Blue Light, Night Mode and Sleep: What Screens Really Do

“Blue light from screens ruins your sleep” is one of those claims that is partly true and widely overstated. Light in the evening really does shift your body clock, and screens are a source of it. But a screen is very dim next to daylight, the color shift of night mode does less than the brightness slider, and the screen’s effect on sleep is tangled up with what you are doing on it. This guide goes through what the research measured.

It is general information, not medical advice. If you have trouble sleeping most nights, talk to a doctor.

The light that sets your body clock

Your eyes do more than see. A small set of cells in the retina contains a pigment called melanopsin and reports overall light levels to the brain’s clock, which uses them to time sleep and the release of melatonin, the hormone that rises in the evening. Melanopsin is most sensitive at around 480 nm, a sky blue, while the cones you see brightness with peak at around 555 nm, a yellow-green. That’s why “blue light” gets the blame: it is the part of the spectrum this system responds to most.

Because ordinary lux measures light as your cones see it, researchers now use a second unit for this system: melanopic equivalent daylight illuminance, or melanopic EDI, defined in the CIE S 026 standard and measured in lux. It answers “how strongly does this light act on the body clock?”

How much light a screen gives off

In 2022 a group of sleep and light researchers (Brown and colleagues) published recommendations in these units:

When	Recommended melanopic EDI at the eye
Daytime	at least 250 lux
Evening, from 3 hours before bed	at most 10 lux
Sleeping	at most 1 lux

Now the screens:

A phone at full brightness showing a white screen 15 cm from your eyes gives about 64 lux melanopic EDI, by Schlangen and Price’s calculation. They describe it as a plausible worst case: darker content, lower brightness or holding the phone further away all reduce it.
An iPad at maximum brightness, held at about 40 cm in Chang and colleagues’ 2015 study, gave about 31 melanopic lux at the eye. A printed book under the room’s dim light gave 0.65.
Brown and colleagues note that displays can provide more than 60 lux melanopic EDI.

Compare daylight. A review by Blume and colleagues gives up to 100,000 lux in direct sun, about 25,000 lux in full daylight, and about 500 lux in a typical office. So a screen is tiny next to the sky, and during the day it barely registers. In a dim room in the evening, though, it can be the brightest thing you look at, and well above the 10 lux the recommendations suggest for the last hours before bed.

What the sleep studies found

The strong result: hours of bright reading in a dim room

The study most often cited is Chang and colleagues’ 2015 paper in PNAS. Twelve young adults lived in a lab for two weeks. For five evenings they read on an iPad at maximum brightness for about four hours before bed in a dim room; for five evenings they read printed books. Compared with the books, the iPad:

cut evening melatonin by about 55%;
pushed the evening rise in melatonin more than 1.5 hours later;
made people take nearly 10 minutes longer to fall asleep (about 26 minutes against 16);
reduced REM sleep, and left them sleepier and slower to wake the next morning.

Total sleep time did not differ. It is a carefully controlled study, and also a worst case: four hours a night, maximum brightness, a dim room and a small group. It shows that bright evening screen light can shift the body clock. It does not show that checking your phone for 15 minutes before bed delays your sleep by 1.5 hours.

Light, or what you are doing?

Outside the lab, the picture blurs. A 2015 review by Hale and Guan of 67 studies of children and teenagers found screen time linked to worse sleep, mostly later bedtimes and shorter sleep, in 90% of them. It named three possible reasons: screens take time away from sleep, the content keeps people alert, and the light acts on the body clock. Most of the studies were observational, so they can’t separate the three, and the review says causation is not confirmed. Blume and colleagues add that nobody yet knows how much of phones’ effect on sleep comes from the light rather than the engagement.

Does night mode help?

Night mode (Night Shift on Apple devices, Night Light on Android and Windows) shifts the screen toward orange, which cuts the blue the body clock responds to most. The studies suggest it helps less than people expect:

Nagare, Plitnick and Figueiro (2019) had 12 people use iPads at maximum brightness with Night Shift at its warmest and least warm settings. Melatonin suppression over two hours was about 10% and 17%, not significantly different, and both still suppressed melatonin. Their conclusion: changing the color without changing the brightness may not be enough.
Duraccio and colleagues (2021) randomly assigned 167 people aged 18 to 24 to a week of using an iPhone with Night Shift, an iPhone without it, or no phone in the hour before bed, and tracked sleep with wrist monitors. There were no significant differences in sleep outcomes between the three groups. In an exploratory analysis, people who slept more than about 6.8 hours a night slept best without the phone.
Brown and colleagues put the effect of spectral changes on screens at roughly a 50% reduction in melanopic output, and expect them to work best combined with dimming.

Brightness is the bigger lever. Blume and colleagues give an example for one phone: at full brightness, Night Shift cut melanopsin activation by about two thirds, while turning the brightness to minimum cut it to less than 1% of the full-brightness level.

So night mode is not useless, but it is the smaller of the two controls. A dim warm screen beats a bright warm one.

Blue-light glasses for sleep

The 2023 Cochrane review of blue-light-filtering glasses (Singh, Downie and colleagues) looked at sleep too. Across six trials with 148 participants, three found better sleep and three found no difference. The review rated the evidence as very low certainty and called the effect on sleep indeterminate. Glasses are not night mode, but the conclusion is similar: if there is an effect, nobody has shown it clearly.

Does screen blue light damage your eyes?

No, on current evidence. O’Hagan, Khazova and Price measured phones, tablets, laptops and monitors at maximum brightness showing a white screen, the worst case. Phones reached 0.15 to 0.38% of the international exposure limit for blue-light damage to the retina over long viewing. A clear June sky in southern England reached 10.4%. None of the screens came near the limits, even for long viewing times.

The American Academy of Ophthalmology says there is no scientific evidence that blue light from digital devices damages the eye, and does not recommend blue-light-blocking glasses for computer use. The sun is by far the largest source of blue light most people get.

What to actually do

Dim the screen in the last hours before bed. It is the biggest single change. Holding the phone further away helps too.
Use night mode if you like it, as an extra, not as the fix.
Prefer dark backgrounds at night. Less lit area means less light.
Mind what you are doing. If the evidence can’t separate light from engagement, the safest assumption is that both matter. Stopping an absorbing game or feed earlier probably does as much as any setting.
Get daylight during the day. The recommendations ask for at least 250 lux melanopic EDI in the daytime, which is easy outdoors and often missed indoors.
Keep the bedroom dark, under 1 lux melanopic EDI while you sleep by the same recommendations, which means no screen lighting up on the nightstand.

One catch on phones with OLED screens: the bottom of the brightness slider is where many of them flicker hardest, because they dim by switching the panel on and off. If very low brightness gives you headaches, screen flicker and PWM dimming explains the trade-off and the workarounds. And if your eyes are tired rather than your body clock, see digital eye strain: what helps.

If you calibrate a monitor, switch night mode off first: it deliberately tints the image, and our monitor calibration steps assume a neutral white.