John Larkin once said that everything electronic that emits something
such as light doesn't live forever. On every project I was involved in
with serious LEDs on there lifetime was one of the agenda items to discuss.
I assume it's like with CFL. Our first round died within the year, well
short of their 10x light bulb claims. So I gave up on this technology
for many years. Then I tried again, this time Philips Marathon. They
seem to last forever.
Well, I didn't want to be as vague as all that. I work with rare
earth phosphors, regularly. And they are pretty hardy materials,
fired usually at highish temperatures and not prone to changing
structure by the fly-speck energies found in near-400nm photon. I
mean, this is only 3eV or so. Most I've worked with are fine at 5eV
almost forever.
I find Martin's response just about exactly how I feel about it. There
are organic/laser dyes that do degrade. But these aren't what I
thought they were using in white LEDs. As he points out in what I
consider to be broad agreement with me, rare earth ceramic/inorganic
phosphors pretty much just work all day long.
I think the LED itself can indeed fail. But not the phosphor. That's
hard for me to imagine, right now. In the case you talk about with
CFLs, I've opened up just about every CFL that has ever failed in my
home and taken a close look -- dismantling every single part, in fact.
There is a particular brand I find to have very dangerous designs (it
is sold at Costco) and those burn up for just about every imaginable
reason. I've had capacitors explode and burn, transistors literally
blow their sides out, inductors which overheated and caught fire, etc.
I think they under-design the entire thing. It's all at risk. On the
other hand, as you say, the Marathon units are much better designed.
hehe. In fact, they also use a more complex method to actually attach
the circuit board between the screw end and the bulb that makes it
harder for me to gain access, too. But when they fail, it seems to be
that the parts do NOT burn up. (I've only had two fail so far from
them and I think both cases related to the fluorescent bulb itself --
but NOT the failure of its phosphor material.)
Jon