The Effect of Rapid Solidification on Microstructure and Corrosion of Advanced Biomaterial Co-Cr-Mo-C Alloy

In this research, the microstructure and corrosion properties of rapidly solidified Co-Cr-Mo-C alloy as an advanced biomaterial alloy were studied. The use of rapid solidification casting method represents significant changes in not only the amount of formed ε-HCP phase, which is strongly influenced by rapid solidification, but also in electrochemical behavior and solidified structure. In this research, rapid solidified Co-Cr-Mo-C alloy is studied using OM, SEM, EDS, XRD, and dynamic potentiostate. Co-alloy ingots were melted into an induction furnace filled by argon gas and casted into a V-shape sand and chill copper molds to prepare rapid solidified samples and its properties were measured in different cooling rates. The microstructure examination demonstrating the structure of alloy is mainly consist of columnar dendritic structure with the distribution of carbides within primary and secondary dendrites arms and finer dendritic structure along with modified carbide distribution will be achieved by rapid solidification. This structure will improve alloy’s corrosion behavior and reduces its corrosion rate when it is tested in Ringer’s solution as an electrolyte.


Introduction
Co-based alloys are increasingly being used in biomedical application for their biocompatibility, excellent corrosion resistance, and mechanical and wear properties [1,2]. Solidified microstructures of as-cast cobalt alloys is mainly FCC dendritic structure in presence of segregation, and second phase precipitates within the matrix and along the interdendritic regions [3]. Since the introduction of Co-based alloys as prosthesis material, a considerable amount of research has been done on examining the possibility of enhancing mechanical properties and biocompatibility of this alloy by modifying the microstructure [1,[3][4][5][6].Overall, Co-Cr-Mo-C alloys made by investment casting are showing poor ductility, high shrinkage porosity, interdendritic segregation, and intermetallic compounds brittleness. Thus, to improve alloy performance for high lifetime service in biomedical applications, modifications in alloy design and casting technology have taken place [7]. Minimizing the formation of brittle intermetallic compounds along the interdendritic regions during solidification have been considered as part alloy design solutions [8].
In pure Co, the γ ↔ ε FCC to HCP allotropic transformation takes place at a Tc = 417 o C (690 K) by a displacive martensitic transformation [9]. The transformation temperature, is even shifted toward higher temperatures (near 970 o C [10]) in Co-27Cr-5Mo-0.05C wrought alloys as Cr and Mo expand the HCP field of stability. The resultant HCP phase from the martensitic transformation is known as ε-martensite [9]. According to phase diagram of Co-Cr; in solidified alloy at room temperature ε-HCP phase and at temperatures higher than 800 o C γ-FCC phase are expected to be thermodynamically stable. However, the γ ↔ ε is strongly restricted by nucleation phenomena and will lead to formation of retained γ-FCC phase in most conventionally solidified Co-Cr-Mo-C alloys at room temperature as well as higher temperature [11,12] because the γ ↔ ε, FCC to HCP, transformation is rather slow under normal cooling conditions [13,14]. In Co alloys, the martensitic transformation can be achieved by (1) isothermal, (2) athermal, and (3) strain induced mechanism.
In athermal transformations the amount of transformed martensite depends primarily on the quenching temperature and not on the residence time at this temperature. Experimentally, it has been found that in conventionally solidified Co-Cr-Mo-C alloys the reported amounts of athermal ε-martensite are not more than 20 vol. % in most cases [15]. The limited extent of athermal martensite induced in Co-Cr-Mo alloys has been attributed to the lack of enough defects necessary for the spontaneous formation of ε-embryos during alloy cooling from the γ -phase [16].
Athermal reaction is controlled by distribution of faulting defects which act as effective nucleation sites for ε-martensite embryos. These faulting defects including tilt boundaries, twin intersections, incoherent twin boundaries or incoherent inclusion interfaces expedite the formation of ε-martensite embryos by having a longrange stress field [6]. Olson et al. [17] have shown that the spontaneous formation of ε-martensite embryos can be achieved when the alloy stacking fault energy becomes zero or negative. The nucleation defect being a faulting mechanism consisting of a group of lattice dislocations where the motion of a Shockley partial on every second FCC closed packed plane gives rise to an ε-phase nucleus. Yet, the estimated density of the proper faulting defects is of the order of 10 5 -10 7 /cm 3 , which explains the sparse amounts of precipitated athermal martensite in cobalt based alloys [17].
Many studies have shown that the metal ions dissolved from prostheses in body because of the corrosion and wear debris in biological environment cause an allergy and implant failure [13,18]. Investigations on the corrosion characteristics of Co-Cr-Mo-C alloy in different body solutions using potentiodynamic method have been done and shown high corrosion resistivity and small passivity region for this alloy in joint fluid and body fluid simulated environments [1,5,7,[19][20][21]. S. Hiromoto et al. [1] considered the effect of forging ratio on corrosion behavior of Co-Cr-Mo alloys in different environment, they found that alloy with low forging ratio shows lower passive current density and also increasing the grain boundaries by forging will decrease the corrosion resistant of alloy. F. Ren et al. [20] investigated the effect of physiological simulated solution on corrosion behavior of ultra-fine grain Co-28Cr-6Mo made by mechanical alloying process. Results demonstrate positive shift in corrosion potential and lower corrosion density as compared to conventional Co-Cr-Mo cast alloys. The study on influence of precipitate carbides on corrosion behavior of a biomedical Co-Cr-Mo-C alloy shows at high anodic potential (0.5-0.7 Vsat Ag/AgCl) metal dissolution occurs at the carbide boundaries and, in the form of etching-like dissolution, on the carbides.
Despite the tremendous volume of research that has been done on the properties and modifications of Co-Cr and Co-Cr-Mo-C alloy over the past several decades, there is no published article on the effect of solidification rate and rapid solidification on the mechanical, physical, and electrochemical properties of Co-Cr-Mo-C alloy. In this investigation, samples were prepared by rapid solidification. Microstructure characteristics of as-cast and rapidly solidified implant alloy were studied. Potentiodynamic method is used to investigate the corrosion behavior of samples in biological Ringer's solution. Experimental results from both microstructure and electrochemical analysis were compared and discussed.

Materials and methods
Cobalt alloy parts containing 28 wt. % Cr, 6 wt. % Mo, and 0.3 wt. % C in conform to ASTM F-75 were melted in an alumina crucible using a vacuum induction furnace. The system was evacuated three times with high purity argon prior to vacuum melting. When the parts turned into liquid at 2273 K (approx. 2000 o C); the chamber was immediately evacuated, and the melt was degassed. On the other hand, the temperature was dropped to 1873 K (approx. 1600 o C) and a slag was collected from the top of the melt. Finally, the molten alloy was cast into a wedge-shaped sand and copper mold. The copper wedge mold sides were covered by an alumina sheet to keep the heat flow predominantly in one direction, (Figure 1). Table 1 shows the chemical composition of cast alloy. Before casting, Pt/Pt-18% Rh thermocouples were centrally inserted into the wedge copper mold at 20, 60, and 90 mm from the tip to the top of the mold cavity, at half thickness locations of 1.0, 3.0, and 4.5 mm, respectively ( Figure  1). The as-cast ingots were then sectioned along the lateral direction in the plane normal to the tip of the wedge and polished followed by electrolyte etching using a 60 vol. % HNO3 + 40 vol. % H2O solution at 6 V and 5 mA for 20 s. Preparedsamples were studied by means of optical microscopy and scanning electron microscopy (SEM). SEM and EDX analysis of specimens were performed using a Topcon SM300 Scanning Electron Microscope at 20KV of intensity.
The samples of as-cast alloy and rapidly solidified alloy from different locations of "V" shaped mold were finished by 1200 SiC abrasive paper and polished by Al2O3 1μm solution for corrosion tests. Biological Ringer's solution consists of 9 g NaCl, 0.43 g KCl, 0.24 g CaCl2, and 0.2 g NaHCO3 in 1 liter of distilled water was used as an electrolyte. Super saturated Ag/AgCl and Pt plate were used as reference electrode and counter, respectively. The sample was fixed in an electron holder with O-ring [22]. To stabilize the potential, samples immersed in the solution for about 15 min. The stabilized potential was used as the open-circuit potential (Eocp) of samples. Samples were polarized in a range of -0.3 V to +1.3 V versus Ecorr with the scanning rate of 10 mV min -1 . In order to measure the corrosion rate, a potentiostat Biologic SP-200 and EC-Lab v10.33 as the corrosion software were used for electrochemical control and data analysis, respectively. The test was repeated three times for each sample to have reliable results. The corrosion current density (icorr) and other corrosion parameters including anodic and cathodic Tafel slopes (ba and −bc), were measured by considering the polarization curves by Tafel extrapolation.

Results and discussion
The microstructure of the as-cast Co-Cr-Mo-C alloy consists of cobalt-rich matrix with dendritic structure and a relatively large volume fraction of interdendritic precipitates. During the solidifying stage, some amount of the metastable phase will retain at the room temperature when Co-Cr-Mo-C alloy solidified through the FCC phase field. The resultant will be cobalt-rich matrix phase consists of residual phase and phase coupled with intragranular striations. The interdendritic precipitate was identified as carbide that also exists continuously at the grain boundaries. This brittle phase corresponding to the higher solute content (Cr, Mo) at grain boundaries causes a stress concentration and leading to the early fracture. The intragranular striations are related to ε-martensite phase formed because of stacking faults. Cr and Mo solutes can diffuse preferentially at the stacking faults with increasing temperature [23,24].
The maximum cooling rate with this experimental measured at: 450 K/s at a distance 20 mm from the tip of the "V" shape ingot; 300 K/s at 60 mm at the center of the ingot; 120 K/s at 90 mm at the top of the ingot; and conventional cooling rate from sand cast sample approximately 10 K/s. Figure 2and 3 show the various microstructures regarding the cooling rate found in the studied locations (at 20 mm, 60 mm and 90 mm from bottom to top in the rapid solidified wedge ingot and top of the sand mold). It can be clearly observed that the dominant microstructure is the columnar dendrites. Microstructure analysis of Co alloy revealed the presence of blocky and lamellar (dendritic) shape carbides in this alloy (Figure 4), depending on its cooling rate. Physical and chemical factors have a great influence on the carbide formation: the main alloying elements are responsible for their composition and structure, and the cooling rate largely affects their size and shapes. According to Tylor and Waterhouse [25] these carbides are most likely to be M23C6 and M6C. Cr has a predominant role in the formation of the carbide type M23C6 (where M = Co, Cr, Mo), but also Mo, when present in large quantities, can promote carbide precipitation in the form of M6C. Figure  4 shows two types of carbides, with their corresponding EDS analysis, formed in Co-Cr-Mo-C alloy due to its cooling rate. It is found that in low cooling rate solidification, carbides are larger and dendritic shape while in high cooling rate solidification; carbides are blocky shape and finer in compare with those in low cooling rate. It can be understood from the EDS results that in low cooling rate the major alloying element in carbides is Mo (Figure 4c), while Cr is the predominant alloying element in high cooling rate Co alloy and represent blocky shape carbides along with Mo (Figure 4d).  Samples were analyzed using X-ray diffraction to measure the quantity of developed γ-FCC and ε-HCP phases. On all samples, XRD peaks originated from γ-FCC and ε-HCP phases were observed. Sage and Guillaud [26] proposed the quantitative model to calculate the developed HCP phase in the Co-alloy. According to their model, only the diffraction patterns corresponding to (002)FCC and (101 � 1)HCP should be considered because they are the only well distinguished and not overlapped diffraction patterns at a given 2θ. The intensities of (002)FCC peak at 50.5 o and (101 � 1)HCP at 46.5 o are used to calculate the weight percentage of γ-FCC and ε-HCP, respectively. The following equation [26] is used to calculate the FCC and HCP relative amounts quantitatively in Co-Cr-Mo-C alloy.
Where is HCP weight percentage, (002)FCC and (101 �1)HCP are integrated areas of the intensity peak correspond to γ-FCC and ε-HCP phases, respectively. Quantitative measurements of ε-HCP developed phase by using Eq. 1 determines the formation of approximately 46% ε-martensite (HCP) phase in the sample with the highest cooling rate and less than 1% in the sample with the lowest cooling rate, Figure 5. Intrinsic stacking faults are known to be the potential sites for ε-HCP phase nucleation as a result of lattice dislocations dissociate into Shockley partials. Because of relatively small stacking fault energy of Co-alloy, this transformation happens in Co-alloy spontaneously [6,27]. This will show that rapid solidification is an effective method to increase the density of faulting defects and consequently prompt transformation of γ → ε. Although, the best wear properties in Co-based alloys are found when the alloy matrix is fully HCP, either in metal on metal wear couples, or in metal on UHDPE wear couples [28,29], the fatigue properties are significantly reduced when the matrix is fully HCP [15]. Apparently, the limited number of slip systems in the HCP matrix severely limits plastic deformation, thus promoting fatigue cracking. Moreover, the mechanical properties (strength and ductility) of Co-based alloys are strongly influenced particularly by the amount of carbon, nitrogen, grain size and by the processing method (casting versus brought or powder processing). Nevertheless, it is worth mentioning that modifications to these properties, particularly ductility have been found through additions of nitrogen and grain size control [30,31]. Thus, It is very important to note the dendrite arm spacing (DAS) and secondary dendrite arm spacing (SDAS) for evaluating microstructure of this alloy. In this research work, SDAS lower than 5 μm were obtained in the bulk microstructure of samples located on 20 mm from copper mold. In addition, precipitates are randomly distributed in the primary and secondary arms of the matrix, which were identified with scanning electron microscopy. As it is shown in figure 2, increasing the cooling rate will decrease the DAS, SDAS, and carbide size which have improvement effects on alloy strength and its hardness. Figure 6 shows the anodic and cathodic polarization curves for Co-Cr-Mo-C alloy in rapidly solidified and sand cast conditions in Ringer's solution. Corrosion parameters such as Ecorr, icorr, ba and bc are measured and reported in Table 2. According to the results, increasing the cooling rate will improve the electrochemical behavior of alloy. The samples with higher cooling rate show higher corrosion potential and lower corrosion rate. The results show that the sand cast sample has the lowest Ecorr and higher icorr than the copper mold which was cast at 20 mm from the tip of the mold sample. The cathodic current densities show a rather linear region with negative slope and then the slope increases to zero. No significant differences are detected in the cathodic current densities depending on the samples. Solidified structure, size and distribution of carbides and the amount of ε-HCP phase characterize each sample. This means that different microstructural properties affect the corrosion behavior of samples. Better corrosion resistivity obtained by highest cooling rate, rapid solidification, is believed to be the result of better distribution of alloying elements, carbides, and lack of grain boundaries. Figure 2 shows the optical image of samples microstructure solidified at different cooling rate. As it shown in Figure 2, increasing cooling rate will highly reduce the density of grain boundaries. It is well known that the excellent corrosion characteristics in Co-Cr alloys are due to the passive layer formed on its surface [1,7,32] which prevents these materials from suffering greater damage even in biological environments. On the other hand, the protectiveness of passive oxide layer in sand cast sample is higher than those with 120 and 230 K/s cooling rate according to its wider passivity region. The reason of narrower passivity region of mentioned samples can be because of carbides size and distribution. The dispersion of carbides causes the increase in local active cell between carbides and base alloy, and then the increase in local corrosion is the result which will lead to pitting corrosion and consequently damaging the passive layer [1]. Local corrosion during passivity causes failure in oxide film and as a result, smaller passivity region. The increase in grain boundaries volume fraction and the precipitates in grain boundaries cause the decrease in protectiveness of the passive oxide film [1].

Conclusions
In this research, an alternative solidification process is proposed which could be used to manufacture Co-Cr-Mo-C alloys with unique microstructures. 1) Dendrite refinement and either elimination or reduction of interdendritic segregation in cast alloys was observed, which can be considered as one of the microstructural achievements due to rapid solidification. The morphology and size of the carbides were highly affected by cooling rate.
2) Rapid solidification effects, such as excess vacancies and the development of numerous stacking faults and corresponding intersections, strongly favored the athermal martensite. The athermal martensite microstructure was confirmed by means of XRD. Results show the increase of formed ε-HCP phase to 46% by rapid solidification.
3) Potentiodynamic polarization curves indicated that corrosion resistance of Co-Cr-Mo-C alloy produced by rapid solidification will increase by increasing the cooling rate. Increasing the cooling rate from 10K/s to 450 K/s will raise the corrosion potential from -280 V to -240 V and reduce the corrosion rate from 0.09 µA/cm 2 to 0.02 µA/cm 2 .