Using MTL » Fabrication » Standard Operating Procedures

E-beam (EML)
STANDARD OPERATING PROCEDURE
CORAL Name: e-beamEML
Model Number: --
Location: EML
What it does: metal evaporator
Introduction:

--

Safety:

(A) Emergency Shut-off Procedure

In case of an emergency, such as arcing in the vacuum chamber, do the following:

  1. Shut of the MAIN and CONTROL circuit breakers on the PAK-8 power supply.
  2. Close the HiVac valve.
  3. Find MTL Staff.

(B) Warnings about this system

You can easily cause this piece of equipment significant damage costing thousands of dollars and months of downtime.  There are few interlocks and it lacks many of the “idiot-proof” safety features that one would take for granted in a machine like this.  The three potentially most damaging operationsare completely unprotected. 

  • Only one valve on the front panel should be open at a time. Opening the “HI-VAC” valve while the “ROUGHING” valve is open will cause the Cryo pump to evaporate the oil from the mechanical pump, contaminating the entire chamber, as oil boils at 10-4 Torr.  This will blow the cryo pump.  Opening the “VENT” before the “HIGH VAC” valve has completely closed will cause the cryo to fail.
  • Always check that the chamber is not under vacuum before raising the Chamber top plate using the mechanical hoist.  You MUST either FEEL or HEAR excess N2 vent air escaping from the chamber BEFORE using the mechanical hoist.  Raising the mechanical hoist while the system is under vacuum will cause the bell jar to break-implode.

Also, the bell-jar gasket sticks to the chamber top plate, so lifting the hoist will often lift the bell jar up into the air.  Only lift the top-plate about a centimeter with the hoist, let the bell-jar drop, then continue to lift the top-plate.

The third and most dangerous component of the system is the electron beam itself.  The electron beam is Powered, Focused and Aligned entirely by the user.  You must make sure the Power dial is not bumped when making positioning adjustments.  Make sure the e-beam is hitting the correctly filled (~2/3 full) and correctly hits the source material, not the crucible wall or hearth, and is not too narrowly focused.   DO NOT OPERATE the e-beam with an empty crucible or hearth: the beam would melt a hole through the machine, causing irreparable damage to the system.
Procedure:

Following are instructions for correct operation of the e-beam evaporator.

(1) Substrate Loading

Vent the system:

  1. From the standby position: (HIVAC and ROUGHING switches on the front panel are closed and the VENT switch is down, or off, on the right side of the chamber, water and circuit breakers OFF), open the VENT valve.  Wait approximately 2 minutes.
  2. Once you can feel N2 Vent gas escaping, or can hear it casing the butyl rubber chamber gasket to resonate, use the toggle switch “mechanical hoist” up to raise the top plate up 1 cm over the bell jar.  The gasket sticks and sometimes the bell jar is lifted high into the air, where it could drop.  If the mechanical hoist is used while the system is still under vacuum, the bell jar will implode/explode. BE CAREFUL.  A new bell jar costs $2000 and takes months to be delivered.
  3. Raise the top chamber plate up the rest of the way, and close the VENT valve.

Clean the chamber/shutter/replace foil:

  • Delaminating films in the chamber can ruin your high-purity deposition material. To clean, use a particle mask, the vacuum, a razor, and fabwipes to clean the chamber wall, top plate, and related fixtures as needed.  Replace the aluminum foil top plate shielding if it is in poor condition.  All surfaces on the top plate must be covered line-of-sight from the crucible by either disposable foil or the substrate plate. Vacuum up all loose debris, so when the chamber is vented or pumped and the air inside swirls, the material doesn’t get deposited in your crucible.  Clean the hearth: any particles in here will cause the crucible to not sit snugly, and prevent the proper cooling of the crucible, making it superheat and either alloy into the evaporation material or break.  Again, as fine particles are much more toxic than bulk materials, wear a mask to protect your lungs.

Load your sample:

  1. Clean, new gloves should be worn when handling parts that go into the vacuum system, or when touching inside the chamber area.
  2. Mount your sample(s) onto a wafer plate with screws and tabs provided.
  3. Load the wafer plate by sliding it into the two parallel rails on the underside of the top plate.  Make sure the shutter blocks the sample when the “closed” button is pushed on the MDC360, and does not block it when “open”.  Open is denoted by the red light on the shutter button being “on”.
  4. Make sure the correct source crucibles are loaded into the hearth. Check that there is enough material in the crucibles so that the electron beam will not burn through the bottom of the crucible.  If the crucible is not reasonably full, you will get poor evaporation results as a larger amount of heat gets directed back into the crucible.  It is good practice to add some source material to the center area of the crucible, on a regular basis to keep the target at an optimal domed profile.  Increase the melt time slightly when you have added source material to the crucible.  When evaporating several layers, take accurate notes of the hearth position for each source. 
  5. Check that several reflective glass slides are in the holder at the bottom of the chamber, and that the arm is in the middle of its range of travel.  These glass slides serve as a mirror in aligning and focusing the electron beam. 
  6. Check the crystal thickness monitor on the MDC360 display: if it shows below 85%, replace it. 
  7. Lightly wipe off the top rubber gasket and the mating metal chamber surface with a clean wipe wetted with a small amount of 2-propanol.
  8. The bell jar is not perfectly flat: lift up the butyl gasket to make up this gap, about 1mm, in the same place you felt N2 escaping when opening the chamber.  Make sure the bottom of the bell jar is well centered within the marked area.
  9. Lower the chamber top plate by pushing the hoist switch to the right.  Guide the chamber plate as it comes down so that it sits on the bell jar properly, taking care to not bang the crystal monitor or trap foil in the seal.  Do not lower the hoist more than 2 cm past when the top plate hits the bell jar, as it will push hard enough to break the jar.

(2) System Evacuation

  1. From the standby position: (HIVAC and ROUGHING closed on the front panel, with the vent toggle switch off, circuit breakers and water valves OFF) open the ROUGHING valve.  If there is a good seal, the mechanical pump should become quieter within 1 minute as less air is being compressed.

    ** make sure the roughing pump quiets down within TWO minutes.  If it doesn’t, you have a bad seal and are going to “blow” the roughing pump oil up the building exhaust pipe – bad for the duct work, bad for the pump.  **
  2. After roughing for several minutes, briefly close the ROUGHING valve and open the vent valve (actually dry N2) for 5 seconds, then re-open the ROUGHING valve.  This will help sweep out moisture and lightweight elements from the chamber, improving pressure and pump down time.  You may want to do this twice but do not have to.
  3. Pump down to approximately 500 mTorr, where there is a marked line on the vacuum gauge, which should take about 5 minutes, then close the ROUGHING valve, and note the leak up rate (Rate Of Rise).  The ROR should not be more than one log-scale increment/min.  If the ROR is bad, vent the chamber and check or clean the seal or backfill with N2 and pump out the chamber again.
  4. With ROUGHING valve CLOSED, OPEN the HIGH VAC valve
  5. Pump a minimum of 30minutes, to 3x10-6 mTorr or lower, as measured by the ion gauge on the left of the e-beam, set to tube#2.  You may wish to toggle the “degas” switch up for 30 seconds to heat and dry the ion gauge bulb.
(3) Programming the MDC-360 Deposition Controller

We will use the controller only to display instantaneous and total film thickness, and to operate the shutter.  To calculate material thickness, you must input the correct material into the MDC360, as follows: with the e-beam current is off (key is off), press START, then choose the material being deposited with the up/down arrow key.  Next press in rapid sequence: START, ABORT, RESET, ZERO, AND SHUTTER.  This will leave the correct material properties in memory, with a zero thickness start and the shutter closed.

(4) Deposition
  1. Before running a deposition check that all of the following conditions have been satisfied. 
    • The vacuum pressure is under 3x10-6 mTorr.
    • The crystal life is more than 85%.
    • Three clean glass slides are in position above the crucible for viewing
    • The MDC-360 has been programmed properly with all the setpoints.
  2. Open brass water valve V3, with a yellow handle behind the e-beam one foot off the floor. All water valves must be opened immediately before your run starts, and must be closed immediately before you open the chamber at the end of a run, after allowing 10 to 15 minutes of active cooling. Check that the vacuum pressure ion gauge does not rise much when the water is turned on, as that signifies a water leak.  If there is a leak, shut down the system per instruction “A” on the front page.
  3. Open the orange plastic water valves (V1, V2) to cool the power supply and crystal oscillator.
  4. Turn on the MAIN and CONTROL circuit breakers.
  5. Set Emission Dial on the PAK remote Gun Control to zero amps.  Center the x-y position at 0 and set the sweep controls at .5, in the middle of their ranges.
  6. Verify the shutter is closed (red light off) and the correct material is displayed.
  7. All interlocks indicated by red lights on the front panel of the Gun Control Module must be lit.
  8. Turn the High Voltage Key to ON, then wait 30 seconds before proceeding.
  9. Press the green Gun Control ON button. You will hear a relay click and see 7KV on the dial; the ON button will flash. 
  10. Turn up the power dial on the PAK remote until current is established, then immediately turn power back down to “0”.  Slowly turn up the power, per instructions following, viewing the “Amps-Current” gauge on the power rack as exclusive feedback of current: dial accuracy is negligible. 
  11. As the source heats and becomes visible, use the x-y position controls on the PAK Remote Gun Control, the two left slide locators, to center the e-beam impact on the source material.  Do this at low power so mistakes in beam aim don’t destroy the hearth! Turn off the room lights to allow easier viewing of the slightly heated crucible.  Note the “X” slide is reversed: move it right and the beam moves left. Once the beam is centered, use the sweep controls, the slides on the top right of the controller, to increase the area heated in the crucible from half way to over 3/4 when doing Si, Al, or any material which sublimes like Cr, SiO2 or other oxides.  These materials are susceptible to having the beam punch holes through to the center of the crucible if the sweep is narrow.
  12. Heat up at approximately .02 A increase per minute.  Watch the deposition rate: use about 2-3 A/s for a thin adhesion layer, or a liftoff film, and ~5 A/s for most others.  Adjust power with the PAK remote as needed.  When the A/s is stable at targeted rate, check the ion gauge pressure.  If the pressure is very high, material dependant, it may signal a vacuum pump or contamination problem. If ok, press SHUTTER open (light on) and ZERO to reset the total thickness.
  13. Visually monitor the vacuum chamber and the interlock lamps for anything unusual, such as:
    1. if the water interlock lamp goes out, from a temporary low water pressure situation, the gun will switch off instantly. You should immediately press the green start button, to reestablish current before the crucible cools. If not done immediately, you must close the shutter and gradually ramp up heat before deposition can begin again.
    2. if the ammeter needle deflects past 0.4 Amps for more than one second, typical of a short circuit typically between the filament and ground, and usually accompanied by a vigorous humming sound, then immediately turn the key off.  Often these shorts are caused by thin delaminating films which are evaporated completely, and you can key back on and re-start your run, but if it again shorts, turn the key back off immediately, and start the 15 min cool down sequence, and notify staff for repairs.
    3. in other emergency situations, execute the Emergency Shut-off Procedure and notify staff. 
  14. Record the run parameters (baseline and run pressure, current, sweep settings, etc.) on the run log sheet while evaporation is occurring. 
  15. When thickness target is reached, push SHUTTER closed (light off) and reduce the filament current by half, wait two minutes, reduce again, then turn to zero on the dial for two more minutes before turning the filament key OFF.  Note: there is a residual current at indicated Zero: you will need to push the “stop” button or key the gun controller “OFF” to stop completely, after 1 min at minimum power.  Leave the circuit breakers, cooling fans, water ON and if this is your last layer, start your 15 min cooling clock now, otherwise rotate to the next layer, and input the correct material into the controller, proceeding at step 8 above.
  16. When done, DO NOT REMOVE the substrate until the system has cooled down with High Voltage filament key OFF, but fans, and water cooling ON.
Shutdown

After having active cooled (all power except HV key on, water on) per above, shut down system and retrieve your sample:

  1. CLOSE HiVac valve
  2. Turn the MAIN and CONTROL circuit breakers OFF
  3. Turn OFF all three cooling water valves
  4. Open the VENT toggle. Proceed to open the chamber per “SUBSTRATE LOADING \ VENT” instructions (1.1 above).
  5. Unload your sample.
  6. Close chamber, and Pump down to 500 mTorr and then turn OFF the ROUGHING valve

The system should now be in the STANDBY status: the system is under vacuum, all water valves are off, and the two rocker switches on the front panel and the side vent are closed/off.

Author: Kurt Broderick, 2/05 rev. <kurt@mtl.mit...>

Fabrication

MTL Annual Research Report 2012: View Online