Intrinsic Silicon

Many people already understand, that two types of silicon exist, N+ -Doped, and P -Doped.

Well I’ve known for some time, that another type of silicon which exists, is called ‘Intrinsic Silicon’. This is a form of silicon, which theoretically contains no dope at all, and which is therefore non-conductive. It’s not even a semiconductor in that state.

This type of silicon might be of some interest in the design of modern Integrated Circuits, especially in the reduction of the capacitance of individual transistors. But there areĀ  essentially two problems with its use:

  1. It’s practically impossible for the silicon to be perfectly pure. The concentrations of Dope, in the N+ or P -Doped silicon, are already extremely low. The concentration of impurities in Intrinsic Silicon is simply lower, industrially, than in the intentionally-doped silicon, not truly zero. And what this means in practice, is that ‘larger pieces’ of Intrinsic Silicon are still partially conductive. In fact, how low the concentrations of N+ or P -Dope can be brought in the industrial process, depends on how low the level of impurities is, in the silicon, to begin with. In either type of intentionally-doped silicon, the concentration of dope must still be at least one order of magnitude greater, than the level of impurities was.
  2. Actually, I think that Intrinsic Silicon is more expensive in bulk, than either type of intentionally-doped silicon, which means, that if the entire wafer needed to be made out of it, since the substrate of the wafer is meant to provide mechanical support as well, then the cost of the manufacturing process would increase.

Yet, small pieces of Intrinsic Silicon, as the following image shows, can still be used to provide lateral insulation, between the P -Doped and the N+ -Doped wells of individual transistors, where a “buried oxide layer” provides vertical insulation between those wells, and the actual wafer:

And, it would be my expectation that because Intrinsic Silicon is ‘non-conductive’, larger pieces of it should also be optically transparent, which means that some people might mistake it for glass.

By definition, glass would be ‘amorphous’, which means ‘not crystalline’, which would make actual glass useless as a semiconductor. However, amorphous forms of silicon can readily be used in the design of wafers, as long as they do not need to participate in the actual semiconductive behavior between N+ and P -Doped silicon.



Why the inter-atomic world only approximates the macroscopic properties of matter.

In a previous posting, I wrote that the microscopic world, in this case implying inter-atomic distances, generates an approximation of the macroscopic, mechanical properties of matter.

What any alert reader should notice, is that in order for this theory to be true, it actually needs to lead to an exact result at some point, and not just to approximate results. And so the question which should follow is, ‘Why only an approximation, the way it was described?’

There is a family of answers to that question, which starts with the fact that not all solids are covalent solids. I was taught that there exist essentially three types of solids:

  1. Molecular Solids,
  2. Covalent Solids,
  3. Ionic Solids.

I feel that the WiKiPedia article I linked to in this list, gives a good explanation for what Molecular Solids are, and also gives links to the other types of solids. If the reader has serious questions, I recommend he read that WiKi next; they explain certain details better than I can.

At the same time, solids which I was taught were covalent solids, are really just a combination of molecular and covalent solids, due to the way molecules could be linked in certain directions, but not linked in other directions, in 3D. This is why the WiKi describes those types of solids as ‘mesh-solids’.

Organic polymers are extreme examples of meshes, while certain structural materials such as beryllium are completely different, being highly covalent, and being much stronger therefore, than organic polymers.

Another reason for which my first description is only an approximation, is the existence of thermal agitation. This means that individual nuclei are always in motion, even if the macroscopic body is not noticeably in motion. Furthermore, due to the involvement of Quantum Mechanics, heat can take the form of transitions between discrete states, instead of all the heat being stored, just as the continuous agitation of the nuclei. Hence, molecules which have a greater number of QM states to occupy, at any given temperature, will also store more heat, as their temperature changes, and will therefore also have greater specific heat. If heat was just the kinetic energy of the nuclei, we should find that all matter have very predictable properties of specific heat, just a function of atomic density, when in fact this is not so.

And, the velocities associated with thermal agitation at room temperature, are often underestimated. They can be enough to break the bonds between molecules by themselves, which is also a reason ‘why ice melts at room temperature’.

Continue reading Why the inter-atomic world only approximates the macroscopic properties of matter.