Experimental Evaluation of Thermo-Mechanical Properties of GRCop-42, Produced by PBF-LB, at Low Temperatures

Abstract

Today, Powder Bed Fusion-Laser Based technology is widely used in many industrial fields, but some high-demanding applications are still not fully investigated, such as low temperatures. In basic physics research, experiments usually use low temperatures to reduce external influences and to increase the sensitivity of particle detectors, accelerators, etc. The production capabilities of this technology have become a standard for manufacturing such components, and the demand for high performance has led to the investigation of new materials, like GRCop-42. It possesses excellent thermal properties and strength at high temperatures, and although several works have been published in recent years, full research on its behaviour at low temperatures is still missing. The aim of the paper is to investigate the mechanical properties of GRCop-42, produced by PBF-LB, from low to room temperature, like Elastic Modulus and Poisson’s ratio, and correlate them with thermal conductivity in the as-built state and after heat treatment. The results showed that the material can maintain high strength even at low temperatures, without losing ductility and the ability to store strain energy; moreover, after heat treatment, it increases its thermal properties due to the way the precipitates are dispersed in the copper matrix.

1. Introduction

Today, Powder Bed Fusion-Laser Based (PBF-LB) technology is widely used in many industrial fields (i.e., aerospace, automotive, medical, etc.), but some high-demand applications are still not fully investigated, such as low temperatures. In basic physics research, experiments usually use low temperatures to reduce external influences and to increase the sensitivity of instruments (i.e., particle detectors, accelerators, etc.) [1,2]; moreover, uncommon conditions, such as ultra-high vacuum level and high radiopurity background [3], are fundamental to investigate theories and find new particles. Radiopurity is a different concept from the purity of a metal, usually characterized by the absence of impurities or undesirable elements, such as oxygen, phosphorus, etc. In fact, it refers to the detection of minimum traces of radionuclides of thorium, uranium, caesium, and potassium. It is related to the natural radioactivity of materials, and it can be influenced, for example, by the manufacturing process [4]. Consequently, the mass of components and their geometric shape play an important role. In this scenario, highly optimized geometries together with high-performing materials to withstand, for example, high thermal gradients are required. The production capabilities of PBF-LB technology have become a standard for the manufacturing of such components, and the demand for high performance has led to the study and investigation of new PBF-LB materials, like precipitation-hardened alloys.

Among them, GRCop-42 (Glenn Research Center Copper, Cu-4 wt.% Cr-2 wt.% Nb) [5] copper alloy showed great potential, and not only for the high-temperature for which it was developed (i.e., combustion chamber components). Its applications involve the design of high-strength and low-weight ratio components, the capability to operate with static and dynamic cyclic loads, the chemical compatibility with fuels, and, finally, the ability to withstand both low (−200 °C) and elevated temperatures (800 °C). GRCop-42 belongs to the Cu-Cr-Nb alloys family that are born as an alternative of other common high strength-conductivity alloys, such as C-18150 (i.e., CuCrZr), NARloy-Z (Cu-3 wt.% Ag-0.5 wt.% Zr), Glidcop-15 (Cu-0.15 wt.% Al₂O₃), and C-18200 (Cu-1.2 wt.% Cr) [6]. It possesses excellent thermal properties, great creep resistance, and strength at high temperatures. The lower amount of Cr and Nb, compared to its predecessor GRCop-84 (Cu-8 wt.% Cr-4 wt.% Nb), makes the production of the powder easier by means of gas atomization without compromising the mechanical properties and ensuring a higher thermal conductivity (about 15–20% more). Also, it has higher thermal expansion due to the half content of Cr₂Nb precipitates that helps to restrain the dilatation. The temperature stability and coarsening resistance of Cr₂Nb precipitates in the copper matrix give the material high strength, while the thermal conductivity is controlled by the amount and the characteristics of this secondary phase [7]. As a result, because the thermal expansion can induce non-negligible strain, the GRCop-42 possesses a lower fatigue life that is, in part, counterbalanced by the higher thermal conductivity that results in a lower operative temperature [8]. Moreover, due to the lower Cr₂Nb phase content, its ductility is higher than that of GRCop-84, as the pure Cu matrix has the main effect on deformation. This effect is more evident above 600 °C, where both GRCop alloys have about the same strength, while below 600 °C, GRCop-84 is clearly stronger. In fact, with respect to other competing copper alloys, the Cu-Cr-Nb family can withstand greater stresses above 600 °C: thanks to oxidation limits, that is up to one order of magnitude lower, no mechanical or microstructural properties change up to 800 °C. In comparison, at high temperatures, the other copper alloys would return to mechanical properties like pure copper. At elevated temperatures, the higher thermal conductivity of GRCop-42 is more evident than GRCop-84; in fact, the decrease is the result of the excess of Cr dissolving in the Cu matrix [9].

The literature of recent years presents some studies, mainly related to GRCop-84. In particular, A.H. Seltzman and S.J. Wukitch [10,11] describe the investigation of this material for RF applications for use in fusion reactors, like the loss analysis in PBF-LB waveguides due to surface roughness and decreased conductivity compared to Oxygen Free Copper (OFC), highlighting how surface finishing (e.g., EDM cut, wet blasting, vibratory finishing, and chemical etching) is an enabling technology for PBF-LB components to provide low loss and a surface compatible with high-power vacuum RF. The same authors [12] explored the resolution and geometric limitations in PBF-LB GRCop-84 structures for a lower hybrid current drive launcher, showing how PBF-LB is an enabling technology for fusion reactor RF components, allowing the manufacturing of complex, high-strength RF structures that would have otherwise been impossible or impractical to fabricate with conventional machining and suggesting several practical recommendations on the manufacturing process. In addition to this, they analysed the braze wetting properties of GRCop-84 compared to CuCrZr alloy and OFC, also demonstrating the possibility of brazing such a material with molybdenum alloys [13]. More specific research on GRCop-42, developed, for example, by G. Demeneghi et al. [14], concerns the size effects of thin wall structures produced by PBF-LB and their impact on fatigue life in the as-built and HIP condition. The study demonstrated the significant implications of the application of thin structures, as reducing the component size (i.e., thickness) can affect its life and performance. On the same subject, S. Sahoo et al. [15] presented a texture analysis of the GRCop-42 thin wall with different thicknesses, which inherently comes with different grain morphology, size, and orientation distributions. The results show a strong correlation between stress localization and the initial grain orientation. In addition, M. H. Rahman et al. [16] propose an ultrasonic-assisted electropolishing to reduce the surface roughness of GRCop-42 heat exchangers. The procedure, employing a combination of HCl, CuCl₂, and SiC, clearly enhanced the surface finish, and the profilometric analysis indicated a notable improvement in surface morphology, characterized by the effective elimination of surface irregularities and the attainment of a more homogenized and smoother surface texture. It should be noted that surface finishing techniques represent an important aspect of post-processing techniques of PBF-LB components, due to the design constraints that several applications require. Research on this subject has been conducted on other common copper alloys, such as C-18150 (i.e., CuCrZr) [17,18]. Results show that the building direction and heat treatments have a significant influence on the machinability of the parts, while a proper combination of process parameters (i.e., layer thickness, building angle, and scanning speed) affects the surface quality macroscopically in contrast to other parameters (i.e., laser pattern) that do not seem to have any contributions. It is not only PBF-LB technology that has been employed to manufacture such Cu-Cr-Nb materials, but also Direct Energy Deposition (DED) [19,20,21]. The resulting material properties change substantially between the two production processes. For example: (i) DED has minimal internal defects in contrast to PBF-LB, which exhibits a substantial amount of porosity, with porosity percentage increasing as the specimen thickness decreased; (ii) the strength of PBF-LB is higher than DED considering the same thickness, thanks to the finer grain size of the material; (iii) the PFB-LB specimens failed primarily due to crack growth and pore coalescence, being facilitated by the high amount of porosity, while DED specimens failed predominantly at the interlayers; (iv) PBF-LB specimens showed columnar grains parallel to the build direction and several grain islands perpendicular to the build direction along the scanning path, while DED specimens demonstrated a zig-zag pattern parallel to the build direction and fine equiaxed grains at the wall edges; (v) surface topography exhibited significant differences: PBF-LB specimens showed significant powder adherence to the surface, while DED specimens displayed minimal to no powder residue on the surface but exhibited surface waviness [22].

Usually, copper alloys present some technical challenges when they are manufactured by PFB-LB technology. Typically, the common machines employ an InfraRed (IR) laser source, which, due to the low energy absorbance of copper at this wavelength, produces a low relative density of the material. In the case of the Cu-Cr-Nb family, the presence of the Cr₂Nb phase increases the powder absorption coefficient at the IR wavelength, resulting in a greater and better energy transfer to the melting pool. Also, because the reflectivity of copper decreases with temperature, the laser source can transfer more heat energy to the powder. These conditions make it able to produce higher relative density materials, with respect to the common copper alloys, typically above 99.0%, without any particular post-treatments, such as the Hot Isostatic Pressing (HIP), that is used both to reduce residual stresses and close any remaining porosity; although, in this case, the overall elongation is significantly reduced with respect to the as-built state condition [9]. The resulting PBF-LB material has a very uniform Cr₂Nb phase distribution, thanks to the rapid cooling of the melting pool through the copper substrate that prevents the segregation of precipitates. Also, with respect to the powder, the Cr₂Nb particles are finer. This aspect is fundamental considering the Ashby–Orowan strengthening mechanisms, where the strength of a material is increased as the average diameter of particles is decreased [9].

Although several works on this material have been published in recent years, full research on its behaviour at low temperatures is still missing. For this reason, the aim of the paper is to investigate the mechanical properties of GRCop-42, produced by PBF-LB, such as Elastic Modulus (E), Poisson’s ratio (ν), Yield Strength (YS), and Ultimate Tensile Strength (UTS), from −150 °C (123.15 K) to room temperature, and correlate them with Thermal Conductivity (TC) in the as-built state and after an appropriate heat treatment (i.e., 700 °C for 30 min. in air, plus furnace cooling). Tensile tests have been conducted by means of a universal testing machine equipped with an environment chamber able to reach −150 °C (123.15 K) using liquid nitrogen, while tensile strains have been evaluated by a biaxial extensometer equipped with multiple independent acquisition channels. On the other hand, the experimental evaluation of TC has been performed in a dry cryostat capable of operating in a vacuum from −243.15 °C (30 K) to 26.85 °C (300 K). Also, tensile data have been used to build a constitutive material model, able to predict the behaviour of the material at low temperatures and to be easily implemented inside of the most common finite element software. The specimens for both experimental tests were produced by a standard PBF-LB machine equipped with a 300 W IR laser source, and they have also been characterized by means of metallographic analysis.

The results showed that the material can maintain high strength even at low temperatures, without losing ductility and the ability to store strain energy; moreover, after a proper heat treatment, it can maintain good thermal properties due to the way the precipitates are dispersed in the copper matrix.

2. Materials and Methods

2.1. GRCop-42 Production

The GRCop-42 was produced on a standard PBF-LB machine (i.e., PRIMA Print Sharp 150, Prima Additive, Turin, Italy) equipped with a 300 W Infrared (IR) laser source and a cylindrical building platform. The main technical characteristics of the machine are listed in

Table 1. Main characteristics of the PBF-LB machine used to produce the GRCop-42 specimens.

Characteristic
Model	PRIMA Print Sharp 150
Building volume	Ø 150 mm × 160 mm
Laser source	1 × 300 W IR fibre laser
Laser spot (d)	35–100 µm
Layer thickness	20–120 µm
Building platform heating	Up to 300 °C

. The process parameters, such as laser power (P), scanning speed (S), hatch distance (H), and layer thickness (L), were evaluated using a well-established experimental procedure [23] able to reduce the experimental data points and identify the optimal levels using a scaling of the Volumetric Energy Density (VED), Equation (1), between PBF-LB machines equipped with different laser spot sizes (d).

$$\mathrm{V}\mathrm{E}\mathrm{D}={\displaystyle \frac{\mathrm{P}}{\mathrm{S}\mathrm{H}\mathrm{L}}}[\mathrm{J}/{\mathrm{m}\mathrm{m}}^3]$$

(1)

The chemical composition and the main physical properties of the GRCop-42 powder, supplied by the Metal4Printing (M4P) company (Klagenfurt, Austria), are shown in

Table 2. Chemical composition and physical properties of the GRCop-42 powder.

Chemical Composition			Physical Properties
Element	Min [wt.%]	Max [wt.%]	Grain size	Flow	App. Density
Al	0.00	0.01	d10 = 13 µm d50 = 27 µm d90 = 53 µm	18 s/50 g	4.67 g/cm3
Cr	3.10	3.40
Cu	Balance
Fe	0.00	0.01
Nb	2.70	3.00
Si	0.00	0.01
O	0.00	0.05

. The process parameters have been evaluated on small cubic samples of 10 mm size. The selection was conducted by means of a full factorial Design of Experiment (DOE) with three levels (i.e., N = 3² = 9 combinations); see

Table 3. Full factorial design experiment (3² = 9) for GRCop-42 process parameters optimization.

#	P (W)	L (µm)	S (mm/s)	H (µm)	VED (J/mm3)	Laser Pattern	ρm ± 2σ (g/cm3)
1	270	60	730	80	77.1	Meander with a rotation of 67° between each layer	8.66 ± 0.05
2			730	90	68.5		8.72 ± 0.03
3 *			730	100	61.6		8.74 ± 0.03
4			750	80	75.0		8.69 ± 0.02
5			750	90	66.7		8.70 ± 0.02
6			750	100	60.0		8.70 ± 0.03
7			770	80	73.1		8.63 ± 0.04
8			770	90	64.9		8.64 ± 0.05
9			770	100	58.4		8.64 ± 0.03

* Optimal combination (theoretical maximum density 8.79 g/cm³).

.

The laser power (P) was set to the maximum operative value allowed by the machine (i.e., 10% less than the nominal maximum power), and the layer thickness (L) was defined as the maximum of the powder grain size distribution (d90) to ensure the proper levelling and compression of the powder grains during the spread on the building platform. On the other hand, the hatch distance (H) was changed according to the defocusing possibility allowed by the PBF-LB machine (i.e., d = ~45 µm) and the scanning speed (S) was adjusted to have a fine variation of the VED. Eventually, the laser pattern chosen was the Meander strategy with 67° of rotation between each layer. The DOE combination and the results of the relative density evaluation in the as-built (AB) state condition are reported in Table 3, while Figure 1 shows the cubic samples after the manufacturing process directly on the building platform where the gas flow and coater direction are highlighted. The placement has been conducted to avoid any possible contamination of unmelted powder or oxides during the laser scanning process on the samples that would have affected the relative density measurements.

The entire production process took place in an inert atmosphere of Ar (i.e., O₂ level < 0.1%) with a recirculation gas speed of 3 m/s. The preheating of the building platform was set to 90 °C.

The estimation of relative density was conducted by means of Archimedes’ principle, see Equation (2), and the measures were repeated three times in distilled water (${\mathsf{\rho }}_{\mathrm{f}\mathrm{l}\mathrm{u}\mathrm{i}\mathrm{d}}=$ 9.9777 gcm³ at ~22 °C). To remove any residual of unmelted powder, the surfaces of the specimens were thoroughly cleaned with an ultrasonic bath, while the lower areas, where the support structures were located, were manually polished with SiC paper. The density calculation was automatically performed by an analytical balance (i.e., Kern ABS 80-4N model (KERN & SOHN, Balingen, Germany): weighing capacity 82 gr, readability 0.100 mg, linearity 0.300 mg, repeatability 0.200 mg) thanks to its density kit and related software.

$$\mathsf{\rho }={\mathsf{\rho }}_{\mathrm{f}\mathrm{l}\mathrm{u}\mathrm{i}\mathrm{d}}{\displaystyle \frac{{\mathrm{m}}_{\mathrm{a}\mathrm{i}\mathrm{r}}}{({\mathrm{m}}_{\mathrm{a}\mathrm{i}\mathrm{r}}-{\mathrm{m}}_{\mathrm{f}\mathrm{l}\mathrm{u}\mathrm{i}\mathrm{d}})}}$$

(2)

The results of the optimal process parameters combination were then confirmed by a metallographic analysis of the side surface parallel to the building direction. Figure 2 shows the optical micrograph of the surface of the #3 cubic samples after polishing. The automatic evaluation of the internal porosity has been performed by means of a high-resolution optical microscope (i.e., KEYENCE VHX 7000, Keyence, Tokyo, Japan) and its analyser software (version 1.4.17.3). The software analysis reveals a residual porosity of ~0.30%. The exposure and contrast values of the image were adjusted to properly recognize the porosity (i.e., irregular black areas) from the presence of Cr₂Nb precipitates (i.e., more regular grey areas). On the other hand, Figure 3 shows the optical micrograph of the same surface where the precipitates inside the Cu matrix are highlighted. They represent ~0.39% of the overall area.

2.2. Experimental Tests Configuration

2.2.1. Mechanical Tests

Mechanical tests were conducted by means of a universal tensile machine (i.e., INSTRON 68FM100, INSTRON, Massachusetts, MA, USA) equipped with a load cell of 100 kN and an environmental chamber (i.e., INSTRON 3119-610, INSTRON, Massachusetts, MA, USA) able to reach −150 °C (123.15 K) using liquid nitrogen. The strains for the evaluation of Elastic Modulus (E) and Poisson’s ratio (ν) in the elastic field were recorded by a biaxial clip-on extensometer (i.e., INSTRON 2650-561, INSTRON, Massachusetts, MA, USA) with a gauge length of 25 mm and multiple independent acquisition channels: two for the axial strain and one for the transversal strain. On the other hand, the strains for the determination of the Yield Strength (YS) and Ultimate Tensile Strength (UTS) up to rupture were determined by a standard axial clip-on extensometer (i.e., INSTRON 2630-105, INSTRON, Massachusetts, MA, USA) with the same gauge length. Due to the temperature limits operation of a standard axial clip-on extensometer, the measurement of YS and UTS was limited at −100 °C/173.15 K. Tests were conducted according to ISO 6892-1 [24] and ISO 6892-3 [25] using a constant strain rate of 0.00025 s⁻¹ (i.e., method A222). The specimens were dimensioned considering the above standards and the available volume of the PBF-LB machine building platform. To optimize the measurement of the different mechanical properties, two types of geometric shapes were utilized: (i) flat-rectangular (Type #A) for the determination of E and ν (Figure 4); (ii) cylindrical (Type #B), with shoulder-ended without thread, for the evaluation of YS and UTS (Figure 5).

Both types of specimens were produced vertically with respect to that platform in the same printing job, without any tilt angle and with support structures only on the bottom surfaces. Figure 6 shows the specimens after the manufacturing randomly placed on the building platform.

After the manufacturing process, specimens were removed from the building platform by Electrical Discharge Machining (EDM). Half of them were left in the AB state condition, while the other half were heat treated (HT). The heat-treatment process has been selected considering the optimisation of mechanical and thermal properties. For this reason, 700 °C for 30 min. in air, plus furnace cooling, has been chosen. This annealing process should guarantee the best trade-off between electricity/thermal conductivity, ductility, and strength [26].

2.2.2. Thermal Tests

The experimental evaluation of TC (λ) has been performed in a dry cryostat capable of operating in a vacuum from −243.15 °C (30 K) to 26.85 °C (300 K). The geometry of the specimens, Type #C (Figure 7), has been customized to guarantee, as much as possible, a linear heat flow along the main direction. In this way, it was assumed that the heat energy is conducted unidirectionally along the cross section of the specimen. As the tensile samples, TC specimens were produced vertically with respect to that platform (Figure 6), EDM cut, and half of them were heat-treated (i.e., 700 °C for 30 min. in air, plus furnace cooling). With respect to tensile specimens, their side surfaces have been polished to ensure perfect thermal contact with the cold finger of the cryostat.

During the measurements, the cryostat (Figure 8) was kept at a pressure of 10⁻⁴ Pa to reduce the contribution of thermal convection. The TC specimen was mechanically fixed with M2 plastic screws to a pure copper cold finger that is thermally connected to a cryocooler (i.e., CRYOMECH PT90, Cryomech Inc., Syracuse, NY, USA). The temperature of the cold finger was adjusted by the reading of a set of Resistance Temperature Detectors (RTD) provided by 100 Ω Platinum Resistors (PT100) with four wires. The temperature was controlled by a control unit (i.e., LAKESHORE 335, Lake Shore Cryotronics, Westerville, OH, USA) feeding the current in two high-power resistors according to a Proportional-Integral-Derivative (PID) algorithm based on the reading of the PT100. Other PT100s were located on the TC specimen in the middle (TC) and at both ends (TA and TB) and on the surface (TD) of the super Multi-Layer Insulation (MLI), which is located externally to minimize the radiation emissivity between the cryostat walls and the specimen surfaces (i.e., <0.003). The different temperatures were acquired by a dedicated acquisition unit (LAKESHORE 218, Lake Shore Cryotronics, Westerville, OH, USA). Finally, a heat source (i.e., metal film resistor), controlled by a power supply unit (i.e., KEYSIGHT E3647A, Keysight, Santa Rosa, CA, USA), has been connected to the TB end of the specimen to regulate the temperature.

A generic measurement cycle starts at −233.15 °C (40 K), after which the cryocooler control unit changes the temperature with a fixed increase of +20 °C/K, up to −73.15 °C (200 K). After an idle time of 1 h, necessary for the overall thermalization of the system, different powers are provided by the heat source to the specimen. For each power setting, a further thermalization period of 1h was necessary for the thermalization. After that, the temperatures (i.e., TA, TB, and TC) are acquired. The uncertainty of the temperature readout was evaluated in ~1 °C/K.

3. Results

3.1. Mechanical Properties

3.1.1. Experimental Data

Mechanical properties (i.e., E, ν, YS, and UTS) have been experimentally evaluated, as described in Section 2. Thanks to the determination of E and ν, the Shear Modulus (G) has also been calculated (see Equation (3)). Average results, determined by means of three specimens (i.e., ID #1–3) under different temperatures and state conditions (i.e., AB and HT), have been reported in

Table 4. GRCop-42: average mechanical properties (E, ν and G) in the AB and HT condition from −150 °C (123.15 K) to room temperature.

Condition	T (°C/K)	$\overline{\mathbf{E}}$ (GPa)	$\overline{\mathsf{\upsilon }}$	G (GPa)
AB	−150/123.15	122.1 ± 0.5	0.369 ± 0.003	44.6
	−100/173.15	118.4 ± 0.2	0.347 ± 0.005	44.3
	−50/223.15	115.3 ± 0.4	0.338 ± 0.002	43.5
	+25/298.15	112.5 ± 0.2	0.328 ± 0.004	42.4
HT	−150/123.15	117.4 ± 0.2	0.373 ± 0.002	42.8
	−100/173.15	115.2 ± 0.1	0.365 ± 0.002	42.2
	−50/223.15	113.5 ± 0.3	0.341 ± 0.006	42.3
	+25/298.15	112.4 ± 0.6	0.328 ± 0.001	42.3

and

Table 5. GRCop-42: average mechanical properties (YS, UTS, and A) in the AB and HT condition from −100 °C (173.15 K) to room temperature.

Condition	T (°C/K)	$\overline{\mathbf{Y}\mathbf{S}}$(MPa)	$\overline{\mathbf{U}\mathbf{T}\mathbf{S}}$ (MPa)	$\overline{\mathbf{A}}$ (%)
AB	−100/173.15	308.6 ± 2.9	513.0 ± 1.20	4.8 ± 0.7
	−50/223.15	303.0 ± 5.5	491.5 ± 19.6	7.9 ± 0.6
	+25/298.15	284.1 ± 2.3	429.9 ± 44.3	7.4 ± 6.3
HT	−100/173.15	307.8 ± 2.1	514.8 ± 28.5	6.9 ± 3.2
	−50/223.15	296.9 ± 4.8	493.6 ± 12.5	7.2 ± 1.1
	+25/298.15	283.5 ± 2.5	435.5 ± 3.07	8.2 ± 0.3

and plotted in Figure 9, Figure 10, Figure 11 and Figure 12 (ID #1 specimen). The estimation of the uncertainty of measurements has been conducted for E, ν, YS, and UTS considering a standard deviation of 2σ (95%).

$$\mathrm{G}={\displaystyle \frac{\overline{\mathrm{E}}}{2(1+\overline{\mathsf{\nu }})}}$$

(3)

As expected, the Elastic Modulus (E) increases with the reduction of temperature, as with the Poisson’s ratio (ν). From room temperature to −150 °C (123.15 K), the variation of Elastic Modulus (E) is more evident for the AB condition with respect to HT one, while for the Poisson’s ratio (ν), the behaviour is almost equivalent. The Poisson’s ratio (ν), determined from Equation (4) as the ratio of transverse strain (${\mathsf{\epsilon }}_t$) to axial strain (${\mathsf{\epsilon }}_a$) shows that as the temperature decreases, it progressively decreases, also producing the increase in Shear Modulus (G). Tensile data at room temperature are comparable with those proposed by literature [5,6,8] except for the elongation (A), which is lower, and it is in the range of 6–8% with respect to a reference value of 20–30%. This can be attributed to the surface quality of the specimens (Ra = 15–20 μm), which were not post-processed and have been left in the AB condition like most of the PBF-LB components. In fact, a rough surface can be a cause for early failure. Moreover, it has been noticed that there is a larger variability of elongation (A) for AB specimens with respect to HT ones, especially at room temperature. This fact can be explained by the random selection of specimens and by the fact that heat treatment unifies the microstructure of the material (i.e., Cr₂Nb precipitates shape and dispersion).

$$\mathsf{\nu }=-{\displaystyle \frac{{\mathsf{\epsilon }}_t}{{\mathsf{\epsilon }}_a}}$$

(4)

Finally, the plots show how the behaviour of Elastic Modulus (E) can be described by a nonlinear interpolation of the data function, whereas the Poisson’s ratio (ν) has an almost linear behaviour. On the other hand, the stress-strain curves present no areas with a clear yield point, and between the AB and HT conditions no significant differences were found. This fact seems to show that the heat treatment we carried out had no apparent effect. The results showed that the material can maintain high strength even at low temperatures without losing ductility and the ability to store strain energy with respect to room temperature.

The failure of all specimens in both AB and HT state conditions occurred as a brittle rupture, with just a small visible necking (Figure 13). The brittle fracture can be interpreted by analysing the presence of Cr₂Nb precipitates. In fact, they can act as nucleation sites for voids that then coalesce into fracture cusps. The phenomenon, typical in PBF-LB due to a high cooling rate, is not observed in GRCop material subjected to HIP or produced by classic processes. Fracture cusps can be microscopic peaks or cavities on fracture surfaces that indicate where voids have formed and coalesced [27].

Usually, rupture is initiated by the fracture of Cr₂Nb precipitates and then continues within the copper matrix. The location of the precipitates is usually near the centre of the cusps. It is possible to correlate the cusps’ size and the dimension of precipitates to understand the influence on material strength. The same mechanism was detected in the specimens tested at low and at room temperatures (Figure 14). With respect to heat-treated material at higher temperature and for a longer time (i.e., 900 °C for 5 h) [27], where transition between polycrystalline and monocrystalline precipitate structure occurred together with a coarsening of precipitates, the out heat-treatment (i.e., 700 °C for 30 min.) does not exhibit any relevant difference that can be detected by an optical microscope analysis.

3.1.2. Constitutive Material Model

Tensile data were used to build a constitutive material model able to predict the behaviour of the GRCop-42 at low temperatures. The Voce–Chaboche (V-C) [28] Equation (5) has been chosen and implemented because it is one of the most common and reliable material models that is possible to find inside finite element analysis software. The model belongs to the “Exponential Law” Equations with a negative exponent, in which the stress progressively decreases with increasing plastic deformation until rupture. In fact, the negative exponent produces a progressive stress levelling during the hardening of the material as the plastic deformation increases. The c1 term is the YS of the material, the c2 term is the saturation term, and the c₃ term is the saturation exponential term, which describes the rate of change of stress. The reported Equation differs from the original V-C Equation by the additional linear term c₄ [29], which allows a better account of the stress increases as plastic deformation progresses. The relationship can, in any case, be returned to the original form by setting this last coefficient equal to zero.

$$\mathsf{\sigma }={\mathrm{c}}_1+{\mathrm{c}}_2\left(1-{\mathrm{e}}^{-{\mathrm{c}}_3{\mathsf{\epsilon }}_{\mathrm{p}}}\right)+{\mathrm{c}}_4{\mathsf{\epsilon }}_{\mathrm{p}}$$

(5)

The coefficients have been determined by means of a “Nonlinear Generalized Reduced Gradient” (GRG) [30]. The algorithm calculates, at each iteration, the gradient of the objective function, i.e., error function (6), with respect to the experimental data and the solution. The definition of the V-C model has been conducted only for the HT state condition because it is of more general interest for applications. Figure 15 shows the plot of the experimental data vs those of the model in terms of true stress and true strain, while

Table 6. GRCop-42: Voce–Chaboche (V-C) calibration coefficients in the HT condition from −150 °C (123.15 K) to room temperature.

Condition	T [°C/K]	c1	c2	c3	c4	c5	e [MPa]
HT	−100/173.15	222.5	161.5	99.5	2743.8	5.4	5.4
	−50/223.15	259.3	161.3	56.0	1428.9	11.2	11.2
	+25/298.15	233.8	157.7	62.8	1034.9	5.1	5.1

reports the coefficients of the V-C model together with the error function mean value. The plot clearly demonstrates how the V-C model can correctly represent the real behaviour of the material at different temperatures.

$$\mathrm{e}={\sum }_{\mathrm{k}=1}^{\mathrm{m}}\sqrt{{\displaystyle \raisebox{1ex}{${\sum }_{\mathrm{j}=1}^{\mathrm{N}}{({\mathsf{\sigma }}_{\mathrm{e}\mathrm{x}\mathrm{p}.}-{\mathsf{\sigma }}_{\mathrm{m}\mathrm{o}\mathrm{d}\mathrm{e}\mathrm{l}})}^2$}\!\left/ \!\raisebox{-1ex}{$\mathrm{N}$}\right.}}$$

(6)

3.2. Thermal Conductivity

The experimental evaluation of TC (λ) has been conducted by the well-known conduction Equation (7), assuming no heat generation or heat loss along the specimen. The term Q is the heat power applied by the metal film resistor (see Section 2), l is the length (i.e., 63 mm) of the specimen between the TA and TB measurement points, w is the width (i.e., 5 mm), and t is the thickness (i.e., 3 mm). The results were reported in

Table 7. GRCop-42: TC in the AB and HT condition from −233.15 °C (40 K) to −73.15 °C (200 K).

Condition	T (°C/K)	λ (W/m K)
AB	−233.15/40	40.5
	−213.15/60	46.2
	−193.15/80	55.4
	−173.15/100	62.5
	−153.15/120	66.0
	−133.15/140	69.1
	−113.15/160	72.6
	−93.15/180	74.0
	−73.15/200	79.6
HT	−233.15/40	252.4
	−213.15/60	266.0
	−193.15/80	274.1
	−173.15/100	275.0
	−153.15/120	277.2
	−133.15/140	279.5
	−113.15/160	285.7
	−93.15/180	286.1
	−73.15/200	291.6

and shown in Figure 16.

$$\mathsf{\lambda }={\displaystyle \frac{\mathrm{Q}\mathrm{l}}{\left({\mathrm{T}}_{\mathrm{A}}-{\mathrm{T}}_{\mathrm{B}}\right)\mathrm{w}\mathrm{t}}}[\mathrm{W}/\mathrm{m}\mathrm{K}]$$

(7)

The plot reveals how the TC, from the AB to the HT state conditions, changes significantly, increasing as the temperature changes by 4–6 times. This evidence indicates how heat treatment (i.e., 700 °C for 30 min. in air, plus furnace cooling) had a significant influence on the microstructure of the material without affecting the mechanical properties. In fact, tensile tests did not find any relevant differences between the AB and HT state conditions. Also, the TC at room temperature for the HT state condition is comparable to that of the literature [31] (i.e., ~320 W/mK), usually reached after HIP post-process.

The mechanisms that increase TC after the heat treatment may be associated with several factors: (i) the precipitates’ solubility [27], (ii) the precipitate coarsening [32], and (iii) the reduction of defects and residual stress [8]. Because the Cr₂Nb precipitates have a low solubility, they do not dissolve in the copper matrix, leaving more pure copper without elements in solid solution that can affect the heat flow. On the other hand, the precipitate coarsening results in less interface within the copper matrix, and fewer interfaces correspond to less thermal dispersion and, thus, higher TC. Finally, heat treatment can reduce dislocations and crystalline defects that would interfere with electron flow, limiting the TC.

To better understand how Cr₂Nb precipitates dispersed in the copper matrix affect the TC after heat treatment, metallographic analyses were conducted on the cubic sample #3, produced for the PBF-LB process parameter setup, in the AB state condition and after the heat treatment. The analysis was performed by means of a high-resolution optical microscope (i.e., KEYENCE VHX 7000). The sample was polished and etched until the melt pools became visible with a mixture of 2 g K₂Cr₂O₇, 8 mL H₂SO₄, 4 drops HCl, and 100 mL water (ASTM E407-07) [33]. Figure 17 shows the optical micrograph of the side surface of the #3 cubic sample in the AB condition after etching. The morphology of the melting pool on the core and contour regions is clearly visible, like the transition area. The contour melt pool can be categorized as “distinct” [34], where a continuous stack of melt pools along the surface of the sample are present. The stack of the contour melt pool is more or less equal to the layer height, while the bulk scan tracks, which are rotated in each layer by 67°, produce a change in the melt pool dimensions. The aspect of the melting pool confirms how the PBF-LB process parameter setup was correctly performed [23], identifying, for example, the optimal hatch distance in relation to the print layer (i.e., H = 100 µm and L = 60 µm).

By analysing the optical micrograph with higher magnification (Figure 18 and Figure 19) on the same #3 cubic sample, it is possible to identify the columnar grain [35] growth along the building direction, together with the Cr₂Nb disperse phase (i.e., Hall–Petch precipitation strength mechanism [36]). In particular, the grey areas are Cr and Nb-rich phases, while dark ones contain Cr and lighter Cr₂Nb [26].

On the other hand, Figure 20 shows the optical micrograph of the same #3 cubic sample after the heat treatment in the cross-section. The figure shows a higher number of Nb-rich phases than the AB condition, probably produced by the dimension of the melting pool, generated by the combination of the laser spot (d = ~45 µm) and the hatch distance (H = 100 µm). The average dimension of the Nb-rich phase in both state conditions is the same within a range of 20–35 µm. The higher presence of this Nb-rich phase in the HT state may reduce the strengthening mechanism by precipitation [26] and may increase the TC due to the less interface within the copper matrix. In fact, the mechanical properties between AB and HT conditions did not show significant changes, but rather a slight decrease for the latter. The significant increase in thermal conductivity with heat treatment may support additional precipitation of Cr and Nb from a solid solution in the copper matrix, which could be countering precipitate coarsening effects on YS and UTS.

4. Discussion

The experimental results demonstrated how the GRCop-42, produced by PBF-LB, can maintain high strength even at low temperatures, without losing ductility and the capability to store strain energy with respect to the behaviour at room temperature. After a proper heat treatment (i.e., 700 °C for 30 min. in air, plus furnace cooling), it increases the TC without affecting the tensile strength. In fact, the TC remains almost constant over the temperature, decreasing only 14%, from ~250 to 290 W/mK, and the UTS remains well above 500 MPa at −233.15 °C (40 K). Moreover, the experimental evaluation of the Elastic Modulus (E) and Poisson’s ratio (ν), by means of a biaxial ex-tensometer, provides important information for use in the design phase that is often estimated by literature with a high degree of uncertainty. In fact, it is well known that materials produced with PBF-LB have different and unique physical properties and characteristics. The variation of the Elastic Modulus (E) is more evident in the AB state condition with respect to HT; it increases from ~112 to 122 GPa. The Poisson’s ratio (ν) also changes, but less noticeably, between the two state conditions. The value of ~0.33 at room temperature is comparable to that known for copper alloys [37], while it increases accordingly as the temperature decreases to ~0.37. This experimental behaviour makes the material a perfect candidate for low-temperature and cryogenic applications (e.g., basic physics research), where it is essential to maintain (i) high TC to ensure a good thermalization of components and (ii) mechanical strength to withstand the self-induced stresses produced by thermal expansion and to resist low thermal fatigue cycles.

A direct comparison between the TC in the HT state condition with pure copper (i.e., C-10200, OFC) [38] and other common copper alloys (i.e., C-18500, CuCrZr) [39], usually used in low-temperature applications (Figure 21), highlights how the TC of the material remains competitive with pure copper (~25% lower) and almost constant as temperature decreases. Also, even in comparison with CuCrZr alloy, the TC seems to be higher (~50 W/mK) below 40 K. This evidence opens new scenarios since, at low temperatures, even a small percentage increase in TC results in large operational advantages. Because of the precipitation strengthening mechanism, other types of heat treatments can be investigated to further improve and optimize this behaviour. Although the material was developed for high-temperature applications, this does not preclude its evaluation for other uses.

On the other hand, considering the relationship between thermal and mechanical properties (

Table 8. Comparison between the thermo-mechanical properties of GRCop-42 in the HT state condition, at −100 °C (173.15 K), with C-12200 (Cu 99.95%) and C-15000 (Cu-Cr) alloy.

Material	$\overline{\mathbf{Y}\mathbf{S}}$ (MPa)	$\overline{\mathbf{U}\mathbf{T}\mathbf{S}}$ (MPa)	$\overline{\mathbf{A}}$	λ (W/mK)
GRCop-42 HT	307.8 ± 2.1	514.8 ± 28.5	6.9 ± 3.2	275
C-12200 (Annealed)	~50	~330	~60	~395
C-15000 (Cold Draw + Aged)	~450	~520	~24	~220

), it can be noticed that the material has a clear advantage compared to both pure copper 99.95% (C-12200) and the Cu-Zr alloy (C-15000) [40], in respect to the pure copper for mechanical performance, while in respect to Cu-Zr alloy for thermal one. The materials chosen as comparisons were selected based on the availability of reliable experimental data. It should also be mentioned that the mechanical properties of pure copper depend on the supply conditions (i.e., annealed, cold draw, etc.) and can change significantly over a wide range.

Future research developments will involve the evaluation of electrical properties at low temperature and the comparison with other reference materials (i.e., OFC). Also, the experimental assessment of the thermal fatigue at low cycles will be conducted together with other types of heat treatments to further improve and optimize the thermo-mechanical properties. Eventually, a special focus will then be made on the chemical composition, using Inductively Coupled Plasma-Mass Spectrometry (ICP-MS) analysis, to check before and after PBF-LB process how the alloying elements in the material change.

5. Conclusions

In this paper, an experimental investigation of the mechanical properties of GRCop-42, produced by PBF-LB, from −150 °C (123.15 K) to room temperature has been presented. Mechanical properties, such as Elastic Modulus (E), Poisson’s ratio (ν), Yield Strength (YS), and Ultimate Tensile Strength (UTS), were correlated with Thermal Conductivity (TC) in the as-built state and after an appropriate heat treatment (i.e., 700 °C for 30 min. in air, plus furnace cooling). The material was produced by a standard PBF-LB machine equipped with a 300 W IR laser source, and it was also characterized by means of metallographic analysis.

The results showed that the GRCop-42 can preserve high strength even at low temperatures without losing ductility and the ability to store strain energy; moreover, after heat treatment, it increases thermal properties due to the way the Cr₂Nb precipitates are dispersed in the copper matrix. In fact, the precipitates have a low solubility, and they do not dissolve, leaving the copper without elements in solid solution that can affect the heat flow. Also, the precipitate coarsening results in less interface within the copper matrix, and fewer interfaces correspond to less thermal dispersion and, thus, to higher thermal conductivity. Finally, a proper heat treatment can reduce defects, like dislocations, that would interfere with electron flow, limiting the thermal conductivity. On the other hand, the heat treatment did not increase the mechanical properties. In fact, the higher presence of the Nb-rich phase after heat-treatment may have reduced the strengthening mechanism by precipitation, typical of this copper alloy.

Future research developments will involve: (i) the evaluation of electrical properties at low temperature; (ii) the experimental assessment of the thermal fatigue at low cycles; (iii) the evaluation of other heat treatments to further improve and optimize the thermo-mechanical properties; and (iv) the chemical composition analysis before and after PBF-LB process to understand how the alloying elements in the material change and influence its performance.