Copper laser melting density: why CuCrZr is hard to fuse

• 1
  Why copper laser melting density is hard to reach
• 2
  CuCrZr widens the copper laser melting density window
• 3
  Parameters that move density, and the ones that lie
• 4
  Which copper features print well and which do not
• 5
  Printing the channel and machining the rest
• 6
  Copper laser melting density vs CNC for copper parts

Most metal laser powder bed machines run a fiber laser near 1,060 nm. Polished copper reflects about 95% of that wavelength at room temperature. Only a few percent of the beam energy enters the melt pool, so the first few layers barely fuse. The same beam that keys into steel leaves copper powder loose and partly sintered.

The second problem is heat. Copper conducts heat roughly 20 times better than steel, so energy that does enter the melt pool spreads sideways and downward within microseconds. A 100 μm spot loses heat faster than the laser can add it. Operators respond with higher power and slower scan speed, which raises spatter and keyhole porosity.

Oxide makes it worse. Copper forms CuO and Cu2O films at elevated temperature, and both absorb less near-infrared light than clean metal. Once an oxide skin forms on the powder or the previous layer, the next pass couples even less energy. Density then drops in bands rather than uniformly.

Green and blue lasers near 515 nm and 450 nm absorb 40% or more on copper, which is why they are used for pure copper. Near-infrared machines still dominate the market, so the practical answer for most shops is an alloy rather than a light source change.

• 1
  ReflectivityAbout 95% of 1,060 nm light is reflected by solid copper.
• 2
  Thermal diffusivityHeat leaves the melt pool before the next scan track arrives.
• 3
  Oxide filmCuO and Cu2O lower absorption further during the build.

CuCrZr, roughly Cu with 0.5–1.5% chromium and 0.05–0.15% zirconium, absorbs more near-infrared light than pure copper and holds its shape better at temperature. The chromium also ties up oxygen, so the melt pool stays cleaner between passes. That is why most near-infrared copper work in industry is done on CuCrZr, not on C101 or C110.

The trade-off is conductivity. Pure copper sits around 58 MS/m. Printed CuCrZr typically lands in the 20–30 MS/m range as built. A solution anneal and age at roughly 470–500 °C for 1–3 hours can push it toward 40–50 MS/m, but the part must survive that cycle without warping. Thin walls and long unsupported spans usually do not.

Chromium and zirconium also evaporate at the melt pool surface. On a small part the loss is minor. On a tall build with a large molten area, the alloy drifts and the top of the part behaves differently from the bottom. Powder reuse across many cycles makes the drift worse unless you blend virgin powder back in.

For tooling inserts and cold plates, printed CuCrZr is often the right answer. For busbars, RF cavities and induction coils where conductivity is the whole point, a machined C101 or C110 part is the safer route.

• 1
  Printed CuCrZr20–30 MS/m as built, up to about 40–50 MS/m after aging.
• 2
  Wrought C101 and C110Around 58 MS/m, machined from bar or plate.
• 3
  Evaporation riskCr and Zr deplete on large molten areas and in reused powder.

Volumetric energy density is the usual starting number: laser power divided by scan speed, layer thickness and hatch spacing. It is a useful bookkeeping figure, not a prediction. Two builds with the same number can differ by 1% in density if one uses a 60 μm spot and the other a 110 μm spot, because the melt pool shape changes.

On near-infrared machines running CuCrZr, the working region is usually 350–500 W, 600–900 mm/s scan speed, 30–40 μm layer thickness and 0.08–0.11 mm hatch. Going below about 600 mm/s tends to keyhole; going above about 900 mm/s leaves lack-of-fusion defects at track overlaps. Remelt or contour passes on the skin raise surface density but do not fix interior porosity.

Atmosphere matters more than most people expect. Below about 1,000 ppm oxygen in the build chamber, spatter and oxide inclusions drop sharply. Argon is standard. Some shops run nitrogen for cost reasons, but nitrogen does not help copper the way it helps stainless.

Post-build hot isostatic pressing at roughly 900–1,000 °C and 100 MPa closes internal pores and lifts density above 99%. It also softens the part and changes dimensions slightly, so plan the HIP step before you quote the tolerance.

• 1
  Energy densityA bookkeeping figure, not a density prediction.
• 2
  Chamber oxygenKeep below about 1,000 ppm to cut oxide inclusions.
• 3
  HIP900–1,000 °C at about 100 MPa closes internal porosity.

Conformal cooling channels inside a mold insert are the classic win. A 1.0–1.5 mm internal channel with a 3 mm wall prints reliably in CuCrZr and cannot be machined at all. That is the case where printing earns its cost.

Thin walls are the weak point. Below about 0.8 mm in CuCrZr the wall may build but the density at the surface drops, and a HIP cycle can distort it. If your design calls for a 0.5 mm copper fin, expect to machine it from plate instead.

Overhangs and large flat areas are also risky. Copper conducts heat away from the melt pool into the solid below, so the top surface of a wide plate cools unevenly and curls. Supports help, but they also conduct heat away. A 45° rule of thumb is not enough; angle the part so that large flats stay off the build plate.

Internal features smaller than about 0.3 mm will not clear powder reliably. If a channel is that small, the powder stays in it and you will not know until the part is cut open. Design for 1 mm minimum internal channel diameter in copper.

• 1
  GoodConformal channels 1.0–1.5 mm with 3 mm walls.
• 2
  PoorFins and walls below about 0.8 mm in CuCrZr.
• 3
  BlockedInternal channels below about 0.3 mm trap powder.

The most practical answer for many copper parts is both processes. Print the CuCrZr near-net shape so the conformal channel exists, then machine the sealing faces, threads and bores to final tolerance. The printed surface is rough and slightly porous at the skin, so the machined faces are what carry the seal.

Allow 0.3–0.5 mm of stock on every machined face. Less than that and the rough as-built skin may not clean up. More than that and you spend build time on material you will cut away, which matters because copper powder is expensive.

Aging after machining is usually better than aging before. Machine the part soft, then run the 470–500 °C age cycle, then finish the critical faces in a light pass. That order keeps the aged part from warping after the final cut.

GreatLight runs the machining side of this route on 5-axis centers holding ±0.005 mm, with the printed blanks supplied by the customer or by a partner. If the part needs both a printed channel and a leak-tight face, this is the sequence to plan.

• 1
  Stock allowance0.3–0.5 mm on every machined face.
• 2
  SequencePrint, machine soft, age, light finish pass.
• 3
  Tolerance±0.005 mm on the machined features.