Another idea, how low can you power the phosphor light? Run it at really low power (moonlight or wherever you can look at it without zapping your eyes) and see what kind of light emission each led has. This may lead you to where the remaining phosphor is.
The UV trick actually worked great, but the was almost no phosphor left, perhaps there was another layer I couldn’t see with the UV that wasn’t removed by scrubbing with denatured alcohol.
I swapped out the emitters last night. Managed to get the PCB out and put it on a hot plate to reflow. It was almost a slam dunk but I dropped the glass on the desk and it chipped. Now I need to figure out how to make it waterproof again, maybe I’ll silicone it.
I’ll post some photos soon.
All that yellow stuff to the sides of the die is phosphor. These chips emit blue light to the side, so that phosphor would be excited.
There is red being rendered in that ‘phosphorless’ photo. Not possible if there wasn’t phosphor
Even if there was some phosphor contaminating peripheral light I would still expect to see fluorescence if the bandwidth is narrow enough. My guess is that without the phosphor these lights still have a FWHM that is too broad for 455nm excitation.
Maybe I’ll give it another shot in the future using a different solvent.
I don’t understand your line of thinking. Even if excitation only occurred in a 1nm range around 455, then ANY light source containing 455nm would yield fluorescence. The problem, as you’ve stated yourself, is when your excitation source spectrum overlaps with the emission spectrum.
If you seriously use a wooden toothpick and remove all traces of phosphor mechanically, you’ll see the difference.
It is plainly obvious from the photos that a meaningful fraction of the output was phosphor converted.
I understand you already swapped in your tried and true XTEs, but the record should be set straight about the phosphor removal process
If your filter is blocking below 500nm light, I don’t see how a 450-455nm source would be that much better than a 450-470nm source. Unless your royal blue XTE spectrum is actually higher energy (lower wavelength) than a de-phoshored white LED. If your absorption spectrum changes a lot around 455nm then that could account for some of the changes you see in your plot in post #10.
However, as Jared pointed out there is most likely still phosphor remaining on your XPL2s which will definitely increase your background. You don’t have to use a solvent. Just scrape it off with a toothpick until you don’t see any more yellow stuff.
Okay I’ll mount one on a PCB when I have time and give it a whirl. Jared, do you have a photo of what your LEDs look like lit up without the phosphor?