Analysis of Microstructure and Mechanical Properties of CoCrMo Alloys Processed by Metal Binder Jetting Multi-Step Technique

Abstract

Metal Binder Jetting (BJT/M) has emerged as a promising additive manufacturing (AM) technology for the realization of complex parts using a wide range of metal alloys. This technology offers several advantages, such as design flexibility, reduced lead times, a high building rate, and the ability to fabricate intricate geometries that are difficult or impossible to achieve with conventional manufacturing methods. Cobalt Chromium Molybdenum (CoCrMo) alloys are particularly suitable for demanding applications in the aerospace, biomedical, and industrial sectors that require high strength and hardness, corrosion resistance, and biocompatibility. In this work, ten cubic and ten tensile samples were printed with a layer height of 50 µm using the shell printing method, debound and sintered at 1325 °C for 4 h, with the aim of investigating the properties of CoCrMo parts made using BJT technology. A density of 7.88 g/cc was obtained from the Archimede’s test. According to the printing and sintering parameters, an average hardness of 18.5 ± 1.8 HRC and an ultimate tensile strength of 520.5 ± 44.6 MPa were obtained. Finally, through a microstructure analysis, an average grain size of 182 ± 14.7 µm was measured and the presence of an intergranular Cr-rich phase and Mo-rich carbides was detected.

1. Introduction

Cobalt Chromium Molybdenum (CoCrMo) alloys are a class of materials that offer a combination of many advantageous properties such as high strength and hardness, excellent wear, corrosion and heat resistance, a high biocompatibility and self-polishing abilities [1]. These properties allowed these kinds of alloys to be adopted for several mechanical applications such as aircraft engines, thermal barrier coatings and artificial joints, as well as surgical implants. Nevertheless, some characteristics of CoCrMo alloy components, such as the high hardness, the wear resistance and the high content of alloy elements, make the use of conventional manufacturing methods very challenging. Consequently, casting and powder metallurgy (PM) are preferred for processing this class of materials. On the other hand, these processes suffer from the high costs associated with the design and manufacturing of the dies and difficulties in rapidly changing the designs. Cobalt alloys also play an important role in the performance of aero- and land-based gas turbines. Vacuum-cast nickel alloys predominate in the hot sections of modern aero turbine engines, but cobalt alloys are routinely used for particularly demanding applications such as fuel nozzles and vanes for industrial gas turbines [2].

Binder Jetting (BJT) is a non-beam additive manufacturing (AM) technology with strong potential for manufacturing large volumes of parts more cost-effectively than other AM technologies [3,4]. When BJT is used for metal materials, it is called Metal Binder Jetting (BJT/M). Compared to the established metal additive manufacturing technologies, the BJT/M process has significantly lower investment costs, up to 60% lower than Laser Powder Bed Fusion (PBF-LB/M), according to [5]. The BJT/M build rate is higher than build rate of both Material Extrusion (MEX) technology and powder-based techniques, such as Laser Powder Bed Fusion and Direct Energy Deposition Laser beam (DED-LB/M). Moreover, the printing operation does not require vacuum or inert gas atmosphere in most cases, and it occurs at room temperature, avoiding residual stresses [3]. During the printing, no support structures and high consumption energies are required [6]. Furthermore, in comparison with Metal Injection Molding (MIM), the BJT/M process allows for a higher part complexity with a higher cost effectiveness for lower volumes, as no tooling molds are required [7].

However, the true potential of BJT has yet to be realized due to a significant knowledge gap about the material–process–property relationships [4]. Indeed, BJT/M, according to the standard UNI EN ISO/ASTM 52900:2022 [8], is classified as multi-step process (BJT-MSt/M) because the geometry of the parts is given by the joining of material with a binder in the primary process step (e.g., printing) which is followed by material consolidation by sintering, with or without infiltration, in subsequent process steps, ensuring that the processes can obtain the final metal parts with a metal allow with the desired properties [9]. The adoption of debinding and sintering (D&S) is well-known from MIM process, and it is fundamental to the whole process of obtaining the final sintered metal part. In this context, it appears to be fundamental to obtaining a deeper understanding of the properties of the metal parts realized via BJT.

The material portfolio of BJT/M covers a wide spectrum of metallic alloys from common stainless steels [10,11] and copper [12] to titanium [3] and cobalt- or nickel-based alloys [13,14]. As additive counterpart MEX or traditional MIM are sinter-based technologies, the main target of previous studies was to optimize the sintering process with the aim of obtaining a nearly fully dense part by varying the main sintering parameters, i.e., temperature, holding time and heating rate [15,16].

In their study, Huber et al. [10] considered the sintering temperature and the holding time. The maximum mechanical properties of 17-4 PH parts were obtained at 1300 °C after 90 min. Furthermore, the authors observed that porosity had a greater influence on the mechanical response of the parts than grain size or the presence of secondary phases. Cabo Rios et al. [11] investigated 316 L samples sintered at 1300 °C and 1370 °C. As expected, at 1300 °C the microstructure revealed a fully austenitic phase but with a large number of pores, while at 1370 °C the samples reached a relative density in the range of 96–99% with an average pore size > 35 µm. Mao et al. [16] investigated the effect of vacuum sintering temperature on 316 L parts. When a sintering temperature of 1380 °C was considered, the authors obtained a relative density of 92.4% and an ultimate tensile strength (UTS) of 473.7 MPa. Khademitab et al. [13] evaluated the properties of CoCrMo parts for biomedical applications when subjected to an aging treatment at 800 °C. When sintered at the highest temperature (e.g., 1380 °C), the maximum value of relative density was reached (99.8%); in contrast, when aged, the parts benefited in terms of microhardness and tensile strength. Stoyanov et al. [17], through the adoption of Hot Isostatic Pressing (HIP), obtained a fully dense microstructure with a grain size in the range of 50–150 µm, without abnormal grain growth. Mostafei et al. [18] successfully realized Co-Cr-based parts using BJT/M. As in previous studies, the sintering temperature was the main parameter investigated, and at 1300 °C the parts reached the highest density (≃99.8%). Recently, Gilmer et al. [19], with the aim of reducing the contamination of residual carbon and limit the applications scenario of the BJT parts, developed an optimized binder. In this context, an optimized concentration of binder allowed the parts to be printed with a decrease in the carbon content of about 90%.

With respect to the known literature, CoCrMo parts were realized using a proprietary system, where from the printing step to the sintering one, the parameters were yet optimized. However, according to authors’ knowledge, no studies have investigated the properties of these parts. In the current study, a comprehensive analysis is reported about the density, mechanical and microstructural properties and roughness of parts realized using pre-optimized parameters for BJT technology to test the reliability and repeatability of the process.

2. Material and Methods

Cobalt-28Chromium-6Molybdenum (UNS R30075) powder was used as feedstock material for the BJT/M process.

Table 1. Chemical composition of CoCrMo [20,26].

Chemical Composition (wt.%)
Source	Co	Cr	Mo	Ni	Fe	C	Si	Mn	W	Al	Tn	N
ASTM F75	Bal.	27.00–30.00	5.00–7.00	0.50	0.75	0.35	1.00	1.00	0.20	0.10	0.10	0.25
TDS	Bal.	27.00–30.00	5.00–7.00	<0.50	<0.75	<0.35	<1.30	<1.00	-	-	-	<0.25
Current study	Bal.	23.39 ± 0.36	7.95 ± 0.36	1.51 ± 0.15	1.45 ± 0.26	1.31 ± 1.05	3.31 ± 0.16	0.43 ± 0.12	-	-	-	-

reports the chemical composition of the powder based on the ASTM F75 standard [20], the manufacturer’s technical data sheet (TDS) and the composition evaluated in this work on a cubic sample. The powder adopted was processed using the proprietary 3D printer from Desktop Metal Inc., Shop System™ (Desktop Metal Inc., Burlington, MA, USA). For the realization of the test samples, both cubic and tensile samples, some predefined printing parameters were assumed: a layer height of 50 µm, a shell saturation of the binder of 53.3% and an inner saturation of 33.3%. The Shop System™ exploits the shell printing method. As has been discussed by Miao et al. [21], during shell printing the binder is deposited in the perimetral area (shell) of the samples with a higher saturation than the inner part. Regarding the samples’ dimensions, the cubic samples were characterized by an edge length of 12 mm, while the tensile samples were designed according to the ASTM E8 standard [22]. In more detail, for the tensile samples, subsize ones were considered, with a thickness of 6 mm, a gauge length of 25 mm and a gauge width of 6 mm. In the printing stage, the cubes and the tensile samples were positioned on the XY plane with the sample axis parallel to the spreading axis of the powder; in contrast, the Z axis was coincident to the printing direction. Samples were automatically oversized using the proprietary slicer software with the aim of compensating for the shrinkage that occurs after sintering. The chemical analysis was performed with a scanning electron microscope (SEM) (LEO EVO-50XVP, Zeiss, Cambridge, Cambridgeshire, UK), using the Energy-dispersive X-ray spectroscopy (EDS) technique. After sintering, the carbon (C) content was significantly higher due to the decomposition of the binder during the debinding and sintering phases, as previously reported in the literature for sinter-based technologies [23,24,25]. At the same time, an enrichment of silicon (Si) of about +2.3 wt.% was registered, due to the presence of silicon oxides (SiO₂) in the matrix.

The debinding and sintering steps were performed in a single cycle in the Desktop Metal PF1 furnace (Desktop Metal Inc., MA, USA). Both steps used predefined parameters and the peak sintering temperature, 1325 °C, was maintained for 4 h in a nitrogen atmosphere.

The density (ρ) achieved after sintering was measured on the cubic samples according to the ASTM B962 [27], using a precision balance (d = 0.0001 g) and distillated water at 26 °C (ρ = 0.9968 g/cc). As reference value for the density of the CoCrMo alloy, 8.30 g/cc was used. It is worth remembering that the density can change with the chemical composition of the alloy.

The Rockwell C hardness (HRC) test was conducted on the same cubic samples using the durometer (Enrst NR3D, Cisam-Ernst s.r.l., Induno Olona (VA), Italy) with a cone diamond penetrator with an opening angle of 120° and a load of 1471 N. Five measurements were performed on each cube to obtain an average value and the relative standard deviation. Once tested, three cubes were prepared, mounted, grinded and polished using abrasive papers from P120 to P2400. In more detail, the cubes with the highest and lowest measured density values, as well as the one with the density closest to the overall mean, were selected for the subsequent analysis using an optical microscope (OM) (Hirox RH-2000, Hirox Co., Ltd., Tokyo, Japan) with an “MXB-10C” optic and an “OL-140II” objective lens. Based on the captured images, the amount of porosity, the size of pores and the grain size were measured using the image processing software ImageJ, Version 1.8.0. Each picture was binarized, and the threshold was manually defined with the aim of considering only the relevant pores.

The surface roughness was also evaluated on the cubic samples using a profilometer Taylor Hobson Surtronic 3P and a cut-off of 0.8 mm. The measurements were conducted on the top surface and on the four lateral ones, parallel to the printing direction. One measurement was performed on each surface and a global average was created.

Finally, for the tensile tests, a universal testing machine with a load cell of 100 kN was adopted (Galdabini s.p.a, Cardano al Campo, Varese, Italy) at room temperature. According to the standard, the preload was set equal to 5 MPa with a crosshead speed of 2 mm/min, while the rest of the test, up to failure, was conducted with a crosshead speed of 1 mm/min.

3. Results and Discussion

3.1. Density Analysis

The densities of the ten cubic samples were determined through the Archimedean method in distilled water at 26 °C and their values, ranging from 93.7% to 97.5%, are reported in

Table 2. Results of the density and relative density of the cubic samples measured via the Archimedean method. * A reference value of 8.30 g/cc was considered.

N° Sample	Density (g/cc)	Relative Density * (%)
1	8.02	96.6
2	7.99	96.2
3	7.85	94.5
4	7.78	93.7
5	7.79	93.9
6	7.82	94.3
7	7.78	93.7
8	7.79	93.8
9	8.09	97.5
10	7.89	95.1

. The average value of 7.88 ± 0.11 g/cc was derived, which was equal to the 94.9 ± 1.4% relative density when compared to the reference value for a CoCrMo cast alloy (8.30 g/cc).

To corroborate the results obtained, the samples with the lowest value (7.78 g/cc, in Figure 1a), the highest value (8.09 g/cc, in Figure 1b) and the value closest to the average (7.89 g/cc, in Figure 1c) were further examined through image analysis. In Figure 1, images of the cross-sections along the XZ plane of the three investigated samples and their relative binarized images are reported. Figure 1a–c confirm the high variability of the density results and the variability of the BJT/M process, despite the use of the same printing parameters.

In more detail, the image analysis revealed an average pores size equal to 24.4 ± 5.3 µm. Considering the investigated cross-sections, the one with the highest number of pores is reported in Figure 1a. Furthermore, after the binarization, the porosities measured on the mounted samples were 3.07%, 0.70% and 1.92%, respectively, for the cross-sections reported in Figure 1a–c.

To better understand the pores’ distribution and their size, the sample with lowest density was further examined. The cross-section reported in Figure 2 shows some pores larger than the average size, characterized by different shapes, i.e., spherical, triangular or irregular, and positioned at grain boundaries and triple points. These pores are typical of sinter-based microstructures, and they are caused by the incomplete densification of the powder during the sintering process [16] or residual binder that evaporated during the thermal debinding (or pre-sintering) [24]. These defects remain a critical factor affecting the mechanical performance and reliability of the process. Further comments on the mechanical and microstructural properties are provided in Section 3.2.

3.2. Mechanical and Microstructure Properties Evaluation

The hardness tests conducted on the cubic samples revealed an average value of 18.5 ± 1.8 HRC, which was close to the manufacturer’s TDS (17.9 HRC). Compared to the available literature on CoCrMo sinter-based technologies, the reported values are lower than those reported by Khademitab et al. [13] (≃29 HRC) for BJT parts and Herranz et al. [15] (≃27 HRC) for MIM parts. Differences can be attributable to the different printing parameters used for the BJT, the manufacturing process (MIM vs. BJT) and also to the different sintering parameters. Previous studies sintered the sample at 1380 °C for 2 h and 1 h, respectively; on the other hand, the samples studied in the present work were sintered with the predefined parameters (1325 °C for 4 h) in the Desktop Metal system. The porosities observed for the parts are strictly related to the sintering parameters. In Figure 3, an inverse correlation between the hardness values and the porosity values can be observed. When increasing the porosity, the content of pores increases, negatively influencing the hardness of the parts. In other words, when the porosity (orange line) decreases, the hardness (blue line) becomes higher.

Concerning the evaluation of tensile properties, ten tensile samples were tested according to the ASTM E8 standard, and a single average ultimate tensile strength (UTS) was achieved. The recorded UTS was lower than the CoCrMo sintered at 1380 °C [13,15], possibly due to a lower density (94.9% versus ≃99.8%), and lower than the standardized value of 655 MPa. On the other hand, the measured elongation was higher than the standard (8%) and approached what Herranz et al. [15] found for MIM-CoCrMo.

To summarize, the average density and the mechanical properties of the parts are reported in

Table 3. Summary of the properties of CoCrMo parts.

Property	Value
Density (g/cc)	7.88 ± 0.11
Hardness (HRC)	18.5 ± 1.8
UTS (MPa)	520.5 ± 44.6
Elongation (%)	12.2 ± 1.8

, while the stress–strain curves with the highest, the lowest and the average value of UTS are shown in Figure 4.

Considering the data, once tested, the fracture surfaces of the tensile sample were observed using an OM. For this analysis, an “MXB-5040RZ” optic with an “AD-5040LOWS” optical lens was used. Two different fracture surfaces were observed, and differences were particularly clear, as shown in Figure 5.

The fracture surfaces at 20× magnification are shown in Figure 5a,b, while in Figure 5c,d their details at higher magnification (60×) are shown. Figure 5a,c shows the fracture surface of a sample with a low content of pores (highlighted with black arrows) that can potentially affect the load bearing section of the sample, decreasing the tensile strength. On the other hand, in Figure 5b,d, the high content of black points could be related to the number of pores. In Figure 5d, the high number of pores identified with black arrows can be considered the main cause of the reduced tensile strength with respect to the counterpart with a reduced number of pores. These differences can be considered to be the consequence of an incomplete sintering of the powder due to the sample’s position in the furnace or a scarce densification in the green condition. Moreover, potential cracks or cleavage are marked with yellow arrows on the right side of Figure 5.

The mounted cubic samples were inspected using the OM and their microstructures were observed in a dedicated image processing analysis. For this, we were required to convert the figure into 8-bit and to use the command “find the edges”. After that, the threshold was defined manually with the aim of highlighting the grain boundaries and measuring the grain size using the ImageJ software. In Figure 6a–c, the polished cross-sections at two different magnifications can be observed; in contrast, in Figure 6b–d, the corresponding processed images are used to measure the grain size and to recognize potential phases.

Considering the three mounted samples investigated, the grains appeared predominantly regular in size and shape, as reported in Figure 6c,d, and an average grain size of 182 ± 14.7 µm was measured. The size measured approached the range found by Herranz et al. [15] on MIM parts and by Khademitab et al. [13] on BJT parts. Despite the different sintering temperature, similarities in the grain size were observed. As a rule, as is widely reported in the state of the art, an increase in the sintering temperature promotes a coarsening of the grains [28]. Despite the similar grain size, differences in tensile properties remained ill defined. According to the authors’ knowledge, during printing, a different level in the same batch, or different batches, of printing could be possible due to a different percentage and number of porosities and pore sizes [10,29]. Furthermore, it the presence of a potential Cr-rich liquid phase was observed at the grain boundaries, as also confirmed by Khademitab et al. [13], and is highlighted in bright green in Figure 6. Round pores or inclusions were dispersed in the grains or at the boundaries as observed in previous works on sinter-based parts [15,30].

A cross-section was observed using SEM for a deeper analysis and is shown in Figure 7. The microstructure appeared to be composed of several characteristics. Firstly, there was a Co-rich matrix in the grains (highlighted with black arrows in Figure 7). Secondly, there was an intergranular Cr-rich liquid phase (see spectrum in Figure 7), as reported above, which was characterized by a darker shade of gray than the Co-rich matrix and highlighted by white arrows in Figure 7. According to the EDS analysis, the Cr content in this area reached 42.3 wt.%. According to Bettini et al. [31], when the Cr content is higher than 25 wt.%, it can promote a rise in the σ-phase, affecting the material properties.

Furthermore, a high content of carbides was detected along the grain boundaries and in the Cr-rich phase (generally M₆C); these carbides, characterized by a bright white color in Figure 7, are rich in Mo (≃40 wt.%), as confirmed by the spectrum reported in the lowest part of Figure 7. A size range of 0.91–2.4 µm was measured. The excessive content of Mo-rich carbides, the potential presence of σ-phase and the presence of pores could be considered to be the potential detrimental causes of the lower tensile properties of the BJT parts with respect to previous works and the TDS. In detail, the pores reduced the size of the load bearing section, and the M₆C and σ-phase had a detrimental effect on the ductility and strength of the parts.

All these aspects can have a detrimental effect on the performance of BJT parts, limiting its adoption in some critical fields. Indeed, contamination must be limited and controlled to ensure the use of BJT components in critical scenarios, such as biomedical ones, where the requirements for the chemical composition of the adopted alloys are very strict [32].

A further analysis was focused on the Cr-rich phase in which the Mo-rich carbides precipitated (Figure 8). In detail, three linear profiles were tracked in three different areas: in the Cr-rich phase in which M₆C precipitated and in the Cr-rich phase without carbides and across the carbides, respectively.

This analysis corroborated the initial results obtained and shown in Figure 7. Along the intergranular phases, a substantial decrease in Co (line blue) with respect to the grains was detected; instead, the Cr and Mo increased (red and orange lines, respectively). A slightly similar trend in the Cr-rich phase was observed. A severe enrichment in Cr and a decrease in Co characterized this area. On the other hand, line data 3 confirmed the spectrum reported in Figure 7. The carbides shown in Figure 7 and Figure 8 explained the higher content of Mo and the substantial decrease in Co.

3.3. Surface Roughness Analysis

The last property investigated was the surface roughness of the BJT parts. BJT is characterized by a higher surface roughness when compared to other Metal AM technologies [33]. In this work, the surface roughness (Ra) was considered along the printing direction and on the top surface. Along the printing direction, an Ra equal to 9.0 ± 1.6 µm was measured; in contrast, on the top surface the Ra was slightly lower at 7.9 ± 1.3 µm. The lateral surfaces were also examined using a 3D optical profiler (Taylor Hobson CCI MP-HS, Leicester, UK), as reported in Figure 9a. Due to the sinter-based nature of the considered process, the surface appeared rougher and more jagged as shown in Figure 9a,b, which was consistent with the Ra results obtained. The investigated profile, parallel to the printing direction, reported in Figure 9c, echoed the layer height used (50 µm) for the manufacturing of the samples, as confirmed by the height profile range of ±50 µm.

4. Conclusions

Metal Binder Jetting technology (BJT/M) was adopted to manufacture parts in CoCrMo using the Desktop Metal proprietary system. Ten cubes and ten tensile samples were printed, debound and sintered and the following properties were derived:

• The average sintered density measured through the Archimedean method was 7.88 ± 0.11 g/cc.
• From the image analysis, the pores, which were randomly dispersed in the different samples, had both irregular and round shapes, with an average size of 24.4 ± 5.3 µm.
• The tensile test revealed an average ultimate tensile strength equal to 520.5 ± 44.6 MPa and an elongation of 12.2 ± 1.8%. The reduction in UTS (−20.5%) compared to the declared value in the standard was due to an incomplete densification during sintering, confirmed by the presence of pores in the bearing sections.
• The microstructure revealed an average grain size of 182 ± 14.7 µm with a prevalence of the Co-rich matrix. An intergranular Cr-rich phase was observed and a high content of Mo-rich carbides along the grain boundaries and in the Cr-rich phase were detected.
• The surface roughness was, in most cases, lower than 10 µm and it was higher along the printing direction (9.0 ± 1.6 µm) with respect to the top surface (7.9 ± 1.3 µm).

The results obtained suggested that the manufacturing of CoCrMo parts is possible using BJT with tailored properties. However, some criticalities, i.e., defects, porosity, microstructural inhomogeneity, still limit the adoption of this technology. In this context, the adoption of thermal post treatments such as Hot Isostatic Pressing (HIP) and heat treatments can reduce the porosity and homogenize the microstructure. In addition, the adjustment of the sintering parameters can also contribute to increasing the density of the parts. In this way, they could be considered potential solutions to overcoming these defects and making the entire processing chain more reliable. Further analyses can be focused on the surface functionalization of BJT parts, and the manufacturing of the thin lightweight structures usually adopted in the biomedical application fields.