Conduction Or Keyhole: What A Laser Actually Does To The Metal
2026/11/14
Conduction or Keyhole: What a Laser Actually Does to the Metal
Two seams can come off the same laser on the same day and be produced by two different physical processes. At low power density the beam heats the surface and heat conduction does the rest: the melt pool is wide and shallow, and the seam is a cosmetic bead. Above a threshold the beam stops conducting and starts drilling. Metal vaporises, a channel opens in the melt, and the beam travels down inside it.
The photograph below is the second kind. Everything a buyer argues about with a laser supplier — penetration, speed, spatter, whether 1.2 kW is enough — comes back to which of the two modes the sample was made in.
A platform laser welder on our floor. The head is the same in both modes described here; what changes is how tightly the beam is focused and how fast it travels.
A Century of Welding in a Millisecond
Bare steel reflects most of a 1 µm beam at room temperature. The figure usually quoted is around two thirds, and it is why the first instants of a laser weld are the difficult ones: the metal has to be hot enough to absorb before it can be welded. Once a melt pool exists, absorption climbs sharply and the process becomes self-sustaining.
This single fact explains a lot of workshop behaviour that otherwise looks like superstition. Why welders tack a start point before running a seam. Why a polished or freshly machined surface is harder to start on than a mill-finish one. Why a thin film of oil or oxide often helps rather than hurts, up to the point where it turns into porosity. Why the same parameter set behaves differently on the second pass than on the first.
The Threshold Is Power Density, Not Power
Keyhole formation is a question of watts per square centimetre, not watts. A 1200 W source focused into a 0.2 mm spot delivers a power density that a 3000 W source spread over a 1 mm spot cannot reach. This is why the number on the cabinet is such a poor predictor of penetration, and why two machines with the same wattage can differ by a factor of two in what they will weld.
It also means the same machine can be run in either mode at will. Defocus the beam and the same 3000 W becomes a wide, gentle heat source. Focus it hard and the identical machine drills. On our platform machines the difference between a cosmetic seam and a structural one is often a few millimetres of focus position and a different travel speed, not a different machine.
Mode |
Power density |
Shape of the weld |
Fails by |
Conduction |
Below the threshold |
Wide, shallow, smooth |
Cold lap on a thick joint |
Keyhole |
Above the threshold |
Narrow, deep, nail-headed |
Porosity and root defects |
Unstable keyhole |
Right at the threshold |
Intermittent, humped |
Spatter, undercut, depth that wanders |
What the Cross-Section Shows
The two modes leave different shapes behind, and the shape is the only honest evidence of which one was used. A conduction weld is a shallow lens: wider than it is deep, with a gentle transition into the parent metal. A keyhole weld is a spike — narrow at the top, deep, and often finishing in a small bulb at the root where the beam finally ran out of energy to keep drilling. Welders call that shape a nail head, and a section showing a nail head and a visible root is the only proof that the penetration on the data sheet was achieved rather than implied.
This is why a photograph of the top of a bead proves very little. A conduction seam and a keyhole seam can look identical from above: same width, same colour, same ripples. The difference is four millimetres of metal underneath. Ask for the section, and ask what it was etched with, because an unetched section hides the boundary that matters.
When Both Modes Are Needed at Once
Many real joints are thick-to-thin: a 3 mm bracket onto a 1.2 mm panel, a boss onto a 0.8 mm cover. Drilling into the thick side is what gives the joint its strength, but the same energy density that opens the keyhole will cut through the thin side long before the thick side has melted. The usual answers are to place the spot so the thin sheet sits at the outside edge of the pool, to defocus slightly and accept a wider, shallower root, or to wobble the beam in a small orbit so the energy is spread rather than concentrated at a point.
All three trade penetration for tolerance, and all three have to be proven on the actual thickness pair rather than on a plate of the thicker material. That last point is the most common mistake we see in sample requests: a customer sends a piece of the 3 mm bracket, we return a beautiful deep weld in 3 mm, and nothing has been learned about the 1.2 mm panel the bracket is attached to.
What Keyhole Buys, and What It Charges for It
The advantages are real: deep penetration in a single pass, a narrow heat-affected zone, and distortion that is an order of magnitude below arc welding because the total heat input is so much lower. A 4 mm butt seam welded in one pass, as in the photograph, is not remarkable for a laser; it is routine.
The costs are equally real, and they are all stability problems. A keyhole is a hole held open by vapour pressure, and anything that disturbs the balance closes it: a gap that widens, a tack in the path, a change in travel speed, a dirty cover glass shifting the focus. When the channel collapses, the metal that was inside it comes out as spatter; when it collapses and does not refill, the result is a root defect that no amount of surface inspection will find.
Conduction Mode Is Not the Failure Mode
It is tempting to treat conduction mode as the compromise you accept when the machine is too small. That is the wrong way round. For visible seams on cabinets, enclosures and domestic products, a wide shallow bead is exactly what is wanted: it is smooth, it takes polishing, and it does not leave the sunken line that a keyhole seam leaves when it is dressed back.
The two requests — "I need 4 mm of penetration" and "I need a seam I can polish out" — cannot be met with the same settings, and a buyer who asks for both in one sentence has not yet decided what the joint is for. That is worth settling before the machine is specified, not after.
Three Ways It Goes Wrong on the Floor
The first is dragging the head like a TIG torch. A laser is fast; travel speeds of one to three metres a minute are normal, and an operator who moves at hand speed overheats the part, opens the keyhole too wide and gets either humping or burn-through. The second is focus drift: a spattered cover glass or a dirty nozzle changes the focus position, and the process quietly migrates from keyhole to conduction without anybody touching a setting. The third is gas. Too little and the plume ionises and blocks the beam; too much and, on thin sheet, the flow blows the keyhole open and the seam becomes a slot.
A butt seam in 4 mm stainless, welded in one pass. Narrow, deep, almost no distortion on either side of it: this is what keyhole mode looks like when it is stable.
What to Ask When You Are Shown a Sample
- Which mode was this weld made in, and what was the focus position and spot size?
- Was the penetration measured on a cross-section, or inferred from the top of the bead?
- What gas, at what flow, and does the machine monitor it?
- How is the focus position maintained over a shift — cover glass changes, nozzle cleaning, or nothing at all?
A supplier who can answer those four questions is describing a process. One who answers only the first is describing a demonstration. The difference matters most in the second year of ownership, when the setting that was proven in the showroom has to survive a night shift, a spattered lens and an operator who was taught by somebody else.
PDKJ continues to provide customized laser welding and resistance welding solutions based on different materials, workpiece structures, welding requirements, and production applications.
Related Products
The finished-goods area in Dongguan, where every machine is run on real parts before it is packed.
Contact PDKJ
Send us your part drawing or a sample, and our engineers will run a welding feasibility trial on your own material and recommend the equipment configured for your application.
WhatsApp: +86 136 1279 3248
Email: jay@gd-pw.com
Website: https://www.spotlaserwelding.com/contactus.html
Address: 1-2F, Building 3, Qichen Industrial Park, No. 26 Luxi 1st Road, Liaobu Town, Dongguan City, Guangdong Province, China