Blue light, sleep and eye safety
What LED lighting and screens do to the body clock and the retina, and what the evidence shows
Blue light is visible light of roughly 400–500 nm. It affects health in two separate ways. Through melanopsin cells in the retina, most sensitive near 490 nm, it shifts the body clock and suppresses melatonin. At very high intensities it can also cause the blue-light hazard, photochemical damage to the retina that peaks near 435–440 nm.
White LEDs make their light with a blue chip (see how LEDs make white light), and screens rely on blue emitters too, so blue light has become a common worry. The research separates two effects. Light in the evening, especially blue-rich light, delays the body clock and suppresses melatonin at the light levels found in ordinary homes. Damage to the retina needs light far brighter than room lighting, such as staring at the sun. This guide sets out what the studies measured, what the standards say and what the numbers mean in practice.
Two different effects: the body clock and the blue-light hazard
Blue light acts on the eye in two separate ways. Melanopsin cells, most sensitive near 490 nm, signal the body clock and suppress melatonin at ordinary indoor light levels. The blue-light hazard is photochemical damage to the retina, peaking near 435–440 nm, which needs long staring at very bright sources such as the sun.
In 2002 two research groups showed that a small set of retinal ganglion cells, the intrinsically photosensitive retinal ganglion cells (ipRGCs), detect light themselves using the pigment melanopsin and pass the signal to the brain's master clock in the suprachiasmatic nucleus. A year earlier, action spectra for melatonin suppression had already pointed to a new photoreceptor most sensitive in the blue, well away from the 555 nm peak of the cone vision on which the lux is based. In 2003 Lockley and colleagues found that 460 nm light delayed the clock twice as much as 555 nm light of equal photon density and concluded that photopic lux is "inappropriate" for measuring the light that resets it. CIE S 026 (2018) now defines the melanopic sensitivity curve, which peaks at 490 nm, and the quantity melanopic EDI: the illuminance of daylight (CIE D65) that would stimulate melanopsin as much as the light being measured.
The blue-light hazard is a different mechanism with a different action spectrum. Its weighting function B(λ), set by ICNIRP and used in IEC 62471, peaks at 435–440 nm and falls to a tenth of its peak by 500 nm. The damage is photochemical rather than thermal, so it builds up with the total dose, radiance multiplied by time. That is why it concerns staring into the sun, a welding arc or a bare high-power emitter, not lighting a room. The CIE has asked that the term "should not be used when referring to circadian rhythm disruption or sleep disturbance".
What the studies show: light, melatonin and sleep
Controlled laboratory studies agree that evening light delays the body clock and suppresses melatonin, and that blue-rich light does more per lux. Ordinary room light did so in a 2011 study, and five evenings of reading on an iPad delayed melatonin onset by about 1.5 hours in 2015. Total sleep time changed little.
| Study | What was tested | Main findings |
|---|---|---|
| Lockley et al. 2003 | 6.5 h of 460 nm (blue) or 555 nm (green) light at equal photon density | Blue light delayed the body clock twice as much and caused twice the melatonin suppression |
| Gooley et al. 2011 | 116 adults aged 18–30 in room light (under 200 lux) or dim light (under 3 lux) for the 8 h before bedtime | Room light delayed melatonin onset in 99% of people and shortened melatonin duration by about 90 min |
| Chang et al. 2015 | 12 young adults reading for about 4 h on 5 evenings, on an iPad (about 32 lux at the eye) or a printed book (about 1 lux) | iPad: melatonin onset more than 1.5 h later, 10 min longer to fall asleep, 12 min less REM sleep; total sleep time unchanged |
| Phillips et al. 2019 | 55 adults aged 18–30, 5 h of evening light from 10 to 2,000 lux | Half-maximal melatonin suppression at about 25 lux on average, from 6 to 350 lux between individuals; 10, 30 and 50 lux delayed melatonin onset by 22, 77 and 109 min |
| Duraccio et al. 2021 | 167 adults aged 18–24, 1 h of iPhone use before bed for 7 nights with Night Shift on, off, or no phone | No difference in sleep between the three groups; among those sleeping more than about 7 h, no phone gave the best sleep |
These are controlled studies of small groups of healthy young adults, usually against a dim background so that the effect of the test light stands out. They establish the mechanism and how sensitive it is: ordinary evening light levels move the clock, and people differ in their sensitivity by more than an order of magnitude. They say much less about long-term health. The European Commission's scientific committee SCHEER concluded in 2018 that there is "some evidence" that light in the late evening, including from LEDs and screens, may affect the circadian rhythm, but that it is "not yet clear" whether this leads to adverse health effects.
Warm and cool white compared
At the same lux, cool white light stimulates melanopsin about twice as strongly as warm white. From CIE data, a 2700 K LED gives 0.41 melanopic EDI per lux, a 4000 K LED 0.63 and a 6500 K LED 0.85; an incandescent lamp gives 0.50. Halving the brightness halves the melanopic EDI.
| Light source | CCT | Melanopic EDI per lux | Lux for 10 melanopic EDI | Blue-light hazard per lux |
|---|---|---|---|---|
| Warm-white LED (LED-B1) | 2733 K | 0.41 | 25 | 0.27 |
| Incandescent (illuminant A) | 2856 K | 0.50 | 20 | 0.31 |
| Warm-white LED (LED-B2) | 2998 K | 0.46 | 22 | 0.32 |
| Neutral-white LED (LED-B3) | 4103 K | 0.63 | 16 | 0.55 |
| Cool-white LED (LED-B4) | 5109 K | 0.69 | 14 | 0.70 |
| Cool-white LED (LED-B5) | 6598 K | 0.85 | 12 | 0.90 |
| Daylight (D65) | 6504 K | 1.00 | 10 | 1.00 |
Melanopic EDI per lux, the melanopic daylight efficacy ratio (melanopic DER), depends only on the spectrum and is 1 for daylight by definition; the blue-light hazard per lux is shown relative to daylight in the same way. LEDlights.org calculated the values above from the CIE's published data: the CIE S 026 melanopic curve, the CIE 015 reference spectra for phosphor-converted white LEDs (LED-B1 to B5) and for incandescent light (illuminant A), and the B(λ) function of IEC 62471. Real products vary with their phosphors: manufacturers' tables give about 0.41 for a typical 2700 K LED with a CRI of 80 and 0.45–0.48 for CRI 90 versions.
Two white LEDs at the same lux, against the eye's sensitivity curves
Two things follow. First, a warm-white LED is no worse for the body clock than the incandescent lamp it replaces: per lux it gives slightly less melanopic light (0.41 against 0.50) and a similar blue-light-hazard weighting (0.27 against 0.31), in line with the CIE's finding that LED and incandescent lamps for general lighting are similar at similar color temperatures. Second, the spectrum is only half the story. Twenty-five lux of 2700 K light at the eye gives the same melanopic EDI as 12 lux of 6500 K light, so a dim cool light can affect the clock less than a bright warm one. Color temperature itself is explained in the CRI and CCT guide.
How much light, and when
A 2022 expert consensus recommends at least 250 lux melanopic EDI at eye level during the day, no more than 10 from three hours before bedtime, and no more than 1 in the bedroom while sleeping. Indoors, 10 melanopic EDI is roughly 25 lux of warm-white LED light measured at the eye.
| When | Melanopic EDI at the eye | Equivalent 2700 K LED light at the eye |
|---|---|---|
| Daytime | At least 250 lux | About 615 lux, or daylight |
| Evening, from 3 h before bed | At most 10 lux | About 25 lux |
| Sleeping | At most 1 lux | About 2.5 lux, or darkness |
| At night, if light is needed | At most 10 lux | About 25 lux |
The recommendations, published by Brown and colleagues in PLoS Biology, are for healthy adults aged 18 to 55. They refer to light measured vertically at the eye, about 1.2 m above the floor for a seated person, not on the desk or floor. By day the target is easiest to meet with daylight, from a window or outdoors. Indoors it takes about 615 lux of 2700 K LED light at the eye, or about 290 lux of 6500 K.
The evening target is often met already. In the measurements Brown's group drew on, typical evening light in homes was about 30 lux, and on nearly half of the occasions melanopic EDI was already 10 or below. Warm-white LEDs reach the limit at about 22–25 lux at the eye; dimmable and dim-to-warm lamps make it easy (see circadian lighting). For sleep, the bedroom should be as dark as practical: blackout blinds help against streetlights, and fully shielded outdoor fixtures keep light out of windows (see dark-sky lighting).
Screens and night modes
A tablet at full brightness and reading distance gave about 32 lux at the eye in a 2015 study, similar to typical evening room light. Night modes warm the colors but do not dim the screen, and in a 2021 trial Night Shift did not improve sleep. Lower brightness and less evening use matter more.
In Chang's study the iPad was read at maximum brightness, 30–45 cm from the eyes, and its light peaked at 452 nm; the printed book was read under about 1 lux. Phillips' results show why a screen at that level matters: 10 to 50 lux of evening light delayed melatonin onset by 22 to 109 minutes. How much a screen adds depends on the room, though. In a dark bedroom it is most of the light reaching the eye, while in a lit living room it adds comparatively little.
Night modes such as Apple's Night Shift shift the screen's colors toward yellow. In Duraccio's trial, 167 students used an iPhone for an hour before bed for a week, with Night Shift on, with it off, or not at all, and wrist-worn sensors tracked their sleep. Sleep did not differ between the groups. Among those who slept more than about seven hours, the students who did not use a phone slept best. The American Academy of Ophthalmology suggests avoiding screens for two to three hours before bed and says night modes may help. Turning the brightness down cuts the melanopic dose in direct proportion.
The blue-light hazard and IEC 62471 risk groups
IEC 62471 classifies lamps by how long a person could stare into them before reaching the blue-light hazard limit: exempt (RG0) beyond 10,000 seconds, low risk (RG1) 100 to 10,000 seconds, moderate (RG2) 0.25 to 100 seconds, and high (RG3) under 0.25 seconds. Household white LED lamps are normally RG0 or RG1.
| Risk group | Staring time to reach the limit | Meaning in IEC 62471 |
|---|---|---|
| RG0, exempt | More than 10,000 s (2.8 h) | No photobiological hazard |
| RG1, low risk | 100–10,000 s | No hazard under normal behavioral limits on exposure |
| RG2, moderate risk | 0.25–100 s | No hazard because people look away from very bright light |
| RG3, high risk | Under 0.25 s | Hazardous even for a momentary or brief exposure |
The times assume fixed staring at the source, which the eye's natural movements and aversion to glare normally prevent. IEC TR 62778 explains how to apply the standard to light sources and luminaires: lamps for general lighting are assessed at the distance at which they give 500 lux, but not closer than 200 mm, and other sources at 200 mm. Under European rules, as ANSES notes, lamps must be in RG0 or RG1, while fixed luminaires in RG2 or above must carry information for users.
The CIE's 2019 position statement says that "practical assessments have shown that the blue light hazard exposure limits are not exceeded under all reasonably foreseeable use conditions" for lighting, and that "the exposure levels are often lower than experienced from viewing a blue sky". It concludes that the blue-light hazard "is not an issue for white-light sources used in general lighting, even for those that are blue-enriched". The exceptions are compact, intense sources viewed up close: ANSES lists flashlights, vehicle lights, some decorative lights and toys among the products that can reach RG2. Do not stare into a bare high-power LED or a flashlight, any more than into the sun.
Children's and older eyes
Children's eyes let more blue light reach the retina, and their body clocks respond more strongly to evening light. The CIE advises against blue indicator lights in toys, and SCHEER flags bright blue LEDs in toys for children under three. With age the lens yellows, so the same lamp gives older eyes less melanopic stimulus.
The lens filters short wavelengths, and the filtering increases through life. ANSES notes that children's lenses let more blue light through than adults' do, and that by 60 the lens filters about twice as much blue light as at 20. CIE S 026 corrects for this with an age adjustment; its standard observer is 32 years old. The evening targets above were set for adults, and Brown and colleagues point out that children's responses to evening light are larger. ANSES recommends limiting children's exposure to blue-rich screens before bedtime and at night.
For the retinal hazard, the CIE recommends against blue indicator lights in toys and, where they are used, a blue-light hazard limit ten times lower than normal. SCHEER warns that blue LEDs, including those in toys, "may be very dazzling and may induce photochemical retinopathy, which is a concern especially for children below three years of age". At the other end of life, SCHEER notes that older people may find blue-rich light uncomfortable.
Do LEDs cause macular degeneration?
No human study has shown that household LED lighting or screens cause age-related macular degeneration. The CIE calls such claims speculative and the EU's SCHEER found no evidence of harm in normal use. France's ANSES disagreed in 2019, concluding that chronic blue-rich exposure raises the risk, while rating warm-white home LEDs low risk.
The CIE's position statement says claims that LED lighting causes age-related macular degeneration (AMD) "are currently speculative and are not supported by the peer-reviewed literature". SCHEER found "no evidence of direct adverse health effects from LEDs emission in normal use (lamps and displays) by the general healthy population", and the American Academy of Ophthalmology says there is "no scientific evidence that blue light from digital devices causes damage to your eye".
ANSES took a more cautious view. Its 2019 opinion concluded that acute exposure to intense blue-rich light is toxic to the retina and that chronic exposure increases the risk of AMD, while finding that warm-white household LEDs are no different from traditional lighting, with a low risk of phototoxicity. Its advice is practical rather than alarming: choose warm white, below 3000 K, for the home, and limit children's exposure to blue-rich light from screens in the evening. The disagreement is about long-term, low-level exposure, which is hard to study in people, and none of these bodies advises against white LED lighting.
Do blue-light glasses help?
The evidence does not support blue-blocking glasses for eye strain or eye health. A 2023 Cochrane review of 17 trials found they may not reduce eye strain from computer use, with sleep effects unclear. The American Academy of Ophthalmology does not recommend them; it suggests the 20-20-20 rule for tired eyes.
The Cochrane review, covering 619 participants followed for between less than a day and five weeks, concluded that blue-light filtering lenses "may not attenuate symptoms of eye strain with computer use, over a short-term follow-up period". The UK College of Optometrists says the best evidence "does not support the use of blue-blocking spectacle lenses in the general population" to improve visual performance, relieve eye strain, improve sleep or protect the macula. When ANSES tested lenses, blue-blocking glasses filtered more than clear lenses with blue-filter coatings, but neither worked as protective equipment, and the coated lenses filtered very little in the melanopic band.
According to the AAO, eye strain at screens comes from how they are used, mainly long spells of close focus with less blinking, rather than from anything the screen emits. Its 20-20-20 rule: every 20 minutes, look at something at least 20 feet (6 m) away for at least 20 seconds.
Flicker and glare
Some LED products flicker at twice the mains frequency, 100 or 120 Hz. Most people cannot see it directly, but it can show up as stroboscopic effects and has been linked to headaches and eye strain. SCHEER lists flicker at 100 Hz and above as a potential concern; IEEE 1789 and EU rules set limits.
Flicker comes from the driver, not from the color of the light, and good drivers keep it very low. The LED drivers and flicker guide explains how it is measured, with percent flicker, Pst LM and SVM, and how to avoid it. Glare is the other comfort issue: a bare LED is a small, very bright source, so diffusers, shielding and indirect lighting make a room more comfortable.
Practical guidance
Get bright light, ideally daylight, during the day. In the three hours before bed, use warm 2700 K or warmer light, dimmed to around 25 lux at the eye or less. Turn screen brightness down, keep the bedroom dark, choose flicker-free products, and never stare into bare high-power LEDs or flashlights.
- Day: spend time outdoors or near a window. Indoors, brighter, cooler light at eye level helps meet the 250 melanopic EDI target.
- Evening: switch to warm, dimmed light; dim-to-warm lamps and tunable-white fixtures do this automatically.
- Screens: lower the brightness, and stop well before bed if you have trouble sleeping. A night mode is fine if you like it, but it is not a substitute for less light.
- Bedroom: as dark as practical, with blackout blinds if street or security lights shine in.
- Eyes: ordinary LED lighting and screens are safe to use; do not stare into bright point sources, and skip blue-light glasses unless an eye doctor advises otherwise.
Frequently asked questions
Is LED light bad for sleep?
Bright light in the evening delays sleep whatever its source, and blue-rich light does more per lux. A warm 2700 K LED gives slightly less melanopic light per lux than an incandescent bulb, so dim, warm LED lighting is no worse for sleep than the lamps it replaced.
Are LED lights bad for your eyes?
Not in normal use. Household LED lamps are exempt or low risk under IEC 62471, and the CIE, SCHEER and the American Academy of Ophthalmology find no evidence of eye damage from normal lighting or screens. Do not stare into bright point sources such as flashlights or bare high-power LEDs.
Is 2700 K or 3000 K better for a bedroom?
Either is fine. Per lux, a 2700 K LED gives about 10% less melanopic light than a 3000 K one (0.41 against 0.46). Dimming makes a bigger difference: in the evening, aim for about 25 lux at the eye or less.
Does Night Shift or night mode help you sleep?
Not much on its own. In a 2021 trial with 167 young adults, using an iPhone before bed with Night Shift on gave the same sleep as with it off. Turning the brightness down and putting the phone away earlier matter more.
Should I wear blue-light glasses?
Most people do not need them. A 2023 Cochrane review found they may not reduce eye strain, and the American Academy of Ophthalmology does not recommend them. For tired eyes, use the 20-20-20 rule.
What is melanopic EDI?
Melanopic equivalent daylight illuminance, in lux: the amount of daylight that would stimulate the eye's melanopsin cells as much as the light being measured. It is defined in CIE S 026 and used for circadian lighting targets.
Sources and further reading
- D. M. Berson, F. A. Dunn, M. Takao, "Phototransduction by retinal ganglion cells that set the circadian clock," Science 295, 1070–1073 (2002); S. Hattar et al., "Melanopsin-containing retinal ganglion cells: architecture, projections, and intrinsic photosensitivity," Science 295, 1065–1070 (2002).
- G. C. Brainard et al., "Action spectrum for melatonin regulation in humans: evidence for a novel circadian photoreceptor," Journal of Neuroscience 21, 6405–6412 (2001); K. Thapan, J. Arendt, D. J. Skene, "An action spectrum for melatonin suppression: evidence for a novel non-rod, non-cone photoreceptor system in humans," Journal of Physiology 535, 261–267 (2001).
- S. W. Lockley, G. C. Brainard, C. A. Czeisler, "High sensitivity of the human circadian melatonin rhythm to resetting by short wavelength light," Journal of Clinical Endocrinology & Metabolism 88, 4502–4505 (2003).
- J. J. Gooley et al., "Exposure to room light before bedtime suppresses melatonin onset and shortens melatonin duration in humans," Journal of Clinical Endocrinology & Metabolism 96, E463–E472 (2011).
- A.-M. Chang, D. Aeschbach, J. F. Duffy, C. A. Czeisler, "Evening use of light-emitting eReaders negatively affects sleep, circadian timing, and next-morning alertness," PNAS 112, 1232–1237 (2015).
- A. J. K. Phillips et al., "High sensitivity and interindividual variability in the response of the human circadian system to evening light," PNAS 116, 12019–12024 (2019).
- K. M. Duraccio, K. K. Zaugg, R. C. Blackburn, C. D. Jensen, "Does iPhone night shift mitigate negative effects of smartphone use on sleep outcomes in emerging adults?" Sleep Health 7, 478–484 (2021).
- T. M. Brown et al., "Recommendations for daytime, evening, and nighttime indoor light exposure to best support physiology, sleep, and wakefulness in healthy adults," PLoS Biology 20, e3001571 (2022).
- CIE S 026/E:2018, CIE System for Metrology of Optical Radiation for ipRGC-Influenced Responses to Light, and the CIE dataset of α-opic action spectra; CIE 015:2018, Colorimetry, 4th edition (illuminants A and D65 and the LED reference spectra).
- IEC 62471:2006 (CIE S 009), Photobiological safety of lamps and lamp systems; IEC TR 62778:2014, Application of IEC 62471 for the assessment of blue light hazard to light sources and luminaires.
- ICNIRP, "Guidelines on limits of exposure to incoherent visible and infrared radiation," Health Physics 105, 74–96 (2013).
- CIE, Position Statement on the Blue Light Hazard (23 April 2019).
- SCHEER (Scientific Committee on Health, Environmental and Emerging Risks), Final Opinion on Potential Risks to Human Health of Light Emitting Diodes (LEDs), European Commission (2018).
- ANSES, Effects on human health and the environment (fauna and flora) of systems using light-emitting diodes (LEDs), opinion and expert report (2019).
- S. Singh et al., "Blue-light filtering spectacle lenses for visual performance, sleep, and macular health in adults," Cochrane Database of Systematic Reviews (2023), CD013244.pub2.
- American Academy of Ophthalmology, "Should You Be Worried About Blue Light?" and "Are Blue Light-Blocking Glasses Worth It?"; The College of Optometrists, position statement "Blue blocking spectacle lenses" (2018).
- IEEE Std 1789-2015, IEEE Recommended Practices for Modulating Current in High-Brightness LEDs for Mitigating Health Risks to Viewers.
- Melanopic conversion tables for commercial LEDs: TRILUX, Human Centric Lighting product information (2023); Zumtobel, Melanopic LED action factors.
- Further reading: EBSCO Research Starters, "Effects of blue light technology".
Image credits
- Bedroom in the evening: AI-generated illustration for LEDlights.org.