How can one measure density of states?

[afrano]

Hello, folks.

Q: How can one measure the density of states of a semiconductor and a conductor? I would imagine you want to measure the charge carrier density and then you can calculate the density of states. If so, what observable(s) can yield the charge carrier density? How can you measure these observables?

The question looks long, but the answer should be condensed into a basic principle...I hope.

 

[ZapperZ]

Try looking into tunneling spectroscopy. One can measure the I vs. V (current vs. voltage) across the tunneling junction, or measure the second derivative dI/dV. This is equivalent to the tunneling conductance. What is interesting here is that the normalized conductance is equivalent or proportional to the density of states, depending on how complicated the tunneling matrix element is for that particular material and tunneling geometry.

Zz.

 

[afrano]

Thanks Zz.

Can you give me a reference where I can see exactly how the second derivative dI/dU relates to the density of states?

Thanks.

 

[ZapperZ]

E.L. Wolf has a definitive text on tunneling spectroscopy in solids, that I used as a graduate student. That might be a good place to start.

Zz.

 

[Modey3]

afrano,

one of the most straight forward ways of measuring the DOS is to use photo electron spectroscopy, which is found in most well equipped surface science labs. We have one our lab. the principle is based on the photoelectric effect and will give you only the LDOS (local density of states) of the atoms near the surface. Mind you, the information obtained isn't bulk information, but it may be sufficient for your purposes.

modey3

 

[ZapperZ]

afrano,

one of the most straight forward ways of measuring the DOS is to use photo electron spectroscopy, which is found in most well equipped surface science labs. We have one our lab. the principle is based on the photoelectric effect and will give you only the LDOS (local density of states) of the atoms near the surface. Mind you, the information obtained isn't bulk information, but it may be sufficient for your purposes.

modey3

We need to be a bit careful here. Photoemission spectroscopy may not necessarily produce a DOS. ARPES, for example, doesn't give you the DOS. What it does give you is the spectral function at a particular momentum. To get the DOS, you have to average out over all momentum values, and this is assuming that the matrix element doesn't play a significant role in transition probability.

Furthermore, even after doing the averaging, it will only tell you the DOS of the occupied side of the band. It cannot probe the unoccupied side the way tunneling spectroscopy can.

Zz.

 

[Modey3]

Zapper,

I was not suggesting going as far as looking at the distribution of occupancies in reciprocal space. I was merely suggesting using photoelectron spectroscopy to get the occupancy at a particular energy which means I don't have to use ARPES. Traditional photoelectron spectroscopy will not tell you anything about the band structure, but it will give you information about the occupied energy levels.

modey3