Making Contacts On GaN grown on Sapphire

[Spinne]

I was looking at a GaN wafer grown on Sapphire with an AFM for a former associate of my professor who now works for a tech company in Boston. The wafer is for a high output, low specteal distribution LED. The basic construction is Sapphire substrate>p-type GaN layer>Quantum Wells consisting of thin alternating GaMgN and GaAlN layers>n-type GaN. The idea is that the alternating thin GaMgN and GaAlN layers form quantum wells with a very high density of states. The p-type GaN and n-type GaN make the well steps inclined so it's easy to inject electrons across the device. The high density of states results in very well defined energy jumps (instead of jumping from the continuum of states in the conduction band to the valence band, you're jumping from a small number of states to another small number of states in each quantum well), so the light emitted has a very low spectral distribution. My question is that since you need to put ohmic contacts on the n-type GaN layer as well, as the (sandwiched between the Sapphire substrate and the quantum wells) p-type layer, how do you make a contact with the p-type layer? When I asked the guy about that, he said that the answer is a secret. So anyone wants to guess?

 

[TuxDave]

Assuming you're only problem is just physically reaching the p-type layer, is there an area that does not contain the quantum wells? I'm imagining that if you looked from the top view the quantum well mask boundary, outside of that you have your p-layer mask because I don't think you can align them both perfectly. Can't you just drop a contact straight down over there? Or maybe diffuse the sapphire substrate beneath with some p-type substance and tap the substrate instead?

Btw, why is he using a sapphire substrate? Is it because that's the only way to make a bond between the GaN?

 

[bobsmith1492]

Electroplate it on with a few seeding molecules placed on the wafer? Or do you mean between the wafer and the sapphire (buried side?) ?

 

[Spinne]

I'm not sure if there are areas with no quantum wells on the periphary. In any case, the ultimate goal is to chop the wafer into diodes, so the periphary won't help much. Sapphire is usually chosen as a substrate for GaN because it is easy to grow GaN on Sapphire and so you can get a heteroepitaxial thin film pretty easily. N has a very low solubility in Ga, so preparing a GaN substrate is hard.

Originally posted by: TuxDave
Assuming you're only problem is just physically reaching the p-type layer, is there an area that does not contain the quantum wells? I'm imagining that if you looked from the top view the quantum well mask boundary, outside of that you have your p-layer mask because I don't think you can align them both perfectly. Can't you just drop a contact straight down over there? Or maybe diffuse the sapphire substrate beneath with some p-type substance and tap the substrate instead?

Btw, why is he using a sapphire substrate? Is it because that's the only way to make a bond between the GaN?

 

[Spinne]

Yeah, I mean between the GaN and Sapphire.

Originally posted by: bobsmith1492
Electroplate it on with a few seeding molecules placed on the wafer? Or do you mean between the wafer and the sapphire (buried side?) ?