Optical principles and energy control of laser marking machines

• 1
  What happens between the laser source and the part
• 2
  Fluence, spot size and the energy control of laser marks
• 3
  How aluminium, stainless and titanium respond
• 4
  Where laser marking stops being safe
• 5
  Mark type against fluence, pulse width and typical use

A marking head is a short optical train. The source emits a beam, a beam expander widens it, two galvanometer mirrors steer it, and an f-theta lens focuses it onto a flat field. Each element changes the energy density that finally lands on the metal.

The f-theta lens does two jobs at once. It sets the focal length, and it keeps the focus on a flat plane across the whole marking field. A 100 mm lens gives a small spot and a narrow field. A 254 mm lens covers more area but spreads the same energy over a larger spot.

Galvo mirrors decide where the beam goes, but they also decide how long it stays there. Scan speed and mirror settling time set the dwell per point. Run the axes too fast and the pulses overlap less, so the mark fades. Run them too slow and the same area absorbs more pulses than it can shed as heat.

Beam quality matters more than headline wattage. A source rated 20 W with a tight, near-Gaussian beam can cut a brighter mark on stainless than a 30 W source with a wide, uneven profile. Ask for the beam parameter product before you compare two machines.

Alignment drifts with heat and vibration. A beam that walks off the lens center will mark one corner of the field darker than the other. On a 200 × 200 mm field, a 0.2 mm drift is enough to show in a contrast check under magnification.

What actually changes the metal is fluence: pulse energy divided by spot area, in J/cm². A 1 mJ pulse focused to a 30 μm spot delivers roughly 140 J/cm². The same pulse at a 100 μm spot delivers about 13 J/cm². Spot size, not power setting, dominates the result.

Every material has a threshold band. Below the lower edge, the surface only heats and oxidizes slightly; the mark washes out after cleaning. Above the upper edge, the metal melts, and you get a raised rim, micro-cracks, or a color shift that will not repeat from batch to batch.

Pulse width sets how fast that energy is delivered. Nanosecond pulses dump heat faster than the lattice can conduct it away, so the mark stays shallow. Picosecond and femtosecond pulses strip material with almost no heat-affected zone, which matters on thin-walled parts and on medical components.

Energy control of laser output is not one knob. Average power, pulse frequency, pulse width and scan speed interact. Doubling frequency at fixed average power halves the energy per pulse. If you also double scan speed, the mark can look unchanged while the heat load per pass drops by half.

Calibrate with a power meter at the work plane, not at the source. Lens transmission, mirror reflectivity and focal offset all eat energy. A 5 percent loss at the source can become a 20 percent loss at the part after a dirty f-theta lens.

Aluminium reflects well at 1064 nm and conducts heat quickly. Marks on 6061 and 7075 usually come from oxidation and controlled melting, so they need higher fluence than steel and tighter pulse control. Anodized aluminium is the easy case: the coating absorbs and the mark lands in the oxide layer.

Stainless 304 and 316 mark well with fiber sources. Dark, high-contrast marks come from oxide growth at moderate fluence, while deep engraving needs repeated passes at higher energy. Pass count changes depth roughly linearly at first, then flattens as debris and plasma shield the surface.

Titanium and Ti-6Al-4V are color-sensitive. A narrow fluence window produces stable blue, gold or grey oxide without cutting into the surface. Step outside that window and the color shifts with small changes in focus or temperature. For medical parts this is a process-control problem, not a setting problem.

Copper and brass absorb poorly at 1064 nm. Green or UV sources at 532 nm and 355 nm absorb far better and give cleaner marks with less subsurface heating. If your part is a copper busbar or a brass fitting, source wavelength matters more than power.

Plastics vary too much to treat as one group. POM and PA mark by carbonization and need low fluence. PC and PMMA can craze or bubble. PEEK tolerates higher energy. Always test on the exact resin and color, because pigments change absorption sharply.

The heat-affected zone is the limit that matters on precision parts. Nanosecond marking on 304 stainless can leave a 20–50 μm recast layer with altered grain structure. On a sealing face or a bearing bore that is ±0.005 mm, that layer is a functional defect even when the mark looks good.

Thin walls are the second boundary. A 0.5 mm aluminium wall under a 30 W beam at low scan speed will warp. The fix is not always lower power. Shorter pulse width, higher scan speed and multiple light passes often give the same contrast with less bulk heating.

Coatings and plated surfaces behave differently from the base metal. Anodized layers, electroless nickel and black oxide absorb and ablate at their own thresholds. Marking through an anodized layer to the aluminium below creates a galvanic path and a corrosion site.

Fatigue-critical parts deserve a separate review. Any melting at the surface starts a crack initiation site. On rotating or cyclically loaded components, engraving should sit away from the highest-stress region, or the mark should be a low-fluence color change rather than a cut.

Position tolerance is its own boundary. Laser marking follows the fixture, not the machined datum. If a mark must sit within 0.1 mm of a bore, the fixture needs the same datum scheme as the machining setup, not a hand-placed stop.