Dear Bob,

Thanks for the input on the boron diffusion,

Some power devices use a long diffusion to diffuse through the entire
slice. sometimes aluminum is used in a vacuum sealed quartz tube and heated
to 1250 deg c for a very fast diffusion.  Some use Gallium phosphide and
lap off one side to make a diode using the correct type of silicon and
resistivity.

The use of the solid planar disc is well known too and for a small R+D
project may not be so economical same as the B2O3.

This deep diffusion of 25 to 75 microns is usually done with diffusion
preforms containing boron for an intimate contact with the silicon and they
can be used to diffuse through the whole slice as you mentioned, a long
diffusion and is sometimes run at 1300 deg c. and a steam clean up for 30
minutes at 800 deg c is often used followed by HF soak to remove the glass
of boron and oxidized silicon or boron silicide.

The  spin on boron is probably the best method for doping at a level which
does not damage the silicon and will yield a deep junction. Levels of about
2.5 ohm per square with no surface damage can be achieved and the diffusion
can be done in about 98 percent nitrogen at a temperature of 1250 deg c
probably for about 9 hours or so. Ramp down of the silicon could be in
oxygen to insure no dark boron stains.

The use of the Solid Planar disc is often employed in device manufacture as
it provides a more uniform diffusion than gas which is difficult to have
flow around the si9licon slices so easily. So the planar source will dope
heavier than Sp-in on boron down to about 1.4 ohm per square without
damage. Planar sources are made several companies but you can find three of
the types of diffusion materials from [email protected] or call 724 352
3790. they have spi9n on sources, diffusion preforms and solid ceramic
boron discs that are excellent in addition to the Spin on glasses some of
which will provide a 1.3 micron thick coating of inorganic material.


Anyway , wish all a pleasant holiday seasoon!!!

Neal



>
>   As far as I know, there is a process used in high voltage diode, which
> drives the boron through the wafer from one side to the other side about
> 250um. It might need one week in a diffusion furnace.
>
>Roger Brennan  wrote:  I believe most (perhaps all)
>people driving boron to 25 um do use a furnace
>"pre-deposition". Boron disks (usually boron-nitride) might be less
>expensive to set up than say, BBr3 or B2O3. You probably need a dose in the
>range of 5E15 to 1E16 atoms/ cm2. Watch out for "boron skin" during the
>deposition. You may need to employ a low temperature "wet" oxidation to
>remove it.