I just measured a pair of LEDs "just for fun" and discovered that they're outputting energy. I held one on the LEDs in the sun's direction and the Multimeter went straight up to 10V. I laughed because I had absolutely no clue what's happening inside those things and how...

How is that possible? Are they "solar panels" in a way, because ... how can this be?

Below are two pictures of my "finding":

10.61V 8.33V

(I tried plugging an Arduino Mega on the other end to see if it would power on but unfortunately it stayed off)

I also measured the amperage between the two leads, but the only mode that work on my multimeter is the "10A" mode. It stated 0.00(A).

EDIT: The LED has this appareance: LED

  • 3
    \$\begingroup\$ All the diodes emit photons (visible spectrum or not) when a current flows through the p-n junction (The thing behind the LEDs). The reverse applies: If you cast a light to a p-n junction, a small current can flow. \$\endgroup\$ – Rohat Kılıç Sep 6 '17 at 19:58
  • \$\begingroup\$ Even the silicon diodes in glass package, can make "noise" if exposed to light. This can be observed in measurement equipment that has to be covered with dark casing to avoid such interference. \$\endgroup\$ – Marko Buršič Sep 6 '17 at 20:10
  • \$\begingroup\$ Physically, the process that allows LEDs to emit light is the exact reverse of the process that allows solar cells to yield current. Ideally you should be able to put a current through a solar cell and get light. I don't know how well it works in practice, though, since a good solar cell probably is a terrible LED. \$\endgroup\$ – Arthur Sep 7 '17 at 8:00
  • \$\begingroup\$ @Arthur I'll be trying this out someday... Might be fun seeing a solarcell glowing \$\endgroup\$ – Fusseldieb Sep 7 '17 at 10:15
  • 1
    \$\begingroup\$ @Fusseldieb Solar cells do have an electroluminescent response to reverse current -- this is actually used to inspect cells for damage that would otherwise not be visible. Whether or not it is visible depends on the type of cell used. Silicon junction cells emit short-wave infrared light the 1100 nm range (see here: pveducation.org/pvcdrom/characterisation/electroluminescence and sinovoltaics.com/quality-control/… ). Other junction types can emit visible wavelengths (though I can't find good imagery) \$\endgroup\$ – Tristan Sep 7 '17 at 14:20

It is common knowledge that all semiconductor diodes when exposed to bright light are current sources. But since the chip is small, it does not generate much current to be useful except to drive a 10M load in a voltmeter.

  • 2
    \$\begingroup\$ I didn't knew that and also found nothing about "LEDs generating energy" on the internet, so I thought I'll ask here \$\endgroup\$ – Fusseldieb Sep 6 '17 at 19:50
  • 1
    \$\begingroup\$ analog.com/en/analog-dialogue/raqs/raq-issue-108.html \$\endgroup\$ – Janka Sep 6 '17 at 19:52
  • \$\begingroup\$ @Janka Great explanation on the site. Can you post that as an answer too? (Maybe the site goes down in the future and then... you know) \$\endgroup\$ – Fusseldieb Sep 6 '17 at 19:54
  • 8
    \$\begingroup\$ This photoelectric current is one reason why most ICs are encapsulated by thick black epoxy resin: to prevent sensitive, high-impedance nodes from going haywire from PN junction current exposed to ambient light. \$\endgroup\$ – MarkU Sep 6 '17 at 20:10
  • \$\begingroup\$ @Fusseldieb Because you should be looking for "photodiode". You've included "LED" which means "light emitting diode" so that's why you got no results. Technically speaking, a LED is not a LED when unpowered. \$\endgroup\$ – Agent_L Sep 7 '17 at 10:36

Yes, LEDs are a sort-of solar panels when exposed to light, just a very small ones and not optimised to produce electricity. The main effect is called "photovoltaic effect", which is one of many physical phenomena related to transport of some carriers across semiconductors and other inhomogenuities between materials with different conductive structure. As the physics of these effects is the same, the effects are symmetrical in some sense. For example, Seebek effect is opposite to Peltier effect. Same with light and p-n junctions, an electric current causes emission of light, which become visible in specially optimized arrangements as LEDs, and the light generates current when a junction of semiconductor materials is exposed to light.

A typical GaP LED would produce DC voltage of about 1 - 1.5 V when exposed to direct sunlight, blue LEDs (typically InGaN based, and white bulbs based on them) will do 2-2.5 V per individual LED, under very light load as a DMM. If your light bulb did produce 10 V, it means that its internal construction has maybe 4-5 LEDs in series.


As already explained, leds can generate some photovoltage, but the available power is small due the low chip area and no optimization for this purpose.

The phenomena has been useful, too. An IR led can be tested without a detector. If there's as much photovoltage output from 2 leds of the same type they probably are as good. This is less needed today, when everyone has a phone camera that can see the light from the IR leds.

A fun exeriment, if you happen to have an old style CRT oscilloscope:

Connect a led to an oscilloscope which does autotriggered sweeps and has quite high input sensitivity (DC). Put the led in front of the screen and see, how the trace gets bended if the led is placed near the trace.

  • \$\begingroup\$ Your last paragraph needs some work. \$\endgroup\$ – Transistor Sep 6 '17 at 21:13
  • \$\begingroup\$ "... and see, how the trace gets bended if the led is placed near the trace." Is someone with an LCD 'scope going to be disappointed? I think this will only work with a CRT type or very bright LED type. Further edit? \$\endgroup\$ – Transistor Sep 6 '17 at 22:45

To be more general, I had a teacher who told us that every emitter can be reversed and used as a current generator.

So lighting a lamp will produce current, as will do turning an electric motor, shouting at a speaker, shocking a quartz, heating a resistor ...

This rule seems a bit simplistic to me, but I didn't find any case where it's not applicable. (comments welcome !)

  • 3
    \$\begingroup\$ Good luck trying to produce current by illuminating an incandescent bulb or heating a resistor. \$\endgroup\$ – Dmitry Grigoryev Sep 7 '17 at 12:18
  • \$\begingroup\$ I just tested the resistor : it's generating a bit of current. \$\endgroup\$ – Technico.top Sep 7 '17 at 12:34

Not the answer you're looking for? Browse other questions tagged or ask your own question.