Re: [Phys-l] Transparency

[John Denker <jsd@av8n.com>]

On 10/10/2008 08:46 AM, Stefan Jeglinski wrote:

> I am curious: how would one typically inject charge into diamond? 
> Could you just deposit a metal electrode on it and have it work 
> efficiently?

Making ohmic contacts to any semiconductor (large gap or small)
is a black art.  If you just deposit a nice layer of metal on
the surface, you're going to get a Schottky barrier diode, not
an ohmic contact.

I've asked quite a few world-class experts how ohmic contacts
work, and never gotten a straight answer.  Various recipes are
well known, but the underlying physics is obscure.  My *guess*
is that it has to do with making lots of small, sharp points
and relying on field emission.

For present purposes, it is probably best to postpone asking
about "typical" or "practical" charge injection, and instead
to start by considering _ideal_ charge injection, such as
could be achieved by using a super-high voltage, from a static 
electricity machine or some such, to inject the charge by 
brute force.

> If I take a chunk of copper, for example, and connect a multimeter to 
> it to measure its resistance, the mere placement of the probe by hand 
> is enough to inject plenty of charge to form a measurable current and 
> make the determination. My intuition might tell me that I would 
> measure an analogous diamond chunk as having very high resistance, 
> and hence would assume it to have insulating properties. But I 
> wouldn't know if my issue was the probe (that is, a poor experiment). 
> Your point is well taken though - high resistance is not charge 
> immobility, although I'm not sure where the line is crossed.

That's the right set of questions.

You can clarify most of the "probe" issues by using a four-point 
probe.  For any semiconductor, if you don't go to the trouble of
making good (black-art) ohmic contacts, touching wires to the 
crystal will give you a resistance reading of the form "zero 
divided by zero", as it should, since the Schottky barrier diodes 
prevent you from injecting any current.

  Actually, with rare exceptions, for *any* serious resistance 
  measurement, you should a four-point probe.
    http://www.av8n.com/physics/img48/four-point-probe.png

Conversely, if you do make good ohmic contacts, the bulk diamond
will carry current just fine.

> I don't know enough about diamond to make good guesses. 

Diamond has the same crystal structure as silicon, so you 
can get a good start by applying everything you know about 
silicon, just with a larger band gap.

As previously mentioned, silicon at liquid nitrogen temperature
has essentially the same physics as diamond at room temperature,
at least as far as the band structure is concerned.


On 10/10/2008 09:07 AM, chuck britton wrote:
> > So a diamond is transparent to an electron beam????
> >
> > 		weird

I don't think diamond is any weirder than silicon.

I say again:  I have built amplifier circuits using silicon
field effect transistors and operated them not just at nitrogen
temperature but even at helium temperature (4 kelvin).


Don't get hung up on the idea that a material is an "insulator"
when the band gap is large compared to kT.  Just because all the
textbooks say it's so doesn't make it so.