2016 Volume 57 Issue 12 Pages 2026-2032
The susceptibility to environment-assisted cracking (EAC) of super-elastic TiNi alloy was investigated as a function of the electrochemical potential and solution pH. The investigation was conducted using a slow-strain-rate tensile test apparatus with a potentiostat. The test solutions were sulfate solutions with various pH values adjusted by H2SO4 or NaOH. The alloy deforming under cathodic reaction fractured under the relatively small strain where the alloy was in the stress-induced martensitic phase. A larger EAC susceptibility was obtained at lower potential and lower pH, which indicates that this is a general feature of hydrogen embrittlement. The severe EAC region of TiNi alloy was different from that of TiAl alloy. The EAC susceptibility was strongly correlated with the cathodic charge density, irrespective of the pH or potential: a charge density below 0.025 MC m−2 yielded almost no EAC; however, above 0.025 MC m−2 EAC was induced, and the EAC susceptibility was independent of the charge density. Hydrogen in solid-solution state was detected in the alloy at a charge density below 0.025 MC m−2, and hydride started to form at a density above 0.025 MC m−2.
TiNi alloy demonstrates unique mechanical properties such as shape memory or super-elasticity. In addition, the alloy is known to show excellent corrosion resistance in dilute aqueous solutions. Therefore, the alloy has been widely employed in various industrial, dental, and medical applications. It has been reported in the dental field, however, that some parts composed of TiNi alloy fracture. For example, orthodontic arch wires and stents made of the alloy have broken down in oral cavities and in vivo, respectively1,2). It has been proposed that the reason for failure is environment-assisted cracking (EAC) in relation with hydrogen absorption, and many studies have been conducted to understand this mechanism3–6). The electrochemical hydrogen absorption technique has been widely employed in these studies to determine the relation between the EAC susceptibility of the alloy and the amount of hydrogen absorbing into it. However, it is also necessary to understand how the hydrogen absorption process depends on electrochemical and circumstantial conditions, such as potential, current density, solute of solution, and solution pH. Therefore, our research group has tried to understand the effects of the electrochemical potential and the solution pH on the EAC susceptibility of TiNi alloy with super-elastic property7); and found that the EAC susceptibility was significantly enhanced for lower potentials as well as lower pH values. In addition, our group previously investigated the EAC behavior of TiAl intermetallic alloy as a function of the electrochemical potential and the solution pH, and successfully produced an EAC map of the alloy based on potential and pH8,9). In comparison with the results of TiNi and TiAl alloys, it was found that the conditions indicating the maximum EAC susceptibility were different each other, although the two alloys were intermetallic compounds with almost the same atomic concentration ratio of Ti. This research, therefore, aims to understand the role of the electrochemical potential and solution pH to the EAC behavior of super-elastic TiNi alloy.
Material was a Ti–50.85 at% Ni alloy and was heated at 673 K for 3.6 ks to have super elastic property. A 1.1 mm thick alloy plate was machined into two types of specimens: one was a tensile specimen with a gauge size of 5.0 mmL × 3.2 mmW for determining EAC susceptibility and the other was a specimen with a size of 10 mm × 10 mm for obtaining electrochemical polarization curves. The surfaces of both types of specimen were mechanically polished with dry emery paper to #6/0 (corresponding to about #800) immediately before carrying out the tests described below.
The test solutions were aqueous sulfate solutions prepared with distilled water and reagents of H2SO4, NaOH, and Na2SO4 (Wako Chemical Co. Ltd.). The sulfate concentration was fixed at 9.85 × 10−3 kmol m−3 and the pH was controlled from 3 to 11. The solutions were used at room temperature (about 298 K) and deaerated with N2 gas from 1.8 ks before the tests to the end of the tests.
The electrochemical polarization curves of the specimens were measured in the test solutions as described above. In the measurements, an Ag/AgCl (3.3 kmol m−3 KCl, room temperature) reference electrode and a Pt counter electrode were employed. After the corrosion potential of each specimen in the test solution stabilized, a constant potential of −2000 mVAg/AgCl was applied to the specimen; then, the potential was scanned to +1500 mVAg/AgCl at 1.6 mV s−1 by a potentiostat (Type2000, Toho Tech. Res. Co. Ltd.).
The EAC tests were conducted using a slow-strain-rate tensile (SSRT) testing apparatus (1000T, Toshin Kogyo Co. Ltd.) with the potentiostat in the test solutions. In the test, an Ag/AgCl reference electrode and a Pt counter electrode were employed. After the corrosion potential of each tensile specimen in the test solution stabilized, a constant potential between −2000 and 0 mVAg/AgCl was applied to the specimen; then, the SSRT test started at an initial strain rate of 1.7 × 10−5 s−1.
Careful observations of hydride formation at the surface of each fractured specimen were conducted using a scanning electron microscope (SEM: JSM-6060LV, JOEL Ltd.) and an X-ray diffraction analyzer (XRD: RINT-2550, Rigaku Co.), respectively. The X-ray emission conditions for the XRD test with a Cu Kα (wavelength: 0.154056 nm) excitation source were 50 kV and 130 mA.
The amount and state of the hydrogen in each specimen were determined by thermal desorption gas spectroscopy (TDS: TDS-200-L, Ulvac Inc.). The specimen was placed in a quartz chamber under ultra-high vacuum (10−6 to 10−7 Pa) conditions and was heated gradually at 5.6 × 10−2 K s−1 from room temperature to 900 K. The desorption rate of hydrogen out of the specimen during the heating process was determined as a function of the specimen temperature using a quadrupole mass spectrometer.
The polarization curves of the specimens in the solutions with various pH values are shown in Fig. 1. A corrosion potential was located in the range from −300 to −700 mVAg/AgCl and was lower in solutions with higher pH. The shape of the cathodic polarization curve was almost independent of the solution pH except for pH 3. The wavy part of the cathodic polarization curve at around 1 A m−2 in the solution with pH 3 was caused by the diffusion-limiting reduction reaction of protons. The shape of the anodic polarization curve varied slightly with pH. A small peak was observed around +200 mVAg/AgCl in the solutions with pH values of 3 and 5; additionally, the data indicated a slight active dissolution around the corrosion potential. Conversely, the peak disappeared in solutions of pH 7 or higher, that is, the specimen was in a passive state under natural immersion conditions. The passive current density slightly decreased with a raise in pH. When a potential higher than +1300 mVAg/AgCl was applied to the specimens, a drastic increase in current density was measured in all solutions. The increase may have been caused by the oxidation reaction of water. A similarly dramatic increase in the current density at potentials higher than +1000 mVAg/AgCl for pH 9 and +800 mVAg/AgCl for pH 11 was observed and may have been caused by the trans-passive dissolution of Ni.
Polarization curves of the specimens in solutions with various pH values.
Curves of the current density as a function of strain on the specimens during the EAC test in the solution with pH 7 are shown in Fig. 27) as an example. The current densities at three selected potentials were negative and were quite similar to those for the cathodic polarization curves, as shown in Fig. 1. The cathodic current densities indicate the reduction reaction rate of protons or water to hydrogen, a proportion of which may have been absorbed by the specimen. Although the stress on the specimens changed during the EAC test as described below, the cathodic current density was relatively steady.
Changes in current density as a function of strain for the specimens during the EAC test in the pH 7 solution at different applied potentials7).
Nominal stress–nominal strain curves of the specimens tested in the solution with pH 7 are shown in Fig. 37). As a potential of −1000 mVAg/AgCl was applied to the specimen, a large tensile strength and a large fracture strain were observed. In addition, the stress–strain curve almost corresponded to that in air. Stress-induced martensitic transformation occurred at 350 MPa. Meanwhile, an applied potential of −1500 mVAg/AgCl or lower caused the small fracture strain and led to a tensile strength of less than 600 MPa.
Stress–strain curves of the specimens during the EAC tests in the solution with pH 7 at various applied potentials7).
The side surfaces of the specimens that fractured in the solution with pH 7 are shown in Fig. 4. The fracture portion of the specimen tested at −1000 mVAg/AgCl (Fig. 4(a)) was not smooth, showing slight necking. The aspect indicates ductile property. When a lower potential was applied to the specimen, its fracture portion was smoother (Fig. 4(b)(c)). A smooth fracture surface strongly suggests that a crack will propagate more easily even for lower stress values. Fracture surfaces of the specimens tested in the solution with pH 7 are shown in Fig. 5. The fracture surface of the specimen tested at −1000 mVAg/AgCl (Fig. 5(a)) was relatively rough. As a lower potential was applied to the specimen, its fracture surface was smoother and flatter (Fig. 5(b)(c)). Magnified views of the edges of the fracture surfaces showed that the edge part consisted of a dimple pattern caused by ductile fracture under an applied potential of −1000 mVAg/AgCl (Fig. 5(d)). Meanwhile, the lower potential formed a quasi-cleavage pattern over the whole fracture surface and the fracture pattern at the edge part had a finer quasi-cleavage aspect (Fig. 5(e)). In addition, the area of the finer or smoother fracture morphology expanded when the applied potential was lowered (Fig. 5(f)).
Side surfaces of the specimens fractured in the solution with pH 7 at applied potentials of (a) −1000, (b) −1500, and (c) −2000 mVAg/AgCl.
Fracture surfaces of the specimens tested in the solution with pH 7 at applied potentials of (a), (d) −1000, (b), (e) −1500, and (c), (f) −2000 mVAg/AgCl. (d), (e), and (f) are magnified views of (a), (b), and (c), respectively.
The EAC behavior of the specimen in the solution with pH 7 can be summarized as follows: a lower applied potential induces a larger cathodic current density along with the formation of hydrogen, larger susceptibility to fracture of the specimen, and easier crack propagation with quasi-cleavage aspect. These findings strongly suggest that the fracture is caused by hydrogen embrittlement. The EAC tests were extensively conducted in solutions with various pH values from 3 to 11, and the trends of the EAC behavior, which depended on the applied potential and pH, were similar to those in the solution with pH 7, as described above.
To understand the role of the electrochemical potential and solution pH to the EAC susceptibility of TiNi alloy, an EAC susceptibility map as function of those two parameters was produced. The EAC susceptibility level was determined by the level of the tensile strength (σTS). The map was superimposed on potential–pH diagrams (i.e., Pourbaix diagrams) of Ni/H2O10) and Ti/H2O11,12), as shown in Fig. 6. The potential–pH diagrams were calculated for ion concentrations of 10−2 and 10−6 kmol m−3, and the correction of potential for different reference electrodes was conducted as follows,
| \[0\,{\rm V}_{\rm Ag/AgCl} = 0.206\,{\rm V}_{\rm SHE}.\] | (1) |
| \[2{\rm H}^+ + 2{\rm e}^- \to {\rm H}_2.\] | (2) |
The EAC susceptibility of TiAl alloy was previously investigated as a function of solution pH and applied potential7,8). As a result, the EAC susceptibility was found to be the highest in the low-pH region and for a potential of around −1000 mVSHE; these conditions corresponded to the Ti ion stable region in the potential–pH diagram of the Ti/H2O system and to the region in which no species of the Al/H2O system are stable. These findings suggested that the surface property relating to the process of hydrogen absorption was not affected by Al, but was affected by Ti. Our research group also previously produced an EAC map of Ti–6Al–4V alloy13) and a hydrogen absorption map of Ti14); and revealed that these Ti–Al alloys, including Ti, absorbed hydrogen most severely in the Ti ion stable region. Conversely, no stable regions for chemical species except for hydrogen could be identified on the potential–pH diagrams of the Ni/H2O and Ti/H2O systems that corresponded to the region of severe EAC susceptibility. This suggests that hydrogen absorption was not affected by the surface properties of TiNi alloy, but by the enhancement of hydrogen generation. In the electrochemical hydrogen permeation technique15), the hydrogen detection side of the specimen is covered with Ni, because Ni and its passive films exhibit almost no barrier to hydrogen permeation16). Therefore, it is proposed that the Ni in TiNi alloy promotes hydrogen absorption much more strongly than Ti, which acts as a barrier under certain conditions. In addition, lower potentials and lower pH values play an important role in increasing the overvoltage and proton concentration, respectively, for hydrogen generation, which enhances hydrogen absorption.
4.2 Cathodic charge affecting the EAC susceptibilityBased on the results of the EAC test, it is proposed that the EAC is caused by hydrogen embrittlement. It is known that susceptibility to hydrogen embrittlement positively depends on the amount of absorbed hydrogen. Since the cathodic current density is considered to be the reduction reaction rate of protons or water to hydrogen, it was expected that the amount of hydrogen being absorbed by the specimen would positively relate to the cathodic charge density, which was calculated by integrating the cathodic current density with time. The results of the tensile strength and the cathodic charge density are summarized in Fig. 7. The figure indicates that the data lined up along a curve, irrespective of the electrochemical potential or solution pH. In the case of a charge density below 0.025 MC m−2, the tensile strength was about 1600 MPa (which is quite similar to that in air) and almost independent of the charge density. The tensile strength quickly decreased to 470 MPa at a charge density of 0.025 MC m−2, and at a density between 0.025 and 100 MC m−2 the tensile strength was constant at around 470 MPa. 470 MPa was about the stress at which the martensitic formation was completed, as shown in Fig. 3. From these results, the features of the EAC of TiNi alloy from the viewpoint of hydrogen absorption can be summarized as follows: (1) the parent phase (< 470 MPa) suffered no EAC even when the phase involved an amount of hydrogen corresponding to a cathodic charge density of 100 MC m−2; (2) the stress-induced martensitic phase (≥ 470 MPa) suffered almost no EAC and fractured at about 1600 MPa when the phase involved an amount of hydrogen corresponding to a cathodic charge density below 0.025 MC m−2; and (3) the stress-induced martensitic phase suffered EAC and fractured at a stress of 470 MPa, at which the phase transformation was completed. The tensile strength was independent of the amount of hydrogen (corresponding to a cathodic charge density of more than 0.025 MC m−2).
Effect of cathodic charge density on tensile stress.
It is important to understand which state of hydrogen, i.e., solid-solution state of hydrogen or hydride, affects the EAC of the alloy. As described before, the specimen receiving small cathodic charge densities fractured in ductile mode under high tensile stresses. However, the specimen to which large cathodic charge densities were applied fractured in brittle mode under quite low stresses of 470 MPa. In addition, the edge part of the fracture surface was finer and smoother as the EAC enhanced. These results suggest the formation of a brittle hydride layer to accelerate EAC susceptibility. Therefore, the specimens fractured in the solution with pH 3 at various applied potentials were probed using XRD.
Figure 8 shows the XRD profiles of the side surfaces for the specimens fractured in the solution with pH 3. The result for the specimen fractured in air is superimposed on the figure. It was found that hydride17,18) was detected as a shoulder of the peak of TiNi below −1000 mVAg/AgCl (i.e., more than 0.19 MC m−2) and that its peak height was quite small, even at −2000 mVAg/AgCl (2.3 MC m−2). These facts indicate that a very thin hydride layer is formed on the TiNi surface.
XRD profiles of the side surfaces for the specimens fractured in the solution with pH 3 as a function of applied potential.
Another test was conducted to analyze the state of hydrogen in the alloy as follows: the specimen was immersed in the solution with pH 3 and then a constant potential of −1000 or −1500 mVAg/AgCl was applied to the specimen under unstressed conditions for periods that were the same as the fracture times in the EAC tests. Thereafter, the specimen was subjected to the TDS test. Changes in hydrogen desorption rate with specimen temperature (i.e., TDS profiles) as a function of applied potential are shown in Fig. 9. The charge density at each potential was slightly larger than that in the EAC test. It was found that there were two peaks irrespective of the applied potential. Yokoyama et al.6) investigated the state of hydrogen in a super-elastic TiNi alloy that was subjected to an electrochemical hydrogen absorption test at a constant current density of 10 A m−2 for 21.6 ks in 0.9 mass% NaCl aqueous solution. The state of hydrogen was confirmed via both XRD and a TDS test at a heating rate of 2.8 × 10−2 K s−1 (which is half the rate used in the present study). Those experiments revealed the following: hydride was clearly detected in the XRD pattern and one broad peak was observed in the TDS profile of the alloy immediately after the hydrogen absorption test. In addition, when the alloy with the hydride was stored in air for 864 ks, the hydride peak disappeared in XRD pattern and a portion of the signal in the lower-temperature region disappeared in the TDS profile. The findings by Yokoyama et al. suggest that there are two states of hydrogen in the alloy immediately after the hydrogen absorption test: one is hydride, which was confirmed by the appearance of a peak in the low-temperature region of the TDS profile, and the other is solid-solution-state hydrogen, which was confirmed by the appearance of a peak in the high-temperature region of the TDS profile. The TDS profiles in Fig. 9 are different from those found by Yokoyama et al., that is, the two peaks were clearly distinguishable from each other and were shifted in the positive temperature direction. The reason for this may be the different heating rates. However, it is proposed that the peaks at low and high temperatures correspond to hydrogen desorption from hydride and solid-solution-state hydrogen, respectively.
Changes in hydrogen desorption rate with specimen temperature as a function of applied potential.
To confirm the decomposition of hydride at low temperatures in the TDS profile, the following test was conducted. A potential of −2000 mVAg/AgCl was applied to the unstressed specimen in the solution with pH 3 until a cathodic charge density of 35 MC m−2 was reached; and then the specimen was probed via XRD. After that, the specimen was heated up to a temperature of 473 K in the TDS system. The temperature region provided the lower−temperature peak in the TDS profile of Fig. 9. Finally, the specimen was measured via XRD again. The XRD patterns before and after TDS are shown in Fig. 10. The XRD pattern of the specimen before the TDS test exhibited remarkable hydride peaks owing to the large cathodic charge density of 35 MC m−2. After the TDS test, the peaks of the hydride disappeared from the XRD pattern owing to decomposition of the hydride when the temperature was elevated to 473 K. These results suggest that the peak at low temperatures in Fig. 9 can be assigned to hydride and that the other peak at high temperatures is associated with solid-solution-state hydrogen.
XRD patterns of the specimen before and after the TDS test heating to 473 K. The specimen was subjected to a hydrogen absorption test immediately before the first XRD test.
Figure 11 shows the TDS profiles of the unstressed TiNi alloy immediately after a hydrogen absorption test at −1500 mVAg/AgCl in the solution with pH 5 as a function of cathodic charge density. It was found in the figure that a small charge density of 0.005 MC m−2 was only able to induce the high−temperature peak. This indicates that hydrogen initially accumulates in the alloy as solid-solution-state hydrogen. The hydrogen in the solid-solution state increases with an increase in cathodic charge density. The low−temperature peak was only observed for charge densities above 0.025 MC m−2, which also induced severe EAC in the alloy.
TDS profiles of the specimen immediately after a hydrogen absorption test without straining at −1500 mVAg/AgCl in the solution with pH 5 as a function of cathodic charge density.
As already described, a large EAC susceptibility was confirmed for the specimens tested at low potentials (Fig. 6). Under these conditions, the specimen had a brittle nature, likely because of hydride (Fig. 8), as well as containing hydrogen in the solid-solution state (Figs. 9 and 11). Since the stress-induced martensitic phase without hydrogen proved to be ductile even for high stress (Fig. 2), it is proposed that this phase hardly suffer brittle fracture that is enhanced by the formation of notches. This would also mean that the effect of the brittle fracture of the phase on the formation of hydride is reduced, even if notches originate from the hydride in this phase during the EAC test. However, the crack rapidly propagated and the whole fracture surface consisted of the quasi-cleavage pattern when the specimen was subjected to a cathodic charge density of more than 0.025 MC m−2 (Figs. 4 and 5). Therefore, it is proposed that the brittle fracture is mainly resulted from hydrogen in the solid-solution state, which is homogeneously distributed in the alloy. The hydrogen may enhance the notch sensitivity against the brittle fracture of the alloy when the parent phase fully transforms to the martensitic phase. In that case, notches may be formed in the hydride layer, at slip steps, and in other similar locations.
This study aimed to understand the role of the electrochemical potential and solution pH to the EAC behavior of super-elastic TiNi alloy. The findings of this research are summarized as follows:
This research was financially supported in part by the Kansai University Research Grants: Grant-in-Aid for Encouragement of Scientists, 2006.
