Effets sur la santé de la lumière artificielle

2. How do artificial lights work?

The SCENIHR opinion states:

3. SCIENTIFIC RATIONALE

3.1. Introduction and scope

Ever since man started to consciously use fire and thus light for improving vision during dark parts of the day, humans have strived to improve on the quality of light sources. With the advent of electricity it was possible to develop the technology that has been used for incandescent light bulbs. The present day society displays a plethora of further developments in light technology (see Annex I – Technical Characteristics of Lighting Technologies) where energy aspects as well as ergonomic and other considerations have their place. It is well established that humans and other biological entities are sensitive to light to various degrees, and that normal physiological processes can be, and are, influenced by light from natural or artificial sources. Typical for the modern society is that the all-encompassing use of artificial light sources disturbs the normal conditions of light at day and darkness at night. This has the potential to disturb circadian rhythms, but to what extent, if any, the common light sources also have negative effects on human health is unclear.

The impact of one type of modern light technology, compact fluorescent lamps (CFLs), on human health issues was covered by an earlier SCENIHR opinion (SCENIHR 2008). The main conclusions from that opinion were that there were no direct scientific data on the relationship between this specific form of energy saving light bulb and a number of symptoms in patients with various conditions. For some of these conditions (epilepsy, migraine, and retinal diseases), it was identified that either flicker and/or UV/blue light could exacerbate the effects. The evidence in regard to the skin conditions chronic actinic dermatitis and solar urticaria was found to be related to UV/blue light emissions only. However, at the time of writing the report, there was very little reliable evidence that emission from fluorescent tubes was a significant contributor. Furthermore, it was noted that certain CFLs can under specific conditions emit UVB and traces of UVC radiation. In addition, CFLs emit a higher proportion of blue light than incandescent lamps. Both these types of emissions can be risk factors for the aggravation of symptoms in some patients suffering from chronic actinic dermatitis and solar urticaria. The purpose of the scientific rationale is to take into account relevant scientific data from the fields of physics, engineering, biology, and medicine, and assess whether optical irradiation from all types of common light sources can cause disease conditions or aggravate already existing conditions.

The use of the light sources discussed in this opinion may also expose the general public to risks not originating from the optical radiation. These potential risks (e.g. fire hazards, cuts, heat, electric shocks, electromagnetic fields, mercury, etc.) are either well known and/or discussed in other opinions from the Scientific Committees (SCENIHR 2008, SCHER 2010). Furthermore, recent media reports on emissions of certain chemicals from energy saving lamps are not discussed in this opinion, but will be dealt with by other public bodies. Also exposure to light from specialized technologies such as operating lamps is not included in this opinion. Thus, the focus of this opinion is on possible effects from optical radiation emanating from artificial light sources.

3.3. Physical characteristics of artificial light sources

3.3.1. Physical principles

Light is electromagnetic (EM) radiation in the range from 400 to 780 nm (1 nm is 10-9 m) that is visible to the intact adult human eye (see also CEI/IEC 62471/2006). Light, like EM radiation in general, is emitted by the transition of quantum states if excess energy is to be released in this specific wavelength range. Light sources experienced in nature include different physical phenomena involving atomic/electronic de-excitation processes induced e.g. by heat, inelastic collisions and nuclear reactions. Examples include: (1) the glowing appearance of fires, flames and other sources e.g. volcanic hot material, where thermal radiation is released; (2) the photochemical light generation of animals such as the glow-worm; (3) the Nordic light (aurora borealis) when showers of elementary particles are trapped by the earth’s magnetic field and hit the outer atmosphere; (4) the bright sensation of the electric discharge through the air in lightning, and last but not least (5) the light emitted by the sun, which emerges from the hot plasma induced by hydrogen to helium fusion.

The light which is incident to a certain user or observer is not only dependent on the initial light emission characteristics in the light source, but also on the typical and often frequency dependent light absorption properties of the medium between the light emission and the observer. This is normally provided by the medium in which light is generated as well as its envelope or surrounding (air). A further aspect, which is important to consider, is the geometric arrangement of the source and the user/observer, possibly wearing eyewear, as well as the geometry and reflective properties of the room and/or the luminaire.

Lighting by flames (e.g. candles and oil lamps) was historically the predominant source of light, until electrical heating of filaments (carbon and then tungsten) came to dominate the field. Common to all these lighting applications is that matter is heated to a suitable temperature in order to emit thermal broadband radiation. Numerous thermal excitation and de-excitation processes occur, and are involved in, the generation of light, which leads to a characteristic “bell shaped” spectrum governed by Planck’s law of radiation. This law predicts that with increasing temperature of the irradiating material, the peak intensity of the irradiated electromagnetic spectrum is observed at higher characteristic frequencies. This implies that at about 5,000 K the emitted spectrum is similar to that of the sun’s radiation through clear skies at midday. Therefore each lamp (and each spectrum) can be associated with a “colour temperature”, which describes the sensation of this light on the human eye and on a photographic film, and also affects the colour perception. Also characteristic to the emission from these sources is an increasing fraction of blue light and ultraviolet radiation with a higher operating temperature.

3.3.2. Artificial light technologies

For centuries, mankind has essentially used burning or heated materials as light sources (incandescence). However, it was well known that light could also be generated without heating (luminescence). Thus, bio-luminescence (fire-flies, glow-worms, glowing mushrooms, etc.), phosphorescent minerals, as well as lightning were observed by prehistoric human beings. Today, flame-operated lamps (mainly kerosene, carbide and gas lamps) and candles are still in use. Such lamps use a chemical reaction to heat material (soot particles in oil lamps, carbide lamps, and barium oxide particles on a glow- body for gas lamps). The emitted spectrum is continuous and further characterized by a correlated colour temperature, which is often low due to the limited temperature of the irradiating component and by a poor luminous efficacy. Beyond flame-operated lamps, which are still used in everyday life by approximately 1.6 billion people who do not have access to an electrical grid, the major part of the world population uses electrical-powered lamps for producing artificial light. In 2005, 3,418 TWh of electricity, which represents roughly 19% of world electricity production, was used for producing 133 Plmh (peta-lumen-hours) of artificial light (Brown 2009, Waide and Tanishima 2006). According to the same authors, on average 43% of this electricity is used for illuminating tertiary-use buildings, 31% for residential lighting, 18% for industrial buildings, and finally 8% for outdoor stationary illumination and signalling. As shown in Figure 1, two technologies are mainly in use today: Incandescent and Luminescent Lamps. The latter category can be further divided into Discharge/Fluorescence Lamps and Solid State Lighting Devices, respectively.

Figure 1 Electrical lighting sources technologies

A thorough overview of different lamp types is presented in Annex I, which includes descriptions of the fundamental technologies and their areas of use as well as some examples of emission spectra when such are available. It should however be noted that there is such a diversity of products among each lighting technology available on the market, that it is, in many cases, very difficult to present emission spectra which are “typical” for a given lamp type.

In the case of some lamp technologies, a second bulb or glass envelope is sometimes present. Most glass types absorb a large fraction of the UV radiation, and the UV transmittance depends on the thickness of the glass. UVB and UVC, as well as the shortest UVA wavelengths, do not penetrate ordinary glass. Whereas pure glass SiO2 does not absorb UV light, soda-lime glass does not allow light at a wavelength lower than 400 nm (UV) to pass. Even Pyrex and other more ordinary forms of heat-resistant glass can be used as shields to block UVB and UVC. Using additional UV-blocking dopants can enhance this blocking behaviour. High efficacy incandescent lamps and some ceramic metal halide lamps often use a second bulb made of soda-lime glass. Compact fluorescent lamps are also made of borosilicate glass and a specific type called “look- alike” has a second shaped envelope in order to mimic the appearance of a classic incandescent lamp. This external envelope is usually made of polycarbonate. Most plastic materials (acrylic, polycarbonate, plexiglass) used for lamp bulbs give more protection than soft glass. Polycarbonate is almost completely transparent throughout the entire visible region until 400 nm, and if intact, successfully blocks UV radiation (UVB, UVC and more than 90% of UVA). Linear fluorescent lamps are also made of soda-lime glass, but it is rather unusual for these lamps to have a second protective envelope. However, this type of lamp can be placed in luminaires that have polycarbonate protection. In addition, it is possible to use a specific filter (GAM 1510 UV shield) that exists in the form of a gel or Rosco film (03114); this can be used to envelop the lamp and eliminates more than 95% of UVA radiation. Last but not least, when a lamp is placed in a luminaire or fixture, UV blocking elements (such as soda-lime glass or polycarbonate layers) may be introduced into the system and drastically reduce UV output. However, as this type of situation is optional rather than a rule, it is suggested to evaluate UV risks for bare lamps.

3.3.3. Lamp emissions

A critical aspect of any risk assessment of the potential health effects of lighting technologies is the availability of exposure data for the general population as well as occupational exposure. Unfortunately, data regarding actual exposure are sparse, which stresses the need for reliable data regarding emissions from the various lamp types. During the writing of this opinion, a call for information was launched regarding inter alia emission data (see section 3.2 for further details regarding the call for information). Relevant information was obtained, based on measurements performed by, or requested from, different stakeholders. Two contributions provided substantial information which could be used in this opinion (spectral UV emissions from the Belgian Federal Public Service of Health with a focus on CFLs and more global and graphical information from the European Lamp Companies Federation). In addition a recent study from Schulmeister et al. (2011) provided valuable relevant detailed spectral information. Measurement data from all three sources are according to the measurement methodologies recommended by Standard EN 62471. Furthermore, detailed emission spectra (with nm resolution) were only provided in the study from Schulmeister and co-workers (2011). There is thus only scattered knowledge regarding the full emission spectrum from all available lamp types. It is not possible at present to perform a much needed comparative assessment of the different lamp types. Based on emissions from the lamp, the Standard EN 62471 (and also IEC 62471 and CIE S009, since they are all identical in this sense) categorizes the lamps according to the photo-biological hazard that they might pose. The different hazards are: 1. Actinic UV-hazard for eye and skin (see section 3.4.3.2); 2. UVA-hazard for the eye (section 3.4.3.2); 3. Blue-light hazard for the retina (section 3.5.2.3); 4. Thermal retina hazard (section 3.4.3.1) and 5. IR-hazard for the eye (sections 3.4.3.1 and 3.4.3.2).

According to the standards, measurements should be performed according to two approaches; viz. at a distance where a light intensity of 500 lx is obtained and also at a distance of 20 cm (see also section 3.4.2.2 for additional discussion regarding measurement). Based on these measurements, lamps are then classified according to the “Risk Group” (RG) to which they belong. RG0 (exempt from risk) and RG1 (minor risk) lamps do not pose any hazards during normal circumstances. RG2 (medium risk) lamps also do not pose hazards because of our aversion responses to very bright light sources, or due to the fact that we would experience thermal discomfort. RG3 (high risk) include only lamps where a short-term exposure poses a hazard. This classification is based on acute exposure responses (a single day, up to 8 hours) and applies only to individuals of normal sensitivity.

The material received from the Belgian Federal Public Service of Health included measurements on 70 CFLs, and also Ecodesign UV functional requirements (Ecodesign regulation 299/2009). Four lamps were classified as RG1 for actinic UV-hazard at 20 cm, whereas lamps were otherwise classified as RG0. However, UVC emissions could not be determined according to Ecodesign, since background levels were higher than the Ecodesign requirements.

The contribution from the European Lamp Companies Federation (ELC) included six lamp types from eight manufacturers, considered by ELC to be “representative lamp types”. Risk group classification was carried out in accordance with EN 62471. Results were presented for the following lamp types:

• Tubular fluorescent (4,000 K and 6,000 K);
• CFL (2,700 K 11W with and without envelope);
• LED (3,000 K, retro-fit, and 6,000 K);
• Halogen (two high voltage, one without UV filter, and three low voltage);
• High pressure discharge (two metal halide and one sodium); and
• Incandescent (60 W clear).

A summary of important parameters for each lamp is shown in Table 1.

Table 1 Lamp parameters supplied by the European Lamp Companies Federation

According to ELC, under normal conditions of exposure, all lamps are classified as RG0 (exempt from risk) or RG1 (low risk) from UV and IR emissions, with the exception of one lamp. This halogen lamp is intended to be used with additional glass shielding, which was tested without the glass shield, and was characterized as RG2-RG3 at 20 cm. The metal halide lamps are RG1 or RG2 at 20 cm, but these are not intended for use at such close distance according to ELC.

Concerning blue light emission, ELC considered all lamps as belonging to RG0 or RG1. This includes the 6,000 K LED (“high power” LED) which is RG0 when analysed as a “small source”. The use of the “small source” approximation is valid because the eye moves rapidly without our knowing it ??).. This means that the image of the source is smeared over a larger area of the retina than the area of the image itself and the light emitted is averaged over an angle of 11 mrad, which is the effective angular subtense taking account of eye movement. This means that treating the LED as a “small source” and averaging over an effective angular subtense of 11 mrad is acceptable in accordance with EN 62471. It follows that the 6,000 K LED in the ELC dataset is correctly classified as RG0.

The metal halide lamps are also RG2 when measured at 20 cm, but these lamps are not intended to be used at such close distance according to ELC. ELC reported that the provided data were measured in an accredited laboratory according to ISO/IEC 17025 and so the measurement procedure should be reliable and the results reproducible. Furthermore, it is stated that the lamps were selected such that they are typical, mid-range samples from the quality control process.

The results presented in the ELC report suggest to SCENIHR that there is little or no risk to individuals of normal sensitivity from the UV, IR or blue light optical radiation emission from lamps which are considered to be “representative” of the type of lamps selected to replace incandescent lamps. SCENIHR however considers that “non-representative” lamps may emit levels that are much higher than those included in the report; however quality control limits applied by lamp manufacturers were not reported. Further consideration should also be given to the “intended” vs. “reasonable foreseeable” use of lamps. For risk assessment purposes, most light sources should be assessed at the distance corresponding to 500 lux illumination. It is inappropriate to classify high output lamps at a distance of 20 cm when they are designed to illuminate a large area, e.g. a factory. Only those lamps that are intended to be used in close proximity to the skin should be assessed at 20 cm.

Further consideration also needs to be given to the risk classification of high power LEDs. Also, halogen lamps that are intended to be used with an external glass filter must not be used without the filter because of the risk of exposure to UV radiation. Schulmeister et al. (2011) measured UV emission characteristics, as far as we can judge according to EN 62471 with a nm resolution, in 96 different types of light sources (including CFLs, LEDs, halogen lamps, fluorescent tubes, high-pressure discharge lamps, and incandescent lamps). One high pressure mercury lamp intended for industrial lighting was classified as RG1 (actinic UV) at 500 lx, whereas some of the high-pressure discharge lamps were assigned to higher RGs at 20 cm. These lamps are however not intended for use at such close distances.

Again, SCENIHR considers that “intended” vs. “reasonably foreseeable” use should be considered for the lamps classified as higher RGs at 20 cm. It is important to know whether the risk categories designed to protect the general public provide adequate protection to photosensitive patients. In a preliminary study, it has been shown that single envelope CFLs may cause an erythematous reaction in patients with a photosensitive disorder (Eadie et al. 2009). The published report does not contain data on the risk classification of the lamp. Subsequent analysis of the lamp used in that investigation (Moseley, personal communication) shows that it is RG1 at 20 cm. In the study reported by Eadie et al. (2009), the lamp was used at a distance of 5 cm because it was argued that in practice this was quite reasonable for task lighting, particularly since there is very little heat emitted. At this distance the lamp would be RG2. Since lamps which are intended to be used in close proximity to the skin are classified at a distance of 20 cm, it is clear that a single envelope CFL classified as RG1 may be hazardous to a photosensitive patient if used closer than 20 cm to the skin. All of the CFL lamps included in the ELC report were RG0 or RG1. It is difficult to predict how an individual patient will respond to light from a particular lamp because of the range of response that individual patients exhibit when exposed to different wavelengths. However, RG1 lamps cannot be considered safe for use by photosensitive patients. Further work into optical radiation emission from 167 CFLs (103 single envelope and 65 double envelope) demonstrated double envelope lamps generally emitted much less than single envelope lamps (Moseley, personal communication). Taking the highest emitting lamp of each model tested, the mean UVB irradiance was 4 mW/m2 (double envelope) and 101 mW/m2 (single envelope).

Schulmeister et al. (2011) also reported on UV emission levels from halogen lamps. These lamps have a smoothly decreasing spectrum at UV wavelengths. Although there are no published data on the effect of exposure of a photosensitive patient to light from a halogen lamp, it is unlikely that there would be a significant risk provided the protective filter was in place. However, it should be noted that some halogen lamps may be used without the filter attached which would increase the chance of an adverse reaction.

Conclusions

Optical radiation emission data from three different laboratories (representing public authority, industry, and a commercial research enterprise) have been obtained and considered for the conclusions in this opinion. Data from more than 180 different lamps were provided and represent all major lamp types that are used for general lighting purposes. Regarding specific lamp types, CFLs are well represented in the samples assessed, whereas LEDs are measured in only a few cases. All other lamp types are represented mostly in small numbers.

The photobiological hazard from each lamp has been determined according to Standard EN 62471. For all investigated hazard outcomes, the absolute majority of lamps are classified as RG0 (exempt from risk). Most other lamps are classified as RG1 (low risk). The lamps assigned to higher risk groups were either measured without a UV-shielding glass cover, or at a short distance (20 cm) which is not the intended use distance for this lamp type.

SCENIHR considers that further consideration needs to be given to the representativeness of the measured lamps and to the question of whether the intended use can be ensured for those lamps classified as RG2 or RG3 at a distance of 20 cm. LEDs were under-represented in the present analysis of lamps. Further assessment of LED retinal hazards should be evaluated at 20 cm taking into account that LED luminaires can be used at this distance for domestic lighting.

Source & ©: , Health effects of artificial light, 19 March 2012,
 3.1. Introduction and scope, 3.2 Methodology, and 3.3 Physical characteristics of artificial light sources, pp. 15-22.