Pure-Copper LPBF Parameters: Building a Window Without False Recipes

Short answer: a credible pure-copper laser powder bed fusion (LPBF) process window is not a table of power and speed values copied from a paper. It is a controlled evidence package that connects one identified material, powder condition, machine, optical source, build setup, specimen geometry, and measurement method to repeatable outcomes. Change one of those items and the evidence may no longer support the same window.

This distinction matters because copper couples weakly and dynamically with many near-infrared systems, conducts heat away rapidly, and can move between lack-of-fusion, stable conduction, and unstable deep-melt behavior over a narrow operating region. A buyer does not need a universal recipe. A buyer needs to know what was varied, what was held constant, how defects and properties were measured, what confirmation was completed, and which changes trigger requalification.

Disclosure: This article was prepared with AI-assisted research and editorial review.

1. Define the deliverable as a bounded process window

A process window is a set of conditions shown to meet named responses within a declared experimental boundary. The boundary should identify at least the copper grade and chemistry, powder lot and history, machine model and serial configuration, source wavelength and beam delivery, nominal spot or profile, layer thickness, scan strategy, atmosphere, recoating system, build layout, specimen type, post-build condition, and inspection method. Without that tuple, the word “window” is too vague for procurement or qualification.

The target responses also need to be explicit. Relative density alone is not enough. A useful window can include defect morphology and location, electrical or thermal conductivity, tensile behavior, dimensional response, surface condition, recoating stability, and monitoring consistency. Select them from the part function: a heat-transfer component and a current-carrying component may reject different pore distributions at similar reported bulk density.

Terminology can hide this boundary. DMLS, SLM, and LPBF are often used inconsistently; the practical distinctions are explained in this copper LPBF terminology guide. For a technical purchase order, name the actual process, machine, energy source, material specification, and controlled parameter revision instead of relying on a marketing label.

2. Freeze the material-machine-measurement tuple before DOE

Before a design of experiments (DOE) begins, freeze the configuration that gives every result its meaning. Record powder supplier, grade, lot, particle-size distribution and chemistry methods, oxygen result and sampling point, virgin/reused blend, sieving history, and storage controls. ISO/ASTM 52907 provides a framework for metal powder characterization and traceability; it does not prescribe a universal pure-copper oxygen limit or reuse rule.

Freeze the equipment state with equal care: machine and source identity, wavelength, measured power, focus convention, beam profile, scanner calibration, recoater, build plate and preheat, gas and oxygen measurement, filter state, software, and parameter-file revision. “Same model” does not prove “same delivered process.” Alignment, contamination, field calibration, and flow distribution can differ.

Finally, freeze the measurement system. Identify the specimen, section plane, preparation procedure, image resolution, thresholding method, sampling plan, conductivity standard, surface preparation, and post-processing condition. ASTM F3637-23 explains why X-ray computed tomography, metallography, and Archimedes-type bulk density measurements do not report the same information. A density percentage without its method, resolution, and sampling is not transferable evidence.

3. Treat absorptivity as a state-dependent response

Copper absorptivity is not a single handbook number that can be inserted into every LPBF calculation. It changes with wavelength, surface condition, powder packing, oxide state, temperature, melt-pool geometry, multiple scattering, and the transition from loose powder to track and keyhole-like depression. Work by Nordet and co-authors measured different optical behavior among pure-copper substrate, powder bed, liquid tracks, and deep depressions under a 515 nm source. The safe inference is that coupling evolves during exposure; it is not that one measured value describes every layer and machine.

Published powder measurements illustrate the scale of the wavelength effect while also showing why the values must remain bounded. Gruber and co-authors reported 515 nm and 1064 nm absorptivity values for two identified gas-atomized copper powders using their stated method. Those results support wavelength-sensitive experiment design, but they are not universal constants for all copper grades, particle distributions, surface oxides, or measurement instruments. The broader energy-source trade space is covered separately in this green, infrared, and electron-beam comparison; this article does not re-rank source technologies.

For DOE, absorptivity should therefore be treated in two ways. First, use material- and state-specific measurements when they are available, with method uncertainty retained. Second, observe process responses that reveal changing coupling: track continuity, melt-pool emission, plume and spatter behavior, pore morphology, remelt evidence, and stability across the scan field. Do not convert nominal laser power into “absorbed energy” by multiplying it by an unverified constant.

4. Separate power, speed, spot, layer, and hatch effects

Power, scan speed, spot size or beam profile, layer thickness, and hatch spacing are independent controlled factors, even though they interact physically. A defensible DOE estimates their main effects and selected interactions instead of assuming that one factor can always compensate for another.

• Power sets nominal optical output, but coupled power also depends on calibration, beam delivery, surface state, and melt geometry.
• Scan speed changes interaction time, melt-pool length, cooling history, and plume dynamics. Doubling speed is not necessarily equivalent to halving power.
• Spot size and beam profile determine irradiance distribution and thermal gradients. Diameter needs a measurement definition; Gaussian, top-hat, and ring profiles cannot be reduced to one diameter alone.
• Layer thickness changes the volume and packing state that each pass must consolidate, plus the overlap with the prior layer.
• Hatch spacing changes lateral overlap, remelting, thermal accumulation, and the probability of unbonded corridors between tracks.

Scan order, contour strategy, rotation, delay, and geometry add further interactions. Limit the initial factor set but document every fixed scan setting. A result such as “400 W worked” is incomplete if speed, focus, layer, hatch, profile, powder, specimen, and density method are absent.

5. Volumetric energy density is an index, not a recipe

The familiar volumetric energy density (VED) expression, E_v = P/(vht), combines power P, scan speed v, hatch spacing h, and layer thickness t. It can be useful as a reporting index within a tightly controlled experiment. It does not contain spot size, beam profile, wavelength, absorptivity, powder packing, scan strategy, gas flow, recoating, preheat, thermal history, or geometry. Conditions with the same calculated VED can produce different irradiance, melt-pool shapes, overlap, spatter, and pore populations.

VED also hides direction. Increasing power and increasing overlap can raise the same index through different physical paths. One may deepen a melt depression; the other may increase cyclic reheating. Likewise, decreasing layer thickness and slowing scan speed can yield the same VED change while altering consolidation and thermal accumulation differently. Plot the original factors and responses, retain interaction terms, and use VED only as a secondary descriptor.

This is why published numerical windows must stay attached to their setup. For example, Stoll and co-authors used a DOE on identified Cu-ETP powder with a 1064 nm system rated to 500 W and reported their best density and conductivity within that study. Jadhav and co-authors examined identified 99.99% copper powder with a 1080 nm source, a reported focal spot, layer thickness, hatch, power range, and speed range. These are valuable original experiments. Neither becomes a starting recipe for a different powder, optical train, machine, coupon, or measurement method.

6. Use a staged DOE with failure gates

A staged DOE prevents attractive but unstable coupons from advancing. Randomize where practical, block nuisance variables such as build position or powder lot, include repeats, and preserve references. Factor levels must come from machine capability, powder evidence, safe trials, and engineering judgment. Published values can inform bounds only after equivalence is demonstrated.

DOE stage	Primary question and responses	Failure gate before advancing	Evidence retained
System baseline	Are powder, optics, atmosphere, plate, recoater, and measurement systems in a known state?	Calibration, contamination, oxygen control, spreading, or measurement-system issue unresolved	Lot records, calibration records, beam/profile evidence, oxygen log, spread checks
Tracks and thin layers	Where are continuity, wetting, width, penetration, balling, plume, and spatter reasonably stable?	Discontinuous tracks, substrate damage, severe ejection, or unstable coupling	Factor matrix, track metrology, images, monitoring channels, build location
Bulk coupons	How do power, speed, profile, layer, hatch, and interactions affect pore type, density, conductivity, surface, and dimensions?	Connected lack-of-fusion, unstable deep pores, unacceptable drift, or response outside declared limits	Raw and analyzed results, section plan, CT settings, conductivity method, statistical model
Feature artifact	Does the candidate window hold at thin walls, channels, overhangs, contours, turns, and changing section mass?	Geometry-specific defects or dimensions fail the part-relevant criteria	Artifact drawing, orientation, location, inspection map, contour/core revision
Confirmation and repeat	Are chosen conditions repeatable across builds, locations, operators, and allowed powder condition?	Confidence interval, control trend, or repeatability fails the approved plan	Confirmation builds, SPC evidence, deviations, nonconformance dispositions

Choose failure gates before looking at results. Otherwise the team can move acceptance limits to preserve a preferred parameter set. The final selected region should include margin around the target, not merely one best-performing point. It should also show which response controls each edge of the region: lack of fusion, unstable deep melting, surface breakdown, conductivity loss, dimensional drift, recoating failure, or another declared mechanism.

7. Wavelength and beam shape redefine the experiment

Changing wavelength changes initial optical coupling and often changes the entire hardware package. Changing from infrared to green can also change available power, optical materials, focal geometry, spot size, beam profile, scanner behavior, calibration method, and maintenance state. A comparison that changes several of these variables is a system comparison, not a wavelength-only causal test.

Nordet and co-authors compared 515 nm and 1080 nm processing of the same stated pure-copper powder in one chamber, but the reported green and infrared beams differed in spot size and profile. That work is useful because it demonstrates real process-system behavior; it should not be reduced to a universal wavelength multiplier. Wang and co-authors reported a dense pure-copper result with a 532 nm system under their powder, 40 µm spot, 30 µm layer, stated power/speed/hatch ranges, coupon, and measurement methods. Those dimensions and values remain study-specific.

Beam shaping is also a first-class factor. Recent original work on pure copper reported high relative density and conductivity with both Gaussian and ring profiles in the authors' system, while surface response differed. The correct conclusion is that beam profile can relocate process boundaries and response trade-offs. It is not that ring, top-hat, or Gaussian illumination always wins. Record the measured profile at a defined plane and power condition, not only the commanded profile name.

8. Gas flow, recoating, and build layout belong in the window

Gas flow removes plume and spatter and interacts with build placement. NIST work on an EOS M290 with Inconel 718 showed downstream spatter transport plus removal and powder-disturbance behavior as flow changed. It is not a copper velocity prescription; it is evidence that flow, local low-speed zones, and layout can affect exposure and must be recorded.

Recoating controls the input layer that the laser actually sees. Track blade or roller type, stiffness, wear, direction, speed, powder feed, packing, streaks, raised edges, collision events, and interruptions. Copper's thermal behavior can create local distortion that later becomes a spreading problem. A nominally correct laser window cannot rescue a missing or nonuniform layer.

Include representative build positions and gas-flow orientation during confirmation. Do not qualify only one small cube in the chamber center and assume equivalence at the upstream edge, downstream edge, or across a densely nested production plate. The general copper AM workflow is summarized in this practical copper printing overview, while powder history and spreading risks are treated in the copper powder and reuse guide.

9. Monitoring detects drift; ex-situ tests establish outcome

Coaxial intensity, off-axis thermal imaging, photodiodes, layer images, oxygen records, pressure, gas-flow indicators, laser command logs, and recoater events can reveal departures from the qualified state. NIST IR 8538 describes relationships between low or high melt signals and lack-of-fusion or keyhole-like risk, while also emphasizing immature capability, false positives, and limits of automatic correction. ASTM E3353-22 similarly treats process-monitoring signatures as indications; an indication is not automatically a flaw, and a flaw is not automatically a rejectable defect.

Monitoring needs a registered evidence chain. Synchronize commands and sensors, map signals to build coordinates, retain threshold revisions, and correlate indications with metallography or CT. NIST work using time-stamped laser commands, coaxial images, and XCT demonstrates the approach; it does not establish a universal copper classifier or replace part acceptance.

Use ex-situ methods to decide whether the physical result met the engineering requirement. Density method, CT voxel size, metallographic sampling, conductivity specimen geometry, temperature, surface preparation, heat treatment, and machining state all affect interpretation. The relationship between density, impurities, thermal history, and conductivity is discussed in this printed-copper conductivity guide. A monitoring alarm can trigger review or quarantine, but disposition needs the approved quality plan.

10. Machine transfer requires change control, not scaling

A process window does not travel automatically between machines, even when model names match. Transfer must compare the complete material-machine-measurement tuple, then demonstrate equivalence through calibration, reference artifacts, confirmation coupons, monitoring distributions, and part-relevant inspection. Scaling power by spot area or wavelength absorptivity is a hypothesis to test, not evidence of equivalence.

Challenge events include source or optical replacement, scanner service, alignment, recoater or gas-system change, sensor change, software update, parameter revision, machine move, powder-source or reuse-policy change, and measurement-method revision. Write the trigger and required response before production.

NASA MSFC-SPEC-3717 provides a strong critical-hardware example: an approved qualification manufacturing plan is controlled, certain machine events negate qualification, and calibration or statistical verification is required to reestablish the state. It is a NASA spaceflight baseline, not a universal commercial obligation. ISO/ASTM 52904:2024 and ISO/ASTM 52908:2023 provide broader production-control and PBF part-qualification frameworks. Contractual applicability and current revisions must be verified for the program.

11. Put these 12 inputs in the RFQ and qualification plan

A useful RFQ asks for evidence needed to judge equivalence and risk. These twelve inputs are a technical starting set; add application-specific safety, cleanliness, regulatory, or traceability requirements.

1. Part function and failure consequence: thermal, electrical, pressure, vacuum, structural, fatigue, leak, or mixed duty.
2. Copper material definition: grade, chemistry limits, temper or final condition, and prohibited substitutions.
3. Powder controls: supplier, lot traceability, size distribution method, morphology, oxygen/chemistry method, reuse and blending rules.
4. Machine identity: manufacturer, model, machine identifier, controlled hardware configuration, and service status.
5. Energy source: wavelength, measured power basis, spot/profile definition, focus convention, and calibration evidence.
6. Layer and scan controls: thickness, hatch, contour/core strategy, scan order, rotation, delays, and parameter-file revision.
7. Atmosphere and flow: gas, oxygen limit and sensor location, pressure, flow verification, filter state, and build orientation to flow.
8. Recoating and layout: recoater type/condition, spreading controls, plate material, preheat, nesting, location, and witness coupons.
9. Required outcomes: defect type/location limits, density method, conductivity, mechanics, dimensions, surface, cleanliness, and leak criteria.
10. Inspection plan: method, resolution, sampling, section planes, calibration/reference standards, data retention, and acceptance authority.
11. Post-processing: stress relief, heat treatment, HIP if any, support removal, machining, cleaning, and sequence controls.
12. Change and nonconformance control: transfer triggers, requalification depth, deviation approval, quarantine, disposition, and record retention.

Procurement decision	Supplier evidence requested	Buyer acceptance or change trigger
Material and powder equivalence	Certificate, lot and reuse history, sampling plan, PSD/chemistry/oxygen methods, storage and sieving record	Declared limits met; new source, grade, lot class, blend, or method reviewed under the approved plan
Machine and optical state	Machine ID, calibration, measured power, beam profile/spot basis, scanner status, maintenance and software revision	Qualified configuration confirmed; service, replacement, move, update, or out-of-tolerance event triggers documented review
Process-window evidence	DOE matrix, original factors, interactions, repeats, response methods, failure edges, confirmation builds	Part-relevant responses meet predeclared limits with sufficient margin and repeatability
Production surveillance	Monitoring plan, oxygen/recoater logs, witness results, traceability, deviation and nonconformance records	Signals remain within approved control rules; excursions are quarantined and dispositioned, not silently averaged
Final part acceptance	Inspection reports, method settings, post-process records, property results, dimensional and cleanliness evidence	Drawing, specification, and contractual acceptance criteria met by the named authority

The supplier's existing process qualification can reduce duplicated work only when its scope matches the RFQ. The copper LPBF qualification evidence guide explains how to distinguish a parameter claim, a machine capability demonstration, a material dataset, and part-specific acceptance.

12. Five forbidden extrapolations before release

1. Do not copy a published parameter row as a production recipe. It remains evidence for the paper's material, powder, machine, source, specimen, and method.
2. Do not treat equal VED as equal physics. Different power, speed, hatch, layer, spot, profile, wavelength, and scan paths can share one VED while producing different defects.
3. Do not convert one absorptivity measurement into universal absorbed power. Coupling changes with powder state, surface chemistry, temperature, melt geometry, and measurement method.
4. Do not transfer a window by machine model, spot-area scaling, or wavelength ratio alone. Demonstrate equivalence and run confirmation evidence under change control.
5. Do not use density or monitoring as a standalone acceptance claim. Retain pore morphology, sampling and resolution, required properties, part geometry, and the approved engineering acceptance basis.

The commercial decision is straightforward: buy demonstrated capability, traceable control, and a defined response to change—not a persuasive number without provenance. To request a technical review, provide the twelve inputs above through the COPPER 3DP RFQ page. A responsible response should state what is known, what still requires coupons or testing, and what cannot be inferred from published work.

Publisher and engineering responsibility: Published by COPPER 3DP / Suzhou Como. This article provides general engineering decision guidance, not a qualified manufacturing instruction or part acceptance. The manufacturer must own parameter development and process control; the responsible buyer, design authority, and quality authority must define and approve material, testing, acceptance, change control, and regulatory requirements for the actual application.

Primary and authoritative sources: ISO/ASTM 52904:2024; ISO/ASTM 52907:2019; ISO/ASTM 52908:2023; ASTM F3637-23; ASTM E3353-22; NIST IR 8538; NIST laser-control and pore-correlation study; NIST spatter-transport study; Nordet et al., dynamic pure-copper absorptance; Gruber et al., green and infrared pure-copper LPBF; Stoll et al., pure-copper DOE; Jadhav et al., pure-copper infrared LPBF; Wang et al., green-laser pure-copper LPBF; Nordet et al., green/infrared system comparison; and original pure-copper beam-shaping research.