Introduction

Due to their low density (in the range of 1.74–1.95 g/cm31) and relatively high strength, magnesium-based alloys are increasingly used in various industries, such as automotive, aerospace, electronics, and sports2,3,4,5. Some magnesium-based alloys (among others Mg-Zn-Ca) are biocompatible with the human body6. Mg2+ ions, naturally present or produced during corrosion of biodegradable implants7,8, are involved in many metabolic reactions and biological mechanisms. An excess of Mg2+ ions is easily excreted in the urine9. Therefore, Mg-based alloys can find applications in medicine as a material for biodegradable implants, namely screws, plates, pegs or stents6,7,8,10,11,12,13. Corrosion of Mg and Mg-based alloys involves reactions (1)–(3)14,15.

$${\rm{Mg}}\to {{\rm{Mg}}}^{2+}+{2{\rm{e}}}^{-}$$
(1)
$${2{\rm{H}}}_{2}{\rm{O}}+{2{\rm{e}}}^{-}\to {2{\rm{OH}}}^{-}+{{\rm{H}}}_{2}$$
(2)
$${{\rm{Mg}}}^{2+}+{2{\rm{OH}}}^{-}\to {{\rm{Mg}}({\rm{OH}})}_{2}$$
(3)

The overall corrosion reaction can be expressed by reaction (4):

$${\rm{Mg}}+{2{\rm{H}}}_{2}{\rm{O}}\to {{\rm{Mg}}({\rm{OH}})}_{2}+{{\rm{H}}}_{2}$$
(4)

Among the Mg-based alloys, the most widely studied and used alloys are Mg-Al and Mg-Zn. In contrast to Al (which can induce various neurological disorders in the human body), Zn plays an important role in cellular metabolism and stabilization of cell membranes. It has a beneficial effect on the wound healing process16,17 and has an anti-rheumatic effect in the case of rheumatoid arthritis16,18. The recommended daily intake of zinc is 10 mg19. Excess Zn induced by biodegradation of implants is neurotoxic20. A review of the literature21,22,23,24,25,26 shows that the microstructure and, therefore, the corrosion resistance of Mg-Zn alloys strongly depend on the Zn content. It was found that the addition of zinc can lead to a higher corrosion potential compared to that of pure Mg21,22 and to a reduction of hydrogen evolution kinetics23. It was also proposed27,28,29 that zinc can reduce the galvanic coupling effect between impurities (Fe, Cu, Ni, Co) and the matrix. The presence of Zn also affects the formation of the protective film during the corrosion process in simulated body fluids24,25,26.

The corrosion rate of biodegradable implants must be sufficiently low to maintain their mechanical properties and to avoid significant physical-chemical changes in the human body (formation of gas cavities and hydroxides) during the period of bone healing. It was proposed that the corrosion rate of biomedical degradable materials should be within the range of 0.2–0.5 mm per year30,31,32,33,34,35, depending on the type of implants and their position in the body. The corrosion rates of well-known crystalline Mg-Zn-based alloys do not meet these requirements36.

Several strategies may be adopted to increase the corrosion resistance of Mg-Zn-based alloys37, including the addition of some other elements (such as Ga, Ag, Ca, Zr, Cu, and rare earth elements) and the application of a prior surface preparation (coatings38,39, for example). The addition of other elements may enhance the glass-forming ability of the alloy40. It also may promote the formation of a wider variety of oxides, leading to a denser corrosion layer that protects the substrate37,41,42. Increasing corrosion resistance by adding other elements may result from microstructural changes that lead to a minimal difference in potentials between metallic phases43. Under these conditions, the microgalvanic coupling effect is reduced. The addition of alloying elements has a strong influence on the corrosion potential value of the alloy. The corrosion potential of Mg-Zn-Ca alloys (which have attracted considerable attention worldwide44) in simulated body fluid at 37 °C can vary by several hundred mV depending on the Zn and Ca contents and the addition of other elements (Cu, Ga, Yb, Ag)6,45,46,47,48,49,50. The tendency to oxidation is reduced by shifting the corrosion potential in the anodic direction.

An alternative route to significantly decrease the corrosion rate is to produce Mg-Zn-based glassy alloys. These materials have an amorphous structure. Over the last five years, several papers proposed a comprehensive review on the processing, mechanical and biological characteristics, and corrosion resistance of Mg-based metallic glasses37,41,43,44. Their very good corrosion resistance is attributed to the absence of microstructural features (grain boundaries, dislocations, metallic phases acting as anodes and cathodes, etc.). The formation of supersaturated solutions with alloying elements (as a result of rapid cooling) also explains the improvement in corrosion resistance43. The presence of certain alloying elements (such as Ga45) can contribute to the formation of a protective layer during corrosion processes, thus improving corrosion resistance. Although the corrosion rates of existing Mg-Zn-based metallic glasses are significantly lower than those of crystalline Mg-Zn-based alloys6,41,44, they do not meet clinical application requirements (biocompatibility and degradation rate).

Taking into account the above facts, it seems justified to continue searching for new compositions of alloys with an amorphous structure. Platinum is a noble element with a high open circuit potential (OCP) value (of +200 mV vs SCE) in Hanks’ solution at 37 °C51. In addition, it has high chemical stability and excellent biocompatibility. In the present paper, Pt is selected as the alloying element of Mg-Zn-based metallic glass alloy. Amorphous and crystalline Mg72Zn27Pt1 samples were produced according to the procedure described in ref. 40. Their microstructure was first investigated using field emission scanning electron microscopy coupled with energy dispersive X-ray spectroscopy (FE-SEM/EDS) and X-ray diffraction (XRD). Next, their corrosion behavior was determined in Hanks’ solution at 37 °C using linear sweep voltammetry (LSV), electrochemical impedance spectroscopy (EIS), and zero resistance ammeter (ZRA) techniques. Values of the corrosion rate were assessed from linear polarization resistance (LPR) measurements. Results derived from corrosion tests were analyzed considering the structure of the samples. They are also discussed on the basis of the literature on other types of Mg-Zn-based metallic glasses.

Results and discussion

Microstructure of crystalline Mg72Zn27Pt1

Figure 1a, b shows FE-SEM images of the microstructure of crystalline Mg72Zn27Pt1, acquired in the LABE mode to reveal the chemical contrast between phases. Five different phases were observed. EDS analyses were carried out in each phase to determine its chemical composition (Table 1).

Fig. 1: FE-SEM images of the crystalline Mg72Zn27Pt1 after mechanical polishing using diamond suspension (1 µm).
figure 1

Magnification (a) 200× and (b) 1000×. Reading the chemical composition at the marked points.

Full size image
Table. 1 Results from EDS analyses (at.%) performed in different points visible in Fig. 1a, b
Full size table

The two main phases were composed of Mg and Zn (points 1a and 1b) and Mg, Zn with a very small amount of Pt (points 2a and 2b). They may correspond to Mg0.7Zn0.3 and Mg0.7Zn0.3(Pt0.002). These two phases can only be distinguished in FE-SEM images by greatly increasing the contrast (Fig. 1a). Under classical conditions, both appear in the same color (Fig. 1b). Three other phases were also found, namely Al0.49-Fe0.46 with a small amount of silicon, manganese, zinc and magnesium (around 1 at.%, points 3a and 3b), Mg0.97Zn0.03 (points 4a and 4b) and Mg0.48Al0.26Pt0.25 with a small amount of Zn (points 5a and 5b).

Figure 2 shows the XRD pattern of the crystalline Mg72Zn27Pt1. Numerous sharp peaks are visible, confirming that they are well-crystallized. The peaks were assigned to metallic phases by taking into account ICCD cards available in the literature and EDS results presented previously. Mg51.04Zn19.8 (ICCD card n° 01-071-9627), whose composition is close to that of Mg0.7Zn0.3 and Mg0.7Zn0.3(Pt0.002) found by EDS, matches numerous peaks of the experimental XRD diagram (blue lines in Fig. 2). The three compounds have a Mg / Zn ratio close to 2.4. The phase containing mainly iron and aluminum was assigned to Al-Fe (ICCD card n° 01-071-9627), whose composition is close to Al0.49-Fe0.46 (with a small amount of silicon, manganese, zinc, and magnesium) found by EDS. Peaks associated with this phase are located by red lines in Fig. 2. The phase containing mainly magnesium (Mg0.97Zn0.03 found by EDS) was assigned to Mg0.971Zn0.029 (ICCD card n° 01-071-9628, green lines in Fig. 2).

Fig. 2: XRD pattern of the crystalline Mg72Zn27Pt1.
figure 2

XRD patterns of ingots.

Full size image

From the XRD analysis described above, the phase ratio was calculated to be 93.4 vol% Mg51.04Zn19.8, 5.6 vol.% Mg0.971Zn0.029, and 1 vol.% Al-Fe. As Al and Fe are present in the alloy as impurities, the very low content of the Al-Fe phase was expected. Only four XRD peaks were not indexed (at 20.819°, 25.828°, 31.671°, and 39.726° in Fig. 2). These peaks could correspond to Mg0.48-Al0.26-Pt0.25 (with a small amount of Zn) revealed by EDS. No ICCD cards were found for compounds with a comparable chemical composition.

Microstructure of amorphous Mg72Zn27Pt1

Figure 3a shows FE-SEM images of the surface of the amorphous Mg72Zn27Pt1 in LABE mode. They exhibit a smooth surface, both at low-spatial (Fig. 3a) and high-spatial (FE-SEM image insert in Fig. 3a) resolutions. The amorphous Mg72Zn27Pt1 was tilted in the FE-SEM chamber (tilt angle of 75°) to obtain backscatter Kikuchi patterns in different locations. No bands were observed, Fig. 3b, confirming their amorphous character. As expected, the XRD pattern exhibits a broad diffuse contribution in the angle range of 2θ from 30 to 45° (Fig. 3c), which clearly indicates the amorphous structure of the samples. Therefore, the amorphous Mg72Zn27Pt1 was characterized by the absence of any microstructural features and classical lattice defects known from crystalline materials, such as dislocations, grain boundaries, stacking faults, etc. However, they contain internal structural inhomogeneities called free volumes52,53.

Fig. 3: Confirmation of amorphousness of Mg72Zn27Pt1 ribbon.
figure 3

a FE-SEM images, (b) backscatter Kikuchi pattern, and (c) XRD pattern of the amorphous.

Full size image

Corrosion behavior of the crystalline and amorphous Mg72Zn27Pt1 in Hanks’ solution

Figure 4a shows the polarization curves of the amorphous Mg72Zn27Pt1 (red curve) and the crystalline Mg72Zn27Pt1 (black curve) in Hanks’ solution. In the anodic domain of both samples, no passive region is identified. The current density increases as the applied potential increases, indicating that they exhibit active electrochemical behavior. The polarization curve of the amorphous Mg72Zn27Pt1 was found to be shifted to noble potential values as compared to that of the crystalline Mg72Zn27Pt1. The difference between the corrosion potential of the two samples is large (about 300 mV). This indicates a lower oxidation tendency of the amorphous Mg72Zn27Pt154.

Fig. 4: Electrochemical corrosion results of Mg72Zn27Pt1.
figure 4

a Polarization curves (1 mV s−1) of the crystalline and amorphous samples in Hanks’ solution at 37 °C. b Values of the OCP (determined from polarization curves) of various Mg-based metallic glass alloys (data from refs. 6,45,46,47,48,49,50) compared to that of the amorphous Mg72Zn27Pt1.

Full size image

The corrosion potential of the amorphous Mg72Zn27Pt1 is equal to −1013 mV vs Ag/AgCl (Fig. 4a). This value was compared to those obtained under similar conditions (in simulated body fluids at 37 °C) on various Mg-Zn-Ca metallic glass alloys6,45,46,47,48,49,50, Fig. 4b. It appears that the presence of 1 at.% Pt in the amorphous Mg72Zn27Pt1 leads to the highest anodic OCP value, and therefore to the lowest oxidation tendency, than those generally found in the literature for binary Mg-Zn alloys with both crystalline and amorphous structures.

Corrosion resistance of the crystalline and amorphous Mg72Zn27Pt1 in Hanks’ solution

EIS and LPR tests were then performed to calculate the polarization resistance and the corrosion current density. The lower the polarization resistance, the higher the corrosion current density and the corrosion rate55. This gives a description of the corrosion resistance of the crystalline and amorphous Mg72Zn27Pt1 in Hanks’ solution at 37 °C, near their open circuit potential.

Figure 5a, b shows the EIS diagrams (Nyquist plots) of the crystalline and amorphous Mg72Zn27Pt1 in Hanks’ solution (37 °C) at the OCP. In both cases, depressed capacitive semi-circles are observed. EIS diagrams were fitted using the electrical equivalent circuit shown in Fig. 5c. This circuit is widely used to describe the corrosion behavior of magnesium alloys56. It contains two time constants Rs_(R1//CPE1)_(R2//CPE2), where Rs is the solution resistance, (R1//CPE1) represents the change transfer process at the surface/electrolyte interface at high frequency, and (R2//CPE2) describes the formation/dissolution of corrosion products at low frequency. The constant phase element (CPE) is used to simulate the capacitive response of a non-ideal capacitor, due to the heterogeneous nature of the electrode surface. The impedance of the CPE is given in Eq. (5).

$${Z}_{{CPE}}=\frac{1}{{Y\left(j\omega \right)}^{n}}$$
(5)

where Y is the value of admittance expressed in (S cm−2 sn), ω is the angular frequency (rad s−1), j is the imaginary number (j2 = (−1)) and n is a dimensionless fractional exponent. When n = 1, 0, or −1, the CPE represents an ideal capacitor, a resistance, and an inductor, respectively. Numerical data derived from fitting are reported in Table 2.

Fig. 5: Impedance diagrams (Nyquist plots) of Mg72Zn27Pt1 in Hanks’ solution at 37 °C.
figure 5

(a) amorphous sample, (b) crystalline sample, (c) scheme of fitting using the equivalent circuit. Circles and triangles represent experimental data, and solid lines represent fitting.

Full size image
Table. 2 Fitting results of the EIS data acquired at the OCP on the crystalline and amorphous Mg72Zn27Pt1 in Hanks’ solution at 37 °C
Full size table

The polarization resistance was calculated as the sum of R1 and R2. Values of 94 and 2440 Ω·cm2 were found for the crystalline and amorphous Mg72Zn27Pt1, respectively. Values of the same order of magnitude were found during LPR measurements (203 Ω·cm2 for the crystalline Mg72Zn27Pt1 and 1700 Ω·cm2 for the amorphous Mg72Zn27Pt1). Average values of the polarization resistance (calculated from EIS and LPR experiments) are 148 and 2070 Ω·cm2 for the crystalline and the amorphous Mg72Zn27Pt1, respectively. The corrosion current density was then calculated from these average values using Eq. (6)57.

$${i}_{0}=\frac{{RT}/{nF}}{{Rp}}$$
(6)

where i0 is the corrosion current density, R is the universal gas constant (8.314 J K−1 mol−1), T is temperature expressed in kelvin (310.15 K), n is the number of electrons transferred in the reaction (n = 2), F the Faraday constant (96,487 C) and Rp the polarization resistance. Values of 88 and 6 µA/cm2 were obtained for the crystalline and amorphous Mg72Zn27Pt1, respectively. The corrosion current density of the crystalline Mg72Zn27Pt1 is nearly 15 times greater than that of the amorphous Mg72Zn27Pt1. This shows that the amorphous Mg72Zn27Pt1 is significantly more resistant to corrosion than the crystalline Mg72Zn27Pt1 in Hanks’ solution at 37 °C. The value found on the amorphous Mg72Zn27Pt1 is of the same order of magnitude as those obtained on Mg65-xZn23Ca4Agx (x = 1, 3) ribbons immersed in Hanks’ solution at 37 °C56. Silver is another noble element.

Corrosion mechanisms in Hanks’ solution at 37 °C

Polarization curves of the amorphous Mg72Zn27Pt1 reveal the presence of a two-stage breakdown mechanism. The first stage (region I in Fig. 4a) occurs for applied potentials between E0 and E1. The current density at the end of this stage is 40 µA/cm2. FE-SEM observations were performed after interrupting the polarization curve within this stage. Filiform corrosion is observed with the presence of long filaments, as shown in Fig. 6a, b. FE-SEM/EDS analyses were performed in filaments (first row in Table 3). They are composed of magnesium and oxygen, with a small amount of zinc and traces of elements from the solution (Na, P, Cl, K, and Ca).

Table. 3 Results from FE-SEM/EDS analyses (at.%) performed in corrosion products
Full size table
Fig. 6: FE-SEM images of Mg72Zn27Pt1 after polarization.
figure 6

Microstructure of the amorphous samples after polarization curves interrupted (a), (b) between E0 and E1, and (c) at E2; (d) and (f) microstructure of the crystalline sample after polarization; and (e) typical EDS spectrum of corrosion products.

Full size image

During stage I, FE-SEM images also show that corrosion starts to propagate in the substrate, perpendicularly to the filaments formed by the filiform degradation. This leads to the formation of nodules (Fig. 6a, b).

The second stage (region II in Fig. 4a) is observed for potentials greater than E1 = −950 mV vs. Ag/AgCl. In region II, the current density increases rapidly, reaching a limiting value of 4 mA cm−2 at E2 = −850 mV vs Ag/AgCl. Surface observations performed after interrupting the polarization curve at E2 show the existence of some strongly oxidized sites (Fig. 6c). These sites extend over a distance of several hundred micrometers. They result from the preferential growth of nodules vs. filaments. Filiform corrosion, which is the main corrosion mechanism in stage I, is only observed at the periphery of strongly oxidized sites (FE-SEM image insert in Fig. 6c). Data derived from FE-SEM/EDS analysis of corrosion products are reported in Table 3 (second row). They are again composed of magnesium, oxygen, and a small amount of elements from the solution (Na, P, Cl, K, and Ca), reactions (1)–(4). As explained in the introduction, these reactions induce an increase of the solution pH. Surface observations of oxidized sites and chemical analyses of corrosion products indicate that corrosion mechanisms involve the same electrochemical reactions (reactions (1)–(4)) during corrosion of the crystalline and amorphous Mg72Zn27Pt1.

The anodic branch of the crystalline Mg72Zn27Pt1 exhibits a single stage, black curve in Fig. 4a. Strongly oxidized sites are visible (Fig. 6d). FE-SEM/EDS analyses show that corrosion products are again composed of magnesium and oxygen (with a small amount of Na, P, Cl, K, and Ca from the solution), third row in Table 3. A typical EDS spectrum is shown in Fig. 6e. Therefore, as in the case of the amorphous Mg72Zn27Pt1, corrosion involves reactions (1)–(4). However, in contrast to the amorphous Mg72Zn27Pt1, numerous cracks are present in strongly oxidized sites (Fig. 6d, f). Oscillations of the current density in the anodic domain of the polarization curve can result from the formation of these cracks, leading to passivity breakdown. These oscillations (and therefore the cracks) appear from the beginning of the anodic domain (Fig. 4a). Corrosion of crystalline Mg-based alloys often leads to cracking, mechanical deformation, and spallation58. It was suggested59 that cracking and mechanical deformation are induced by hydrogen uptake by the substrate. Hydrogen is produced by reaction (2). Increasing mechanical deformation accelerates corrosion, as it was already shown for other Mg-based crystalline alloys60. Such a mechanism involving hydrogen uptake by the substrate, followed by the increase of mechanical deformation promoting corrosion, may occur in the case of the crystalline Mg72Zn27Pt1 immersed in the Hanks’ solution. To confirm this assumption, further measurements will be performed to measure the amount of hydrogen uptake by the crystalline Mg72Zn27Pt1 (by means of TOF-SIMS, for example61) and the increase of mechanical deformation (using XRD and stress analysis).

Galvanic couple in Hanks’ solution at 37 °C

Figure 7 shows the evolution vs. time of the galvanic current density (red curve) and the galvanic potential (black curve) when the amorphous Mg72Zn27Pt1 is coupled with the crystalline Mg72Zn27Pt1 (ground). After immersion for 1.3 h, both parameters reach a steady state. The galvanic potential is around −1200 mV vs Ag/AgCl. This value is very close to the coupled potential that can be determined from polarization curves in Fig. 4a. Therefore, the crystalline Mg72Zn27Pt1 is more anodically polarized compared to that immersed alone in the solution (previous sections). Negative values of the current density were measured, indicating that electrons flew from the crystalline Mg72Zn27Pt1 (acting as anode) to the amorphous Mg72Zn27Pt1 (acting as cathode).

Fig. 7: ZRA measurements of Mg72Zn27Pt1.
figure 7

Evolution vs. immersion time of the galvanic potential and galvanic current density between the crystalline and amorphous Mg72Zn27Pt1 in Hanks’ solution at 37 °C.

Full size image

Surface observations of the crystalline Mg72Zn27Pt1 after ZRA test show that the initiation of corrosion can be associated with the Mg0.971Zn0.029 phase. Figure 8a, b shows sites containing Mg0.971Zn0.029 grains in the early stages of corrosion, while other grains are already strongly corroded. These sites do not contain Mg0.48-Al0.26-Pt0.25(Zn), indicating that this phase certainly does not play a critical role in corrosion processes. Once corrosion proceeds, cracks are generated in the main Mg51.04Zn19.8 phase (Fig. 8c, d). According to refs. 58,59, this may be due to hydrogen uptake by this phase. Corrosion propagates preferentially along these cracks (Fig. 8d). FE-SEM/EDS analysis in point 1 yields: 35.5 at.% O, 32.6 at.% Mg, 9.3 at.% Zn, 12.8 at.% C, 5.7 at.% P, 1.4 at.% Cl, 1.5 at.% Ca and 0.8 at.% Na. Therefore, corrosion products are mainly composed of magnesium and oxygen (reactions (1)–(4)).

Fig. 8: FE-SEM images of Mg72Zn27Pt1 after ZRA test.
figure 8

a–e Crystalline samples, and (f) amorphous samples. Reading of chemical composition in area 1 marked in (c).

Full size image

After ZRA test for 1.3 h, numerous sites are strongly oxidized in the crystalline Mg72Zn27Pt1. These sites extend over several hundred micrometers, both in Mg0.971Zn0.029 and Mg51.04Zn19.8 phases (Fig. 8e). Filiform corrosion is never observed. Therefore, corrosion mechanisms are slightly different from those observed when samples are immersed alone (no coupling processes). By contrast, the amorphous Mg72Zn27Pt1 remains intact (cathode) (Fig. 8f).

Methods

Synthesis of alloys and surface preparation

Crystalline Mg72Zn27Pt1 was prepared by induction melting of magnesium, zinc, and platinum in an argon atmosphere (NG-40 induction generator). High-purity metals (99.9% Mg, 99.9% Zn, and 99.95% Pt) were provided in the form of rods by Goodfellow (diameter of 10 mm for Mg and Zn and 1 mm for Pt). According to the certificates, a small amount of impurities can be found in Mg (280 wt. ppm Fe, 170 wt. ppm Mn, 70 wt. ppm Al, 50 wt. ppm Si, and 20 wt. ppm Cu), Zn (<200 wt. ppm Cd, <300 wt. ppm Pb, <100 wt. ppm Fe), and Pt (50 wt. ppm Pd).

A first set of crystalline Mg72Zn27Pt1 was mechanically polished using silicon carbide (SiC) papers (down to 4000). Between each step, samples were cleaned in ethanol in an ultrasonics cleaner for 1 min. These polished samples were used for microstructural characterizations and electrochemical investigations.

A second set of crystalline Mg72Zn27Pt1 was placed in the quartz crucible with slit nozzle (0.4 × 10 mm), heated inductively by high frequency generator and melt spun into rotating copper wheel (melt spinner HV from Edmund Buehler GmbH, Germany)62. The distance between the crucible nozzle and the wheel surface was set to 0.4 mm. The linear speed and ejection pressure were set to 50 m s−1 and 300 mbar, respectively. Microstructural and electrochemical investigations were performed on melt-spun ribbons with a thickness of about 150 µm (amorphous Mg72Zn27Pt1).

Microstructural characterizations

The microstructure of crystalline and amorphous Mg72Zn27Pt1 was characterized by means of XRD. These tests were carried out using CuKα (λ = 1.54 Å) as the radiation source. Measurements were carried out with a Bruker D8-A25-Advance diffractometer and a LynxEye detector. The full angular range was investigated (2θ in the range of 10–100°). XRD patterns were analyzed by the Topas software package and the Rietveld method (structural model)63.

A FE-SEM (JEOL 7600 F) with an integrated EDS was also used. Images were acquired in the low-angle back-scattered electron (LABE) mode. This detector is capable of producing qualitative compositional images with a very high degree of atomic number contrast. Images were also acquired in lower secondary electron imaging (LEI) mode to have topographical information. Chemical analyses in metallic phases were performed at an accelerating voltage of 15 kV. Under these conditions, the electron penetration depth in Mg72Zn27Pt1 alloy is 260 nm (predicted by Monte Carlo simulations using Casino v2.51). This indicates that analyses are actually performed in single phases. Chemical analyses of corrosion products were performed at an accelerating voltage of 5 kV. The electron penetration depth is close to 50 nm (simulations performed using Casino v2.51).

Corrosion tests

The corrosion behavior of crystalline and amorphous Mg72Zn27Pt1 was investigated in Hanks’ solution at 37 °C. The chemical composition of the Hanks’ solution is: 8 g/L NaCl, 0.4 g/L KCl, 0.140 g/L CaCl2, 0.350 g/L NaHCO3, 0.217 g/L NaH2PO4, 0.06 g/L Na2HPO4, 0.406 g/L 1.42 MgCl2•6H2O, 0.029 g/L MgSO4, and 1 g/L D-glucose. The solution is not buffered. The solution pH was measured, and a value of 7.2 was found. A classical three-electrode cell consisting of a silver/silver chloride reference electrode (Ag/AgCl, 3 M KCl), a platinum counter-electrode (Pt plate with a surface area of 4 cm2), and a working electrode (samples) was used to perform corrosion tests. PGSTAT128 Potentiostat/Galvanostat with the Nova 2.1 software package (Metrohm Autolab, Utrecht, The Netherlands) was also used.

Prior to all electrochemical measurements, the OCP was measured for 30 min (to reach steady state). LSV curves of crystalline and amorphous Mg72Zn27Pt1 were plotted from −200 mV vs. OCP to the anodic direction, at a potential scan rate of 1 mV/s. They were stopped at an applied potential of +200 mV vs. OCP. EIS measurements were performed on crystalline and amorphous Mg72Zn27Pt1 at the OCP. A sinusoidal AC wave of ±10 mV amplitude was applied from 50 kHz to 0.1 Hz. Fitting of experimental EIS spectra was carried out using the Z-View 4 software package (Scribner Associates, Southern Pines, NC, USA). LPR was determined by polarizing samples (crystalline and amorphous Mg72Zn27Pt1) in the potential range of ±15 mV vs. OCP at a potential scan rate of 1 mV s−1 (LPR tests). The current-potential plot within the investigated potential range was found to be linear, and the value of the LPR was calculated from the slope. In addition, ZRA tests were carried out to measure vs. time the galvanic potential and the galvanic current between crystalline and amorphous Mg72Zn27Pt1.