Head of Lighting and Electromagnetism Division, Department of Health and Comfort Centre Scientifique et Technique du Bâtiment - CSTB Grenoble, France
About us


This article has been read 208 times
Share this content

Refer this article as: Martinsons, C., Light Emitting Diodes (LEDs) and the blue light risk, Points de Vue, International Review of Ophthalmic Optics, N68, Spring, 2013

Light Emitting Diodes (LEDs) and the blue light risk

Online publication :
Reading time :
12 min


Traditional lighting sources such as the well-known incandescent lamp and the compact fluorescent lamp are rapidly being replaced by products based on light emitting diodes (LED) (fig. 1).

Fig. 1: Photographs of several types of solid-state lighting products.
a: Directional luminaire (spot light) using an LED.
b: SSL lamp based on three LEDs and used to replace an incandescent lamp.
c: Outdoor high power SSL luminaire using 121 LED modules.
d: Typical single LED component, used in many SSL products. This type of LED consumes about 1 W of electricity and generates a luminous flux of about 100 lm. Its luminance can be as high a 10 7 cd/m².

The so-called “solid-state lighting” (SSL) presents many advantages such as longer lifetime, reduced energy consumption and lower environmental impact. Many governments have therefore started to progressively ban older lighting technologies, paving the way for the massive usage of LEDs in the general lighting market. As a matter of fact, leaders of the lighting industry believe that over 90% of all lighting sources in the world will be based on SSL products and LEDs by 2020. As any new and emerging technologies, SSL products should be proven to be at least as safe as the products they intend to replace. Furthermore, some unique properties of LEDs such as their compactness have generated many new lighting applications for which older technologies could not be employed. For instance, some kinds of toys and clothes now incorporate LEDs. The safety of products using LEDs should be assessed considering the interactions with the human body in existing and new ways of using them.

The potential adverse effects of optical radiation on the skin and on the eyes are known as photobiological hazards. LEDs currently used in lighting applications have the advantage of emitting a negligible amount of ultraviolet (UV) and infrared (IR) radiation 1. The only photobiological hazards to consider when assessing the safety of LEDs are linked to visible light, and more particularly the blue part of the spectrum.

Several health agencies such as ANSES 2 and SCENIHR 3 have investigated and reviewed the scientific literature on photobiological hazards related to the use of LEDs. Two key features of LEDs have drawn the attention of experts:

  •  LEDs are very bright small sources of visible light, which can be glaring. Due to their high brightness, LEDs also have very high radiance (a photometric quantity expressing the “concentration” of light), which in turn produces a high illuminance level upon the retina.
  • The vast majority of white LEDs producing white light rely on a chip emitting blue light associated with layers of fluorescent materials (luminophores) to produce longer wavelengths. As a consequence, the emission spectrum of a white LED consists in a narrow primary blue peak and a large secondary peak in the yellow-orange-red part of the spectrum. The two peaks are separated by a region of very low emission in the blue-green part of the spectrum (fig. 2).

Fig. 2: The blue curve represents the typical emission spectrum of a white LED. The blue peak reaches its maximum value at about 435nm. It corresponds to the primary light generated by the LED semiconducting structure itself (the LED die). The secondary peak reaches a maximum value at 550nm (yellow colour) and is the secondary light emitted by luminophores excited by the blue light (fluorescence). The combination of the direct blue light and the yellow/red secondary light produces white color. The red curve is a plot of the blue light retinal phototoxicity function. It reaches a maximum value at wavelengths corresponding to the blue light peak emitted by LEDs.


Visible light on the retina can cause thermal damage and photochemical damage. The exposure levels needed to result in thermal damage on the retina cannot be met with light emitted by LEDs of current technologies. The photochemical risk is associated with blue light retinal illuminance. Due to the high brightness of LEDs, the retinal illuminance levels are potentially high and must be carefully considered. In general, the photochemical damage of the retina depends on the accumulated dose to which the person has been exposed, which can be the result of a high intensity short exposure but can also appear after low intensity exposures repeated over long periods. Blue light is recognised as being harmful to the retina, as a result of cellular oxidative stress. 
Blue light is also suspected to be a risk factor in age-related macular degeneration (ARMD).

Retinal blue light exposure can be estimated using the ICNIRP 4 guidelines. A quantity called the blue-light weighted radiance LB can be estimated as a function of the viewing distance and the exposure time. Maximum permissible exposure values (MPEs) were set by ICNIRP to provide limits for LB as a function of exposure time. 
For the past three years, blue light exposure data about LEDs have been provided by LED manufacturers and professional lighting associations but also by independent laboratories and governmental agencies.
It was found that the retinal blue light exposure levels LB produced at a distance of 200mm from the user by blue and cold-white LEDs (bare LEDs and LEDs equipped with a focusing lens) exceed the MPE limits set by ICNIRP after an exposure time comprised between a few seconds for high power blue LEDs to a few tens of seconds for high power cold-white LEDs. As a consequence, the potential toxicity of some LED components viewed at short distances cannot be neglected. However, when the viewing distance is increased to one metre, the maximum permissible exposure time rapidly increases to a few thousands of seconds, up to a few tens of thousands of seconds. These very long exposure times provide a reasonable safety margin to assert that there is virtually no possible blue light retinal damage caused by LEDs at longer viewing distances (statement valid for state of the art LEDs at the time of writing).

Several classes of products and applications based on bare LEDs or LEDs covered by a focusing lens (collimator) are directly related to a potentially high level of retinal blue light exposure when short viewing distances are possible. Examples are (but are not limited to):

  • Tests and adjustments of high power blue and cold white LEDs by operators in lighting manufacturing facilities or by lighting installers
  • Toys using LEDs, given that the higher degree of transparency of the crystalline lens of children makes them more susceptible to higher blue light retinal exposures
  • Automotive LED daytime running lights when activated near children and other sensitive subjects
  • Some types of directional LED lamps sold for home applications. These lamps can be viewed from distances as short as 200m

The conclusions drawn for single LED components or LED modules cannot be extended to all SSL applications because the photobiological safety of a final SSL product must be assessed independently of its LED components. As a matter of fact, the LB value of an SSL product is generally very different from the LB value of the LED components that it uses. For instance, a higher LB can be obtained with a lamp using an assembly of low LB LEDs. Reversely, a lower LB can be obtained with a lamp using a diffuser in front of a high LB LED. For all LEDs and products using LEDs, a photobiological blue light risk assessment must be carried out to determine whether or not the MPEs can be exceeded in the conditions of usage. Such risk assessments can be performed by test laboratories specialised in light sources photometry such as CSTB 5 and LNE 6 in France.

The main tool used to perform photobiological risk assessment is the CIE 7 S009 publication whose content was included in an international standard (IEC 62471) and other national standards (IESNA RP27, JIS C8159, etc.).


This standard deals with the photobiological safety of lamps and devices using lamps and includes a classification of the light source in several risk groups. The standard considers all of the photobiological hazards that may affect the skin and the eye (thermal and photochemical hazards) from ultraviolet to infrared wavelengths. Four risk groups are defined: Risk Group 0 (RG0, no risk), Risk Group 1 (RG1, low risk), Risk Group 2 (RG2, moderate risk), Risk Group 3 (RG3, high risk). The risk group depends on the maximum permissible exposure time (MPE time) assessed at a given viewing distance.


IEC 62471 defines two different criteria to determine the viewing distance. Light sources used in general lighting should be assessed at a distance corresponding to an illuminance of 500 lx. Other types of light sources should be assessed at a fixed distance of 200mm. For LED components, there is no ambiguity in the distance since LED components are not used per se in general lighting. In this case, IEC 62471 requires using the distance of 200mm. The application of the IEC 62471 measurement technique at 200mm leads to RG2 classification (moderate risk) for some high power blue and cold white LEDs.

However, the choice of the viewing distance in IEC 62471 is sometimes ambiguous and not realistic in the context of the real usage conditions. For instance, in the case of stage lighting (theatres, concert halls) where artists are exposed to an illuminance level higher than 500 lx. Applying the 500 lx criterion would underestimate the exposure while the 200mm criterion would largely overestimate it. In a more usual situation, directional household lamps fall under the 500 lx criterion, which corresponds to a typical viewing distance of a few metres. It is however quite common to have shorter viewing distances, as short as 200 or 500mm at home. Another example is street lighting where the illuminance level is much lower than 500 lx, typically a few tens of lx. Assessing the exposure to blue light emitted by a street lighting luminaire at the distance giving an illuminance of 500 lx is clearly not appropriate. A future revision of IEC 62471 should bring a more accurate definition of the distance at which the risk group is determined.

It is interesting to note that the strict application of CIE S009 and IEC 62471 to indoor LED lamps and luminaires lead to RG0 and RG1 classifications, similar to traditional indoor light sources (fluorescent lamps, incandescent and halogen lamps). Nevertheless, when the 200mm viewing distance is chosen, several measurement campaigns reveal that a small number of indoor LED lamps and luminaires belonged to RG2 while traditional indoor light sources (fluorescent and incandescent) were still in RG0 or RG1. This result shows that LED technology potentially raises the blue light risk in home applications where the viewing distance is not limited and light sources are accessible to children and other sensitive people. At the time of publication, the general public remains unaware of potential risks to the eye since no mandatory labeling system is currently in place for consumer SSL products.

The notion of a safety distance would actually be more appropriate to communicate to installers and to users, especially the general public. The safety distance of an SSL product would be the minimum distance for which the blue light hazard risk group does not exceed RG1. Measurement campaigns carried out by several laboratories showed that the vast majority of indoor LED lamps and luminaires have a safety distance of 200mm which is compatible with most lighting applications. 

It is important to note that other widely used lighting sources, particularly high intensity discharge lamps used for outdoor lighting are in RG2 (moderate risk). However, these lamps are intended for clearly identified uses and can only be installed by professionals who should be aware of the safety distance required to limit the exposure.


The maximum exposure limits defined by the ICNIRP and used to define the Risk Groups in both IEC 62471 and CIE S009 are not appropriate for repeated exposures to blue light as they were calculated for a maximum exposure in one 8-hour day. They do not take into account the possibility of exposure over an entire lifetime. Neither CIE S009 nor IEC 62471 takes into account the sensitivity of certain specific population groups, which can be characterised by an accrued sensitivity to visible light:

  • People having pre-existing eye or skin conditions for which artificial lighting can trigger or aggravate pathological symptoms
  • Aphakic (people with no crystalline lens) and pseudophakic people (with artificial crystalline lenses) who consequently either cannot or can only insufficiently filter short wavelengths (particularly blue light)
  • Children
  • Elderly people as their eyes are more sensitive to optical radiation

The photobiological standards for lighting systems should be extended to cover children and aphakic or pseudophakic individuals, taking into account the corresponding phototoxicity curve published by the ICNIRP in its guidelines.

In addition to proven photochemical damage of the retina resulting from acute exposure to blue light, uncertainty still remains surrounding the effects of chronic exposure at low doses. These effects are still being investigated by ophthalmologists, biologists and optical scientists.
In France, the RETINALED project 8 is investigating the effects of chronic low exposure of rodents to light emitted by LEDs.

Certain categories of workers are exposed to high doses of artificial light (long exposure times and/or high retinal illuminances) during their daily activities (examples: lighting professionals, stage artists, etc.). Since the damage mechanisms are not fully understood yet, exposed workers should use appropriate individual means of protection as a precautionary measure (glasses filtering out blue light for instance).


Due to their unique light emission properties, LEDs are currently on the verge of becoming the dominant lighting source of this century. However, the risks posed by these new sources of light are also rooted in their intrinsic characteristics: high optical output in a small package (producing a high radiance level) associated with a significant blue light emission. The combination of these two factors can potentially increase the risk of photochemical damage of the retina, in comparison with the incandescent lamp and the fluorescent lamp.

Lighting industry leaders are well aware of the photobiological safety of their products. Many lighting products using LEDs now emit warmer shades of white light (reduction of the blue light content in the spectrum) or use diffusers to reduce glare (reduction of the radiance). Most lighting products are found to present low risks or no risk at all for the general population when the viewing distance is equal to or greater than 200mm. 

However, measurement campaigns carried out by independent agencies pointed out a few lighting products with significantly higher risk levels below a distance of one metre or more. At the present time, no mention is made by lighting manufacturers of a “safety distance”. It is therefore impossible for the public to identify lamps or luminaires with a higher risk level.

The blue light risk assessment related to LEDs can be performed by test laboratories using the IEC 62471 standard which is not perfectly clear about the viewing distance to consider. In addition, this standard does not consider sensitive populations such as children, aphakic, pseudophakic and elderly people, despite the fact that these populations are exposed to a higher level of blue light on the retina. The current knowledge of the mechanisms of blue light phototoxicity is far from being complete. The effects of chronic exposure and accumulated low exposure over very long periods of time are still an active subject of research. As far as LEDs are concerned, the better comprehension of the possible long term effects of the blue light on the retina is fundamental to guaranteeing that the “LED revolution” will not compromise our vision of the future. 


Footnote page:
1. As they emit negligible amounts of UV and IR, LEDs should not be expected to contribute to the onset of photokeratitis and cataract.
2. Agence nationale de sécurité sanitaire de l’alimentation, de l’environnement et du travail (French National Agency for Food, Environmental and Work Safety).
3. Scientific Committee on Emerging and Newly Identified Health Risks.
4. International Commission for Non-Ionising Radiation Protection.
5. Centre Scientifique et Technique du Bâtiment (French Technical and Scientific Research Center on Construction and Buidling).
6. Laboratoire National de Métrologie et d’Essais (National Testing and Metrology Laboratory).
7. Commission Internationale de l’Eclairage (International Commission on Illumination).
8. The RETINALED project is carried out by INSERM, CSTB and ENVA. It is supported by ADEME (French Environmental and Energy Management Agency).



This article presents an overview of the knowledge regarding the potential toxicity of light emitted by LEDs on the retina. Due to their high brightness and their emission spectrum containing a significant fraction of blue light, the so-called “blue light hazard” has been considered and studied for several years. Several independent studies carried out by health agencies have shown that the risk posed by LEDs used as general lighting sources is low, but cannot be neglected in the case of some sensitive populations, considering the increasing optical performances of LEDs and their fast mass market distribution.

Head of Lighting and Electromagnetism Division, Department of Health and Comfort Centre Scientifique et Technique du Bâtiment - CSTB Grenoble, France
About us


This article has been read 208 times
Share this content

Refer this article as: Martinsons, C., Light Emitting Diodes (LEDs) and the blue light risk, Points de Vue, International Review of Ophthalmic Optics, N68, Spring, 2013

Continue reading