All now were fond of nanotechnology. The term "Nanotechnology" means very small distances, is thinner, than the hair of the Neanderthal man, but is more than "Infinitesimal" than Frenchman Cauchy. However all forgot about small corners. It is necessary to pay more attention to nanocorners, then everything at you will turn out dear Pisek Kultavewuti.

Sincerely yours.

Alexander Yurkin

Specialty tools are important for reproducible and yield. However, a diamond scribe can work well for many materials. GaAs is relatively easy to cleave. Silicon is tougher. Sapphire c-plane is hard, but you can sometimes get nice facets on a-plane. It depends on the material I usually don't rub the diamon scribe along the wafer back and forth. Instead I put down a chem wipe, and wrap a microscope slide with another chem wipe. I align the edge of the chip  and on the back surface and put a small scratch using the diamond scribe. Then I flip the chip over, and place it on either the edge of the microscope slide, or on several folded layers of chem-wipes, then using the rounded ball on the back of the diamond scribe apply a rolling pressure. Placing the edge of the chip on the folded paper seems to help. I have the mark suspended between folded paper and the chem-wipe, There are other tricks for laser diodes people used to bond the chip to a foil with wax , thin the wafer by polishing,  then under a microscope then make a series of tick marks. Then wrapping the foild around a cylinder like a pencil, the laser facets could be cleaved. A dicing saw was used to cut the other dimension. and have more non-reflective edges.

Cleaving optical fibers can also done with a scribe, but I recommend using a tool if possible. 

In all the above cases, if done carefully it is possible to get a very nice cleaved surface. Yield though can be a problem..

I have't really tried to quantify the surface roughness by AFM or other technique, but mirror like surfaces are a reasonable expectation if you have a crystal plane to work with, or in the case of glass fibers, you can apply tension and let the crack propagate.

Several quick questions regarding working with free-standing substrates of Si and GaN:

1. What is meant by stating that a sample of GaN is "C plane (0001)"?  Why are there four digits in designating the plane?  I have seen this reference with several vendors.

2. With GaN or Si, can I scribe a line with any angle on that flat surface and then cleave, or must the the line have a specific orientation?  I would appreciate an answer telling me why as well.  

3. I see excellent texts on solid state physics, but I have not found any text describing the methods for working with the substrates--including Google, a library search, and Amazon.  I would appreciate your referring me to such a book.   Thank you.