Dials from silicon wafers

[dadistic]

Rather than to continue to hijack @Barsomn's intro thread, I thought I'd float this idea here.

I think it would be quite interesting to make a dial out of an etched silicon wafer. I used to have some blank ones that I got from a friend that worked at a company that built lapping machines for polishing them, but unfortunately they are long gone. My memory of them (long time ago) was that they might have been sturdy enough to work with to cut out a dial, but I'm not sure.

I have worked with more recent raw triple junction solar cells, and those darn things were so thin that they would crack if you sneezed at them.  Not something I would consider trying to make a dial out of. 

Anyone out there have any reject wafers that they can describe?  I'm just trying to figure out if this is feasible.

Cheers!

 

[spectre6000]

I think my buddy's lab was doing stuff with experimental PV cells recently. He is working remotely and at great distance these days, so direct access might be a challenge, but I'm still in the area and might be able to work out a handoff. I'll reach out to see if he's got any leads.

[dadistic]

Thanks!

I was thinking IC, but a PV wafer could be very cool, too. The colors can be quite striking.

Cheers!

 

[spectre6000]

I hooked up with him yesterday. He said if they're not immediately binned, they might end up at the university's surplus auctions, but neither of us recall seeing any such things there.

You mean the green (or purple or blue or black or whatever) circuit board silicon?

[dadistic]

I'm talking about the wafers from an IC fab, that have been etched (maybe partially) but not finished, and so not cut apart and encapsulated etc.  I would expect to see at least a few of these around, maybe from when processes are started up and being debugged.

I've attached a jpg of an etched wafer.  This one has a nice diffraction effect.

[dadistic]

Found some wafers! These are 4 inch wafers from the 80's, supposedly Rockwell chips. I've got five of them so that one or two of them can be sacrificed to my ham handed ministrations. They are about 0.5mm thick, and seem fairly sturdy. 

They are very reflective and hard to photograph

 

 

 

 

 

[AndyHull]

Silicon wafers are tricky things to cut. You will almost certainly need some diamond abrasive cutting tools.

https://www.waferworld.com/thin-silicon-wafers-methods-silicon-wafer-dicing/

 

 

[dadistic]

That's why i wanted some to mangle. I was going to try scribing like they were glass, and try a high speed rotary tool, otherwise known as my trusty Dremel. I really don't know how successful I'll be, but I thought it would be fun to try. Good thought about the diamond abrasive tools, thanks.

[AndyHull]

I'm looking forward to seeing the results.

Don't forget to wear your dust mask if you are cutting ceramics. You may find cutting with the Dremel and a diamond disk goes better if you use a bit of water or oil as a cutting fluid (and this will also cut down on any potentially irritating dust too).

Silicon wafers are doped with various metals including potentially antimony, gallium and arsenic. You are not going to be exposed to hazardous quantities of these materials, but it is worth keeping that in mind.

Ceramic wafers, like glass and other ceramics also produce extremely sharp splinters when broken, so a pair of nitrile gloves or similar when handling them may be a good idea to protect you from small splinters. They will also keep the fingerprints at bay, as silicone wafers are fingerprint magnets too.

https://en.wikipedia.org/wiki/Doping_(semiconductor)

Edited August 23, 2021 by AndyHull

[LittleWatchShop]

I have spent 40+ years in the semiconductor industry and have a number of wafers ranging from 2" to 8"

Wafers are made from crystalline silicon--meaning the atoms are perfectly ordered in the lattice (that is the goal anyway). Cutting a hole, or making a circle is probably impossible.  You can make orthogonal cuts and I suppose, in principle, you could make many small orthogonal cuts to approximate a circle or hole.  You would not do this with a Dremel tool however!

Rather than physically cutting, you could chemically etch the wafer, but even then, not something you would do at home.

Your best bet would be to dice up a wafer and then mount the die on a dial substrate.

[AndyHull]

If you can drill a hole in a glass light bulb, then I suspect you can drill a hole in a silicon wafer.. you will need to do so very carefully though.

As @LittleWatchShop states, this material is in essence a single crystal, so it will tend to shatter very easily, so bonding it to a substrate to enhance its structural integrity before you start may be necessary.

Epoxying it to a brass or aluminium sheet might work.

Would etching not require some pretty exotic etchants (nitric or hydrofluoric acid or whatever). Mechanical abrasion might be safer if somewhat more fiddly.

[dadistic]

Thanks,  @LittleWatchShop. My plan was to first simply try drilling with a carbide or diamond tipped drill. If I can do that, then the next step is to try grinding the edge of the wafer.  If that looks feasible, then the plan would be to do as you say above, make orthogonal cuts to approx a circle, then mount in my lathe and grind the edge. I have a rough idea for an arbor type fixture that might hold the piece and provide a diameter guide, and perhaps keep the piece from cracking

What makes me think this *might* be possible is how thick and sturdy these 4" wafers are. 

Thanks!

[dadistic]

I like the idea about bonding to a substrate. Stabilizing the piece could make a real difference.

Thanks!

[AndyHull]

If you want to bond, epoxy might be OK if you are not worried by thickness and intend to leave the substrate attached, but superglue or perhaps PLA or something soluble might be better as you can then dissolve it and remove the substrate when done.

[dadistic]

I was thinking superglue,  as I think acetate will dissolve it. I'm not sure about PLA, as the piece may be submerged in or sprayed with water.

 

Cheers!

[dadistic]

Well, tried drilling a hole today and had some success. Got some diamond coated drills, and tried out a 2mm one. Didn't do anything special,  just cranked the Dremel up to 10K and had at it.  There was some chipping as the drill broke through the hole. I think some tape front and back, some lubrication with water, and a lighter hand might help with that.  I could also see that there was quite a bit of runout in the drill bit, I'm  sure that didn't help. I'll have to see what I can do about that, but given I'm using a Dremel, it might not be much

 

[LittleWatchShop]

Wow.  Frankly I am shocked.  Well done.

[AndyHull]

Drilling a hole is a pretty good start, since you can then use the hole to centre whatever additional cutting method you use to cut your dial disk. Keeping the drill bit as short as possible and perhaps a gentle pecking motion rather than a full on push from the drill may help with the chatter and the chipping. I'm assuming the drill is in some kind of drill press.

[dadistic]

It's a Dremel drill press,  my Dremel rotary tool mounts in it, and turns it into a nice little bench drill press. Nice thing about it is the high speed. Yes, pecking is essential. 

I also figured out my runout problem, I tried a different collet and the runout is much better. I think the one I was using was distorted from being over tightened, they are just dinky little aluminum bits and hand tight is likely plenty.

I have some center drills buried in storage, I think a #0 or #1 will fit in the Dremel collet (1/8th"), and using one first might help a bit, too. 

Thanks!