Why Diode Lasers Struggle Where CO₂ Lasers Cut Easily
By Brandon Cullum Last updated Aug 5, 2026
My 50 W CO₂ laser cuts 6 mm birch plywood in one pass. Clean edges. The 20 W diode laser sold as its equal took fifteen passes on the same wood, and the edges came out charred.
I wasted $600 in test materials finding out why.
I review laser cutters for a living, so when diode companies started claiming their 20 W units matched a CO₂ on wood (same performance, half the price) I bought three models to test. By test sheet number 30 I had tried every combination of power, speed and focus I could think of. Same result every time. The diode worked, technically. But it took 15 passes where the CO₂ took one.
The spec sheets compared watts. Watts were not the thing that was different.
That is when I stopped trusting spec sheets and started asking why.
The short answer, before the long one: light wood is only 10–30% chromophores, the dark compounds that can absorb blue light, so 70–90% of a blue diode's beam lands on molecules with no way to use it. A CO₂ laser at 10,600 nm matches the C-O bonds that make up almost all of the wood, so nearly every photon does work. Wavelength decides that, not watts, and there is no setting for wavelength.
Every answer leads to another why. It is like that thing with kids, where you keep asking until you hit quantum mechanics and even physicists shrug. This guide is that rabbit hole, in three parts. Stop reading when you have gone deep enough.
- This page. What each laser type does in a shop, and what a beam is actually made of. Answers most people's question completely.
- Wavelength and materials is why wood absorbs 10.6 μm and reflects 450 nm, down to the bonds doing the absorbing.
- The laser physics deep dive is how each laser type generates its light in the first place, and the quantum mechanics underneath all of it.
What you can see: the four laser types
Start with what you already know from using these machines. No theory yet, just the patterns you have seen in your own shop. I have tested dozens of lasers over the years and there are four main types you will run into. Each one has a personality: things it loves cutting, and things that make it look stupid.
| Laser | Dominates | Struggles with | What you actually see |
|---|---|---|---|
| CO₂10,600 nm · 40–150 W | Wood, acrylic, leather, paper, cardboard, fabric. Clean cuts at speed, glass-like polished edges on acrylic, sealed edges on fabric. | Metals (the beam bounces off), glass (passes right through), PVC (melts instead of cutting), anything transparent. | Point it at birch plywood: clean cut, slight brown edge, smells like a campfire. Point it at aluminum: nothing happens, and the beam reflects off like a bathroom mirror. |
| Fiber1,064 nm · 20–60 W | Steel and stainless (marks, engraves, cuts thin sheet), anodized aluminum, brass and copper, coated metals, ABS plastic. | Wood (chars the surface, no clean cut), acrylic (melts and deforms), leather and paper, anything transparent. | Hit stainless steel with it and you get a permanent white mark in milliseconds. Try the same settings on wood and the surface turns black and crusty with no actual cut. It is the exact opposite of CO₂. |
| Blue diode450 nm · 5–40 W | Dark wood engraving, leather, painted and coated surfaces, cardboard and paper, dark plastics. | Light wood (many passes, rough edges), metals (mostly reflects, marks only with a coating), thick material, transparent or reflective stock. Slow compared to CO₂ for cutting. | My Atomstack X20 at 20 W engraves walnut beautifully, and the contrast is gorgeous. Try cutting 1/4" birch plywood with it and you are making 15–20 passes for rough edges. |
| UV355 nm · 3–20 W | Almost anything, including glass and ceramics. Marks metal without heat damage, marks PCBs without damaging components, ultra-precise detail on wood and plastic. | Price ($3,000–15,000+), slower than CO₂ for cutting, low wattage means thin material only, and alignment is critical. | I have seen UV lasers engrave glass wine bottles with detail that would make CO₂ and fiber lasers cry. They break bonds directly instead of heating the material, which is why it is called cold ablation. At $8K for an entry-level unit, most of us are window shopping. |
Wavelengths and typical wattages are the standard figures for each type. The "what you actually see" column is from my own testing.
The difference shows up in the cut itself, not just the clock. Run the same 6 mm birch through my 50 W CO₂ and you get a U-shaped kerf 0.3 mm wide with a light brown edge. Run it through the 20 W diode and you get a V-shaped kerf 0.5 mm wide, black all the way down. Two different shapes of hole, which is the first sign that these are not the same process at different speeds.
Blue Diode Laser
15 passes required • Surface absorption
Surface-concentrated: Each pass chars the surface, creating chromophores that absorb better on subsequent passes. Builds V-groove from top down.
CO₂ Laser
1 pass • Resonant absorption
Uniform depth: Resonant frequency matches C-O bonds throughout material. Energy absorbed evenly from top to bottom. Clean parallel walls.
Look up your exact material The full compatibility matrix lives in the material guide, one row per material with the shop reality and the physics reason for each laser type.
See which marks your laser can actually make Pick your machine and material and get a computed verdict per mark type, including an honest "we have not modelled this" where the answer is not known.
But why? Why does wood see infrared but ignore visible blue? Why does metal laugh at CO₂ but respect fiber? To answer that you first have to know what a laser beam actually is.
What a laser beam is: energy per photon vs how many
Sixty watts does not tell the whole story, because it says nothing about how that power gets delivered. Two beams can carry identical wattage and do completely different things to the same board.
Wavelength sets the energy per photon
Light is not continuous. It arrives in discrete packets called photons, and your beam is a stream of trillions of them hitting the material every second.
Each photon carries a specific amount of energy, and wavelength is what determines how much. Shorter wavelength, more energy per photon. This is locked in by physics. There is no setting for it.
The math, which you do not need to memorise
E = hc/λ
h is Planck's constant, c the speed of light, λ the wavelength. All that matters here: shorter wavelength means more energy per photon.
UV photons carry 30× more energy than CO₂ photons. So UV must cut better, right? Not necessarily, because power is not energy per photon. It is total energy delivered per second.
Power is total energy per second
When you see 60 W on a laser, that is 60 joules of energy per second. Always. Does not matter if it is CO₂, fiber, diode or UV.
But you can deliver those 60 joules two ways. Send trillions of low-energy photons so each one contributes a little and the count does the work. Or send far fewer high-energy photons that each pack a bigger punch. The math always balances: energy per photon × photons per second = watts.
60 W CO₂ laser
Energy per photon: 0.117 eV (weak)
Photons per second: 3,201 quintillion
Strategy: overwhelm with quantity
60 W UV laser
Energy per photon: 3.5 eV (strong)
Photons per second: 107 quintillion
Strategy: precision with power
I ran the numbers on my own two machines. The 50 W CO₂ at 10,600 nm puts out 2.7 × 10²¹ photons per second, each carrying 0.117 eV. The 20 W diode at 450 nm puts out 4.5 × 10¹⁹ photons per second, each carrying 2.75 eV. So the CO₂ delivers about 60 times more photons, and each of those photons is 23 times weaker.
The spec sheets compared the watts and stopped: 20 W is 40% of 50 W, expect roughly 40% of the cutting speed. The arithmetic is fine. What it leaves out is whether wood can do anything with 2.75 eV photons in the first place.
The key insight
At the same wattage, CO₂ fires 30× more photons per second than UV, and each one is 30× weaker. Identical total energy, delivered completely differently.
Interactive Photon Energy Simulator
Adjust the wavelength and power to see how photon energy and quantity change. The math always balances: Energy per photon × Photons per second = Power (Watts)
Quick Select:
Energy Per Photon
0.117
electron volts (eV)
Photons Per Second
3,200
quintillion photons/second
Total Power Output
60W
always constant
Why this matters for cutting
Materials are picky about which photon energies they can absorb. Match it and more photons means faster cutting. That is why CO₂ dominates on wood: wood molecules absorb 0.117 eV photons efficiently, and having 3,201 quintillion of them per second means the cut happens fast.
Miss it, like a blue diode on light wood, and most of those photons bounce off or pass straight through. You are firing 600 trillion photons per second and only 10–30% of them do any work. Hence the multiple passes.
Low energy per photon is not a weakness, it is a different strategy. Like shooting 1,000 BBs instead of 30 bullets. Same total energy. The BBs work great if the target absorbs BBs, and the bullets work if you need to punch through directly. Wrong tool, wasted energy.
What to do with this
Watts tell you how much energy per second. Wavelength tells you whether the material can use it. You need both numbers, and only one of them is on the box.
This is why a 40 W CO₂ cuts wood better than a 40 W diode even though diode photons carry 23× more energy each. It is not photon strength. It is whether the material can absorb what you are sending.
They listed watts because watts sell. They didn't list wavelength because then you'd ask questions.
But why are materials picky? Why does wood absorb 0.117 eV photons efficiently but not 2.75 eV photons? What is inside wood that decides which photon energies work?
Wavelength and materials picks it up there, from the chemistry through to the moment a bond actually breaks.
