K.S.Durán ,N.Hernández ,L.M.Rued ,C.A.Hernández-Brrios ,A.E.Coy ,F.Viejo,?

aGrupo de Investigación en Desarrollo y Tecnología de Nuevos Materiales(GIMAT),Universidad Industrial de Santander,Bucaramanga,Calle 9 Cra.27,Colombia

bGrupo de Investigación de las Ciencias de las Ingenierías(GICI),Universidad de San Buenaventura,Cartagena,Calle Real de Ternera,Diagonal 32 # 30–966,Colombia

Abstract The present study aims to develop multilayer barrier-bioactive hybrid sol-gel coatings from a mixture of the silane precursors tetraethylorthosilicate(TEOS)and glycidoxypropyltriethoxysilane(GPTMS)deposited on the Elektron 21 magnesium alloy.The purpose of the inner layer(barrier coating)was to provide corrosion protection to the magnesium alloy,whereas the outer layer(bioactive coating)was doped with different Ca and Mg contents to produce a bioactive material.The coatings were characterised using scanning electron microscopy(SEM)and their corrosion behaviour was evaluated by anodic polarisation and electrochemical impedance spectroscopy after immersion in simulated body fluid(SBF)at 37±0.5°C.The experimental results showed that the multilayer coatings increased the corrosion resistance of the alloy up to three orders of magnitude during immersion in SBF solution.On the other hand,the presence of Ca and Mg in the bioactive coating promoted the growth of apatite-like phases.However,an increment of salt content favoured the formation of porous coatings,which allowed the access of the electrolyte to the substrate leading to their rapid deterioration.Despite the latter,this research endorses the premise that the TEOS-GPTMS hybrid system represents a promising alternative to produce bifunctional barrier-bioactive coatings.

Keywords: Magnesium;bioactivity;coating;sol-gel;EIS;corrosion.

1.Introduction

In the last decades,metallic biomaterials have gained increasing interest in the research and development of orthopaedic implants used to repair or replace bone tissue[1–3].Common metallic biomaterials include titanium and cobalt–chromium-based alloys,as well as the 316L stainless steel.However,a critical limitation of these materials is associated with their relatively high Young’s modulus and density(E=100-200GPa andρ=4.4-9.2g/cm3)compared to bone(3 -20GPa andρ=1.8 to 2.1g/cm3)that can lead to the occurrence of stress-shielding [4–6].This phenomenon takes place when the bulk load is supported mainly by the implant,which prevents the surrounding bone tissue from receiving suitable stimuli for its growth [7].This condition gives rise to severe clinical consequences such as bone resorption and an early loosening and failure of the implant [6].Consequently,any reduction in stiffness of the implant might adapt the redistribution of the load to the adjacent bone tissues,minimizing the appearance of the stress shielding effect [8].

One of the most promising strategies is the employment of magnesium alloys due to their high specific mechanical properties with regard to the conventional materials [1,7].Mg is a lightweight metal with Young’s modulus(E=41–45GPa)and density(1.74–1.84g/cm3)close to those of natural bone[9–11].Moreover,magnesium exhibits excellent biocompatibility,it being the fourth most abundant cation in the human body,with an estimation of 1mol of magnesium in a 70kg adult,where approximately half of the total physiological magnesium is found in bone tissue [11,12].

Nevertheless,an important aspect to consider is that not all Mg alloys are suitable for the design of orthopaedic implants,since some alloying elements may cause adverse effects on human health [1].Particularly,the Elektron 21 alloy(Mg-Zn-Re,Re≡Nd,Gd)has been identified as an attractive candidate for biomedical applications [13–16].In this regard,zinc is known to be one of the most important trace elements in the human body that serves as a cofactor for enzymes in bone and cartilage [17].On the other hand,many RE elements possess anti-carcinogenic characteristics [18].

Unfortunately,the rapid corrosion rate of the Mg alloys in physiological environments represents a great challenge for their application in the field of biomaterials [8,9].Mg is a thermodynamically unstable element with a standard electrode potential of?2.37V [19,20],which makes it easily corrodible in the human body.Moreover,the corrosion process of the magnesium alloys promotes the development of a magnesium hydroxide(Mg(OH)2)layer that does not offer effective protection in the physiological environment,in which it reacts with chloride ions and produce soluble magnesium chloride(MgCl2)[21–23].All the above leads to a severe deterioration of the mechanical integrity of the metallic implant before the tissues have enough time for healing,hence it is crucial to design a barrier coating between the alloy and the corrosive medium [22,23].

An additional factor to take into account in the development of orthopaedic implants entails the need for the surface to be bioactive,which is related to the capacity to create a natural union through chemical bonding between bone and metallic surface,favouring cell adhesion and growth of new tissue[1,8,24].With the purpose to meet this requirement,the development of bioactive ceramic coatings such as hydroxyapatite(HA)and bioglasses has been extensively studied due to their ability to facilitate the integration of the implant into de surrounding tissue [25,26].Hydroxyapatite has been preferred as bioactive coating due to its chemical,structural and biological similarity to the human bone as well as its capacity to support osteoblast adhesion and proliferation [27,28].On the other hand,bioglasses are characterised by possessing excellent biocompatibility and osteogenic capacity,as they are composed of minerals that are naturally found in the body(SiO2,CaO,Na2O)[29,30].However,the main drawbacks of both bioactive coatings are their high cracking tendency,limited corrosion protection and,the elevated temperatures required during the curing stage(600°C -1000°C),a critical restriction for alloys with low melting point such as magnesium alloys(650°C)[6,30].

Considering the foregoing,it is essential to develop a multilayer coating with bifunctional response on the magnesium alloy that acts as a barrier against corrosion and,simultaneously,promotes the rapid healing of the damaged tissues.Specifically,the development of TEOS-GPTMS hybrid sol-gel coatings represents a promising alternative since TEOS provides good adhesion and high corrosion resistance.Meanwhile,GPTMS enhances flexibility,preventing the cracking susceptibility and allowing the production of multilayer coatings [31,32];its presence also reduces the curing temperature below 150°C [6,30].Additionally,GPTMS augments the nanoporosity of the SiO2network,which can be used to modify the surface activity through the incorporation of different species within the coating structure [33].

In particular,the incorporation of calcium and magnesium salts is an effective method to produce biologically active materials that stimulate bone repair and regeneration [34,35].It is well known that calcium ions play a special role in bone metabolism,during proliferation,differentiation,and mineralization of the osteoblast cells [36].Likewise,magnesium has been described as a vital mineral element in a variety of biological processes such as bone remodelling and skeletal development.Moreover,magnesium acts actively in the regulation of calcium transport [34,36].Finally,Ca and Mg can also stimulate the growth of crystalline hydroxyapatite in contact with body fluid [15,25].In this regard,different investigations of silica-based coatings have reported that both elements could favour a rapid ions exchange,leading to the formation of silanol(Si–OH)groups on the glass surface,which generally act as nucleation sites for apatite phase precipitation [37–40].

Therefore,the purpose of the present work was to design multilayer hybrid sol-gel coatings with dual barrier-bioactive functionality based on the TEOS-GPTMS system and applied on the Elektron 21 magnesium alloy for its potential application in the manufacture of temporary orthopaedic implants.Special attention was focused on the effect of Ca and Mg on the morphology and the surface reactivity of the multilayer system in SBF solution.

2.Experimental

2.1.Material

The chemical composition(wt.%)of the Elektron 21 alloy was 1.0–1.7 Gd;2.6–3.1 Nd;0.2–0.5Zn;saturated amount of Zr and balance Mg.The alloy was provided by Magnesium Elektron Ltd.(UK)in the form of 20mm thick square plates that were cut with an abrasive metal-cutting disc to obtain specimens with dimensions of 25×20×3mm.

2.2.Microstructural and elemental characterisation

The microstructure of the Elektron 21 alloy was examined by scanning electron microscopy(SEM)employing a QUANTA FEG 650 microscope equipped with energy dispersive X-ray(EDX)analysis and backscattered electron(BSE)detectors.

Fig.1.Schematic diagram of the preparation of the BS and DS hybrid sol-gel coatings.

2.3.Preparation and characterisation of the hybrid sols

2.3.1.Barrier sol

Based on previous works [31,32,41,42],the barrier sol(namely BS)was synthesised by mixing TEOS and GPTMS precursors with a molar relation of 3:1,using ethanol as solvent and acetic acid as catalyst in volume proportions of 1:1 and 0.45:1 with regard to the TEOS-GPTMS precursors.The ageing time of the sol before the deposition stage was set at 21 days with the aim to ensure the adequate progress of the hydrolysis and condensation reactions.

2.3.2.Doped sols

Doped sols(DS)were produced under the same conditions established for the preparation of the BS sol,in which calcium nitrate tetrahydrate(Ca(NO3).4H2O)and magnesium nitrate hexahydrate(Mg(NO3).6H2O)were added in different molar proportions with regard to the TEOS-GPTMS mixture(Table 1).Additionally,the effect of the ageing time(from 1 to 7 days)was evaluated.In this regard,the viscosity evolution of the DS sols as a function of ageing time was measured using a rotational rheometer Brookfield DV.III with a shear rate of 240 s?1.The viscosity measurements were taken three times for each Ca-Mg molar ratio and ageing time.

Table 1 Experimental conditions for the synthesis of Ca-Mg doped sols.

Table 2 Electrochemical parameters obtained from the anodic polarisation curves.

2.4.Deposition and characterisation of the hybrid coatings

Before the deposition process of the hybrid sols,Elektron 21 specimens were ground on water lubricated SiC abrasive papers from 120 to 1200 grit size,and then rinsed with ethanol and dried in warm air.Later,specimens were pretreated in 4 vol% HF solution at room temperature for 24h to promote the development of a fluoride conversion layer that avoided the direct contact of the metallic substrate with the acidic sols during the deposition stage [31].Afterwards,BS and DS sols were deposited by dip-coating,obtaining a multilayer coating that consisted of a bilayer of the BS sol and an external single layer of the DS sols.For that purpose,the pretreated specimens were immersed in the sols for 15s with a drying time of 20s between each layer at an immersion-extraction velocity of 2mm/s.After deposition,specimens were heat treated in air atmosphere at 60°C for 2h,followed by 120°C for the same time with a heating rate of 4.5°C/min.Finally,the examination of the surface morphology and elemental composition of the different hybrid sol-gel coatings deposited was performed by SEM-EDX(see Section 2.2).

A schematic representation of the preparation of the BS and DS hybrid sol-gel coatings is illustrated in Fig.1.

2.5.Corrosion evaluation of hybrid coatings

Corrosion performance was assessed by potentiodynamic anodic polarisation and electrochemical impedance spectroscopy,utilising an AUTOLAB PGSTAT302N potentiostat/galvanostat connected to a conventional three-electrode electrochemical cell:the coated specimens,as the working electrode,with an exposed area of 1.0 cm2,whereas Ag-AgCl and platinum were the reference and counter electrode respectively.The electrolyte selected was simulated body fluid(SBF)at 37±0.5°C,which was prepared from the following reagents:8.04g/l NaCl,0.36g/l NaHCO3,0.23g/l KCl,0.23g/l K2HPO4·3H2O,0.31g/l MgCl2·6H2O,1.0M HCl(39–44ml),0.29g/l CaCl2and 0.07g/l Na2SO4in distilled water at pH 7.4 employing 6.12g/l of TRIS(CH2OH)3CNH2and 1.0M HCl [43].A thermostatic bath was used to keep the temperature constant during the tests.The electrochemical measurements were repeated at least twice to confirm the reproducibility of the results.For comparison purpose,the BS coating and the bare alloy were evaluated in this stage.

In order to reach a stable potential,the open circuit potential(OCP)was monitored during the first hour of immersion of the specimens in the corrosive solution.The potentiodynamic anodic polarisation test was undertaken at a scan rate of 0.3mV/s,starting from?0.1V to 2.0V,with regard to the OCP.The testing was finalised once an anodic current density of 1.0mA/cm2was attained.On the other hand,electrochemical impedance measurements were conducted after immersion of the specimens in the test solution for times of 1,3 and 5h,applying a sinusoidal potential signal with an amplitude of 10mV with regard to the OCP over a frequency range from 100kHz to 0.01Hz.Once the electrochemical impedance response was completed,specimens were rinsed with distilled water and dried in warm air and,then,examined by SEMEDX.

2.6.Bioactivity evaluation of hybrid coatings

In order to evaluate the bioactive character of the hybrid sol-gel coatings and identify the precipitation of calcium phosphates on the surface coatings,an immersion test was carried out in SBF solution at 37±0.5°C for a period of 48h.The morphology and elemental composition of the corrosion products formed on the coated specimens were determined by SEM-EDX.

3.Results and discussion

3.1.Microstructural characterisation of the Elektron 21 alloy

Fig.2 shows the microstructure of the Elektron 21 alloy that consisted ofα-Mg matrix with secondary coarse Mg12(NdxGd1-x)particles located at the grain boundaries.Moreover,two types of precipitates were found:i)irregular precipitates rich in Zr with a size of about 1–2μm;and ii)precipitates with acicular morphology,containing Zn and Zr in their composition [44–46].

Fig.2.SEM micrograph(BSE mode)of the microstructure of the Elektron 21 alloy.

3.2.Characterisation of the bs coating

Fig.3a discloses the SEM micrograph of the Elektron 21 specimen pretreated in HF solution(4 vol%?24h)before the deposition of the BS coating.It was noticed that the conversion pretreatment promoted the formation of a relatively uniform layer.However,detailed surface examination revealed the presence of cracks in the Mg12(NdxGd1-x)phase,that were possibly produced by the partial dissolution of this phase during the immersion in HF.

Despite the above,the characterisation of the specimens after further deposition of the BS coating(Fig.3b)displayed that the alloy surface was homogeneously covered,without apparent presence of defects in the sol-gel coating.Complementarily,the cross-sectional SEM micrograph of the protective system(Fig.3c)allowed discerning the fluoride layer and the BS coating,with thicknesses of about 0.5–0.6 μm and 1.0–1.1μm respectively.

3.3.Corrosion evaluation of the Elektron 21 alloy and the bs coating

3.3.1.Potentiodynamic anodic polarisation

Fig.4 and Table 2 present the results obtained from the potentiodynamic anodic polarisation tests of the bare alloy and the BS coating after 1h of immersion in SBF solution.The bare alloy showed a relatively high corrosion current density value(icorr)(2.3×10?4A/cm2),indicating that the material underwent an accelerated corrosion process in the test solution.This fact gave rise to the development of a corrosion film on the metallic substrate,mainly in form of magnesium oxide/hydroxides,which exhibited a passivation range(Epit-Ecorr)of 450mV.Nevertheless,its passivation current density(ipass)was considerably high(above 10?3A/cm2on average),which signifies that this corrosion film is unstable in SBF solution and,therefore,it only provides partial protection [23].On the other hand,the BS coating significantly enhanced the electrochemical behaviour of the alloy in the corrosive medium by diminishing the icorrvalue by about three orders of magnitude,whereas both the corrosion and pitting potentials were displaced towards more noble values.

Table 3 EIS Fitting parameters for the Elektron 21 alloy after different immersion times in SBF solution.

Table 4 EIS Fitting parameters for the BS coating after different immersion times in SBF solution.

Fig.3.SEM micrographs(BSE mode)of the a)pretreated Elektron 21 alloy with 4 vol% HF,b)BS coating and c)cross-section.

3.3.2.Electrochemical impedance spectroscopy(EIS)

The EIS spectra of the Elektron 21 alloy and the BS coating are depicted in Fig.5.The total impedance modulus at low frequencies(0.01Hz)determined for the bare alloy reached a value of 2.5×102Ω·cm2at early times of immersion,confirming the poor corrosion resistance of the alloy(Fig.5a).However,after 5h of immersion,it increased up to 5.6×102Ω·cm2due to the growth of a partially protective corrosion layer in accordance with the results obtained from the anodic polarisation testing.Further,the Bode phase angle diagram exhibited three time constants that correspond to three different interfacial processes,Fig.5b.The first time constant,located in the high frequency range(～105Hz),was assigned to the presence of a corrosion products layer at the electrolyte/metal interface.The second one,at medium frequencies(～102Hz),was attributed to the charge transfer and electrical double-layer.Finally,the third time constant,identified in the low frequency region(10?1–10?2Hz),was associated with the diffusion of ions through the surface and adsorption-desorption processes of species such as Mg+,Mg(OH)+,or Mg(OH)2[19,47,48].

Concerning the Bode magnitude diagram for the BS coating(Fig.5c),the total impedance modulus remained at values around 1.1×105Ω·cm2during the first 3h of immersion,approximately three orders of magnitude higher than that determined for the bare alloy,which is indicative of corrosion protection provided by the BS coating.Nevertheless,it showed a moderate decrease along the immersion time(7.9×104Ω·cm2after 5h),suggesting the presence of structural defects in either the barrier coating or the fluoride conversion layer.

Fig.4.Anodic polarisation curves of the bare Elektron 21 alloy and the BS coating after 1 h of immersion in SBF solution.

The Bode phase angle diagram corresponding to the coated specimens(Fig.5d)allowed identifying four time constants for all the immersion times evaluated.The first one,at high frequencies(～105Hz),was attributed to the response of the barrier properties of the BS coating.Further,the second and third constants are overlapped in the medium frequency region(～10–102Hz),and correspond to the electrochemical response of the fluoride layer and the charge transfer along with the electrical double-layer effects at the metal/coating interface.Finally,the fourth constant,in the low frequency range(10?1–10?2Hz),was related to the adsorption/desorption processes of intermediate species [49,50].

Fig.5.Bode diagrams of the a)-b)bare Elektron 21 and c)-d)SB coating after 1,3 and 5h of immersion in SBF solution.

Fig.6.Equivalent circuits of the a)bare Elektron 21 and b)BS coating after immersion in SBF solution.

In order to complete the analysis of the EIS spectra,the impedance data were fitted to the equivalent electrical circuits disclosed in Fig.6.In these models,constant phase elements(CPE)were used to describe a non-ideal capacitive behaviour associated with heterogeneities of the system.The respective fitting parameters were interpreted as follows:Rsrepresents the SBF solution resistance,Rcorrand CPEcorrcorrespond to the resistance and the capacitance of the corrosion products layer respectively.Rctand CPEdlrelated to the charge transfer resistance and the electrical double-layer capacity at the substrate/electrolyte interface.Likewise,L and RLare associated with the inductance and the inductance resistance that are correlated to the relaxation process of adsorbed species at the surface free of corrosion products,implying the initiation of localised corrosion.Lastly,RBS,CPEBSand Rfl,CPEflelements represent the resistance and capacitance of the BS coating and the fluoride layer respectively.

The EIS fitting results of the bare Elektron 21 alloy and BS coating immersed in SBF are comprised in Tables 3 and 4.In the case of the alloy,the formation and growth of a partially protective corrosion layer along the immersion time led to an augmentation of the Rcorrvalues that was accompanied by a decrease in CPEcorr.The latter resulted in a diminution in the production of adsorbed intermediate species,it being supported by the rising of the inductance resistance RLfor immersion times from 3h [50].Regarding the BS coating,RBSexperienced a reduction from 5.3×104Ω·cm2to 3.0×104Ω·cm2during immersion in SBF solution.This trend can be explained by the progressive deterioration of the barrier properties of the coating due to the possible incidence of localised defects that served as pathways for the access of the electrolyte.

Finally,the SEM-EDX analysis of the bare alloy and coated specimens after 5h in SBF solution is depicted in Fig.7.Elektron 21 alloy displayed a cracked corrosion layer mainly composed of Mg,P,Ca and O probably in form of oxides and hydroxides of Mg,as well as calcium phosphates(Fig.7a and b).In contrast,concerning the BS coating,there was no evidence of corrosion products or pits on the surface morphology.Nonetheless,it was detected the presence of cracks around the intermetallic phase,possibly associated with the defects originally generated by the partial dissolution of the intermetallic phase during the pretreatment of the alloy in HF solution(Fig.7c).This observation is in accordance with the hypothesis previously discussed where a reduction of the coating resistance RBSwas related to the presence of localised defects.

3.4.Characterisation of the doped sols

Table 5 shows the viscosity values of the doped sols as a function of the ageing time as well as the Ca and Mg contents.As expected,the viscosity of the doped sols increased along the ageing time due to the progress of the hydrolysis,condensation and polymerisation reactions of the precursors[31].Regarding the addition of Ca and Mg,the presence of both elements in the sols augmented the viscosity as a result of their incorporation within the silica structure,which promoted the formation of nucleation sites and,therefore,accelerated the gelation process [29,35,37].

Table 5 Viscosity variation of the doped sols as function of Ca-Mg molar relation and ageing time.

Fig.7.SEM micrographs of the a)bare Elektron 21 alloy(BSE mode),b)EDX analysis of the bare alloy surface(red sign),c)BS coating(SE mode)after 5h of immersion in SBF solution.(For interpretation of the references to colour in this figure legend,the reader is referred to the web version of this article.)

3.5.Characterisation of the BS-DS coatings

The SEM micrographs of the different barrier-doped coatings(BS-DS)are displayed in Fig.8.In general,it was evidenced that the coatings were apparently continuous for all the synthesis conditions employed(Fig.8a-b).Nevertheless,for 7.5 mol% Ca-Mg and 5 days of ageing,the BS-DS coating exhibited a porous morphology,with a pore diameter comprised between 200 and 300nm approximately(Fig.8c).This fact may be associated with the elevated gelation rate of the sols together with the opening of the silica network triggered by the incorporation of Ca and Mg ions.On the other hand,the BS-DS coatings synthesised at different Ca-Mg molar ratios and equal total salt content did not show significant changes in their surface morphology,which is in agreement with the negligible fluctuation of the viscosity values of the sols under these experimental conditions(Table 5).

Table 6 EIS Fitting parameters for the BS-DS coatings after different immersion times in SBF solution.

Fig.8.SEM micrographs(BSE mode)of the surface of the BS-DS coatings synthesised under different ageing times with:a)2.5 and b)-c)7.5 mol% Ca-Mg;d)Cross-section of the BS-DS coating with 2.5 mol% Ca-Mg.

Complementarily,the cross-sectional examination revealed that thicknesses of the doped coatings ranged from 1.4 to 1.7μm,Fig.8d,without evidencing a specific trend regardless of the synthesis condition studied.To finish,it is important to mention that the coatings with the highest total salt concentration and the longest ageing time(7 days)are not presented in the manuscript since they could not be deposited on the metallic substrate due to the excessive gelation degree of the corresponding sols.

3.6.Corrosion evaluation of the BS-DS coatings

The Bode magnitude diagrams of the coatings with 1.25Ca-1.25Mg and 3.75Ca-3.75Mg after 1,3 and 5h of immersion in SBF solution are displayed in Fig.9.Initially,the BS-DS coatings obtained from the sol with 1.25Ca-1.25Mg and 1 d of ageing exhibited a total impedance modulus at low frequencies(0.01Hz)of about 1.6×105Ω.cm2,which was slightly higher than that determined for the BS coating(Fig.9a).However,the increase of the total salt content or ageing time led to a moderate deterioration of the barrier properties of the multilayer hybrid system.Accordingly,the hybrid coatings with 3.75Ca-3.75Mg(1 d)and 1.25Ca-1.25Mg(7 d)showed a total impedance modulus of 1.2×105Ω.cm2after 1h of immersion(Figs.9b and c).Moreover,the combination of both parameters contributed to a more accentuated decrease in the total impedance(3.2×104Ω.cm2for the condition 3.75Ca-3.75Mg and 5 days of ageing)(Fig.9d).The latter is probably correlated to the inherent porosity observed in these coatings(Fig.8c)that facilitated the penetration of corrosive species and promoted the deterioration of the metallic substrate.

Fig.9.Bode magnitude diagrams of the BS-DS coatings synthesised under different ageing times with:a)-c)1.25Ca–1.25Mg(mol%)and b)-d)3.75Ca-3.75Mg(mol%)after 5h of immersion in SBF solution.

Fig.10.Impedance modulus evolution at low frequencies(0.01Hz)of the BS-DS coatings as a function of the Ca and Mg content after 1 and 5h of immersion in SBF solution.

Regarding the influence of the immersion time in SBF solution,two opposite trends were evidenced as a function of the total salt content regardless of the ageing time employed during the deposition of the hybrid coatings.For low concentrations,the total impedance of the BS-DS coatings experienced a progressive reduction whereas,for high salt concentrations,this parameter slightly augmented as the immersion time progressed.In this regard,it is worth to highlight that the condition 3.75Ca-3.75Mg(1 d)exhibited the highest total impedance value after 5h of immersion,reaching a value of 1.8×105Ω.cm2.Furthermore,for equal total salt content,the impedance modulus of the BSDS coatings increased as the magnesium concentration was greater,which suggests that this element plays an active role in the electrochemical activity of the hybrid system(Fig.10).

Fig.11.Bode phase angle diagrams of the BS-DS coatings synthesised under different ageing times with a)-c)1.25Ca–1.25Mg(mol%)and b)-d)3.75Ca-3.75Mg(mol%)after 5h of immersion in SBF solution.

Fig.12.Equivalent circuits of the BS-DS coatings synthesised under different ageing times and total salt contents.

Fig.13.SEM micrographs(BSE mode)of the BS-DS coatings synthesised under different ageing times with a)-b)-d)1.25Ca–1.25Mg(mol%)and c)3.75Ca-3.75Mg(mol%)after 5h of immersion in SBF.

Fig.14.SEM-EDX analysis of the BS-DS coatings synthesised under different ageing times with 1.25Ca–1.25Mg(mol%)and 3.75Ca-3.75Mg(mol%)after 48 h of immersion in SBF:a)-b)surface micrographs(BSE mode),c)detail at higher magnification of(a);d)EDX spectrum of the precipitates observed in(c);e)detail at higher magnification of(b).

On the other hand,Fig.11 discloses the Bode phase angle diagrams of the different systems previously analysed.For the BS-DS coatings synthesised with the sol 1.25Ca-1.25Mg(1 d),four time constants can be identified(Fig.11a):i)the first one located at high frequencies(～105Hz)and ascribed to the response of the BS-DS coating;ii)the second and third constants,overlapped in the region of medium frequencies(～1–102Hz)are again attributed to the presence of the fluoride layer and the charge transfer as well as the electrical double layer effects and iii)a fourth constant at low frequencies(～0.1–0.01Hz)correlated to the ionic diffusion and the adsorption/desorption processes of intermediate species.For greater ageing times or total salt contents(Fig.11b and c),the phase angle diagram did not show substantial differences in the time constants.Finally,the presence of porosity produced due to high salt content and long periods of ageing allowed the transport of the electrolyte through the coating,favouring the subsequent dissolution of the fluoride layer and the interaction of the metallic substrate.As a result,a new phase constant was noticed around 103Hz,possibly associated with the nucleation of an intermediate corrosion layer,in the form of oxides and hydroxides of Mg together with calcium phosphates,at the coating-metallic substrate interface(Fig.11d).

The different electrochemical systems discussed above were fitted using three models of equivalent circuits,which are presented in Fig.12,whereas their respective fitting parameters are summarised in Table 6.In these models,RBS-DSand CPEBS-DSare referred to the resistance and non-ideal capacitance associated with the barrier-bioactive coating(BSDS).The values obtained for each parameter of the proposed equivalent circuits evidence a clear consistency with the previous discussion of the EIS results.

Fig.13 displays the SEM micrographs of the BS-DS coatings after 5h of immersion in SBF.The coatings under the synthesis condition 1.25Ca-1.25Mg(mol%)and 1 d of ageing were apparently homogeneous.Nevertheless,a detailed examination allowed identifying the presence of small cracks around the intermetallic phase(Fig.13a and b).The occurrence of these defects became greater as the total salt content or the ageing time increased,causing the partial detachment of the coating(Fig.13c and d).The latter results are in agreement with those disclosed in the EIS analysis.However,the immersion time(5h)is not long enough to discern clearly the degradation-bioactivity behaviour of the BS-DS system.

3.7.Bioactivity evaluation of BS-DS coatings

Fig.14 depicts the surface morphology of the BS-DS coatings after 48h of immersion in SBF solution.The electronic micrographs showed that all the hybrid coatings exhibited remarkable delamination regardless of the synthesis condition(Fig.14a and b).Despite this,it was noticed the growth of nodular precipitates on the coated surface,which EDX analysis revealed that they were mainly composed of Ca and P(Fig.14c and d),probably in the form of apatite-like phases with a Ca/P ratio of 0.74±0.05 that was calculated after repeating this analysis for three times.Further,for high total salt content,those precipitates coalesced to produce a denser and more continuous layer(Fig.14e).Accordingly,some studies have reported that the structural role of Ca and Mg has a significant effect on the bioactive nature of the surface of the sol-gel-derived glasses.The latter is based on the fact that both elements are network modifiers that enter into the glass structure forming non-bridging oxygen bonds(≡Si-OCa-O-Si ≡and ≡Si-O-Mg-O-Si ≡).Once the surface of the glass interacts with the SBF solution,an ion exchange between Ca2+and Mg2+ions from glass and H3O+ions from the SBF solution occurs,leading to the production of silanol groups(Si-OH)that act as nucleation sites for the precipitation of the apatite phase [35,37–40].Furthermore,a greater Ca content causes an augmentation in the porosity of the glasses,which facilitates the apatite nucleation on the coating surface [51].

In retrospect,the experimental results demonstrated that the TEOS/GPTMS hybrid sol-gel coatings represent a promising approach to produce bifunctional systems with barrierbioactive properties since they might:i)allow controlling the corrosion rate of the Elektron 21 Mg alloy at early times of immersion in simulated body fluid(SBF)solution;and ii)increase the superficial reactivity of the metallic substrate,promoting the precipitation of calcium phosphates-rich species through the incorporation of Ca and Mg into the silane-based network.Nevertheless,further investigations would be necessary to optimise the protective system,making a special effort on the improvement of the conversion pretreatment either by the study of the synthesis variables or by the employment of different compounds,with the aim to avoid the selective attack on the zones where structural defects were observed.

4.Conclusions

The barrier coating(BS)provided a significant protective effect to the Elektron 21 alloy against corrosion processes by augmenting the impedance modulus at low frequencies(0.01Hz)by about three orders of magnitude at early times of immersion in simulated physiological fluid.However,the corrosion resistance of the coating partially decayed as the immersion time in SBF solution was increased.This result was possibly caused by the formation of defects around the intermetallic phase during the fluoride conversion pretreatment,leading to cracking and delamination of the BS coating.

An augmentation of the total salt content and the ageing time in the synthesis of DS sols favoured the gelation reactions of the sols,inducing the appearance of porosity in the surface of the BS-DS coatings.The latter enabled the access of the electrolyte to the substrate during initial periods of immersion in SBF solution,giving rise to the consequent reduction in corrosion resistance of the coated specimens.Nevertheless,Ca and Mg acted as network modifiers promoting the precipitation of calcium phosphates after immersion in SBF solution,probably in the form of apatite-like phases,indicating that the BS-DS coatings possess a potential bioactive response.

Data availability

All research data supporting this publication are directly available within this publication.

Declaration of competing interest

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

Acknowledgements

The authors wish to thank the Vicerrectoría de Investigación y Extension of the Universidad Industrial de Santander,Colombia(grant number 2508)for the financial support of the present work.