/
Index Guided GaSb based MQW laser

Index Guided GaSb based MQW laser

Index Guided GaSb Laser Flow Adopted from Process Flow for Tunnel Junctions and TLM Written by NTS 01/23/13. Modified by NTS 131114, 140602, 150115.

[Note: Sample must be cleaved and in sample holders to begin.]

Part 1: Lithography and top contacts

Lithography, Laser 2013 Mask 1, and metal evaporation for metal liftoff that results in top contacts and hard mask for ICP etching.

Laser Stripe Lithography

Prep

  • Take photoresist out of the refrigerator because it needs to be at room temp

    • Crack PR fridge to let fumes get vacuumed up, then open.

    • AZ 5214E-IR

  • Check MJB4 lamp power

    • Check the power unit is working and on CH1/365

    • Insert empty mask holder, raise the stage, and expose for 10-20 seconds

    • Look at the output power listed on the power unit, it should be ~7.5 mW/cm^2

    • If it is not, you need to email Mertech to calibrate the lamp.

    • Power less than 6.5 can ruin your litho. 6.5-7.5 is not ideal but OK for process dummies.

  • Check IR oven temperature is at or fluctuating around 110 C.

    • Adjust knob slightly if needed

Solvent Clean

  • Solvent Clean (ACE, IPA, DI: 2 min)

    • Create a clean workstation before beginning.

      • Spray bench with IPA if needed, then wipe clean

      • Do this by getting clean wipes from the dispenser and lay them down anywhere you plan to work.

      • For Regrowth, layer wipes from back to front, and right to left, and go up to/over lip of bench

    • Setup solvent beakers and dipper basket

    • Note: The dipper baskets are labeled. Only use those that are appropriate for your immediate task. Otherwise, keep them in the box and close the box to prevent any contamination.

    • Always fill each subsequent beaker higher than the previous

    • Note: Gentle agitation during cleaning process is recommended

    • Acetone: 2 min

    • 2-propanol: 2 min

      • Go get UHV foil for boat

      • Scrap UHV foil is ideal, never trust other person’s boat

    • DI water cascade rinse: 2 min

    • Make foil boat

    • N2 blow dry

    • Put samples in boat

  • Dehydration bake

    • 150C oven: 5 min

    • Let Samples cool in boat

Spin PR

  • Spinner Prep

    • Make sure spinner chuck is clean

      • Spray wipe with Acetone and wipe chuck off

      • N2 dry chuck

      • If chuck is not at spinner you want it at, assume it’s dirty

    • Make sure you have pipettes for PR

    • Note: Set spinner to max time, always use your timer!

    • Check the spinner by placing a junk wafer piece on it. Make sure it spins correctly and at the correct RPM (4k) before placing on your actual wafer.

  • Spin on resist (AZ 5214E-IR, 4k rpm, 40 sec)

    • Check Sample 1 and N2 blow

    • Put Sample 1 on chuck and center

    • Check centering and RPM

    • Note: When putting PR onto the wafer, try not to put on so much that it leaks off the sides because this can get into the vacuum and mess things up. Also, make sure that there is an adequate amount, as to completely coat the wafer with an even amount.

    • Put PR (AZ 5214E-IR) in pipette carefully and drop a drop out on spinner wipes (to eliminate air)

    • Drop on PR, no bubbles, and Spin

      • 4k rpm

      • 40 sec

    • Check wafer when done. No streaks.

    • Repeat as needed

    • When done, empty pipette into PR container and return to fridge

  • Pre-Bake (evaporate solvents before lithography)

    • 90C oven: 10 min

    • Let samples cool in boat while you prep aligner

Exposures

  • First exposure (Laser 2013 Mask 1, 13 sec)

    • Choose stage for your sample

    • Prepare height of aligner stage to be close to (but lower than) what you will want with the sample.

    • Put Laser 2013 Mask 1 (laser lines) onto holder, ~1 mm from edge, turn on vacuum, slide into aligner, and tighten screws. Only touch the mask corners and edges.

    • Note: When placing mask onto the holder, place it so that the chrome side of the mask faces away from the holder. (Brown side faces sample after placement.) Keep your hand under the mask when moving the loaded mask holder.

    • Load sample onto stage and lower stage a few turns

    • While looking from the front of the aligner at the sample (between the mask and stage), rotate contact lever, and then raise the stage til it looks like the sample and its reflection on the mask are about to touch.

    • Align features, then raise stage til contact:

      • The best way to tell contact is by looking (through the microscope) for rings to appear at a corner of the sample or an edge

      • Can also look for a brightening of the mask metal, meaning there’s pressure on the mask

      • Whichever corner or edge displays contact first on your first sample will probably be the corner/edge to display contact first for the other samples

      • Note: Will feel a tiny increase in resistance of the knob

    • Parameters

      • Align + expose

      • Hard contact: 1 sec

      • 13 sec on MJB4 contact aligner

    • Note: Do not face the aligner when the wafer is being exposed. Also, make sure to move your other samples away from the aligner, as not to ruin them from scattered UV light. Usually this means putting a toolbox or bin between them and the aligner.

    • Lower the stage using the front knob about four numbers (knob number increases)

    • Rotate contact lever and center the stage rotation knob

    • Unload sample and repeat as needed

  • IR bake (Image Reversal Bake)

    • 110C blue oven: 2 min 15 sec

    • Note: Very important step, do not over/under bake

    • Remove mask from aligner

  • Second exposure (Flood exposure = w/o mask)

    • 30 sec on MJB4 aligner

    • Do not use hard contact

    • Note: Only put one wafer on at a time so as to not ruin wafers with scattered UV light.

Develop and Inspect

  • Develop with AZ 726 mif

    • 60 sec followed immediately by DI cascade rinse: 2min

    • Note: Gentle agitation the whole time.

    • Note: Does not stop developing after you take the wafer basket out of the developer solution, make sure to immediately dunk into DI water and rinse

    • N2 blow dry

    • Put stuff away and clean up bench

  • Inspect Pattern

The key parameter for this process is the bake. If the pattern looks like its features are way to big then the sample was under baked. If the features look too small then it was over baked.

Top Contact Metallization

  • Refer to CHA #1 procedure for more detail

  • Start Vent of CHA #1 (takes about 12 min)

  • Go to acid hood, prepare and use surface cleaning solution

    • HCL:DI (1:10, 40:120 mL)

    • Dip samples for 15 sec in solution (precision is not needed here)

    • Rinse for 30 sec in DI

    • Blow dry with N2

    • This part takes about 10 minutes, so start it once you vent CHA #1

Note: This takes 20-45 minutes depending on number of samples.

  • Take samples immediately to CHA #1

    • Vacuum and clean hearth and sealing surface

    • Load samples

    • Load sources (Ti, Pt, Au)

    • Pump down

  • Deposit contact metal (for top contact near epilayer we contact p-GaSb

    • p-GaSb: Ti/Pt/Au (100/100/2000 A)

  • Lift off metal in Acetone or KwikStrip

    • 60-90 min in Acetone or 40-60 min 90C KwikStrip at minimum

      • 6 hours or more is OK. Do not force metal off.

    • Gently spray with acetone or kwikstrip if possible

    • Collect all waste, as it is metal contaminated

  • 2-propanol

  • DI rinse

  • Blow dry with N2

Part 2: ICP etch of laser ridges and immediate conformal SiNx PECVD

Note: The etching must be immediately followed by SiNx PECVD! The laser material is extremely sensitive to oxygen, and if it is not passivated by SiNx immediately, the device might die.

Note: These two procedures are to be done in parallel, and timed so that the PECVD chamber is ready for use once the laser is etched! Start the PECVD work about 30 minutes after you start the ICP work (when the clean is ending). It is better to have the PECVD ready and waiting than the laser waiting for the PECVD.

ICP Etch of Laser Ridges

  • Refer to Oxford ICP procedure

    • Never run a recipe without a carrier wafer

    • Run a 35 min O2 clean

  • Etch (separately) three GaSb samples (call them A, B, C) for 8 min each

    • Recipe 1 (BCl3_a): Alter recipe AJ GaAs BCl3/Ar Etch

    • P = 5 mT, Strike = 5 mT, BCl3 = 15 sccm, Ar = 0, RF = 100 W, ICP = 500 W

      • Up to here should take 1.67 hours (100 minutes)

    • Load laser (sample D)

    • KwikStrip and Dektak the second two samples (B, C)

      • Sample B should be KwikStripped when C is etching

  • Determine etch rate and time for laser (D), 1790 nm

    • How do you do this? Look at the last page in the current GaSb_BCl3 etch document (version 7f)

    • We target an etch depth of 1790 nm (ignoring Metal Stack), although this number was chosen so that even if there’s variation in the etch we will still land in the safe acceptable etch depth range. The safe acceptable range is between 100 and 200 nm of the laser cladding remaining. This means an etch depth of 1740-1840 nm.

    • Pro tip, for experienced etchers, use with caution. It is acceptable to etch further into the cladding, but it is critical not to go into the core! If you could choose between 250 and 50 nm of cladding left, 50 nm is better. But between 250 nm or 0 nm, 250 nm is better. The non-safe acceptable etch depth is 1740-1920 nm.

    • Based on the change in rate between sample B and C, use the information to predict what the change between B and D, and C and D will be. The two answers you get for the expected rate of D should almost match. Use learned intuition to lean toward one prediction or the other.

    • Pro tip, for experienced etchers, use with caution. As a check to make sure you don’t over etch, calculate what the etch depth will be if you run the etch for the determined time, but the etch rate only changes by half of the predicted value or ten, whichever is less. If that is less than 1920 nm, you should be safe. For example, if your expected change d1 = 16 and you determine the next etch rate er4 and time t, find the etch depth if d1 = 8 (giving er4’) and you etch for time t, and make sure this is <1920 nm. This ( >1920 nm) should only occur if d1/er4 > 0.1453 (roughly). Also see what you’d get with d1/2-2 and d1/2-4 to account for even worse case scenarios. This is anecdotal advice.

      • Up to here should take 2.83 hours (170 minutes)

    • Get second pair of tweezers for quick laser unloading and transfer

  • Run etch on laser (D)

    • Immediately transfer laser into waiting PECVD chamber and deposit SiNx

      • Up to here should take ~3.083 hours (185 minutes)

    • Run O2 clean of proper length

  • Remember there is a 120 nm metal stack on top of the laser bars.

Silicon Nitride Deposition

  • Refer to PECVD procedure

    • Wipe out PECVD chamber and heat up (once the ICP clean ends)

    • O2 clean chamber for 30 minutes (Full Recipe on Wiki)

      • Up to here should take 1.583 hours (95 minutes (from Part 2 start))

    • Prime chamber: Recipe AMC_SIN 5 minutes (Full Recipe on Wiki)

      • Up to here should take 1.83 hours (110 minutes)

      • Or you can wait to run the prime til closer to etching the laser; just be sure to start the prime at least 10 minutes before loading the laser into the ICP (~18 minutes for the prime and venting the PECVD).

    • Remember to have chamber vented for laser before it comes out of the ICP

      • Vent the PECVD when the ICP vent counter is at 150 seconds

  • Deposit ~200 nm of SiNx onto sample: run AMC_SIN for 10 minutes

    • Run O2 clean for 20 minutes

Part 3: Lithography and SiNx etching

Lithography, Laser 2013 Mask 1.5, to pattern PR that will allow us to RIE etch windows in the SiNx; top contacts will be exposed.

Lithography

  • Same as in Part 1, but use Laser 2013 Mask 1.5

SiNx etch

  • Refer to PT1 RIE procedure

    • Wipe out RIE chamber

    • O2 clean chamber for 30 minutes (Full Recipe on Wiki)

    • Prime chamber: Recipe SIN_SF6b for 3 minutes (Full Recipe on Wiki)

      • P =150 mT

      • RF = 100 W

      • SF6 = 25 sccm

  • Can etch a dummy sample if you want to check the rate of the day

  • Etch: Run SIN_SF6b for 8 minutes

    • Make sure etch succeeded. This can be done by dektaking the Laser near the evaporator clip spot, or put in a dummy sample with the laser. (must be patterned with Mask 1 or Mask 1.5).

    • If it did not, etch for 0.5-2 more minutes. The top contact serves as an etch stop.

    • We might get thicker SiN at the edges where the PR was thicker, but this is of secondary concern.

    • Run O2 clean for 30 minutes

  • Kwikstrip PR for X minutes (3<X<3000, this seems to vary based on PR toughness)

Part 4: Lithography and Evaporation

Lithography, Laser 2013 Mask 2, to pattern PR that will divide each laser ridge and allow us to evaporate more metal on top of the top contacts.

Lithography

  • Same as in Part 1, but user Laser 2013 Mask 2

More top contact metallization

  • Same as in Part 1, but metal stack is Ti/Au (100/1000 A)

 

++++++++++++++++++++++++++++ Everything after here is a little questionable

Lap Substrate

  • Cleave laser and crystal bond pieces to holder as needed (forget what this holder is)

  • Lap laser pieces as needed

  • Release laser pieces from crystal bond using acetone

  • Bond pieces to glass slides

Bottom Contact Metallization

  • Same as in Part 1, but metal stack is for n-GaSb

  • Deposit contact metal (for bottom contact we contact n-GaSb)

    • n-GaSb: Au/Sn/Ni/Au (50/50/?/1000 A)?

    • nGaSb: Au/Sn/Au (50/50/1000 A) or Au/Ge/Ni/Au (100/100/100/1000)?

  • Lift off metal in Acetone or Kwik Strip and clean

Final Packaging

  • Release piece from glass slide and cleave into individual lasers

  • Indium mount bottom contact to copper piece

Process Flow

image-20240603-085751.png

 

Confluence Documentation | Web Privacy Policy | Web Accessibility