Processing of Zn-3Mg alloy by equal channel angular pressing for biodegradable metal implants

2.1 Materials preparation and ECAP process

A Zn-3Mg alloy (3 wt% Mg) was prepared from pure Zn (99.99% purity) and pure Mg (99.95% purity) ingots (Goodfellow Inc., UK) using a conventional casting method. The ingots were melted at 550 °C using an induction furnace protected by a flow of argon. The molten metal was mechanically stirred for 2 min before cast into a preheated (∼150 °C) rectangular shaped cast iron mould and then cooled to room temperature. Billets having dimension of 8 × 8 × 120 mm were cut out from the cast and subsequently homogenized under vacuum at 370 °C for 15 h followed by water quenching. The billets were then subjected to ECAP at one pass (1-ECAP) and two passes (2-ECAP) using a hardened steel dies having 8 × 8 mm of square channel section and intersection angle of 120°. A progressive pressing force was applied at a constant speed of 1 mm/s using an Instron-5569 universal testing machine. Molybdenum disulfide was used as the solid lubricant to reduce friction between the die and the billets. The ECAP dies was heated at 200 °C to allow a smooth passage of the billets through the channel without cracking. Cracks were observed on billets processed at 100–190 °C while increasing the temperature above 210 °C led to partial melting of the billets.

2.2 Characterization and mechanical testing

All samples for microstructure observation were longitudinally cut (parallel to the ECAP direction). Their microstructure was observed under an optical microscope (Nikon Microphot-FXL, Japan) after being ground, polished with diamond suspension (from 3 µm to 1 µm) and colloidal silica suspension (0.04 µm), and finally etched with acid solution containing 50 g CrO₃, 15 g Na₂SO₄ in 1 L distilled water at 80 °C. The optical micrographs were used to measure the average grain size and percentage of porosity by using image analysis software (Microvisual Advance, IMT iSolution, Canada) with three replications. The average grain size was measured using ferret diameter which is the average caliper lengths of the grains. Microstructural observation and elemental analysis at higher magnification were done by using a Field Emission Scanning Electron Microscope (FESEM, Zeiss Supra VP35, Germany) equipped with Energy Dispersive X-ray Spectroscopy (EDS, Oxford Instrument, UK). Phase identification was conducted using an X-ray diffraction machine (XRD, Siemens-D500, Germany) with Cu Kα line generated at 40 kV and 35 mA. Further analysis of the XRD pattern was performed to determine the crystallite size of detected phases on both homogenized and ECAP samples using Debye–Scherrer equation: $T=\frac{C\lambda }{\mathit{Bcos}\theta }$ , where T is the crystallite size (nm), λ is the wavelength of the X-rays (i.e. 1.541 Ǻ for Cu Kα radiation), C is the shape factor equal to 0.9, θ is the Braggs angle, and B is the full-width at half-maximum (FWHM) of the selected peak (radians). Microhardness measurement was carried out using a Vickers hardness tester (DVK-2 Matsuzawa, Japan) by applying 5 kg load at room temperature for 15 s at five different locations on each polished specimen with three replications. Tensile test was conducted by using a universal testing machine (Instron-5569, Japan) at a constant strain rate of 0.5 mm/min on flat specimens having 9 mm gage length, 40 mm length and 3 mm thickness with at least three replications. The test and the determination of tensile strength, yield strength, maximum elongation and Young’s modulus were done in accordance with the ASTM E8M standard (Standard Test Methods for Tension Testing of Metallic Materials).

2.3 Corrosion testing

Corrosion behavior of the metal samples was assessed using both simple immersion and electrochemical polarization methods. Test solution was prepared based on Kokubo’s simulated body fluid (SBF) containing the following ions in mmol/L: 142 Na⁺, 5 K⁺, 2.5 Ca⁺ 1.5 Mg⁺, 4.2 ${\text{HCO}}_3^{-}$ , 103 Cl⁻, 1 ${\text{HPO}}_4^{2-}$ and 0.5 ${\text{SO}}_4^{2-}$ at adjusted pH of 7.4 (Kokubo and Takadama, 2006). For immersion test, metal samples of 7.9 × 7.8 × 3 mm were prepared, weighed and immersed (in triplicate) in stagnant SBF solution at 37 ± 1 °C. The samples were removed after 21 days and rinsed with distilled water then allowed to dry at room temperature. Prior to corrosion product cleaning, surface morphology, elemental compositions and formed corrosion products were analysed and evaluated. Corrosion products were then removed from the samples by a chemical cleaning as specified in the ASTM G1 standard (Standard Practice for Preparing, Cleaning, and Evaluating Corrosion Test Specimens). Cleaned samples were then weighed to determine the weight loss. A flame atomic absorption spectrometry (Perkin Elmer, HGA 9000 model) technique was used to determine the concentration of Zn and Mg ions released by the samples into the SBF solution during the immersion test period. Corrosion rate of the samples was calculated by using equation: Corrosion rate = (KW)/(ATD), where the coefficient K = 8.76 × 10⁴, W is the weight loss (g), A is the sample area exposed to the SBF solution (cm²), T is the exposure time (h) and D is the density of the exposed material (g/cm³). Electrochemical corrosion test was conducted using potentiodynamic polarization and electrochemical impedance spectroscopy (EIS) methods by means of a potentiostat (VersaStat-3, Princeton Applied Research, US). The three-electrode system was used with the metal sample; graphite rod and saturated calomel electrode (SCE) served as the working, counter and reference electrodes, respectively. The test was carried out at 37 °C after 15 min stabilization time at the open-circuit potential (OCP) at a potential range from −200 to +700 mV against the OCP and a scan rate of 2 mV/s. Tafel fitting as suggested by Atrens et al. (2011) was used to approximate the corrosion potential and corrosion current which then used to calculate the corrosion rate of the alloy (CR) using the equation: $\mathit{CR}=3.27\ast {10}^{-3}\frac{i_{\mathit{corr}}\mathit{EW}}{\rho }$ , where EW is equivalent weight and ρ is density of the alloy.

3.1 Microstructure observation

Fig. 1 shows microstructure of the as cast, homogenized and ECAP samples of Zn-3Mg alloy. The as cast alloy consisted of Zn-rich dendrites dispersed in segregated form within the Mg₂Zn₁₁ phase (Fig. 1a). The dendrites were mostly dissolved during homogenization (Fig. 1b), with some of undissolved dendrites appear to have been experienced breakdown and transformation into connected precipitates. The 1-ECAP alloy possesses a near equiaxial microstructure with finer grain size but some elongated dendrites are still observed (Fig. 1c). In 2-ECAP alloy, the grain size is further refined with microstructure composed of complete equiaxed grains (Fig. 1d).

Close

Fig. 1

Zn-3Mg alloy microstructure: (a) as cast, (b) homogenized, (c) 1-ECAP, and (d) 2-ECAP.

Export to PPT

Table 1 details the results of grain size measurement of the alloy. In average, the grain size was decreased progressively from about 48 µm (as cast) to 30 µm (homogenized), 2.3 µm (1-ECAP) and 1.8 µm (2-ECAP). The grain size is distributed within a range of 30–150 µm (as cast) to 20–80 µm (homogenized) and 0.9–12 µm (2-ECAP) indicating the effectiveness of the process in refining the grains. Furthermore, percentage of porosity was also measured and shows a decrease from about 7% (as cast) to 4% (homogenized) and 0.04% (1-ECAP) till 0.02% (2-ECAP) indicating the effectiveness of the process in eliminating the defect generated by the primary process (casting).

Fig. 2a shows XRD pattern of homogenized, 1-ECAP and 2-ECAP Zn-3Mg alloy detecting the presence of hexagonal structure of Zn-rich phase (JCPDS: 00-004-0831) and cubic structure of Mg₂Zn₁₁ phase (JCPDS: 03-065-1853). The study by Li et al. (2003) reported the presence of intermetallic phases of Zn and Mg₂Zn₁₁ which formed above 325 °C during casting of Zn and Mg metals. There was a decrease of crystalline size of homogenized Zn-3Mg alloy after subjected to ECAP process. The crystallite size of Zn-rich phase on homogenized, 1-ECAP and 2-ECAP sample was 24.86 nm, 10.36 nm and 9.51 nm whilst for Mg₂Zn₁₁ phase was 14.30 nm, 10.75 nm and 10.45 nm, respectively.

Close

Fig. 2

(a) XRD pattern and (b) tensile curves of the Zn-3Mg alloy samples.

Export to PPT

3.2 Mechanical testing

Fig. 2b shows tensile curves of the alloys at different metallurgical conditions indicating an obvious improvement on tensile strength as well as elongation of the Zn-3Mn alloy after subjected to ECAP. As detailed in Table 2, an increase in tensile strength, yield strength and elongation was achieved by the 2-ECAP sample at about 160%, 215%, and 400% higher than those of the as cast alloy, respectively. Results from Vickers microhardness measurement indicates a similar trend where hardness decrease from 200 Hv (as cast) to 175 Hv (homogenized) and then a slight increase to 180 Hv (1-ECAP) and 186 Hv (2-ECAP).

Fig. 3 shows fracture surface of the samples after subjected to tensile tests. The morphology of both as cast (Fig. 4a) and homogenized alloy (Fig. 4b) is characterized by a relatively flat facet surface indicating a brittle fracture behavior. More dimples are noticed on the surface of ECAP samples (Fig. 4c–d) indicating a transitional fracture behavior from fully brittle to slightly ductile.

Close

Fig. 3

Fracture surface of Zn-3Mg alloy samples after tensile tests: (a) as cast, (b) homogenized, (c) 1-ECAP, and (d) 2-ECAP.

Export to PPT

Close

Fig. 4

Surface morphology of: (a) homogenized, (b) 1-ECAP and (c) 2-ECAP samples after immersed in SBF solution for 21 days, and (d) their corresponding XRD pattern.

Export to PPT

3.3 Corrosion testing

The corrosion rate of homogenized, 1-ECAP and 2-ECAP Zn-3Mg alloy samples after being immersed in the SBF solution for 21 days is listed in Table 3. Corrosion rate of the homogenized sample is similar to that of as cast alloy previously reported (Dambatta et al., 2015). A decrease on the corrosion rate is observed for the ECAP samples and it is supported by the decrease in Zn and Mg ions concentration released from the samples into the SBF solution during the immersion test.

Fig. 4 shows surface morphology of the Zn-3Mg alloy samples after being subjected to the immersion test. A porous corrosion layer dominates the surface morphology of homogenized sample (Fig. 4a). A rather granulated corrosion layer characterizes the surface of 1-ECAP sample (Fig. 4b) which seems different to that of 2-ECAP where crystal-like structure is formed (Fig. 4c). The EDS analysis revealed the presence of Cl, Zn, Mg, Ca, P, and O elements on the corroded surfaces. The XRD pattern (Fig. 4d) further detected the presence of zinc hydroxide [Zn(OH)₂] (JCPDS: 38-0385) and hydrozincite [Zn₅(OH)₆(CO3)₂] (JCPDS: 54-0047). Previous studies have shown that Zn-Mg alloys (with Mg content of 1–16 wt%) immersed in aggressive chloride containing solution usually formed corrosion products consisting of insoluble hydrozincite [Zn₅(OH)₆(CO₃)₂] and traces of simonkolleite [Zn₅(OH)₈Cl₂·H₂O] (JCPDS: 07-0155) Prosek et al., 2008.

Fig. 5a depicts Nyquist plots of the three Zn-3Mg alloy samples. The plots of ECAP samples indicate the existence of two capacitive loops at high and mid frequencies. Previous studies reported that capacitive loop observed at high frequency corresponds to charge transfer reaction in the electric double layer that occurred at the exposed surface and solution interface (Prosek et al., 2008; Zucchi et al., 2006). Fig. 5b shows potentiodynamic polarization curves of the alloy that indicate a little change to corrosion potential of the alloy regardless their metallurgical condition resulting into a narrow range of corrosion rates at around 0.24–0.30 mm/year (Table 4). The dielectric properties of the layer were analyzed by proposing an equivalent circuit (inset Fig. 5a) and the values are presented in Table 4. In addition to the presence of Warburg impedance, the charge transfers resistance (R_ct) value of 1-ECAP is higher than that of 2-ECAP samples. Conversely, homogenized sample gives the least value of R_ct compared to ECAP samples. It is well established that higher R_ct value corresponds to higher corrosion resistance (Rondelli et al., 2005).

Close

Fig. 5

(a) Nyquist plots and (b) polarization curves of homogenized, 1-ECAP, and 2-ECAP Zn-3Mg alloy samples tested in SBF solution at 37 °C.

Export to PPT

4.1 Microstructure evolution and mechanical properties

The dendritic structure of the as cast Zn-3Mg alloy was dissolved during homogenization and transformed into equiaxed grains during the 2 steps ECAP process (Fig. 1). The grain size refinement during ECAP is attributed to the interaction of shear and recrystallization that occurs along the equal channel intersection angle (Verlinden, 2005). The process gives also similar effect on the reduction of casting defects such as porosity. The initial coarse grain size of the as cast alloy was reduced by 96% due to the combining process of homogenization and two passes of ECAP (Table 1). This notable grain refinement can be attributed to the first partial refinement by the homogenization that speeds up the process in ECAP. It was observed that the initial grain size prior to ECAP plays a vital role in controlling the rate of refinement at the final stage of ECAP process (Mostaed et al., 2014). The present combined technique reduces the number of ECAP pass thus potentially reduces production cycle, time and cost. Most of the reported grain size refinement in the range of 1–2 µm for Zn-Al alloys was achieved after conducting 8–12 ECAP passes (Pürçek, 2005; Al-Maharbi et al., 2010; Aydın, 2012). In addition, there is still a room for optimization such as by progressively decreasing the ECAP temperature in the range of 190–210 °C to allow higher plastic straining and add more passes without causing cracks. This manipulation was found useful in the ECAP process for Mg-Zn alloy (Mostaed et al., 2014). The XRD pattern reveals that there is no major variation in phase formation between the homogenized and the ECAP Zn-3Mg alloy. Phase identification on both conditions indicates the presence of Zn-rich and Mg₂Zn₁₁ phases (Fig. 2a). The crystallite size of Zn-rich and Mg₂Zn₁₁ phases of homogenized alloy decreases after being processed through ECAP which correlates very well with the SEM observation on the reduction of grains size (Fig. 1). Similarly, remarkable reduction of porosity was observed after 2-ECAP passes. The samples processed up to 2-ECAP passes exhibit a reduction of percentage porosity of about 300 times compared to as cast sample (Table 1). This implies that reduction of the observed porosity after homogenization followed by ECAP process contributes also toward improving the alloy’s ductility. Therefore, as expected the refined grains structure improves mechanical properties of the alloy (Fig. 2b). The combined homogenization and the two passes of ECAP increases both strength and ductility (Table 2), in contrast with the results obtained by strain hardening treatment such as extrusion that increase the strength in expense of ductility. The increase in strength and ductility at the same time could be attributed to fine microstructure and high number of grains size that lead to increase of volume density of high-angle grain boundaries in the Zn-3Mg alloy after ECAP. The grain boundaries impinge dislocation movement and hence increase the alloy’s strength but at the same time the large number of individual grains increases the possibility of slip to occur and hence improves the alloy’s ductility (Chuvil’Deev et al., 2004; Valiev et al., 2002). The increase of ductility may also be a result of the breakdown and homogenous distribution of hard Mg₂Zn₁₁ intermetallic phase in addition to the reduction of porosity. The small change in hardness may indicate a nearly saturated effect of ECAP that leads to no more increase of hardness with additional ECAP passes. This is in consistent with previous finding on Zn-Al alloy subjected to ECAP (Al-Maharbi et al., 2010). The fracture morphology of ECAP samples show a ridge-like features radiating outwards with lesser flat facet area but more dimples for 2-ECAP (Fig. 3). This feature indicates an occurrence of typical brittle fracture that took place with little appreciable deformation (Callister and Rethwisch, 2012). However, the indication of a transitional fracture behavior from full brittle to slight ductile (Fig. 3) correlates very well with the increase of elongation (Fig. 2b).

4.2 Corrosion behavior

The immersion and electrochemical corrosion tests revealed that corrosion rate of homogenized and ECAP samples (0.24–0.30 mm/year) is comparable to that of as cast alloy (0.23 mm/year) Dambatta et al., 2015. This can be attributed to the effect of Mg alloying on the formation of a relatively stable oxide layer on the alloy’s surface. In more detail, the results show that both ECAP samples exhibited lower corrosion rate when compared to homogenized sample (Table 3). This could be attributed to the uniform distribution of the Mg₂Zn₁₁ intermetallic phase after ECAP in addition a homogenous grain structure. A slight increase in corrosion rate of 2-ECAP compared to 1-ECAP (Table 4) can be related to the existence of finer grains size observed in the microstructure of 2-ECAP sample (Fig. 1d) leading to the formation of a less resistive corrosion layer (Table 4). The measured slight higher ions concentration released from 2-ECAP compared 1-ECAP sample (Table 3) further supports this finding. This is in agreement with previous corrosion study that reported that samples with fine-grained microstructure may likely to corrode faster than those consisted of coarse-grained microstructure (Kutniy et al., 2009). The analysis on EIS result indicates the presence of a more resistive corrosion layer formed on both ECAP samples compared to the homogenized sample (Fig. 5a and Table 4). The presence of mid-frequency capacitive loop on ECAP samples implies corrosion mechanism occurred due to mass transport in solid phase such as diffusion of ions through oxide layers (Walter and Kannan, 2011). Conversely, the absence of mid frequency loop was noticed on the homogenized sample which could be attributed to poorly formed protective layer as observed on the surface morphology (Fig. 4a). The corrosion layer, as revealed by the XRD analysis (Fig. 4d), is composed of hydrozincite where its formation may be related to high local pH magnesium rich corrosion product. Earlier investigation reported a stable hydrozincite at higher pH value (Lindström et al., 2002; Volovitch et al., 2009).