Clarifying the influence of albumin on the initial stages of magnesium corrosion in Hank’s balanced salt solution

Di Mei,Cheng Wng,Svitln V.Lmk,Mikhil L.Zheludkevich,b

a Magnesium Innovation Centre-MagIC,Institute of Materials Research,Helmholtz-Zentrum Geesthacht,Geesthacht 21502,Germany

bInstitute for Materials Science,Faculty of Engineering,Kiel University,Kiel 24103,Germany

Abstract The corrosion behavior of Mg,a promising biodegradable metallic material,in protein-containing pseudophysiological environment is still not fully understood.In this work,the influence of albumin on the corrosion behavior of commercially pure magnesium(CP Mg)is investigated during short-term tests in Hank’s Balanced Salt Solution(HBSS).This work focuses on the reactions at the Mg/medium interface from the perspective of the interactions among albumin,media components and substrate.Hydrogen evolution tests demonstrate that the physiological amount of albumin(40 g L−1)accelerates Mg corrosion in HBSS during the first few hours but slows down the degradation afterwards.The presence of albumin decreases the concentration of free Ca2+in HBSS and delays formation of protective co-precipitation products on Mg surface.The evolution of local pH(differs from typically monitored bulk pH)near Mg/medium interface in albumincontaining HBSS is reported for the first time.Comparison of local and bulk pH values elucidates the pH buffering effect of albumin during the immersion period.Based on these results,adsorption,chelating and pH buffering effects are summarized as three important aspects of albumin influence on Mg corrosion.Additionally,constant replenishment of medium components is shown to be an influential factor during the Mg corrosion tests.© 2020 Chongqing University.Publishing services provided by Elsevier B.V.on behalf of KeAi Communications Co.Ltd.This is an open access article under the CC BY-NC-ND license(http://creativecommons.org/licenses/by-nc-nd/4.0/)Peer review under responsibility of Chongqing University

Keywords:Magnesium;Corrosion;Albumin;Biodegradable;Local pH.

1.Introduction

During the service period of metallic implant materials,the interaction between the implant and surrounding environment influences its surface states,which is an important factor for the degradation behavior of biodegradable metallic materials,such as Mg and its alloys[1–10].Ion exchange,corrosion product formation and precipitation influenced by organic compounds are the important parts of these interactions.These are the research hotspots in the field of biomedical magnesium[8,9,11–18].The synergy between inorganic ions and the formation of corrosion products on Mg surface has been clarified[19–23],the Ca2+,carbonate and phosphate were found to have significant influences on Mg corrosion.Furthermore,to deepen the understanding of the in vivo degradation behavior of magnesium,a number of experimental works have focused on the influence of amino acids,vitamins,saccharides and other bio-relevant small molecule organic compounds on the Mg corrosion[24–26].The results showed that these organic compounds influenced the corrosion behavior of Mg by altering the formation of corrosion products.However,at low,physiological concentrations,these compounds did not critically influence Mg corrosion rate.

In addition to small molecule organic compounds,the influence of macromolecule organic compounds,such as proteins,on the corrosion of magnesium has attracted a lot of deserved attention.However,there are discrepancies in the research results regarding the influence of protein on Mg corrosion in published works[6,27–29].Liu[30]et al.investigated the corrosion behavior of Mg-Ca in albumincontaining NaCl solution.The results indicated that albumin at high concentrations significantly inhibited the corrosion of this alloy.Zhang[31]et al.reported that proteins slowed down the corrosion rate of Mg-Nd-Zn-Zr in cell culture medium.In contrast,Li[32]et al.investigated the influence of albumin on the corrosion behavior of Mg in Hank’s Balanced Salt Solution(HBSS)and found that albumin inhibited Mg corrosion in the first 6h but accelerated corrosion after long immersion times.Harandi[33]et al.provided similar results that BSA(bovine serum albumin)slowed the corrosion rate of magnesium in HBSS during the first 2 days and then increased it.Besides,Walker[27]et al.reported that the corrosion rate of several magnesium alloys in protein-containing MEM was significantly higher than that in Minimum Essential Medium(MEM),Earle’s balanced salt solution(EBSS)and in vivo.Similar results were also proposed by Gu[34]et al.Certainly,these discrepancies may be attributed to different composition of test media,tested materials,and test conditions,e.g.static vs.dynamic.In addition,long-term tests in protein-containing media in an open environment allow for the growth of microbial life,which may significantly affect Mg corrosion[35,36].

At the present stage,the mechanistic understanding of the protein influence on the corrosion of Mg is still insufficient[5].The barrier effect caused by protein adsorption has been used to explain its inhibitory effect.Hou[4]et al.experimentally investigated the influencing factors of protein adsorption on Mg in HBSS and Dulbecco’s Modified Eagle Medium(DMEM).Protein absorption was found to slow down Mg degradation.Computational methods have also been employed to gain the insights into the action mechanism of organic molecules(such as amino acids,peptides and proteins)on Mg,the alloying elements and the surface status of samples were found to influence the adsorption of the organic molecules on Mg[37–41].It is noteworthy that the adsorption effect of protein cannot be applied toward explaining all relative research results since protein is also found to accelerate Mg corrosion[27,28,34,42].The chelating effect of protein on cations(e.g.Mg2+and Ca2+)was emphasized decades ago[43–45].The consumption of Mg2+by binding in chelate complexes undoubtedly shifts the chemical equilibria towards Mg corrosion reaction.Willumeit[46]et al.found that the addition of protein led to a less dense but thick corrosion product layer on the corroded Mg and noted that the binding of ions to proteins was the key for corrosion product formation in protein-containing media.Heakal[47]et al.also mentioned the possible binding effect of protein.It is noteworthy that the protein may not only chelate the dissolved Mg2+from substrate,but also has affinity to other cations,like Ca2+,in the complex corrosive media.This may be another influencing factor for Mg corrosion,given the principal role of Mg2+and especially Ca2+in interface stabilization.It has been found that Tris and streptomycin,as Ca2+complexing agents,inhibit the formation of a protective layer on Mg surface in simulated body fluid[19,20,24].This promotes Mg corrosion.Chelating effect of proteins in the course of Mg corrosion is worthy of further investigations.In addition,recent works reported that local pH in the diffusion layer near the interface of Mg exposed to simulated body fluid(SBF)and HBSS,differs significantly from the pH monitored in bulk electrolytes[19,20].During the corrosion process of Mg,generated OH−in the diffusion layer participates in the formation of corrosion products(Mg(OH)2but also Ca-P containing precipitates)and hence influences the corrosion kinetics.Thus,the pH buffering effect caused by functional groups in the chemical structure of protein,should be taken into account[48].The buffered pH in protein-containing solution alters the test environment,which influences the corrosion behavior of Mg.At the present stage,understanding of the pH buffering effect of protein during Mg corrosion is still incomplete,especially,its influence on local pH at Mg/medium interface has not been reported.

Although,as previously mentioned,a number of publications investigated the influence of proteins on Mg corrosion,most of them focused on the interaction between proteins and substrates.Recently,the interaction between proteins and medium components was caught in a spot light of research interest.Yan[49]et al.reported that the synergistic effects of glucose and albumin decreased the corrosion rate of Mg in NaCl solution.However,as Höhn[5]et al.mentioned in a recent review that the understanding of protein interactions with Mg surfaces and their influence on Ca-P products formation in pseudophysiological solutions is still limited.Thus,we believe it is necessary to clarify the potential action mechanisms of protein on Mg corrosion in a complex pseudophysiological medium.It will be beneficial for the comprehensive understanding on the influence of protein on the reactions at the Mg/medium interface and its influence on the formation of Ca-P products.Furthermore,it will also help to elucidate the origin of different opinions about the protein influence on Mg corrosion in the previously published works.

In our recent works,a continuous electrochemical impedance spectroscopy(EIS)test has been proven to be an effective approach to investigatein-situthe formation of corrosion products and precipitation of protective layers on the Mg surface[20,24].In addition,a scanning ion-selective electrode technique(SIET)has been employed to monitor the localized pH values near the Mg/medium interface during the corrosion process[19,20].As compared to the pH monitoring of bulk electrolytes,the local pH has been proven to more accurately reflect the corrosion behavior of Mg[19,20].Although these tests in protein-containing solution cannot be maintained for a long time due to the risk of contamination,short-term tests can still provide information,which contributes to understanding the influence of protein.In this work,hydrogen evolution test,EIS monitoring and local pH measurement are complementary employed to investigate the corrosion behavior of Mg in albumin-containing HBSS during a short-term test.Possible mechanisms of protein interaction with Mg are discussed.This work improves our understanding of the corrosion behavior of Mg in a complex pseudophysiological environment.

Table 1The elemental composition of commercially pure Mg(CP Mg).

Table 2The composition of Hank’s balanced salt solution(HBSS).

2.Experimental

In this work,commercially pure Mg(CP Mg)was employed,its elemental compositions,obtained by optical discharge emission spectroscopy(SPECTROLAB with Spark Analyser Vision software,Germany),is listed in Table 1.Although CP Mg is not regarded as a potential implant material because of its unsatisfactory mechanical properties and corrosion resistance,it was selected as the research subject here to exclude the possible influence of alloying elements on corrosion and passivation mechanisms.In addition,due to the high risk of microbial contamination in protein-containing medium in the open environment,this fast corroding material(CP Mg)was employed here to shorten the test time and to obtain reliable results for analyzing the influence mechanisms of albumin.Small chips of CP Mg,which have a large surface area of 47.7±5.0 cm2/g,were used for the hydrogen evolution tests(electrolyte volume to sample surface area was 500mL/23.85 cm2=21mL/cm2).Eudiometers from Neubert-Glas,Germany(Art.Nr.2591–10–5)combined with an electronic balance(OHAUS,SKX series)were used for a custom-made automated recording of the water weight displaced by evolved hydrogen[24,50].Note that this is a closed setup,thus only the air initially entrapped(ca.70ml,equal to ca.15ml O2)in the test bottle and the air initially dissolved in the electrolyte can possibly contribute to the oxygen reduction reaction[51],which otherwise might lead to a significant underestimation of the total cathodic reaction.The detailed description of the hydrogen evolution test can be found in our previous work[24].

HBSS was used as the corrosive medium;its detailed composition is listed in Table 2.40 g L−1Bovine serum albumin(BSA,Carl Roth,8076.3)was added to the HBSS to investigate its influence.BSA-containing HBSS was referred to as HBSS+BSA.We choose to add 40 g L−1of protein based on the concentration of albumin in human serum(35–52 g L−1).The initial pH value of the HBSS was 7.2–7.3 and HBSS+BSA was 7.0±0.1.

A conventional three-electrode setup was used for the EIS measurements.A CP Mg sample with a surface area of 0.5 cm2was exposed to 350mL of electrolyte.A Pt wire coil was used as the counter electrode and a Ag/AgCl in saturated KCl solution was used as the reference electrode.The ground plates of 13×13×4mm(SiC paper up to 1200 grit)were used for the EIS measurements.Measurements were performed at open circuit potential(OCP)with an amplitude of 10mV RMS over a frequency range from 100kHz to 0.1Hz by a Gamry Interface 1000 potentiostat/galvanostat under constant stirring conditions(200rpm).

Scanning electron microscopy(TESCAN,Vega3 SB)at accelerating voltage of 15kV was employed to observe the corrosion products morphology after a 24h immersion in HBSS and HBSS+BSA.The elemental compositions of the corrosion products was tested by energy dispersive X-ray spectrometry(EDS).

Local pH measurements were performed in the HBSS electrolyte,in the diffusion layer directly adjacent to the surface of the corroding magnesium.The probes were positioned 50±5μm above the sample surface.The measurements were performed under hydrodynamic conditions(Medorex TL15E peristaltic pump)at a flow rate of 1.0 mL min−1assuring full electrolyte renewal in the cell every 5min.A commercial instrument for a scanning ion-selective electrode technique(SIET from Applicable Electronics)was used and controlled by LV4 software from Science Wares.A pH-sensitive microelectrode with a tip made of pH-sensitive glass from Unisense(pH-10,tip diameter was 10μm,tip length was 50μm)was employed.The samples were prepared by embedding CP Mg machined to a rod of 2±0.2mm in transparent epoxy resin(Buehler 20–3430064/20–3432–016)that,upon solidification,was ground up to 4000 grit.

The concentration of free Ca2+in the medium was measured potentiometrically,employing Ca2+-selective electrode(Perfection TM)and ion meter(SevenExcellence)both from Mettler Toledo.

The weight loss of CP Mg was measured three times after immersing the samples in HBSS and HBSS+BSA in a closed cell either under static or hydrodynamic conditions.In static conditions,the medium was changed once a day and the ratios of electrolyte volume to sample surface area(V/A)were 3:1 mL cm−2and 10:1 mL cm−2.In contrast,the medium was constantly renewed by a peristaltic pump at a rate of 0.5 mL min−1under hydrodynamic conditions.After 3 days of immersion,the samples were cleaned with chromic acid(200 g L−1)to remove the corrosion products,dried and weighed.

Fig.1.(a)Hydrogen evolution curves of CP Mg during 24h immersion in HBSS with and without 40 g·L−1 BSA;(b)Instantaneous hydrogen evolution rate derived from(a).

3.Results

3.1.Hydrogen evolution test

Fig.1 shows the cumulative hydrogen evolution curves and instantaneous hydrogen evolution rate of CP Mg during 24 hours’immersion in HBSS and HBSS+BSA.Initially high degradation rate of Mg stabilized at lower,quasi-constant values after 5(HBSS)to 10(HBSS+BSA)hours of immersion.It further slightly decreased after 15 h for both samples and remained quasi-constant until the end of the measurement.During this time,the components of HBSS,such as Ca2+,HCO3−and HPO42−/H2PO4−,were consumed[20]in the confined volume of the eudiometers.The presence of albumin in HBSS has a distinct effect on the corrosion behavior of Mg.During the first few hours of immersion,the Mg corrosion rate in HBSS+BSA was significantly higher than that in simple HBSS,but it gradually deceased over the immersion time.As shown in Fig.1(b),initially,the instantaneous hydrogen evolution rate of CP Mg in HBSS+BSA was very high,but it became lower than the corrosion rate in a simple HBSS after 13 h of immersion.Although the Mg corrosion rate in HBSS+BSA was lower than that in simple HBSS at the second half of the measurement,according to the volume of evolved hydrogen after 24 h,BSA accelerates the corrosion of CP Mg in HBSS.

Another noteworthy point is that the pH change of the bulk electrolyte during measurements in HBSS and HBSS+BSA.What needs to be emphasized is that the initial pH value of the two media is similar(shown in Fig.1(a),HBSS:7.2;HBSS+BSA:6.9–7.1),while the volume of evolved hydrogen in HBSS+BSA higher than that in HBSS(meaning that the amount of generated OH−was also bigger).In this case,the smaller pH change of the albumin-containing solution(approximately 2 units in HBSS+BSA;3 units in HBSS)indicates an additional pH buffering effect associated with albumin.Besides,the initial pH value of HBSS+BSA(6.9–7.1)is lower than that of HBSS(7.2–7.3).This is attributed to the weak acidity of albumin[52–54].

3.2.EIS evolution

Continuous EIS monitoring is an effective approach to investigate the formation of corrosion product layers and estimate their protective properties.Fig.2 shows the evolution of Bode plots measured over Mg during 24h immersion in HBSS with and without BSA.A striking feature of the Bode plots in simple HBSS was the presence of an additional time constant at the frequency of approximately 10kHz(marked by arrow in Fig.2(a,b)).A similar phenomenon has been emphasized in our previous research,performed in SBF[20,24].This time constant was attributed to the growth of a semi-protective continuous layer containing Ca2+,Mg2+,HPO42−/H2PO4−and HCO3−,which provides an important additional protection and slows down the corrosion rate.

However,the presence of BSA changed the Bode plots.During the first hour of immersion in HBSS+BSA,no additional time constant appeared in the high frequency range,as shown in Fig.2(c).Even after 24 h of immersion,the additional time constant was still not easily observable at the frequency of 10kHz,a typical manifestation frequency of the coprecipitation layer in the Bode plots obtained in HBSS+BSA(as shown in Fig.2(d),marked by a dotted arrow).

In simple HBSS electrolyte,a rapid formation(within less than 15min)of the co-precipitation layer provided protection against Mg corrosion,evidenced by a rapid growth of a high frequency time constant.However,in albumin-containing HBSS,its formation was delayed significantly.After 24 h in HBSS+BSA,the phase angle component of the Bode plot in high frequency domain has just taken the shape similar to that for the sample exposed to HBSS after 15 min.In addition,the evolution of the low frequency impedance of samples in the two media differs,while the initial Bode plots(upon first immersion)are very similar for the samples immersed in HBSS and HBSS+BSA.The low frequency value of impedance modulus increased during the first 3–7 h and then remained at the value of ca.18 kOhm·cm2until the end of the measurement for the sample immersed in HBSS.In contrast,the presence of albumin in HBSS caused the impedance to grow continuously but much lower during the entire test period,so that the low frequency impedance modulus was ca.4 kOhm by the end of the immersion test.These EIS results correlate with the hydrogen evolution tests to some extent.At the early stage of immersion,the limited adsorption of protein and the suppressed formation of the co-precipitation layer caused the rapid corrosion of CP Mg in the albumin-containing HBSS.With increased immersion time,the synergy between the protein adsorption and formation of the precipitation layer provided protection against the corrosion of CP Mg,slowing down the corrosion rate.

Fig.2.Bode plots evolution of CP Mg during 24h immersion in HBSS(a,b)and HBSS+BSA(c,d).

3.3.Corrosion morphology

The corrosion morphology of CP Mg after a 24h immersion in HBSS with and without albumin was investigated by SEM/EDS.Fig.3 shows the typical corrosion morphology of CP Mg in HBSS.The elemental compositions of selected points and regions in Fig.3 given by EDS are listed in Table 3.Similar morphologies have been reported in our previous works about Mg corrosion in SBF[20,24].As shown in Fig.3(a,b),a distinct feature of the corrosion products was that the products with high Ca-P content formed clusters(point 1 and 2 in Fig.3(b))and a compact product layer with low Ca-P content(region 3 and 4 in Fig.3(b)).The grinding scratches on Mg surface remained recognizable after the immersion,suggesting limited corrosion attack.

Table 3Elemental composition of corrosion products on the CP Mg given by EDS after 24h immersion in HBSS and HBSS+BSA.

However,the addition of albumin in HBSS causes different corrosion morphologies of CP Mg.The typical corrosion morphologies of CP Mg in HBSS+BSA are shown in Fig.4.The EDS results of the selected points and regions in Fig.4 are also listed in Table 3.

Fig.3.Typical corrosion morphology of CP Mg after 24h immersion in HBSS,SEM images.

The loose products with the layer structure in Fig.4 were associated with the fast corrosion rate of magnesium in HBSS+BSA at the early stage of the hydrogen evolution test.This feature is not typical for the corrosion morphology of magnesium after immersion in simple HBSS.After the comparison between Figs.3 and 4,it could be concluded that the addition of BSA inhibited the formation of product clusters with high Ca-P content.In contrast,the whole outer layer of the corrosion products(point 5 and 6 in Fig.4)contained a high Ca-P content,while the inner layer(point 7 in Fig.4)contained a low Ca-P content.In addition,the EDS mapping results of region 8 in Fig.4 supported the different elemental compositions between outer and inner layers of the corrosion products formed on corroded surface of Mg exposed to HBSS+BSA.

The cracked morphology of the corrosion products in HBSS+BSA correlates well with the EIS results.Even though this morphology can be related to the drying and shrinkage,the corrosion products formed in HBSS+BSA is more characteristic for thicker but looser as compared with dense corrosion products that shown in Fig.3.The less dense corrosion products in presence of albumin(Fig.4)verified the delayed and limited formation of a semi-protective continuous layer of Ca2+,Mg2+,HPO42−/H2PO4−and HCO3−(described by the change of a high frequency time constant in EIS evolution in Fig.2).

3.4.Evolution of local pH near the corroded Mg surface

All the above measurements(the hydrogen evolution tests,EIS and immersion tests)were performed under stirring conditions but without constant renewal of the medium.To mimic the real service environment of the implant materials,the local pH near the surface of corroded CP Mg was measured during the corrosion test while constantly refreshing of the media.

Fig.5 shows the point scan of local pH measurement on the surface of CP Mg immersed(under hydrodynamic conditions)in HBSS with and without BSA.The measurements captured the initial seconds of electrolyte contact with the sample surface.Technically,the pH micro-probe was positioned on the prepared sample and the measurement has started.Some seconds later,the electrolyte was added.This explains that the pH values in Fig.5 are recorded from 50 s point on.Local pH values increased from the initial value of ca.7.2 to above 9 within several seconds.The local pH in the HBSS increased sharper and to higher values,up to 9.9,while it reached 9.6 in HBSS+BSA electrolyte.The measurement points for each sample are marked with a white cross in Fig.5.The optical micrographs showed that the H2bubbles that appeared on the Mg surface in HBSS+BSA were smaller but more densely populated than those in simple HBSS,that testifies for higher rate of hydrogen generation,in line with similar conclusion reached by analyzing the hydrogen evolution by eudiometers,Section 3.1.The rapid corrosion of CP Mg was shown at the early stage of immersion in HBSS+BSA.Normally,rapid corrosion would indicate faster growing,high local pH[19,20].However,the result in this work was different,and it confirmed the assumption about the pH buffering effect of albumin-containing HBSS.Meanwhile,the pH buffering mechanism of albumin was different from that of synthetic buffer(i.e.Tris,HEPES).While Tris and HEPES were not efficient in buffering the local pH[19,20],albumin effectively buffered both the local and bulk pH due to sufficient solubility and effective adsorption on the Mg surface.

Fig.4.Typical corrosion morphology by SEM(a,b)and EDS mapping(c)of CP Mg after 24h immersion in HBSS+BSA.

In addition to the single point pH evolution,line scans of local pH measurements in HBSS with and without albumin were also performed under hydrodynamic conditions,Fig.6.In simple HBSS,the local pH measured at the distance of 50μm from the surface of sample was around 8.5 after first 10min of immersion and slightly decreased to around 8.3–8.4 after 30min of stabilization.Similarly,we have previously reported the fast stabilization of local pH in HBSS-like electrolytes over high purity Mg and Mg-2Ag,Mg-1.2Ca and Mg-Gd-Nd-Ca alloys[19,20].After 1 hour’s immersion,filiform corrosion occurred on the sample(marked by arrow in optical image in Fig.6(a)).With propagation of filiform corrosion from the edge to the center of the sample,the line scan after 1.5 hours’immersion showed higher local pH values in the area of filiform corrosion.However,due to the component replenishment by constant renewal of the medium and continuous formation of passive layer containing Ca2+,Mg2+,HPO42−/H2PO4−and HCO3−,the local pH decreased again afterwards.This indicates important protection effect of the components in HBSS on Mg corrosion.

In HBSS+BSA,as the immersion time elapses,the highest local pH decreased gradually to around 8.3 after 30min.The highest local pH value was maintained below 8.5 within 6 h of immersion.Although the local pH in HBSS and HBSS+BSA showed similar maximal value,the local pH in HBSS+BSA always showed a marked fluctuation during the first 3 h of immersion,which was different from the evolution of local pH in HBSS.This pointed out the different originations of low local pH in HBSS and HBSS+BSA.The formation of a dense co-precipitation layer of Ca2+,Mg2+,HPO42−/H2PO4−and HCO3−slowed down the corrosion rate of Mg in simple HBSS and led to the stable and low local pH value in HBSS.In HBSS+BSA,as discussed in Section 3.2,BSA delayed the formation of the aforementioned dense protective precipitation layer.The addition of BSA activated the corrosion reaction at the Mg/medium interface and led to faster corrosion of CP Mg in HBSS+BSA than that in simple HBSS(as shown in Fig.1).However,although the local pH did not show a higher value due to the pH buffering capacity of BSA,during the first 3 h of immersion,the distribution of local pH was uneven because of the activation effect of BSA on Mg corrosion(Fig.6(c,d)).As the immersion time increased,the combined albumin adsorption and corrosion products formed on the corroded Mg contributed to the stabilization of local pH in HBSS+BSA(as shown in Fig.6(d)after 4.5 hours’immersion).

Fig.5.Evolution of single point local pH at 50μm from the surface of CP Mg immersed in HBSS or HBSS+BSA during the initial stage of immersion.Performed under hydrodynamic conditions,flow rate 1.0mL/min.Optical micrographs show the visual appearance of two samples during the test.Multiple round features are the bubbles of evolved H2.

4.Discussion

4.1.The importance of component replenishment/medium renewal during the tests

During the corrosion tests under static conditions,the medium composition changes depending on the substrate dissolution and precipitation of corrosion products.The latter contains the components of the medium,thus depleting its ionic load.This is an important notion that,if left without deserved attention,might distort our understanding of the corrosion behavior of magnesium alloys.The importance of a constant medium composition during corrosion tests have been already pointed out previously[55,56].In spite of this,in most published works on immersion tests,the media were only refreshed after a fixed time interval(e.g.1 day or 2–3 days)instead of continuous renewal.In commonly referred standards about immersion tests,there are several recommended V/A values.For example,V/A of 1.25–6.0(depends on sample shape)in ISO 10993–12:2012[57]and 20–40 in ASTM G31–72(2004)[58]or ASTM G31–2012a[59].

Fig.7 compares the weight loss of CP Mg after 3 days of immersion in HBSS with and without BSA under static and hydrodynamic conditions.With the timely renewal of the media,the weight loss of CP Mg in HBSS decreased significantly.In addition,the stability of the media compositions has an impact on the reproducibility of the parallel tests to some extent.A stable medium composition has a significant influence on the corrosion test results of Mg in HBSS,especially for long-term measurements.However,the test conditions do not have significant influence on Mg corrosion in HBSS+BSA.In this work,BSA decreased the weight loss of CP Mg under static conditions with a V/A of 3:1,while it did not show apparent effect on the weight loss under static conditions with a V/A of 10:1 and increased the weight loss under hydrodynamic condition.These results,demonstrating susceptibility to experimental conditions,highlight the urgency of standardization for the test protocols of biodegradable metallic materials.

Fig.6.Local pH measurements(line scan)at 50μm from the surface of CP Mg immersed in(a,b)HBSS;(c,d)HBSS+BSA during immersion.All the measurements are performed under hydrodynamic conditions,flow rate 1.0 mL·min−1.Optical micrographs show the visual appearance of two samples during the test.Multiple round features are the bubbles of evolved H2.Dotted lines show the exact location of pH line scan.

Fig.7.Weight loss of CP Mg after 3 days immersion in HBSS with and without BSA under static conditions and hydrodynamic conditions.

To mimic the real service environment of implant materials,a timely,ideally constant renewal of the media should be performed.This should be considered as one of the essential factors when upgrading the setup for the Mg corrosion test.

4.2.Summary of the protein effect

In addition to the often-mentioned adsorption effect,the other two effects of albumin,including cation complexing and pH buffering,need to be discussed based on the results of this study.Complementary to the works previously published by different scientific groups,the real-time measurement methods used in this work provided detailed information about the influence of albumin on Mg corrosion in HBSS during up to 24 h period.Based on the above results and a number of relevant literatures,albumin interaction with corroding magnesium can be summarized to three factors:the barrier(adsorption),Ca2+/Mg2+chelation and pH buffering effects.Indeed,we did not provide direct evidence to prove the albumin adsorption in this work.However,the adsorption effect of proteins has been investigated and discussed for long time in the literatures.We deem that there is no need to provide even more results to support this recognized fact yet again.

With further research, biomaterials modeled based on porcupine quills could provide a new class of adhesive materials, says Robert Langer, the David H. Koch Institute Professor at MIT and a senior author of the study, which appears this week in the Proceedings of the National Academy of Sciences.

It is noteworthy that the hydrogen evolution rate in both cases became similar after 10h of immersion,which does not corresponded to the EIS results,that showed different low frequency impedance modulus value and time constant(s)after 10h immersion.This discrepancy is most likely attributed to the different test conditions(such as V/A ratio,shape of tested material)of these two test methods.However,it could be found that the results obtained from two methods showed similar trend.The rapid hydrogen evolution of CP Mg in HBSS+BSA for the first few hours correlates with the delayed and limited formation of co-precipitation layers composed of Ca2+,HPO42−/H2PO4−,HCO3−and Mg2+(as described by the high frequency response in impedance spectra).Then,with the increase of immersion time,the synergy between albumin adsorption and co-precipitation layer formation inhibits further corrosion of CP Mg in HBSS+BSA(demonstrated by the decreased rate of hydrogen evolution).The second stage was verified to some extent by the dimly appeared additional time constant at high frequency range(marked by a dotted arrow in Fig.2(d))and the continuous increase of low frequency impedance modulus value.However,the opposite result has been reported by Wang[60]et al.The causes of this discrepancy are apparent.Apart from the tested material,the presence of Tris/HCl in the media[60]also has a significant influence.It has been shown in the literatures that Tris(as well as HEPES)accelerate degradation of Mg[20,61].

As presented in our previous works[19,20,24],compared with simple NaCl solution,SBF,HBSS,MEM or similar media possess pH buffering effects that originate from carbonate and phosphate components.The pH buffer capacity in these SBF-like electrolytes was attributed to the buffering effect of several ions(i.g.carbonates and phosphates)and the decreased concentration of generated OH-induced by the suppressed corrosion owing to the formed protective products layer.It is different from the buffering mechanism of common pH buffers,such as Tirs,HEPES or HCO3−/CO2.Combining the hydrogen evolution and local pH results,at the initial stage of immersion in albumin-containing HBSS,the rapid corrosion of Mg did not lead to a high local pH value.This is because of the strengthened buffer capacity of HBSS by adding albumin.Approximately one-third of the amino acid functional groups in proteins can provide or accept hydrogen ions[48].More specifically,the sidechains of glutamic and aspartic acid are comprised of carboxyl groups;lysine sidechains include an amino group;arginine sidechains contribute guanidyl groups;phenolic groups,sulfhydryl groups,imidazolyl groups can be found in the sidechains of tyrosine,cysteine and histidine,respectively.Almost all proteins contain terminalα-amino andα-carboxyl groups on polypeptide chains.All the above functional groups make protein a strong pH buffer,influencing the pH value of media during the corrosion test and then influencing the corrosion behavior of Mg.

The basic corrosion reaction of magnesium in a pseudophysiological environment can be described by the following equations[20,51,62–67]:

Dissolution of the matrix:

Anodic reaction

Table 4Concentration of free Ca2+in HBSS and HBSS+BSA albumin measured potentiometrically with Ca2+-selective electrode.The nominal concentration of Ca2+in HBSS is 1.3mM.

Combining Eq.(2)–(4)and the obtained results,it can be concluded that the addition of albumin shifts the chemical equilibria of these reactions.More specifically,the consumption of OH−caused by the pH buffering effect of albumin shifts the chemical equilibria of the reactions described by Eq.(2)–(3)to the right.Similarly,it also influences the formation of the protective co-precipitation product layer(described by MgmCan(PO4)x(CO3)y(OH)z),as shown in Eq.(4).The concentration of free Ca2+(measured potentiometrically with Ca2+-selective electrode)in HBSS with and without albumin is listed in Table 4.The addition of albumin significantly reduces the concentration of free Ca2+in HBSS(from 1.23±0.06mM to 0.81±0.07mM).This is a direct evidence for binding Ca2+by albumin,as suggested earlier by different methods[44,45].The complexation between albumin and metal cations shifts the chemical equilibrium of the reactions described by Eq.(3)to the left thus delaying the formation of the protective co-precipitation layer MgmCan(PO4)x(CO3)y(OH)z.We would like to emphasize that the decreased concentration of free Ca2+is not the only cause of the delayed formation of dense and protective coprecipitation layer on Mg surface in HBSS+BSA.The adsorption of albumin changes the surface state of Mg,which probably also influences the formation reaction of the product layer at the interface.Although the effect of protein may be caused by a number of factors,there is no doubt that the presence of albumin finally delays formation of dense and protective co-precipitation layer.

Fig.8 depicts the mentioned mechanisms of albuminmagnesium interaction at the initial stages of immersion in HBSS.The barrier(adsorption),chelating and pH buffering effects are summarized.Based on the hydrogen evolution results in this work,at the initial stage of immersion in albumincontaining HBSS,the absence or delayed formation of coprecipitation layer leads to the fast corrosion of CP Mg.With increasing immersion time,the adsorption effect of albumin appears gradually.The barrier effect caused by the combination of adsorbed protein and tardily formed inorganic corrosion products slows down the corrosion.However,the above statements may not be regarded as the general description of the corrosion process of Mg in protein-containing pseudophysiological media.The question of whether the addition of protein accelerates or inhibits the corrosion of Mg cannot be answered unambiguously.As previously reported by other research groups,protein was found to inhibit Mg corrosion at the initial stage of test and accelerate it after long time immersion[32,33].Even in this work,BSA showed different influence on the corrosion rate of Mg in different measurements.BSA was found to accelerate Mg corrosion based on 24 h hydrogen evolution test,while it showed different influence on Mg corrosion in 3 days weight loss measurements under various test conditions.It is clear that the testing media,tested materials,and test conditions significantly affect the final results of the albumin influence on Mg corrosion in a pseudophysiological environment.In any case,we do believe that the ultimate effect of protein on Mg corrosion originates from the synergy of three aforementioned mechanisms.It is reasonable to assume that the dominating factor(s)in these three mechanisms might differ depending on experimental conditions.

Fig.8.Schematic illustration of influential factors of albumin on corrosion of CP Mg during short corrosion test in HBSS+BSA.

4.3.Outlook

Although the findings in this work were obtained from the tests with CP Mg,we believe they can be extended,at least in some parts,to other Mg alloys.Undoubtedly,it can be concluded that the addition of albumin significantly changes the corrosive environment and corrosion behavior of Mg.However,it does not mean that protein should be introduced into all corrosive media when studying magnesium for biomedical applications.For a long-term test in an open environment,protein-containing media may not be suitable.Specifically,after several days of immersion,the acidification of media induced by microbial proliferation significantly changes the test environment under unsterile conditions.That may also be one of the causes of the discrepancies between results in previously published works.Additionally,in some published works[30,68–70],the source of albumin is not mentioned.Fig.9 shows cumulative hydrogen evolution curves of CP Mg in HBSS and HBSS with two different albumins:40 g·L−1,BSA(from Carl Roth,Ref.8076.3)and chicken egg albumin(CEA)(from Alfa Aesar,Ref.A16951).Although both of BSA and CEA significantly accelerated hydrogen evolution at first 2 h,they showed different acceleration capacity.According to the volume of evolved hydrogen after 24 hours’test,CEA slightly inhibited Mg corrosion,while BSA accelerated Mg corrosion.Based on this finding,it is evident that different sources of albumin may be one of the factors that influence the research results.

Fig.9.Hydrogen evolution curves of CP Mg during 24h immersion in HBSS,HBSS+BSA and HBSS+CEA.

As Höhn[5]et al.outlined in a recent review,the understanding of protein interactions with Mg surfaces and their influence on Ca-P product formation is still limited.We hope the results of this work open yet another facet of Mg-albumin interaction and improve the general understanding of these complex processes.

5.Conclusion

In this study,we investigated the corrosion behavior of CP Mg during up to 24 h in albumin-containing HBSS.The conclusions are listed below.

(1)The presence of BSA in HBSS buffers the pH of the electrolyte during the corrosion test.The bulk pH of HBSS+BSA only increases by approximately 2 units after the hydrogen evolution test,while the pH value increases by 3 units after immersion in simple HBSS.Even though the corrosion of CP Mg in HBSS+BSA is much faster,at the initial stage of immersion(first 1–2 min),the increase of local pH in HBSS+BSA is slower than that in HBSS.

(2)Although the local pH in HBSS and HBSS+BSA showed similar maximal value(8.5),the local pH in HBSS+BSA always showed a marked fluctuation during the first 3 h of immersion in contrast to stable local pH in HBSS.This indicates that different factors control low local pH in HBSS and HBSS+BSA.

(3)The addition of BSA decreases the concentration of free Ca2+in HBSS by ca.35%(from 1.23 to 0.81mM)due to a chelating effect.During the first hours of immersion,the presence of BSA slows down the formation of a co-precipitation protective layer on the corroded CP Mg and leads to rapid corrosion,as shown by the hydrogen evolution test.

(4)Influence of BSA on the corrosion of Mg is susceptible to experimental conditions.Importance of replenishment of medium components during the corrosion tests was emphasized here and should be taken into account.

(5)The influence of albumin on Mg corrosion can be summarized in three aspects,namely,adsorption(the formation of a barrier layer),Ca2+/Mg2+chelation and pH buffering effects.The synergy of these three factors ultimately affects the corrosion of Mg in protein-containing HBSS.

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.

Acknowledgment

Mr.Di Mei and Mr.Cheng Wang thank China Scholarship Council for the award of fellowship and funding(No.201607040051;201806310128).“MMDi”IDEA project funded by HZG is gratefully acknowledged.The technical support of Mr.Volker Heitmann and Mr.Ulrich Burmester during this work is gratefully acknowledged.