Panasonic schottky......or not ?

That data sheet says "schottky" at the top. The forward and reverse curves sure look schottky.

I don't think that epitaxial is incompatible with schottky.

The V_forward graph looks like a Schottky.

George h.

So does the reverse curve, both in magnitude and shape.

Schottkies can have real reverse recovery, if they have a p-n guard ring that conducts at higher currents. But Panasonic may just be a little weird.

Huh, thanks, I never knew Schottkys were so leaky in reverse. There must be some physics reason for that...?

Re: p-n guard ring, well that's doubly beyond my knowledge, though edge effects in photodiodes is something I see all the time.

George H.

Dunno about the physics, but they are very leaky and leakage increases smoothly with voltage. Regular diodes are more constant-current leakage then a steep zener thing.

It's a steep avalanche thing. Zener breakdown only happens for breakdown voltage less than about five volts.

Sure, Schottkys are leaky. If they weren't, there will be far less use for silicon diodes.

Vf = 350mV at 100mA, near the 200mA max, has to be Schottky.

A diffused guard ring creates the builtin field of an unbiased PN junction, in the right direction to pull surface dirt away from the (sensitive) edge regions of the Schottky metallization. The same kind of thing helps HV diodes (lessens leakage and aging effects), when simple passivation isn't enough.

Half the problem in semiconductors is that you can make resistors, and amplifiers, and switches, easily, but insulators have to be reverse biased PN junctions, and wires have to be mapped onto one surface (if two wires have to cross, there's cost or trickery to do that).

The other half, is that the finished item has all the working parts buried, but unburied surfaces ALSO work, usually against your design goals. You can't keep surfaces absolutely clean. Some of those surfaces have to sink heat, through bear grease, too..

Guard rings are a behind-the-scenes necessity, and (from the physics point of view) rather elegant.

Of course it's a Schottky diode, but can't a metal contact be deposited on epitaxially grown silicon?

One reason early diodes (of the 50s and 60s) were so fragile (relatively low voltage rating, avalanche causes death): where the junction meets the surface, it's either cleaved (in which case, who knows what dirt, and particular dopants and surface states, are going on), or early planar diffused with similar problems, poor material purity, homogeneity and such.

Today, even schottky diodes are rated for avalanche (though not all that much, it's just the guard rings), which sure is handy.

Tim

Lower barrier. :^)

Or something to do with the metal being full of free charge carriers, I don't remember.

The high voltage kinds (over 100V I think) leak almost as little as the PN kind, because of another bit of tech: the schottky junction is placed on a pillar, surrounded by gate oxide and metallization (or just bare, using ambient electric field without needing extra field plates). Basically they put a depletion MOS in series with it, increasing Vf and Rs a little bit once again (hey, no free lunch), in exchange for much higher breakdown voltage and lower leakage.

Tim

Right that was my first guess. (But I'm not sure.)

This is OK

Says it's due to majority carriers on metal side. I should make my own Schottky I-V graph, that would help.

George H. (I'll take Sze home tonight and read chapter 5 again.)

And a very differently shaped curve.

OK I got my parts ordered and did an I-V. Well two, first I did a

It's a semi-log plot (with log of the voltage) so it looks different from the panasonic data sheet. (making it semi-log the other way and it's mostly just a flat line.)

George H.

Sure. Epitaxial silicon is no different from bulk.

Cheers

Phil Hobbs

Well, to the extent that a schottky junction cares. (There are important differences with epitaxy, but that's subtle compared to "does it junction" or not.)

Tim

No, there aren't, except for doping, and not necessarily even that. Epi is monocrystalline silicon, grown on top of the bulk with no boundary. That's the whole point of the exercise--building clean, dislocation-free junctions with way more control than old-timey diffusion or even ion implantation.

Cheers

Phil Hobbs

Yeah, that's what I mean :)

Tim