Luxeon Vf shift over time

Candle Power Flashlight Forum

Help Support CPF:

Newbie,

Both lux and Vf measurements are taken with the luxeon running at the test current for at least 5 seconds. At high currents, I did notice an initial drop from powering the luxeon of about 0.01-0.02V, and 0.2-0.3 lux, then the reading stayed steady. I leave things running for a few seconds after the reading just to make sure the reading I took was a steady value.
 
Please take my calculation with a grain of salt... the color shift, for the measured Vf drop of a blue emitter, should be about 10 nm, or a bit less, toward the green, at 25 C.
It cannot be appreciated by naked eye, but it could with the CS-1000 NewBie can access. The key to the success of this measurement is the temperature control, that cannot be correctly achieved by passive heatsink.

Anthony
 
Though a K-type thermocouple embedded into the slug of the emitter, hooked into a servo loop with a peltier could do a great job of holding the temp. If the target temp was slightly warmer than normal, you could also do the loop to control a fan...

Today I took a 461nm CREE, and a 458nm CREE, and a 470nm CREE (XL7090 series), and saw 1nm shift from 5% to 95% duty cycle at 350mA in all cases. I took one and ran it over night and saw no output shift the next day.

I'll have to take in a blue LumiLEDs Luxeon with no time on it and see if it exhibits a different behavior.
 
evan9162,
I lost your E-mail address but not your mailing address. You have a couple T bin luxeon stars in J Vf as well as a couple in L Vf coming your way. Burn 'em up! /ubbthreads/images/graemlins/nana.gif
 
Even without knowing the actual mechanisms involved in the Vf drop, if I were to make a WAG I would guess that the higher Vf binned emitters would on average exhibit a greater drop in Vf. This is based solely on the fact that they have further to go before bumping up against the minimum possible Vf defined by the quantum physics involved. [see formula in Newbie's earlier post]. BTW, for you quantum mechanics enthusiasts, I believe that it is possible for an LED to emit a *very small* amount of light at a Vf below the theoretical minimum determined by the formula in Newbie's posts due to the tunneling phenomenon associated with the Heisenberg uncertainly principal. Note, this is purely conjecture on my part.
 
After reading some notable quotables on their take on quantum physics, I know enough to run and hide if you guys start talking about it here! /ubbthreads/images/graemlins/nana.gif
 
Don, you can run but you can't hide. /ubbthreads/images/graemlins/wink.gif And BTW, /ubbthreads/images/graemlins/thumbsup.gif /ubbthreads/images/graemlins/thumbsup.gif /ubbthreads/images/graemlins/thumbsup.gif for your generous contibutions to this effort.
 
[ QUOTE ]
Doug S said:
BTW, for you quantum mechanics enthusiasts, I believe that it is possible for an LED to emit a *very small* amount of light at a Vf below the theoretical minimum determined by the formula in Newbie's posts due to the tunneling phenomenon associated with the Heisenberg uncertainly principal. Note, this is purely conjecture on my part.

[/ QUOTE ]
If you measure VF with a DMM (with something like 1mA) you get typical readings of 2.4..2.5V which is only sufficient to generate green light, but the LED is glowing white => seems to be the effect you describe.
But I have attributed this to a different effect: The thermal energy at room temperature is kt=25meV. And there is a small chance, that the electons or holes gain some kt energy (by absorbing phonons) before radiative recombination => these thermally activated electrons and holes can reach the necessary 2.6V and are responsible for the small light emission.

My first measurements also indicate that lower VF has smaller drift. I think we see here a change of the resistance overhead, but not a change of the light emitting pn junction itself.
R2k: 123mV@24h
QWOH: 90mW@70h
 
[ QUOTE ]
PeterB said:
If you measure VF with a DMM (with something like 1mA) you get typical readings of 2.4..2.5V which is only sufficient to generate green light, but the LED is glowing white => seems to be the effect you describe.
But I have attributed this to a different effect: The thermal energy at room temperature is kt=25meV. And there is a small chance, that the electons or holes gain some kt energy (by absorbing phonons) before radiative recombination => these thermally activated electrons and holes can reach the necessary 2.6V and are responsible for the small light emission.


[/ QUOTE ]
Peter, your explaination sounds plausible to me. Now, has anyone seen that Don fellow?
 
Peter,

Using an LM317, the ADJ pin outputs about 50uA. In my current supply, I have switches that I can switch on (and combine) different current levels. When all are turned off, I essentially have a 50uA current supply, as it's simply being supplied by the ADJ pin. At that current, even royal blue Luxeons still light up. I don't have my data with me, but I also see around 2.4-2.5V Vf at that current.

One thing that is useful for is I can make a good general prediction of the color of the emitter (white luxeons) at high current levels, but without blinding myself in the process. One day, I need to also see if there is a correlation between brightness at 50uA and brightness at high current levels.
 
evan,

For the royal blue (center is around 450nm, right?) the photon energy is even 2.75V! Now we have to show, that the LED cools down due to the Phonon absorption /ubbthreads/images/graemlins/smile.gif
 
[ QUOTE ]
PeterB said:
My first measurements also indicate that lower VF has smaller drift. I think we see here a change of the resistance overhead, but not a change of the light emitting pn junction itself.
R2k: 123mV@24h
QWOH: 90mW@70h

[/ QUOTE ]

Peter, data posted by Darin of Vf drop vs current shows that the drop is not linear with current. This at least demonstrates that the change is not simply a reduction of bulk resistance in the device. Reduction in bulk resistance may be occuring but it appears that additional phenomena are involved.
 
Planck formula is only a "relationship" to connect the color of the emitted light to what we call Vf but is in reality the energy band gap of the PN junction.

Tunnel effect in semiconductors requires a doping 100 times stronger than that used in LEDs.

Band gap energy (the Vf) can estract photons from the cristalline reticule also at lower value than nominal for blue, due to a phenomena named "photon scattering" (due to borrowed energy) and the theory of the liquid drop. Those "extractions" are statistically irrelevant for what we are looking for.

I believe is not the case to start a course of quantum physics here, it is not needed since we, IMHO, are looking for experimental data. 2 €c on this.

Anthony
 
Newbie,

I noticed that shining light into a Luxoen produces voltage across the terminals. I tried shorting it through the 10A scale on my DMM, but didn't read any current flowing, so it might have been less than 10mA (the lower limit on that scale).

Any idea how much current can be generated?

[edit]
I just measured a white Lux III. Putting the bare dome of a white LuxV at 750mA up to the LuxIII dome, I measured 1.4V across the Lux III. On the 2mA current scale, I measured 0.002mA (or 2uA).

I tried lighting up the lux III with a 35W MR16 halogen light, which didn't cause any voltage. A Red-Orange Luxeon also didn't produce any voltage on the Lux III, neither did a royal blue Luxeon V.
 
[ QUOTE ]
evan9162 said:
I noticed that shining light into a Luxoen produces voltage across the terminals. I tried shorting it through the 10A scale on my DMM, but didn't read any current flowing, so it might have been less than 10mA (the lower limit on that scale).

[/ QUOTE ]
I just tried this. With a white Luxeon I get very little voltage and maybe about 0.1 uA of current when placed right next to a 20W halogen desk lamp bulb. Presumably the phosphor keeps most of the light from reaching the die. An amber Luxeon gives me about 1.6V open circuit, and a current of close to 10 uA under the same conditions. In effect the LED is acting like a very inefficient solar cell.
 
Here's the 2 TY0Ls:

l3ty0lVffinal.png


and the data:

<font class="small">Code:</font><hr /><pre>
#1 #2
current Vf Vf delta (Vf) Vf Vf delta (Vf)
(mA) 0 hrs 100 hrs 0 hrs 100 hrs
130 3.24 3.08 0.16 3.29 3.04 0.25
310 3.52 3.30 0.22 3.57 3.25 0.32
620 3.78 3.53 0.25 3.83 3.47 0.36
910 3.94 3.68 0.26 3.98 3.61 0.37
1540 4.19 3.92 0.27 4.23 3.84 0.39


</pre><hr />


Vf didn't drop as much as the TV1Ks, but, it looks like the slope is more gradual. I may consider running these two for an additional 100 hours to see if they drop more.

TX1Js are on now. Will have results by Christmas (may not get posted then).
 
You must log in or register to reply here.

Latest posts

Back
Top