notify Staff.
4.5.15. Repeat this operation with source guns A and B until all samples are sputtered.
4.5.16. Press POWER button to power off supply.
4.6. Chamber Venting and Unloading
4.6.1. Turn off the ion gauge by pressing the ON button.
4.6.2. Close gate valve by turning it all the way clockwise.
4.6.3. Turn the gas inlet T valve toward “Argon Closed” (Quarter turn clockwise).
4.6.4. Turn Argon MFC switch to OFF.
4.6.5. On the “Metalica Target Tester Panel”, switch toggle to the up position “Test”
4.6.6. Verify that the roughing valve and gate valve are closed (clockwise).
4.6.7. Open the N2 vent line and wait until the READY light is lit (~ 1 min.), plus 1 additional minute. Verify that it appears fully vented by looking for decompression of the seal ring.
4.6.8. Lift the bell jar by tapping on the hoist up button initially (the seal ring may come off if the bell jar is lifted too quickly). Use one hand to steady the bell jar in order to avoid jolting the inner assembly. Guide and adjust the seal ring if necessary.
4.6.9. Turn the N2 vent valve clockwise to close.
4.6.10. Inspect the targets and surrounding areas for target wear, flaking, or other problems.
4.6.11. Unload samples and targets. Clean silver paste from the back of targets, cathode and magnetic tables using IPA from the wbsolvent bench and a clean wipe.
4.7. Shutdown Procedure
4.7.1. Lower the bell jar with the hoist down toggle button
4.7.2 Switch roughing pump toggle up (red and green light will come on) before opening rouging valve.
4.7.3 Open roughing valve, pump will be nosy at first and taper off in one minute.
4.7.4. When the base pressure (as monitored on the t/c vacuum gauge on the front panel) reaches 150 mT or less, close the roughing valve. DO NOT use the cryo pump unless another batch of wafers need to be pumped down for deposition.
4.7.5. Verify that the system is in the following STANDBY condition:
126.96.36.199. Gate valve closed
188.8.131.52. N2 vent valve closed
184.108.40.206. Argon valve closed
220.127.116.11. Power supply off
18.104.22.168. Ion gauge off.
22.214.171.124. Toggle in up position on the “Metalica Target Tester panel”
4.7.6.. Measure the step height of the test wafer with the Gold Contaminated Alpha Step 500 and record in logbook. Depositions of less than 500A may be difficult to measure accurately with the lift-off test wafer or tape method. Please enter data as soon as it is known.
4.7.7. Finish filling out the logbook.
4.7.8. Record any “Metalica” problems on Coral and in logbook.
4.7.9. Turn the tag to “EMPTY” on the front panel of the machine.
4.8. Disable “Metalica” on Coral.
5. Process Monitoring and Machine Qualification
5.1. Metalica Qualification Run
To ensure continued machine performance the Metalica is regularly put through a qualification procedure meant to capture basic functionality and potential errors in the system.
The machine qualification is ran once per month under normal operation. This test is performed down quarterly. Additionally a qualification run is performed after any major maintenance to the system. The resulting test data will be placed in a binder in the lab. At this time the data is from Prometrix (sheet resistivity) and in the future it will be converted to thickness in Angstroms .
A deposition of approximately 100nm of Al is performed under the following conditions:
126.96.36.199. The qualification run is performed by the quality circle to eliminate the user-to-user variation inherent in the system as much as possible.
188.8.131.52. Load Al target in s-gun A (or B alternating each month) using silver paste to ensure good thermal and electrical contact.
184.108.40.206. One “dummy” wafer and one diff clean 100 Si test wafer are loaded into the 1 and 2 slots of the wafer platen. Wafers should be oriented so that the Bottom (wafer flat) is towards the outside of the chamber.
220.127.116.11. Pump down chamber to a base pressure of 1.0e-6 Torr, or as close as possible.
18.104.22.168. Clean the targets by depositing Al (1 minute at 100W) onto the dummy wafer.
22.214.171.124. Deposit Al on wafer 2 and / or 3 position (10 minutes at 100W).
126.96.36.199. After deposition, the test wafer are tested for sheet resistance in the Prometrix, and (after metal etch) step height is recorded using the gold contaminated Alpha Step.
5.1.3. Qualification Notes
188.8.131.52. The sheet resistance measurement should capture either deposition rate inaccuracies or contamination.
184.108.40.206. The profilometer measurement determines if the layer deposited is of an appropriate thickness but of incorrect conductivity, to differentiate between errors in deposition rate and contamination.
220.127.116.11. Aluminum is used because of their substantially different deposition rates in the Metalica system. The Metalica utilizes too many different sources to make it feasible to check each one. But if one material can be deposited correctly and accurately, the basic functionality of the tool has been demonstrated.
Performance of the qualification procedure is primarily the responsibility of the process members of the tool’s quality circle. The super-users are available to assist in times of need with running, testing, and reporting of the qualification procedure.
5.1.5. Recommendation to users with critical processes
If a user has a critical deposition it is in their best interest to perform their own run of the qualification procedure before entrusting their precious process wafers to the Metalica. A run on the Metalica is time consuming and the schedule is often very full, so appropriate “cost-benefit analysis” should be considered.
5.2. Machine Status States (Red, Yellow, Green and What They Mean)
The Metalica is in red-light condition and should not be used. It can be in this state for any of the following reasons
18.104.22.168. The machine failed to pass the qualification procedure within spec
22.214.171.124. The machine is currently under repair
126.96.36.199. The machine will not pump down to a base pressure below 1.0e-6 Torr in a usable amount of time.
188.8.131.52. A plasma will not strike during sputtering, or arcing occurs on two sequential runs with different targets and substrates.
The Metallica is in yellow-light condition and should only be used after careful consideration of posted comments and evaluation of likely effect to your process. NOTE: if the machine is in Yellow or Green it has successfully passed the qualification procedure in the most recent test.
184.108.40.206. Problems with measured conductivity or thickness achieved for any deposited metal as compared to the crystal monitor.
220.127.116.11. Excessive pump down times.
18.104.22.168. User reported problems or issues.
Te Metalica is in green-light condition and is free of any known problems. NOTE: this does not mean that any specific ultra-critical process will work immediately without flaw, and users should take appropriate care with their wafers.
5.3. Process Monitoring Results
Below is profilometric thickness measurements of qualification runs for the first 3 months of 2008.
6. How To Become an Metalica User
6.1. Read the Metalica operating procedures thoroughly.
6.2. Contact SNF training contact on the Equipment Summary page to register for a training session.
6.3. Once training has been completed and the Trainer and Trainee both feel comfortable that the Trainee can run Metalica on his/he own, the trainer will qualify the trainee.
6.4. Upon passing the final training exam you have full access to use Metalica. Marginal passing performance of the final exam can result in a request to operate the tool only during staff hours or coordinated with a super-user presence for your first operation.