Nichia soon to release 460nm diodes and DI green diodes

[Lasermad]

Just had a look at the website of nichia , in japan , with is not a cheap production country like china and LOOK at the amount of factories and officies they have , they clearly employ many thousands of people , never mind the amount of cash they must throw into development
http://www.nichia.co.jp/en/about_nichia/locations.html

In short the laser show industry volume will not even compute on their rader , worldwide laser show related purchases are likely 0.001 % of their turnover

By the looks of things they aint happy .... you need to AGREE to their terms just to look at their products ... http://www.nichia.co.jp/en/product/laser.html

and they have made it clear ...........
"Nichia prohibit Purchaser from reverse engineering, disassembling, or taking any other steps to derive the structure or design of the LD"

These developments wont be along "cheap" soon ...maybe if iphone with Picolasers or Pico projecotrs take off but i dont think we will all be kicking around with ziggawatt RGBS in the near future :-(

Paul

[White-Light]

Thats a standard clause. It actually relates to reverse engineering the diode to find out how its made, (with a view to copying its design), not reverse engineering what its put into.

[ixfd64]

I met with Tonto from Nichia last Thursday - he mentioned that a 460nm, 1 Watt diode source will be available within the next 6 months, similar price point to the 445nm diodes (ie, 3000-4000 per unit in sample/single quantities).

Too bad they'll probably be multi-mode.

[pullbangdead]

Too bad they'll probably be multi-mode.

It's all relative, and it depends. A little company called Soraa just released a paper detailing single-mode blue diodes that were quite impressive. The blue device they showed in the paper was a "non-c-plane" single-mode diode at 447nm that was outputting 750mW CW. Danged impressive. The same paper showed single-mode 520nm diodes running CW at >60mW.

Of course the commercial stuff is still on c-plane now and you're right, it'll likely be pretty close to the performance of the 445nm diodes we're seeing now. But still, the future is likely quite bright (pun intended).

[ixfd64]

This is a stupid question, but what does "c-plane" mean, in the context of laser diodes?

[mixedgas]

This is a stupid question, but what does "c-plane" mean, in the context of laser diodes?

Ah, The part of the physical geography class that caused me to do my best to skip and cut class. Crystallography.

If you have a material that is non symmetrical at the atomic level, , it has different materials properties along different planes that intersect certain orientations of the crystal.

Traditionally, natural Quartz, which has a lot of funky optical properties, and was the easiest thing to find and work with, is the reference, and Quartz' longest axis in a natural crystal is the optically active one, (I think) so it became Z when "they" developed the co-ordinate system.

C plane, on a wafer, is the easiest usable plane to cut and polish out of the whole X-Y-Z orientation system for gallium arsenide wafers.

The laser diode is grown on top of the wafer, and each of its layers have different materials properties. This causes stress at the junction between the layer and the wafer.
So far, the shorter the wavelength you go, the worse the mismatch from strain. It was thought other less optimal, harder to make orientations of the wafer would have to be used to eliminate the strain for materials with quantum wells in the green, if it could be done at all.

So if some one came up with a way to deposit layers on the C plane with less stress, green just became as easy to make as blue and violet.

I'm sure some one will be along to correct me when cali opens for business. The correction will be welcome. Cue our resident diode grower.


Wiki tyme:

http://en.wikipedia.org/wiki/Crystal_structure

Quote
"The crystallographic directions are fictitious lines linking nodes (atoms, ions or molecules) of a crystal. Similarly, the crystallographic planes are fictitious planes linking nodes. Some directions and planes have a higher density of nodes."
End quote.

Start reading the wicki right after the above line.

Steve

[pullbangdead]

Ah, The part of the physical geography class that caused me to do my best to skip and cut class. Crystallography.

If you have a material that is non symmetrical at the atomic level, , it has different materials properties along different planes that intersect certain orientations of the crystal.

Great description.

Traditionally, natural Quartz, which has a lot of funky optical properties, and was the easiest thing to find and work with, is the reference, and Quartz' longest axis in a natural crystal is the optically active one, (I think) so it became Z when "they" developed the co-ordinate system.

C plane, on a wafer, is the easiest usable plane to cut and polish out of the whole X-Y-Z orientation system for gallium arsenide wafers.

The laser diode is grown on top of the wafer, and each of its layers have different materials properties. This causes stress at the junction between the layer and the wafer.
So far, the shorter the wavelength you go, the worse the mismatch from strain. It was thought other less optimal, harder to make orientations of the wafer would have to be used to eliminate the strain for materials with quantum wells in the green, if it could be done at all.

So if some one came up with a way to deposit layers on the C plane with less stress, green just became as easy to make as blue and violet.

And to this I'll add some and correct a slight bit.

So GaAs is used as the substrate on the red end of the spectrum. But in the blue end of the spectrum, sapphire, silicon carbide, or gallium nitride itself is used as the substrate. The most common is growing gallium nitride devices on sapphire. Commercially, when they do this, you get c-plane GaN on the sapphire wafer. Exactly as you say though, strain is a BIG problem. You get a lot of misfit dislocations due to strain, and it's worse as you go from violet to blue to green. GaN is more forgiving than the materials used in the red end, it's pretty amazing it works as well as it does to be grown on sapphire. They can do things to get rid of dislocations though, like ELO (epitaxial lateral overgrowth, which cuts off and removes many of the dislocations). But there's another big problem with using c-plane in gallium nitride violet/blue/green devices: polarization.

The material itself is polarized in the c-direction, it's piezoelectric, and if your device is on c-plane, there's a built-in electric field between each layer of atoms that is opposing your device. It's especially bad in the quantum well, where it leads to QCSE, the quantum-confined Stark effect. Basically, the electric filed is holding the holes and electrons apart, which causes them to be spatially separated and makes it harder for them to recombine. In GaN, if you can use a different plane to grow your device, you can orient that built-in electric field so that it doesn't oppose your device. If you go to m-plane or a-plane, that field is perpendicular to your device, so those are referred to as non-polar planes. Other planes at different angles inbetween have some polarization field, so they're referred to as semi-polar planes, and they have some other advantages over the non-polar planes, in properties such as strain/strain relaxation.

So, if we can make a pure GaN crystal/substrate of arbitrary size, we can eliminate the strain problems of growing on sapphire, and we can also slice it in different directions in order to work around the QCSE that makes the device inefficient.

Everything GaN commercially available now is on c-plane, because c-plane is the only plane on which full-size 2-inch wafers are available. They can actually make 2-inch GaN wafers by using HVPE on a 2-inch sapphire wafer, lifting it off, and then slicing that GaN "hockey-puck" into 2-inch wafers, which is generally what they do for laser diodes (they often don't bother with that with LEDs, they just grow on the sapphire itself). They can make other planes by slicing that same hockey puck in different directions, but clearly those slices can't be as large, and therefore aren't commercially viable. But maybe one day soon we'll be able to make full-sized wafers on non-polar and semi-polar planes of GaN, and device performance will improve.

Although it is absolutely amazing what Nichia (and some other companies, mostly Nichia) has been able to do on "traditional" c-plane GaN, where other companies have only been able to achieve many of the same things by using non-polar or semi-polar GaN.

I'm sure some one will be along to correct me when cali opens for business. The correction will be welcome. Cue our resident diode grower.

Who's the "resident diode grower" around here? I am one, but I don't think you knew that and I didn't realize there was another.

[ixfd64]

Thanks for the detailed explanations, mixedgas and pullbangdead. Too bad I can't +rep you here.

[mixedgas]

PBD,

PM sent with the username.

Thank you for the correction, I helped build thousands of a product that used Nichia GAN on Sapphire leds when they first came out. More often then not I am using Ga-As semiconductors in RF, I crossed the mental streams at 5 am when I could not sleep.

Again, thank you for a great post.

Steve

[ixfd64]

I wonder if those "non c-plane" diodes will have better beam specs (i.e., divergence and diameter) than their traditional counterparts...