The Naked Scientists

The Naked Scientists Forum

Author Topic: Is there any potential for long term eye damage from LED lighting?  (Read 4528 times)

Offline graham.d

  • Neilep Level Member
  • ******
  • Posts: 2208
    • View Profile
I was not sure whether to ask this in the physics section or here because it crosses both field to some extent. Incandescent lights (the old filament bulbs) give out light over a wide spectrum which we perceive as white because the detectors in the eye see an even amplitude in each of their RG and B detection ranges. Fluorescent lights do not but, to an even greater extent neither do LEDs. Although some LED lamps also rely on a fluorescent coating, generally these lights broadcast in 3 very narrow spectral regions. Is there any reason to be concerned that the very sharply defined spectral lines would affect our eyes - at least some of the receptors - by having a high intensity in these small spectral regions? Whilst natural light is also "spiky" because it comprises emssion spectra there is a lot of spectral lines involved. This is not so with LEDs.


 

Offline Bored chemist

  • Neilep Level Member
  • ******
  • Posts: 8667
  • Thanked: 42 times
    • View Profile
" Is there any reason to be concerned that the very sharply defined spectral lines would affect our eyes - at least some of the receptors - by having a high intensity in these small spectral regions?"
No, not really. There's no mechanism by which the eye can tell if the lines are sharp or diffuse. Also, every photon that hits the eye is "spikey"
In my experience, unlike fluorescent lights, the LED lamps don't have narrow emission lines. There's a fairly sharp blue emission from the led itself then a yellowish "white" light from the phosphor.
http://en.wikipedia.org/wiki/File:White_LED.png

http://bmb.lcd.lu/science/compared_spectra.jpg

There is possibly some concern over the relatively large amount of blue light they produce.
http://en.wikipedia.org/wiki/Light_effects_on_circadian_rhythm#cite_note-Warman-3
but that could be avoided (at the cost of a slight drop in efficiency) by using a different phosphor
 

Offline graham.d

  • Neilep Level Member
  • ******
  • Posts: 2208
    • View Profile
I did not think all LEDs used a phosphor. If so then they would probably be no worse than a fluorescent lamp. In fact, from the link you posted, BC, they are somewhat better though not as flat a spectrum as the incandescent black-body spectrum. Another spectral graph for RG&B LEDs is shown here:

http://en.wikipedia.org/wiki/File:Red-YellowGreen-Blue_LED_spectra.png

This has a much spikier spectrum. I know the eye can't tell if a spectrum is spikey or diffuse. I was wondering more about how the specialise colour receptors (cones) in the eye may react longer term to a high intensity in narrow wavelength ranges compared with what would be that from natural light. Would there be any compensatory dimming of the response at these wavelengths after long term exposure? I don't recall ever seeing any studies on this but then I probably don't read the appropriate journals.

BTW, I see that you work in spectrophotometry (in another recent post). My first job in 1970-72 was as a designer for Rank Precision Industries (formerly Hilger and Watts) and I designed the analog electronics for a revision of the scanning spectrophotometer (it scanned photographic plates of spectra measuring the intensity of the different spectral lines). I also worked on their huge emission spectrometer which was used to analyse (for example) the chemical content of steels so as to provide on-line feedback in its manufacture. It had an "arc unit" that operated and a dangerous voltage which, by arcing onto the sample, produced the light, and the measurement unit had a diffraction grating with an array of about 80 photomultipliers behind critically positioned slits to measure the intensity of specific spectral lines. The software then made corrections to remove interfering effects between the spectral lines. It was quite advanced at the time; the software was all written (not my me) in assembler for a PDP8.

I guess spectrometer are rather more sophisticated today.
 

Offline Bored chemist

  • Neilep Level Member
  • ******
  • Posts: 8667
  • Thanked: 42 times
    • View Profile
There have been changes, though some things have changed less than others. The ICP/MS at work still uses a triode valve to generate the RF power for the plasma. (3CX1500D3- I have an old one which I use as a paperweight)
The changes have made it easier to use new kit, but that's still no substitute for good basic design.
Also, while I may get laughed at for saying this on a science site, the new stuff has no "soul".
It's all beige boxes.
I can't argue with the fact that they do the job, and do it much better than their predecessors, they just lack "something".
Nostalgia ain't what it used to be.

(And I think the first lab I worked in may well have had one of those beasts in it or a close relative)
« Last Edit: 22/03/2013 11:17:32 by Bored chemist »
 

Offline CliffordK

  • Neilep Level Member
  • ******
  • Posts: 6321
  • Thanked: 3 times
  • Site Moderator
    • View Profile
There seem to be two competing modes of making LED lighting.  One with a single bright bulb, vs an array of smaller bulbs. 

It should be entirely possible to build an array of LED bulbs with varying colors, so the sum still appears to be white. 

However, as with sodium lights, part of their advantage is to be extremely bright in a narrow spectrum, and thus overall the lights appear to be brighter. 

Broader spectrum lights will give better color definition.
 

Offline evan_au

  • Neilep Level Member
  • ******
  • Posts: 4123
  • Thanked: 245 times
    • View Profile
Commercial colour LEDs used a very pure compound semiconductor material which produced a narrow line spectrum: infra-red, then visible red, yellow, green, blue and violet, in rough order of difficulty & date of manufacture. You can almost date the decade of an electronic product by the colour of the LEDs it uses - the first decade of the 2000s was the decade of the blue LED.

With a red, green and blue LED packaged with an electronic controller chip, you can independently control the brightness of each LED, producing the incessantly changing colours of toys and electronics which is the hallmark of the current decade.

There are two common ways to produce a "white" LED:
  • Combine a red, green and blue LED in a single package, with the intensity of each adjusted to produce a "white" light with 3 narrow bands. In principle, the colour could be adjustable.
  • Use a blue LED, with a phosphor to downconvert some of the high-energy blue photons into a diffuse spectrum of red and green photons, which produces a "white" light with a broader spectrum. These can't adjust the colour.
The human visual system has 3 colour sensing pigments, each cone cell sensitive in a slightly different (but overlapping) region of the spectrum. It does not matter if this retinal pigment is triggered by photons of a narrow spectrum or a broad spectrum, the result is the same: a nerve impulse is registered, which is then correlated with adjacent cells to detect edges, motion, and eventually fused into a 3D colour image in the brain.

The main health risks are documented in the international laser safety regulations:
  • High average power: This can burn your skin as well as your eyes.
  • Infra-red and ultraviolet light sources: These are not visible to the human eye, and so they do not invoke the protective blink reflex - they can damage your eye without you knowing it.
  • Extremely high peak power: Despite a low average power, these sources can damage your eye before you can blink (in 100ms) and well before your pupil can contract (this takes a few seconds).
  • Very small-area sources: These can concentrate damaging intensity on a small part of the retina, even though the total power is low. It is best to use a diffuser on the light source (this is more of a risk with a coherent source such as a laser than with a LED).
  • Apparently, infants under 9 months don't have a blink reflex, so extra care should be taken with them.
See http://en.wikipedia.org/wiki/Colour_vision#Physiology_of_color_perception
« Last Edit: 23/03/2013 05:27:37 by evan_au »
 

Offline evan_au

  • Neilep Level Member
  • ******
  • Posts: 4123
  • Thanked: 245 times
    • View Profile
The cost of the 3-colour LED is likely to be higher than the Blue+phosphor LED because the 3-colour LED needs:
  • Photons of 3 different colours, each have different energies (measured in electron-Volts), which must come from materials with a different band-gap
  • This requires mounting and wiring 3 different chips of 3 different semiconductor materials to produce the 3 different colours
  • http://en.wikipedia.org/wiki/LED#Colors_and_materials
  • A different bandgap means you want a different supply voltage to the chip (measured in Volts) to maintain efficiency
  • So the power-supply chip (usually made of silicon, a 4th semiconductor) must produce 3 separate output voltages
  • The efficiency of each LED material is different, so the output power must be balanced between the 3 colours.
It may be possible to simplify the design by putting all 3 LEDs in series, but it is likely that the Blue LED+phosphor design will be cheaper to manufacture.
 

Offline graham.d

  • Neilep Level Member
  • ******
  • Posts: 2208
    • View Profile
Evan, I think you answered the question I had with saying "It does not matter if this retinal pigment is triggered by photons of a narrow spectrum or a broad spectrum". I had not thought about it very much and was thinking that it was possible that the receptors might be a conglomerate of narrower band detectors, which is not the case.

However, I do get a little concerned about making leaps into widespread use of a technology without long term studies on the possible effects. It is one thing having nice and efficient torches and quite another incorporating LEDs into lighting which millions of people will be living with for many hours of the day. A brief scan of the web did not find very many such studies though this one on the effects on the retinas of mice was interesting...

http://www.lifesciencesite.com/lsj/life0901/072_8366life0901_477_482.pdf

Another study (not specifically about LEDs) looked at the effects of flicker (50 or 100Hz) which we are all fairly used to and to which the human brain is fairly immune. This particular study looked at the effects on other animals, particularly birds, who do not filter such flicker out. We tend to take technology advances for granted and when something is clearly advantageous there is huge pressure to make it ubiquitous. It is different from drug testing which, even if such testing fails to find and significant detrimental issues (which it often does) generally only affects a small percentage of a population. My old chemistry teacher (an eccentric gent who wanted to teach us lot despite being an FRS) expressed a similar view when it was decided to introduce fluorination of the water supplies. He may have been pessimistic about any detrimental effects but I think the principles behind his concerns were generally OK.

« Last Edit: 23/03/2013 11:04:45 by graham.d »
 

Offline JP

  • Neilep Level Member
  • ******
  • Posts: 3366
  • Thanked: 2 times
    • View Profile
From what I know of laser safety (I work a lot with visible lasers as well as LED sources), the main danger is burns.  Since the eye focuses light entering it, a laser that wouldn't burn you if it hit your skin can burn your retina as a result of focusing.  Lasers and LED sources both tend to be very narrow bandwidth (single frequency), but I've never seen any warnings about that--just about intensity.

There's a nice list on wikipedia of different types of damage that lasers can cause:
http://en.wikipedia.org/wiki/Laser_safety#Damage_mechanisms

Presumably LEDs as sufficient brightness could cause these issues as well, but it would take far more power, since LEDs are not "coherent" sources and can't project their light along a narrow beam.
 

Offline techmind

  • Hero Member
  • *****
  • Posts: 934
  • Un-obfuscated
    • View Profile
    • techmind.org
In general the spectral output from modern "tri-phosphor" fluorescent tubes (and compact fluorescents) has significantly narrower spectral emission-lines than any LED source - so you couldn't argue that LEDs would be worse than FL. That said, I can't see much scope for physcial damage to the eye from this kind of narrow-band spectral light from domestic fittings.

However, all narrowband sources (FL and to a lesser extent LEDs) when used for general lighting, affect "colour rendering" - the way object-colours appear. I think I might be justified in arguing that there is scope for some concern that for an infant developing substantially under such artificial light (with little exposure to natural light) runs some risk of aquiring some kind of subtle developmental disorder in the analysis and recognition of colours - particularly of subtle colour gradations.
 

Offline graham.d

  • Neilep Level Member
  • ******
  • Posts: 2208
    • View Profile
In fact I read a paper, whilst looking into this subject, that showed that the balance of numbers of the different colour receptors in the eye varies hugely from one person to another. Although this might be expected to lead to people's perception of colours to be different, tests on perception showed remarkable correlation. If subjected to a different hue of background light (or wearing tinted spectacles for long periods) the brain ultimately compensated to enable them, once again, to agree on particular colours. They also eventually reverted back to their normal perception when the light was normal again or the specatacles removed. The conclusion was that the brain compensates hugely from the stimuli it receives from the cone colour receptors either when the eye has many more or many less receptors of one type or because the external colour balance is changed. Maybe this would would indicate there is less chance of developmental issues in infants subjected to unnaturally coloured light sources.
 

The Naked Scientists Forum


 

SMF 2.0.10 | SMF © 2015, Simple Machines
SMFAds for Free Forums