# Red Light Therapy Wavelength Guide

_Updated 2026-03-20_

The optimal wavelengths for red light therapy are 660nm (visible red) and 850nm (near-infrared). These correspond to the peak absorption wavelengths of cytochrome c oxidase, the mitochondrial enzyme responsible for photobiomodulation. 660nm targets superficial skin cells for collagen synthesis, while 850nm penetrates deeper for anti-inflammatory and circulatory benefits.

## Electromagnetic Spectrum Overview

Visible light occupies a narrow band of the electromagnetic spectrum between approximately 380nm (violet) and 700nm (deep red). Beyond 700nm lies the near-infrared (NIR) range, which extends to about 1400nm. Red light therapy operates within the "therapeutic optical window" — the wavelength range between 600nm and 1100nm where biological tissue is most transparent to light penetration.

Below 600nm, light is heavily absorbed by hemoglobin and melanin before reaching therapeutic targets. Above 1100nm, water absorption dominates and converts light energy to heat rather than photochemical stimulation. Within the optical window, two "sub-windows" show peak efficacy: 620–680nm (red) and 780–870nm (near-infrared), corresponding directly to the absorption peaks of cytochrome c oxidase as mapped by Karu (2010).

Understanding this spectrum is essential for evaluating any light therapy device. Marketing claims about "full-spectrum" or "multi-wavelength" devices are meaningless unless the specific wavelengths fall within the CCO absorption peaks. A device emitting at 730nm, for example, falls in the absorption trough between the red and NIR peaks and delivers significantly less photobiomodulation effect per photon than one at 660nm or 850nm.

> **Key takeaway:** The therapeutic optical window (600–1100nm) contains two efficacy peaks matching cytochrome c oxidase absorption: 620–680nm (red) and 780–870nm (near-infrared).

## Red Light (620–700nm)

The red light spectrum from 620nm to 700nm represents the most extensively studied wavelength range in photobiomodulation. Within this range, 660nm has emerged as the gold-standard wavelength based on decades of research. Barolet (2009) demonstrated that 660nm LED exposure increased type-I procollagen expression by 31% while simultaneously reducing MMP-1 (collagenase) activity — a dual mechanism that both builds new collagen and protects existing collagen from degradation.

The penetration depth of red light at 660nm is approximately 2–3mm in human tissue, which places it squarely within the papillary and reticular dermis — the tissue layers where fibroblasts (collagen-producing cells) reside. This makes 660nm ideally suited for skin rejuvenation applications: wrinkle reduction, skin texture improvement, and age spot reduction. The periorbital skin, being the thinnest on the body (0.5mm at the tear trough), is particularly responsive to 660nm treatment.

Other wavelengths within the red range have been studied but show inferior results. 630nm (the wavelength used in many inexpensive consumer devices) is 20–30% less efficiently absorbed by CCO compared to 660nm. 700nm sits at the transition to NIR and loses the strong CCO absorption peak. This is why Mota et al. (2023) specifically used 660nm — not a broader red range — to achieve the 31.6% periocular wrinkle reduction that represents the gold-standard clinical outcome.

> **Key takeaway:** 660nm is the gold-standard red wavelength — 20–30% more efficiently absorbed by cytochrome c oxidase than 630nm, with the ideal penetration depth for dermal fibroblast stimulation.

## Near-Infrared (700–1100nm)

Near-infrared (NIR) light spans 700nm to 1100nm, but the therapeutically relevant range for photobiomodulation is primarily 780–870nm, centered on the 850nm absorption peak of cytochrome c oxidase. NIR is invisible to the human eye but penetrates significantly deeper than visible red light — up to 5–10mm in tissue, reaching subcutaneous fat, muscle fibers, blood vessels, and in thin areas like the periorbital region, even bone.

The deeper penetration of 850nm NIR makes it the preferred wavelength for inflammatory conditions. Hamblin (2017) demonstrated that NIR photobiomodulation reduces pro-inflammatory cytokines (TNF-α, IL-1β, IL-6) and promotes the transition of macrophages from pro-inflammatory (M1) to anti-inflammatory (M2) phenotypes. For periorbital treatment, this mechanism directly addresses the vascular inflammation responsible for dark circles and the fluid retention causing puffiness.

Lee et al. (2007) provided critical evidence that combined red (633nm) and NIR (830nm) wavelengths produce superior outcomes compared to either wavelength alone — 36% wrinkle reduction and 19% elasticity improvement in a double-blind RCT. This synergy occurs because 660nm targets superficial collagen production while 850nm addresses deeper inflammation and microcirculation, producing results across the full dermal thickness.

> **Key takeaway:** 850nm near-infrared penetrates 5–10mm deep, targeting inflammation and microcirculation. Combined with 660nm red light, it produces synergistic results that exceed either wavelength alone.

## Why 660nm Is Optimal for Skin

The selection of 660nm as the optimal visible red wavelength is not arbitrary — it represents the precise intersection of three key factors: maximum CCO absorption, optimal tissue penetration depth for dermal targets, and the strongest clinical evidence base. Karu (2010) mapped the action spectrum of cytochrome c oxidase and identified 660nm as sitting within the primary absorption peak of the enzyme's copper-A (CuA) center.

At the tissue level, 660nm light achieves what photobiologists call "target layer selectivity" — it penetrates deep enough to reach the dermis (where fibroblasts reside) but is sufficiently absorbed that most of its energy is deposited within the target tissue rather than passing through to deeper structures. This means a higher proportion of incident photons contribute to therapeutic effect compared to longer wavelengths that pass through the dermis.

The clinical evidence supports this specificity. Mota et al. (2023) used 660nm for periocular wrinkle reduction (31.6%). Barolet (2009) used 660nm for procollagen stimulation (31% increase). Wunsch & Matuschka (2014) used a range centered on 660nm and confirmed collagen density increase via ultrasound. The consistency of positive outcomes specifically at 660nm — across different research groups, geographies, and study designs — constitutes a robust evidence base for this wavelength.

> **Key takeaway:** 660nm sits at the intersection of maximum CCO absorption and ideal dermal penetration depth — with the strongest clinical evidence base of any PBM wavelength for skin rejuvenation.

## Why 850nm for Deep Tissue

The 850nm wavelength targets the second major absorption peak of cytochrome c oxidase — specifically the binuclear center containing heme a3 and CuB. At this wavelength, photons penetrate tissue approximately 3–5 times deeper than at 660nm due to reduced scattering and absorption by superficial chromophores like melanin and hemoglobin. This deep penetration makes 850nm the wavelength of choice for targets beyond the superficial dermis.

For periorbital applications, 850nm reaches the orbicularis oculi muscle (responsible for dynamic wrinkles), the periorbital vasculature (where stagnant blood pooling causes dark circles), and the lymphatic channels (where impaired drainage causes puffiness). Goldberg et al. (2006) demonstrated that combined 633nm and 830nm LED treatment achieved 80% patient-reported improvement in periorbital wrinkle softening — a result that required the deep-tissue effects of NIR light.

The anti-inflammatory properties of 850nm are particularly relevant for the thin periorbital tissue, which is highly susceptible to chronic low-grade inflammation. Hamblin (2017) showed that NIR PBM suppresses NF-κB activation — the master switch for inflammatory gene expression — and reduces circulating levels of TNF-α, IL-1β, and IL-6. This creates an anti-inflammatory microenvironment that reduces edema, improves vascular tone, and slows the collagen degradation caused by chronic inflammaging.

> **Key takeaway:** 850nm penetrates 3–5x deeper than 660nm, reaching the periorbital muscle, vasculature, and lymphatics — directly addressing dark circles, puffiness, and dynamic wrinkle formation.

## Common Wavelengths Compared

Consumer devices commonly use wavelengths ranging from 620nm to 1000nm, but not all perform equally. At 620nm, CCO absorption is moderate but penetration is limited — suitable for very superficial epidermal effects but insufficient for dermal collagen stimulation. At 630nm (the most common in budget devices), efficacy drops 20–30% compared to 660nm due to the absorption curve falloff. The 635nm lasers used in some clinics are closer to optimal but still inferior to LED arrays at 660nm for broad-area coverage.

In the NIR range, 810nm is used in some transcranial PBM devices targeting brain tissue, where slightly shorter NIR wavelengths improve skull penetration. At 850nm, the absorption-to-penetration ratio is optimized for soft tissue targets. Devices claiming "1000nm" capability often fall outside the CCO absorption spectrum entirely, producing primarily thermal (heating) effects rather than photochemical stimulation.

Dual-wavelength devices combining 660nm and 850nm represent the current evidence-based standard. Ngoc et al. (2023), in a systematic review screening 554 articles, concluded that combined red and NIR LED therapy demonstrates the strongest evidence for skin rejuvenation outcomes. Single-wavelength devices deliver incomplete therapy — missing either the superficial collagen stimulation (if NIR-only) or the deep anti-inflammatory effect (if red-only). The clinical evidence from Lee et al. (2007) confirms the synergistic superiority of the dual-wavelength approach.

> **Key takeaway:** Dual-wavelength (660nm + 850nm) devices represent the evidence-based gold standard. Single-wavelength devices and off-peak wavelengths (630nm, 1000nm) deliver incomplete or inferior therapy.

## FAQ

**What is the best wavelength for wrinkle reduction?**

660nm is the most clinically validated wavelength for wrinkle reduction, with Mota et al. (2023) demonstrating 31.6% periocular wrinkle volume reduction. Combined with 850nm, synergistic effects produce even greater improvement (Lee et al., 2007 showed 36% wrinkle reduction with dual wavelengths).

**Does 630nm work as well as 660nm?**

No. While 630nm produces some photobiomodulation effect, cytochrome c oxidase absorption is 20–30% lower at 630nm compared to 660nm. The largest and most successful RCTs (Mota 2023, Barolet 2009) all use 660nm specifically.

**Can I use only one wavelength?**

You can, but clinical evidence shows dual-wavelength protocols (660nm + 850nm) outperform single-wavelength treatments. Lee et al. (2007) demonstrated this in a double-blind RCT where combined wavelengths achieved superior wrinkle and elasticity outcomes.

**What wavelength is best for dark circles?**

850nm near-infrared is the primary wavelength for dark circles, as it penetrates deep enough to reach the periorbital vasculature where blood pooling occurs. The anti-inflammatory effects documented by Hamblin (2017) directly address the vascular inflammation component.

**Is 940nm or 1000nm effective for skin?**

These wavelengths fall outside the primary CCO absorption peaks and produce weaker photobiomodulation effects for skin targets. Water absorption increases significantly above 900nm, converting light energy to heat rather than photochemical stimulation. 850nm remains the evidence-based optimal NIR wavelength.

**Why do some devices use blue or green light?**

Blue light (400–470nm) targets porphyrins in bacteria and is used for acne treatment. Green light (520–560nm) targets melanin for pigmentation concerns. Neither blue nor green light produces significant photobiomodulation via cytochrome c oxidase — they work through completely different mechanisms.

**How important is wavelength precision?**

Very important. Even 10nm of wavelength drift from the 660nm peak can reduce CCO absorption by 20–30%. Clinical-grade devices use binned LEDs sorted to ±2–5nm tolerance, while consumer devices may vary by ±15–20nm, significantly reducing therapeutic efficacy.