Technical guide

IPL Filters Explained: 530, 560, 590 and 640nm

IPL filters are the discs of coated glass that decide what your machine can actually treat. Each one blocks everything below its cut-off wavelength and passes the rest out toward the infrared, so a 560nm filter and a 640nm filter turn the same xenon lamp into two different machines. Which band does what, and why the glass in the box matters more than the console.

One lamp, several machines

Take the covers off an IPL platform and you find a xenon flashlamp. It doesn't emit a wavelength. It emits a spectrum, roughly 400 to 1200 nm in the description Lipp and colleagues use in the Journal of Drugs in Dermatology, all of it at once. Raw, that light is close to useless clinically, because it heats melanin, haemoglobin, water and collagen together and gives you no way to favour one.

Glass solves that. Our device manuals describe a dual filtration system specified to block wavelengths below 400 nm before light ever leaves the lamp housing, so ultraviolet output at the skin should be negligible in normal use. Treat that as a specification to verify, not a promise. Coatings have finite out-of-band rejection, and they age. Ask any supplier, us included, for the transmission specification of the housing filter and of the handpiece filter, and re-check it if a filter is cracked or sitting badly in its seat. Then comes the working piece, sitting over the output window: the cut-off filter. It blocks every wavelength below its number and passes the rest.

Read the label properly, though. A 560nm filter doesn't give you 560 nm light. It gives you 560 nm and everything above it, up to whatever the lamp and light guide still carry. Same lamp. Same supply. Same console. Different treatment.

Why the cut-off wavelength changes what you can reach

Two things shift as you climb the scale. Scattering drops, and melanin's grip loosens.

Our engineering archive states it flatly: scattering falls as wavelength rises, and roughly 600 to 1200 nm is the optical window into skin, because light there scatters little and few natural pigments compete for it. Below 800 nm, any wavelength preferentially heats melanin, and much of it scatters in the epidermis before reaching anything worth treating. The same archive defines penetration depth as the distance at which the beam has dropped to 37 per cent.

So the trade is fixed and you can't cheat it. Short cut-off means strong absorption by surface targets, shallow reach, and an epidermis fighting you for the same photons. Long cut-off means weaker absorption, deeper delivery, calmer epidermis.

Anderson and Parrish set the underlying rule in Science back in 1983: pick a wavelength your target absorbs better than its neighbours, keep the pulse short enough, and the target cooks while the surroundings don't. On a broadband source, the filter is how you make that choice. Our archive puts haemoglobin's absorption peaks at 418, 542 and 577 nm, which is precisely why the vascular cut-off sits where it does rather than somewhere rounder.

What each band is actually for

Cut-off	Typical job	What it costs you
430-1200 nm	An option in our manuals; the retained blue-green is what acne protocols want	Maximum melanin competition. Fair skin, low fluence, careful hands
530-1200 nm	Flat epidermal pigment, blotchy tone, general rejuvenation, fine telangiectasia	Surface melanin absorbs hard; reserved for lighter phototypes
560-1200 nm	The general photofacial band; an option in our manuals and the default on many platforms	Slightly less surface bite than 530, slightly less punch on pale freckling
585-1200 nm	Facial vessels and diffuse redness. Opens just past the 577 nm haemoglobin peak in our archive	Gives up green-band absorption in exchange for depth
610-1200 nm	Hair reduction, and darker phototypes generally. The hair filter in our manuals	Poor on superficial pigment. It's not a rejuvenation filter
640 nm and longer	Market convention on hair-first platforms; not part of our standard three	Deepest reach, lowest epidermal risk, and little useful pigment or vascular work left

Lipp and colleagues ran 515, 560 and 590 nm filters, plus a 900 to 1200 nm vascular filter, through a study of clinical endpoints rather than photographs. Right instinct. The endpoint decides the glass, not the other way round.

The filter set in the box is your treatment menu

Here's the most useful thing a first-time buyer can absorb. A platform's capability isn't the console. It's the glass that ships with it.

Our device manuals put three cut-offs into one changeable handpiece, 530-1200 nm, 585-1200 nm and 610-1200 nm, across an 8 by 40 mm output window, with 560-1200 nm and 430-1200 nm listed as extras. Buy the platform with one filter and you've bought a third of the machine. Pigment work with only the hair filter fitted is a slow, expensive disappointment for everybody in the room. Vascular work without the vascular band is worse, because you'll push fluence to compensate, and that's how people get hurt.

Our archive is equally direct about choosing by patient: longer cut-offs for deep targets, longer again on darker skin to protect the epidermis, shorter ones kept for lighter phototypes. A single machine can hold two of those positions only if you own two filters.

The Pmise MF-05 E-Light IPL and RF platform is our filtered flashlamp answer, and the treatment side lives on the skin rejuvenation solution page and the pigment and lentigo removal page. Choose filters against the two services you'll genuinely sell, not the twelve on the brochure.

One handling note from our manuals, since it costs nothing to follow: fit the handpiece filter only after the machine is switched on. Filters break at the worst moment.

Spares, and how to read a quotation that hides its filters

Filters are consumables that pretend not to be. They sit in a wet, warm, hand-held part that tired people put down on trolleys. Price the spare on the original order, because glass in a crate you're already paying freight on is a different number from glass airfreighted in a panic. Ask lead time before price. Buy the spare for the band you run most: if hair pays the lease, the 610 filter must never be missing. The sum worth doing isn't filter cost against filter cost. It's filter cost against a day with the machine idle.

Most quotations you'll receive describe a chassis. Cooling, screen size, a photo. Filters appear as the word "filters", plural, unnumbered. That isn't an oversight, it's margin. Send it back with these.

Which cut-offs are included, stated in nanometres, and how many pieces of each?
Which handpiece does each filter fit? A filter for a window you didn't buy is decoration.
What does a spare cost delivered, and how long does it take to arrive?
Are parameter tables supplied per filter and per Fitzpatrick type, or is one table meant to cover everything?
What optical density are the supplied goggles rated to, across the band each filter passes?
What is the regulatory status in your market? In the United States these systems are Class II devices handled under 21 CFR 878.4810 and cleared through the 510(k) route.
Who replaces a cracked filter, and does doing it yourself void anything?

A supplier who answers all seven in writing is worth buying from. One who says "standard configuration" has told you plenty.

Skin type, risk, and the limits of glass

A longer cut-off lowers epidermal risk. It doesn't remove it. The StatPearls chapter on IPL therapy keeps shorter wavelengths for fair skin and treats eye protection as non-negotiable for patient and operator alike. Nothing fires inside the bony orbital rim, whatever filter is fitted.

Post-inflammatory hyperpigmentation is the complication you'll meet most. It clusters at higher fluences and on darker phototypes, it fades slowly, and no filter choice makes it impossible. Test patch, wait, then treat conservatively. Melasma deserves particular caution, since heat itself can drive it. Anything undiagnosed, changing, asymmetric or bleeding goes to a clinician first and possibly to biopsy; treating it destroys the evidence. Pigment clearance is gradual and runs over a course of sessions, with our clinical guide spacing photorejuvenation by body region. This page is educational material for equipment buyers, not medical advice.

Questions buyers ask us

Is a 560nm filter better than a 530nm filter for a photofacial?

Not better. Different. The 530 band keeps more of the green that surface melanin and haemoglobin absorb hardest, so it hits pale freckling and fine vessels harder, and competes harder with the epidermis. The 560 band trades some of that bite for margin, which matters on type III and IV skin. Clinics with a busy photofacial column usually end up owning both.

Do I need a 640nm filter if I already have 610?

Usually no. Both sit in the region where melanin competition is low and depth is good, and our own hair band is 610-1200 nm. A 640nm filter earns its place mainly on hair-first platforms working darker phototypes at higher fluences, where every extra nanometre of margin at the surface is welcome. If hair on type V skin is your core business, an 808 nm diode is the better instrument anyway.

Can I run everything from one filter?

You can run one thing well from one filter. The rest you'll do badly, and you'll be tempted to fix it with energy, which is the wrong lever. Our parameter sheets set fluence, pulse count and sub-pulse width by indication and skin type, and they assume the right band is fitted. Change the glass, not the joules.

How do I check the filter is doing anything?

Compare treatment response, not the colour of the glass. Swap filters at fixed settings on a test patch and the response should change visibly and predictably; if it doesn't, suspect the filter seating or a lamp near the end of its life. Ask for the transmission specification in writing too. If nobody can produce one, you're buying coloured glass on trust.

Evidence & further reading

Educational material for equipment selection and operator training. It is not medical advice, a treatment protocol or a promise of clinical outcome.

Request a quotation More guides