Qs: How Group and Period affect semiconductors

[JeffEvarts]

Qs: How "Group" and "Period" affect semiconductors

Background: I have a BS in CompSci, not Physics, so if this question stuns you in its naivete, please forgive me. In the last couple weeks, I have been reading up on PN junctions and how they work. I've also seen the standard Si/Ge (Group IV) + Tri/Penta-valent (Group III/V) doping examples.

Despite their efforts to escape I think I have those basics trapped within my brain.

Wikipedia has a nice "list of semiconducting materials" list, and I was perusing that, and some oddities stuck out: mostly about how the overall size/atomic-weight/periodic-row affect how semiconductors are fashioned and doped. My questions stem from there.

Tin and Lead for instance, are in the same column as Silicon, but I don't ever hear about them being used as semiconductors by themselves. Tin and Lead are metallic/oid, of course, while carbon and silicon are not, but Germanium, which is also frequently used as a semiconductor, seems to be a metal as well, or at least as "metallic" as Tin and Lead.

Observation 1: The Group "III-IV-V" examples in the all textbooks don't seem to hold as the period increases.

Q1: Why?

Observation 2: Larger/heavier elements from group IV are sometimes paired with Sulfur or Selenium from Group VI to turn them into semiconducting strata, but not with Oxygen (also from Group VI)

Q2: Why?
Q3: Do they "dope" these semiconductors with the same group III/V elements to make them P and N? (I seem to find that the answer is YES, but I'm asking anyway)

Observation 3: Even stranger (3-and-4-element) combinations, (CuInGaSe) are also used.

Q4: What's the underpinning logic/science of choosing these compounds?

Metaquestion: All the textbooks seem to make these bright-shiny-clean generalizations about 4 valence electrons in their explanations, but that seems to fall well short of the actual physics, possibly TOO far to warrant such inaccurate explanations. Is there a slightly less general description of semiconductor structure which might prepare the dilettante such as myself for the realities of semiconductors?

I thank you for your patience,
-Jeff Evarts

 

[JeffEvarts]

Anyone? Bueller? I've been banished to the second page... 170+ views and no replies.

-Jeff

 

[Physics Monkey]

Since no one else has answered I will say what I can. This is not my area of expertise, but I think I can answer some of your questions. You asked several things so maybe I'll start out slow and we'll see what we can do.

You asked about the progression of silicon, germanium, tin etc. Silicon is of course a very common semiconductor. The other elements in this group may have relatively similar bulk electronic properties. However, germanium, for example, has pretty different surface physics as compared to silicon. When you create a crystal surface you often modify the bulk properties dramatically. Germanium has a fairly metallic surface and "reconstructs" (reforms its bond structure) different from silicon. Since the physics of these exposed surfaces (chemical reactivity, etc) is often quite important for semiconducting devices you can imagine that these surface variations can change usability quite a bit. So to comment on one of your comments, perhaps its fair to say that even if the bulk chemistry of the elements in a group are similar, much more matters for a semiconductor than just the bulk physics. This means that the bulk generalization isn't necessarily wrong, it's just not the whole story in the presence of surfaces, say.

There are other factors as well. Lead is toxic, for example.

Does this help at all? More to come later.

 

[PhaseShifter]

There's also the matter of the size of atoms. A III-V semiconductor essentially has the same structure as a IV semiconductor, with half the atoms replaced by group III atoms and the other half replaced by group V atoms.

This works well assuming the atoms are comparable in size, but for really vast differences you may end up with a different crystal structure altogether.

Also, smaller atoms tend to have more tightly bound electrons than larger atoms--so larger atoms tend to lead to more metallic materials, while smaller atoms tend to lead to materials that resemble insulators.

This is somewhat dependent on temperature as well--most people would think of diamond as an insulator at room temperature, but at a few hundred degrees kT is much closer to the bandgap.

 

[JeffEvarts]

PhysicsMonkey: Thank you for (the beginning of) your reply. I'm learning that a lot of this seems to center around a "suitable band gap", and that band gap seems to be determined experimentally rather than predicted theoretically. Thus, at regular temperatures, C's band gap is too wide, and that of Pb and Sn is negatively sized (they're conductors) So it's less about the rigidity of the 4-valence-electron-crystal, and more about the close-but-not-too-close band gap. It does seem odd that all the examples seem to stress the former, when the latter appears to be the "core" of the matter.

PhaseShifter: That makes sense, BUT... then I'd expect Tin to be paired with Iodine, but I find a lot more references to SnS than SnI. Not that "counting google references" is the same thing as doing actual research. :)

Thanks to BOTH of you for replying, and I look forward to more, Physics Monkey.

 

[marcusl]

I know a little about semiconductors (SC's) and not much about the rest of the periodic table, so I won't be much help beyond this: lead is a metal, tin and graphite are semimetals, while germanium is a semiconductor and not a metal at all. The differences arise from the band structure of the solid, and this is a topic that is not easy to understand without some quantum mechanics and at least a brief introduction to statistical mechanics. If you want to proceed, you may find a systematic (as opposed to spotty Wikipedia) approach will be more satisfying. Here are two books that are classics in their fields and have served generations of students:

Kittel, Introduction to Solid State Physics. (Up to its 7th edition, although I've heard some say the older early editions are better).
Sze, Semiconductor Physics, 2nd ed.

The first is a general course in solid state physics, very well written, and maybe just the thing since it seems you want to delve into the topic broadly and deeply. The second is a physics book oriented towards engineers, dealing only with semiconductors.

For a broader view of elements in general, Walter Harrison wrote a book called "Electronic Structure and the Properties of Solids: The Physics of the Chemical Bond". I've not read it but he is very solid and his book on Solid State is another one of the grad-level classics. It's a Dover book so inexpensive.

Hope others chime in here with additional information, but if not these might get you started.

 

[PhaseShifter]

That makes sense, BUT... then I'd expect Tin to be paired with Iodine, but I find a lot more references to SnS than SnI. Not that "counting google references" is the same thing as doing actual research. :)

Keep in mind iodine also is one valence electron short of a full octet--You would expect its compounds to not form network solids at all.