Chapter 4 Results and Discussion
4.5 Simulated body fluid electrochemical response of Zr-based and TiZr-based
4.5.1 Structural characterization and thermal analysis
The thermal properties of specimens have to be determined for confirming the amorphous structures and ideal annealing temperature. The DSC curves of Zr53Cu30Ni9Al8
and Ti42Zr40Si15Ta3 amorphous alloys reveal obvious glass transition (Tg) and crystallization (Tx) temperatures, as shown in Figure 4.27, which means that the structure of as-cast specimens are amorphous, and the Tg and Tx are 420 and 495 oC (or 693 and 768 K) for Zr53Cu30Ni9Al8 and 526 and 606 oC (or 799 and 879 K) for Ti42Zr40Si15Ta3, respectively. The proper annealing temperatures should be between Tg and Tx so as to avoid the formation of too pronounced phases which will cause complex electrochemical analysis. Fig. 4.28(a) and
Fig. 4.29(a) show the DSC scanning curve of the as-cast and annealed specimens to different time periods, the crystallinities of the annealed specimens can be calculated from the enthalpy release (ΔH) of the annealed specimens by DSC analysis program. The ΔH of the as-cast Zr53Cu30Ni9Al8 and annealed for 21, 23, and 40 min at 713 K (20 oC higher than Tg) are
−59.62, −39.25, −22.29, and 0 J/g. In addition, the ΔH of the as-cast and annealed Ti42Zr40Si15Ta3 for 10 and 20 min at 818 K (19 oC higher than Tg) are 48.27, -33.16, and -19.07 J/g. The evaluations of crystallinities of specimens are carried out by following equation:
as cast annealed
as cast
H H
Crystallinity
H
. (4.9)
The volume fractions of the crystallinities for Zr53Cu30Ni9Al8 are determined to be 0, 34, 63, and near 100%, respectively. In parallel, the volume fractions of the crystallinities for Ti42Zr40Si15Ta3 are determined to be 0, 31, and 60%, respectively. From the above results, the nanocrystallized metallic glasses were successfully fabricated and the crystallinite volume fractions were precisely determined.
Next, it needs to confirm which phases were precipitated out. Figure 4.28(b) and Figure 4.29(b) show the XRD patterns of the as-cast and annealed Zr53Cu30Ni9Al8 and Ti42Zr40Si15Ta3. The crystallization peaks reveal the formation of the predominant Zr2Cu and minor Zr2Ni phases in the Zr53Cu30Ni9Al8 glassy matrix, and the beta-Ti phase in the Ti42Zr40Si15Ta3 glassy matrix. The crystal structure of Zr2Cu and Zr2Ni are tetragonal with lattice constants of a=0.457 nm and c=0.359 nm and a=0.651 nm and 0.493 nm, respectively.
nm. For longer annealing time, Zr2Cu and β-Ti appears as the major phases for Zr53Cu30Ni9Al8 and Ti42Zr40Si15Ta3, respectively, due to the higher atomic percentage of Zr/Cu and Ti in these two MGs. Based on our previous studies on the Zr-Cu-Ni-Al MGs [158,174-177], the Zr2Cu nanocrystal size and volume fraction would increase with increasing heat treatment time. The Zr2Cu particles start from 2 nm and gradually grow to over 20 nm. In parallel, the TEM dark field and bright filed images, as well as diffraction patterns, taken from Ti42Zr40Si15Ta3 after 10 min heat treatment show that the nanocrystalllized beta-Ti phase forms in the amorphous matrix homogeneously, and the crystal size is about 10 to 20 nm. In addition, the HRTEM image (Figure 4.30) of Ti42Zr40Si15Ta3 after 10 min heat treatment also shows the clear boundaries between β-Ti nanocrystals and amorphous matrix, which is a high energy site for the electrochemical reaction.
4.5.2 Electrochemical response
The representative bio-corrosion potential versus immersion time for the as-cast and annealed specimens are first studied by the open circuit potential curve (also called the OCP or E–t curve), as shown in Figures 4.31 and 4.32. The open circuit potential of as-cast and annealed specimen is stabilized over immersion time, indicating the formation of an oxidative passive layer for all alloys after a 5,400 s for Zr53Cu30Ni9Al8 and 12,000 s for Ti42Zr40Si15Ta3
immersion in the Hank's solution. Because the open circuit potential is equal to corrosion potential approximately, the as-cast Zr53Cu30Ni9Al8 in Figure 4.31 possesses lower corrosion activity (at −0.183 V) than the annealed samples for 21 min (at −0.360 V), 23 min (at −0.380 V), and 40 min (at −0.436 V), due to the higher stabilized open circuit potential, as compared in Table 4.8. This result suggests that the as-cast fully amorphous Zr53Cu30Ni9Al8 MG would
exhibit higher bio-corrosion resistance than the partially crystallized ones.
In comparison, the as-cast Ti42Zr40Si15Ta3 own higher corrosion activity (at about
−0.469 V) than the samples annealed for 10 min (at about −0.381 V) and 20 min (at about
−0.345 V), as listed in Table 4.9. The trend in Ti42Zr40Si15Ta3 is in fact opposite to what is seen for Zr53Cu30Ni9Al8.
In comparison of the above OCP curve results, the bio-corrosion activity increases with increasing annealing time period for Zr53Cu30Ni9Al8, but decreases with decreasing annealing time period for Ti42Zr40Si15Ta3. This implies the negative (or more reactive) nature of the e Zr2Cu and Zr2Ni crystallized phases, and the positive (or more inert) effect of the crystallized β-Ti nanocrystals in terms of the corrosion resistance of metallic glasses.
More important corrosion parameters of as-cast and annealed specimens were also determined by the potentiodynamic polarization measurement, as shown in Figure 4.33 and Figure 4.34. The corrosion potential (Ecorr) is used to evaluate driving force for bio-corrosion, and the system with a higher Ecorr means it needs more energy to start a corrosion reaction.
For the Zr53Cu30Ni9Al8, the average Ecorr of the as-cast, partially (annealed for 21 min and 23 min), and fully (annealed for 40 min) crystalline specimens are about −0.214, −0.333, −0391, and −0.423 V, respectively (i.e., gradually decreasing in Table 4.8). In comparison, for the Ti42Zr40Si15Ta3, the average Ecorr of the as-cast, partially (annealed for 10 min), and fully (annealed for 20 min) crystalline specimens are about −0.455, −0.391, −0321, respectively (i.e., gradually increasing in Table 4.9). Hence, the annealed partially and fully crystallized Zr53Cu30Ni9Al8 possess greater opportunities to start the corrosion reaction than the
fully crystallized Ti42Zr40Si15Ta3 own higher Ecorr reading for initiating a polarization process in SBF. The above results show that the presence of reactive Zr2Cu and Zr2Ni phases in amorphous matrix will promote corrosion reaction, but the presence of high corrosion-resistant β-Ti will retard the corrosion activity.
The corrosion current density (Icorr) is another important parameter for determining the activity of corrosion reaction.The average Icorr is determined to be about 0.06 × 10−5 A for the as-cast Zr53Cu30Ni9Al8 and about 0.198 × 10−5, 0.433 × 10−5 and 1.134 × 10−5 A for the specimens annealed for 21 to 40 min. The overall difference or increment from the fully amorphous to fully crystallized is about 19 times, a significant increment effect. The much higher Icorr values of the annealed specimens again stand for the greater degree of electrochemical activity as compared with the as-cast specimen. The induced corrosion current of the fully crystallized sample is almost 20 times that of the fully amorphous one, directly indicating that the bio-corrosion reaction is significantly more severe in the crystallized samples. Both the above Ecorr and Icorr data consistently indicate that the fully amorphous metallic glass would exhibit the strongest bio-electrochemical resistance than the crystallized one with reactive phases such as Zr2Cu and Zr2Ni. With an increasing degree of reactive phases in glassy matrix, the sample corrosion resistance becomes progressively degraded.
The average Icorr of the as-cast Ti42Zr40Si15Ta3 and annealed for 10 min and 20 min are about 0.049 × 10−6, 0.067 × 10−6 and 0.087 × 10−6 A, respectively. The overall difference or increment from the fully amorphous to fully crystallized is only 1.8 times, a very minor effect.
This result suggests that the Icorr of partially crystallized Ti42Zr40Si15Ta3 MG is very slightly higher than the amorphous one, presumably due to the formation of the interface boundaries
between nanocrsytalliine beta-Ti phases and amorphous matrix. Generally, interfacial boundaries for nanocrystalline phases in amorphous matrix in Ti42Zr40Si15Ta3 are not as sharp as those grain boundaries in crystalline cp Ti. Therefore, the difference of Icorr between the amorphous and crystallized Ti42Zr40Si15Ta3 is very minor.
In the SBF, the higher ionic strength of chloride makes serious pitting reaction on amorphous materials. The pitting potential (Epit) is a more critical index in determining whether the chloride-induced pitting would occur from the sudden signal rise of polarization curve, as defined in Figure 4.33 and Figure 4.34. For the as-cast Zr53Cu30Ni9Al8, the pitting reaction is seen to start from a high potential at +0.035 V, as shown in Figure 4.33 and listed in Table 4.8. The Epit of partially and fully crystalline specimens all lie in a much lower potential range from −0.037 to −0.060 V. Epit seems to be slightly lower with an increasing crystalline volume fraction, i.e., pitting would occur slightly earlier in the crystallized samples. Nevertheless, the difference of Epit is not significant (only 0.095 V from +0.035 to
−0.060 V), suggesting that both the amorphous and crystalline materials cannot avoid pitting reaction in SBF when the applied potential reaches this level.
For the Ti42Zr40Si15Ta3 case, the Epit readings of the as-cast and annealed for 10 and 20 min Ti42Zr40Si15Ta3 are 1.130, 1.345, and 1.765 V, respectively, as shown in Figure 4.34 and listed in Table 4.9. Interestingly, the pitting resistance of Ti42Zr40Si15Ta3 is improved by forming the nanocrystalline beta-Ti in the amorphous matrix. This result can be explained by the oxide-bridge model [178], in which the authors assumed that the build-up of a pitting-resistant alloying element in the remaining amorphous phase would make the concentration of oxide increase for improving the resistance of pitting initiation [178, 179]. In
and structurally heterogeneous beta-Ti nanocrystals from the pitting initiation and stabilization process [179]. The SEM image and EDS analysis reveal (Figure 4.35) the pitting morphology of the as-cast Zr53Cu30Ni9Al8. We can find that the Zr and Ni were selectively dissolved from the surface. That is because the Cu is electrochemically nobler than Zr.
Therefore, the galvanic corrosion will occur. From the polarization curves and pitting morphology, we can conclude that formation of reactive Zr2Cu nanocrystalline phases in the amorphous matrix would reduce the bio corrosion resistance more seriously. This galvanic corrosion will not happen in the partially crystallized Ti-based MG with pure -Ti. Thus, the crystallization in the Zr-based or Ti-based MG results in a reversed consequence. The former becomes degraded and the latter becomes upgraded in bio-corrosion resistance.
The passive region (ΔE = Epit − Ecorr) is a region for the formation of the oxidative film during anodic polarization. Although the crystalline materials appear to possess a slightly wider region (about 0.1 V in Table 4.8) for the formation of the passive film, the Epit levels of crystallized Zr53Cu30Ni9Al8 were all low. Thus ΔE does not act as a solid indication on protection capability of the oxidation film for Zr53Cu30Ni9Al8. The passivation current density, Ipass, has to be considered. The fully amorphous Zr53Cu30Ni9Al8 possesses a much lower passive current density (about 10−7 A in Table 4.8) than the crystalline ones (about 10−4 A in Table 4.8, or a factor of 1000 times increment), meaning that a more protective and dense passive film has formed on the surface of the fully amorphous specimen in SBF [37]. The partial and fully crystalline specimens with reactive Zr2Cu and Zr2Ni are more sensitive and prone to the chloride induce pitting reaction. Based on the above results, both the Epit and Ipass
data again consistently indicate that the fully amorphous Zr53Cu30Ni9Al8 is inherent with the strongest bio-electrochemical resistance.
On the other hand, the ΔE values of the as-cast and the annealed for 10 and 20 min Ti42Zr40Si15Ta3 are 1.586, 1.736, and 2.086, respectively (in Table 4.9). The crystallized Ti42Zr40Si15Ta3 owns a wider region for forming protective passive film, due to the superior pitting resistance. The Ipass readings of Ti42Zr40Si15Ta3 exhibit a nearly fixed trend, which are 2.913 x 10-6, 3.683 x 10-6, and 3.983 x 10-6 for the as-cast and the annealed specimens for 10 and 20 min, meaning the forming ability of protective passive film for nanocrystallized Ti42Zr40Si15Ta3 is similar to amorphous one. Therefore, the formation of beta-Ti in amorphous matrix will not change the homogenous structure of the passive film.
Finally, a common electrochemical impedance method is adopted to ensure the above findings. It utilizes the so-called Nyquist plot [57] to illustrate the electrochemical impedance spectra (EIS) for evaluating the polarization resistance (Rp). The equivalent circuit model for fitting the curves of the Nyquist plot is shown in Figure 4.36. The impedance of constant phase element (CPE) is defined as ZCPE = 1 / Q(jω)n, where ω is angular frequency, Q is pre-factor of CPE, and n is its exponent with the range 0 ≤ n ≤ 1. In an AC circuit, the impedance Z is more appropriate to represent the system's entire resistance. From Euler's relations, we know that Z = Z0 (cosθ +jsinθ ). Hence, the impedance can be described by the real and imaginary part (Z = Z’ + jZ’’), and the Nyquist plot is often represented by Z’ versus Z’’. The relation between Rp and Z is given by [180]:
2
2 2
(1 cos( )) - sin( )
2 2
1 2 cos( ) ( ) 1 2 cos( ) ( )
2 2
n n
p p p
s
n n n n
p p p p
n n
R R Q R Q
Z R j
n n
R Q R Q R Q R Q
, (4.10)
The fitted data are presented as the Nyquist plot in Figure 4.37 and Figure 4.38 for the as-cast and nancrystallized Zr53Cu30Ni9Al8 and Ti42Zr40Si15Ta3. In the Nyquist plot, the larger diameter of the semi-circle represents a higher corrosion resistance [57]. For the Zr53Cu30Ni9Al8, it is clear that the as-cast specimen owns the highest corrosion resistance.
The Rp reading is 6.82 × 105 Ω for the as-cast specimen, 0.21 × 105 and 0.06 × 105 for the partially crystalline 21 and 23 min ones, and 0.04 × 105 Ω for the fully crystalline 40 min sample. The Rp reading of the fully amorphous MG is 170 times higher than that of the fully crystallized sample, exhibiting significant corrosion resistance in SBF for the as-cast fully amorphous Zr53Cu30Ni9Al8 sample. In comparison, the Rp readings of as-cast and the annealed for 10 and 20 min Ti42Zr40Si15Ta3 are 11.69 × 105 Ω, 9.80 × 105 Ω, and 3.33 × 105 Ω, respectively. The values only show slight decrement (only about 3 times, not as the above 170 times for the Zr-based MG). Although there are new interfaces induced in the partially crystallized Ti42Zr40Si15Ta3, the negative influence from the interfaces would be offset by the nobler -Ti precipitates.
The results of Rp readings show the similar results with the Icorr, the reactive nanocrystalline Zr2Cu and Zr2Ni phases reduce the corrosion resistance of Zr53Cu30Ni9Al8. In comparison, the corrosion-resistant beta-Ti phases do not affect significantly the polarization resistance for Ti42Zr40Si15Ta3.
As for the underlying reasons for the lower bio-corrosion resistance in the crystallized specimens, it is believed to be caused by the structure and composition of the precipitated phases. Because Cu and Ni is both nobler than Zr [37,181], the dominant Zr2Cu and minor Zr2Ni nanocrystalline phases will induce serious galvanic corrosion. With increasing
annealing time, the size and quantity of the reactive Zr2Cu and minor Zr2Ni nanocrystals both increase. Preferential electrochemical reaction would undergo more and more intensely around the precipitated phases. With increasing annealing time and thus an increasing crystallite volume fraction, the bio-corrosion on Zr53Cu30Ni9Al8 becomes more severe. On the other hand, the β-Ti would not induce no occurrence of the galvanic corrosion. Furthermore, pure Ti is also a pitting-resistant element, the formation -Ti in the amorphous matrix will improve the pitting resistance of Ti42Zr40Si15Ta3 under chloride-rich corrosive media. In general, with the crystallite volume fraction increases, the ability of long-term immersion of Ti42Zr40Si15Ta3 under SBF is improved. Figure 4.39 shows schematic diagram demonstrating that the non-protective and protective passive layer will form on the surface of nanocrystallized MGs with reactive Zr2Cu and Zr2Ni in Zr53Cu30Ni9Al8 and pitting resistant β-Ti in Ti42Zr40Si15Ta3.
Chapter 5 Summary and Conclusions
Based on the present research, the potential metallic glasses for biomedical uses are rapidly screened by cyclic voltammetry, potential state measurement, in vivo and in vitro biocompatibility test. The TiZr-based metallic glasses with or without low concentrations of biologically unfavorable Cu were fabricated successfully. Then their superior corrosion resistance and biocompatibility were also determined. Moreover, the Cu effect on metallic glasses was also studied for clear profile of corrosion mechanism under the SBF. Finally, the corrosion behaviors of nanocrystalline Zr-based and TiZr-based metallic glasses were investigated. Conclusions for each studies reached through the experimental results are summarized below.
(1) This research reports the biocompatibility of various metallic glasses under in-vitro tests.
Results show that the Zr61Cu17.5Ni10Al7.5Si4 metallic glass exhibits good corrosion resistance compared to other Fe-based and Mg-based metallic glasses. Although there is a small electrochemical response in the simulation body-fluid of Hank's solution, Zr-based metallic glass still shows its potential for biomedical application. The rapid screening method developed in the present research provides a simple and rapid way for testing the biocompatibility of metallic glasses.
(2) The rapid screening method was used again to screen the potential materials for biomedical applications out of seven Ti-based, Zr-based, and Ta-based metallic glasses.
Electrochemical responses of the produced MGs were highly correlated with the possibility for potential corrosion in bio-environments. The electrochemical activity of the MGs was first evaluated with simulation body fluid of Hank's solution and human
serum. Results indicated that the copper content in the MG played a role on the electrochemical activity of the material. MGs with the copper content higher than 17.5%
showed significant electrochemical activity in all electrochemical tests. The MGs of the bulk Ti65Si15Ta10Zr10 and thin-film deposited showed that minor electrochemical response Ta57Zr23Cu12Ti8 exhibited excellent electrochemical stability in comparison with the reference material of pure titanium. All the bulk MGs did not show acute cytotoxicity in the MTT tests utilizing murine bone marrow stem cells, D1, in 72 h of incubation.
Nevertheless, the corrosion released ions from the MGs with significant electrochemical activity exhibited significant cytotoxicity in the MTT tests. The Ti65Si15Ta10Zr10 MG has shown its potential for biomedical applications due to its very low electrochemical response and very low cytotoxicity.
(3) This research also systematically investigated the electrochemical activities, cytotoxicity and in vivo response of the three MGs including two EC stable MGs and one Cu-rich MG.
The electrochemistry property and biocompatibility of Ti42Zr40Si15Ta3 and Ti40Zr40Si15Cu5 were compared with the pure Ti. Results showed that the two newly developed MGs of Ti42Zr40Si15Ta3 and Ti40Zr40Si15Cu5 exhibited low electrochemical responses in both the cyclic voltammetry and potential state tests. The potentiodynamic polarization measurements indicated that these two MGs would form passivation layers to prevent them from further corrosion. On contrast, the Cu-rich MG of Ti45Cu35Zr20 showed significant electrochemical responses in the SBF and resulted in toxic effects on the MTT assay. The ICP-MS results confirmed that the Cu-rich MG released a significant amount of ions after the potential state test. Micro-CT observations showed that all the three MGs exhibited no significant inflammatory effect after one month of implantation. The low
bone ingrowth/ongrowth property of the implanted MGs. The developed TiZr-based MGs, Ti42Zr40Si15Ta3 and Ti40Zr40Si15Cu5, are demonstrated promising potentials for biomedical implant applications.
(4) The systematic characterization of the bio-corrosion response of the Cu-free Ti45Zr40Si15
and Cu-containing Ti40Zr40Si15-Cu5 and Ti45Zr25-Cu30 metallic glasses in the Hank's solution is conducted. Based on the open circuit potential (OCP), potentiodynamic polarization Tafel, electrochemical impedance measurements, as well as cytotoxicity MTT testing, the following conclusions can be drawn. Firstly, the positive effect of the addition of Cu in the Ti-based metallic glasses comes from the nobler nature of Cu, which would shift the oxide corrosion to a higher voltage level, for example, from Ecorr ~ -0.4 V for the Cu-free Ti45Zr40Si15 to Ecorr ~ -0.2 V for the Cu-containing Ti45Zr25-Cu30. Secondly, the negative effect is the induction of severe Cu-induced galvanic corrosion under higher applied voltage levels, resulting in continuous pitting and the depletion of Ti, Zr or Si in the alloy. Thirdly, the bio-corrosion mechanism is modeled in Figure 6. The pitting reaction between Cu and Cl ions in the simulated body fluid would induce CuCl, Cu2O and CuO. And resulting Cl ions will in-turn continuously react with Ti, Zr and Si, enlarging the pitting areas. The role of Cu in the Ti-based metallic glasses is systematically established. Fourthly, it is clear that the presence of nobler Cu will impose the above two opposite effects. Since the minor positive shift of Ecorr for forming surface oxide layers is in fact not of a major issue, the negative effect on local pitting and ion release would cause a major drawback. It is suggested that the Cu content should be kept as low as one can, less than about 10 at% to ensure the absence of pitting during long-termed exposure in human body.
(5) Metallic glasses have recently been considered to be applied for bio-applications. It is always inevitable that some metallic glasses would possess a certain degree of nano-scaled crystalline phases. It is critical in application and of interest in scientific understanding whether the fully amorphous or the partially crystallized alloys would exhibit better bio-corrosion resistance in body fluid. In this study, through controlled thermal annealing, the Zr-based metallic glasses with the crystalline phase volume fractions (predominantly Zr2Cu plus minor Zr2Ni) of 0, 34, 63, and near 100% and Ti-based metallic glasses with the crystalline phase volume fractions (predominantly beta-Ti plus minor alpha-Ti) of 0, 31 %, and near 71 % are prepared.
(6) Based on the bio-corrosion voltage and current, as well as the polarization resistance, it is concluded that the fully amorphous Zr53Cu30Ni9Al8 exhibits the highest bio-electrochemical resistance. With increasing annealing time and thus an increasing degree of partial crystallization, the corrosion resistance becomes progressively degraded.
Preferential electrochemical reaction would undergo more and more intensely around the precipitated phases. Besides, the fully amorphous metallic glasses can form a more protective and denser passive film on the metallic glass surface. Generally, the formation of reactive nanocrystalline Zr2Cu and Zr2Ni phases, which themselves would induce serious galvanic corrosion, in the amorphous matrix would reduce the bio-corrosion resistance.
(7) For the Ti42Zr40Si15Ta3, the corrosion resistance is improved with increasing degree of crystallization. The corrosion activity of nanocrystallized Ti42Zr40Si15Ta3 is slightly larger than amorphous one, due to the structural defect between the nano-sized alpha-Ti and
phases in amorphous matrix will improve the anti-pitting ability of Ti42Zr40Si15Ta3. Generally, the formation of pitting-resistant nanocrystalline Ti phases, which is lack of galvanic corrosion and less prone to pitting reaction, in the amorphous matrix would enhance the bio-corrosion resistance.