2.1

Alloys and selective laser melting (SLM)

Three different SLM processing strategies were employed to fabricate fully dense (a density of approx. 8.3 g cm ⁻³ ) CoCrMo alloys (1 cm × 1 cm × 0.1 cm) from a commercial alloy powder (CobaltChrome MP1™, EOS GmbH, Germany). The powder size was 25 μm (5–40 μm) based on scanning electron microscopy measurements. The nominal bulk alloy composition of the powder is given in Table 1 . Based on supplier information, this powder is recommended for dental applications and has a reported Young’s modulus of 200–210 GPa, a tensile strength of 1150–1400 MPa, and a yield strength of 880–980 MPa when standard SLM strategies for fully dense materials are employed .

Three SLM processing strategies were applied using an EOS M270 apparatus: (i) L20, a layer thickness of 0.02 mm, scanning parallel traces in each layer; (ii) R20, layer thickness of 0.02 mm, scanning traces rotated at an angle of 60° every second layer, and (iii) R40, layer thickness of 0.04 mm, scanning traces rotated at an angle of 60° every second layer, allowing a fast formation speed and yielding a increased surface roughness compared with R20 and L20. R20 is the strategy for an optimal surface, and R40 for an optimal and rapid manufacture of a fully dense alloy. Commercially available SLM manufactured dental alloys differ in laser strategies due to business secrets. Detailed SLM process parameters for this study are given in Table 2 .

A cast biomedical grade high carbon CoCrMo alloy (ASTM F75 ) intended for orthopedic implants, denoted cast (Ø 14 mm, thickness 2 mm) was investigated for comparative reasons. Its nominal bulk composition is given in Table 1 , and its corrosion properties and microstructure described elsewhere . After casting, the alloy was processed by hot isostatic pressing (1200 °C, 4 h, 103 MPa, slow cooling) and subsequently heat-treated (1200 °C, 4 h, fast cooled to below 800 °C). The presence of M ₂₃ C ₆ (M = metal) carbides in the grain boundaries of the cast CoCrMo alloy and microstructural features of main importance for corrosion and dissolution are reported elsewhere .

2.2

Microstructural investigation

The microstructure was investigated on a macro-scale using optical microscopy (Olympus SZX12; reflective mode), and on a micro-scale by means of scanning electron microscopy (SEM, JEOL JSM-7000F). The SLM fabricated alloys were mechanically polished and further chemically etched (immersion etching) in 36% HCl (in water) for 8 h prior to the SEM investigation. Specimens for TEM characterization were prepared by either electrolytic polishing at room temperature using a TenuPol-5 electrolytic polishing system (90% acetic acid and 10% perchloric acid; 25 V), or by argon ion beam milling with an accelerating voltage of 6 keV. Transmission electron microscopy (TEM, JEOL FEG-2100F) with energy dispersive spectroscopy (EDS) was employed to investigate differences in composition with high lateral resolution (<1 μm) for the alloy R20. X ray powder diffraction (Panalytical X’Pert PRO) in reflection mode was used to obtain crystallographic phase information. Diffractograms were recorded from 20° to 120° (2 θ ) using monochromatic Cu Kα ( λ = 1.5405981 Å) radiation (40 mA, 45 kV) and step sizes of 0.013° (polished SLM samples) and 0.026° (unpolished SLM samples), respectively.

2.3

Electrochemical properties and surface analysis

Corrosion susceptibility was investigated for all samples by employing electrochemical potentiodynamic polarization using a microcapillary set-up (glass capillary, Ø 159 μm, filled with 1 M NaCl). The advantage of micro-electrochemistry compared with conventional macroscopic electrochemistry is that micron-sized inhomogeneities within the material can be detected. Measurements were performed on at least three different surface areas for each material. The microcell is described in detail elsewhere . In short, the auxiliary electrode (platinum wire) is located inside the glass capillary with a reference electrode (Ag/AgCl (3 M KCl)) connected via an electrolytic bridge. The tip of the glass capillary is polished and a silicone rubber O-ring is used to seal between the capillary and the sample surface. A high resolution potentiostat (Jaissle IMP 83 PC T-BC) with a current detection limit of 10 fA was used with a sampling rate of 9 Hz. The capillary was positioned in an objective holder to enable the microscopic investigation (Leica DM LS2, lateral resolution approximately 1 μm) of the same area. One day prior to the measurements the samples were polished to 1 μm using diamond paste, ultrasonically cleaned in ethanol for 5 min, dried in air using a fan and stored in air at ambient conditions. This pre-treatment and aging process were made for all samples in the same way to allow an equal growth of the native surface oxide. The start potential was set to −600 mV (Ag/AgCl), and the samples were anodically polarized at 1 mV s ⁻¹ to 1000 mV, followed by cathodic polarization back to −600 mV. Before applying the potential, the system was allowed to equilibrate for 10 ± 5 min. After polarization, compositional surface analysis (Co, Cr, O, Fe, C, Na, Cl (Mo not included)) was made on a corroded area (250 μm × 200 μm) of one of the samples, the L20 alloy, using scanning Auger electron microscopy (SAM, Physical Electronics (PHI) 670, 10 kV). The main aim of this investigation was to identify enriched or depleted elements as a result of corrosion.

2.4

Metal release (biocompatibility) investigations

Prior to exposure, the samples were ground using 1200 grit SiC, ultrasonically cleaned in acetone and isopropyl alcohol for 7 min, respectively, dried with cold nitrogen gas, and aged (desiccator, room temperature) for 24 ± 1 h to achieve a defined and stable surface oxide, as described elsewhere . The samples were entirely immersed in a vertical position (a surface area to solution volume ratio of 1 cm ² mL ⁻¹ ) in phosphate buffered saline (PBS, pH 7.3 ± 0.1), or PBS + 10 g L ⁻¹ bovine serum albumin (PBS + BSA, pH 7.3 ± 0.1), in closed acid-cleaned polyethylene vessels for 2 and 168 h (1 week), respectively. The vessels were positioned in a Stuart platform-rocker incubator with bilinear shaking (25 cycles min ⁻¹ ) at 37 ± 0.5 °C. Triplicate samples and one blank sample (without alloy sample) were exposed in parallel for each time period and fluid. The chemical composition of PBS was 8.77 g L ⁻¹ NaCl, 1.28 g L ⁻¹ Na ₂ HPO ₄ , 1.36 g L ⁻¹ KH ₂ PO ₄ , 370 μL L ⁻¹ 50% NaOH and ultrapure water (18.2 MΩ cm). PBS of slightly varying compositions is widely used in the scientific literature as a physiological reference solution . The fluids were justified from their relevance for human blood, and previous deliberations and findings by some of the authors on protein adsorption and protein-induced metal release on chromium metal, stainless steel, and cast CoCrMo . After exposure, the solution was separated from the samples, and acidified (65 vol.% ultrapure nitric acid) to a pH < 2. All chemicals were of at least analytical grade (Sigma–Aldrich), and all equipment and vessels were acid-cleaned by 10 vol.% nitric acid for 24 h before rinsed four times with ultrapure water. Solution metal analysis (total metal) was performed using atomic absorption spectroscopy (AAnalyst 800, Perkin Elmer) with flame (for Co) and graphite furnace (for Co, Cr and Mo) modes. Calibration curves were based on at least four calibration standards and quality control samples of known concentration were analyzed regularly. Limits of detection (LOD, 3 times the highest standard deviation of blank samples) were 1 μg L ⁻¹ , for Co, Cr, and Mo, respectively, in the graphite mode. Measured metal concentrations were significantly higher than their corresponding limits of detection for all time periods, exposure conditions and analyzed metals.

2.5

Statistical significance test method

A Student t -test of unpaired data with unequal variance was employed between two different data sets to identify the significance of observed differences. Differences are counted as significant when p < 0.05, with higher significance for a smaller p -value.

2.1

Alloys and selective laser melting (SLM)

Three different SLM processing strategies were employed to fabricate fully dense (a density of approx. 8.3 g cm ⁻³ ) CoCrMo alloys (1 cm × 1 cm × 0.1 cm) from a commercial alloy powder (CobaltChrome MP1™, EOS GmbH, Germany). The powder size was 25 μm (5–40 μm) based on scanning electron microscopy measurements. The nominal bulk alloy composition of the powder is given in Table 1 . Based on supplier information, this powder is recommended for dental applications and has a reported Young’s modulus of 200–210 GPa, a tensile strength of 1150–1400 MPa, and a yield strength of 880–980 MPa when standard SLM strategies for fully dense materials are employed .

Three SLM processing strategies were applied using an EOS M270 apparatus: (i) L20, a layer thickness of 0.02 mm, scanning parallel traces in each layer; (ii) R20, layer thickness of 0.02 mm, scanning traces rotated at an angle of 60° every second layer, and (iii) R40, layer thickness of 0.04 mm, scanning traces rotated at an angle of 60° every second layer, allowing a fast formation speed and yielding a increased surface roughness compared with R20 and L20. R20 is the strategy for an optimal surface, and R40 for an optimal and rapid manufacture of a fully dense alloy. Commercially available SLM manufactured dental alloys differ in laser strategies due to business secrets. Detailed SLM process parameters for this study are given in Table 2 .

A cast biomedical grade high carbon CoCrMo alloy (ASTM F75 ) intended for orthopedic implants, denoted cast (Ø 14 mm, thickness 2 mm) was investigated for comparative reasons. Its nominal bulk composition is given in Table 1 , and its corrosion properties and microstructure described elsewhere . After casting, the alloy was processed by hot isostatic pressing (1200 °C, 4 h, 103 MPa, slow cooling) and subsequently heat-treated (1200 °C, 4 h, fast cooled to below 800 °C). The presence of M ₂₃ C ₆ (M = metal) carbides in the grain boundaries of the cast CoCrMo alloy and microstructural features of main importance for corrosion and dissolution are reported elsewhere .

2.2

Microstructural investigation

The microstructure was investigated on a macro-scale using optical microscopy (Olympus SZX12; reflective mode), and on a micro-scale by means of scanning electron microscopy (SEM, JEOL JSM-7000F). The SLM fabricated alloys were mechanically polished and further chemically etched (immersion etching) in 36% HCl (in water) for 8 h prior to the SEM investigation. Specimens for TEM characterization were prepared by either electrolytic polishing at room temperature using a TenuPol-5 electrolytic polishing system (90% acetic acid and 10% perchloric acid; 25 V), or by argon ion beam milling with an accelerating voltage of 6 keV. Transmission electron microscopy (TEM, JEOL FEG-2100F) with energy dispersive spectroscopy (EDS) was employed to investigate differences in composition with high lateral resolution (<1 μm) for the alloy R20. X ray powder diffraction (Panalytical X’Pert PRO) in reflection mode was used to obtain crystallographic phase information. Diffractograms were recorded from 20° to 120° (2 θ ) using monochromatic Cu Kα ( λ = 1.5405981 Å) radiation (40 mA, 45 kV) and step sizes of 0.013° (polished SLM samples) and 0.026° (unpolished SLM samples), respectively.

2.3

Electrochemical properties and surface analysis

Corrosion susceptibility was investigated for all samples by employing electrochemical potentiodynamic polarization using a microcapillary set-up (glass capillary, Ø 159 μm, filled with 1 M NaCl). The advantage of micro-electrochemistry compared with conventional macroscopic electrochemistry is that micron-sized inhomogeneities within the material can be detected. Measurements were performed on at least three different surface areas for each material. The microcell is described in detail elsewhere . In short, the auxiliary electrode (platinum wire) is located inside the glass capillary with a reference electrode (Ag/AgCl (3 M KCl)) connected via an electrolytic bridge. The tip of the glass capillary is polished and a silicone rubber O-ring is used to seal between the capillary and the sample surface. A high resolution potentiostat (Jaissle IMP 83 PC T-BC) with a current detection limit of 10 fA was used with a sampling rate of 9 Hz. The capillary was positioned in an objective holder to enable the microscopic investigation (Leica DM LS2, lateral resolution approximately 1 μm) of the same area. One day prior to the measurements the samples were polished to 1 μm using diamond paste, ultrasonically cleaned in ethanol for 5 min, dried in air using a fan and stored in air at ambient conditions. This pre-treatment and aging process were made for all samples in the same way to allow an equal growth of the native surface oxide. The start potential was set to −600 mV (Ag/AgCl), and the samples were anodically polarized at 1 mV s ⁻¹ to 1000 mV, followed by cathodic polarization back to −600 mV. Before applying the potential, the system was allowed to equilibrate for 10 ± 5 min. After polarization, compositional surface analysis (Co, Cr, O, Fe, C, Na, Cl (Mo not included)) was made on a corroded area (250 μm × 200 μm) of one of the samples, the L20 alloy, using scanning Auger electron microscopy (SAM, Physical Electronics (PHI) 670, 10 kV). The main aim of this investigation was to identify enriched or depleted elements as a result of corrosion.

2.4

Metal release (biocompatibility) investigations

Prior to exposure, the samples were ground using 1200 grit SiC, ultrasonically cleaned in acetone and isopropyl alcohol for 7 min, respectively, dried with cold nitrogen gas, and aged (desiccator, room temperature) for 24 ± 1 h to achieve a defined and stable surface oxide, as described elsewhere . The samples were entirely immersed in a vertical position (a surface area to solution volume ratio of 1 cm ² mL ⁻¹ ) in phosphate buffered saline (PBS, pH 7.3 ± 0.1), or PBS + 10 g L ⁻¹ bovine serum albumin (PBS + BSA, pH 7.3 ± 0.1), in closed acid-cleaned polyethylene vessels for 2 and 168 h (1 week), respectively. The vessels were positioned in a Stuart platform-rocker incubator with bilinear shaking (25 cycles min ⁻¹ ) at 37 ± 0.5 °C. Triplicate samples and one blank sample (without alloy sample) were exposed in parallel for each time period and fluid. The chemical composition of PBS was 8.77 g L ⁻¹ NaCl, 1.28 g L ⁻¹ Na ₂ HPO ₄ , 1.36 g L ⁻¹ KH ₂ PO ₄ , 370 μL L ⁻¹ 50% NaOH and ultrapure water (18.2 MΩ cm). PBS of slightly varying compositions is widely used in the scientific literature as a physiological reference solution . The fluids were justified from their relevance for human blood, and previous deliberations and findings by some of the authors on protein adsorption and protein-induced metal release on chromium metal, stainless steel, and cast CoCrMo . After exposure, the solution was separated from the samples, and acidified (65 vol.% ultrapure nitric acid) to a pH < 2. All chemicals were of at least analytical grade (Sigma–Aldrich), and all equipment and vessels were acid-cleaned by 10 vol.% nitric acid for 24 h before rinsed four times with ultrapure water. Solution metal analysis (total metal) was performed using atomic absorption spectroscopy (AAnalyst 800, Perkin Elmer) with flame (for Co) and graphite furnace (for Co, Cr and Mo) modes. Calibration curves were based on at least four calibration standards and quality control samples of known concentration were analyzed regularly. Limits of detection (LOD, 3 times the highest standard deviation of blank samples) were 1 μg L ⁻¹ , for Co, Cr, and Mo, respectively, in the graphite mode. Measured metal concentrations were significantly higher than their corresponding limits of detection for all time periods, exposure conditions and analyzed metals.

2.5

Statistical significance test method

A Student t -test of unpaired data with unequal variance was employed between two different data sets to identify the significance of observed differences. Differences are counted as significant when p < 0.05, with higher significance for a smaller p -value.