Process development for laser powder bed fusion of GRCop-42 using a 515 nm laser source

Copper is widely used in high heat flux and electrical applications because of its excellent electrical and thermal conductivity properties. Alloying elements such as chromium or nickel are added to strengthen the material, especially for higher temperatures. Cu4Cr2Nb, also known as GRCop-42, is a dispersion-strengthened copper-chromium-niobium alloy developed by NASA for high-temperature applications with high thermal and mechanical stresses such as rocket engines. Additive manufacturing (AM) enables applications with complex functionalized geometries and is particularly promising in the aerospace industry. In this contribution, a parametric study was performed for GRCop-42 and the AM process laser powder bed fusion (PBF-LB/M) using a green laser source for two-layer thicknesses of 30 and 60 μ m. Density, electrical conductivity, hardness, microstructure, and static mechanical properties were analyzed. Various heat treatments ranging from 400 to 1000 °C and 30 min to 4 h were tested to increase the electrical conductivity and hardness. For both layer thicknesses, dense parameter sets could be obtained with resulting relative densities above 99.8%. Hardness and electrical conductivity could be tailored in the range of 103 – 219 HV2 and 24% – 88% International Annealed Copper Standard (IACS) depending on the heat treatment. The highest ultimate tensile strength (UTS) obtained was 493 MPa. An aging temperature of 700 °C for 30 min showed the best combination of room temperature properties such as electrical conductivity of 83.76%IACS, UTS of 481 MPa, elongation at break (A) at 24%, and hardness of 125 HV2.


I. INTRODUCTION
The development of Cu-Cr-Nb (also known as Glenn Research Copper-GRCop) alloys started in 1980 with the motivation to create a copper alloy with high thermal conductivity, elevated temperature strength, and long low-cycle fatigue (LCF) life. 1 The benchmark alloys were Cu-Cr, Cu-Cr-Zr, and NARloy-Z (Cu-3 wt.% Ag-0.5 wt.% Zr) which mostly lose tensile strength above 400 °C, whereas Cu-Cr-Nb still shows good strength at operating temperatures of 700 °C.At a 2:1 atomic ratio of chromium to niobium, the intermetallic compound Cr 2 Nb forms in the pure copper matrix.Until now, two technically relevant alloy compositions have been developed by NASA: Cu-8 at.% Cr-4 at.% Nb (GRCop-84) and Cu-4 at.% Cr-2 at.% Nb (GRCop-42).The dominant strengthening mechanisms are Hall-Petch and Orowan. 1 The reduced content of chromium and niobium leads to higher thermal conductivities and easier powder metallurgy, which is why today GRCop-42 is the preferred alloy over GRCop-84. 2As-extruded GRCop-84 shows an electrical conductivity of 67%IACS at room temperature.PBF-LB/M is an AM process where powder particles are selectively solidified layer by layer by a laser source to form complex parts (see Fig. 1).
The main process parameters are laser power P L , scan velocity v scan , hatch distance h, and layer thickness LT.The volumetric energy density (VED) can be calculated according to (1), For each process parameter study for a specific machine or material, there is an optimal VED process window where complete melting of the powder is possible and overheating is avoided.A comparison of the VED between different lasers and machines is difficult since the absorptivity of the material and the spot size can differ, having an impact on the coupling of the energy into the material.Hayes et al. 3 have achieved 99.9% relative density for PBF-LB/M-processed GRCop-84 and very high mechanical static properties without any heat treatments (see Table I).No information is available on the chosen process parameters.Seltzman and Wukitch have used the parameter sets shown in Table II in several publications, [7][8][9][10][11] and it seems that this was the developed parameter set for GRCop-84 by the Glenn Research Center on a Concept Laser M2.Their application is lower-hybrid launchers, and they have tested the nuclear response, brazing characteristics, microstructure, and radio frequency (RF) losses for use in hybrid launchers in a fusion environment.
Cooper et al. 5 processed GRCop-42 with a Concept Laser M2 with different productivity rates and energy densities and compared the properties with GRCop-84 values (see Table II).Layer thickness was increased to 45 μm and scan velocity to 1025 mm/s for higher productivity.The relative density was around 99%, which is why the HIP process was added for tensile testing (see Table I).This parameter set was then transferred to a EOS M400 machine with a larger build volume by Gradl et al. 15 Gradl et al. 2,[15][16][17][18][19] and other works, with the participation of either Gradl or Ellis, [20][21][22][23] report on NASA research activities for GRCop-84 and GRCop-42 for achieving the complete additive manufacturing process chain of launcher rocket engines.Wilms and Rittinghaus 6 printed 5 × 5 × 5 mm 3 cubes from a CuCrNb alloy with Y 2 O 3 nanoparticles at a higher VED compared with NASA process parameters, with a resulting relative density of >99.4% (see Table II).In Table I, mechanical properties such as yield strength (YS), ultimate tensile strength (UTS), and elongation (A) at room temperature for CuCrNb alloys are summarized from several publications for different material conditions.The strength is higher in the as-built state after PBF-LB/M than in the as-extruded state, and ductility is improved by hot isostatic pressing (HIP) treatment or heat treatment at 425 °C (SR = stress relieved).
The electrical conductivity σ corresponds to the thermal conductivity λ according to the Wiedemann-Franz law (2) for pure metals, with the Lorenz number L being approximately 2:44 Â 10 À8 WΩK À2 , 14 For GRCop-84, an electrical conductivity of 75%IACS is reported vs. 85%IACS for GRCop-42 and a thermal conductivity of 290 W/m K vs 350 W/m K, all at room temperatures. 15o far, only few publications have been found on process parameter studies for GRCop-42, subsequent heat treatment, and their effect on microstructure and mechanical properties.This study aims at providing more details on the process-structureproperty relationships of GRCop-42 processes with PBF-LB/M using a green laser system.

A. Materials and machine
The GRCop-42 powder used in this study had mean particle diameters of d10 = 13, d50 = 27, d90 = 48 μm, and a chemical composition of chromium 3.14 wt.% and niobium 2.79 wt.%.For manufacturing the samples, the TruPrint1000 Green Edition equipped with a 515 nm frequency doubled disk laser with a maximum laser power of 500 W and a spot diameter of 200 μm was used.The build volume was Ø 97 × 100 mm 2 build height.As inert gas, argon was used, and the stainless-steel 316L build plate was actively watercooled.The study consisted of three consequent studies summarized in Fig. 2.

B. Parameter study
A total of 20 single tracks with a length of 10 mm and 10 layers of LT = 60 and 30 μm were printed with the maximum laser power of 500 W and varying scan velocity from 240 to 1200 mm/s in steps of 50 mm/s.The single tracks were analyzed with a Keyence microscope and image analysis to measure the track width.In the second stage of the parameter study, density cuboids of 10 × 15 × 10 mm 3 were built according to Table III in three build jobs for each layer thickness with approximately 12-14 cuboids per build plate (see Fig. 3).
The parts were tilted 30°to the wiper to reduce stresses in interaction.During the print, some apparent overheating was visible, but the cuboids seemed unharmed and were, therefore, separated from the build platform via wire electrical discharge machining (wire-EDM) for further analysis.The electrical conductivity σ was measured on the machined surface using the eddy-current method with the Helmut Fischer Sigmascope350 and probe FS40 by averaging 10 single measurements per sample.All other side surfaces were grinded to obtain the bulk properties for subsequent Archimedes density measurement using a Sartorius Quintix 125D-1x scale.Each cube was measured three times and averaged.
Selected promising candidates were then microsectioned in the XZ plane along the build direction and embedded and polished for porosity measurement via image analysis.For each layer thickness, the best parameter set was selected as P1 and P2 for further characterization.The samples were vibration-polished for 15-20 h to remove the deformation layer and etched with Adler consisting of water, hydrochloric acid, ferric chloride, and diammonium tetrachlorocuprate.

C. Heat treatment study
To further improve conductivity and hardness, different heat treatments were tested according to the test plan in Table IV on 10 mm thick cylindrical segments with 16 mm diameter manufactured with the best parameter set P1 from the density study [see Fig. 3(b)].
The heat treatments were performed using a LOBA 1200-60-600-3 tube furnace (HTM Reetz GmbH) for HT1 to HT6 as well as HT9 and HT10.The HIP treatment was performed on the Quintus QIH 15L URQ, and the heat treatment of AMS5663 INCONEL718® was performed using the LOBA tube furnace for solution annealing and air cooling.For double aging, the Xerion X.VAC 3945 oven was used.Scanning electron microscopy (SEM) was performed using the JEOL JSM-7800F with the energy dispersive x-ray spectroscopy (EDS) device, Oxford X-Max, and electron backscattered diffraction was performed using a NordlysNano EBSD detector.Hardness HV5 was measured on the polished  v scan (mm/s) surfaces using the EMCO-M4 U 024 machine and hardness HV2 was measured using the LECO AMH55 machine at an average of 10 single measurements per sample.

D. Tensile testing
For P1 and P2, three samples were built in the horizontal (cuboid 10 × 10 × 60 mm 3 ) and vertical (Ø 10 mm × 60 mm) build direction and then machined to the geometry B5 × 25 according to DIN 50125.For the heat treatments HT3, 4, 7, and 8, 1 tensile sample each in the vertical build direction were built, machined, and tested.All tensile tests were performed by using the inspekt Table 50 kN machine.

A. Powder
The powder used in this study was spherical with low amounts of satellites [see Fig. 4   spots smaller than 1 μm, which were identified via EDS as Nb and Cr rich segregations within the Cu matrix.For these regions, the formation of Cr 2 Nb intermetallic phases can be suggested, as it has already been reported for gas-atomized powders with similar compositions. 23[Figs.4(d)-4(g)].From the investigation, the powder was assumed suitable for the PBF-LB/M process.It could be the case that the high reactivity of niobium led to oxide formation after preparation of the sample and contact to air.

B. Parameter study-Single tracks
The single tracks showed regular weld beads for both layer thicknesses and all chosen scan speeds.The measured widths are shown in Fig. 5.With border counts, the single track with agglomerating particles and without border only counts the molten track in the middle of the bead.The single tracks show a clear molten zone in the middle and agglomerated particles on the sides.
Counting in the agglomerated particle border zone, bead widths for LT = 60 μm were averaged 58 μm lower than for LT = 30 μm.The core width was independent of the layer thickness and gave a good indication to estimate suitable hatch distances for further volume trials.

C. Parameter study-Density cuboids
The Archimedes density of the manufactured cuboids is visualized for both layer thicknesses in Figs. 6 and 7 over the scan velocity and hatch distance.
The overall density values for LT = 30 μm are higher than those for LT = 60 μm because of the higher VED and deeper penetration depth from the laser into the previous layer.Increasing the hatch distance above 0.2 mm results in lower densities because of insufficient overlap of the single tracks.Hatch distances of 0.1 or 0.15 mm seem most favorable, with a higher tendency toward 0.15 mm for both layer thicknesses.Scan velocities should be below 900 mm/s for higher densities.Two parameter sets for LT = 30 and 60 μm were chosen for further microstructural characterization based on the Archimedes measurements and micrographs: P1 with P L = 500 W, v scan = 600 mm/s, h = 0.15 mm, LT = 30 μm, VED = 185.18J/mm 3 and P2 with P L = 500 W, v scan = 500 mm/s, h = 0.1 mm, LT = 60 μm, VED = 166.67J/mm 3 (see Table V), since those samples showed the lowest porosity in the metallographic sections and high values in the Archimedes measurements.Archimedes results were influenced by air bubbles forming at the top surface on which the sample labels were printed.Both parameter sets are conservative with regard to the scan velocity compared with literature values listed in Table I and use a significant higher volumetric energy density.
When plotting the electrical conductivity over the volumetric energy density, a clear trend can be seen: the higher the VED, the higher the electrical conductivity (see Fig. 8), with an upper limit in electrical conductivity inherent to the PBF-LB/M process of around 29%IACS.The electrical conductivity depends on defects in the matrix such as precipitates or pores as well as on the crystal structure.Interestingly, cuboids with LT = 30 μm and h = 0.15 mm, all with a density above 98%, show a wide range of conductivity between 20.4 and 28.5%IACS.Therefore, apart from the density, the microstructure resulting from the energy input has an overlying effect on the electrical conductivity.Higher standard variations were observed mainly at lower densities.A further increase in electrical conductivity demands a subsequent heat treatment.

D. Parameter study-Microstructure
In Fig. 9, the etched as-built microstructure of P2 shows elongated grains in the z-building direction as well as fine segregation in the melting zones.The agglomerated gray particles are either elongated with measured widths of around 40 μm (1, 2, and 4 in Fig. 9) or round shaped (5 in Fig. 9) with a size of 20 μm.The agglomeration of this size can have negative effects on mechanical properties and has not been reported in the literature.EDS mapping confirmed the gray areas to be Cr, Nb-rich phase with a dendritic structure similar to that found in Ref. 24, where the dark phase mainly consisted of chromium and the lighter phase of Cr2Nb.
Because the used green laser has a larger spot size than other commercial infrared laser machines, the resulting larger melt pool, in combination with the high VED, could lead to a direct precipitation of CrNb within the melt and local concentration and agglomeration of the eutectic CrNb phase with high Nb content.This results in larger "defects" of CrNb phases and not fine precipitations needed for the Hall-Petch strengthening mechanism.It is unclear whether the oxygen pick up happened during the AM build process or during SEM sample preparation under atmospheric conditions.

E. Heat treatment study
Macro hardness HV2 and electrical conductivity σ were measured after performing the different heat treatments (see Figs. 10  and 11).
With increasing annealing temperature, the electrical conductivity increases from 25%IACS in the as-built state (at t = 0) to over 85%IACS when annealing above 700 °C.The influence of annealing temperature is much higher than the influence of annealing time.The HIP treatment at 900 °C results in a lower electrical conductivity than at annealing temperatures higher than 500 °C, indicating an additional effect caused by the high pressure.The AMS5662 INCONEL718® heat treatment results in the highest electrical conductivity and lowest hardness.Hardness peaks at 500 °C annealing temperature at 219 HV2.Further increased annealing temperatures result in decreasing hardness.Hardness at 1000 °C annealing temperature, HIP at 900 °C, and solution annealing at 980 °C for the AMS5662 heat treatment results in even lower hardness values than the as-built state.When overlaying both properties, the best compromise between high hardness and high conductivity is reached by annealing at 700 °C for 30 min.Figure 12 shows the SEM analysis of the samples after three different heat treatments HT1, HT3, and HT7.A highly directional alignment of the individual grains along the built direction can be clearly observed, Figs.12(a), 12(f ), and 12(k).A high density of Nband Cr-rich segregations is characteristic of melt track boundaries, as marked by red and orange arrows in Figs.12(b), 12(g), and 12(l).For the sample after HIP treatment at 900 °C for 120 min, a clear overlap of the signals for Cr and Nb can be observed, Figs.12(n) and 12(o).Here, the formation of intermetallic Cr 2 Nb phases can be assumed; however, a thorough structural analysis of the precipitates is required for a precise estimation.Furthermore, a preferred formation of Cr precipitates was observed in the microstructure of the HIP-treated sample at grain boundaries [black particles marked with yellow arrows in Figs.12(l) and 12(o)].Very fine black particles (Cr precipitates) are also visible in the sample heattreated at 700 °C for 30 min, Fig. 12(g).The coarsening of Cr precipitates can be explained by increased temperature and time of the heat treatment, as this process is diffusion-driven.Among the three samples metallographically characterized by SEM-EDS, the one heat-treated at 400 °C for 30 min has the highest hardness.This is possibly due to the more pronounced solid solution strengthening effect of Cr and Nb.This assumption is in good agreement with the measured conductivity values of the samples, where the sample heat-treated at lower temperature shows the lowest σ-value.It is a sign for a higher content of alloying elements in the copper matrix, which are detrimental for electrical conductivity of the alloy.Furthermore, precipitation of very fine Cr particles or clusters and Cr 2 Nb intermetallics can be assumed as an additional strengthening mechanism.However, a high-resolution structural analysis of the microstructure by means of TEM analysis is required to verify this assumption.After heat treatment at a higher temperature of 700 °C for 30 min, fine Cr precipitates became visible, while the hardness decreased slightly.Through an increase of temperature and time, as is the case for HIP treatment, and a corresponding coarsening of Cr precipitates, there was a further decrease in hardness.

F. Tensile testing
When comparing the stress-strain curves at room temperature for both layer thicknesses and both building directions (see Fig. 13), the horizontally built samples show a higher elongation at break than the vertically built samples.The highest UTS of 493 MPa was observed for LT = 30 μm built in the vertical direction, which at the same time, also showed the lowest elongation at break A of 18.3%.
When comparing the effect of the heat treatment on the static mechanical properties (see Fig. 14), a longer annealing time leads to reduced strength and ductility.The higher the annealing temperature, the lower the YS and UTS.The best combination of ductility and strength was reached with annealing at 700 °C for 30 min, which was already the best compromise for electrical conductivity and hardness.The HIP treatment significantly decreased electrical conductivity, also translating into a lower thermal conductivity with lower strength.Therefore, contrary to what is reported in the literature, where HIP is the usually performed heat treatment, annealing at lower temperatures at ambient pressure shows a better combination of properties (Table VI).

IV. CONCLUSIONS
• PBF-LB/M processing of GRCop-42 with a green laser source can produce highly dense samples for layer thicknesses of 30 and 60 μm.• Highest densities were observed at a hatch distance of 0.15 mm, and there should be a further investigation if the scan velocity can be further increased for higher productivity.

FIG. 1 .
FIG.1.Schematic view of a PBF-LB/M process with a green laser.
FIG. 3. LT = 30 μm density cuboids build job (a) cylinder build job for heat treatment study and tensile testing (b).

•
Heat treatment is needed to tailor the microstructure and mechanical properties depending on the final application.• Homogeneously distributed fine Cr 2 Nb particles are hard to obtain, and agglomerated Nb and Cr-rich phases larger than 10 μm in size are formed during PBF-LB/M.• The volumetric energy density input can be used as an in-situ heat treatment to tailor microstructure.• Annealing temperature of 700 °C for 30 min shows the best trade-off between electrical conductivity, ductility, and strength.• The process-microstructure-property relationship including precipitates and crystal structure should be investigated further, especially after analyzing different process parameters and subsequent heat treatments.More tensile samples should be tested for the purpose of statistical verification of the findings

TABLE I .
Mechanical properties for GRCop alloys.

TABLE II .
Known process parameters of GRCop alloys.

TABLE III .
Sample plan for density cuboids with P L = 500 W, LT = 30 and 60 μm, X = sample was built.

TABLE IV .
Sample plan for heat treatments.FC, furnace cooling; AC, air cooling.

TABLE V .
Properties of parameter sets P1 and P2.
FIG. 8. Electrical conductivity vs VED for all density cuboids with selected relative density labels.FIG.9. Etched microstructure of P2.