versão On-line ISSN 0104-6632
Braz. J. Chem. Eng. v.18 n.1 São Paulo mar. 2001
IN-VITRO CORROSION RESISTANCE STUDY OF HOT WORKED Ti-6Al-7Nb ALLOY IN A ISOTONIC MEDIUM
(Received: October 20, 1999 ; Accepted: November 16, 2000)
Abstract - This investigation reports the results of linear polarization of hot upset Ti-6Al-7Nb bar samples. Current-potential curves recorded in Hanks solution were analyzed by correlating characteristics of passivation and microstructures obtained after processing. Results have shown that it is important to select temperature process and deformation rate as parameters when more noble potential values are required. Low deformation rate facilitates the formation of beta phase that is retained in the structure at room temperature shifting the corrosion potential to more positive values. However, samples hot compressed from 750 °C to 1030 °C showed passive layer stability over a wide range of potentials extending from 0.15 V to 1.75 V vs SCE . Furthermore, passive films grown onto the Ti-6Al-7Nb samples surfaces have shown no sign of rupture for the processing conditions selected for this study.
Keywords: titanium alloys, metallic prothesis, corrosion of titanium alloys, prosthetic materials.
Titanium and titanium alloys were introduced in orthopedic surgery in the sixties. Since then titanium-based alloys have been used for orthopedic implants due to their outstanding properties of biocompatibility, corrosion resistance, and high strength resistance to weight ratio. These materials exhibit instantaneous surface passivation, and bony tissue therefore grows onto the rough oxide film. The idea of developing a Ti-Al-Nb alloy for use in implants, rather than the traditional Ti-6Al-4V, was based on suspicions that vanadium oxide is thermodynamically unstable, releasing vanadium which diffuses through the tissues of the human body. Requirements for the new alloy were biocompatibility, the presence of all alloying elements in the passive layer, a stabilized a + b structure allowing superplastic deformation, and mechanical properties comparable to those offered by the Ti-6Al-4V alloy. The composition which met these initial requirements is Ti-6Al 7Nb (Semlitsch et al.,1992). While pure titanium is a monophase, physiologically inert, and a non toxic metal, ternary titanium alloys containing Al and V or Nb also present an a + b phase structure that has attractive mechanical properties. Indeed, titanium alloys may be grouped into five alloy types categories, depending on alloy composition and heat treatment: alpha, near-alpha, alpha-beta, near-beta and beta. Pure titanium undergoes an allotropic transformation at about 885 oC, changing from an alpha phase (h.c.p.) to a beta phase (b.c.c.), and the alloying elements may either lower or raise the phase transformation (beta transus) temperature. Aluminium, as alpha phase stabilizer, is highly soluble in both the a and b phases. Vanadium and niobium are both beta phase stabilizers and also highly soluble in both phases. As such, the beta phase tends to be enriched in Nb or V and leaner in Al, compared to alpha phase (Boyer and Gall, 1985; Collings, 1984). As the prosthetic devices are manufactured by hot forging under different temperature conditions and deformation rates, presenting regions with greater deformation, processes of recovery or partial recrystallization may or may not take place during processing. In these cases different microstructures and a/b phase ratios are expected, resulting in varying mechanical properties and corrosion resistance. As a consequence, processes involving metal release into tissues by passive dissolution or wear may be facilitated.
Even considering that during the last three decades many alloys have become available for orthopedic surgery, the study of new titanium alloy compositions is still a subject of interest. While the literature provides studies on mechanical properties and corrosion resistance and their relationship with microstructure, processing and surface treatment for Ti-Al-V and Ti-Al-Nb alloys (Aragon and Hulbert, 1972; Semlitsch et al, 1985; Noort, 1987; Shepard and Norley, 1988; Streicher, 1990; Semlitsch, 1987; Khan et al, 1996; Simpson, 1986; Oliveira et al, 1998), new Al-free compositions are also being considered. Beta-type Ti-Nb-Ta-M alloys, with M being Zr, Mo or Sn appear to be the current choice for medical implants (Niinomi, 1998; Akahori and Niinomi, 1998; Kuroda et al, 1998) , due to their greater strength and lower elastic modulii.
Considering the knowledge gained by perusal of these papers and the results they present, we have become interested in investigating the electrochemical performance of the Ti-6Al-7Nb alloy after processing by hot compression, considering several temperatures and deformation rates that may produce different degrees of deformation under compression. Furthermore, this investigation reports the results of linear polarization of hot upset Ti-6Al-7Nb samples, and I-E curves recorded in Hanks solution were analyzed by correlating characteristics of passivation and microstructures obtained after processing.
IMI-317 Ti-6Al-7Nb alloy was supplied by IMI Titanium LTD. in 17 mm diameter cylinders in the hot rolled and annealed (700 oC during 1 hour and then air cooled) condition, with the composition shown in Table 1.
Samples were hot upset in a MTS model 810 testing machine, equipped with model 418 a function generator. Samples were heated in an electrical resistance furnace, in air. Three thermocouples monitored the temperatures of the samples, with two thermocouples on the upper and lower dies, and on the sample itself. The test matrix consisted of four temperatures (750, 850, 950 and 1030 oC) and three deformation rates ( 0.01, 0.1 and 1.0 mm s-1) for hot upset forging as outlined in Table 2. Heating time ranged from 15 to 45 minutes with a soak time of 10 minutes. After upsetting, the samples were air cooled.
Two groups of samples were used for the deformation experiments: cylinders with a diameter of 12 mm and of 20 mm for a temperature of 1030 ºC, and cylinders with a diameter of 8 mm and a height of 12 mm for the other temperatures. Upset samples were cut in the longitudinal section, embedded in bakelite for 20 minutes at 220 ºC, and then ground and polished. This operation comprised the use of emery paper of 600 to 1200 grid and polishing cloth with 6 mm, 1mm, 0.5 mm and 0.05 mm diamond paste. After being etched for 5 seconds with Kroll solution, the samples were examined and photographed in an optical microscope. For the polarization experiments, only the first part of the above procedure, of abrading and polishing, was used followed by sonication in acetone.
The electrochemical cell used in the polarization experiments, shown in Figure 1, was a three-electrode cell with a large surface platinum foil counter electrode and a saturated calomel reference electrode (SCE). The body of the cell consisted of a glass tube with a two-inch diameter and a length of 10 cm, supported by two PTFE end plates. The reference electrode entrance had a Luggin capillary termination placed at a distance of 2 mm from the working electrode. Figure 1 also shows a hole drilled at the center of the left end plate such that the exposed working electrode area was 0.95 cm2. The Ti-6Al-7Nb alloy samples were pressed into this hole, facing the Luggin capillary, by a copper current feeder.
All the polarization experiments were carried out at room temperature (27 ± 2 oC), controlled by an EG&G 273 A potentiostat and M352 Corrosion software. Distilled/deionized water was used to prepare all solutions. The electrolyte was prepared daily mixing the five solutions used to obtain the final composition of the Hank synthetic sinovial fluid, according to the procedure described elsewhere (Flower, 1990). Prior to polarization experiments the samples were left in open circuit for fifty minutes until stabilization of the rest potential. All I-E curves were recorded at a scan rate of 0.2 mV s-1, from 0.8 V to 2.0 V related to the SCE reference electrode, and tripled.
RESULTS AND DISCUSSION
Characterization of the as Received Ti-6Al-7Nb Bar
Each microstructure, morphology and phase distribution represents a set of properties characteristic of the process temperature and deformation rate. Some orthopedic devices, such as hip joints, present regions with a different morphology due to several degrees of deformation in the body of the prothesis. Grain size, grain boundary area and internal stress may also be responsible for different behavior under polarization. Figure 2 shows a representation of the workpiece and the microstructure of the cross and longitudinal section for samples 01 and 02, respectively. Both surfaces present an a + b structure, in which a is the lighter phase. At the grain boundaries (dark regions), b phase surrounds acicular a . The longitudinal section shows that the grains are deformed in the rolling direction, and the cross section exhibits a larger grain boundary perimeter.
The polarization curves for both samples are shown in Figure 3. Scanning of the potential was started at 0.8 V vs SCE and continued to more positive values at a rate of 0.2 mV s-1. Decreasing values of cathodic current density are observed as potential approximates the combined equilibrium potential. In the anodic portion of the voltammogram, I-E data present typical characteristics of a metal capable of passivation by the growth of a passivating layer on the surface. After an active oxidative process, anodic current density stopped increasing, remaining constant in the range of the selected potentials. Table 3 presents values of open-circuit potential (EOC), the potential in which cathodic current changes to anodic current (ECA) and current density in the passive state (IP). From now on both, EOC and ECA, are referred as corrosion potential in the text due to the reproducibility and correspondence between the two values.
From Table 3 it is possible to note that corrosion potential for sample 02 is more noble than that for sample 01. On the other hand, sample 01 presents a greater area of grain boundary, which means a higher fraction of beta phase. This may be an indication that, in fact, beta phase is less noble than alpha phase and it may exhibit anodic behavior in the metallic a/b couple. As shown in Table 3, sample 02, taken from the longitudinal section, also passivates first at a more negative potential. However, current density in the passive state is practically the same (i.e., about 10-6 A cm-2), although sample 01 presents some noise in the current measurement probably due an artifact of the experiment.
Implant materials should remain in the passive state in a range of potentials extending over about 500 mV, from 200 mV to 300 mV vs. SCE, in isotonic saline solutions (Simpson, 1986). As a common feature both samples presented this behavior. Furthermore, in the range of potentials used in this experiment there is no evidence of the passive layer breakdown or pit initiation. Titanium alloys exhibit high pitting potential. Particularly Ti-6Al-4V and Ti-6Al-7Nb have shown breakdown potentials greater than 6.0 V vs SCE and 10.0 V vs SCE, respectively, in a 2 M HCl solution at 37 oC. In a phosphate buffered saline solution, those values are within the interval of 3.0 to 4.0 V vs SCE (Khan et al, 1996).
Series of Samples Upset at 750oC
Sample 03, which resulted from a light degree of deformation (15% at 1 mm s-1) still exhibited a microstructure similar to that observed in the longitudinal section of the as received material. Dark grains of the beta phase, containing alpha acicular, are still surrounding grains of the light alpha phase. Due to the low percentage of deformation, recrystallization was not observed. However, for samples 04 and 05, with 41% and 62% deformation, respectively, the deformed grains in the direction of previous rolling (longitudinal section) disappeared. Figure 4 presents the microstructure of samples 04 and 05. The first shows grains oriented in the direction of the applied stress. Smaller and well-defined grains are observed, indicating recrystallized grains. Dynamic recrystallization took place in the highly deformed sample 05, but not sufficiently fast to allow grain growth. As a consequence, smaller and refined grains have a greater grain boundary area, and should produce a less noble corrosion potential.
Figure 5 presents polarization curves for samples 03, 04 and 05, and Table 4 depicts a summary of data taken from the curves. As a common feature all curves present characteristics of an active oxidation process at the beginning of the anodic portion, followed by passivation. As shown in Table 4, sample 03, with a low degree of deformation, exhibits a corrosion potential very close to that presented by sample 02. Samples 04 and 05 have corrosion potentials which are more positive than that of sample 03. However, due to the greater area of grain boundary that means a higher fraction of beta phase, sample 05 is less noble than sample 04.
Despite the fact that the surface of sample 04 passivates sooner, in a more negative potential, differences are not too significant. Current density in the passive state is 10-6 A cm-2 for sample 03, and 7.9 x 10-7 A cm-2 for both samples 04 and 05. It is interesting to note that both samples 04 and 05, which were deformed at 750 oC, present corrosion potentials nobler than that of as received sample 02. Furthermore, hot working apparently does not change the current density in the passive state, which is confirmed by comparing the data presented in Tables 3 and 4.
Series of Samples Upset at 850oC
At this temperature the effect of the deformation was more pronounced. Figure 6 presents microstructure after compression for samples 06 (34% at 1 mm s-1), 07 (26% at 0.1 mm s-1), 08 (62% at 0.1 mm s-1), and 09 (54% at 0.01 mm s-1). Sample 06 presents a structure with elongated grains parallels to the direction of deformation while in sample 07, which has a similar degree of deformation, this effect is not so marked. The slower deformation rate for sample 07 permitted partial recrystallization. In samples 08 and 09, in which greater deformation was imposed, microstructures are markedly recrystallized. Light globular grains of alpha phase surrounded by beta phase are present in sample 08. The slow process of deformation used for upsetting sample 09 allowed complete recrystallization and grain growth, in such a way that light alpha grains with poorly defined grain boundaries can be seen and dark islands of beta phase are spread all over the structure. These differences in the final microstructure are reflected in the results of the polarization experiments. Figure 7 shows I-E curves for the four samples upset at 850 oC. Contrary to what was expected, partially recrystallized microstructures (samples 06 and 07) presented more noble corrosion potentials than those that were totally recrystallized (samples 08 and 09), although they passivate at more positive potentials. Furthermore, in these two groups, samples deformed at slower rates presented more positive corrosion potentials as summarized in the Table 5.
While in the passive state, current density was around 10-6 A cm-2 for all samples deformed at 850 oC, despite some noises observed at 1.5 V.
Series of Samples Upset at 950 oC
The effect of the deformation rate on the microstructure can be seen in Figure 8 since the samples presented a similar degree of deformation. In all the samples dynamic recrystallization ocurred at various degrees, resulting in a structure where globular alpha phase grains and transformed beta matrix grains containing a-acicular phase with well-defined grain boundaries. For the slower process of compression at 0.01 mm s-1, Figure 8-c shows larger alpha grains resulting from the longer process time available for solid state diffusion and grain growth. The comparison with microstructures obtained at lower temperatures (see Figures 4 and 6) shows that higher process temperatures facilitate the growth of beta phase grains with precipitated a -acicular.
Figure 9 shows voltammograms recorded for microstructures in Figure 8 in the range of potential from -0.8 V to 2.0 V vs SCE. The I-E curves present the same characteristics as those for samples compressed at lower temperatures. Critical values that characterize the corrosion resistance of samples 10, 11 and 12 are summarized in Table 6, where the corrosion potential shifts more negative for slower deformation rates. While stable, the passive layer presents no significant differences in the current density, as shown in Table 6.
Series of Samples Upset at 1030 oC
Temperature of 1030oC is above the beta transus of a + b to b in this alloy (Flower, 1990) and the deformed samples exhibited similar microstructures as seen in Figure 10. No preferential grain orientation was observed and complete recrystallization and beta grain growth took place, despite the fact that sample 13, upset at a higher deformation rate, presents larger grains, and sample 15, upset at 0.01 mm s-1, shows small grains of recrystallized alpha phase at grain boundaries. A common feature of microstructures is that they are composed of parallel needles of the alpha and beta phases called the Widmanstätten structure. Comparison with samples deformed at lower temperatures shows that samples upset at 1030 oC present higher fractions of beta phase.
Figure 11 presents I vs E data recorded during potential scanning at 0.2 mV s-1, and data taken from these voltammograms are summarized in Table 7. Again, the phenomenon of shifting the corrosion potential to more negative values for the slower process of deformation is observed.
GENERAL COMMENTS AND DISCUSSION
Analysis of data obtained for the conditions of upsetting, shown above, were contrary to what was expected. Higher process temperatures and slower deformation rates usually facilitate stress relief, structure recovery, solid state diffusion, nucleation of new phases, which means recrystallization and grain growth. In these cases metallic materials should present better corrosion resistance and also improved behavior during polarization in aqueous media. However, all we can observe in this paper is the shifting of corrosion potentials to less noble positions (more negative) as the deformation rate decreases. Figure 12 summarizes, as a qualitative comparison, the shifting of corrosion potentials as a function of deformation rate in a plot of potentials versus the logarithm of the deformation rate, where we have excluded data for samples 03 and 07 due to the low degree of deformation obtained. Values of potential in which cathodic current changes to anodic current (ECA) were used in this figure since the values of ECA and the open circuit potential (EOC) never differ by more than 17 mV, except for the sample 06.
Poorer results during polarization may be related to the increasing amounts of beta phase in the microstructure of the Ti-6Al-7Nb alloy at room temperature, after hot processing. Higher temperatures and slow processing facilitates the allotropic transformation of alpha phase to beta phase. However, this transition is, apparently, not totally reversible when the material is air cooled to room temperature and beta phase is retained in the structure. An increasing amount of the beta phase is observed at a process temperature of 950 oC and the increase in the b/a ratio occurs in detriment of corrosion resistance. The first evidence of this behavior was obtained during polarization of the alloy as received (see Table 3 and Figure 3) where the cross section, with a larger exposed area of the beta phase, presented poorer results during polarization. On the other hand, passive layer stability is absolutely unsensible to the deformation rate process temperature and severity of deformation. No sign of passive layer breakdown has been observed during the polarization experiments.
The conditions used for hot working the Ti-6Al-7Nb alloy determine its properties of corrosion resistance. It is important to select temperature process and deformation rate as parameters when more noble potential values are required. Low deformation rate facilitates the formation of beta phase that is retained in the structure at room temperature. According to the evidence found in this study, beta phase presents poor corrosion resistance and increasing amounts of this phase shift the corrosion potentials to more positive values. However, samples hot compressed from 750 oC to 1030 oC showed passive layer stability over a wide range of potentials extending from 0.15 V to 1.75 V vs SCE . Furthermore, passive films grown onto the Ti-6Al-7Nb samples surfaces have shown no sign of rupture for the processing conditions selected for this study.
Semlitsch, M. F., Weber, H., Streicher, R. M. and Schön, R., Joint Replacement Components Made of Hot-Forged and Surface Treated Ti-6Al-7Nb Alloy, Biomaterials, 13, 781 (1992). [ Links ]
Boyer, H. E. and Gall, T. L., Metals Handbook (American Society for Metals: OH, USA) (1985). [ Links ]
Collings, E. W., The Physical Metallurgy of Titanium Alloys (American Society for Metals: OH, USA) (1984). [ Links ]
Aragon, P. J. and Hulbert, S. F., Corosion of Ti-6Al-4V in Simulated Body Fluids and bovine Plasma, J. of Biomedical Materials Research, 6 155, (1972). [ Links ]
Semlitsch, M. F., Weber, H. and Staub, F., Titanium-Aluminium-Niobium Alloy, Development for Biocompatible, High Strength Surgical Implants, Biomed. Technik, 30, 334 (1985). [ Links ]
Noort, V., The Implant Material of Today, J. of Materials Science, 22, 3801 (1987). [ Links ]
Shepard, T. and Norley, J., Deformation Characteristics of Ti-6Al-4V, Materials Science and Technology, 4, 903 (1988). [ Links ]
Streicher, R. M., Schön, R. and Semlitsch, M., Investigation of the Tribological Bahaviour of Metal-on Metal Combination for Artificial Hip Joints, Biomed. Tech., 12, 107 (1990) 125. [ Links ]
Semlitsch, M. F., Titanium Alloys for Hip Joint Replacements, Clinical Materials, 2, 1 (1987). [ Links ]
Khan, M. A., Williams, R. L. and Williams, D. F., In Vitro Corrosion and Wear of Titanium Alloys in the Biological Environment, Biomaterials, 17, (1996) 2117. [ Links ]
Simpson, J. P., The Electrochemical Behaviour of Titanium Alloys With Respect to Their Use as Surgical Implant Materials, in Biological and Biochemical Performance of Biomaterials, ed. by P. Chritel, A. Meunier and A. J. C. Lee( Elsevier, Amsterdan 1986). [ Links ]
Oliveira, V., Chaves, R. R., Bertazzoli, R., Caram, R., Preparation and Characterization of Ti-Al-Nb Alloys for Orthopedic Implants, Brazilian Journal of Chemical Engineering, 15, 326 (1998). [ Links ]
Niinomi, M., Mechanical Properties of Biomedical Titanium Alloys, Materials Science and Engineering- A, 243, 231 (1998). [ Links ]
Akahori, T. and Niinomi, M., Fracture Chacteristics of Fatigued Ti-6Al-4V a an Implant Material, Materials Science and Engineering- A, 243, 237 (1998). [ Links ]
Kuroda, D., Niinomi, M., Morinaga, M., Kato, Y. and Yashiro, T., Design and Mechanical properties of New Beta Type Titanium Alloys for Implant Materials, Materials Science and Engineering- A, 243, 244 (1998). [ Links ]
Flower, H. M., Microstructural Development in Relation to Hot Working of Titanium Alloys, Materials Science and Technology, 6, 1082 (1990). [ Links ]
*To whom correspondence should be addressed