[labnetwork] CVD Diborane Diffusion

[John Shott]

On 8/9/2012 4:44 AM, Siva Penmetsa wrote:

You are talking about an atmospheric pressure pre-deposition correct?  
While it's been many, many years, we used to use diborane as a 
predeposition source and my failing memory recalls that we would expect 
a sheet resistance of about 45-50 ohms/square following a 30 minute 
predep at 1000 degrees C prior to drive in.

I trust that you are actually using fairly dilute diborane in a carrier 
gas, correct?  What is the diborane concentration of your source gas and 
what is the carrier gas?

While my recommendations might vary somewhat depending on those answers, 
one thing jumps off the page at me when I read your process details:  2 
SLPM of oxygen is a very large flow oxygen when your total flow is about 
5 SLPM.

While you need enough oxygen to fully oxidize the amount of diborane 
injected to convert it into B2O3 ... which is ultimately what does the 
doping, too much oxygen is a bad thing because it creates an ever 
increasing oxide layer at the silicon surface that tends to isolate the 
B2O3 layer (that is deposited on top of that oxidizing surface) from the 
silicon surface.  I believe that it is likely that you would see lower 
sheet resistance and better reproducibility if you lowered your oxygen 
flow significantly .... while I can't quite recall details from my 
youth, I'm guessing that an oxygen flow of 100-200 SCCM is probably a 
lot closer to where you want to be to fully oxidize the diborane but not 
have such a high concentration of oxygen that the thermally grown oxide 
prevents the glassy B2O3 layer from doping your silicon surface.

You need to look carefully at how much diborane you are actually flowing 
at various flow conditions .... assuming that you are using a 
comparatively low concentration of diborane.  As I recall, for example, 
we used to use something on the order of 1000 ppm diborane in nitrogen 
as our source .... certainly well less than 1%.  If you are using a 
comparable concentration, it only takes a few SCCM of oxygen to fully 
oxidize your diborane.  Rather than adjusting diborane flow, I'd fix 
your diborane flow at 40 SCCM and then back down your oxygen flow 
significantly.  I suspect that you'll see lower sheet resistance and 
better reproducibility because the oxidation of the silicon surface will 
no longer be "winning".

This is all from many years ago from someone whose memory is not that 
great, but that is where I would start.

Good luck,

John

> Hi All,
> We are trying to optimise Diborane furnace in a new CVD equipment.
> We have performed few trials(all the souces are gases) at
> Temperature 1000 C
> Diborane Flow rate in the range 40 sccm to 100 sccm
> Oxygen flow rate 2 SLPM
> Nitrogen flow rate is 3 SLPM
> Time 30 minutes
> Instead of decrese in resistance we observed that the resistance 
> incresed to around 500 ohm/sq from around 50 ohm/sq
> We apprecite your experience and inputs on how we can resduce the 
> resisitivity with better Diborane diffussion.
>
> -- 
> Thanks & Regards,
> *Siva Prasad Raju Penmetsa*
> /Senior Facility Technologist/
> National Nano Fabrication Center
> /Indian Institute of Science(IISc)/
> Bangalore, India 560 054

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://mtl.mit.edu/pipermail/labnetwork/attachments/20120809/d53f017f/attachment.html>