Wavelengths in Red Light Therapy: What They Mean, Why They Matter, and How to Evaluate a Device

Deep dive

The science: why red and near-infrared

What is a wavelength, and where does red light fit?

Light travels in energy waves, and a wavelength is simply the distance between the peaks of those waves. We measure them in nanometers (nm), where one nanometer is a microscopic one-billionth of a meter.

Visible light is part of a much bigger family. X-rays, ultraviolet (UV) light, microwaves, and the radio waves that carry TV broadcasts are all the same kind of energy. The only thing that changes is the wavelength (NASA Science).

• Short wavelengths (gamma rays, X-rays, and UV) carry enough energy per photon to damage DNA. That’s why you wear a lead apron for an X-ray and why UV causes sunburn.
• Visible light runs from about 400 nm (violet) to about 700 nm (red). Red is the long end of the visible spectrum.
• Near-infrared starts just past 700 nm, where human vision stops. TV remotes use it, which is why you can’t see the remote’s light (NASA Science).
• Longer still are mid- and far-infrared (felt as heat), then microwaves, then radio and TV waves.

Figure 1. Red and near-infrared light are a narrow slice of the electromagnetic spectrum, between visible light and the heat-carrying far infrared. Band edges are approximate. Diagram: MedcoVet, based on NASA Science and *Chung et al., 2012.*

Red and near-infrared light used in therapy are non-ionizing (Frankowski et al., 2025) (unless in doses over 1M times stronger than typical treatments). Unlike X-rays, they don’t carry enough energy to knock electrons out of atoms or damage DNA (US EPA). At therapeutic doses, their main effect is photochemical, not heating (Chung et al., 2012).

These boundaries are just guidelines, not strict biological limits. A 670 nm photon does not act differently from a 669 nm photon. Also, device output has a range, so the label number usually refers to the peak, not every photon the device emits.

Cool fact: you can see near-infrared light with your phone

Your eyes can’t see 850 nm light, but most phone cameras can. Try it with a TV remote: point it at your phone’s camera, press a button, and watch the screen. The LED flashes on screen, even though it looks dark to your eye.

You can do the same with a red light therapy device. If your device lets you switch the red LEDs off, hold the near-infrared LEDs up to the camera. The “dark” LEDs light up on the screen. If the back camera doesn’t show it, try the selfie camera. Many back cameras have a stronger infrared-blocking filter.

One caution: don’t stare into therapy LEDs, and never point them at your pet’s eyes. And remember that the glow on your screen shows the light is there. It doesn’t measure how much power the device delivers.

Choosing a wavelength comes down to two steps:

1. Start with light that has a biological impact. That means light that does something you want, like reducing inflammation, destroying bacteria, or increasing collagen production. Most of that research focuses on red and near-infrared light, but research on blue and green light is growing.
2. Remove light that doesn’t make it to the target tissue. Light absorbed by pigment, blood, or water along the way can’t help the target.

How common PBM wavelengths compare

Once you group wavelengths by what actually changes, the picture gets simpler. Within red (about 620 to 680 nm), the differences are small. The same is true within near-infrared (about 800 to 900 nm). The big jumps happen between colors.

Band	Common examples	How far it gets	Commonly used for	The catch
Blue ~400–495 nm	405, 415, 450 nm	Least. About 1 mm in skin (Ash et al., 2017)	Killing surface bacteria; acne; infected wounds (Wang et al., 2017c)	Heavily absorbed by melanin and blood. Slowed stem-cell growth where red stimulated it (Wang et al., 2017a)
Green ~495–570 nm	525, 532, 540 nm	Shallow. Strongly absorbed by hemoglobin	Pain and migraine research, mostly through the eyes (Ibrahim et al., 2017; Martin et al., 2021)	Not a way to reach muscles or joints
Red ~620–680 nm	630, 633, 635, 660, 670 nm	Several millimeters; best for surface and near-surface targets	Wounds, skin, and shallow joints (paw, carpus, tarsus). Also incisions and surface inflammation; 670 nm in neurologic and retinal research	660 nm passes a bit better than 630 nm because blood absorbs less, but no consistent evidence shows one red beats another
Near-infrared~800–900 nm	808, 810, 830, 850, 875 nm	Deepest of the four. Still, more than 90% of laser energy was absorbed within the first 10 mm in one ex vivo study (Kaub and Schmitz, 2022)	Wounds too, and the only band that reaches deeper joints, the spine, muscles, tendons, ligaments, and nerves	810 vs. 850 nm differences are small. Nearby-wavelength evidence is supportive, not identical

Why only one red row and one near-infrared row? Because the numbers inside each band are close. Melanin absorbs only about 1.2 times more at 630 nm than at 660 nm, and about 1.2 times more at 810 nm than at 850 nm. Compare that with about 2.4 times more at 660 nm than at 850 nm (Jacques, 2013). Hemoglobin shows a bigger gap within red: whole blood absorbs about 1.6 to 1.9 times more at 630 nm than at 660 nm (Prahl, OMLC), which is one reason 660 nm is so popular. The action spectrum also shows separate peaks near 620 and 670 nm (Karu and Kolyakov, 2005). But no consistent head-to-head clinical evidence picks a winner inside either band.

Some longer near-infrared wavelengths are special cases. They behave differently enough that they deserve their own rows. The WAIT Test™ counts only 800 to 900 nm as near-infrared, so 875 nm counts. Superpulsed 905 nm lasers are scored separately, against WALT’s superpulsed chart, and continuous light at 940 nm and longer doesn’t count:

Wavelength	Typical use	Evidence	Main caution
904 to 905 nm	Often used in superpulsed laser systems	Clinical and preclinical research exists for specific pulsed protocols	Peak power, pulse width, frequency, and average power are central to interpreting the wavelength
940 to 980 nm	Longer NIR designs	Can produce biological effects through mechanisms that may differ from 810 nm	Water absorption and heating become more important; evidence from 810 or 850 nm cannot simply be transferred
1,064 nm	Used in some deeper-tissue and neurologic systems	Some transmission and clinical research is available	It sits outside the best-known PBM action bands and needs use-specific support

These tables provide scientific context. They are not a treatment chart.

Red light: commonly used for targets closer to the surface

Visible red PBM commonly uses wavelengths from about 620 to 680 nm. Frequently studied examples include 633, 635, and 660 nm. The original PBM workhorse, the helium-neon laser, emitted at 632.8 nm (Chung et al., 2012).

Red light is often selected for:

• Wounds and incisions
• Skin and other superficial tissues
• Surface inflammation or irritation
• Structures located close to the skin
• Shallow joints such as the paw, carpus, and tarsus
• Treatment areas where visible placement feedback is useful

Red wavelengths still go into tissue. Calling them “superficial” just means they target areas near the surface, not that red light only works at the surface.

With irradiance and surface energy density held constant, 635 and 810 nm improved healing measures, while 730 and 980 nm did not under the same tested settings. The strongest response occurred at 810 nm (Gupta et al., 2014).

Non-invasive red and near-infrared exposure has also been reported to support wound healing in animal models of impaired cutaneous repair (Yadav and Gupta, 2017).

In cultured human adipose-derived stem cells, both 660 and 810 nm increased ATP and proliferation under the study conditions (Wang et al., 2017a).

These studies support the biological activity of red and near-infrared bands. They also show why a result belongs to the wavelength, dose, model, and outcome that were actually tested.

Near-infrared light: commonly used for deeper targets

Near-infrared light is outside normal human vision. PBM devices use several near-infrared bands, with 808, 810, 830, and 850 nm among the most common in research and consumer devices.

For inflammation below the skin, near-infrared reaches the target tissue. It is often selected for:

• Muscles
• Joints
• Tendons and ligaments
• Nerves
• Deeper areas of pain or inflammation
• Anatomical sites with several layers above the target

Wavelengths around 800 to 900 nm usually pass through tissue better than visible red light. But that doesn’t mean there’s a set treatment depth, or that all light reaching deeper tissue is still useful. And not all infrared is the same: the longer wavelengths covered below work differently.

Does research support 850 nm?

Yes, with an important qualification.

Wavelengths around 810 to 850 nm sit within a well-studied near-infrared region. Laboratory, animal, and human studies report biological effects across this range. A randomized, double-blind, placebo-controlled crossover trial used an LED cluster combining 660 and 850 nm before a knee-extensor fatigue test and found less torque loss after exercise (Baroni et al., 2010).

That trial supports a 660 and 850 nm combination protocol. It cannot isolate the effect of 850 nm because both wavelengths were delivered together. Evidence from nearby wavelengths, especially 810 and 830 nm, strengthens the scientific basis for the region but should not be presented as direct proof for every 850 nm device.

This difference matters when you look at product claims. Evidence for a wavelength range can back up a design choice. But evidence for a specific emitter, output, dose, and clinical use is even stronger.

Is a dual-wavelength device better?

Sometimes, a dual-wavelength device is the smarter design. But you have to prove it’s better, not just assume.

A device that combines red and near-infrared light can provide:

• Coverage over a broader range of tissue depths
• A way to address treatment areas containing skin, connective tissue, muscle, and joint structures
• Simpler home use when the user should not have to select a wavelength for each placement
• Access to two well-studied portions of the PBM spectrum

Research on multiwavelength PBM reports that combinations can produce different effects from single wavelengths, including additive, complementary, or occasionally opposing responses. Results depend on the wavelengths, sequence, timing, dose from each source, and biological model (Lima, Sergio, and Fonseca, 2020).

The takeaway: using two wavelengths can be a smart way to cover different tissue depths. But more wavelengths don’t automatically mean better quality. For example, adding blue or green to a deep-tissue device wastes power on light that won’t reach the target.

What a dual-wavelength company should disclose

Ask:

• Do both wavelengths operate at the same time?
• Can they be selected separately?
• How many emitters produce each wavelength?
• How much average optical power comes from each wavelength?
• Is the stated irradiance measured with both wavelengths active?
• Does the cited research use both wavelengths together?
• Is the dose calculated from the combined output or from one channel?

A device may advertise five wavelengths while delivering most of its power through only one. An emitter count also does not answer the power question because different LEDs can have different outputs.

What changes when the patient has fur?

Fur blocks light in two ways. Its pigment absorbs light (that’s the melanin from Step 2), and the coat’s density and structure physically block and scatter it. Choosing a longer, near-infrared wavelength can reduce the pigment-related loss somewhat, since fur melanin absorbs less at longer wavelengths just as skin melanin does. But it does not fix the physical blocking caused by a thick or unshaved coat.

In a study of 47 dogs, wavelength, coat color, shaving, and device power significantly affected measured light transmission. Under the tested conditions, darker and unshaved coats transmitted less light. Coat length itself was not a significant predictor in that study (Hochman-Elam, Heidel, and Shmalberg, 2020).

This doesn’t mean coat length never matters. In this study, pigment and whether the dog had hair explained more of the difference.

In practice, getting light past fur comes down to application: parting the coat, contact, and placement. The Target Tissue guide covers that in detail.

Skin pigment also affects absorption and heating. A human study comparing red and infrared laser transmission found that skin color and tissue thickness influenced transmittance, reflectance, and temperature response (Souza-Barros et al., 2018). The same surface setting should not be assumed to deliver the same target-tissue dose in every patient.

What does veterinary research show about wavelength?

Veterinary PBM research supports continued clinical interest, but it does not identify one best wavelength for every species, condition, or tissue.

A 2023 systematic review included 45 veterinary articles published from 1980 through 2020: 24 involved dogs, one involved cats, and the remainder involved horses. The authors found conflicting results, wide variation in wavelength, laser class, dose, session frequency, and treatment duration, and generally moderate or high risk of bias (Millis and Bergh, 2023; Bunch, 2023).

That evidence gap matters in three ways:

• A human study can support a mechanism or wavelength rationale, but it is not direct clinical proof for dogs or cats.
• A canine study using one full protocol cannot show that wavelength was the only reason for the result.
• Feline evidence is especially limited, so dog protocols should not be treated as proven cat protocols.

The good news: the cell-level mechanism from Step 1 is the same in people, dogs, and cats. Mitochondria and cytochrome c oxidase work the same way across mammals. What differs is the path to the target: coat, skin thickness, body size, and anatomy.

Newer veterinary research continues to add condition-specific data. A 2026 prospective multicenter pilot study tested a dual 808 and 905 nm PBM protocol in 22 dogs with otitis externa. Clinical scores and yeast counts improved across five sessions, with no adverse events reported. The study had no sham control; some ears also received a cleanser, and the authors called for controlled trials, so read this as early evidence for that exact protocol rather than proof for all dual-wavelength devices (Gesuete et al., 2026).

The honest answer is that choosing wavelengths for veterinary care should be based on the best available evidence from tissue optics, lab studies, human research, and veterinary studies. It is important to clearly state what kind of evidence supports each claim.

Technical reference

Technical details

How far does red or near-infrared light penetrate?

There is no simple, universal answer for how many millimeters red or near-infrared light can penetrate. Statements like “light penetrates 10 mm” or “850 nm reaches 40 mm” treat a changing photon distribution as if light stops at a fixed line. It does not. Light generally weakens as it travels. A small fraction may be measured beyond the depth where most of the original energy has already been absorbed or scattered away.

Penetration depends on:

• Wavelength
• Tissue type and thickness
• Blood content and oxygenation
• Water and fat content
• Melanin and skin pigment
• Fur or hair between the emitter and skin
• Contact versus non-contact use
• Beam size and geometry
• Angle and pressure
• Power and irradiance at the surface
• The definition of “penetration” used by the researcher

That last point matters more than it sounds. When researchers attach a precise meaning to “penetration depth,” the standard diffusion-theory definition is the depth at which irradiance falls to 1/e (about 37%) of its surface value (Jacques, 2013). Other studies instead report the depth at which only 10% or 1% remains, which yields a much larger number for the same light and tissue. The skin-model figures above, for example, use the 1% definition (Ash et al., 2017). A bare figure in millimeters that doesn’t say which of these it means can’t be compared to a different bare figure in millimeters.

A review of head-tissue studies found very large differences in measured red and near-infrared transmission across species, anatomical sites, and experimental methods (Salehpour et al., 2019). In an ex vivo study of several animal tissues and two commercial near-infrared laser systems, more than 90% of measured light energy was absorbed within the first 10 mm, with very little left after 15 to 20 mm under the tested conditions (Kaub and Schmitz, 2022). That result should not be converted into a rule for every device or patient. It shows why surface output and target-tissue dose are different quantities.

Beam size matters too, but only up to a point. In the same skin model, a 10 mm-wide beam delivered 73% to 88% of the light an infinitely wide beam delivered at 1 to 3 mm deep. Going wider than that added little (Ash et al., 2017).

Penetration and effectiveness are not the same thing

A comparative animal study illustrates the point. At the spinal cord injury site, more 830 nm photons arrived than 670 nm photons. Yet 670 nm performed better for some optic-nerve and retinal outcomes, and neither wavelength improved the tested traumatic brain injury or spinal cord injury outcomes under those protocols (Giacci et al., 2014).

Deeper transmission can be useful when the target is deep. Biological response, delivered dose, and protocol are still essential.

What is different about 904, 905, 980, and 1,064 nm?

904 and 905 nm

These wavelengths commonly appear in superpulsed laser systems. A very short pulse can have high peak power while the average power across the full treatment remains much lower. When reading research on these systems, consider pulse width, repetition rate, duty cycle, average power, beam area, and dose. Quoting the wavelength or peak power alone leaves out most of the protocol.

Water absorption is somewhat higher at 904 to 905 nm than at 850 nm, which alone would suggest slightly shallower penetration. But water is only one factor. Melanin and hemoglobin absorb less at 904 to 905 nm than at 850 nm, and scattering is somewhat lower too, so these effects roughly offset the water difference. The net penetration difference between these two wavelengths is modest and depends on the tissue, so neither is a clear win. The bigger practical issue is duty cycle. A superpulsed laser can deliver many watts of peak power at a duty cycle well under 1%, leaving the average power delivered to tissue far lower, often by several orders of magnitude, than a continuous-wave source at a similar wavelength. Whatever small penetration difference exists between these wavelengths is easily outweighed by that gap in total photons delivered.

For scoring, The WAIT Test™ follows WALT: a superpulsed 905 nm laser is judged against WALT’s superpulsed chart, which uses lower doses (3 J/cm² for a hip instead of 9), and its contribution is combined with the device’s other light by its share of the dose. A continuous 905 nm source doesn’t count.

980 nm

At 980 nm, you’re sitting right on water’s small absorption bump (see Figure 2), so water absorbs more light than it does around 810 nm. In cell research comparing 810 and 980 nm, the authors reported evidence of different dominant mechanisms: 810 nm effects were linked mainly to cytochrome c oxidase. In contrast, 980 nm effects were linked more strongly to temperature-sensitive calcium channels (Wang et al., 2017b).

This does not mean 980 nm is ineffective. It means you should not assume results from 810 or 850 nm apply to 980 nm without careful review. Pay close attention to heat and all device settings.

1,064 nm

Research has tested 1,064 nm for deeper transmission and for selected clinical uses. One ex vivo comparison found better transmission through the upper 10 mm of tested tissues at 1,064 nm than at 905 nm, while the advantage did not continue uniformly at greater depths (Kaub and Schmitz, 2023).

As with every less-common wavelength, the right question is not whether 1,064 nm can enter tissue. Ask whether the company can support its use for the claimed target with the same wavelength and a comparable delivery protocol.

Peak wavelength, tolerance, and spectral width

This is an advanced topic, and most people don’t need to worry about it. Most light sources sold today are well within what red light therapy needs.

For completeness, here’s what’s going on. When a light is sold as a certain wavelength, such as 660 nm, that number is the peak, where output is strongest. A good LED usually hits its rated peak within a few nanometers and spreads a little light to either side.

Light sources didn’t always look like this. Before LEDs took over, many light therapy devices used filtered lamps that spread their output across hundreds of nanometers. LED-based PBM only started appearing regularly in research after about 2001 (Heiskanen and Hamblin, 2018). Today’s LEDs put nearly all of their output inside the band they’re sold for.

Lasers have always been very narrow. A diode laser’s output is typically just 1 to 2 nm wide (Heiskanen and Hamblin, 2018). Modern LEDs are still a bit wider, but their output stays within the same red or near-infrared band. That’s why today’s LEDs are much more comparable to lasers than older light sources.

Laser versus LED: does the wavelength mean the same thing?

A nanometer is the same unit whether the source is a laser or an LED. The sources differ in other ways.

Lasers typically produce a narrow spectral band and can be highly directional and coherent. LEDs generally produce a more extensive spectral band and a more divergent beam. Device design affects beam shape, treatment area, irradiance, heating, and how light reaches the surface.

Neither lasers nor LEDs work just because of their name. A review comparing the two concluded that PBM doesn’t depend on lasers or coherence, though few head-to-head comparisons were well matched (Heiskanen and Hamblin, 2018). LEDs have shown results in controlled studies, like the 660 and 850 nm muscle trial mentioned earlier. You cannot assume results from laser studies apply to LED devices with different beam areas, outputs, distances, or doses, and the reverse is also true.

When comparing studies or products, match the full treatment as closely as possible (Jenkins and Carroll, 2011):

• Wavelength
• Light source
• Average power
• Irradiance at the treatment surface
• Beam or treatment area
• Continuous or pulsed delivery
• Treatment time
• Energy and energy density
• Contact, distance, and pressure
• Anatomical site
• Number and timing of sessions

Why wavelength is necessary but not sufficient

Choosing the right wavelength tells you the type of light you are using, but it does not tell you whether enough useful light is actually hitting the target area.

Plenty of devices proudly list standard 660 nm and 850 nm wavelengths, yet they can end up delivering vastly different results in practice.

How much light a device puts out, and how concentrated it is at the skin, are covered in the Average Power and Irradiance guides. Wavelength only sets what kind of light it is.