Mechanical and Microstructural Characterization of the Ti-6Al-4V Alloy Processed by Additive Manufacturing for Overdenture Prosthesis

Abstract

The main objective of this work is to show the capabilities of additive manufacturing to obtain arches and overdentures from titanium alloys. Overdentures are obtained mainly by subtractive techniques in both titanium alloys and Co-Cr-Mo. Obtaining these overdentures in Ti-6Al-4V, with better biocompatibility than Co alloys, by additive manufacturing (AM), by both laser and electron beam techniques, is of increasing interest. However, adequate mechanical and microstructural characterization is necessary to bring them closer to the alloys obtained by forging and machining. Parts obtained by selective laser melting (SLM) have been developed, which show mechanical properties like those of casting and plastic deformation, although their plasticity decreases significantly. Its lamellar microstructure can be modified by thermal treatments that improve the plasticity of AM alloys, which currently present a deformation slightly lower than that required by the American Society for Testing and Materials (ASTM) F2924-2021 standard. Therefore, there is a need to improve this property through appropriate thermal treatments. Its lamellar microstructure can be modified through heat treatments that can improve the plasticity of MA alloys, which currently have a deformation slightly lower than that required by the ASTM F2924-2021 standard. Hence, there is a need to improve this property through thermal treatments.

1. Introduction

Additive manufacturing (AM) or 3D printing, is a recent technique that makes it possible to produce low-cost materials in a short time, using computer-aided design CAD, 3D Slash, 3DPrinterOS, FreeCAD, Fusion 360, or TinkerCAD prototypes that can be transformed into 3D objects, which are built layer by layer [1, 2] or scans of a real object, based on the tissue of the host [3]. In this field, there are several technologies capable of using these complex materials in engineering, medicine, aerospace, automotive, construction, and dentistry.

The American Society for Testing and Materials (ASTM) defines AM as the “process of joining materials to make parts from 3D model data, usually layer upon layer” [4]. More specifically, in the biomedical area, AM techniques have been used to create solid or porous metallic implants, through different methods of AM. The conventional manufacturing methods present some disadvantages in terms of machining, standard sizes and make it difficult to obtain porous structures that facilitate osseointegration process, which is the mechanical stability of the metal within the bone tissue.

The complex parts must efficiently serve as excellent biocompatibility, biofunctionality, good hardness, low elastic modulus, low susceptibility to corrosion, wear resistance, and fatigue strength [5]. One of their main applications has been the manufacture of personalized medical implants using computerized tomography scans of a patient’s affected region [6]. This device is quite difficult to obtain through conventional methods such as computer numerical control (CNC) machining. This is why this technique is being widely used in different areas. During fabrication, it includes various near-net shaping routes capable of manufacturing 3D geometries directly from raw materials, demanding little follow-on post-processing [7].

In addition to powders, wire-type materials can also be used as feedstock materials, and wire-based AM methods such as wire arc additive manufacturing are gaining significant attention in the additive manufacturing of (α + β) Ti6Al4V alloys [8]. The common feature of these AM processes is the use of geometrical data, which is sliced into layers with a defined thickness depending on the application. Following the sliced pattern, a focused, high-power laser or electron beam scans and melts the pioneer powders, producing a molten pool. As the heating source moves away, the molten pool cools down faster and solidifies to form a track bead. This procedure is continual until a final geometry is completed.

Newly, AM technique uses titanium alloys to produce custom-made implants for hip joints and dental prosthesis [9, 10], which is the subject of this chapter. The main AM technique used nowadays to create complex geometric implants is powder bed fusion (PBF). The PBF technique employs a high-energy source that selectively melts a bed of powdered material (preferably metallic materials) layer by layer. The PBF technique can be separated into two procedures depending on the type of energy source used [11]. Figure 1 indicates a schematic procedure of the PBF technique, which is formed by two fabrication routes: electron beam melting (EBM) and selective beam melting (SLM).

Figure 1.

Schematic procedures of selective laser melting and electron beam melting used for fabrication of complex parts by powder bed fusion technique.

The first one is called electron beam melting (EBM), which uses an electron beam, Figure 2. The EBM was proposed and commercialized first by Arcam AB in 2001 [12]. The EBM is also a powder bed fusion process, but its heating source is an electron beam instead of a laser beam. Because of the special working nature of electron beam, EBM builds parts in a high-vacuum environment of 10⁻⁴ mbar or greater, providing an ideal contamination-free environment for manufacturing reactive materials (such as titanium) that have a high affinity to oxygen and nitrogen. Additionally, this technique can generate a faster build rate due to its superior energy input and fast scan speed [13]. The high build temperature of 600–750°C also leads to different manufacturing features.

Figure 2.

Schematic procedures of electron beam melting used for fabrication of complex parts.

Besides, the EBM method offers the advantage of manufacturing prostheses with open-cell structures (lattice structures) to obtain properties like those of a bone (e.g., low elastic modulus). For instance, some works have been showed that EBM technology can produce porous of (α + β) Ti4Al6V alloy with a compression strength and a Young’s modulus compatible with those of a natural bone (1.03–17.5 GPa) [14, 15, 16, 17]. This technique uses a high-intensity energy source (a fiber laser). To fabricate the parts, a layer of material powder with a thickness varying between 0 and 100 mm is spread over the table. Then, the recoating blade distributes the powder across the build table that was preheated.

Subsequently, the laser selectively melts the powder layer based on the CAD model [18]. The parts produced by means of this technique are built in two steps. First, the contour of the part is built, and then, the powder inside the contour is melted until the complete part is obtained [19]. For the EBM process, the temperature in the molten pool was estimated to reach 2700°C [20]. Price et al. [21] conducted a specific study using thermography and measured a temperature of 2500°C in the molten pool. Al-Bermani et al. [22] calculated the cooling rate of an EBM-fabricated Ti6Al4V bead on a stainless-steel substrate and obtained a cooling rate between 103 and 105 K/s. Antonysamy et al. [23] also obtained a cooling rate in the order of 10⁴ s through simulation.

The second technique that PBF includes is selective laser melting (SLM), Figure 3, which employs a high-power laser [19]. It involves melting a powder bed using an electron beam as the energy source, and the whole process is carried out in a vacuum chamber [24]. Arcam AB, a Swiss brand founded in 1997 and later acquired by GE Additive in 2017, is currently the leader in EBM prototyping, with both companies being pioneers in this technique [12]. The SLM process started in 1995 at the Fraunhofer Institute ILT [25]. In SLM, metallic powders are uniformly spread on the building platform by rake instead of being blown out from nozzles. A focused laser beam scans the surface according to the prescribed path and selectively melts the powders in this layer, after which a new layer of powders is spread after lowering the building platform to the distance of the layer thickness. The layer height of SLM is on a scale of tens of microns, which is much thinner than that of EBM products [26]. The non-melted powders are left in the powder bed to support the subsequent layers. Powders surrounding the deposited parts are affected by the thermal process and cannot be reused due to the change in physical properties.

Figure 3.

Schematic procedures of selective laser melting used for fabrication of complex parts.

Besides being employed to manufacture customized dental implants [27], complex biomedical structures [28], and porous parts (e.g., bone substitutes) [29], this method has shown promising results in hip arthroplasty (particularly resurfacing hip arthroplasty [RHA]) [30]. However, in laser technology, materials melt while the parts are at relatively low temperatures, which may induce residual stresses that can cause the components in service to fail. To reduce the effects of excessive stress, some important factors to consider include implant’s rigidity, as well as implant–bone fit and adhesion.

Like SLM, the components fabricated using EBM are also surrounded by partly melted metallic powders, and thus, conduction dominates the heat transfer in the EBM process. Heat loss via radiation also plays a role but can be neglected as compared to conduction. Due to the highly concentrated energy source and extremely short interaction time, a high temperature and cooling rate will be produced in the molten pool.

Yadroitsev et al. [31] reported a maximum temperature of about 2710 K in the Ti6Al4V molten pool created by SLM, which also produces a cooling rate in the range of about 104–10⁶ K/s [32]. To precisely understand the thermal behavior of AM processes, the authors’ group developed a laser direct deposition model to analyze the temperature history for a three-track laser directly deposited Ti6Al4V. The model considers laser-powder interaction, mass addition, heat transfer, and fluid dynamics in the molten pool occurring in the laser direct deposition process.

Actual deposition parameters and temperature-dependent material properties were used as inputs in the model. The extracted free surface, molten pool, and heat-affected zone boundaries that are superposed on the experimental micrograph, which show an excellent agreement. The temporal and spatial temperature fields in the molten pool generated at 11 different locations along the dashed line on the micrograph are also shown. A maximum temperature of approximately 2600 K is generated in the molten pool, and it decreases to room temperature within about 0.25 s as the laser beam moves away.

The average cooling rate in the molten pool is about 10⁴ K/s, and even in the heat-affected zone, an average cooling rate of about 5 × 10³ K/s is still obtained. As the laser scan speed for this deposition is 15 mm/s, even for scanning a small layer of 1 cm², the temperature has dropped to near room temperature before the laser travels back to the same location for the subsequent layer. Stevens et al. [33] have studied binder jetting of Ti-6Al-4V alloy and showed that the sintered density at the edges and regions with significant topological curvature is lower and higher than in the other areas, respectively. The microstructural heterogeneity is ascribed to the powder-binder interactions during printing. Dilip et al. [34] have studied the feasibility of TiAl fabrication by binder jetting of Ti-6Al-4V/Al powder mixture followed by reactive sintering. They have demonstrated the formation of TiAl with various other intermediate phases.

Titanium and its alloys are used in biomedical implants due to their excellent biological performance, such as biocompatibility (non-toxic and low allergenic properties) and osseointegration [35]. It is pertinent to point out that the Ti system is complex; hence, the properties of Ti alloys greatly vary with the chemical composition, level of impurities, phase constitution, grain size, etc. [36]. Nevertheless, demands for orthopedic implants (artificial knees, hip joints, elbows, bone plates, and screws for fracture fixation), as well as dental implants (removable prostheses, maxillofacial prostheses, and supporting materials), are increasing.

According to ISO 5832 standards, 26 groups of titanium alloys have been used in biomedical applications; among them, Ti6Al4V (grade 5) is the most used [2]. This grade is heat treatable with decent combination of properties and fabricability. The remarkable biocompatibility of titanium alloys makes them ideal candidates to replace hard tissues and implant them in the body. This chapter will describe and discuss the Ti64 Gd23 fabricated by powder by fusion (PBF), as well as the influence of the process on the mechanical and electrochemical properties.

2. A brief description of Ti-6Al-4V powders and its process of fabrication by PBD and its characterization

The powder used was LPW Ti64 Gd23, supplied by LPW Technology Ltd., in accordance with the ASTM B348/B348M-19 standard [37]. Figure 4A,B indicate the morphology of the Ti64 powder and its granulometry, which presents many of it of size 40 μm with a collection of tiny satellites and some agglomerate (Figure 4A), as can be seen in the size distribution of Figure 4B. In this granulometric distribution, a small peak is observed corresponding to a small volume fraction of about 10 μm, resulting in values of d(0.1) = 25.6 μm, d(0.5) = 38.1 μm, and d(0.9) = 103.2 μm.

Figure 4.

(A) Secondary electron (SE) image of the powder used in printing components (green arrows indicate small satellites and yellow arrows irregular melts) and (B) granulometric distribution of the powder.

The test specimens have been obtained by the PBF process using a Trumpf TruPrint 1000 equipment from TRUMPF Laser- und Systemtechnik GmbH, operating with 200 W TRUMPF fiber lasers (wavelength: 1.070 nm), in an argon atmosphere and a scanning speed of 70 mm/s. The thickness of the layers has been 20 μm. Tensile specimens have been obtained in the Ti-6Al-4V ELI (Extra Low Interstitial) with PBF. The tensile tests are carried out in accordance with ASTM E8/E8M–22 [38], with the dimensions indicated in Figure 5. The compression tests have been carried out in accordance with the ASTM E9–18 standard [39] in cylinders obtained in the Z axis with a diameter of 13 mm and heights of 25 mm, with a ratio of 2.0 and 38 mm for a ratio of 3.0.

Figure 5.

Dimensions of the tensile specimen obtained in the X/Y and Z axes.

The structural analysis has been carried out using the Bruker D2Phaser X-ray equipment, using copper Kα radiation (1.5406 Å) at 30 kV and 15 mA, at a 2θ angle between 20° and 90°, with a step of 0.02° and a scanning speed of 0.0025°s⁻¹. The analysis of the results was carried out using the Rietveld method with the Materials Analysis Using Diffraction (MAUD) software version 2.9993. The microstructure was studied, after its metallographic preparation, using a NIKON LV100 optical microscope and a Zeiss Ultra 55 field emission scanning microscope (FE-SEM) equipped with X-ray spectroscopy (EDS) microanalysis from Oxford Instruments Ltda. To reveal the microstructure, it has been etching the surface polished using Kroll’s reagent (HNO₃ 6 mL, HF 3 mL, 100 mL distilled water).

Elastic modulus (E) was measured using the impulse excitation technique (IET, ATCP-Sonelastic®). The ATCP software Sonelastic 3.0 was used to analyze the data. Tensile test was performed on samples with a calibrated geometry following the ASTM standard. A Shimadzu universal testing machine model, Autograph GX Plus 100 kN was used with an optical extensometer, at a crosshead speed of 0.5 mm·min⁻¹. The compression tests were carried out on an Ibertest MEH 2000 kN universal testing machine. The hardness has been determined using a Centaur HD9-45 durometer on the HR15T superficial Rockwell scale, applying 147 N for 15 seconds. The microhardness scans were carried out in a Shimadzu HMV microhardness tester with an applied load of 2.94 N g and an application time of 15 s.

Corrosion tests were performed by potentiostat AUTOLAB PGSTAT204 with a three-electrode cell configuration (Ag/AgCl electrode, platinum wire, and the investigated alloy) at 37°C and a scanning rate [40]. The Ringer–Hartmann solution (NaCl 5.97 g/L, KCl 0.37 g/L, CaCl₂ 0.22 g/L, Na lactate 3.25 g/L, pH 6.5) was used as the electrolytes [41]. Electrochemical impedance spectroscopy (EIS) measurements, from 10 mHz to 100 kHz, were fitted by the ZView program (version 3.5a from Scribner Associates Inc., USA) for two-layer equivalent circuit simulation.

3. Physical, geometrical, and structural properties of powders and pieces obtained by PBF

The properties of the alloy obtained by additive manufacturing are comparable to those of the forging alloy, according to the ASTM F3001-14 (2021) standard [42]. The general density of the forging Ti-6Al-4V ELI alloy is 4.43 g cm⁻³. The determination of the densities obtained in the analyzed samples is shown in Table 1.

Alloy and process	Open porosity (%)	Close porosity (%)	Archimedes density (g·cm^−3)	Relative density (%)
Ti-6al-4V ELI forge	—	—	4.432 ± 0.008
Ti-6Al-4V AM cylinder piece	0.20 ± 0.08	1.96 ± 0.41	4.412 ± 0.020	98.70 ± 0.44

Table 1.

Densities obtained in Ti-6Al-4V ELI alloys from forging and additive manufacturing samples.

As shown in the table, the 3D printed material has a slightly lower density than the wrought alloy, so it can finally incorporate up to 1.3% porosity. This can affect the mechanical properties, but above all, it can also affect its plasticity and fatigue resistance. This porosity is somewhat higher than that described by Dong et al. using powder modification through mechanical alloying and printing by selective laser melting (SLM) [43].

The X-ray diffraction is similar for both the powder and the forging material and the additive manufacturing material (see Figure 6). The diffraction angles are practically the same, although with a different intensity. While for the starting powder, the β-Ti phase is practically negligible, and the highest intensities occur in the (101) α-Ti plane, in additive manufacturing, the (200) β-Ti angle and the maximum intensity peaks correspond to the α-Ti phase to the (101) (102) (110) planes. However, the forging material presents its maximum intensity in the plane (101) corresponding to the α-Ti phase, next to the plane (100) of the same phase. However, works such as those of Dong et al. [43] use SLM to observe the most intense peaks at (100) and (101), more like the peak intensity of wrought Ti-6Al-4V. However, this depends on the orientation of the laser during printing and the direction of growth of the layers. In our case, there are notable differences in the intensity of the peaks corresponding to each family of planes, which are more similar to the powder used in the process.

Figure 6.

X-ray diffraction patterns of Ti-6Al-4V ELI alloy from the powders, forging and additive manufacturing pieces.

The hardness obtained in the pieces for the tensile test manufactured in the Z-direction has presented an average of 74.1 ± 5.8 HR15T, equivalent to 98.2 ± 8.0 HV. For the specimens manufactured in the X/Y direction, the hardness obtained was 79.1 ± 1.7 HR15T equivalent to 118.0 ± 6.9 HV. However, the hardness obtained in the first layers obtained was 51.7 HR15T (43.2 HV). This macroscopic hardness contrasts with the results of the microhardness scans, expressed on the Vickers (HV) scale, which are shown in Table 2, both in its radial and longitudinal directions. Figure 7 presents the evolution of the measurements carried out along the diameter and longitudinally to the central axis of the specimens obtained in the direction of the Z axis. The average value obtained similarly in both axes is represented in solid lines, 353 ± 6 HV (longitudinal) and 347 ± 7 HV (radial), and very similar to that of the forge.

Process	Hardness HR15T	Microhardness HV	Elastic Moduli E (GPa)
Ti-6al-4V ELI forge	89 ± 1	358 ± 2 HV	110 ± 1
Ti-6Al-4V AM Z axis	74 ± 6	347 ± 7 HV	104 ± 2
Ti-6Al-4V AM X/Y axis	79 ± 2	353 ± 6 HV	90 ± 1

Table 2.

Hardness, microhardness, and elastic modulus for the forged Ti-6Al-4V alloy and that obtained by additive manufacturing.

Figure 7.

Representation of the microhardness scans, both in the radial (X/Y axes) and longitudinal (Z axis) directions.

The elastic modulus obtained by the Sonelastic® method has been carried out both in the tensile samples, whose results are collected in Table 2, as in the cylindrical samples with a height of 38 mm, whose final average is 118 ± 2 GPa. There are real differences between the wrought alloy and the printed one with lower moduli, most likely due to the microstructural modification in the form of sheets and the porosity present. However, in the cylindrical pieces, there is a greater modulus that is very similar to that found during the tensile tests, as shown in Table 3. The results finally obtained depend greatly on the printing direction but also on the method of determining the elastic modulus, while in Sonelastic, the lowest value is obtained in the X/Y direction, with 90 GPa in the tensile test this elastic modulus increases to 117 GPa. However, in the Z-direction, both methods present values of 104 GPa, although in the case of the tensile test, the standard deviation is very high, approximately 27 GPa.

Sample	Elastic modulus (GPa)	YS (MPa)	Elastic def. (%)	UTS (MPa)	Def. Max. (%)
Ti-6Al-4V AM (X/Y axis)	117 ± 9	942 ± 38	1 ± 0	1020 ± 46	1 ± 0
Ti-6Al-4V AM (Z axis)	104 ± 27	974 ± 27	1 ± 0	1100 ± 17	6 ± 1

Table 3.

Results obtained in the tensile tests in the manufacturing X/Y and Z positions.

The tensile test was carried out in both the X/Y and Z positions. Table 3 shows the results obtained in the X/Y position, in which the tensile elastic modulus is expressed in GPa, the limit elastic in MPa, the elastic deformation in %, the maximum resistance in MPa, and the maximum deformation in %. In this test, the deformation has been determined using a high-precision optical extensometer, so the final deformation corresponds perfectly to the real deformation.

As seen in Table 3, the mechanical properties are similar in the different directions, but the elasticity in this case is much higher, with deformation averages of 6% for position Z compared to 1% for position X/Y. Figure 8 indicates the curves of stress–strain of the samples manufactured in both directions with a clear difference in the plastic behavior of the material. UTS is slightly lower than those obtained by Rafi et al. [44] of 1219 ± 20 MPa or the 1246 ± 143 MPa obtained by Wysocki et al. [26] and a lower yield strength. The measurement of the elongation (%) at the end of the breakage of the specimens, although with less precision than extensometer, has provided comparable results in any case. For the samples processed in X/Y, the average was 3 ± 0.4%, while, for the samples processed in Z, the average was 7 ± 0.4%. Considering the characteristics required by the ASTM F3001-14 standard [43], for class F, it more than meets the resistant properties, but not the elongation in the Z-direction. However, the maximum deformation reported by Wysocki et al. [26] is 3.2 ± 0.5% for the X/Y direction and 1.4 ± 0.5 in the Z-direction. In our tests, optical extensometry yields a plastic strain in the Z direction of 6 ± 1% higher than that reported by Wysocki et al. and of the order of that indicated by Rafi et al. [44] of 4.9 ± 0.6% in the built vertical direction. Much of these strain values are collected by Ngyuyen et al. [45] in their review carried out on additive manufacturing of Ti-6Al-4V alloy.

Figure 8.

Stress–strain diagram for the additive manufacturing samples in directions X/Y and Z.

This lack of plastic deformation is mainly due to the defects that can occur between the laser fusion layers, as can be seen in Figures 9A,B and 10A,B, in which a lack of fusion is observed in some areas.

Figure 9.

Fractographies of the tensile fracture of the sample obtained by additive manufacturing in the X/Y direction. (A) General appearance of the fracture. (B) Detail of ductility with the formation of dimples.

Figure 10.

Fractographies of the tensile fracture of the sample obtained by additive manufacturing in the Z-direction. (A) General appearance of the fracture. (B) Detail of the formation of ductile failure dimples.

The lack of plastic deformation is mainly due to defects that can occur between the layers fused by the laser. In Figure 11A,B, some of the most defined trajectories of the superposition of the fusion layers are represented with a dashed line, although they do not correspond to each one of them as they are much thinner.

Figure 11.

(A) Infused material and cracking between layers in the X/Y direction. (B) Indications of crack paths between fusion layers in the X/Y direction.

The results obtained in the compression tests of the additive manufacturing samples, with a D/H ratio of 3.0, present a yield strength of 838 ± 85 MPa, with an elastic deformation of 1.95 ± 0.07%, somewhat higher than those obtained with traction. However, it has not been possible to determine the ultimate compressive strength (UTS), but the stress when deformation was 0.03 mm/mm, corresponding to a displacement of 1.14 mm of 953 ± 54 MPa. For the samples with D/H ratio = 2.0, a yield strength of 925 MPa, an elastic deformation of 2.4%, and a stress when the deformation was 3%, corresponding to a displacement of 0.75 mm of 1004 MPa, have been obtained.

The microstructure observed by optical microscopy, after etching with the Kroll solution, is shown in Figure 12. For the samples manufactured in the X/Y direction, in which some important defects originating from the fusion of the layers are observed. Mainly infused particles and small pores located between the fusion layers. Figure 13 shows the microstructure of the printed samples in the Z-direction. This clearly shows the formation of the different fusion layers and how the lamellar microstructure originates through the layers, creating large grains corresponding to the β phase generated at high temperature.

Figure 12.

Optical microscopy image of the microstructure of the Ti-6Al-4V alloy, obtained by additive manufacturing and printed in the X/Y direction.

Figure 13.

Optical microscopy image of the microstructure of the Ti-6Al-4V alloy, obtained by additive manufacturing printed in the Z-direction.

However, the microstructure is much better appreciated after immersion corrosion tests. Figure 14A–D shows an important degradation located mainly in the α-Ti phase, especially when the magnification is higher. The microstructure of the formed alloy is mainly composed of equiaxed α phase grains with randomly positioned fine β phase grains. In the additive manufacturing samples, the formation of lamellar α + β grains is observed, with a strong presence of α phase at the grain boundary. This different distribution of the phases is what differentiates the behavior of the alloy depending on its production process. These microstructural differences are also observed when their microstructure and crystalline orientation are analyzed by backscattered electron diffraction. Figure 15A–D shows the far scattered diffraction (FSD) image. Figure 15A represents the band contrast, showing the morphology of the grains and the phase contrast. The distribution of the phases is indicated in Figure 15B. The inverse pole figure (IPF) is indicated in Figure 15C. In Figure 15D shows a slight texture in the (1010) plane. However, in AM Ti-6Al-4V, Figure 16A–D, at the magnification used, the size of the β phase grains obtained during solidification is not appreciated, and only the distribution of α phase sheets are appreciated, in which observe the formation of these sheets at angles of 60° or 120°, forming a Widmanstätten microstructure. Although a defined texture cannot be seen clearly, the orientation of these sheets bears no resemblance to the alloy obtained by forging. Therefore, the resistance and plasticity properties can be different between both types of alloys. This microstructure is more like that obtained by EBM [26, 44] than that obtained by these same authors by SLM, whose main microstructure is α’ martensite.

Figure 14.

Microstructure after immersion corrosion in Ringer-Hartmann electrolyte: (A) and (B) forged Ti-6Al-4V, (C) and (D) AM Ti-6Al-4V.

Figure 15.

Microstructure of forging Ti-6Al-4V ELI: (A) far scattered diffraction image. (B) Map of phase distribution (red represents the α-phase and the blue represents the β-phase). (C) Inverse pole figure map. (D) Pole figure in plans (0001) and (1010).

Figure 16.

Microstructure of AM Ti-6Al-4V ELI: (A) far scattered diffraction image. (B) Map of phase distribution. (C) Inverse pole figure map. (D) Pole figure in plans (0001) and (1010).

Corrosion resistance also presents differences between wrought and AM alloys. The potentiodynamic polarization curves, Figure 17, show great similarity between the CP-Ti samples and the wrought Ti-6Al-4V ElI alloy. However, the response of the additive manufacturing alloy, even presenting a more anodic potential and a lower passivation intensity than the two standard materials, shows a breakdown at a potential of 1.2 V and significantly increasing passivation intensity up to the end of the test. This could indicate that the rupture of the TiO₂ oxide layer initially formed. Although it could subsequently form again, limiting the current flow. The results of the corrosion parameters are shown in Table 4, which summarizes the open circuit potential (E_OCP), with the corrosion potential (E_corr), the current density (i_corr) obtained by the Tafel extrapolation method [46], the anodic (b_an) and cathodic (b_cat) Tafel constants, the polarization resistance (R_p), and finally the corrosion rate (C_r) expressed in μm·year⁻¹.

Figure 17.

Potentiodynamic polarization curves of the Ti-6Al-4V AM alloy in Ringer–Hartmann artificial saliva solution, and of CP-Ti and the wrought alloy.

Sample	EOCP (V)	Ecorr (V)	icorr (A·cm^−2)	ban	bcat	Rp (kΩ·cm^2)	Cr (μm·year^−1)
Ti-6Al-4V AM	−0.08	−0.03	1.70 × 10^−7	0.12	0.12	361	1.68
CP-Ti	−0.24	−0.25	5.37 × 10^−8	0.10	0.11	109	4.59
wrought Ti-6Al-4V ELI	−0.02	−0.18	1.47 × 10^−7	0.11	0.12	231	1.25

Table 4.

Kinetics parameters obtained from the potentiodynamic polarization curves in the Ti-6Al-4V AM alloy, CP-Ti and the wrought Ti-6Al-4V ELI alloy.

The Nyquist diagrams of the experiments carried out on Ti-6Al-4V AM and wrought Ti-6Al-4V ELI, together with the Ti CP, are shown in Figure 18. These diagrams have a semicircular arc shape, characteristic of passive metals, with important differences in them, finding the behavior of the AM alloy superior to CP-Ti but slightly inferior to wrought Ti-6Al-4V ELI alloy due to the greater length of the semicircle of this last alloy is associated with a greater resistance R_p.

Figure 18.

Nyquist diagrams of the Ti-6Al-4V AM alloy in Ringer–Hartmann artificial saliva solution, and of CP-Ti and wrought Ti-6Al-4V ELI alloy.

In the Bode diagrams, Figure 19, a similar behavior also occurs with a phase angle close to 80°, which indicates a mainly capacitive behavior, although especially at low and medium frequencies. The impedance modulus is practically the same for the three materials, but the phase angle is slightly lower in Ti-6Al-4V AM, which would indicate a lower resistance, although it is quite like CP-Ti.

Figure 19.

Bode diagrams (modulus and phase) as a function of frequency for Ti-6Al-4V AM, CP-Ti, and wrought Ti-6Al-4V ELI alloy.

The results were compared with an equivalent double-layer circuit, using ZView software (version 3.5f), considering the non-ideal behavior with the electrolyte resistance (R_s), charge transfer resistance (R_ct), film resistance (R_film), and double-layer polarization resistance (CPE_dl,) and film’s (CPE_film). This circuit has been used by other authors, representing the equivalent circuit that best adapts to the composite corrosion mechanism [47]. In the same way, we can analyze the CPE exponents (n_dl and n_film) related to the capacities of the dense layer (C_dl) and the porous layer (C_film) [48]. These results are shown in Table 5 for the three alloys analyzed, Ti-6Al-4V AM, CP-Ti, and wrought Ti-6Al-4V ELI.

Sample	Rs (Ω·cm^2)	CPEdl (μF/cm^2)	ndl	Rct (kΩ·cm^2)	CPEfilm (μF/cm^2)	nfilm	Rfilm (kΩ·cm^2)
Ti-6Al-4V AM	46	26	0.85	146	17	0.38	1100
CP-Ti	89	21	0.94	0.13	9	0.93	1380
Wrought Ti-6Al-4V	100	54	0.51	8.06	16	0.91	1190

Table 5.

Parameters obtained by electrochemical impedance spectrometry (EIS) in Ti-6Al-4V AM, Ti CP and wrought Ti-6Al-4V ELI alloys.

The values obtained for R_s are similar in all samples, also considering the obstacle of standardizing the resistivity of the electrolyte in the different tests. The charge transfer resistance is always lower than that of the film (R_film), which is related to the internal oxide layer as it is much more protective than the porous external layer [49]. However, the n_dl and n_film exponents were approximately 1 for CP-Ti. For the Ti-6Al-4V alloy, the n_dl exponent was 0.51 for the wrought alloy and the n_film of 0.38 for the AM alloy. Besides that, together with higher R_p, it ensures better behavior against corrosion compared to the CP-Ti. The fits obtained had a χ² in the order of 10⁻³, which would indicate that the selected equivalent circuit is representative of the oxide layers formed in these alloys [49].

4. Discussion

Additive manufacturing offers excellent opportunities in the biomedical sector [50]. The selective laser melting (SLM) or direct metal laser sintering (DMLS) technology also offers great versatility and greater ease than other powder bed fusion (PBF) processes, such as electron beam melting (EBM) [51]. The obstacle of machining titanium alloys, as well as the residual stresses that they can accumulate, makes AM a perfectly applicable technology in different sectors mainly in the biomedical sector [52], where the customization of products is found. This is mostly justified, mainly in maxillofacial surgery or implant-supported prostheses [53]. The microstructure formed is basically laminar in practically all cases, with the formation of α phase at the grain boundary of solidification β phase. These β grains are large as they remain at high temperatures for some time in PBF processes [54]. From the grain boundary, the formation of the α phase begins in sheets with a Widmästaetten-type microstructure, Figure 14. However, the crystal structure is only modified by the cooling direction of the sample, which in our case enhances the (101) (102) and (110) planes, Figure 6.

The formation of the microstructure in PBF processes can originate, especially in laser-assisted processes, SLM or DMLS, metallurgical transformations to the martensite α’. Although this phase presents laminar α phase (HCP) microstructure, the mechanical properties obtained after the AM process are very similar to those presented by wrought Ti-6l-4V ELI alloy. However, the main difference is the ductility of the material, since it is decreased compared to the commercial forged alloy. However, it depends on the direction of the effort depending on the growth of the layers, Table 3. Similar results have been obtained in the works of Shunmugavel et al. [55], who, after a vacuum thermal treatment at 750°C obtained ductility properties that were also lower in the transverse direction, although in this case somewhat higher than those obtained after AM. In addition, a preferential formation of the primary β-phase can be available, as analyzed by Wei et al. using multi-laser and analyzing the mechanical properties of overlap regions. The transformation to α’ phase is evident from the energy in the melting and cooling of the different layers [56]. The effect of energy in the process is analyzed by Cepeda-Jiménez et al. with a columnar formation of the β-phase prior to the α-α’ transformation [57]. The application of a sub-β-transus heat treatment, the α phase is stabilized, with somewhat lower frequency at misorientation angles of 60°. Infusion defects are analyzed by Andani et al., who evaluate the spatter formation in the SLM process using multi-laser technology and conclude that most of the particle shape of the spatter is spherical and that this can significantly affect the fatigue properties of the components [58]. In fact, when determining their mechanical properties, a decreased elongation is present, especially when two-laser melting is used.

The microstructure conditions formed obviously affect the mechanical properties, which are also affected by external parameters such as surface roughness and possible internal defects in the material. These mechanical properties depend critically on the posterior heat treatment used. Although in all cases, the resistance decreases slightly, as the formation of the phase α phase stabilizes. At the same time, it is true that the longitudinal and transversal properties are getting closer together. Nevertheless, the biggest influence is on ductility, which decreases greatly when the load is applied perpendicular to the melting of the layers, i.e., in the Z-direction. Therefore, the stability of the α phase, simultaneously with the growth of the thickness of the α phase lamellae, promotes an increase in the ductility of the component [58], especially in the Z-direction. Rafi et al. [59] report only 5% elongations in both directions after heat treatment at 650°C for 4 hours Ar. Simonelli, Tse, and Tuck [60] obtain elongations higher than 10% with a treatment with N₂ at 730°C for 2 hours, while with an annealing treatment at 843°C for 2 hours and furnace cooling, Brandl, Leyens, and Palm [61] obtain elongations of 17% in the longitudinal direction and 15% in the transverse direction. However, it is true that in this case, the yield strength is 810–840 MPa and UTS between 860 and 920 MPa. Zhang et al. [11] obtain elongations higher than 20% when, after obtaining by L-PBF a phase distribution α’ + (α + β), they subject the material to a sub-β-transus heat treatment at 950°C for 2 h, with subsequent cooling in air or in the furnace. However, one of the fundamental problems of AM is the lower reliability of the product due to the possible internal defects that can occur in the material. The differential melting at some points of the different layers usually generates irregular pores, Figure 11, or by microporosities derived from Ar absorption, although in this case, the pores are micrometric in size and rounded.

The elastic modulus obtained by impulse excitation technique in the cylindrical sheets is like that determined in the tensile tests in the X/Y direction, around 117 GPa, while in the Z-direction, it decreases to 104 GPa. The yield strength and maximum load exceed the minimum established by the ASTM standard, and the hardening index (YS/UTS) is 0.92 in the X/Y direction. For the Z-direction, use the YS/UTS. The microhardness does not present appreciable differences in both directions and with the wrought alloy, less than 2%.

However, the greatest difference is found in the plastic deformation where in the Z-direction, an elongation greater than 6% is obtained, as indicated in the ASTM F3001-14 standard. However, in the X/Y direction, a loss of 79% of the elongation with respect to Z is obtained, which is being far from the minimum indicated in the standard. This effect is seen in fractography where the influence on the fracture of the orientation of the layers and the influence of internal defects in the material are observed. Therefore, one of the biggest differences with wrought alloy is the greater difference in the results, which is expressed by high standard deviations, somewhat higher in the X/Y direction than in the Z-direction. Surface roughness also has an important effect on the resistant properties of the parts, which, jointly with possible internal defects, provide higher standard deviations than in wrought parts. This is one of the limitations of the technique in biomedical and especially aeronautical applications. A preliminary determination of the roughness obtained, using white light profilometry, has provided a R_a value of around 48 μm. This roughness and possible internal defects play a fundamental role in fatigue resistance and, hence, there is need to determine the number, shape, and size of these defects [62] and try to minimize them through thermal post-treatments or by hot isostatic pressing (HIP). Thermal treatments, as already indicated, have a secondary effect on grain growth in the β phase and thickening of the microstructural lamellae in the α phase, but they can improve the ductility, which is necessary in these cases, by closing the internal pores, as gas porosity and other small fusion defects. Leuders et al. [63, 64] analyze the effect of ductility on the fatigue resistance of Ti-6Al-4V processed by SLM. In this sense, it is essential to evaluate the differences in microstructure in both static and dynamic properties. Greitemeier et al. [65] compare the fatigue resistance between electron beam melting (EBM) and direct metal laser sintering (DMLS), reducing the surface roughness, which possibly increases the fatigue resistance in both cases, with values close to 550 MPa for DMLS and 260 MPa for EBM, after 10⁷ cycles.

The chemical and electrochemical behavior is also similar between the alloys obtained in both processes, wrought and AM. In immersion tests, small pits appear mainly in the α phase. On the other hand, due to the different equiaxial microstructure grains in one case and laminar in AM, the contact surface between the α and β phases is higher in AM, and therefore, the release of Ti ions is slightly higher. However, in its electrochemical behavior, the AM alloy has a lower E_corr than both the wrought alloy and CP-Ti, with great similarity in terms of i_corr and resistance to polarization R_p. Despite this, while wrought alloys present passivation potentials that extend beyond 3 V, the AM alloy presents a breakdown potential at 1.2 V, although it subsequently passivates again with higher passivation intensity values. In any case, a similar passive TiO₂ layer is developed in all cases that must be more widely studied. In any case, the potentiodynamic tests present variability in the three samples tested for each material. Thus, the value of the corrosion ratio is very similar for the two alloys Ti-6Al-4V ELI wrought and AM. In the electrochemical impedance spectrometry tests, it is observed that in the Niquist diagrams, the AM alloy is between the CP-Ti and the wrought alloy, but with similar behaviors. In the Bode diagrams, all the alloys show similar behaviors. The phase angles are around 80°, although in the case of the wrought alloy, it moves to low frequencies. With the results, an equivalent double-layer circuit has been obtained, the internal compact and the external porous, also used by other authors [44, 49, 66, 67, 68] with varying success. With the circuit and once the different resistive and capacitive components have been obtained, it is important to determine the CPE exponent, since an exponent greater than 0.8 confers a merely capacitive character and with values less than or close to 0.5 a diffusive character. As seen in the results table, very similar values are presented between the three alloys studied, and the AM alloy, it has a somewhat higher charge transfer resistance R_ct, although much lower than the resistance R_film, where the exponent n_film remains at 0.38.

5. Conclusions

The additively manufactured Ti-6Al-4V alloy has mechanical properties like the Ti-6Al-4V ELI wrought alloy. Its elastic modulus (104–90 GPa compared to 110 GPa for Ti-6Al-4V ELI), microhardness (347–353 HV compared to 358 HV for wrought alloy), yield strength (943–974 MPa), and maximum tensile strength (1020–1100 MPa) differ very little and, after some heat treatments, are even higher in the AM alloy.

Ductility, however, is limited and although the standards reduce the minimum ductility for AM alloys, these are not achieved when manufacturing is carried out in some specific directions, X/Y, since the maximum deformation is only 1.5%.

The microstructure is completely different, and the AM is very similar to that of casting, as it develops large β phase grains that, in their allotropic transformation like the Widmanstätten type α phase, but far from the microstructure of equiaxial grains in the wrought mill annealing alloy.

The X-ray diffraction indicates a distribution of similar phases in wrought and AM alloys, but with differences in textures that depend on the direction of growth of the layers but do not affect the mechanical properties except ductility.

The corrosion resistance is similar between AM and wrought Ti-6Al-4V alloy with very similar behavior and a similar corrosion ratio (1.68 μm·year⁻¹ for AM alloy compared to 1.25 μm·year⁻¹ for wrought alloy).

Therefore, no significant differences have been found in the behavior of AM alloys obtained from 3D printing, except for their microstructure that can hardly be modified, and the ductility derived from the manufacturing direction, internal defects, and surface roughness mainly. This makes the technique widely used in the manufacture of personalized parts in the biomedical sector, and possibly through subsequent heat treatments or hot isostatic processes, the reliability of the material is improved. In that case, its application in both overdentures and maxillofacial surgery, which require customization for each patient, can offer important advantages.