Skip to content
MFG Processes

Laser Plastic Welding

Laser plastic welding passes a beam through a transmissive part into an absorbing one, melting only the interface between them.

Part
Joining
Revised
2026-08-11

At a glance

Family
Thermal
Typical tolerances
Clamped joint gap ≤ 0.004 in (0.1 mm) required across the seam; quasi-simultaneous collapse controlled to about ±0.001 in (±0.025 mm)
Typical volumes
1,000 to 5,000,000+ per year; contour welding is viable from prototype quantities
Lead time
Seconds to a few tens of seconds per part for contour welding, under 5 s for simultaneous; 4–10 weeks for transparent clamping tooling and part-specific optics
Materials
Plastic

What it is

Laser plastic welding passes a near-infrared beam — 808–1,064 nm from a diode or fiber source — straight through a transmissive upper part and into an absorbing lower part, where it converts to heat. Only the interface melts. The upper part stays cold enough to hold a mirror finish, a printed graphic or a bonded film through the weld.

That selectivity is the whole value: no vibration, no particulate, no flash, no heat anywhere except a seam 0.02–0.08 in (0.5–2 mm) wide. It is the standard route for microfluidic cartridges, sensor housings, medical devices and electronics enclosures with parts that other welding processes would shake apart.

The controlling constraint is contact. The clamped gap between the two parts must stay under about 0.004 in (0.1 mm) across the whole seam, because the beam heats only what it reaches — there is no melt pool to bridge with.

How it works

  1. Material pairing. One part must transmit the beam and the other must absorb it. The absorber is usually pigmented with 0.1–0.5% carbon black, or with a near-IR absorbing additive that leaves the part visually clear or color-matched.
  2. Clamp. A transparent fixture — glass or acrylic — presses the parts together through the beam path. Contact pressure closes the gap and, once melting starts, pushes the parts together as the joint collapses.
  3. Deliver the beam. 10–500 W of diode or fiber laser power passes through the upper part with little absorption, then deposits energy in the first few thousandths of an inch of the absorber.
  4. Melt and conduct. The absorber melts, heat conducts back into the transmissive part, and a melt layer forms on both sides of the interface. Collapse of 0.002–0.008 in (0.05–0.2 mm) confirms fusion.
  5. Cool under clamp. Pressure is held while the seam solidifies.

The four scanning strategies

Contour welding traces the joint line once with a moving spot at 0.4–20 in/s (10–500 mm/s). Simple, flexible, and suits long or one-off seams; no collapse is possible because the seam behind the spot has already frozen.

Quasi-simultaneous welding uses galvanometer mirrors to sweep the entire joint line many times per second, so the whole seam is molten at once. That allows genuine collapse — the parts close under clamp pressure, taking up molded-part variation — and gives the best hermetic sealing. It is the workhorse for small and medium parts.

Simultaneous welding illuminates the whole joint at once through a fixed array of fiber-coupled diodes or a mask. Cycle times drop under a couple of seconds, but the optics are part-specific tooling.

Mask welding projects the beam through a mask onto flat parts, producing seams down to a few hundredths of an inch — used for microfluidic and lab-on-chip devices where channel walls must be sharply defined.

Design guidelines

Close the gap — this is the design problem

Laser welding has no capacity to bridge a gap: any place the parts do not touch is a place that will not weld. Design the clamped gap to stay under 0.004 in (0.1 mm) along the entire seam. That means supporting the joint from both sides of the beam path, avoiding long unsupported spans, and designing the parts so clamp force closes them rather than bowing them.

Give the beam a clean path

Everything the beam passes through attenuates it. Keep the transmissive wall as thin as function allows, keep it free of ribs, bosses, gate marks, sink and text within the beam footprint, and mold it with a smooth surface — a textured or frosted surface scatters the beam and widens the seam unpredictably. Approach the joint perpendicular where possible; steep angles cost transmission and smear the focus.

Design the joint for collapse

With quasi-simultaneous or simultaneous welding, specify 0.004–0.010 in (0.1–0.25 mm) of collapse into the design, and set the finished assembly height off a positive stop or the machine's collapse control. Collapse is what absorbs molding variation and squeezes out any residual gap, and it is the single biggest reliability lever available.

Material transmission

Material at the jointRoleNotes
Natural / unpigmented PP, PE, PA, PC, ABS, PMMATransmitterGood transmission at 808–1,064 nm
Carbon-black pigmented resin (0.1–0.5%)AbsorberThe standard, cheapest absorber
Near-IR absorbing additiveAbsorberLets the absorber stay clear or color-matched to the transmitter
TiO2-pigmented white resinPoor transmitterScatters the beam heavily; white-on-white is difficult
30%+ glass filled gradePoor transmitterFibers scatter; transmission falls sharply with fill and wall thickness
Carbon black in the upper partWill not workThe beam never reaches the joint
Dissimilar resin pairCase by caseNeeds chemical compatibility and close melt temperature, same as any plastic weld

Wall thickness, gate placement and rib rules that govern the beam path are on the injection molding design guidelines chart.

Inspection

The transmissive part is an inspection window: the weld seam can be seen and measured optically through it, which no other plastic welding process allows. Beyond that, collapse distance is logged per cycle and is the primary process signal for quasi-simultaneous welding; pressure-decay or helium leak testing verifies hermetic parts; burst and pull testing on lot samples verifies strength; and infrared thermography during welding detects cold spots caused by local gaps.

FeatureRecommendedLimitWhy
Clamped joint gap0.001 in (0.025 mm)0.004 in (0.1 mm)The beam cannot bridge what it does not reach
Transmissive wall in the beam pathAs thin as function allowsEvery millimeter of path attenuates
Surface finish in the beam pathSmooth, untexturedTexture scatters the beam and widens the seam
Designed collapse (quasi-simultaneous)0.006 in (0.15 mm)0.004–0.010 in (0.1–0.25 mm)Absorbs molding variation and closes residual gaps
Weld seam width0.04 in (1 mm)0.02–0.08 in (0.5–2 mm)Sets joint strength; wider needs more power or slower travel
Absorber pigment0.1–0.5% carbon blackEnough to absorb in the first few thousandths of an inch
Clamp supportBoth sides of the seamUnsupported spans bow open under clamp force

Cost drivers

Capital is the main cost. A laser source, scanner or fiber array, transparent clamping tooling and a Class 1 enclosure represent a larger investment than an ultrasonic or hot plate welder, and simultaneous welding adds part-specific optics on top. Against that, the cycle is short, there are no consumables, and the process has essentially no reject mechanism once the gap is under control.

The economics work when the alternative fails on quality rather than cost: parts that cannot tolerate vibration, assemblies where particulate is a contamination issue, seams that must be narrow and clean, or products where the visible upper surface must survive the joining operation. Comparing laser welding to ultrasonic welding on price per part alone will always favor ultrasonics — the comparison only makes sense against scrap and warranty.

Tooling cost is dominated by the clamping fixture, which must be optically transparent, dimensionally accurate and stiff enough to close the gap everywhere along the seam.

  1. Use contour welding for prototypes and low volume — no part-specific optics, just a program change.
  2. Design one absorber and one transmitter across a product family so the same material pair and parameters carry over.
  3. Spend on the clamping fixture, not the laser. Gap control is what determines yield.
  4. Use an IR-absorbing additive instead of carbon black only where the part must be clear or color-matched; it costs more.
  5. Check that quasi-simultaneous collapse is available before tightening molded part tolerances — collapse is cheaper than tooling precision.

Questions

5 questions
How does laser welding melt only the joint and not the top part?

The upper part is transmissive at the laser wavelength (808–1,064 nm) and passes the beam with little absorption; the lower part contains carbon black or a near-IR absorbing additive that converts the beam to heat in the first few thousandths of an inch. Melting therefore starts exactly at the interface and conducts back into the upper part from there.

How tight does the fit-up need to be for laser plastic welding?

The clamped gap must stay under about 0.004 in (0.1 mm) across the entire seam. There is no melt pool to bridge a gap — anywhere the parts do not touch simply does not weld. Quasi-simultaneous welding relaxes this somewhat because the whole seam is molten at once and clamp pressure can collapse the joint by 0.004–0.010 in (0.1–0.25 mm).

Can you laser weld two clear plastic parts?

Yes, using a near-IR absorbing additive applied at the interface or compounded into the lower part instead of carbon black. The additive absorbs at the laser wavelength but is nearly invisible, so both parts stay clear or color-matched. It costs more than carbon black, which is why carbon black remains the default where appearance permits.

Why would I choose laser welding over ultrasonic welding?

When the part cannot tolerate 20 kHz vibration — populated PCBs, sensors, crystals, filled microfluidic channels — or when particulate generation is a contamination problem, or when the visible surface must survive the joining step. On price per part alone ultrasonic welding usually wins; laser welding justifies itself against scrap, contamination and warranty.

Why are white plastic parts hard to laser weld?

Titanium dioxide, the pigment that makes plastic white, scatters near-infrared light strongly, so the beam disperses inside the transmissive part instead of reaching the joint. Glass fill above about 30% causes the same problem. Both can sometimes be worked around with thinner walls, higher power or a different pigment package, but they should be flagged at design review rather than discovered at process validation.