Sciencemadness Discussion Board

Which boron for p-type doping (diffusion of solar cells)

SupaVillain - 19-1-2015 at 23:48

For the silicon solar cells manufacture they have to make n-type doped side and a p-type doped side, for the n-type side they use a phosphorus they say, which i found to be pocl3 specifically, and for the p-type side they use boron....

but what kind of boron is it?

i think it's pretty much the same whether youre doing diffusion or ion implantation

Jylliana - 20-1-2015 at 02:35

I don't know much about this myself, but my father works quite directly with solar cells and the chemistry around them. I'm sure he will know what boron compound is used. I will ask him as soon as I can :)

[Edited on 20-1-2015 by Jylliana]

Jylliana - 20-1-2015 at 03:30

The compound they use is boron tribromide(BBr3).
In a special diffusion chamber they bubble N2 through a POCl3 solution and the N2 acts a a carrier gas for the POCl3. The same happens with BBr3.

In the case of ion implantation, they use diborane(B2H6) in gaseous state.
Using a high voltage, they create an electron bundle and with the help of a magnetic field they bend/curve the bundle to shoot the right molecules towards the silicon wafers(a bit the same principle as a mass spectrometer). Too heavy or too light molecules will go past the wafer, while the diborane ends up on the wafer.
For the ion implantation on n-type semiconductors they often use phosphine(PH3) instead of POCl3.

My english is not great, and I find it a tricky subject to translate and explain properly, but I hope I could be of some help to you :)

[Edited on 20-1-2015 by Jylliana]

SupaVillain - 20-1-2015 at 11:15

Yes thats actually everything i needed to know thank you so much!

DrMario - 20-1-2015 at 13:33

The cheapest method for doping silicon is spin-on-glass (abbreviated as SOG). Even the poorest, least equipped cleanroom, or even non-clean-room lab would be able to do SOG doping. My colleagues make actual, working silicon solar cells with spin-on-glass doping, so I know for certain that it is a viable route. The only difficult detail is the drive-in process which is done in a furnace at high temperatures of at least 800C for a realistic diffusion time, and even so, it takes hours. So you need a pretty serious furnace. But this problem is the same pretty much regardless of the doping method you use. Even with ion implantation, you need to anneal your silicon wafers. So, I imagine that you have that part of the equation solved somehow, if you are intent on doping silicon.

Incidentally, what kind of silicon wafers do you plan to use?

DrMario - 21-1-2015 at 08:16

And another thing: you must do at least a thorough RCA-1 clean before doping and drive in. You must understand that even the smallest metallic impurities will be driven in during the drive-in (doh, you say) - aluminium, for instance, is an N type dopant, and many, MANY transition metals will render the PN junction poorly functioning or disrupt it altogether (they add energy levels around the middle of the band-gap of silicon). Gold is the worst, BTW.

I would think that this level of cleanliness is outside of the capability of most amateurs, but then again, I thought the same about making a functioning magnetron sputter - and I was proven wrong.

SupaVillain - 2-2-2015 at 14:26

Uhhhhh i was assuming a bulk order of crystalline silicon wafers from ebay they had some square shaped ones but honestly after looking at the solar panels that get 46% efficiency as opposed to the 14%-20% of normal panels. I really am just going to have to do more research into that. If you check out
youll see what im talking about but im unsure if thats 46% efficiency for the surface area of the mirror or of the layered solar panel.

Thank you for the doping information on cleanliness I know nothing of this metallic impurity issue. Ive been studying basic manufacturing processes basically... of the different types, including spray on flexible solar panels

Also yes magnetron sputtering is very DIY thats what Im working on now to make AFM microscope probes/tips

DrMario - 5-2-2015 at 15:55

What material will you use to make the AFM tips?

Regarding those "square shaped" silicon wafers, those are very likely polysilicon wafers. They are kind of cool, much cheaper than single-crystal silicon, and the yare EXTREMELY fragile. One of my colleagues was involved with the evaporation of copper conductors on solar cells made from polysilicon wafers, and he broke half a dozen in a month, even though he knew what he was doing. For an amateur, I would very highly recommend to buy single crystal Si wafers, as they are easier to handle. However.... the issue of contamination is one not to take lightly: you don't necessarily know the hitory of those wafers you get on eBay.

Also, did you say you have a drive-in furnace for the doping, or not? We're talking of 12-24 hours at 800+ degrees C.

And I am not at all convinced you have the cleanliness conditions to not accidentally dope your wafers with some transition metal that will act as recombination point.

[Edited on 6-2-2015 by DrMario]

SupaVillain - 5-8-2015 at 23:53

I've read that RCA clean that you linked for me, I don't understand how that would be difficult at all.

RCA clean (only one step)

clean room DIY

I'm stuck trying to decide between thin film CIGS type cells and the monocrystalline silicon kind. I've heard amorphous silicon is much more expensive so I'll choose one of the two above.

AFM tip sI would make with silicon wafers, and unfortunately they would require some silicon nitride films made.

SupaVillain - 10-8-2015 at 18:26

Anybody that's interested in making their own silicon solar cells, check out doping them with phosphoric acid instead of phsophorus oxychloride, efficiency of 18.7% has been achieved, no need for white phosphorus hassle