Do LED face masks work? What the evidence shows

Acne

LED masks listing 415 nm blue light, the wavelength in most acne trials: CurrentBody Series 3, Dr. Dennis Gross FaceWare Pro and Shark CryoGlow.

• Blue plus red light against acne: the classic randomised trial Randomised trial, observer-blinded · 2000
  In this randomised trial, 107 people with mild to moderate acne used blue light, blue-plus-red light, white light or benzoyl peroxide for 12 weeks. Blue-plus-red light reduced inflammatory spots by 76% on average, better than white light throughout and better than benzoyl peroxide at weeks 8 and 12, with no significant short-term side effects.
• Does adding red light to blue help acne at home? A 2026 trial Open-label randomised controlled trial · 2026
  Twenty-three people with mild to moderate inflammatory acne used either a combined 660 nm red and 415 nm blue LED device or a blue-only device at home, daily for eight weeks. Both reduced inflammatory spots. Only red-blue reduced total spots significantly, but the two did not differ significantly on lesion counts.
• What 31 randomised trials say about LED light for skin Systematic review of 31 RCTs · 2018
  This systematic review analysed 31 randomised controlled trials of LED treatment for skin. Acne, cold sores and wound healing earned a grade B recommendation; other uses, including rejuvenation, got grade C or D. Common weaknesses were small samples, no blinding, no sham and varied settings.

Wrinkles and collagen

LED masks listing red light (620–670 nm): CurrentBody Series 3, Dr. Dennis Gross FaceWare Pro, Dr. Renú SpectraLift, Omnilux Contour and Shark CryoGlow.

• 830 nm, 633 nm or both? A double-blind, sham-controlled LED trial Randomised, placebo-controlled, double-blind, split-face · 2007
  Seventy-six people with facial wrinkles were randomised to 830 nm, 633 nm, both combined, or a sham light, twice a week for four weeks. Treated groups showed wrinkle reductions of up to 36% and elasticity gains of up to 19% compared with baseline, plus more collagen and elastic fibres under the microscope.
• A sham-controlled trial of a red LED mask: two sessions a week seemed enough Randomised, sham-controlled, double-blind · 2025
  Ninety-five women aged 45–60 used a 660 nm red LED mask two or three times a week for four weeks, or a sham mask. Dermatologist wrinkle scores showed no difference between groups, but image measurements found shorter frown and eye wrinkles with real light. Satisfaction was higher than sham, and three sessions weekly were no better than two.
• Red light increased collagen density against an untreated control Randomised controlled trial · 2014
  In this controlled trial, 113 people received 30 sessions of red or broad red-to-near-infrared light and were compared with 23 untreated controls. Treated groups showed better skin roughness, ultrasound-measured collagen density and blinded photo ratings. Broad-spectrum light gave no advantage over red light alone.
• Red and amber LED light both reduced eye-wrinkle volume by about 30% Randomised split-face trial · 2023
  In a randomised split-face trial, 137 women aged 40–65 received 10 sessions over four weeks of red (660 nm) light on one side and amber (590 nm) on the other, at the same dose. Wrinkle volume fell 31.6% with red and 29.9% with amber. Neither improved hydration or elasticity.
• The early clinic study behind red plus near-infrared LED masks Clinical study, no control group · 2005
  Thirty-one people received nine clinic treatments combining 633 nm red and 830 nm near-infrared LED light. By week 12, skin-surface measurements had improved significantly, 52% showed a 25–50% improvement in photoaging scores, and 81% reported improved eye-area wrinkles. There was no control group.
• A commercial three-wavelength LED mask tested on men Open-label study of a commercial mask · 2023
  This study assessed a commercially available face mask using 633, 830 and 1072 nm light in men over six weeks. Participants reported improvement in every category, especially fine lines, texture and youthful appearance, and digital photo analysis found improvements in wrinkles, spots and pores. No serious problems were reported.

All 12 studies at a glance

Newest first. “Control” says what the treated group was compared with; studies with no control group cannot separate the treatment’s effect from placebo or natural change.

Dose: what studies used and what devices can deliver

Dose (fluence) is irradiance multiplied by time, in J/cm². It applies to non-heating light such as LED and low-level laser. Studies summarised here that report a dose: A sham-controlled trial of a red LED mask: two sessions a week seemed enough (8.05 J/cm² per session at 6.4 mW/cm² for 21 minutes (660 nm)), Red and amber LED light both reduced eye-wrinkle volume by about 30% (3.8 J/cm² per session (red and amber)), Red light increased collagen density against an untreated control (About 9 J/cm² in the 611–650 nm range; irradiance and duration varied) and The early clinic study behind red plus near-infrared LED masks (126 J/cm² (633 nm) and 66 J/cm² (830 nm)).

Only two of the five LED masks publish enough to estimate a dose. No estimate is possible for CurrentBody Series 3, Dr. Dennis Gross FaceWare Pro and Shark CryoGlow, because no numeric irradiance is published on the checked page.

Estimates are ours, not the brand’s. They assume the published irradiance reaches the skin evenly for the whole session; brands do not say how irradiance was measured. A similar dose does not show a device produces a study’s results.

What the evidence cannot tell you

• Most trials used clinic LED panels or research devices, not the commercial masks compared here.
• Only some trials used a sham or untreated control; uncontrolled studies cannot separate the light’s effect from placebo or natural change.
• Many LED trials are small: the 2018 review notes many had fewer than 20 participants.
• Settings vary between trials, so results at one dose do not transfer to another.
• Some research is industry-funded or written by authors linked to device brands; each research page discloses this where known.