Last updated: July 2026 · Written by the Dermfix phototherapy team
Short answer: Red light in the 600–700nm range is absorbed close to the skin's surface, while near-infrared light in the 780–950nm range travels further into tissue, reaching muscle and joint depth. This split isn't marketing shorthand — it comes from how differently these wavelengths scatter and absorb in living tissue, and it's why most panels are built to offer a range across both bands rather than a single wavelength.
What is red light (600–700nm)?
Red light sits in the visible part of the spectrum. According to a widely cited mechanistic review of low-level light therapy (LLLT) by Michael R. Hamblin of Harvard Medical School's Wellman Center for Photomedicine, wavelengths in the 600–700nm range are generally chosen for treating superficial tissue, precisely because they don't travel as far before being absorbed or scattered.
The reason light behaves this way at all comes down to tissue optics. Hemoglobin and melanin — the main light-absorbing molecules in skin — strongly absorb wavelengths shorter than 600nm, while water only starts absorbing heavily above roughly 1150nm. Between those two points lies what Hamblin's review calls the tissue's “optical window,” the band where red and near-infrared light penetrate most effectively rather than being immediately absorbed or scattered away.
What is near-infrared light (780–950nm)?
Near-infrared (NIR) sits just beyond what the human eye can detect. Per the same review, wavelengths between 780 and 950nm are typically selected for deeper-seated tissues, since longer wavelengths within the optical window travel further through skin, fat, and muscle before being absorbed.
Interestingly, the paper notes that wavelengths between 700 and 770nm — the gap between the red and NIR bands — are not considered to have much biological activity at all. That's a useful detail: it's not simply “longer wavelength, more penetration” in a straight line. There's a dip in effectiveness right between the two useful bands, which is part of why panels tend to cluster their output around red and NIR specifically rather than spreading evenly across the spectrum.
At the cellular level, the review points to cytochrome c oxidase — an enzyme in the mitochondrial respiratory chain — as the likely primary light-absorbing molecule (chromophore) responsible for these effects, since its absorption spectrum closely tracks the wavelengths found to be biologically active. One study cited in the review pinpointed four specific peaks within the action spectrum: roughly 613–624nm, 668–684nm, 751–772nm, and 813–846nm — which is why red and NIR devices tend to cluster their diodes around these specific ranges rather than picking wavelengths at random.
Do you need both?
Based on how the two bands are described in the literature, yes — a panel offering only one band limits you to either surface-level or deeper coverage, not both. Some devices go further and combine a red wavelength with an NIR wavelength on the reasoning that pairing the two can have additive effects and lets a single device be used across a broader range of applications. This is consistent with why “red + NIR together” has become the baseline combination for general-purpose panels rather than either wavelength alone.
What about dose and other parameters?
Wavelength is only one piece of the puzzle — and notably, the review describes it as the parameter with the most agreement in the field, compared to dose, pulsing, and coherence, which remain far more contested. On dosing, red wavelengths for superficial use are typically delivered in the range of 1–10 J/cm² (with 4 J/cm² commonly cited), while NIR wavelengths for deeper structures tend to run higher, in the 10–50 J/cm² range. Treatment is usually repeated daily or every other day over a course of around two weeks.
It's also worth flagging what the review calls a biphasic dose response: more light isn't automatically better. Effectiveness rises with dose up to an optimal point, but doses pushed too high can produce a diminished — or even negative — outcome. This is part of why manufacturers publish specific irradiance and session-time guidance rather than leaving dose open-ended.
Frequently asked questions
Is 660nm/red or 850nm/NIR “better”?
Neither — they're suited to different depths. Red for surface-level tissue, NIR for deeper muscle and joint tissue. Reviews on the underlying mechanisms consistently treat them as complementary rather than competing.
Can I see near-infrared light?
No — NIR sits outside the visible spectrum, which is why NIR-only devices are often paired with a faint visible indicator so users can confirm the panel is active.
Does the wavelength combination matter more than the number of wavelengths?
Based on the mechanistic literature, yes. The strongest area of scientific agreement is that red (600–700nm) and NIR (780–950nm) are the two bands worth prioritizing, with the 700–770nm gap between them showing little activity. A long list of wavelengths matters less than solid, well-documented output across the ranges that are actually shown to be active.
Sourced from: Hamblin, M.R., “Mechanisms of Low Level Light Therapy,” Department of Dermatology, Harvard Medical School / Wellman Center for Photomedicine, Massachusetts General Hospital. Specific findings on action-spectrum peaks reference Karu, T.I. and Kolyakov, S.F., “Exact action spectra for cellular responses relevant to phototherapy,” Photomed Laser Surg 23 (2005), as cited within the review.
Disclaimer: this article summarises published research for general information. It is not medical advice and does not describe the intended purpose of any Dermfix product.