The detailed corrosion performance of bioresorbable Mg-0.8Ca alloy in physiological solutions

A.S.Gnedenkov, S.L.Sinebryukhov, V.S.Filonina, V.S.Egorkin, A.Yu.Ustinov, V.I.Sergienko,S.V.Gnedenkov

Institute of Chemistry of FEB RAS, 159 Pr.100-letiya Vladivostoka, Vladivostok, 690022, Russia

Abstract

Keywords: Magnesium-calcium alloy; Biodegradation; Corrosion rate; Electrochemistry; Weight loss; Minimum essential medium.?Corresponding author.

1.Introduction

1.1.Biomaterials

Nowadays, a tendency to increase in life quality leads to both an extension of materials sciences research area and the development of new technologies for biomedical appliances.Such technologies are orientated to the development of advanced methods and tools for one’s health resumption, using minimal medical interventions and collateral damage.For example, the appliance of traditional materials for implants(such as stainless steels, titanium, and its alloys, ceramics,polymer materials, etc.), faces the challenge of biodegradation unavailability [1–3].These permanent implants remain in the patient throughout his life but in some cases require removal.This generates a need for surgical reintervention for implant extraction [4–8].

An increase of interest in bioresorbable metallic materials,in particular, magnesium and its alloys is caused by a combination of their properties.Values of elastic modulus and compressive yield strength of magnesium are close to ones of natural bone[9].The presence of magnesium as a nutrient element of human body, makes its degradation process relatively harmless [9–11].Nevertheless, the properties of magnesiumbased bioresorbable materials are characterized by high electrochemical activity in physiological liquids(i.e.human blood plasma), which may cause a rapid loss of general integrityof an implant.Moreover, magnesium and most magnesium alloys are sensitive to local and heterogeneous degradation,which can cause a strength deterioration [12].Thus, there is a necessity for development and research of new magnesium alloys for biomedical applications.To inhibit the corrosion process and increase biocompatibility, different methods can be applied, including formation of protective coatings and the design of new alloys with different alloying elements[13–19].

1.2.Magnesium-calcium alloys study

Amongst all of the magnesium alloys, binary Mg-Ca alloys are of greatest interest for needs of implant surgery.Such alloys show a good biocompatibility and are non-toxic because calcium is also a natural constituent in a human body,like magnesium, which gives calcium an undeniable advantage over other alloying elements[20–24].Von Der Hoh et al.[20]tested the degradation behaviour of Mg-0.8Ca alloy in rabbits and established a good tolerance of this material for up to 6 months.Li et al.[21]revealed good biocompatibility of Mg-Ca alloys viain vitroandin vivotests using cytotoxicity evaluation and implantation into the rabbit femoral shafts, respectively.Feser et al.[22]studied the effects of Mg–Ca alloys on dendritic cell function and found excellent biocompatibility of the degradable material.Mohamed et al.[24]established the excellent biocompatibility and bioactivity of the Mg-0.8Ca alloy, which promote the attachment of human cell lines to the implant and assist in the bone healing process.Chen et al.[23]evaluated the high degradability,antibacterial property, biocompatibility and osteogenic activity of three Mg-Ca alloys, which can appear to be the good choice for orthopaedic implants.

However, a controversial issue today is the optimal concentration of calcium in the alloys to achieve the best mechanical and electrochemical properties.Different scientific groups [21,25–29]during the study of the properties of MgxCa alloys in a wide range of calcium concentrations(x=0–16.2 wt.%) revealed a significant decrease in the corrosion resistance of the material with an increase in the amount of calcium.Blajan et al.[30]identified that in simulated body fluid(SBF) medium, the corrosion destruction of the Mg-0.8Ca alloy slows down over time due to a uniformly growing layer of corrosion products while the Mg-1.8Ca alloy sample undergoes more intense destruction, which can negatively affect the healing process.Considering the consistency with the previously mentioned studies, it can be concluded that the best corrosion resistance is provided at a calcium concentration up to 1 wt.%.The results of Zheng et al.[31]show that when the weight concentration of calcium reaches 1 – 3 wt.%, the tensile strength and stretch elongation begin to decrease, and at high values of Ca concentration, these characteristics of the material may be insufficient for its operation as an implant.This allows us to conclude that amongst other magnesiumcalcium alloys Mg-0.8Ca possesses the most optimal mechanical characteristics and also has good electrochemical properties, which increases the urgency of its further and more detailed study.

Mei et al.[32]carried out a comparative analysis of the corrosion activity of the Mg-0.8Ca alloy and commercially pure magnesium in variousin vitrosolutions and stated that similarity of the corrosion behaviour of materials in SBF and MEM (minimum essential medium) depends on the formed corrosion products, which are controlled by the inorganic components of the electrolyte.In addition toin vitrotests,in vivotests were also carried out, confirming not only high biocompatibility but also a satisfactory level of mechanical and corrosive properties of the Mg-0.8Ca alloy [20,33–35].

Unfortunately, not all of the mentioned works present the corrosion rate measured by mass loss of the studied magnesium alloy.Some of them show only degradation rates calculated from the potentiodynamic polarization measurements.However, it is very interesting to measure this steady-state value and compare it with the intrinsic corrosion rate of highpurity magnesium in concentrated chloride solution (0.3 mm year–1) estimated using weight loss by Atrens et al.[36].Corrosion rate established by Chen et al.using mass loss tests for various magnesium alloys was significantly higher than the intrinsic corrosion rate: pure Mg shows the degradation rate equal to 1.8 mm year–1after immersion in MEM for 24 h[26].Deng et al.produced highly corrosion-resistant Mg using Ca micro-alloying [37].These Mg-Ca alloys show low corrosion rates ～0.1 mm year–1in 3.5 wt.% NaCl solution,which is lower than the intrinsic corrosion rate of magnesium.Some of the aforementioned studies also present the corrosion rate in various dimensions.However, for fair comparison of these values, it is better to use mm per year.

Moreover, the forecited studies are mainly aimed at studying the electrochemical properties of the alloy at the macrolevel using methods that reflect the totality of all processes occurring in the studied area(hydrogen evolution test,electrochemical impedance spectroscopy, potentiodynamic polarization,etc.).For more accurate understanding of the biodegradation processes of the Mg-0.8Ca alloy, it is necessary to study the local zones indicating the structure and composition features of the material in detail.This will make it possible to identify the regions and the sequence of formation of corrosion damage, as well as to reveal the zones of the highest electrochemical activity.

Nowadays, for the localin situstudy of biodegradation processes at the electrode/electrolyte interface, such methods as scanning ion-selective electrode technique (SIET) and scanning vibrating electrode technique (SVET) are actively used [38–45].The SVET study is carried out by measuring the potential difference between heterogeneous areas of the sample surface in solution, arising from the ion flux provoked by the occurrence of electrochemical reactions on the active surface.These potential differences are subsequently converted and final results are presenting as local current density distribution maps.SVET measurements make it possible to record the current intensity and direction in the solution.However, the concentration of specific ions involved in the corrosion process cannot be fixed.For such measurements,there is a scanning ion-selective electrode technique.SIET,which is micropotentiometry-based method, enables one torecord changes in the concentration of specific ion (for example, H+, Mg2+, Ca2+, etc.) in a solution using a special ion-selective electrode.This technology has high sensitivity and is capable of detecting the evolution of the ion concentration at the picomolar level at a distance of 10–50 μm from the studied surface at a high spatial resolution,which is set by the dimensions of the capillary of the SIET microelectrode.The ability to measure local current densities, as well as pH values at the microscale is of great importance from the standpoint of studying the effect of alloy composition and, in particular, intermetallic compounds on the corrosion behaviour of a material.This is necessary to interpret the process of alloy degradation in biological environment, where both the mechanism of corrosion destruction, and the forming reaction products are of particular importance.Moreover, the process of biodegradation of metals and their alloys can be studied most extensively by quasi-simultaneous conducting of SVET and SIET measurements.

The electrochemical behaviour of the Mg2Ca (the secondary phase in Mg-0.8Ca alloy) compound as part of a galvanic pair with pure magnesium was studied in Hanks solution using SVET/SIET [46].However, this experiment presents only a simulation of the conditions arising during the study of the intermetallic phase / metal matrix interface,and the behaviour of the material in other solutions simulating physiological conditions (for example, inorganic and organic composition of human blood plasma) was not revealed.Taking these features into account, the study of the effect of the Mg2Ca intermetallic phase on the corrosiveness of the alloy in MEM and NaCl solution by means of local methods, especially SVET/SIET, remains relevant.

1.3.Objective of current contribution

SVET/SIET are powerful techniques, which show the corrosion process development on a microscale (including study the corrosion initiation, pitting formation, microgalvanic corrosion, cut-edge corrosion, microbiologically influenced corrosion, corrosion degradation of the welded joints of metals and alloys) as compared to traditional electrochemical methods that give an average information about the sample degradation in aggressive media.Therefore SVET/SIET are key methods, which govern the novelty of this paper.

It should be noted that the use of traditional direct and alternating current electrochemical methods, such as potentiodynamic polarization and electrochemical impedance spectroscopy, in combination with local scanning electrochemical technologies makes it possible to study the effect of local heterogeneous mesoscale inclusions and morphological features on the intensity of exchange processes at the sample interface.The application of these methods enables one to obtain detailed information about the electrode surface (including formation and degradation of the protective layers,the evolution of the surface film) and establish the mechanism of electrochemical reactions and material’s corrosion.Therefore,it is very important to use traditional electrochemical methods as complementary, along with local methods, when studying corrosion processes of such complicated systems, as an implant / physiological solution.Moreover, to obtain valuable and complete results revealing the local corrosion behaviour of the material in detail the scanning Kelvin probe force microscopy(SKPFM)measurements should be provided.On the basis of measurements of Volta potential difference (which is correlated with the electrochemical properties of metals in aqueous solutions) from the sample surface at a local scale SKPFM identifies the role of intermetallics in corrosion propagation as well as determines the corrosion propensity of the different elements of the material microstructure.

The research carried out within the framework of the presented study are orientated to determine, using traditional and local scanning electrochemical techniquesin vitro, the effect of composition and heterogeneity of the bioresorbable calcium-containing magnesium alloy (Mg-0.8Ca) on the corrosiveness of the material.The nature and features of electrochemical processes occurring on the alloy surface were established in this work.

Concerning the recent progress in the development of methods of investigation of corrosion activity of electrochemical active material it is very important to show all opportunities of SVET/SIET localized methods.Study the effect of microstructure on the degradation behaviour of the materials in various corrosive solutiuons is one of the key point which should be done before development of the ways of corrosion protection of materials.The results of this study show that localized techniques can be efficient instruments, which provide this information.

The novelty of this study is based on the determination using local electrochemical methods the influence of the microstructure of Mg-0.8Ca alloy on itsin vitrocorrosion behaviour.SVET/SIET studies performed in this manuscript:1) enable one to show the effect of secondary phases on the overall corrosion degradation trend of Mg alloy; 2) are in agreement with SKPFM results.The operating modes of SVET/SIET measurements were proposed.During the research, the mechanism of the alloy bioresorptionin vitrowas established and the influence of the corrosion products formed on the surface of the material on the rate of its biodegradation was determined.

2.Experimental

2.1.Samples

Experimental rectangular samples (1.5 cm × 2 cm ×0.1 cm) were made of Mg-0.8Ca magnesium alloy (wt.%:0.83 – Ca, 0.004 – Fe, 0.002 – Cu, 0.001 – Ni, 0.02 – Al,0.05 – Mn, 0.001 – Ce, 0.005 – Zn, 0.045 – Si, 0.003 – Zr,balance–Mg).Optical emission spectroscopy(PDA-MF Plus,Shimadzu, Japan) was applied to check the elemental composition of the alloy.In order to standardize the sample surface for further measurements, all specimens were mechanically ground and polished according to the methodic described below.Then samples were washed with deionized water, degreased with ethanol and dried in air.

Specimens for SVET/SIET tests were wire cut into a cuboid with a dimension of 4 mm × 1 mm × 4 mm.To ensure the best adhesion of the polymer material to the alloy under study, the samples were pressed into an epoxy resin of increased hardness with mineral fillers EPO moulding compound (Metkon Instruments Ltd., Turkey) using a Metapress-M (Metkon Instruments Ltd., Turkey).The inner diameter of the mould for pressing was 25 mm, which is insufficient for carrying out the studies by localized scanning electrochemical methods in terms of the location of the test sample on a platform movable in the X-Y plane.For this reason, the hot-pressed specimens were additionally cold-poured by ViaFix epoxy(Struers,Denmark)into a 30 mm diameter mould.Further, the samples were mechanically processed using silicon carbide-based (SiC) grinding paper with a decrease in the abrasive grain size to 14 – 20 μm (P1000) and polished on a Tegramin-25 grinding and polishing machine (Struers A/S,Denmark).The polishing steps included processing with a fine grinding abrasive MD-Largo and polishing cloths MD-Mol and MD-Nap on magnetic disks using sequentially diamond suspensions of 9 μm, 3 μm, and 1 μm.

2.2.Testing medium

Aqueous NaCl solutions (0.3 wt.% and 0.9 wt.%) and a mammalian cell culture medium, MEM (powder No.61,100,Gibco?, Thermo Fisher Scientific, USA) were used as media for carrying out electrochemical tests.The composition of MEM is shown in Table 1.

Table 1Composition of mammalian cell culture medium, MEM.

2.3.Surface study

2.3.1.Optical microscopy and SEM-EDX analysis

The microstructure of the alloy was studied using an Axiovert 40 MAT metallographic microscope (Carl Zeiss Group,Germany).

The morphology of the calcium-containing magnesium alloy and the distribution of elements over the surface area of the sample before and after exposure of the material to aggressive environment were studied employing scanning electron microscopy (SEM, EVO 40 device, Carl Zeiss Group, Germany) and energy dispersive X-ray analysis (EDX, INCA Xact analyser, Oxford Instruments, USA).SEM was equipped with Silicon Drift Detector X-MaxN 80 (Oxford Instruments,USA).SEM-EDX measurements were performed through AZtec 3.0 SP2 software (Oxford Instruments, USA).

2.3.2.Scanning Kelvin probe force microscopy (SKPFM)

SKPFM was utilized to analyse the volta-potential distribution on the surface between theα-Mg matrix and theβ-phase of the Mg alloy.The scanning Kelvin probe force microscope(Dimension Icon, Bruker Nano Inc.) was used.Prior to the SKPFM measurements, the alloy was polished using 0.1 μm diamond paste, rinsed with deionized water and ultrasonically cleaned in ethanol.Polished samples were immediately studied using SKPFM to avoid oxidation.Measurement was conducted in open air at room temperature and a controlled rela-tive humidity of 45 – 55%.Volta potential distributions were probed in work function mode using a scanning frequency of 0.8 Hz, a pixel resolution of 512 × 512.

Tapping/lift mode sequence with an interleave lift height of 100 nm was used to conduct simultaneous acquisition of surface topography and surface potential maps.A Pt coated silicon tip with 20 nm thickness (SCM-PIT) (k = 3.0 N/m,Bruker Nano Inc.,CA,USA)was used.Results were analysed by using NanoScope Analysis 1.5 software.

2.3.3.Raman spectroscopy

Raman spectroscopy was applied to study the chemical composition of the corrosion products formed on Mg alloy in MEM.The Raman spectrometer alpha 500(WITec,Germany)equipped with 532 nm laser (20 mW power) and Zeiss EC“Epiplan” microscope with a 100 × objective was used.The spectral range was from 100 up to 1200 cm-1.Micro-Raman spectra were collected for 50 min (50 accumulated spectra).

2D map depicts the intensity distribution of components of the corrosion film was obtained using the scanning mode.Spectra were acquired from the area 40 × 34 μm, which contains 40 × 40 micro-Raman spectra.The integration time was 1 s.WITec Control program was used for analysis of acquired spectra.

2.3.4.XPS analysis

X-ray photoelectron spectroscopy (XPS) was applied for chemical analysis of the corrosion products formed on Mg-Ca alloy in MEM.The measurements were carried out using the spectral complex SPECS (Germany) with hemispherical energy analyser PHOIBOS-150 in ultra-high vacuum environment (0.5 μPa).The non-monochromatic Al Кαradiation with the energy of 1486.6 eV was used.All spectra were charge referenced using aliphatic carbon (C1s line) at the binding energy 285.0 eV.The examined surface was etched by Ar+beam (surface cleaning) during 5 min with an energy of 5000 eV in the scanning mode.The 3 nm layer was removed after the ion etching.

2.4.Electrochemical measurements

2.4.1.Localized electrochemical tests

The study of local electrochemical behaviour was carried out using the SVET/SIET system (Applicable Electronics, USA).SVET measurements were performed using a platinum-iridium alloy probe with a 10 μm diameter spherical tip coated with platinum black.Scanning took place at a distance of 100 ± 5 μm from the surface with vibration of 128 Hz and 325 Hz in the horizontal (X-axis) and vertical(Z-axis) directions respectively; the vibration amplitude was 20 μm.The research results were interpreted taking into account the vibration of Z-component directed along the vertical axis [47,48].Local pH changes were recorded using a scanning ion-selective (H+) electrode at a distance of 40 ± 5 μm from the surface.The constituent parts of the SIET electrode are a cylindrical glass capillary (outer diameter – 1.5 mm,diameter of the conical tip – 2.0 ± 0.5 μm) and silver wire coated with silver chloride.The capillaries were pulled using a P-97 Flaming/Brown Micropipette Puller (Sutter Instruments Company, USA).Then the capillaries were silanized in a special chamber at a temperature of 220 °C by adding 200 μL of N,N-dimethyltrimethylsilylamine.Further preparation of the microelectrode includes filling the tip of the capillary with an ion-selective membrane (0.5 wt.% polyvinylchloride,9.9 wt.% hydrogen ionophore I – tridodecylamine, 88.9 wt.%2-nitrophenyloctyl ether and 0.7 wt.% potassium tetrakis (4-chlorophenyl)borate)[47].The rest of the capillary space was filled with an internal solution (0.01 M KH2PO4in 0.1 M KCl).The length of the membrane column in the microelectrode was 70μm.The process of filling with a membrane and an internal solution was carried out using an optical microscope, two 3D micromanipulators and reagents produced by ?Selectophore? from Fluka (USA).A silver chloride reference electrode was placed in an internal solution in a capillary.Further calibration of the SIET electrode was carried out in solutions of the investigated media (MEM, 0.3 wt.%NaCl) with fixed pH value, according to the Nernst equation.The Nernst slope was (56.0 ± 0.5) mV pH–1(NaCl solution),(57.0 ± 0.6) mV pH–1(MEM).In addition to the installed SVET/SIET microelectrodes, the system includes an external reference electrode – a silver chloride (Ag / AgCl / 0.1 M KCl, 0.01 M KH2PO4) one.The scanning process was carried out using LV-4 software (ScienceWares, USA).The potential values were measured using a preamplifier with an input impedance of 1 PΩ.When conducting quasi-simultaneous tests, the SVET probe was placed at 50 μm along the X-axis,25 μm along the Y-axis, and 60 μm along the Z-axis from the SIET electrode to avoid possible electrolyte mixing and failure of the glass SIET electrode due to the multidirectional vibration of the SVET probe.The potential drift during the experiment was taken into account using a Sentron-SI pHmetre with a MiniFET electrode (pH-electrode, # 9202–010,SENTRON, Netherlands).The solution was renewed every 0.5 h in order to replenish the MEM components consumed in the corrosion process, to stabilize the medium conductivity equal to 2.5 mS?cm–1and to decrease the rate of microbial growth in MEM.It should be noted that the conductivity of the NaCl solution (2.6 mS?cm–1) was comparable to the value of this parameter for MEM.Before measurement, the SVET system was calibrated in each of the used media, taking into account the indicated conductivity values.To avoid uncontrolled evaporation of the medium, the SVET/SIET cell(4 ml) was in contact with the flask (500 ml) containing the test solution (NaCl, MEM).Tests were performed at room temperature [49].

To establish the electrochemical activity of the Mg-0.8Ca alloy during the exposure, and to identify the stages of material degradation in various corrosive media, an analysis of the change in the integral (total) anodic and cathodic currents was used,calculated by Eqs.(1,2)[50–52].The total currents(Ianodic, Icathodic) were calculated by integrating the values of the current density (iz),established during SVET scanning the studied area (xmax, xmin, ymaxand yminare the coordinates of this area).

To select and optimize the scanning modes, SVET/SIET tests were carried out on samples with different areas of the studied zones equal to 5 mm2and 0.005 mm2(Table 2).Area value of 0.005 mm2was achieved by isolating the sample surface with beeswax.Local corrosion characteristics were determined at given values of the number of steps of microelectrodes in the X-Y plane in certain time intervals.For a sample with 5 mm2scanning area, total scanning duration in NaCl solution achieved 38.5 h, the number of steps (within one scan) – 55 × 55.Scan duration for samples in MEM with 0.005 mm2areas was 2 h, and the number of steps was 35 × 10.Localized currents and pH values were recorded continuously for the entire time of the sample exposure to the test medium.Summarized information on scanning modes using SVET/SVET is presented in Table 2.

Table 2Input parameters of different scanning modes.

2.4.2.Conventional electrochemical tests

In addition to local electrochemical tests, traditional electrochemical measurements were carried out.Potentiodynamicpolarization (PDP) and electrochemical impedance spectroscopy(EIS)methods with recording the open circuit potential (OCP) over time were performed using VersaSTAT MC electrochemical system (Princeton Applied Research, USA).The tests were carried out at room temperature in a threeelectrode cell in MEM and 0.9 wt.% NaCl solution.The area of the studied surface was 1 cm2.The counter electrode was a platinized niobium grid, the reference electrode was a silver chloride (Ag/AgCl) electrode (the potential versus standard hydrogen electrode is equal to 0.197 V).To establish the corrosion potentialEc, the samples were held in the solution for 60 min.Potentiodynamic measurements were carried out at a sweep rate of 1 mV s–1.The sample was polarized in the anodic direction in the potential range fromEc– 0.25 V toEc+ 0.5 V.To estimate the values of the corrosion potentialEcand corrosion current densityIc, the Levenberg-Marquardt method was used following the Eq.(3).This method is known as suitable for describing the corrosion parameters of metals with a surface oxide layer, in particular, magnesium and its alloys [53,54].The frequency values during the recording of the impedance spectrum varied in the range from 1 MHz to 0.1 Hz with a logarithmic sweep of 10 points per decade.The duration of the experiment was 42 h for the sample in MEM.EIS data were recordered every 2 h of sample exposure.The spectra for the sample in NaCl are shown to compare the electrochemical parameters of the surface film formed in two media after 60 min of exposure.

2.5.Weight loss tests

The samples weight loss after 7 days immersion in MEM and 0.9 wt.% NaCl solution was measured by comparing the initial and final weight of the specimens using the sensitive analytical balance (Shimadzu AUW120D, Japan).All samples were exposed to corrosive media at room temperature with daily pH monitoring.Such temperature was chosen in accordance with the paper [49], where the influence of temperature and solution flow for DMEM on the corrosion behaviour of commercially pure Mg was investigated.Wagener et al.[49]established that temperature increase (from room temperature to 37 °C) only shows a minor influence on the corrosion behaviour of Mg.Chromic acid solution(200 g L-1CrO3+ 10 g L-1AgNO3) was used to remove the corrosion products.The degradation rate (DR, mm year-1) was calculated in accordance with previous works [55–57].Four samples (with a total surface area of about 28 cm2) were tested in both media (500 ml of solution).The ratio of the sample surface area to solution volume was 1:18 cm2mL–1.Partial solution refreshment was performed every 12 h (to reduce the bacteria growth rate in MEM).The initial values of bulk pH for 0.9 wt.% NaCl solution and MEM were 7.2 and 7.3, respectively.Tests were performed in triplicates to ensure data reproducibility.

Fig.1.Optical and SEM images of the Mg-0.8Ca alloy microstructure and the distribution of elements (Mg+Ca, Mg, Ca) over the samples surface.

3.Results and discussion

3.1.Microstructure characterization

The results of studying the sample microstructure are shown in Fig.1.Analysis of the obtained optical images indicates the presence of elongated grains formed during alloy rolling.It was established [58–61]that as a result of the integration of calcium into pure magnesium, the grain size decreases, and an intermetallic compound, the Mg2Ca phase, is formed along the grain boundaries.This secondary phase is distributed not only along the grain boundaries, but also concentrates within the magnesium matrix as small inclusions.The analysis of obtained SEM images and EDX-maps of the elements’ distribution over the sample surface confirmed thepresence of calcium at grain boundaries and in the composition of inclusions (Fig.1).The lower concentration of magnesium in the corresponding areas (darker areas in Fig.1),in comparison with theα-magnesium matrix, confirms the presence of calcium in the intermetallic phase.

Fig.2.Optical images of the sample with the investigated area of 5 mm2 before and after SVET/SIET tests and the corresponding maps of local current density distribution after 1.5 h, 6 h, 13.5 h, 16.5 h, 20 h, 24.5 h, 29.5 h and 38.5 h of exposure to 0.3 wt.% NaCl solution.Evolution of total anodic and cathodic currents established using SVET during the samples immersion in 0.3 wt.% NaCl solution for 38.5 h (plot).

3.2.Local electrochemistry analysis

3.2.1.Mg alloy degradation in NaCl solution

Fig.2 presents SVET diagrams obtained by scanning the sample with a surface area of 5 mm2in 0.3 wt.% NaCl solution for 38.5 h.0.3 wt.% NaCl solution was used since it provides the accurate value of the local current density due to the high signal/noise ratio.Anodic-cathodic zones are identified on SVET maps, with the most distinct localization observed in the first hour of the sample exposure to a corrosive solution.After 1.5 h of exposure, a high electrochemical activity of the material is observed, as evidenced by high peak local current density values in the anodic (zones with more positive current density, shown in red) and cathodic zones (zones with more negative current density, shown in blue).The change in current during first immersion can be due to the dissolution of the film formed during specimen preparation and its replacement by a film characteristic of the solution.With an increase in the exposure time, the maximum difference in the current density, recorded in the anodic and cathodic regions of the alloy (Δimax), starts to decrease gradually: 140 μA·cm–2(1.5 h), 63 μA·cm–2(6 h) and stabilizes at a value of 45 μA·cm–2(10–38.5 h).High values of the anodic and cathodic curent density indicate high corrosivity of the sample in 0.3 wt.% NaCl solution.Evolution of total anodic and cathodic currents, established using the SVET method, confirms the tendency to a decrease in the rate of corrosion processes occurring on the surface of the material during the immersion of the sample in 0.3 wt.% NaCl solution for 38.5 h (plot in Fig.2).An obtained result, along with the decrease of colour intensity and a narrowing the area of heterogeneous electrochemically active regions, indicates gradual and partial passivation of the material due to the formation of a Mg(OH)2surface film, according to the reactions (4–6).Nevertheless, the protective properties of magnesium hydroxide film are quite low and not sufficient to protect the material from degradation in the aggressive environment, therefore, the Mg-0.8Ca sample continues to degrade over time, as evidenced by the high value of the local current densities at the end of the experiment (Fig.2).Moreover, the location of the active spots on SVET maps is constant indicating the low passivation rate of the alloy in NaCl solution.

Analysis of the change in the state of the sample surface during exposure (optical images in Fig.2) confirms the intensive degradation of the material.Besides, the regions with the highest number of defects and pittings are in good agreement with high anodic activity sites on SVET maps.

The evolution of the local pH distribution(maps)and maximum values of local pH (plot) over time is shown in Fig.3.Using SIET an alkaline region was recorded over the entire surface of the studied sample (the pH value reaches 11),due to the intensive course of cathodic reactions of water reduction with hydrogen evolution (5) and reduction of oxygen dissolved in the electrolyte (7).These local pH values are compatible with the low solubility of Mg(OH)2equal toKs= 1.2?10–11[62].

In spite of the large area of the sample and the long duration of one complete scan, SVET established a non-uniform colour distribution in the anodic zones (yellow-red areas), associated with the probable influence of the structural heterogeneity of the material, as well as the presence of an anodic phase around the cathodic region in the right part of the studied zone (Fig.2).These data indicate the influence of the structural features of the Mg-0.8Ca alloy on its corrosiveness.However, SIET did not reveal the clear boundaries of the electrochemically active areas of the alloy as a result of the intensive medium alkalization.For clearer determination of the intensity of electrochemical processes associated with the material structure and the presence of secondary phases,it is necessary to reduce the investigated area and scanning steps of SVET/SIET microelectrodes.Application of a lowcorrosive medium can also increase the degree of localization since the specimen will be corroded at a low rate.

Fig.3.Optical images of the sample with the investigated area of 5 mm2 before and after SVET/SIET tests and the corresponding maps of local pH distribution after 1.5 h, 6 h, 13.5 h, 16.5 h, 20 h, 24.5 h, 29.5 h and 38.5 h of exposure to 0.3 wt.% NaCl solution.Evolution of maximum values of local pH established using SIET during the samples immersion in 0.3 wt.% NaCl solution for 38.5 h (plot).

Fig.4.SEM-EDX images of the sample surface after 22 h of exposure to 0.3 wt.% NaCl solution.

The results of SEM-EDX analysis of the film formed on the surface of the material after 22 h of the immersion in 0.3 wt.% NaCl solution indicate a high concentration of magnesium and oxygen in the composition of corrosion products Fig.4).During the exposure of the material to an aggressive environment, a layer of Mg(OH)2is formed on its surface according to Eqs.(4–6).Further immersion results in cracks formation in the corrosion products layer.Insignificant inclusions of calcium are associated with its partial dissolution from the alloy and incorporation into the composition of the surface film.In accordance with [37]due to the electrolyte carbonation as a result of CO2dissolution from ambient air– calcium carbonate may be one of the components involved in the surface film formation during Mg-Ca alloy corrosion in NaCl medium.The presence of calcium on the EDX map can also be the result of electron beam penetration through the layer of corrosion products (with the following X-ray emission through this layer).Fig.4 presented for data comparison with the chemical composition of corrosion film formed on the alloy in MEM.

3.2.2.Mg alloy degradation in MEM

To establish the mechanism of corrosion processes on implant materials, it is necessary to perform the experiment in a solution with ionic composition close to the environment of the human body, in particular, in the medium for cultivation of mammalian cells (minimum essential medium).Such an approach will facilitate the search for optimal ways in the development of coatings that protect the material from intensive degradation caused by specific processes in the human body.In this work,in vitrostudies of the local electrochemical activity of a magnesium alloy, promising for application in implant surgery, were carried out using MEM.

To establish the peculiarities of the material degradation process at the early stage of its exposure, the scanning area was significantly limited by beeswax(S=0.005 mm2,Figs.5,6).This method provides to minimize the scanning time, reduce the step of microelectrodes and increase the SVET/SIET measurement accuracy.For an accurate recording the influence of secondary phases on the corrosion process of the alloy, the movement of the microelectrodes was carried out with a step of 3–5 μm, while the time of one complete scanover the entire study area was 12 min, which made it possible to record the evolution of the electrochemical activity of the material at the micro-level in more detail.Hereinafter,the scan area is marked with a frame.The difference in the position of the SVET and SIET maps on the optical image of studied area (Figs.5, 6) corresponds to the distance between the microelectrodes indicated above.

Fig.5.Optical images of the sample before and after 2 h exposure to MEM.Evolution of the local current density distribution over the sample surface with the area of 0.005 mm2 after 12 min, 48 min, 84 min, 120 min of exposure to MEM (maps).Evolution of total anodic and cathodic currents established using SVET during the samples immersion in MEM for 2 h(plot).

The intensification of the material degradation process is observed during the first 12 min of exposure, as evidenced by the high value ofΔimax= 40 μA?cm–2, decreasing to 3 μA?cm–2over time (Fig.5), according to the SVET data.The formation and movement of the anodic-cathodic regions over the surface was recorded.The analysis of evolution of the integral anodic and cathodic currents indicates rapid passivation of the material during the first hour of immersion in MEM (plot in Fig.5).The level of total currents indicates a significantly lower corrosive activity of the material in MEM, in comparison with the degradation of the alloy in 0.3 wt.% NaCl solution (Fig.2).Inspite of higher concentration of sodium chloride in minimum essential medium as compared to NaCl solution (0.9 wt.% vs.0.3 wt.%), the presence of various compounds in MEM (mainly inorganic ones)inhibits the corrosion degradation of the alloy.

An obtained result agrees with the data of [63], where higher protective properties of the surface film formed on the MA8 alloy in MEM in comparison with 0.83% NaCl solution were shown.However higher level of current that was established in [47]using SVET for the Mg sample in MEM as compared to NaCl solution was not registered in the present work.This can be the result of the effect of solution refreshing every 0.5 h,which does not lead to microbial growth(as compared to electrochemical tests, where only a peristaltic pump was used for MEM renewal and the bacterial contamination of MEM was observed[47]).Therefore the extra currents related to bacterial metabolism and associated extracellular electron transfer were not detected by SVET in this work.

SIET established a rapid change in the corrosion pattern during the sample exposure for 2 h Fig.6).The formation of microgalvanic pairs and anodic zones on the surface of the investigated area was observed.Besides,the activity of the anodic and cathodic regions changed sharply during scanning,as indicated by the rapid relocation of these zones during the exposure of the material to MEM.It was established that local pH values change in a narrow lightly alkaline range from 7.4 to 7.5, which confirms the low intensity of alloy degradation revealed using SVET.The exception is the result of the first scan of the sample, according to which the pH varied from 6.3 to 9.0 over the sample’s surface, which agrees with the maximum electrochemical activity of the material in the first 12 min of exposure established by SVET.Lower values of the local pH in the investigated area of the alloy, in comparison with the value of this parameter for the sample in NaCl solution are explained due to the formation of a hydroxyapatite(HA) layer on the surface of the material in MEM according to reaction (8).The OH–ions formed during the corrosion of magnesium (reaction (5) are consumpted to the formation of Ca10(PO4)6(OH)2and therefore the traditionally high pH values observed during the corrosion of magnesium and its alloys in sodium chloride solution are not achieved.Moreover,taking into account that HA is a poorly soluble compound(Ks= 1.6 × 10–58), the corrosion of the magnesium alloy is slowed down as a result of its inclusion into the formed surface layer, which confirms the low values of the local current density during the immersion of the sample.

SIET maps contain regions with a clearly delineated local minimum of pH, which is the result of the probable intensive formation of HA.Due to pitting corrosion of the alloy (anodic phase dissolution) the released ions participate in OH–binding according to reaction (8) that significantly decreases the local pH.Closely located zones with different pH (anodic and cathodic regions) in Fig.6 are the microgalvanic pairs,caused by the influence of the alloy structure.Therefore, formation of electrochemically active areas on the investigated surface, which occurred during the first minutes of exposure of the material to MEM, was revealed and studied by localized methods.

3D images of the local pH evolution, shown in Fig.6,clearly demonstrate a sharp decrease in the range of its values

and, accordingly, the intensity of corrosion processes caused by the passivation of the material surface due to the formation of a corrosion products layer.The low corrosiveness of the magnesium alloy, recorded in MEM by SVET/SIET, indicates that after 2 h the protective surface film has already formed.

Fig.6.Optical images of the sample before (a) and after (b) 2 h exposure to MEM.Evolution of the local pH distribution over the sample surface (in 2D and 3D maps) with the area of 0.005 mm2 after 12 min (c, d), 48 min (e, f), 84 min (g, h), 120 min (i, j) of exposure to MEM.

It should be noted that zones of electrochemical activity –pittings, recorded using SIET (Fig.6c, e, g, i), are in good agreement with regions formed on the material surface during immersion and shown in optical image at the end of the experiment (Fig.6b).This confirms the correlation between the electrochemical activity of the alloy and its composition and microstructure features (effect of the secondary phases).Moreover, the analysis of experimental data indicates a correspondence of the anodic areas, which were established by SIET, with calcium-containing phases of magnesium alloy which were found using optical microscopy and SEM-EDX analysis (Fig.1).As mentioned above, as a result of the integration of calcium into magnesium, an intermetallic compound, the Mg2Ca phase, forms along the grain boundaries and in the Mg matrix, which contributes to an increase in strength and yield stress at high temperatures.However, the presence of the Mg2Ca intermetallic compound also has a significant effect on the electrochemical behaviour of the alloy.The obtained result shows high opportunity of localized electrochemical methods in the investigation of corrosion behaviour of the material on microscale.These techniques enable us to identify that corrosion occurs preferentially in the sites of secondary phase and Mg2Ca dissolves first.

3.3.Analysis of localized Volta potential distribution

Nowadays there are works dealing with investigation of the corrosion behaviour of Mg2Ca phase [64–69].Some authors considered that Mg2Ca is a cathodic phase paired with magnesium[21,67,70,71].Opposite works using electrochemical analysis proved a more negative corrosion potential of this phase in comparison with theα-Mg matrix [64–66,72–77].Therefore, these conflicting results should be cheked.Mohedano et al.[67],who showed the small cathodic effect of the Mg2Ca, indicated that controversial results regarding the electrochemical activity of Mg2Ca can be explained by the presence of impurities and matrix segregation.To confirm the influence of these parameters a systematic SKPFM/SEM/EDX study is necessary to perform.SKPFM is a powerful tool to study the weak force of atoms on the surface of the tested alloy, which reflects the surface potential of the sample in correlation with its surface morphology [64–66,78–87].

The localized electrochemical techniques used in this work have already showed the anodic behaviour of the secondary phase of magnesium-calcium alloy.Nevertheless, further confirmation of this result should be made to show the agreement of the SVET/SIET and SKPFM data with respect to Mg-Ca alloy studied in this work (the level of alloy impurities was the same for studied specimens).

The local nobility/activity of the different constituents in the alloy microstructure can be determined using SKPFM.The following analysis of the Volta potential distribution provides to predict the potential cathodes and anodes that can be formed during specimen exposure to aggressive medium.Therefore, SKPFM measurements were carried out to further support the above SVET/SIET results regarding the anodic behaviour of the secondary phases(with respect toα-Mg matrix)of the alloy, which leads to preferential corrosion initiation of the Mg2Ca phase.

Fig.7 shows the surface topography (a) and potential distribution (b) maps from the region of the Mg-0.8Ca alloy(55 × 65 μm) that contains bothα-Mg grain and Mg2Ca phases.The analysis of the topography map shows height difference between the grain and secondary phase (this difference in peak heights is 300 nm).Such a height difference can result from the hardness differences of these constituents.β-Mg2Ca is softer compared toα-Mg matrix.Hence,the hardα-Mg matrix is less removed by grinding during the polishing process.Following topography measurement,the potential distribution is mapped over the same area.

The thermodynamic tendency of the alloy and the role of the phases in an electrochemical reaction can be determined using such electron sensitive parameter as Volta potential.The obtained map indicates that secondary phase formed in the alloy has different Volta potentials from the matrix.Fig.7b shows that secondary Mg2Ca phases have a darker colour than the surrounding eutectic mixture ofα-Mg matrix (the light region).When operating in the work function mode of SKPFM, the bright areas indicate zones with more positive potential, while the dark ones have relatively more negative Volta potential.The obtained result shows that the second phase Mg2Ca is more electrochemically active thanα-Mg matrix and possesses lower nobility in course of microgalvanic corrosion.

The corresponding Volta potential line profile (marked AB in Fig.7b) is shown in Fig.7c.The average potential of the Mg2Ca changes from about –40 to –85 mV which is lower than that ofα-Mg.Based on Volta potential analysis the Mg2Ca andα-Mg matrix form a strong galvanic coupling,where the second phase acts as micro-anode and preferentially corrode in aggressive environment.

SKPFM results indicate that secondary phase is anodic to theα-matrix and creates a localized microgalvanic cell that facilitates the dissolution of Mg2Ca at grain boundaries(and withinα-Mg grains) during the corrosion process until these inclusions are separated from the matrix.This results in faster degradation, higher susceptibility to pitting corrosion and, therefore, increased corrosion of the Mg-0.8Ca alloy compared to magnesium.Therefore, SVET/SIET results were verified using SKPFM analysis.The anodic effect of Mg2Ca precipitations on the corrosion behaviour of the alloy was confirmed.

3.4.Chemical analysis of the corrosion film formed in MEM

3.4.1.SEM-EDX data

SEM-EDX was used to analyse the corrosion products formed on the magnesium alloy in MEM after 24 h of im-mersion of the area, which was previously studied using SVET/SIET (Figs.5, 6).The results of morphology and elemental composition evaluation are shown in Fig.8.The obtained data indicate the formation of a Ca-P layer over the oxide-hydroxide film on the surface of the magnesium alloy.As indicated above, the formed Ca-P compounds include magnesium-and-carbonate substituted hydroxyapatite according to the generalized reaction (9) [88].

Fig.7.Surface topography map (a) and Volta potential distribution map (b) for Mg-0.8Ca alloy.Volta potential line profile (marked AB in b) drawn through a Mg2Ca phase and α-Mg matrix (c).

The results of SEM-EDX analysis show the formation of the corrosion products on the areas of high electrochemical activity established using SVET/SIET.The formation of calcium phosphate compounds is observed mainly on the surface of the anodic zones (round areas in the SIET diagram in Fig.6 and SEM-EDX images), due to their more intense dissolution and participation in the supply of Mg2+and Ca2+ions (Fig.8).The obtained results show a good agreement of SEM-EDX and SVET/SIET data.Moreover, the elemental analysis additionally confirmed that anodic areas are the Mg2Ca phases, which were observed using SIET.The small discrepancy between the optical (Fig.6b) and SEM images(Fig.8) is the result of different immersion times of the sample in MEM, which were 2 h and 24 h, respectively.

3.4.2.XPS data

Fig.9a represents the XPS spectra for the Mg alloy sample with corrosion film formed after 30 days of immersion in MEM before and after the Ar+beam etching.Analysis of XPS data indicates that upper layer before etching contains high concentration of C (42.7 at.%, Table 3) in aliphatic state.To establish the chemical state of elements in the film composition the samples were Ar+etched and the cleaned surface was analysed.After the 5 min etching carbon concentration decreased (down to 15.7 at.%, Table 3).Detailed analysis of C 1s high resolution spectrum shows that about half amount of carbon is associated with oxidized state (Fig.9b): CO32–,O–C–O, C=O.This indicates that the presence of C is associated not only with surface contamination, but also with the real composition of the formed film.

Table 3Binding energy (eV) and elemental composition (at.%, in parentheses) of the surface film formed after 30 days immersion in MEM.

Table 4The calculated parameters of the equivalent electrical circuits (EEC) obtained by fitting the impedance spectra during exposure to MEM and 0.9 wt.% NaCl.

Phosphorus observed in the corrosion layer is obviously included in the composition of calcium and magnesium phosphates, as indicated by the characteristic binding energy of P 2p electrons (Fig.9c).In the surface film, the phosphorus content is almost uniform (ca.12 at.%).

Analysis of Mg 2p, Ca 2p high resolution spectra (Fig.9d,e) also shows that Mg and Ca present in the corrosion film as phosphates: Mg3(PO4)2and Ca3(PO4)2as well as in oxide state: MgO, CaO, which follows from a comparison of the total amount of Ca (13.1 at.%) and Mg (9.5 at.%) with P amount.It is possible that a part of the observed magnesium is in the metallic form (substrate response), as indicated by the spectrum of Mg Auger electrons (Fig.9f).O 1s spectra (Fig.9g) depicts the presence of carbon and metals (Mg and Ca) in an oxidized state at 532 and 531 eV, respectively(MeOx, Me – stands for metal).The component at 532 eV also indicates that layer of corrosion products contains metal phosphates Mex(PO4)y.

Fig.8.SEM-image of the samples surface structure after 24 h of exposure to MEM and EDX maps of the elements distribution over the sample surface.

Presence of nitrogen(3.3 at.%)in the surface layer is associated with organic substances adsorbed on the sample surface during immersion in MEM.Before etching N concentration was higher (4.1 at.%), which indicates the sorption of the organic compounds as the top-most layer.

Presence of carbon, phosphorous, magnesium and calcium in the aforementioned state also confirms the probable formation of Mg2+-and-CO32––substituted hydroxyapatite.However, further analysis of the surface film should be done.

3.4.3.Raman spectroscopy data

To confirm the presence of hydroxyapatite in the composition of corrosion film formed on Mg alloy after 30 days of immersion in MEM and to show the intensity of apatitephase distribution on the material surface the Raman spectroscopy was used.Analysis of the optical image indicates the complex morphology of the corrosion film with many cracks in its structure formed as a result of corrosion degradation during continuous immersion in aggressive medium(Fig.10).

Fig.9.XPS spectra of a Mg-0.8Ca sample after the exposure to MEM for 1 month (a) before and after Ar+ beam etching for 5 min, and C 1s (b), P 2p (c),Mg 2p (d), Ca 2p (e), Mg-auger (f), O 1s (g) high resolution spectra.

To obtain 2D intensity map the micro-Raman spectra were collected from the surface of corrosion film (framed area in Fig.10a).The acquired spectra were limited in the range(from 950 cm-1up to 970 cm-1) corresponding the band which is the most representative for HA-containing samples[89,90].

Fig.10.The optical image of the sample’s studied area (a), and 2D map(obtained using the scanning mode of micro-Raman spectroscopy) of HA intensity distribution in the composition of the surface film (b).Points 1 and 2 indicate places on the surface, where micro-Raman spectra (c) were collected.

Analysis of plotted 2D intensity map (Fig.10b) shows areas of high HA concentration in the composition of film.The heterogeneous distribution of HA corresponds to the artefacts of the surface film formed as a result of its degradation.The highest content of apatite phase is found in the centre of the studied zone (the compact and defect-free part of the film,point 1), whereas the lowest concentration is observed in the cracks of the film (point 2).

Two micro-Raman spectra collected in point 1 and point 2 are presented in Fig.10c.The low intensity spectrum presented only for comparison.The high intensity spectrum acquired from point 1 shows bands related to vibrations of the phosphate group (PO43-) at about 432 cm-1and 583 cm-1(corresponding tov2andv4bending modes of the O-P-O bond).The peak at 959 cm-1is attributed to theν1nondegenerate symmetrical stretching mode of HA [90–93].The band detected around 1061 cm-1represents v3a triply degenerate asymmetric stretching mode of phosphate group (of the P–O bond).[94,95].

The presence of peak ～895 cm-1confirms the presence of P-O-H vibration in hydrogen phosphate phase (HPO42-).The peak observed around 809 cm-1is related to the brushite phase Mg(OH)2[96].

In the Raman spectra, MgO peak at 994 cm-1can be observed [97–99].Two peaks at about 349 and 480 cm-1can be observed due to the formation of Mg(OH)2[100–103].

Raman spectra demonstrate the existence of symmetric stretching mode of the carbonate group, which is characterized by peaks at 1117, 726, 696 and 400 cm-1[104].This might be attributed to the magnesium substituted hydroxyapatite and magnesium carbonate as the corrosion product on the Mg alloy [24,91,98,105,106].

Analysis of these results indicates the formation of magnesium-and-carbonate substituted hydroxyapatite and confirms XPS and SEM-EDX data.

3.5.Analysis of corrosion performance using conventional electrochemical techniques

In order to study electrochemical parameters evolution during the immersion of a magnesium alloy sample, the traditional electrochemical methods OCP,EIS and PDP were used.

The change in the electrode potential during 42 h of the material exposure to MEM is shown in Fig.11a.During 20 h of sample immersion, an increase in potential values is observed from–1.720 V to–1.475 V due to formation of a Ca-P film.After 30 h of the alloy exposure, there is a slight decrease in potential associated with partial degradation of the surface layer.

Evolution of impedance spectra,presented in 3D images of Nyquist (dependence of the imaginary impedance component on the real one, Fig.12a) and Bode plots (dependence of the impedance modulus,|Z|, and phase angle,θ, on frequency,f,Fig.12b and 12c, respectively), clearly reflects the change in the state of the Mg-0.8Ca alloy surface during exposure to MEM.

These data are in good agreement with OCP variation over time and indicate an increase in the protective properties of the film due to the formation of the Ca-P layer.During the first 30 h of the exposure, there is an increase in the diameter of the half-cycle on the complex plane (Fig.12), as well as the value of the impedance modulus measured at a low frequency (|Z|f→0Hzin Fig.11a).The further increase in time of the immersion in MEM resulted in a decrease of these parameters, caused by the degradation of the corrosion products film, due to the lack of the medium replenishment with such ions as Ca2+, HPO42–, HCO3–, which are consumed during the formation of a layer of hydroxyapatite-like products [63].

In Fig.12 and further, impedance spectra contain experimental values and theoretically calculated curves, which are represented by symbols and solid lines,respectively.Fitting of the spectra for the sample in the MEM was carried out using the equivalent electrical circuit (EEC) with a series-parallel connection of twoR–CPEcircuits (insertion in Fig.12a), taking into account the previously established two-layer structure of the surface film [63].This circuit includes the electrolyte resistance,RS, elementsR1–CPE1, which are the resistive and capacitive components of the first time constant, responsible for the outer layer of corrosion products,andR2–CPE2,which are associated with second time constant, describing the innerlayer.TheCPE(constant phase element) was used instead of the ideal capacitance due to the heterogeneity of the investigated surface.TheCPEimpedance was established using the Eq.(10), in whichωis the radial frequency (ω=2π?),jis the imaginary unit,nis the exponential factor (–1 ≤n≤1),andQis theCPEcoefficient [107–109].EIS spectra were described with a high degree of accuracy (χ2= 1 × 10–4).

Fig.11.Evolution of the OCP values and the impedance modulus measured at a low frequency |Z|f→0 Hz during 42 h of the sample immersion in MEM (a).Dynamic of the change in calculated parameters (presented in Table 4) during the exposure of the specimen to MEM (b, c).

Fig.12.Evolution of impedance spectra, presented in the form of 3D images of Nyquist plots – dependences of the imaginary impedance component on the real one (a) and Bode – dependences of the impedance modulus, |Z|, (b) and phase angle, θ, (c) on frequency during exposure to MEM for 42 h.Equivalent electrical circuit used to fit the impedance spectra obtained during the study of the samples in MEM (insertion in Fig.12a).

The change in the fit parameters is shown in Table 4 and Fig.11b,c.Analysis of the obtained data confirms the growth of the outer and inner layers of corrosion products during 30 h of the immersion, as evidenced by decrease in theCPEcoefficient (Q1andQ2) and an increase in the resistance of the surface film (R1andR2).Fluctuations in the calculated parameters with a trend to reduce in resistance and an increase in the capacitive component ofCPEare observed after 30 h of exposure of the sample to MEM, due to partial destruction of the corrosion layer.These calculated data(Fig.11b,c)confirmed the corrosion trend presented in Fig 11a.The results of this work are in agreement with the data obtained in [63],in which similar processes of a surface film formation on the MA8 alloy were observedin vitro.However, the moment of the corrosion layer degradation on MA8 alloy occurred after 50 h of the sample immersion, which is due to the higher protective properties of the Mg-Mn-Ce alloy, in comparison with the Mg-Ca alloy studied in the present work.

Table 5The main electrochemical parameters obtained by analysing the polarization curves and impedance spectra after 1 h of exposure to MEM and 0.9 wt.%NaCl solution.

During the specimen immersion in a corrosive medium,there is a noticeable increase in the amplitude of the first time constant of the phase angle and a corresponding increase in the impedance modulus in this region on the Bode plot(Fig.12b, c), which determine the evolution of the outer Ca-P layer.

Evolution of the first time constant is not so expressed,which is the result of low rate of the formation and degradation of the Mg(OH)2– the inner layer of corrosion products(layer with an increased concentration of Mg and O in Fig.8).This layer is actively formed only at the initial stage of sample immersion, as evidenced by high pH values (Fig.6c, d)in the first 12 min of exposure of the sample to MEM.In spite of the Mg(OH)2layer also grows over time as a result of the gradual degradation of the Ca-P layer and the penetration of aggressive components of the medium to the material substrate, this process is not intensive and cannot sufficiently change the corrosion behaviour of the sample.

To compare the protective parameters of the film formed in MEM and 0.9 wt.% NaCl solution, PDP tests were carried out and the impedance spectra recorded in both media were studied.The recorded polarization curves shown in Fig.13 indicate a lower corrosion current density for the sample in MEM (9.5 μA?cm–2) in comparison with the alloy tested in 0.9 wt.% NaCl solution (110 μA?cm–2).A decrease in the corrosion potential for a sample in MEM is due to a more complex composition of this medium (as compared to trivial NaCl solution) and, as a consequence, a change in the potential-determining reaction that occurs during the degradation of the sample with an increase in the thickness of the surface layer.

Fig.13.Polarization curves obtained after 1 h of exposure of the Mg-0.8Ca sample to MEM and 0.9 wt.% NaCl solution.

The results of impedance spectroscopy (Fig.14) confirm the higher protective properties and lower corrosion rate of the sample in MEM, in comparison with these parameters for the sample in 0.9 wt.% sodium chloride solution.It was recorded that impedance modulus at a low frequency was 45 folds higher during the sample exposure to MEM as compared to exposure to NaCl solution.The summarized results of the analysis of PDP and EIS experiments are presented in Table 5.For the spectrum of the sample in NaCl, two time constants are observed, which is due to the formation of a Mg(OH)2film as a result of reactions (4–6) (high frequency range) and charge transfer Faraday processes(low frequency range).This spectrum was described using the EEC shown in insertion in Fig.14b, containingRS,R1–CPE1andRL–L,whereRLandLare the inductive reactance and inductance, related to the impedance modulus decrease at a low frequency, caused by intense dissolution of the alloy, according to reaction (4).The calculated parameters of the EEC elements are presented in Table 4.The total resistance of the surface film formed as a result of the sample immersion in MEM for 1 h (R1+R2,9.01 kΩ?cm2) was 41.5 times higher than the value of this parameter for the corrosion products layer on the sample in NaCl (0.22 kΩ?cm2).

3.6.Corrosion degradation rate (DR)

It is well known that corrosion rates of Mg alloys measured using electrochemistry have often not agreed with those measured by weight loss tests [18,62].However, analysis of the weight loss enables one to calculate the accurate value of the steady-state corrosion degradation rate of magnesium and its alloys [36].Fig.15 shows the results of magnesium alloy corrosion after 7 days immersion in 0.9 wt.% NaCl solution and MEM.TheDRof the alloy in 0.9 wt.% NaCl solution are presented for comparison.BothDRvalues were higher than the intrinsic corrosion rate of high-purity magnesium in concnetrated chloride solution (0.3 mm year–1).In MEM, theDRof Mg-0.8Ca alloy is two times lower than that in NaCl solution (1.44 ± 0.15 mm year–1and 2.83 ± 0.49 mm year–1,respectively).These results are in agreement with those from the conventional electrochemical tests, which show the higher protective properties of the corrosion film formed on Mg alloy in MEM as compared to that formed in sodium chloride solution.Corrosion rate calculated from PDP curves is 2.5 mm year–1, and 0.2 mm year–1, for Mg alloy specimens exposed to NaCl solution and MEM, respectively.This Ca-P film significantly slower corrosion rate of magnesium, which is confirmed by optical images of the samples after 7 days exposure to agressive media (low ammount of corrosion products after immersion and more smooth surface after etching for specimens in MEM as compared to ones in NaCl solution).During the mass loss tests the average value of the bulk pH for the samples in MEM and 0.9 wt.% NaCl solution were 7.6 ± 0.4 and 10.8 ± 0.3, respectively.pH values of the sample in MEM practically did not change due to partial solution refreshing and its buffer properties.However, the medium renewal does not have such influence on the sample exposed to NaCl solution.pH rapidly increases from neutral value due to more intensive corrosion degradation of the material in NaCl solution as compared to the sample in MEM.

As mentioned above the corrosion rate determined using PDP measurements very often low correlated with long-term weight loss tests [88].The difference inDRvalues roughly estimated from PDP curves and mass loss test can be related to the contribution of oxygen reduction reaction to the total cathodic process [110,111]and the influence of negative difference effect [112,113].It should be noted thatDRmeasured using PDP is an instant value, which contains information about material behaviour at a certain time and doesn’t reflect the data about corrosion rate evolution [114].At the same time, the long-term weight loss test provides the steady-state corrosion rate values estimated on the basis of days of exposure.Atrens et al.[36]indicated that PDP givesDRfor Mg alloys less than the corrosion rates estimated using mass loss or hydrogen evolution since electrochemical measurements are usually carried out soon after specimen immersion in the solution, before there is steady-state corrosion behaviour.TheDRsoon after sample immersion can be orders of magnitude smaller than the steady-state corrosion rate [115,116].This effect was also established in the current work.It should be noted that part of the investigated sample surface can be

isolated by evolved hydrogen (as a result of cathodic partial reaction), which decreases the corrosion rate measured by electrochemical methods [54,57,117].

Fig.14.Impedance spectra presented in the form of Nyquist (a) and Bode (b) plots, obtained after 1 h of keeping the samples in MEM and 0.9 wt.% NaCl.EEC used to fit the impedance spectrum obtained during the study of the samples in 0.9 wt.% NaCl solution (insertion in Fig.14b).

Fig.15.Optical images of Mg alloy samples before 7 days immersion in 0.9 wt.% NaCl solution and MEM, after extraction, after etching.Weight loss and corrosion rate of Mg-0.8Ca alloy after 7 days immersion in either 0.9 wt.% NaCl or MEM.

In this work, the disparity ofDRvalues estimated using weight loss and PDP tests for specimens in MEM and correlation of these values for samples immersed in NaCl solution can be due to different trends of corrosion rate evolution and different real corrosion rates of the material in these media at various times.For the sample in MEM, theDRcan sufficiently change during immersion probably due to the formation and degradation of the hydroxyapatite-containing protective film.Therefore, the initialDRvalue, which was determined using the PDP test is a result of the real low corrosion of the alloy (as a result of the protective film formation)at that period of time.At the same time, higher values of steady-state corrosion rate measured after 7 days of exposure to MEM using mass loss can be due to an increase of the material’s corrosion degradation (breakdown of the protective film) during the experiment time [63].Whereas theDRcorrelation estimated according to two different measurements for samples in NaCl solution is related to quasi-constant dissolution rate of the Mg (as a result of the barely protective magnesium hydroxide film) in this medium.Therefore, the initial instant rate (according to PDP) and steady-stateDR(according to mass loss) were similar to each other for specimens in NaCl solution.

Since the level of impurities can sufficiently affect the degradation behaviour of the alloy in aggressive media it is necessary to compare the obtained results with the data of similar studies of the corrosion rate of pure Mg in MEM.The analysis of publications shows contradictory information.Mei et al.[32]and Chen et al.[26]studied the corrosion rate of commercially pure Mg in MEM after 24 h of exposure using hydrogen evolution and mass loss tests, respectively, and showed theDRvalue equal to ca.1.8 mm year–1, which is similar to the result of current work.Kirkland et al.[118]used weight loss tests for a range of Mg alloys during specimen exposure to MEM and found the degradation rate of highly purity Mg equal to ca.2 mm year–1(after 7 days immersion).However, Walker et al.[119]and Hou et al.[120]showed a lower corrosion rate (estimated using mass loss test) of pure Mg after 7 days of immersion in MEM equal to 0.73 and 0.57 mm year–1, respectively.Myrissa et al.[121]measured the degradation rate using the weight loss analysis of pure Mg after 1 week of immersion in DMEM (Dulbecco’s modified eagle’s medium), which was 0.75 ± 0.45 mm year–1.Walker et al.[119]also showed a lower degradation rate of Mg-0.8Ca alloy after 7 days of immersion in MEM equal to 0.94 mm year–1as compared to the result of this work.The above-mentioned disparity inDRvalues can be due to different qualities of used Mg and its alloys as well as different content of impurities and experimental conditions.

3.7.Corrosion mechanism

Fig.16.Model of the mechanism of Mg-0.8Ca alloy corrosion degradation process during the exposure of the samples to MEM (a) and 0.9 wt.% NaCl solution (b).

Based on the obtained experimental data, the mechanism of the Mg-0.8Ca alloy corrosion degradation in two different media is suggested.During the sample immersion in MEM,the Mg(OH)2layer grows according to reaction (6) and Ca-P compounds, including magnesium-and-carbonate substituted hydroxyapatite, formed due to the synergistic interaction of Ca2+, Mg2+, H2PO4–/ HPO42–, CO32–/ HCO3–ions, according to the reaction (9) (Fig.16a).It should be noted that Ca2+ions involved in the formation of HA are the components of MEM and are also formed during the corrosion of the Mg2Ca anodic phase, which is responsible for the alloy degradation process in accordance with SVET/SIET/SKPFM measurements.After 30 h of the exposure, partial destruction of the Ca–P layer occurs and the MgO / Mg(OH)2film begins to form intensively as an inner sublayer, which is consistent with the results of [63].In case of the alloy corrosion in 0.9 wt.% NaCl solution (Fig.16b), the surface layer of corrosion products includes MgO / Mg(OH)2according to EDX analysis data.In Fig.16 anodic phase is presented as a combination ofα-Mg+Mg2Ca.It is in good agreement with the results obtained in [24]where Mg2Ca compound is characterized by inclusions of a size of 1.8 ± 0.6 μm, located throughout eutectic microstructure in Mg matrix and at grain boundaries.The oxide/hydroxide layer formed on the alloy surface (Fig.16b) is not dense and begins to crack.Therefore, the corrosion process in NaCl, has a more active and prolonged effect than in MEM, for which the higher protective properties of the sample are explained by the formation of Ca-P surface film of hydroxyapatite-like products.The possible formation of CaCO3as the component of the corrosion film on Mg-Ca alloy in NaCl solution as a result of CO2dissolution from the ambient air was not included in the present model.

4.Perspectives and suggestions

Nowadays there is increasing interest in developing and studying magnesium–calcium alloys [122].It is proved that binary Mg–Ca implants have such biological peculiarities,as osteoconductivity, possess high potential for cell adhesion and stimulation of cell growth on the material surface [123].Mohamed et al.[24]indicated that Mg–Ca alloys are promising for application as orthopaedic implants.Calcium is the essential element that forms hydroxyapatite complex, which is responsible for the rigidity and stability of the bone.During the degradation process, co-releasing of magnesium and calcium ions has a good effect on the bone healing and regeneration process [21].

In this work, Mg-0.8Ca alloy was used as the test material.Mg-0.8Ca alloy is commonly used in biomedical research and it is also representative for Mg-Ca types of magnesium alloy[32].Mg–0.8Ca alloy is a promising biocompatible, bioresorbable implant material, which shows high bioactivity and compatibility with human cell lines [24].

However, the addition of alloying elements (like Ca or Zn) intended to improve the mechanical properties of the Mg biomaterial also results in corrosion activation increase, since promoting the micro-galvanic couple formation [21,124,125].The results of this work also indicate that the addition of Ca to the alloy system is responsible for its intensive dissolution as compared to the pure Mg.

In the current work, the detailed corrosion analysis performed using novel scanning electrochemical techniques(scanning vibrating electrode technique coupled with scanning ion-selective electrode technique) combined with SKPFM and SEM-EDX measurements revealed the effect of precipitations on the corrosion behaviour of the Mg-Ca alloy.SVET/SIET firstly showed that Mg2Ca phase, which forms at the grain boundaries and in theα-Mg matrix, is anodic to the Mg matrix.In vitroelectrochemical studies performed in physiological solutions (MEM, 0.9 wt.% NaCl solution) established the faster degradation of Mg2Ca phase and higher susceptibility to pitting corrosion of the Mg-0.8Ca alloy as compared to pure magnesium.It was established using SKPFM that this secondary phase possesses a more negative electrochemical potential than theα-Mg phase.

The results of these studies provided new insights in the analysis of the corrosion behaviour of the sample.This paper showed additional opportunities of SVET/SIET localized methods in the precise investigation of corrosion activity of electrochemically active material(like Mg,Al,Zn etc.).These techniques provide a better understanding the effect of microstructure on thein vitrodegradation behaviour of the materials in various physiological solutions.This is the key information, which ensures the development of effective ways of corrosion protection of materials.

By varying the operating modes of SVET/SIET this study presented the effect of secondary phases on the overall corrosion degradation trend of Mg alloy.On the basis of the electrochemical analysis, the mechanism of the alloy bioresorptionin vitrowas established.The results of this study also indicated that application of the local electrochemical methods and traditional electrochemical methods, along with weight loss tests, SKPFM and SEM-EDX analysis, micro-Raman spectroscopy, and XPS measurements provides to obtain the complete corrosion scenario and shows the influence of the corrosion products formed on the surface of the material on the rate of material biodegradation.

It is stated that despite the lower corrosiveness of the sample in the mammalian cell culture medium, in comparison with 0.9 wt.% sodium chloride solution, the obtained results indicate low corrosion resistance of the Mg-0.8Ca alloyin vitro.To promote the use of this magnesium-calcium alloy as temporary implants designed for reliable fixation of bone fragments during the period of their healing, it is necessary to develop a method for the formation of anticorrosive bioactive layers on the surface of resorbable biomaterials.There are different ways of the protective coating formation, which can suppress the biodegradation rate of the Mg-based material and provide the desired bioresorption, biocompatibility and bioactivity [124,126,127].These methods include Ca-P conversion coatings (including HA-containing ones) [128,129],sol-gel coatings [130], and plasma electrolytic oxidation coatings [131–137].

Nowadays there are several companies (BIOTRONIK(Magmaris?)[138],Syntellix AG(MAGNEZIX?)[139],and U&i (RESOMET?) [140]), that provide certified commercially available implants made from Mg alloy.They use highpurity magnesium (99.99%) implants, Mg-Ca or Mg-Y-RE-Zr alloys [141]for the treatment of bone fractures and vascular stenosis[142].To the best of our knowledge,these companies do not apply the above-mentioned ways of coating formation.However, these implant materials also have protective layers on their surface intended to decrease the degradation rate and to promote better bioactivity to the material.For example,Mg scaffold material produced by Magmaris? used a dualcoated system including active biodegradable poly(l-lactide)-containing coating [138].Therefore, the development of new effective ways of bioactive coating formation is an important task for corrosion protection of Mg-based implant material.

Another promising way to control the degradation behaviour of the magnesium implants is to design the new composition of Mg-Ca alloys.One of these studies shows the promising way of formation of ‘‘stainless” Mg alloys with low amounts of Ca: 0.05, 0.1 and 0.15 wt.%, whichDRvalues are 3 times lower than the intrinsic corrosion rate of magnesium [37].

5.Conclusion

In this work, we have demonstrated the detailed analysis of corrosion behaviour of Mg-0.8Ca alloy in physiological media (MEM, NaCl solutions).The effect of composition,microstructure and heterogeneity of the bioresorbable magnesium alloy on the corrosion performance of the alloy at the micro- and meso–level was fully evaluated via localized elec-trochemical technique,gravimetric method,conventional electrochemical measurements and spatially resolved technique.The following can be highlighted:

?SVET and SIET with pH-selective microelectrode are efficient instruments which enable to show the effect of secondary phases on the overall corrosion degradation trend.Taking into account the known controversial results regarding the electrochemical activity of Mg2Ca – in this work using localized electrochemicalin situanalysis and SKPFM measurements the anodic behaviour of the Mg2Ca phase was proved.This secondary phase creates a localized microgalvanic cell withα-matrix that facilitates the dissolution of Mg2Ca compounds at the grain boundaries(and withinα-Mg grains) during the corrosion process.This provides faster degradation, higher susceptibility to pitting formation and, therefore, increased corrosion of the Mg-0.8Ca alloy.The parameters of local electrochemical methods were optimized forin vitrotests of the bioresorbable material surface.It was established that the maximum electrochemical activity of the magnesium alloy is revealed at the initial stage of its exposure to MEM (the first 12 min), after which there is a decrease and stabilization of the corrosion process associated with material passivation.

?The complementary SEM-EDX, XPS, Raman microspectroscopy analysis showed the formation of magnesium-andcarbonate substituted hydroxyapatite in the composition of corrosion film.The nonuniform distribution of the apatite phase in the corrosion film formed on alloy in MEM was shown.

?The effect of corrosion products formed on the surface of a magnesium-calcium alloy on the rate of implant material resorption was established.Formation of Mg2+-and-CO32–– substituted hydroxyapatite on the surface of the material in MEM provides the low local pH of the medium nearly the sample surface (ca.7.4 – 7.5), in contrast to the typically high value (ca.11) for the magnesium and its alloys in NaCl solution.A lower electrochemical activity of the Mg-0.8Ca alloy in MEM was detected in comparison with the 0.3 wt.% and 0.9 wt.% NaCl solutions.As a result of HA-containing layer formation the difference in the local current density recorded in the anodic and cathodic regions of the surface decreases (from 40 μA?cm–2to 3 μA?cm–2), and corrosion of the magnesium alloy slows down (Icdecreases and|Z|f→0Hzincreases by more than one order of magnitude).This Ca-P film containing HA reduces the degradation rate of Mg-0.8Ca alloy in two times as compared to Mg(OH)2corrosion layer formed on the sample in NaCl solution (1.44 ± 0.15 mm year–1and 2.83 ± 0.49 mm year–1, respectively).TheseDRvalues were higher than the intrinsic corrosion rate of high-purity magnesium in concnetrated chloride solution(0.3 mm year–1), which indicates the low possibility of application of Mg-0.8Ca alloy without coating protection in implant surgery.

?The mechanism of the material bioresorptionin vitrowas established and a model of Mg-0.8Ca alloy biodegradation in MEM and NaCl solution was suggested.The main electrochemical and time parameters of the surface film evolution, describing the growth stage (the first 30 h of exposure) and its subsequent degradation, were established.

Acknowledgements

The study of material’s structure, composition, and corrosion processes kinetics was supported by the Grant of Russian Science Foundation, Russia (project no.20–13–00130, https://rscf.ru/en/project/20–13–00130/).SKPFM, weight loss tests and local electrochemical measurements were supported by the Grant of Russian Science Foundation, Russia (project no.21–73–10148, https://rscf.ru/en/project/21–73–10148/).Raman spectra were acquired under the government assignments from the Ministry of Science and Higher Education of the Russian Federation, Russia (project no.0205–2021–0003).The Far Eastern centre for Electron Microscopy of National Scientific Centre of Marine Biology FEB RAS (Vladivostok,Russia) is gratefully acknowledged.