Materials Science and Engineering C 49 (2015) 517–525
Contents lists available at ScienceDirect
Materials Science and Engineering C journal homepage: www.elsevier.com/locate/msec
Investigation on the microstructure, mechanical property and corrosion behavior of the selective laser melted CoCrW alloy for dental application Yanjin Lu a, Songquan Wu a, Yiliang Gan a, Junlei Li b, Chaoqian Zhao a,c, Dongxian Zhuo a, Jinxin Lin a,⁎ a b c
Fujian Institute of Research on the Structure of Matter, Chinese Academy of Sciences, 155 Yangqiao Road West, Fuzhou, China Institute of Metal Research, Chinese Academy of Sciences, 72 Wenhua Road, Shenyang, China College Materials Science and Engineering, Fujian Normal University, Fuzhou, China
a r t i c l e
i n f o
Article history: Received 25 September 2014 Received in revised form 4 January 2015 Accepted 6 January 2015 Available online 8 January 2015 Keywords: Selective laser melting CoCrW alloys Dental application Scanning strategy Corrosion resistance Metal release
a b s t r a c t In this study, an experimental investigation on fabricating Ni-free CoCrW alloys by selective laser melting (SLM) for dental application was conducted in terms of microstructure, hardness, mechanical property, electrochemical behavior, and metal release; and line and island scanning strategy were applied to determine whether these strategies are able to obtain expected CoCrW parts. The XRD revealed that the γ-phase and ε-phase coexisted in the as-SLM CoCrW alloys; The OM and SEM images showed that the microstructure of CoCrW alloys appeared square-like pattern with the fine cellular dendrites at the borders; tensile test suggested that the difference of mechanical properties of line- and island-formed specimens was very small; whilst the outcomes from the electrochemical and metal release tests indicated that the island-formed alloys showed slightly better corrosion resistance than line-formed ones in PBS and Hanks solutions. Considering that the mechanical properties and corrosion resistance of line-formed and island-formed specimens meet the standards of ISO 22674:2006 and EN ISO 10271, CoCrW dental alloys can be successfully fabricated by line and island scanning strategies in the SLM process. © 2015 Elsevier B.V. All rights reserved.
1. Introduction Cobalt–chromium based alloys have been widely used in various orthopedic implants such as artificial hip, knee joints and dentistry because of their excellent mechanical properties, high corrosion resistance, and good biocompatibility [1]. For dental application, cobalt–chromium based dental alloys are required to have yield stresses higher than the standard ISO 22674:2006 of 500 MPa. In recent years, improving the mechanical properties of cobalt–chromium based dental alloys therefore has been extensively conducted [2–4]. On the other hand, materials used in dental applications should consist of small grains. This is because chipping failure will occur in machined components with coarse grain structures as consequence of reducing the accuracy of the marginal fit of restorations [5]. Because the Ni can cause skin allergies or cancer in living organisms [3,6], note that novel Ni-free CoCrW has been extensively studied because of their excellent biocompatibility and strong bonding strengths of metal-porcelain, which has high potential for dental applications [7]. Yamanaka et al. [8] prepared Co–29Cr–9W cast alloys with carbon concentrations in the range 0.01–0.27 mass%. The effects of carbon concentration on the microstructures and tensile properties of the Ni-free Co–29Cr–9W cast alloys used in dental applications were reported. Meanwhile, Yamanaka et al. also ⁎ Corresponding author. E-mail address: [email protected] (J. Lin).
http://dx.doi.org/10.1016/j.msec.2015.01.023 0928-4931/© 2015 Elsevier B.V. All rights reserved.
investigated the effects of nitrogen on the high-temperature deformation and microstructure of biomedical Ni-free CoCrW alloys during the process of hot deformation. The main conclusion of their work was that the solute segregation and resulting carbide precipitation at the grain boundaries were promoted by increased carbon concentrations and were considered as the origin of such grain refinement [5]. In general, dental restorations are fabricated by casting and computer-aided manufacturing technologies. However, cast cobalt– chromium based alloys generally have large grains and exist quite non-homogeneous microstructure and solidification defects, which result in inadequate mechanical properties and inhomogeneous material quality. Therefore SLM technique, a kind of additive manufacturing method that fabricates metal products directly from CAD data in an additive layer-by-layer manner by selectively fusing together metal powder, has been applied in prosthetic dentistry to construct individual restorations with special anatomical characteristics and complex geometry for patients [9,10]. Basic research on cobalt–chromium based alloys for dental applications has been reported using the SLM process [11,12]. As suggested by a previous study that CoCr alloys formed by SLM existed higher corrosion resistance, lower metal release, and lower cell proliferation compared with the cast CoCr alloys. This was due to the formation of fine microstructure with cellular dendrites in as-SLM CoCr alloys during the very high rapid solidification [13]. According to Takaichi et al. [4], dense parts could be obtained and existed better mechanical property and lower amounts of metal release when input
518
Y. Lu et al. / Materials Science and Engineering C 49 (2015) 517–525
energy of laser scan is increased more. Obviously, optimization of process parameters, such as scanning speed, laser power, scanning strategies etc., for the fabrication of CoCr alloy is very important, which can affect their microstructure and properties [14]. Among these processing parameters, most concerns of scanning strategies focus on the line scanning strategy that laser moves in bidirectional or unidirectional way across part surfaces [15,16] illustrated in Fig. 1a–c. Actually, due to high local heat input in a short time during the SLM process, the part experiences residual stresses, which can introduce deformation, initiate cracks, even accuracy problem. Additionally, it was reported by Kruth et al. [17], who investigated the influence of vector length in the xdirection on curling of Ti–6Al–4V test part, that the residual stresses decreased when lower vector lengths were applied. In view of this, note that island scanning strategy is recently used to reduce the thermal residual stresses by shortening the individual tracks and dividing each layer into a number of smaller islands [18]. These smaller islands then are randomly scanned in an attempt to produce a more even heat distribution as a consequence of decreasing the residual stresses. The vectors in the neighboring islands are perpendicular to each other. In the subsequent layer, the island is shifted by 1 mm in both the X and Y directions, seen in Fig. 1d–f. The patented process allows solid and large-volume components to be generated with low warping [19]. So far some works have been performed to determine the relationship between the island strategy and properties of selective laser melted specimens [20]. Thijs et al. [21] applied line and island scanning strategies to investigate the microstructure and controllable texture of as-SLM AlSi10Mg parts. Kruth et al. [22] used island scanning strategies with different island sizes to measure the residual stresses of maraging steel. However, they suggested that the size of the islands did not seem to influence the residual stresses. In contrast with the relatively large body of published works on CoCrMo and CoCrW formed by traditional processes, there is rather limited information on the microstructure, mechanical property, and corrosion resistance evaluations of the CoCrW alloy fabricated by SLM. In this study, CoCrW alloys were fabricated by selective laser melting using line and island scanning strategies to identify whether these parameters are able to obtain expected CoCrW parts for dental applications. The line and island-formed CoCrW alloys were also evaluated in terms of microstructure, hardness, mechanical property, electrochemical behavior, and metal release.
2. Experimental detail 2.1. Materials and sample preparation The commercial CoCrW alloy powders were prepared for this study. The particle size distribution is 10–45 μm. The chemical compositions are listed in Table 1. Based on the supplier information, the density of the CoCrW alloy is 8.6 g/cm3.
line
a
b
The samples were fabricated by the SLM technique (Mlab-R, CONCEPTLAER, Germany). Two kinds of laser scanning strategies, i.e., line scanning and island scanning strategy, were applied for the fabrication of selective laser melted CoCrW alloys, as illustrated in Fig. 1a–f. The detail processing parameters are listed Table 2. 2.2. Microstructural observation and phase analysis Microstructural observation was carried out using optical microscope (Axio Vert.A1, ZEISS) and scanning electron microscope (SEM, JSM-6700F). Phase identification was studied by X-ray diffractometer (XRD, D/MAX-2500PC) with CuKα radiation. Specimen preparation was firstly mechanically polished, and then etched in HCl for 7 h [14]. 2.3. Mechanical property characterization Tensile tests were performed on a universal testing machine (MTS E45.105) at room temperature, with an initial strain rate of 2 mm/s. The schematic diagram of tensile specimen is shown in Fig. 2. Fractography were examined by using SEM (SEM, JSM-6700F). Vickers hardness (HV) of the specimens was also carried out using the hardness tester (HX-1000TM), with a load of 100 gf and a dwell time of 12 s. The density of the samples with the dimension of 10 × 10 × 10 mm was measured by applying Archimedes' principle. 2.4. Electrochemical test Electrochemical test was conducted to evaluate the corrosion resistance in the PBS and Hanks solution at 37 °C. A three-electrode cell was used for the electrochemical measurements. The counter electrode was made of platinum and a saturated calomel electrode (SCE) was used as the reference electrode. The sample with an exposed area of 1 cm2 was taken as the working electrode. The polarization scan was started from anodic region from 250 mV below open circuit potential at a constant voltage scan rate of 0.5 mV/s. 2.5. Immersion test The static immersion test was performed in accordance with the currently specified EN ISO 10271 standard for metallic biomaterials. Samples with the dimensions of Φ10 × 30 mm were fabricated by SLM. Prior to performing this study, the samples were ground with waterproof emery paper to 1200 grit under running water, then ultrasonically cleaned in acetone for 15 min, rinsed in distilled water and finally dried at room temperature. The static immersion tests were conducted using PBS and Hank's solutions, respectively. A 12.56 ml of each solution was poured into the polypropylene bottles each containing a plate specimen. All the sealed bottles were placed inside an incubator at 37 °C for 7 d. The concentrations of various metals released into solution were determined in ppb (mg/ml) by inductively coupled plasma-atomic emission spectroscopy (ICP-AES, Ultima2).
c 2.6. Statistical analysis The SPSS15.0 statistical software package was used for data analysis. P N 0.05 was considered as no statistically significant difference.
island d
e
3. Results
f
Fig. 1. The illustration of the line scanning strategy (a–c) and island scanning strategy (d–f).
3.1. XRD study Fig. 3 shows the XRD patterns of the unpolished top-view of the lineformed and island alloys. As can be seen, both CoCrW alloys exhibit duplex phases of γ (fcc) and ε (hcp) phase. Interestingly, the islandformed alloy exhibits much higher diffraction intensity assigned to ε phase compared with the line-formed alloy.
Y. Lu et al. / Materials Science and Engineering C 49 (2015) 517–525
519
Fig. 2. Schematic diagram of tensile specimen.
3.2. Microstructure observation Fig. 4 shows the morphologies of the unpolished and polished topview of both alloys. From the morphologies of the unpolished topview in Fig. 4a–b, it is visible that the melted tracks are obviously presented. The solidified lines marked by the red line of the line scanning strategy and island scanning strategy are approximately 70 and 100 μm width, respectively. Some tens-micron-scale un-melted powders are visible on the unpolished surface of line-formed alloy (Fig. 4a) and island-formed alloy (Fig. 4b). However, the more unmelted powders can be observed on the surface of island-formed alloy than that of line-formed one. Fig. 4c–d shows the optical micrographs of the polished top-view of both alloys. Higher porosity is observed in the island-formed alloy (Fig. 4d) compared with the line-formed alloy (Fig. 4c), which may be related to more un-melted powders in the criss-cross regions when used island scanning strategy. Fig. 5 shows the optical micrograph of the top-view of both alloys. Both alloys exhibit distinctive square-like pattern with the surrounding border as denoted by white lines in Fig. 5a–b. The edge length of the square is 60 ± 5 μm, whilst the width of the border is 20 ± 3 μm. The border presents to be darker whereas the square-like structure (central zone) is brighter. The square-like pattern of the line-formed specimen appears to be more uniform compared with island-formed specimen. A furthermore observation of metallographic structure from the top view was obtained by scanning electron microscope as shown in Fig. 6. Striations were detected inside the square-like structure (Fig. 6a–b), indicating that the structure may consist of γ- and ε-phases [23]. The magnified microstructures of regions ①–④ (see Fig. 6a–b) are shown in Fig. 6c–f. It is evident that the border structure shows the presence of fine homogeneous cellular dendrites with a submicron size (Fig. 6c– d), whereas elongated lamellae seem to grow along a direction in the square-like structure (Fig. 6e–f), revealing that grain orientations were different.
3.3. Mechanical property and hardness Fig. 7 shows the typical nominal stress–strain curves of the selective laser melted CoCrW specimens formed by the two kinds of laser
scanning strategies. Both stress–strain curves exhibit uniform elongation followed by a sudden fracture without any macroscopic necking. Such tensile deformation was also observed in the traditional CoCrW and CoCrMo orthopedic alloys [3,24]. The mechanical properties of the alloys listed in Table 3 indicate that the line-formed alloys exhibit slightly higher tensile strength and yield strength when compared with island-formed alloys (p N 0.05), whereas the elongation shows no significant differences (p N 0.05). The yield strength of both CoCrW alloys meets the standard ISO 22674:2006 for dental restorations (N 500 MPa). Fig. 8 shows the SEM images of the fractured surface after tensile test. It can be found that both the alloys exhibit facet-like fracture surfaces. Likewise, striations and wedge type cracks are also clearly found on both fractured surface (Fig. 8a–b), which are dominant for both CoCrW specimens. Under the higher magnification, the dimple-like patterns can be observed on the line-formed alloy (Fig. 8c) and islandformed alloy (Fig. 8d). These results indicate that the fractured mechanisms of all the investigated alloys have the same nature. The average density of the line-formed and island-formed specimens is about 8.537 and 8.526 g/cm3, respectively, that is, the alloys formed at line scanning strategy show slightly better dense than the island-formed alloy (P N 0.05). However, both relative densities exceed 99%, indicating that the two kinds of scanning strategies applied in the study can fabricate dense parts. For the hardness, the line-formed alloys have slightly higher average value of hardness than that of islandformed alloys, 570 ± 3.4HV and 564 ± 2.9HV, respectively. 3.4. Electrochemical test The polarization curves for selective laser melted CoCrW specimens formed at different scanning strategies in PBS and Hanks solutions are presented in Fig. 9. Polarization parameters, including corrosion potential (Ecorr), corrosion current density (Icorr), polarization resistance (Rp), and corrosion rate were obtained from the polarization curves listed in Table 4. It is clearly seen that the typical polarization curves are found to show the representative corrosion behaviors of both CoCrW specimens. The anodic branches of polarization curves for both specimens in PBS and Hanks solutions exhibit a similar passive region at a similar potential range, which are at the potential of 0 to 0.5 V. This passivation behavior is closely related with the formation of protective oxide layer [25]. Also, it has to note that there is a distinct oscillation around Table 2 The processing parameters of SLM.
Table 1 The chemical composition of the CoCrW alloy powders. Chemical Composition (Wt%) Co
Cr
W
Si
Mn
N
Nb
Fe
Ni
Be
60.5
28.0
9.0
1.5
b1.0%
b1.0%
b1.0%
b1.0%
–
–
Scan strategy
Laser power (W)
Scan speed (mms−1)
Powder thickness (μm)
Island size (mm)
Line Island
95 95
700 700
25 25
– 2×2
520
Y. Lu et al. / Materials Science and Engineering C 49 (2015) 517–525
solution have higher corrosion current density than in Hanks solution as well as corrosion rate, indicating that the alloys in Hanks solution appeared to be more susceptible to corrosion than in PBS solution independent of laser scanning strategy. 3.5. Metal release
Fig. 3. XRD pattern of the selective laser melted CoCrW alloys.
0.60 V in the anodic curve of both samples in PBS solution as denoted by rectangle in Fig. 9a. Likewise, a small region of oscillation around 0 V is also found in the anodic curve of line-formed alloy in Hanks solution (Fig. 9b). However, such phenomenon is not observed in the anodic curve of island-formed alloys in Hanks solution. As given in Table 4, the line-formed specimens exhibit a slightly higher corrosion potential (Ecorr) but the slightly lower corrosion current density (Icorr), indicating that the corrosion resistance of the island-formed alloys was slightly greater than that of line-formed specimens. This is further verified by the higher values of polarization resistance Rp and lower corrosion rate of the island-formed alloy in comparison with the line-formed alloy in both simulated body solutions. On the other hand, given the differing solutions the alloys in PBS
a
70µm
The released amounts of the metal in different simulated body solution are shown in Fig. 10 after a 7-day immersion. Regarding the different scanning strategies, the released amount of Co from line-formed specimen is more than that of island-formed alloy in regardless of the differing simulated body solutions seen in Fig. 10a. This indicates that it was easier for elements released from the line-formed specimen in simulated body solutions as compared with island-formed specimen. The difference in the released amount of Cr and W, especially that of Cr, in both solutions is very small (Fig. 10b–c). On the other hand, metal release of the selective laser melted alloys was also affected by simulated body solutions chemistry. As can be seen from Fig. 10a–c, the released amounts of base metal Co and each alloying element differed depending on the type of solutions. The released amount of metal in PBS solution is higher than that in Hanks solution independent of laser scanning strategy, which is consistent with the electrochemical test. In general, the total amounts of metal release were in the range of 0.223–0.276 μg/cm2/7 d, which is significantly lower than that of the standard EN ISO 10271 of 200 μg/cm2/7 d. 4. Discussion 4.1. Phase XRD results indicate that the main phases of both selective laser melted alloys were the γ-phase and ε-phase. Generally, the γ-phase is stable at high temperature above 1120 K, while the ε-phase exist as
b 100µm
Unmelted powder
Unmelted powder
line
200µm
island
c
d
line
island
200µm
Fig. 4. (a) Surface morphology of unpolished line-formed alloy; (b) surface morphology of unpolished island-formed alloy; (c) optical micrograph of polished line-formed alloy; (d) optical micrograph of polished island-formed alloy.
Y. Lu et al. / Materials Science and Engineering C 49 (2015) 517–525
a
521
b
Fig. 5. Top view of metallographic structure: (a) 100× optical micrograph of polished line-formed alloy; (b) 100× optical micrograph of polished island-formed alloy.
an equilibrium phase at room temperature [23,26]. The ε-phase is formed as a consequence of γ → ε transformation while the alloys cooled from high temperatures. Considering the quick cooling rate during the SLM process, the γ-phase could exist at room temperature in this
study. It is noteworthy from the XRD patterns that the peak intensity of island-formed alloy from the ε-phase is greater than that of line-formed alloy. This fact may attribute to the more un-melted powder on the surface of island-formed alloys.
Fig. 6. SEM images of metallographic structure of CrCoW alloy: (a) the microstructure of line-formed alloy; (b) the microstructure of island-formed alloy; (c) the magnified microstructure of region ① in Fig. 6a, (d) the magnified microstructure of region ③ in Fig. 6b; (e) the magnified microstructure of region ② in Fig. 6a; (f) the magnified microstructure of region ④ in Fig. 6b.
522
Y. Lu et al. / Materials Science and Engineering C 49 (2015) 517–525 1200
patterns of island-formed specimen are irregularity and uneven compared with line scanning strategy. The microstructure of both alloys shows a homogeneous cellular structure with cell boundaries. Similar findings also have been reported by some previous reports [28]. This is because of the rapid solidification and strong temperature gradients of molten pool during the SLM [14]. However, the further formation mechanism of the microstructure should be clarified in the further study.
Island Line
tenslie strength/MPa
1000
800
600
4.3. Mechanical property and hardness
400
200
0
4
8
12
16
20
24
28
32
36
40
strain % Fig. 7. Nominal stress–strain curves of the selective laser melted CoCrW alloys.
4.2. Microstructure and possible formation mechanism The microstructure of both selective laser melted CoCrW alloys exhibited two distinct structures, square-like structure and border structure, which were rarely observed in cobalt–chromium based alloys processed by a traditional method. The illustration of possible formation mechanism is shown in Fig. 11. It can be found from the optical micrograph of slightly etched top-view surface that the layer N and layer N + 1 are criss-cross regularly, as can be seen in Fig. 11a. The possible formation mechanism can be deduced as follows. During the SLM, the thermal gradients and growth rates were differing over the melt pool due to the movement of the heat source: the highest ones were located in the central regions of the melting track and then decreased when going towards the border of the melting tracks [27]. The prior melted tracks were heavily remelted and crossed on the subsequent laser tracks, and suffered more sufficient over-heating. In fact, the remelting could be served as a solution annealing at a higher temperature for the prior melted tracks, thereby promoted the growth of elongated grains, as indicated by ① in Fig. 11b–c. In consideration of the width of solidified tracks (approximately 70–100 μm), it is reasonable to form a square pattern with the edge length of approximately 60 μm after the deduction of the width of the overlaps between adjacent melted tracks. Namely, the width of the solidified overlap between adjacent melted tracks was about 10–30 μm, which is nearly correspond with the width of border structure. Regarding the border structure, however, the solidified overlap ③ between aside adjacent melted tracks was partial remelted due to the existence of slight concave ②, when compared with criss-cross central melted tracks ① (see Fig. 11b–c). On the other hand, because of the decrease of thermal gradients and growth rates at borders of the laser scanning passes that even reached zero at the melting border [27]. Correspondingly, the overlaps ③ suffered less over-remelting and overheating as a consequence of impeding the grain growth. Hence, the above analysis can be used to explain why the microstructure of border regions appeared cellular dendrites, while the elongated lamellar structures were located inside the central regions. Additionally, it has been described in the Introduction that the island scanning strategy conducts random laser exposure in each layer; subsequently the island pattern is shifted by 1 mm in both the X and Y directions on the next layer, which can be used to explain why square-like Table 3 The mechanical properties of the both CoCrW alloys. Samples
UTS, (MPa)
σ0.2 (MPa)
E, (%)
Line Island
1158.22 ± 21 1115.56 ± 19
850 ± 23 825 ± 14
9.80 ± 0.10% 10 ± 0.10%
The different laser scanning strategies used in the present study caused slightly different mechanical properties in both specimens that the line-formed alloys exhibit slightly higher UTS and 0.2% yield strength when compared with island-formed alloys, whereas the elongation shows no significant differences. It is well known that mechanical properties are affected by many factors, such as phase, grain orientation, grain boundary conditions, defects, etc. [16]. Given that the phase, grain orientation, and grain boundary conditions of both specimens are very similar, the slight difference in UTS and 0.2% yield strength is obviously associated with the difference in defect and grain size. It is considered that the more dense part without pores exhibits excellent mechanical property; conversely the presence of more pores is a sign of poor mechanical properties, and larger grains are also prone to inducing cracks. In the present work, the island-formed specimens show slightly more pores and larger grain size than that of line-formed one. As a result, the cracks may easily propagate from those pores and larger grains for island-formed specimen. Although it is well recognized that presence of mechanical anisotropy exists in SLM parts, most of them attributed this slightly difference in this study to the inner porosity [29]. Indeed, compared with the selective laser melted parts as reported by previous study [4], the CoCrW alloys fabricated by differing laser scanning strategies exhibited better 0.2% yield stresses and tensile strength. This is due to the homogeneous microstructure with diamond patterns that endowed with adequate mechanical property. However, both alloys showed relatively poor ductility as compared with previous study. The ε-phase is known to enhance the strength and wear resistance of the CoCr-based alloy. But the formation of ε-phase also leads to poor ductility [30]. During the tensile test, strain-induced martensite ε-phase intersected with the preexisting athermal ε martensite may induce the stress concentration at their interfaces and/or grain boundaries resulting in premature fracture before the onset of plastic instability, consequently caused detrimental to tensile ductility [23]. As a result shown by Yamanaka et al. [26], the CoCrW alloys exhibited the better tensile ductility when microstructures contained the less martensite ε phase. During the SLM process, it is recognized that molten pool undergoes a fast cooling rate, leading to the restraining transformation of metastable γ-phase to the martensite ε-phase. However, such transformation would be occurred when the previous solidified layers were reheated and remelted by the subsequent layer process [14]. That is, the γ-phase and ε-phase coexisted in the alloys. Therefore, the reason for the relatively lower elongation in the present study is the major result from the rich ε-phase in alloy. From fracture surface observation in Fig. 8a–b, facet-like fracture surfaces and wedge type cracks are clearly observed on fracture surfaces similar to the previous reports [31]. As suggested by a previous study [23], this kind of fracture mode observed herein can be recognized as a quasi-cleavage type fracture since a true cleavage fracture cannot occur in fcc polycrystalline materials [24,32]. The type fracture was enhanced by the athermal ε martensite during the tensile test [33]. Wedge type cracks may be formed by the cleavage fracture along the {1 1 1}, while the dimple-like patterns may be induced from the cellular dendrites [24]. Hardness is an important characteristic in a way to evaluate whether alloys combine with greater elastic modulus to allow the manufacture of longer prostheses. Result on hardness measurement indicates the
Y. Lu et al. / Materials Science and Engineering C 49 (2015) 517–525
a
b
z
z
x
20µm
x
c
d
z
z x
5µm
x
523
20µm
5µm
Fig. 8. SEM images of the fracture surface of the selective laser melted CoCrW alloys after tensile test. (a) fracture surface of line-formed; (b) fracture surface of island-formed;(c) the magnified fracture surface of Fig. 8a; (d) the magnified fracture surface of Fig. 8b.
hardness value of line-formed alloys was slightly higher than islandformed alloys. This fact may be due to the slightly higher density of line-formed alloys. Henriques et al. [2], in their study on the hardness of hot pressed CoCrMo samples for surgical applications, also reported higher hardness in hot pressed specimens than in cast specimens. They demonstrated that the greater hardness of hot pressed specimens was related with low porosity relatively to cast ones. Therefore, the line scanning strategy can fabricate more dense part in comparison with island strategy.
4.4. Corrosion resistance and metal release In vitro studies, i.e., electrochemical and metal release tests, were conducted in simulated body solutions give an overview of the corrosion resistance and biocompatibility (metal release) of the selective laser melted alloy. The anodic polarization curves show that both alloy formed by line and island scanning strategies were in the passive state, which was a characteristic of materials with high resistant to corrosion. The oscillation around 0.60 V recognized as an oxidation peak
Fig. 9. The polarization curves of selective laser melted CoCrW specimens formed at different scanning strategies in PBS solution (a) and Hanks solution (b).
524
Y. Lu et al. / Materials Science and Engineering C 49 (2015) 517–525
Table 4 The polarization parameters obtained from the polarization curves. Samples
Icorr (nA)
Ecorr (mV)
Rp (Ω)
Corrosion rate (10−4 mm/year)
Line–PBS Island–PBS Line–Hanks Island–Hanks
20.20 19.70 12.00 8.02
−201 −200 −284 −327
1.790 1.855 1.325 1.914
4.60 4.50 2.70 1.80
may be attributed to the competition between formation and dissolution of passive film [34]. The corrosion potential (Ecorr) of the lineformed alloys was shifted to more noble potentials (more than 40 mV). However, the corrosion current density (Icorr) of the lineformed alloys was slightly greater than that of island-formed alloys. This indicates that the slightly higher corrosion rate occurred in the line-scanning alloys in simulated body solutions, which was verified by the fact that the line-formed alloys exhibited slightly lower polarization resistance Rp and higher corrosion rate with the respect to islandformed alloys. Results from the metal release revealed that Co was more active than Cr and W independent of processing parameters and simulated body solutions. Such similar results were found in previous studies that Co is preferentially released from CoCrMo alloy [24]. The quantity of the Co released in PBS solution from selective laser melted CoCrMo in previous report was approximately 0.45 μg/cm2/7 d [14]. In our test, the amount of Co release from the selective laser melted CoCrW alloys into PBS was less than 0.22 μg/cm2/7 d. Likewise, the quantity of Cr and W released from the CoCrW alloys used in this study was very small. It is noteworthy that the quantity of Cr was lower than that of W in both simulated body solutions, indicating that elements were not absolutely determined among component elements of the selective laser melted alloys. Otherwise, a higher amount of Co released from line-formed alloys was detected than that of island-formed alloys in Hanks and PBS solutions. Although it has been determined that alloy that existed more porosity that corresponded to bigger surface area are regarded for higher corrosion current densities, and that the surface quality could affect the corrosion resistance [35]. However, based on the studies of the electrochemical and metal release tests, the porosity and surface quality are not in the scope of the factors. Assuming that thermal residual stress, which had accumulated from the repeated rapid heating and cooling of molten pool during the process, can be used to explain why the difference in corrosion behavior. As mentioned above, the island scanning strategy make the laser randomly exposed on each individual layer, thereby ensures a significant reduction in stresses within the alloy. An ongoing quantified examination should be conducted to assess how the thermal residual stress influents on the corrosion resistance.
The corrosion behavior of cobalt-based alloy is very sensitive to the solution composition [36]. The CoCrW alloys in the PBS solution had slightly higher corrosion rate than in Hanks solution. According to the electrochemical and metal release analysis, both CoCrW alloys showed the higher release amount of Co and lower corrosion current density in PBS than in Hanks solution. This suggested that the alloys in PBS solution appeared to be more susceptible to corrosion than in Hanks solution regardless of laser scanning strategy. Generally, when CoCrMo alloy is immersed in Hanks and PBS solution a passive layer is formed on surface. The passive layer was characterized as containing oxides of cobalt, chromium and some hydroxides [28,37], which played an important role as an inhibitor of ion release. Namely, the corrosion reaction of the materials with fluid usually occurs at interfaces. According to previous study [38], the activity of the cobalt oxide layer decreased due to the formation of insoluble cobalt phosphate when CoCrMo alloy was placed in the presence of NaH2PO4. Since the Hank's solution contains both NaH2PO4 and Na2HPO4, it is probable that the insoluble cobalt phosphate was formed in the early stage. This is coherent with the early appearance of the oscillation in anodic polarization curves of line-formed alloy as indicated by the rectangle in Fig. 9b. Actually, corrosion resistance of metallic biomaterials depends on the passive layer and the compositions of the layer are usually important for biocompatibility [34]. Therefore, the surface structure and composition in the passive layer of the selective laser melted CoCrW alloy should be analyzed for further investigation, such as XPS study. It has to be mentioned that the residual stress is generated during the SLM in some degree regardless of line and island laser scanning strategies, which will affect the accuracy of the as-SLM parts. For the practical application, the heat treatment should be conducted to release the residual stress. Therefore, further studies focusing on microstructure, mechanical properties, and precision of the SLM parts after heat treatment will be investigated. 5. Conclusions In the study, two dense CoCrW alloys were formed by line and island scanning strategy, respectively. Due to the rapid cooling rate and strong temperature gradient during the laser melting process, the metallographic structure of the top-view presented square-like pattern with fine cellular dendrites. Results in terms of tensile, hardness, density, electrochemical, and metal release tests suggested that there was no or small difference between the two CoCrW alloys. This should be attributed to their similar microstructure. The tensile test indicated that the yield strength of both alloys met the standard ISO 22764:2006 for dental restorations. The failure type was recognized as a type quasicleavage fracture. The electrochemical and metal release tests suggested that the island-formed alloys showed slightly better corrosion
Fig. 10. Released amounts of the metal in PBS and Hanks solution.
Y. Lu et al. / Materials Science and Engineering C 49 (2015) 517–525
525
Fig. 11. (a) Slightly etched top view surface; (b) illustration of formation of square-like structure in section view of Y-Z axis; (c) illustration of formation of square-like structure in top view of X–Y axis.
resistance than line-formed ones, i.e., lower current densities (Icorr), corrosion rate and lower amounts of metal release in PBS and Hanks solution. The total quantity of Co, Cr and W used in this study was in the range of 0.223–0.276 μg/cm2/7 d for both alloys. Based on the results obtained in this study, line and island scanning strategy in the SLM process for the CoCrW alloy is a promising candidate for use in dental applications. Acknowledgments This work was financially supported by National pilot projects (No. XDA09020301). References [1] G. Manivasagam, D. Dhinasekaran, A. Rajamanickam, Recent Patents Corros. Sci. 2 (2010) 40–54. [2] B. Henriques, D. Soares, F.S. Silva, J. Mech. Behav. Biomed. Mater. 12 (2012) 83–92. [3] K. Yoda, Suyalatu, A. Takaichi, N. Nomura, Y. Tsutsumi, H. Doi, S. Kurosu, A. Chiba, Y. Igarashi, T. Hanawa, Acta Biomater. 8 (2012) 2856–2862. [4] A. Takaichi, Suyalatu, T. Nakamoto, N. Joko, N. Nomura, Y. Tsutsumi, S. Migita, H. Doi, S. Kurosu, A. Chiba, N. Wakabayashi, Y. Igarashi, T. Hanawa, J. Mech. Behav. Biomed. Mater. 21 (2013) 67–76. [5] K. Yamanaka, M. Mori, A. Chiba, Mater. Des. 57 (2014) 421–425. [6] M. Mori, K. Yamanaka, A. Chiba, J. Alloys Compd. 590 (2014) 411–416. [7] A. Karaali, K. Mirouh, S. Hamamda, P. Guiraldenq, Mater. Sci. Eng. A 390 (2005) 255–259. [8] K. Yamanaka, M. Mori, A. Chiba, Mater. Lett. 116 (2014) 82–85. [9] Z. Zhang, D. Jones, S. Yue, P.D. Lee, J.R. Jones, C.J. Sutcliffe, E. Jones, Mater. Sci. Eng. C 33 (2013) 4055–4062. [10] B. Song, S. Dong, P. Coddet, H. Liao, C. Coddet, Mater. Des. 53 (2014) 1–7. [11] B.-A. Lee, O.-S. Kim, M.-S. Vang, Y.-J. Park, J. Prosthet. Dent. 109 (2013) 95–105. [12] N. Xiang, X.-Z. Xin, J. Chen, B. Wei, J. Dent. 40 (2012) 453–457.
[13] X.Z. Xin, N. Xiang, J. Chen, B. Wei, Mater. Lett. 88 (2012) 101–103. [14] Y.S. Hedberg, B. Qian, Z. Shen, S. Virtanen, I. Odnevall Wallinder, Dent. Mater. 30 (5) (2014) 525–534. [15] B. Song, S. Dong, H. Liao, C. Coddet, Int. J. Adv. Manuf. Technol. 61 (2011) 967–974. [16] F. Liu, X. Lin, C. Huang, M. Song, G. Yang, J. Chen, W. Huang, J. Alloys Compd. 509 (2011) 4505–4509. [17] J.P. Kruth, J. Deckers, E. Yasa, R. Wauthle, Proc IME B J Eng Manufact 226 (2012) 980–991. [18] B. Vrancken, V. Cain, R. Knutsen, J. Van Humbeeck, Scr. Mater. 87 (2014) 29–32. [19] Concept laser, http://www.concept-laser.de/2014. [20] M.F. Zaeh, G. Branner, Production Engineering 4 (1) (2009) 35–45. [21] L. Thijs, K. Kempen, J.-P. Kruth, J. Van Humbeeck, Acta Mater. 61 (2013) 1809–1819. [22] J.P. Kruth, M. Badrossamay, E. Yasa, J. Deckers, L. Thijs, J. Van Humbeeck, Proceedings of the 16th International Symposium on Electromachining, 2010. [23] K. Yamanaka, M. Mori, A. Chiba, Mater. Sci. Eng. A 528 (2011) 5961–5966. [24] K. Yamanaka, M. Mori, A. Chiba, J. Mech. Behav. Biomed. Mater. 37 (2014) 274–285. [25] D. Sun, J.A. Wharton, R.J.K. Wood, L. Ma, W.M. Rainforth, Tribol. Int. 42 (2009) 99–110. [26] K. Yamanaka, M. Mori, A. Chiba, Mater. Sci. Eng. C 40 (2014) 127–134. [27] L. Thijs, F. Verhaeghe, T. Craeghs, J.V. Humbeeck, J.-P. Kruth, Acta Materialia 58 (9) (2010) 3303–3312. [28] X. Xin, N. Xiang, J. Chen, D. Xu, B. Wei, J. Mater. Sci. (2012) 1–8. [29] C. Qiu, N.J. Adkins, M.M. Attallah, Mater. Sci. Eng. A 578 (2013) 230–239. [30] K. Yamanaka, M. Mori, Y. Koizumi, A. Chiba, J. Mech. Behav. Biomed. Mater. 32 (2014) 52–61. [31] K. Yamanaka, M. Mori, A. Chiba, J. Mech. Behav. Biomed. Mater. 29 (2014) 417–426. [32] J. Morris, Z. Guo, C. Krenn, Y.-H. Kim, ISIJ Int. 41 (2001) 599–611. [33] T. Matković, P. Matković, J. Malina, J. Alloys Compd. 366 (2004) 293–297. [34] T. Hanawa, Mater. Sci. Eng. C 24 (2004) 745–752. [35] W.C. Rodrigues, L.R. Broilo, L. Schaeffer, G. Knörnschild, F.R.M. Espinoza, Powder Technol. 206 (2011) 233–238. [36] M. Metikoš-Huković, R. Babić, Corros. Sci. 49 (2007) 3570–3579. [37] A. Hodgson, S. Kurz, S. Virtanen, V. Fervel, C.-O. Olsson, S. Mischler, Electrochim. Acta 49 (2004) 2167–2178. [38] L. Burke, M. Lyons, O. Murphy, J. Electroanal. Chem. Interfacial Electrochem. 132 (1982) 247–261.
Contents lists available at ScienceDirect
Materials Science and Engineering C journal homepage: www.elsevier.com/locate/msec
Investigation on the microstructure, mechanical property and corrosion behavior of the selective laser melted CoCrW alloy for dental application Yanjin Lu a, Songquan Wu a, Yiliang Gan a, Junlei Li b, Chaoqian Zhao a,c, Dongxian Zhuo a, Jinxin Lin a,⁎ a b c
Fujian Institute of Research on the Structure of Matter, Chinese Academy of Sciences, 155 Yangqiao Road West, Fuzhou, China Institute of Metal Research, Chinese Academy of Sciences, 72 Wenhua Road, Shenyang, China College Materials Science and Engineering, Fujian Normal University, Fuzhou, China
a r t i c l e
i n f o
Article history: Received 25 September 2014 Received in revised form 4 January 2015 Accepted 6 January 2015 Available online 8 January 2015 Keywords: Selective laser melting CoCrW alloys Dental application Scanning strategy Corrosion resistance Metal release
a b s t r a c t In this study, an experimental investigation on fabricating Ni-free CoCrW alloys by selective laser melting (SLM) for dental application was conducted in terms of microstructure, hardness, mechanical property, electrochemical behavior, and metal release; and line and island scanning strategy were applied to determine whether these strategies are able to obtain expected CoCrW parts. The XRD revealed that the γ-phase and ε-phase coexisted in the as-SLM CoCrW alloys; The OM and SEM images showed that the microstructure of CoCrW alloys appeared square-like pattern with the fine cellular dendrites at the borders; tensile test suggested that the difference of mechanical properties of line- and island-formed specimens was very small; whilst the outcomes from the electrochemical and metal release tests indicated that the island-formed alloys showed slightly better corrosion resistance than line-formed ones in PBS and Hanks solutions. Considering that the mechanical properties and corrosion resistance of line-formed and island-formed specimens meet the standards of ISO 22674:2006 and EN ISO 10271, CoCrW dental alloys can be successfully fabricated by line and island scanning strategies in the SLM process. © 2015 Elsevier B.V. All rights reserved.
1. Introduction Cobalt–chromium based alloys have been widely used in various orthopedic implants such as artificial hip, knee joints and dentistry because of their excellent mechanical properties, high corrosion resistance, and good biocompatibility [1]. For dental application, cobalt–chromium based dental alloys are required to have yield stresses higher than the standard ISO 22674:2006 of 500 MPa. In recent years, improving the mechanical properties of cobalt–chromium based dental alloys therefore has been extensively conducted [2–4]. On the other hand, materials used in dental applications should consist of small grains. This is because chipping failure will occur in machined components with coarse grain structures as consequence of reducing the accuracy of the marginal fit of restorations [5]. Because the Ni can cause skin allergies or cancer in living organisms [3,6], note that novel Ni-free CoCrW has been extensively studied because of their excellent biocompatibility and strong bonding strengths of metal-porcelain, which has high potential for dental applications [7]. Yamanaka et al. [8] prepared Co–29Cr–9W cast alloys with carbon concentrations in the range 0.01–0.27 mass%. The effects of carbon concentration on the microstructures and tensile properties of the Ni-free Co–29Cr–9W cast alloys used in dental applications were reported. Meanwhile, Yamanaka et al. also ⁎ Corresponding author. E-mail address: [email protected] (J. Lin).
http://dx.doi.org/10.1016/j.msec.2015.01.023 0928-4931/© 2015 Elsevier B.V. All rights reserved.
investigated the effects of nitrogen on the high-temperature deformation and microstructure of biomedical Ni-free CoCrW alloys during the process of hot deformation. The main conclusion of their work was that the solute segregation and resulting carbide precipitation at the grain boundaries were promoted by increased carbon concentrations and were considered as the origin of such grain refinement [5]. In general, dental restorations are fabricated by casting and computer-aided manufacturing technologies. However, cast cobalt– chromium based alloys generally have large grains and exist quite non-homogeneous microstructure and solidification defects, which result in inadequate mechanical properties and inhomogeneous material quality. Therefore SLM technique, a kind of additive manufacturing method that fabricates metal products directly from CAD data in an additive layer-by-layer manner by selectively fusing together metal powder, has been applied in prosthetic dentistry to construct individual restorations with special anatomical characteristics and complex geometry for patients [9,10]. Basic research on cobalt–chromium based alloys for dental applications has been reported using the SLM process [11,12]. As suggested by a previous study that CoCr alloys formed by SLM existed higher corrosion resistance, lower metal release, and lower cell proliferation compared with the cast CoCr alloys. This was due to the formation of fine microstructure with cellular dendrites in as-SLM CoCr alloys during the very high rapid solidification [13]. According to Takaichi et al. [4], dense parts could be obtained and existed better mechanical property and lower amounts of metal release when input
518
Y. Lu et al. / Materials Science and Engineering C 49 (2015) 517–525
energy of laser scan is increased more. Obviously, optimization of process parameters, such as scanning speed, laser power, scanning strategies etc., for the fabrication of CoCr alloy is very important, which can affect their microstructure and properties [14]. Among these processing parameters, most concerns of scanning strategies focus on the line scanning strategy that laser moves in bidirectional or unidirectional way across part surfaces [15,16] illustrated in Fig. 1a–c. Actually, due to high local heat input in a short time during the SLM process, the part experiences residual stresses, which can introduce deformation, initiate cracks, even accuracy problem. Additionally, it was reported by Kruth et al. [17], who investigated the influence of vector length in the xdirection on curling of Ti–6Al–4V test part, that the residual stresses decreased when lower vector lengths were applied. In view of this, note that island scanning strategy is recently used to reduce the thermal residual stresses by shortening the individual tracks and dividing each layer into a number of smaller islands [18]. These smaller islands then are randomly scanned in an attempt to produce a more even heat distribution as a consequence of decreasing the residual stresses. The vectors in the neighboring islands are perpendicular to each other. In the subsequent layer, the island is shifted by 1 mm in both the X and Y directions, seen in Fig. 1d–f. The patented process allows solid and large-volume components to be generated with low warping [19]. So far some works have been performed to determine the relationship between the island strategy and properties of selective laser melted specimens [20]. Thijs et al. [21] applied line and island scanning strategies to investigate the microstructure and controllable texture of as-SLM AlSi10Mg parts. Kruth et al. [22] used island scanning strategies with different island sizes to measure the residual stresses of maraging steel. However, they suggested that the size of the islands did not seem to influence the residual stresses. In contrast with the relatively large body of published works on CoCrMo and CoCrW formed by traditional processes, there is rather limited information on the microstructure, mechanical property, and corrosion resistance evaluations of the CoCrW alloy fabricated by SLM. In this study, CoCrW alloys were fabricated by selective laser melting using line and island scanning strategies to identify whether these parameters are able to obtain expected CoCrW parts for dental applications. The line and island-formed CoCrW alloys were also evaluated in terms of microstructure, hardness, mechanical property, electrochemical behavior, and metal release.
2. Experimental detail 2.1. Materials and sample preparation The commercial CoCrW alloy powders were prepared for this study. The particle size distribution is 10–45 μm. The chemical compositions are listed in Table 1. Based on the supplier information, the density of the CoCrW alloy is 8.6 g/cm3.
line
a
b
The samples were fabricated by the SLM technique (Mlab-R, CONCEPTLAER, Germany). Two kinds of laser scanning strategies, i.e., line scanning and island scanning strategy, were applied for the fabrication of selective laser melted CoCrW alloys, as illustrated in Fig. 1a–f. The detail processing parameters are listed Table 2. 2.2. Microstructural observation and phase analysis Microstructural observation was carried out using optical microscope (Axio Vert.A1, ZEISS) and scanning electron microscope (SEM, JSM-6700F). Phase identification was studied by X-ray diffractometer (XRD, D/MAX-2500PC) with CuKα radiation. Specimen preparation was firstly mechanically polished, and then etched in HCl for 7 h [14]. 2.3. Mechanical property characterization Tensile tests were performed on a universal testing machine (MTS E45.105) at room temperature, with an initial strain rate of 2 mm/s. The schematic diagram of tensile specimen is shown in Fig. 2. Fractography were examined by using SEM (SEM, JSM-6700F). Vickers hardness (HV) of the specimens was also carried out using the hardness tester (HX-1000TM), with a load of 100 gf and a dwell time of 12 s. The density of the samples with the dimension of 10 × 10 × 10 mm was measured by applying Archimedes' principle. 2.4. Electrochemical test Electrochemical test was conducted to evaluate the corrosion resistance in the PBS and Hanks solution at 37 °C. A three-electrode cell was used for the electrochemical measurements. The counter electrode was made of platinum and a saturated calomel electrode (SCE) was used as the reference electrode. The sample with an exposed area of 1 cm2 was taken as the working electrode. The polarization scan was started from anodic region from 250 mV below open circuit potential at a constant voltage scan rate of 0.5 mV/s. 2.5. Immersion test The static immersion test was performed in accordance with the currently specified EN ISO 10271 standard for metallic biomaterials. Samples with the dimensions of Φ10 × 30 mm were fabricated by SLM. Prior to performing this study, the samples were ground with waterproof emery paper to 1200 grit under running water, then ultrasonically cleaned in acetone for 15 min, rinsed in distilled water and finally dried at room temperature. The static immersion tests were conducted using PBS and Hank's solutions, respectively. A 12.56 ml of each solution was poured into the polypropylene bottles each containing a plate specimen. All the sealed bottles were placed inside an incubator at 37 °C for 7 d. The concentrations of various metals released into solution were determined in ppb (mg/ml) by inductively coupled plasma-atomic emission spectroscopy (ICP-AES, Ultima2).
c 2.6. Statistical analysis The SPSS15.0 statistical software package was used for data analysis. P N 0.05 was considered as no statistically significant difference.
island d
e
3. Results
f
Fig. 1. The illustration of the line scanning strategy (a–c) and island scanning strategy (d–f).
3.1. XRD study Fig. 3 shows the XRD patterns of the unpolished top-view of the lineformed and island alloys. As can be seen, both CoCrW alloys exhibit duplex phases of γ (fcc) and ε (hcp) phase. Interestingly, the islandformed alloy exhibits much higher diffraction intensity assigned to ε phase compared with the line-formed alloy.
Y. Lu et al. / Materials Science and Engineering C 49 (2015) 517–525
519
Fig. 2. Schematic diagram of tensile specimen.
3.2. Microstructure observation Fig. 4 shows the morphologies of the unpolished and polished topview of both alloys. From the morphologies of the unpolished topview in Fig. 4a–b, it is visible that the melted tracks are obviously presented. The solidified lines marked by the red line of the line scanning strategy and island scanning strategy are approximately 70 and 100 μm width, respectively. Some tens-micron-scale un-melted powders are visible on the unpolished surface of line-formed alloy (Fig. 4a) and island-formed alloy (Fig. 4b). However, the more unmelted powders can be observed on the surface of island-formed alloy than that of line-formed one. Fig. 4c–d shows the optical micrographs of the polished top-view of both alloys. Higher porosity is observed in the island-formed alloy (Fig. 4d) compared with the line-formed alloy (Fig. 4c), which may be related to more un-melted powders in the criss-cross regions when used island scanning strategy. Fig. 5 shows the optical micrograph of the top-view of both alloys. Both alloys exhibit distinctive square-like pattern with the surrounding border as denoted by white lines in Fig. 5a–b. The edge length of the square is 60 ± 5 μm, whilst the width of the border is 20 ± 3 μm. The border presents to be darker whereas the square-like structure (central zone) is brighter. The square-like pattern of the line-formed specimen appears to be more uniform compared with island-formed specimen. A furthermore observation of metallographic structure from the top view was obtained by scanning electron microscope as shown in Fig. 6. Striations were detected inside the square-like structure (Fig. 6a–b), indicating that the structure may consist of γ- and ε-phases [23]. The magnified microstructures of regions ①–④ (see Fig. 6a–b) are shown in Fig. 6c–f. It is evident that the border structure shows the presence of fine homogeneous cellular dendrites with a submicron size (Fig. 6c– d), whereas elongated lamellae seem to grow along a direction in the square-like structure (Fig. 6e–f), revealing that grain orientations were different.
3.3. Mechanical property and hardness Fig. 7 shows the typical nominal stress–strain curves of the selective laser melted CoCrW specimens formed by the two kinds of laser
scanning strategies. Both stress–strain curves exhibit uniform elongation followed by a sudden fracture without any macroscopic necking. Such tensile deformation was also observed in the traditional CoCrW and CoCrMo orthopedic alloys [3,24]. The mechanical properties of the alloys listed in Table 3 indicate that the line-formed alloys exhibit slightly higher tensile strength and yield strength when compared with island-formed alloys (p N 0.05), whereas the elongation shows no significant differences (p N 0.05). The yield strength of both CoCrW alloys meets the standard ISO 22674:2006 for dental restorations (N 500 MPa). Fig. 8 shows the SEM images of the fractured surface after tensile test. It can be found that both the alloys exhibit facet-like fracture surfaces. Likewise, striations and wedge type cracks are also clearly found on both fractured surface (Fig. 8a–b), which are dominant for both CoCrW specimens. Under the higher magnification, the dimple-like patterns can be observed on the line-formed alloy (Fig. 8c) and islandformed alloy (Fig. 8d). These results indicate that the fractured mechanisms of all the investigated alloys have the same nature. The average density of the line-formed and island-formed specimens is about 8.537 and 8.526 g/cm3, respectively, that is, the alloys formed at line scanning strategy show slightly better dense than the island-formed alloy (P N 0.05). However, both relative densities exceed 99%, indicating that the two kinds of scanning strategies applied in the study can fabricate dense parts. For the hardness, the line-formed alloys have slightly higher average value of hardness than that of islandformed alloys, 570 ± 3.4HV and 564 ± 2.9HV, respectively. 3.4. Electrochemical test The polarization curves for selective laser melted CoCrW specimens formed at different scanning strategies in PBS and Hanks solutions are presented in Fig. 9. Polarization parameters, including corrosion potential (Ecorr), corrosion current density (Icorr), polarization resistance (Rp), and corrosion rate were obtained from the polarization curves listed in Table 4. It is clearly seen that the typical polarization curves are found to show the representative corrosion behaviors of both CoCrW specimens. The anodic branches of polarization curves for both specimens in PBS and Hanks solutions exhibit a similar passive region at a similar potential range, which are at the potential of 0 to 0.5 V. This passivation behavior is closely related with the formation of protective oxide layer [25]. Also, it has to note that there is a distinct oscillation around Table 2 The processing parameters of SLM.
Table 1 The chemical composition of the CoCrW alloy powders. Chemical Composition (Wt%) Co
Cr
W
Si
Mn
N
Nb
Fe
Ni
Be
60.5
28.0
9.0
1.5
b1.0%
b1.0%
b1.0%
b1.0%
–
–
Scan strategy
Laser power (W)
Scan speed (mms−1)
Powder thickness (μm)
Island size (mm)
Line Island
95 95
700 700
25 25
– 2×2
520
Y. Lu et al. / Materials Science and Engineering C 49 (2015) 517–525
solution have higher corrosion current density than in Hanks solution as well as corrosion rate, indicating that the alloys in Hanks solution appeared to be more susceptible to corrosion than in PBS solution independent of laser scanning strategy. 3.5. Metal release
Fig. 3. XRD pattern of the selective laser melted CoCrW alloys.
0.60 V in the anodic curve of both samples in PBS solution as denoted by rectangle in Fig. 9a. Likewise, a small region of oscillation around 0 V is also found in the anodic curve of line-formed alloy in Hanks solution (Fig. 9b). However, such phenomenon is not observed in the anodic curve of island-formed alloys in Hanks solution. As given in Table 4, the line-formed specimens exhibit a slightly higher corrosion potential (Ecorr) but the slightly lower corrosion current density (Icorr), indicating that the corrosion resistance of the island-formed alloys was slightly greater than that of line-formed specimens. This is further verified by the higher values of polarization resistance Rp and lower corrosion rate of the island-formed alloy in comparison with the line-formed alloy in both simulated body solutions. On the other hand, given the differing solutions the alloys in PBS
a
70µm
The released amounts of the metal in different simulated body solution are shown in Fig. 10 after a 7-day immersion. Regarding the different scanning strategies, the released amount of Co from line-formed specimen is more than that of island-formed alloy in regardless of the differing simulated body solutions seen in Fig. 10a. This indicates that it was easier for elements released from the line-formed specimen in simulated body solutions as compared with island-formed specimen. The difference in the released amount of Cr and W, especially that of Cr, in both solutions is very small (Fig. 10b–c). On the other hand, metal release of the selective laser melted alloys was also affected by simulated body solutions chemistry. As can be seen from Fig. 10a–c, the released amounts of base metal Co and each alloying element differed depending on the type of solutions. The released amount of metal in PBS solution is higher than that in Hanks solution independent of laser scanning strategy, which is consistent with the electrochemical test. In general, the total amounts of metal release were in the range of 0.223–0.276 μg/cm2/7 d, which is significantly lower than that of the standard EN ISO 10271 of 200 μg/cm2/7 d. 4. Discussion 4.1. Phase XRD results indicate that the main phases of both selective laser melted alloys were the γ-phase and ε-phase. Generally, the γ-phase is stable at high temperature above 1120 K, while the ε-phase exist as
b 100µm
Unmelted powder
Unmelted powder
line
200µm
island
c
d
line
island
200µm
Fig. 4. (a) Surface morphology of unpolished line-formed alloy; (b) surface morphology of unpolished island-formed alloy; (c) optical micrograph of polished line-formed alloy; (d) optical micrograph of polished island-formed alloy.
Y. Lu et al. / Materials Science and Engineering C 49 (2015) 517–525
a
521
b
Fig. 5. Top view of metallographic structure: (a) 100× optical micrograph of polished line-formed alloy; (b) 100× optical micrograph of polished island-formed alloy.
an equilibrium phase at room temperature [23,26]. The ε-phase is formed as a consequence of γ → ε transformation while the alloys cooled from high temperatures. Considering the quick cooling rate during the SLM process, the γ-phase could exist at room temperature in this
study. It is noteworthy from the XRD patterns that the peak intensity of island-formed alloy from the ε-phase is greater than that of line-formed alloy. This fact may attribute to the more un-melted powder on the surface of island-formed alloys.
Fig. 6. SEM images of metallographic structure of CrCoW alloy: (a) the microstructure of line-formed alloy; (b) the microstructure of island-formed alloy; (c) the magnified microstructure of region ① in Fig. 6a, (d) the magnified microstructure of region ③ in Fig. 6b; (e) the magnified microstructure of region ② in Fig. 6a; (f) the magnified microstructure of region ④ in Fig. 6b.
522
Y. Lu et al. / Materials Science and Engineering C 49 (2015) 517–525 1200
patterns of island-formed specimen are irregularity and uneven compared with line scanning strategy. The microstructure of both alloys shows a homogeneous cellular structure with cell boundaries. Similar findings also have been reported by some previous reports [28]. This is because of the rapid solidification and strong temperature gradients of molten pool during the SLM [14]. However, the further formation mechanism of the microstructure should be clarified in the further study.
Island Line
tenslie strength/MPa
1000
800
600
4.3. Mechanical property and hardness
400
200
0
4
8
12
16
20
24
28
32
36
40
strain % Fig. 7. Nominal stress–strain curves of the selective laser melted CoCrW alloys.
4.2. Microstructure and possible formation mechanism The microstructure of both selective laser melted CoCrW alloys exhibited two distinct structures, square-like structure and border structure, which were rarely observed in cobalt–chromium based alloys processed by a traditional method. The illustration of possible formation mechanism is shown in Fig. 11. It can be found from the optical micrograph of slightly etched top-view surface that the layer N and layer N + 1 are criss-cross regularly, as can be seen in Fig. 11a. The possible formation mechanism can be deduced as follows. During the SLM, the thermal gradients and growth rates were differing over the melt pool due to the movement of the heat source: the highest ones were located in the central regions of the melting track and then decreased when going towards the border of the melting tracks [27]. The prior melted tracks were heavily remelted and crossed on the subsequent laser tracks, and suffered more sufficient over-heating. In fact, the remelting could be served as a solution annealing at a higher temperature for the prior melted tracks, thereby promoted the growth of elongated grains, as indicated by ① in Fig. 11b–c. In consideration of the width of solidified tracks (approximately 70–100 μm), it is reasonable to form a square pattern with the edge length of approximately 60 μm after the deduction of the width of the overlaps between adjacent melted tracks. Namely, the width of the solidified overlap between adjacent melted tracks was about 10–30 μm, which is nearly correspond with the width of border structure. Regarding the border structure, however, the solidified overlap ③ between aside adjacent melted tracks was partial remelted due to the existence of slight concave ②, when compared with criss-cross central melted tracks ① (see Fig. 11b–c). On the other hand, because of the decrease of thermal gradients and growth rates at borders of the laser scanning passes that even reached zero at the melting border [27]. Correspondingly, the overlaps ③ suffered less over-remelting and overheating as a consequence of impeding the grain growth. Hence, the above analysis can be used to explain why the microstructure of border regions appeared cellular dendrites, while the elongated lamellar structures were located inside the central regions. Additionally, it has been described in the Introduction that the island scanning strategy conducts random laser exposure in each layer; subsequently the island pattern is shifted by 1 mm in both the X and Y directions on the next layer, which can be used to explain why square-like Table 3 The mechanical properties of the both CoCrW alloys. Samples
UTS, (MPa)
σ0.2 (MPa)
E, (%)
Line Island
1158.22 ± 21 1115.56 ± 19
850 ± 23 825 ± 14
9.80 ± 0.10% 10 ± 0.10%
The different laser scanning strategies used in the present study caused slightly different mechanical properties in both specimens that the line-formed alloys exhibit slightly higher UTS and 0.2% yield strength when compared with island-formed alloys, whereas the elongation shows no significant differences. It is well known that mechanical properties are affected by many factors, such as phase, grain orientation, grain boundary conditions, defects, etc. [16]. Given that the phase, grain orientation, and grain boundary conditions of both specimens are very similar, the slight difference in UTS and 0.2% yield strength is obviously associated with the difference in defect and grain size. It is considered that the more dense part without pores exhibits excellent mechanical property; conversely the presence of more pores is a sign of poor mechanical properties, and larger grains are also prone to inducing cracks. In the present work, the island-formed specimens show slightly more pores and larger grain size than that of line-formed one. As a result, the cracks may easily propagate from those pores and larger grains for island-formed specimen. Although it is well recognized that presence of mechanical anisotropy exists in SLM parts, most of them attributed this slightly difference in this study to the inner porosity [29]. Indeed, compared with the selective laser melted parts as reported by previous study [4], the CoCrW alloys fabricated by differing laser scanning strategies exhibited better 0.2% yield stresses and tensile strength. This is due to the homogeneous microstructure with diamond patterns that endowed with adequate mechanical property. However, both alloys showed relatively poor ductility as compared with previous study. The ε-phase is known to enhance the strength and wear resistance of the CoCr-based alloy. But the formation of ε-phase also leads to poor ductility [30]. During the tensile test, strain-induced martensite ε-phase intersected with the preexisting athermal ε martensite may induce the stress concentration at their interfaces and/or grain boundaries resulting in premature fracture before the onset of plastic instability, consequently caused detrimental to tensile ductility [23]. As a result shown by Yamanaka et al. [26], the CoCrW alloys exhibited the better tensile ductility when microstructures contained the less martensite ε phase. During the SLM process, it is recognized that molten pool undergoes a fast cooling rate, leading to the restraining transformation of metastable γ-phase to the martensite ε-phase. However, such transformation would be occurred when the previous solidified layers were reheated and remelted by the subsequent layer process [14]. That is, the γ-phase and ε-phase coexisted in the alloys. Therefore, the reason for the relatively lower elongation in the present study is the major result from the rich ε-phase in alloy. From fracture surface observation in Fig. 8a–b, facet-like fracture surfaces and wedge type cracks are clearly observed on fracture surfaces similar to the previous reports [31]. As suggested by a previous study [23], this kind of fracture mode observed herein can be recognized as a quasi-cleavage type fracture since a true cleavage fracture cannot occur in fcc polycrystalline materials [24,32]. The type fracture was enhanced by the athermal ε martensite during the tensile test [33]. Wedge type cracks may be formed by the cleavage fracture along the {1 1 1}, while the dimple-like patterns may be induced from the cellular dendrites [24]. Hardness is an important characteristic in a way to evaluate whether alloys combine with greater elastic modulus to allow the manufacture of longer prostheses. Result on hardness measurement indicates the
Y. Lu et al. / Materials Science and Engineering C 49 (2015) 517–525
a
b
z
z
x
20µm
x
c
d
z
z x
5µm
x
523
20µm
5µm
Fig. 8. SEM images of the fracture surface of the selective laser melted CoCrW alloys after tensile test. (a) fracture surface of line-formed; (b) fracture surface of island-formed;(c) the magnified fracture surface of Fig. 8a; (d) the magnified fracture surface of Fig. 8b.
hardness value of line-formed alloys was slightly higher than islandformed alloys. This fact may be due to the slightly higher density of line-formed alloys. Henriques et al. [2], in their study on the hardness of hot pressed CoCrMo samples for surgical applications, also reported higher hardness in hot pressed specimens than in cast specimens. They demonstrated that the greater hardness of hot pressed specimens was related with low porosity relatively to cast ones. Therefore, the line scanning strategy can fabricate more dense part in comparison with island strategy.
4.4. Corrosion resistance and metal release In vitro studies, i.e., electrochemical and metal release tests, were conducted in simulated body solutions give an overview of the corrosion resistance and biocompatibility (metal release) of the selective laser melted alloy. The anodic polarization curves show that both alloy formed by line and island scanning strategies were in the passive state, which was a characteristic of materials with high resistant to corrosion. The oscillation around 0.60 V recognized as an oxidation peak
Fig. 9. The polarization curves of selective laser melted CoCrW specimens formed at different scanning strategies in PBS solution (a) and Hanks solution (b).
524
Y. Lu et al. / Materials Science and Engineering C 49 (2015) 517–525
Table 4 The polarization parameters obtained from the polarization curves. Samples
Icorr (nA)
Ecorr (mV)
Rp (Ω)
Corrosion rate (10−4 mm/year)
Line–PBS Island–PBS Line–Hanks Island–Hanks
20.20 19.70 12.00 8.02
−201 −200 −284 −327
1.790 1.855 1.325 1.914
4.60 4.50 2.70 1.80
may be attributed to the competition between formation and dissolution of passive film [34]. The corrosion potential (Ecorr) of the lineformed alloys was shifted to more noble potentials (more than 40 mV). However, the corrosion current density (Icorr) of the lineformed alloys was slightly greater than that of island-formed alloys. This indicates that the slightly higher corrosion rate occurred in the line-scanning alloys in simulated body solutions, which was verified by the fact that the line-formed alloys exhibited slightly lower polarization resistance Rp and higher corrosion rate with the respect to islandformed alloys. Results from the metal release revealed that Co was more active than Cr and W independent of processing parameters and simulated body solutions. Such similar results were found in previous studies that Co is preferentially released from CoCrMo alloy [24]. The quantity of the Co released in PBS solution from selective laser melted CoCrMo in previous report was approximately 0.45 μg/cm2/7 d [14]. In our test, the amount of Co release from the selective laser melted CoCrW alloys into PBS was less than 0.22 μg/cm2/7 d. Likewise, the quantity of Cr and W released from the CoCrW alloys used in this study was very small. It is noteworthy that the quantity of Cr was lower than that of W in both simulated body solutions, indicating that elements were not absolutely determined among component elements of the selective laser melted alloys. Otherwise, a higher amount of Co released from line-formed alloys was detected than that of island-formed alloys in Hanks and PBS solutions. Although it has been determined that alloy that existed more porosity that corresponded to bigger surface area are regarded for higher corrosion current densities, and that the surface quality could affect the corrosion resistance [35]. However, based on the studies of the electrochemical and metal release tests, the porosity and surface quality are not in the scope of the factors. Assuming that thermal residual stress, which had accumulated from the repeated rapid heating and cooling of molten pool during the process, can be used to explain why the difference in corrosion behavior. As mentioned above, the island scanning strategy make the laser randomly exposed on each individual layer, thereby ensures a significant reduction in stresses within the alloy. An ongoing quantified examination should be conducted to assess how the thermal residual stress influents on the corrosion resistance.
The corrosion behavior of cobalt-based alloy is very sensitive to the solution composition [36]. The CoCrW alloys in the PBS solution had slightly higher corrosion rate than in Hanks solution. According to the electrochemical and metal release analysis, both CoCrW alloys showed the higher release amount of Co and lower corrosion current density in PBS than in Hanks solution. This suggested that the alloys in PBS solution appeared to be more susceptible to corrosion than in Hanks solution regardless of laser scanning strategy. Generally, when CoCrMo alloy is immersed in Hanks and PBS solution a passive layer is formed on surface. The passive layer was characterized as containing oxides of cobalt, chromium and some hydroxides [28,37], which played an important role as an inhibitor of ion release. Namely, the corrosion reaction of the materials with fluid usually occurs at interfaces. According to previous study [38], the activity of the cobalt oxide layer decreased due to the formation of insoluble cobalt phosphate when CoCrMo alloy was placed in the presence of NaH2PO4. Since the Hank's solution contains both NaH2PO4 and Na2HPO4, it is probable that the insoluble cobalt phosphate was formed in the early stage. This is coherent with the early appearance of the oscillation in anodic polarization curves of line-formed alloy as indicated by the rectangle in Fig. 9b. Actually, corrosion resistance of metallic biomaterials depends on the passive layer and the compositions of the layer are usually important for biocompatibility [34]. Therefore, the surface structure and composition in the passive layer of the selective laser melted CoCrW alloy should be analyzed for further investigation, such as XPS study. It has to be mentioned that the residual stress is generated during the SLM in some degree regardless of line and island laser scanning strategies, which will affect the accuracy of the as-SLM parts. For the practical application, the heat treatment should be conducted to release the residual stress. Therefore, further studies focusing on microstructure, mechanical properties, and precision of the SLM parts after heat treatment will be investigated. 5. Conclusions In the study, two dense CoCrW alloys were formed by line and island scanning strategy, respectively. Due to the rapid cooling rate and strong temperature gradient during the laser melting process, the metallographic structure of the top-view presented square-like pattern with fine cellular dendrites. Results in terms of tensile, hardness, density, electrochemical, and metal release tests suggested that there was no or small difference between the two CoCrW alloys. This should be attributed to their similar microstructure. The tensile test indicated that the yield strength of both alloys met the standard ISO 22764:2006 for dental restorations. The failure type was recognized as a type quasicleavage fracture. The electrochemical and metal release tests suggested that the island-formed alloys showed slightly better corrosion
Fig. 10. Released amounts of the metal in PBS and Hanks solution.
Y. Lu et al. / Materials Science and Engineering C 49 (2015) 517–525
525
Fig. 11. (a) Slightly etched top view surface; (b) illustration of formation of square-like structure in section view of Y-Z axis; (c) illustration of formation of square-like structure in top view of X–Y axis.
resistance than line-formed ones, i.e., lower current densities (Icorr), corrosion rate and lower amounts of metal release in PBS and Hanks solution. The total quantity of Co, Cr and W used in this study was in the range of 0.223–0.276 μg/cm2/7 d for both alloys. Based on the results obtained in this study, line and island scanning strategy in the SLM process for the CoCrW alloy is a promising candidate for use in dental applications. Acknowledgments This work was financially supported by National pilot projects (No. XDA09020301). References [1] G. Manivasagam, D. Dhinasekaran, A. Rajamanickam, Recent Patents Corros. Sci. 2 (2010) 40–54. [2] B. Henriques, D. Soares, F.S. Silva, J. Mech. Behav. Biomed. Mater. 12 (2012) 83–92. [3] K. Yoda, Suyalatu, A. Takaichi, N. Nomura, Y. Tsutsumi, H. Doi, S. Kurosu, A. Chiba, Y. Igarashi, T. Hanawa, Acta Biomater. 8 (2012) 2856–2862. [4] A. Takaichi, Suyalatu, T. Nakamoto, N. Joko, N. Nomura, Y. Tsutsumi, S. Migita, H. Doi, S. Kurosu, A. Chiba, N. Wakabayashi, Y. Igarashi, T. Hanawa, J. Mech. Behav. Biomed. Mater. 21 (2013) 67–76. [5] K. Yamanaka, M. Mori, A. Chiba, Mater. Des. 57 (2014) 421–425. [6] M. Mori, K. Yamanaka, A. Chiba, J. Alloys Compd. 590 (2014) 411–416. [7] A. Karaali, K. Mirouh, S. Hamamda, P. Guiraldenq, Mater. Sci. Eng. A 390 (2005) 255–259. [8] K. Yamanaka, M. Mori, A. Chiba, Mater. Lett. 116 (2014) 82–85. [9] Z. Zhang, D. Jones, S. Yue, P.D. Lee, J.R. Jones, C.J. Sutcliffe, E. Jones, Mater. Sci. Eng. C 33 (2013) 4055–4062. [10] B. Song, S. Dong, P. Coddet, H. Liao, C. Coddet, Mater. Des. 53 (2014) 1–7. [11] B.-A. Lee, O.-S. Kim, M.-S. Vang, Y.-J. Park, J. Prosthet. Dent. 109 (2013) 95–105. [12] N. Xiang, X.-Z. Xin, J. Chen, B. Wei, J. Dent. 40 (2012) 453–457.
[13] X.Z. Xin, N. Xiang, J. Chen, B. Wei, Mater. Lett. 88 (2012) 101–103. [14] Y.S. Hedberg, B. Qian, Z. Shen, S. Virtanen, I. Odnevall Wallinder, Dent. Mater. 30 (5) (2014) 525–534. [15] B. Song, S. Dong, H. Liao, C. Coddet, Int. J. Adv. Manuf. Technol. 61 (2011) 967–974. [16] F. Liu, X. Lin, C. Huang, M. Song, G. Yang, J. Chen, W. Huang, J. Alloys Compd. 509 (2011) 4505–4509. [17] J.P. Kruth, J. Deckers, E. Yasa, R. Wauthle, Proc IME B J Eng Manufact 226 (2012) 980–991. [18] B. Vrancken, V. Cain, R. Knutsen, J. Van Humbeeck, Scr. Mater. 87 (2014) 29–32. [19] Concept laser, http://www.concept-laser.de/2014. [20] M.F. Zaeh, G. Branner, Production Engineering 4 (1) (2009) 35–45. [21] L. Thijs, K. Kempen, J.-P. Kruth, J. Van Humbeeck, Acta Mater. 61 (2013) 1809–1819. [22] J.P. Kruth, M. Badrossamay, E. Yasa, J. Deckers, L. Thijs, J. Van Humbeeck, Proceedings of the 16th International Symposium on Electromachining, 2010. [23] K. Yamanaka, M. Mori, A. Chiba, Mater. Sci. Eng. A 528 (2011) 5961–5966. [24] K. Yamanaka, M. Mori, A. Chiba, J. Mech. Behav. Biomed. Mater. 37 (2014) 274–285. [25] D. Sun, J.A. Wharton, R.J.K. Wood, L. Ma, W.M. Rainforth, Tribol. Int. 42 (2009) 99–110. [26] K. Yamanaka, M. Mori, A. Chiba, Mater. Sci. Eng. C 40 (2014) 127–134. [27] L. Thijs, F. Verhaeghe, T. Craeghs, J.V. Humbeeck, J.-P. Kruth, Acta Materialia 58 (9) (2010) 3303–3312. [28] X. Xin, N. Xiang, J. Chen, D. Xu, B. Wei, J. Mater. Sci. (2012) 1–8. [29] C. Qiu, N.J. Adkins, M.M. Attallah, Mater. Sci. Eng. A 578 (2013) 230–239. [30] K. Yamanaka, M. Mori, Y. Koizumi, A. Chiba, J. Mech. Behav. Biomed. Mater. 32 (2014) 52–61. [31] K. Yamanaka, M. Mori, A. Chiba, J. Mech. Behav. Biomed. Mater. 29 (2014) 417–426. [32] J. Morris, Z. Guo, C. Krenn, Y.-H. Kim, ISIJ Int. 41 (2001) 599–611. [33] T. Matković, P. Matković, J. Malina, J. Alloys Compd. 366 (2004) 293–297. [34] T. Hanawa, Mater. Sci. Eng. C 24 (2004) 745–752. [35] W.C. Rodrigues, L.R. Broilo, L. Schaeffer, G. Knörnschild, F.R.M. Espinoza, Powder Technol. 206 (2011) 233–238. [36] M. Metikoš-Huković, R. Babić, Corros. Sci. 49 (2007) 3570–3579. [37] A. Hodgson, S. Kurz, S. Virtanen, V. Fervel, C.-O. Olsson, S. Mischler, Electrochim. Acta 49 (2004) 2167–2178. [38] L. Burke, M. Lyons, O. Murphy, J. Electroanal. Chem. Interfacial Electrochem. 132 (1982) 247–261.
