Ga removal from backside of GaAs wafers

When receiving GaAs wafers, they often come with a layer of gallium on the backside. This layer has turned out to cause quite some troubles, as it introduces an uneven surface on the backside of the wafers. The rough gallium on the backside tends to give high sample inclination and makes it difficult to handle small samples. You can request the growers to remove this Ga layer before shipping. Meanwhile, you need to remove the Ga from the wafers.

Techniques from John Watson

• Spin and soft bake photoresist (2x AZ1505 resist spun at 4000RPM for 1 min sec, baked 1 min @ 110C, giving a thickness of ~1 um). This protects the chip surface during the subsequent steps
• Then set the chip face down on a clean wipe and wipe off as much Ga as possible (usually almost all) with a cleanroom cotton swap (sometimes it is necessary to set the wipe on a slightly warm hot plate to keep the Ga from freezing)
• Etch the chip in full strength HCl for ~ 3 min to get the residual Ga off the back of the chip.
• Rinse the photoresist off the chip with acetone. I personally hold the chip with tweezers over a "dirty" beaker while I rinse it off with acetone and then immediately (before the acetone can dry) stick it in a clean beaker of acetone.  The idea here is to prevent any Ga that may have been on the photoresist surface from settling on the chip surface (a good way to ruin a chip).
• You could probably skip the wiping step and just etch the Ga off with HCl, but it would have to sit in the acid a lot longer (probably 30-60 min) to get all the Ga off.

We see that at Harvard they used a technique provided by Xanthe, who again got his technique from Purdue, so I guess from some of Manfra's former students.

Ga removal tips from Xanthe

• My method of getting the gallium off is to mount the sample face down on a dummy wafer with some thick photoresist (I just glop about 2-3 drops of AZ9260 on the dummy wafer without spinning) and push the sample around on the dummy wafer on the hotplate (bake 2 min @ 125C on hotplate) to get good resist coverage on the sample.
• Then I wipe as much gallium off as possible (probably >95%) with cleanroom q-tips. If the gallium re-freezes, just stick it back on the (still warmish) hotplate for a few seconds.
• If the engineering staff in your cleanroom are really picky about not wanting any gallium in any of their machines (like they are at Purdue), then you can just stick the sample-dummy wafer sandwich in full strength HCl for 2-3 minutes to get the last little gallium residue off.
• The sample will come off the dummy wafer after a few seconds in sonicated acetone.

I initially used the q-tip method, with AZ6612 resist - I baked it for about 2 minutes at 95C. Wiping with q-tips seemed to remove most of the gallium, but I wanted to be sure it was all off. For the HCl dunk, I actually just spun a bilayer of LOR 5A and AZ6612 resist, baking them for appropriate times (AZ same as before, LOR 5A for 5 minutes on 170C) ... I didn't worry about the sample-dummy wafer sandwich. This seemed to work, but the resist was pretty thin following the acid dunk, so next time I think I'll play it safer and use the sandwich method. Also, I would remove the sample from the acid every 30 seconds or so just to check it's okay.