Very strange LED die behavior

[saabluster]

In the process of binning LEDs with a setup which can be seen here I found one R2 that exhibited a very strange behavior. I would be very interested to hear your thoughts and opinions as to why this is happening.

These pictures were very hard to get. What you see is a localized area of significantly higher light output than the rest of the die. Also unusual is the fact that the spot was of a cyan color.

The spot remained regardless of current to the LED.

Is this a current spreading issue? Is it some impurity in the die? Is this bad or good? Could make for one heck of a good thrower behind the DEFT lens with the higher surface brightness.:cool:

 

[Canuke]

I'd guess at an imperfection in the dopant distribution. You say it's a cyan color, so I take that as meaning that there's a tiny spot that is outputting at a slightly lower photon energy. That means that the wavelength is longer, and the Vf a bit lower. That's why the color shifts to cyan, and why it gets brighter sooner than the rest as V increases. I don't recall whether it's more or less dopant which would do that.

 

[Fallingwater]

The die underneath the phosphor is actually cyan/blue. What that is is probably a small deficiency in the phosphor (perhaps a bubble?). In that tiny spot there's less phosphor than elsewhere, so you're seeing the bare output of the die underneath.

 

[Illum]

DM51 would say its an aurora phenomenon when a chip weevil accidentally crossed into the electron stream between the substrate and the phosphor:D

hmm...interesting:)

 

[saabluster]

The die underneath the phosphor is actually cyan/blue. What that is is probably a small deficiency in the phosphor (perhaps a bubble?). In that tiny spot there's less phosphor than elsewhere, so you're seeing the bare output of the die underneath.

The dies they make these white LEDs from are royal blue which is a deep blue color. I admit that a flaw in the phosphor was one of the first things that popped into my head but it does not look like you are peering at the die itself when you see it in person. I have actually seen a bare die before so I know what it looks like. It looks more like a doping issue.

 

[2xTrinity]

The dies they make these white LEDs from are royal blue which is a deep blue color. I admit that a flaw in the phosphor was one of the first things that popped into my head but it does not look like you are peering at the die itself when you see it in person. I have actually seen a bare die before so I know what it looks like. It looks more like a doping issue.

The fact that the "flaw" is brighter, relative to its surrounding areas, when the overall output is low suggests a doping issue. A region with different doping concentration such that the LED emits at a longer than normal wavelenth (cyan not deep blue). In the shots showing the phosphor more completely lit up, that relative "bright spot" goes away.


Also, the cyan would also be a too long of a wavelength to excite the phosphor, which is why it appears as a briht cyan, and not just a (relatively) brighter white.

 

[LukeA]

As mentioned before, it's a chip weevil burrow that's broken through the surface. What you're seeing is the photons escaping from the weevils' winter stash.

 

[csshih]

... not those chip weevils again. ;)

 

[Ryanrpm]

Notice that the spec of brightness is right over the wire woven through the chip. Like you said, if only there was a way to have the entire surface covered with that brightness!! That's what?.....about 5 times brighter?

 

[bshanahan14rulz]

it's lasing. or a chip weevil. perhaps the light emitting region right there is really thin. or a hole in the chip allows the current spreader to be closer to the other "electrode" on the bottom of the chip. No, no, it's definitely those pesky weevils

 

[Illum]

yep, heres a mug shot


you see the phosphor powder on the snout! ha, an oval evil weevil swirl of fur:shakehead

 

[R33E8]

Just wondering, do you know the forward voltage of when it just starts to light?

 

[Ryanrpm]

Just wondering if we ever came to a definite conclusion on what was causing this behavior....

 

[saabluster]

Just wondering, do you know the forward voltage of when it just starts to light?

1.9V

Just wondering if we ever came to a definite conclusion on what was causing this behavior....

I am fairly certain their is a dopant issue. It does seem to go away after you get the current up to normal levels. The color is still there but the dramatic difference in surface brightness is not there at higher levels. I used two different ways to look at the die at higher current levels. I took polarized lenses and turned them where almost all of the light is blocked and looked at it with the naked eye. Then I took an optic and projected the image of the die on a white piece of paper. This turned out to be the better method as I could still see the artifact at the higher current when I could not see it with the polarized lenses. With the image projected you could still see the greenish cast but it was being washed out by the white around it.

 

[LukeA]

Have you looked at the LED driven at a moderate level through a filter that allows you to see the die?

 

[saabluster]

Have you looked at the LED driven at a moderate level through a filter that allows you to see the die?

I don't know what you consider moderate. It started to get washed out after about 30-40mA. Beyond that and you really could not tell it was there even looking through the filter. It was not a high quality filter however(two used lenses from polarized glasses turned off axis from one another). It might work better if I had some type of dark cyan filter but I don't know if that exists.