Most specs for light bulbs list luminous efficiency (or, more accurately sometimes, luminous efficacy) but not energy efficiency. Luminous efficacy tells you how bright the light is as observed by a human eye per unit of input electricity. Unfortunately, the human eye is terrible at measuring power and its very hard to convert between lumens/spectrum/human eye/radiant power.

I'm curious as to which types of lights have the highest energy efficiency, which I'm defining the amount of energy that would be absorbed by a perfect absorber surrounding the light bulb divided by the electric input. In this metric, the light is just a one form that the energy takes between electricity and its absorbed form (eg, heat). Obviously the exergy is degraded in this process, but I'm not worried about that here.

I would expect that incandescents are near 100% efficient by this metric since the only losses should be conduction/convection from the element through the near-vacuum and wires. This weakly scientific analysis suggests that LEDs are probably only 40% efficient by this metric because all their luminous inefficiencies are thermal losses in the cells. I think fluorescent lights might be extremely efficient by this metric but I can't really find any data.

I found the phrase "radiant efficiency" somewhat useful but still can't find much in manufacturers data sheets. Is there a better name for the metric I've described or a good database for what values might be?

In response to comments: I am interested in the power radiated away from the light bulb in the form of electromagnetic radiation divided by the electric power that goes in to the lightbulb. I don't care about the wavelength of the light (even though I understand that it changes both how much you can see and how much can be recaptured in another form later). I do care about thermal losses inside the bulb that don't radiate power across a distance.

  • $\begingroup$ Incandescent lights produce a lot of heat because they are essentially a small electrical heating element in glass bulb - a very small version of a bar electrical radiator. By your criteria they are definitely not "near 100% efficient". $\endgroup$
    – Fred
    Feb 21 '17 at 7:54
  • $\begingroup$ Have you ever touched a incandescent bulb?! They convert electrical energy into 98% heat and 2% light. Of course, some of that heat is carried off as IR radiation. $\endgroup$
    – MSalters
    Feb 21 '17 at 9:25
  • 1
    $\begingroup$ Your question isn't clear. Do you want to count only visible wavelengths, or not? $\endgroup$
    – 410 gone
    Feb 21 '17 at 12:12
  • 2
    $\begingroup$ Your question reads like, "What light bulb is most efficient at converting electricity into anything other than electricity?" That is because you seem to be including infrared (heat) in your output along with the visible spectrum. $\endgroup$
    – hazzey
    Feb 21 '17 at 14:32
  • $\begingroup$ aren't they all 100% efficient by this metric? Where do you think LED thermal losses go? $\endgroup$
    – agentp
    Feb 22 '17 at 19:09

By your metric, a electrically-heated black body radiator in vacuum would be very efficient.

Incandescent bulbs are close, but run at high temperatures to make (at least some) visible light. However, the filament is not in vacuum, so incandescent bulbs have significant heat conduction losses. A typical bulb gets too hot to touch after running for a few minutes. The filament is run so hot that if it were in a vacuum, it would evaporate too quickly and coat the inside of the bulb with the filament material.

A modified incandescent that runs a bit cooler with the filament in vacuum should work nicely. At dimly glowing red temperatures, the evaporation of the filament is greatly reduced, and you should get useful run times in vacuum. The spectrum would be shifted even lower than typical incandescent light bulbs, but your spec says nothing about wavelength restrictions.

The only losses would be a little heat conduction thru the filament connections to the outside, and transmission losses thru the bulb at the wavelengths actually emitted.

Another way to do this is to start with a standard incandescent light bulb and place it within another bulb that is evacuated. Now there can't be conduction losses from the incandescent bulb, except the little bit thru the feed wires. The problem with this setup is that the inner bulb will likely overheat at full power, since it was probably designed with some air cooling assumptions. You could compensate by running the inner bulb at lower than rated power. That will shift the light more toward longer wavelength, but probably doesn't need to be as far as in the first example.

  • $\begingroup$ Thanks for the excellent discussion of incandescents. I didn't realize they were no longer under vacuum. Argon is a terrible heat conductor though so I wonder how much heat they can actually transfer. Any thoughts about fluorescents? $\endgroup$
    – ericksonla
    Feb 21 '17 at 14:00
  • $\begingroup$ @eric: We know heat gets transferred to the glass because it gets so hot you can't touch it for long. Flourescents should be less efficient by your metric because there is inherent inefficiency in converting the electrical power to mercury UV emission, then more loss in the phosphor in converting that to visible light. The most efficient by your metrics would be a "black light", which is a flourescent without the phosphor. Since you didn't constrain the radiation to visible light, the mercury emissions directly would be the most efficient. $\endgroup$ Feb 21 '17 at 17:12

I don't think your metric is particularly useful.

You even elude to the flaw yourself, incandescent bulbs should be more efficient than LEDS in your system, because they create more heat than light.

You are thinking of efficiency wrong in this case. It isn't just the ratio of energy out to energy in. It's the ratio of useful energy out to energy in.

In the case of light bulbs the main factor you want is light, not heat. That means a good measure of efficiency is the amount of light vs. Energy input (not heat vs. Energy input).

Incandescents are more efficient than LEDS when it comes to home heating, but then a lot of things heat better than lights anyways.

Sorry I couldn't directly answer your question, I'm not sure where exactly you would find the information on bulb radiation. Energy management texts would likely be a good start, I know lighting and heat management come up in that quite a bit.

  • $\begingroup$ My metric is useful to me. Telling me whether my question is stupid is not helpful. Off the top of my head, systems where you need this metric include airplane IR deicers, concentrated power solar simulators, and lightsail testing. $\endgroup$
    – ericksonla
    Feb 21 '17 at 13:53
  • 2
    $\begingroup$ @ericsonla I never said the question is stupid. You asked about the "energy efficiency" of lights. I explained that efficiency is a measure of energy input to useful energy out. In the case of lights, energy efficiency deals with the light output not the heat output, since the purpose of lights is to light, not heat. What I'm getting at is that it's not a common value, so it's unlikely to be standard in data sheets. The information is useful, which is why I suggested looking at energy management texts, where they deal with topics which can include heat due to lighting. $\endgroup$
    – JMac
    Feb 21 '17 at 14:23
  • $\begingroup$ @ericksonla It looks like you need to define what output you are looking for. Don't ask about "light bulbs" and then say that you are considering "IR deicers". As JMac said, efficiency is useful output over energy in. What is the "useful output" that you are looking for? $\endgroup$
    – hazzey
    Feb 21 '17 at 14:39
  • $\begingroup$ @hazzey True, currently airplane IR deicers use hydrocarbon fuels. But they could use lightbulbs. We can still call it a lightbulb even if its IR. $\endgroup$
    – ericksonla
    Feb 21 '17 at 14:59

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