Silicon samples

For a product, or just for your own use?

Worcester Polytechnic University used to have a research fab -- perhaps there are universities that would do runs of these for you? If it's for a commercial product it'll be an unreliable source, but it's a thought.

At one point there were wafer aggregation companies out there: you made your custom design, and sent it to the wafer aggregation company, and then when they had a wafer full of design they'd run it through fab. I know that it was less expensive than doing a full wafer, but I have absolutely no clue how affordable of a "less expensive" that is, nor how you'd find one.

There are fabs out there that specialize in making obsolete components on obsolete processes -- if you're doing enough, maybe a 4" wafer full of Hall bars the size of 2N2222's wouldn't be too expensive?

--

Tim Wescott 
Wescott Design Services 
http://www.wescottdesign.com

ve

r

of

t I

For a product. I could imagine ordering 100 of them. (well after checking out that they do work.)

Yeah there are some schools that could do this, but it gets complicated if it's for a commercial product. And then of course the one guy at the schoo l who knows how to run the wire bonder leaves... and we're left hanging.

OK I should search for fab companies then? Part of my problem is I don't know what I should be searching for. Maybe i t's just a guy who has some lithography and deposition equipment in his gar age. (I've got some old pieces of Si kicking around at home.. maybe I'll start t here.)

Thanks Tim,

How about a piezoresistive pressure sensor? Those use diffused resistors, and come in nice packages if you like.

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs 
Principal Consultant 
ElectroOptical Innovations LLC 
Optics, Electro-optics, Photonics, Analog Electronics 

160 North State Road #203 
Briarcliff Manor NY 10510 

hobbs at electrooptical dot net 
http://electrooptical.net

Maybe this is what you want:

I think you need aluminium deposited before you can wire-bond to the sample, so I doubt those places could help directly.

You will get a greater Hall voltage with a thinner sample, or putting the current through a very thin diffusion on the surface of a thick sample.

I wonder if you can use a piece of solar cell. Those usually have a thin diffusion on the front with contact fingers that make ohmic contact to it. Perhaps you can cut or break a cell in such a way that the fingers break into separate contacts in a useful configuration, though I can't think of a way to make a hall effect device from lots of parallel contact fingers.

Otherwise you might be able to make your own! :

Chris

e are easy to find on the web. What I?d really like are Si samples that have been sliced with ohmic contacts added. This will be for a stude nt lab where they will look at conductivity and the Hall effect as a functi on of temperature (and doping). It?d be great if I could have a Ha ll bar patterned onto the device. (But perhaps that is asking too much?) Back in grad school we use to get samples (GaAs/AlAs QW) from the Cornell nanofabrication facility, and they always came with a hall bar and some C(V ) pads so these are perhaps standard? Searching the web I didn?t h ave much luck. And I thought of asking here. Besides names of companies e tc. even just a good search term would help. Should I look for people who do wire bonding for example? Will I need an epilayer put on the samples to protect them? (See, I?m not even sure exactly what I want.) Any i deas, thoughts would be much appreciated!

Hmm Something like this?

I think I might even have some of these.... So are you suggesting that I could just use the diffused p region as a samp le? Or can I somehow access the n-epi region too?

Thanks, George H.

e ar

hat

could

ought

I found that also. Making my own just sounds too expensive... time-wise.

There use to be a clean room at UB (where I went to grad school) that I had access to. I think I could still get in, but then it's a whole project.

Sure I didn't mean the wire bonding as a serious suggestion... but just tha t maybe I should be searching for something other than Si wafers.

e.

Yeah, I guess I'm not so much interested in big signals as much as showing the effect. Letting students measure thickness with some calipers would be OK.

Yeah a commercial device that I could bend to my needs would be ideal cost- wise.

Hmm I saw that before... HF scares me, but it's interesting that she just u ses supermarket cleaner.... That might work. Maybe give up on hall bars an d just do a van der Pauw type measurement.

Thanks,

George H.

The trick is to get an actual ohmic contact versus a Schotky, not to mention the removal of oxide that will form naturally for silicon just laying around a long time. These silicon valley analytical labs that do wafer profiling probably know more about contacting bare silicon than a semi company, or at least one that doesn't own a materials lab. [Probably TI has this technology in house since they seen to do everything in house.]

This is actually more in line with materials science than semiconductors. Maybe a company like Evans Analytical could set you up with something.

Now I suppose if you could get some junk wafers with test patterns on them, you might find a suitable test structure, but this presumes you have micromanipulators to make contact.

The old Motorola ones just pinned out the piezoresistive elements. I forget if the two sections were isolated or not, but if they weren't, it ought to be a poor man's Hall sensor.

Cheers

Phil Hobbs

--
Dr Philip C D Hobbs 
Principal Consultant 
ElectroOptical Innovations LLC 
Optics, Electro-optics, Photonics, Analog Electronics 

160 North State Road #203 
Briarcliff Manor NY 10510 

hobbs at electrooptical dot net 
http://electrooptical.net

Hmm, OK It's not something I know much about. We use to diffuse Indium into GaAs, but maybe that doesn't work as well on silicon. I realized I was thinking about the Hall bar in the wrong way. For the Hall bar I need to deposit something on a mostly insulating substrate. But I was also thinking about using wafers where one can measure the thickness. So I thne had the following idea... making a hall bar from a wafer. Say I take a wafer and break off ~1cm rectangular pieces. With four freshly cleved edges. On the narrow ends I'll make a contact down the entire length. And then along the long edges I'll deposit two (or three for zero offset) little contacts.

No micromanipulators. I'd like wires, wire bond? onto a piece of pcb, with solder pads.

I'll send Evans analytical an email... can't hurt.

Thanks miso, George H.

OK, we've got honeywell sensors, I can crack one open and see what's inside.

George H.

You would have to settle for wirebond onto a sidebraze package. [Die gets a scrub attach or expoxy mount to the package, then they bond it.] In the dark ages of NMOS (as in depletion load analog), some chips would have a substrate connection. There was a school of thought that filtering the substrate would improve analog performance. That kind of chip, well provided you could find one, would have vertical current flow through silicon. I don't know if that would meet your needs.

Also in the early days of RAM design,the substrate would be pumped in an effort to reduce parasitic capacitance. [You would tweak the VT to compensate for back body.] I don't know if they ever brought the substate out to a pin.

Ya know, the 80's were a long time ago.