Department of Manufacturing Engineering,Annamalai University,Annamalainagar,608 002 Chidambaram,Tamil Nadu,India

Comparison of the corrosion behaviour of AZ31B magnesium alloy under immersion test and potentiodynamic polarization test in NaCl solution

D.Thirumalaikumarasamy*,K.Shanmugam,V.Balasubramanian

Department of Manufacturing Engineering,Annamalai University,Annamalainagar,608 002 Chidambaram,Tamil Nadu,India

This paper reports the comparative evaluation of the corrosion behaviour of AZ31B magnesium alloy under immersion and potentiodynamic polarization test in NaCl solution at different chloride ion concentrations,pH value and exposure time.The specimens were exposed to immersion and polarization environments in order to evaluate their corrosion rates.Empirical relationship was established to predict the corrosion rate of AZ31B magnesium alloy.Three factors,f i ve level,central composite rotatable design matrix was used to minimize the number of experimental conditions.Response surface methodology was used to develop the relationship.The developed relationship can be effectively used to predict the corrosion rate of AZ31B magnesium alloy at 95%conf i dence level for both the testing.This research work proves a better corrosion resistance of AZ31B magnesium alloy at the alkaline solution than the acidic and the neutral solutions,moreover,low corrosion rate was found at low concentrated solution and higher exposure time respectively.

Magnesium alloy;Immersion tests;Potentiodynamic polarization tests;Response surface methodology;Corrosion rate

1.Introduction

The desire to use lightweight metallic alloys in the automobile and aerospace industries has increased in recent years as the search for lightweight solutions has become amplif i ed. Magnesium alloys are one of these lightweight metallic alloys currently being investigated,because of its low density,1.74 g/ cm3,and high mechanical stiffness.The mechanical benef i ts of magnesium,however,are contrasted by a high corrosion rate as compared to aluminium or steel.Because of magnesium’s electrochemical potential,as illustrated in the galvanic series,it corrodes easily in the presence of seawater.The high corrosion of magnesium has relegated the alloy to use in areas unexposed to the atmosphere,including car seats and electronic boxes[1,2].However,the corrosion resistance of the Mg-based alloys is generally inadequate due to the low standard electrochemical potential-2.37 V compared to the SHE (Standard Hydrogen Electrode)and this limits the range of applications for Mg and its alloys.Therefore,the study of corrosion behaviour of magnesium alloys in active media, especially those containing aggressive ions,is crucial to the understanding the corrosion mechanisms,and hence,to improving the corrosion resistance under various service conditions.The reason for the less corrosion resistance of magnesium and its alloys results primarily from two mechanisms:(i)oxide f i lms forming on the surface is not perfect andprotective;(ii)galvanic or bi-metallic corrosion can be caused by impurities and secondary phases[3].

NomenclatureP pH valueC chloride ion concentration,moleT time,hrs CR corrosion rate,mm/year RSM response surface methodology

This research focused on comparing immersion testing with potentiodynamic polarization testing,which are the two main techniques for corrosion studies,in an effort to expose the magnesium alloy to environments similar to those environments experienced by automotive engine blocks[4].It is well known that Mg alloys are susceptible to corrosion such as pitting and stress cracking corrosion(SCC).Major studies shows that the SCC susceptibility of Mg alloys is increased in solutions containing chloride[5].

The galvanic couples formed by the second phase particles and the matrix are the main source of the localized corrosion of magnesium alloys[6].The corrosion of AZ31 magnesium alloy in simulated acid rain solution is controlled by the rate of anodic dissolution and hydrogen evolution,and the corrosion rate of AZ31 increases with increasing concentration of Clion[7].The corrosion attack of Mg and its alloy in dilute chloride solutions depends on both Al content and alloy microstructure[8].Yingwei song et al.[9],investigated the effect of second phases on the corrosion behaviour of wrought Mg-Zn-Y-Zr alloy and they found that the increase of exposure time,the second phases can promote the corrosion rate signif i cantly and cause pitting corrosion.Rajan Ambat et al.[10],studied the evaluation of micro structural effects on corrosion behaviour of AZ91D magnesium alloy and they reported that size and morphology of β phase and coring were found to have signif i cant inf l uence on corrosion behaviour of AZ91D alloy.Pardo et al.[11],explored the inf l uence of microstructure and composition on the corrosion behaviour of Mg/Al alloys in chloride media and it was found that the aluminium enrichment on the corroded surface for the magnesium alloy,and the β-phase(Mg17Al12),which acted as a barrier for the corrosion progress for the magnesium alloys. The corrosion product consisted of magnesium hydroxide, fallen β particles and magnesium-aluminium oxide;the amount of each component was found to be a function of chloride ion concentration and pH[12].

From the literature reviews[6-12],it is understood that most of the published information on corrosion behaviour of Mg alloys were focused on general corrosion of magnesium alloys.Moreover,there is no literature available related to comparative study of corrosion behaviour of magnesium alloys under immersion and potentiodynamic polarization test in NaCl solution.Hence,the present investigation was carried out to investigate the comparative evaluation of corrosion behaviour of AZ31B magnesium alloy under immersion testand polarization test and henceforth,to develop an empirical relationship to predict the corrosion rate of AZ31B magnesium alloy using response surface methodology.

Table 1aChemical composition(wt%)of AZ31B Mg alloy.

2.Experimental work

The base material,AZ31B magnesium alloy used in this investigation was an extruded cylindrical rod of 16 mm diameter.The chemical composition and mechanical properties of the base material are presented in Table 1(a)and(b), respectively.The specimens were cut to the dimensions of 16 mm × 4 mm to evaluate the corrosion rate by immersion corrosion test method.The corrosion test specimens were ground with 500#,800#,1200#,1500#grit SiC paper washed with distilled water and dried by warm f l owing air.The optical micrograph(OM)of base metal is shown in Fig.1(a)and it basically contains equiaxed grains of 10 μm diameter (average).Fig.1(b)shows scanning electron microscopy (SEM)image of AZ31B magnesium alloy.

2.1.Finding the limits of corrosion test parameters

From the literature[6-12],and the previous investigation carried out in our laboratory[13]the predominant factors that have a greater inf l uence on the corrosion behaviour of magnesium alloy were identif i ed.They are:(i)pH value of the solution,(ii)chloride ion concentration and(iii)exposure time.Large numbers of trial experiments were conducted to identify the feasible testing conditions using AZ31B magnesium alloy under immersion tests and potentiodynamic polarization test conditions.The following inferences were obtained:

1.If the pH value of the solution was less than 3,the change in chloride ion concentration did not considerably affect the corrosion.

2.If the pH value was in between 3 and 12,there was inhibition of the corrosion process and stabilization of the protective layer.

3.If the pH value was greater than 12,then blocking of further corrosion by the active centres of protective layer.

4.If the chloride ion concentration was less than 0.2 M,then the visible corrosion did not occur in the experimental period.

Table 1bMechanical properties of AZ31B Mg alloy.

Fig.1.Micrograph of AZ31B Mg alloy.

5.If the chloride ion concentration was in between 0.2 M and 1 M,then there was a reasonable f l uctuation in the corrosion rate.

6.If the chloride ion concentration was greater than 1 M, then the rise in corrosion rate may hesitate and decrease a little.

7.If the exposure time was less than an 1 h,then the surface was completely covered with thick and rough corrosion products.

8.If the exposure time was in between 1 and 8 h,then the tracks of the corrosion can be predicted.

9.If the exposure time was greater than 8 h,then the tracks of corrosion f i lm were diff i cult to identify.

2.2.Developing the experimental design matrix

As the range of individual factor was wide,a central composite rotatable three-factor,f i ve-level factorial design matrix was selected.The experimental design matrix consisting 20 sets of coded condition and comprising a full replication threefactor factorial design of 8 points,6 star points,and 6 centre points was used.Table 2 represents the range of factors considered,and Table 3 shows the 20 sets of coded and actual values used to conduct the experiments.The upper and lower limits of the parameters were coded as+1.682 and-1.682, respectively.Thus,the 20 experimental runs allowed for the estimation of the linear,quadratic,and two-way interactive effects of the variables.The method of designing such a matrix is dealt with elsewhere[14].The coded values for intermediate levels can be calculated from the relationship.

where,

Xiis the required coded value of a variableXandXis any value of the variable fromXmintoXmax.

Xminis the lower level of the variable.

Xmaxis the upper level of the variable.

2.3.Recording the responses(corrosion rate evaluation)

Solution of NaCl with concentrations of 0.2 M,0.36 M, 0.6 M,0.84 M,and 1 M were prepared.The pH value of the solution was maintained as pH3,pH4.82,pH7.5,pH10.18,& pH12 with concentrated HCl and NaOH respectively.The pH value was measured using a digital pH metre.Two different tests environments were used in this study:immersion test and potentiodynamic polarization tests.For immersion testing,the test method consists of immersing the specimens in an aquarium f i lled with NaCl solutions at room temperature as per the ASTM standard G31-72[15].

The corrosion rate of the AZ31B alloy specimen was estimated by weight loss measurement.The original weight (w0)of the specimen was recorded and then the specimen was immersed with the solution of NaCl for different immersion times of 1,2.42,4.5,6.58 and 8 h.For both the testing,the corrosion products were removed by immersing the specimens for one minute in a solution prepared by using 50 g chromium trioxide(CrO3),2.5 g silver nitrate(AgNO3)and 5 g barium nitrate(Ba(NO3)2)for 250 ml distilled water.Finally,the specimens were washed with distilled water,dried and weighed again to obtain the f i nal weight(w1).The weight loss (w)can be measured using the following relation,

Table 2Important factors and their levels.

Table 3Design matrix and experimental results.

where,

w=weight loss in grams.

wo=original weight before test in grams.

w1=f i nal weight after test in grams.

For potentiodynamic polarization tests,the specimens were treated with metallographic polishing prior to each experiment,followed by washing distill water and acetone, and f i nally dried in warm air.The polarization measurements were carried out in corrosion test cell containing 500 ml solution.Solution of NaCl with concentrations of 0.2 M, 0.36 M,0.6 M,0.84 M,and 1 M were prepared.The pH value of the solution was maintained as pH3,pH4.82,pH7.5, pH10.18,& pH12 with concentrated HCl and NaOH respectively.The pH value was measured using a digital pH metre.To evaluate the corrosion tested specimen with the method as per ASTM G5-94[16].An electrochemical polarization experiment was carried out using a potentiostat GILL AC.The electrodes for this purpose were prepared by connecting a wire to one side of the sample that was covered with cold setting resin.The other side of the specimen, whose area is 1 cm2,was exposed to NaCl solution with different pH and chloride ion concentration for different exposure times of 1,2.42,4.5,6.58 and 8 h.Electrochemical corrosion tests were carried out using a computer controlled Gill AC potentiostat/frequency response analyser to evaluate the deterioration process of AZ31B magnesium alloy specimens in NaCl solutions.A typical three electrode cell,with a saturated Ag/AgCl(saturated with KCl)as reference electrode,a platinum mesh counter electrode and the AZ31B magnesium alloy as the working electrode(1 cm2exposed area)were used in the tests.The specimens were exposed in the test solution,and a polarization scan was carried out towards more noble values at a rate of 300 mV/min after allowing a steady state potential to develop.The scan scope was set from-600 mV to+600 mV vs.OCP.All electrochemical tests were conducted in triplicate in order to ensure the reproducibility of results.The corrosion potential was developed and observed from the open circuit potential. Furthermore,corrosion current densities for all tests were measured directly from the tangent slope and it was recorded.

To calculate the alloy equivalent weight,the following approach may be used.Consider a unit mass of alloy oxidized.The electron equivalent for 1 g of an alloy,Qis then:

where,

ni-the valence of theithelement of the alloy;

fi-the mass fraction of theithelement in the alloy;Wi-the atomic weight of theithelement in the alloy.

Therefore,the alloy equivalent weight,EW,is the reciprocal of this quantity,

The corrosion rate can be calculated using Faraday’s Law in terms of penetration rates as per ASTM G102-89,

where,

K-Corrosion constant(K=0.00327 if the corrosion rate in mm/yr);

Icorr-Current density in μA/cm2;

EW-Equivalent weight of the alloy;

ρ -Density of the AZ31B alloy(1.74 g/cm3)

Micro structural examination of the corroded specimens was carried out using a light optical microscope(VERSAMET-3) incorporated with image analysing software(Clemex vision). Corroded surfaces were evaluated by scanning electron microscopy(SEM)using a JEOL JSM-6400 microscope equipped with Oxford Link energy-dispersive X-ray(EDX)microanalysishardwareinordertostudythemorphologyandevolutionof corrosion products formed on the surface of the material.Phase analysis of the corroded surfaces were performed on Philips 3121X-raydiffractrometerusingCuKα radiationwhichwasset at 40 kV and 20 mA for the XRD analysis and the data were recorded in the 2θ range 10°to 80°in steps of 2°/min.

3.Developing an empirical relationship

In the present investigation,to correlate the immersion tests and potentiodynamic polarization test parameters with the corrosion rate of AZ31B specimens,a second order quadratic model was developed.The response(corrosion rate)is a function of pH value(P),chloride ion concentration(C), exposure time(T)and hence it can be expressed as

The empirical relationship chosen includes the effects of the main and interaction effect of all factors.The construction of empirical relationship and the procedure to calculate the values of the regression coeff i cients can be referred elsewhere [17].In this work,the regression coeff i cients were calculated with the help of Design Expert V 8.1 statistical software.After determining the coeff i cients(at a 95%conf i dence level),the f i nal empirical relationship was developed using these coeff i cients.The f i nal empirical relationship to estimate the response is given below:

For immersion tests:

For potentiodynamic polarization tests:

3.1.Checking the adequacy of the model for immersion testing

The Analysis of Variance(ANOVA)technique was used to f i nd the signif i cant main and interaction factors.The results of second order response surface model f i tting in the form of Analysis of Variance(ANOVA)are given in the ‘Table 4’.The determination coeff i cient(r2)indicated the goodness of f i t for the model.The ModelF-value of 14.64 implies the model is signif i cant.There is only a 0.01%chance that a “ModelFValue” this large could occur due to noise.Values of“Prob>F”less than 0.0500 indicates model terms are signif i cant.In this caseP,C,T,P2,C2,T2,are signif i cant model terms.Values greater than 0.1000 indicate the model terms are not signif i cant.If there are many insignif i cant model terms (not counting those required to support hierarchy),model reduction may improve the model.The “Lack of FitF-value”of 5.04 implies the Lack of Fit is not signif i cant relative to the pure error.There is a 5.02%chance that a “Lack of FitF-value”this large could occur due to noise.Non-signif i cant lack of f i t is good.The “PredR-Squared” of 0.9420 is in reasonableagreement with the “AdjR-Squared”of 0.9754. “Adeq Precision”measures the signal to noise ratio.Pratio greater than 4 is desirable.Our ratio of 12.764 indicates an adequate signal. Each of the observed values compared with the experimental values shown in the ‘Fig.2’.

Table 4ANOVA test results for immersion testing.

Fig.2.Correlation graph for response(corrosion rate)for immersion tests.

3.2.Checking the adequacy of the model for potentiodynamic polarization testing

The Analysis of Variance(ANOVA)technique was used to f i nd the signif i cant main and interaction factors.It was thus conducted in this study in order to determine whether the above mentioned second order polynomial equation was signif i cant f i t with the experimental results.The results of the second order response surface model f i tting as Analysis of Variance(ANOVA)are given in the ‘Table 5’.The determination coeff i cient(r2)indicated the goodness of f i t for the model.The ModelF-value of 71.14 implies the model is signif i cant.There is only a 0.01%chance that a “ModelFValue” this large could occur due to noise.Values of“Prob>F”less than 0.0500 indicate model terms are signif i cant.In this caseP,T,C,PC,TCare signif i cant model terms.Values greater than 0.1000 indicate the model terms were not signif i cant.If there are many insignif i cant model terms(not counting those required to support hierarchy), model reduction may improve the model.The “Lack of FitF-value” of 1.89 implies the Lack of Fit was not signif i cant relative to the pure error.There was a 25.11%chance that a“Lack of FitF-value”this large could occur due to noise.Nonsignif i cant lack of f i t is good.The “PredR-Squared”of 0.9128 is in reasonable agreement with the “AdjR-Squared” of 0.9568.“Adeq Precision”measures the signal to noise ratio.Pratios greater than 4 are desirable.Our ratio of 29.887 indicates an adequate signal.All of this indicated an excellent suitability of the regression model.Each of the observed values compared with the experimental values shown in the‘Fig.3’.

4.Discussion

4.1.Effect of pH value on corrosion rate

Table 3 shows the corrosion rates obtained from immersion test and potentiodynamic polarization tests at different pH value,chloride ion concentration and exposure time.From this table,the following points can be inferred.The pH value has an inversely proportional relationship with the corrosion rate; i.e.,if the pH value increases,the corrosion rate decreases [18].At all pH values,the specimen exhibited a rise in corrosion rate with the decrease in pH value.In the neutral pH, the corrosion rate remained constant approximately and a comparatively low corrosion rate was observed in alkaline solutions.Also,there exists a correlation between pH,chloride ion concentration and exposure time.The rate of corrosion increases with the increase of chloride ion concentration and found a corrosion resistant with the decrease in chloride ion concentration.At lower pH values,the specimen exhibited a rise in corrosion rate with an increase in chloride ion concentration.But the quantity of this rise was different in such away that,the change in chloride ion concentration at lower concentrations affected the corrosion rate much more as compared to that of higher concentration.It showed that with the increase in chloride ion concentration,the rising rate at corrosion rate decreased that is,the inf l uence of chloride ion concentration was much lower at higher concentrations. Consequently,the rate of corrosion decrease slightly with the increase in exposure time.It resulted from the increase in hydrogen evolution with an increase in exposure time;the idea attributed to the corrosion occurring over an increasing fraction of the surface,which is the insoluble corrosion product [19].The insoluble corrosion product on the surface of the alloy could slow down the corrosion rate.

Table 5ANOVA test results for potentiodynamic polarization testing.

Fig.3.Correlation graph for response(corrosion rate)for potentiodynamic polarization tests.

Fig.4.Effect of pH on corrosion morphology of immersion test specimen.

Fig.4 shows the effect of pH on corrosion morphology of AZ31B magnesium alloy exposed in 0.6 M concentration of NaCl for 4.5 h with different pH values of pH3,pH7.5 and pH12 for immersion testing.From the f i gure,it can be seen that,at lower pH values,surface of the AZ31B magnesiumalloy was completely corroded,corrosion pits almost distribute on the entire surface.Visible corrosion became slighter as the pH value increased.In the solution with pH3, the bubbles overf l owed and broke away from the specimen surface.The metallic lustre of the specimen surface was also gradually lost,and corrosion became much more severe than in other specimens.Pits are also observed which are suspected to be the sites with the β phase.Mg17Al2has a higher standard potential and exhibits a more passive behaviour over a wide pH range in chloride solutions than either Al or Mg.Due to selective attack along the β phase networks,they are gradually attacked and peel from the surface forming the pits is expected to take place at these sites.It meant that the pH value was one of the major factors of corrosion rate[20].

Fig.5 shows the effect of pH on the corrosion morphology and pit morphology of the specimen exposed in 0.6 M concentration of NaCl for 4.5 h with different pH values of pH3, pH7.5 and pH12 for potentiodynamic polarization tests.From the f i gure,it was found that at higher pH values,pit corrosion has been observed only at the edges of the surface and also several small corrosion pits formed on the surface of the AZ31B magnesium alloy.Nevertheless it still suffered much less corrosive attack when compared to AZ31 specimen at lower pH values.The surface of the specimen exposed to low pH solution constituting morecorrosion products,thus corrosion occurs severely which would enhances the corrosion rate.The density of the pit formed in exposing lower pH (acidic)solution is quite high,comparing with the neutral and alkaline solution[21].The diameter of the pit increases with the decrease of pH value.Thus the pit depth increases with the decrease of pH as shown in Fig.5.This is due to the corroding surface which had a higher rate of hydrogen evolution.This means that,if a corroding area is adjacent area to a noncorroded area,there will be a galvanic cell causing the galvanic acceleration of the corrosion rate of the non-corroded area. Thus, once the corrosion starts,there is anelectrochemical driving force for the spread of the corrosion across the surface.This is indeed what is observed experimentally.The galvanic acceleration of the corrosion across the non-corroding areas is balance by the galvanic protection of the corroded areas,so that the corrosion tends to be rather shallow in the corroded areas.

Fig.5.Effect of pH on corrosion morphology and pit depth of potentiodynamic polarization test specimen.

Fig.6 shows the effect of pH on Tafel plots,from the f i gure it could be inferred that with the decrease in the pH value of the solutions,the anodic curve of the materials showed a shift to higher current density values.The inf l uence of pH on corrosion needs to take into account the magnesium E-pH diagram predicts there should be no f i lm on a magnesium surface in a solution with a pH lower than 10.5 because Mg(OH)2is not stable under such conditions.However,it is not thermodynamically stable at low pH values,the dissolution kinetics may be slower and a surface f i lm may be formed provided the dissolution kinetics is slower than the formation kinetics.It was observed that the corrosion rate usually increased with the decrease in the pH value of the solutions [22].The dissolution of magnesium in aqueous solutions proceeds by the reduction of water to produce magnesium hydroxide.The reduction process was mainly water reduction, thus forming a Mg(OH)2protective layer.Higher the pH value favours the formation of Mg(OH)2which protects the alloy from corrosion.

4.2.Effect of chloride ion concentration on corrosion rate

Fig.6.Effect of pH on Tafel plots.

The inf l uence of chloride ion concentration on corrosion rates of AZ31B magnesium alloy for both immersion and potentiodynamic polarization testing in NaCl solution are displayed in ‘Table 3’.It is seen that the corrosion behaviour is consistent with the current understanding that the corrosion behaviour of magnesium alloys is governed by a partially protective surface f i lm with the corrosion reaction occurring predominantly at the breaks or imperfections of the partially protective f i lm.This is consistent with the known tendency of chloride ions to cause f i lm break down,and the known instability of Mg(OH)2in solutions with pH less than 7.5. However it was observed that,with the increase in chloride ion concentration,the rising rate at corrosion rate decreased.The increase in corrosion rate with increasing chloride ion concentration may be attributed to the participation of chloride ions in the dissolution reaction.Chloride ions were aggressive for magnesium.The adsorption of chloride ions to oxide covered magnesium surface transformed Mg(OH)2to easily soluble MgCl2thus destroying the compactness of the corrosion product f i lm and resulting in pitting corrosion[23].The anodic specimen exhibited a rise in corrosion rate with increase in Cl-concentration and thus the change of Cl-concentration affected the corrosion rate much more in higher concentration solutions than that in lower concentration solutions.When more Cl-in NaCl solution promoted the corrosion,the corrosive intermediate(Cl-)would be rapidly transferred through the outer layer and reached the substrate of the alloy surface.Hence,the corrosion rate was increased[24].

The effect of chloride ion concentration on corrosion morphology and pit depth of the specimen exposed in NaCl solution with different chloride ion concentrations for both immersion and potentiodynamic polarization testing are presented in ‘Figs.7 and 8’respectively.At lower chloride ion concentrations,less corrosion pits were formed on the surface of the AZ31B magnesium alloy.The pit formation depends on the change in the chloride ion concentration of the solution. Here,with the increase of chloride ion concentration of the solution,the density of the pit increases.It is found that the whole α-grains undergo a majority of corrosion and the border of the corroded area is mostly surrounded by the β-phase.It might be attributed to the more negative potential of the αphase than that of the β-phase in NaCl solution.There is a tendency for the corrosion rate of the α-phase to be accelerated by micro-galvanic coupling between the α-phase and the βphase[25].The increasing trend of the pit depth with the increase of chloride ion concentration is attributed to the attack of Cl-ions on the surface leading to the anodic dissolutions of Mg(Fig.8).The chloride ion concentration increases,some obvious pits appear on the surface of the specimen.Thereby, increasing the corrosion rate with the increase of chloride ion concentration.

Fig.9 shows the effect of chloride ion concentration on Tafel plots.From the f i gure,it was observed that with the increase of chloride ion concentration of the solutions,the anodic curve of the materials showed a shift to higher current density values and corrosion potential shifted to more negative (active)values with the increase in chloride ion concentration, which may explained by the adsorption of chloride ion on the alloy surface at weak parts of oxide f i lm.Thus the increase in corrosion rate with the increasing chloride ion concentration attributed the participation chloride ions in the dissolution reaction[26].

4.3.Effect of exposure time on corrosion rate

Fig.7.Effect of chloride ion concentration on corrosion morphology of immersion test specimen.

Table 3 depicts the inf l uence of the exposure time on the corrosion rate of AZ31B magnesium alloy obtained from immersion test and potentiodynamic polarization test in NaCl solution.From the table,it can be seen that the corrosion rate decreased with the increase in exposure time.The increase in exposure time enhanced the tendency to form the corrosion products,which accumulated over the surface of the samples. These corrosion products which in turn depressed the corrosion rate due to the passivation in the medium immersion[27]. It results that there was an increase in hydrogen evolution with the increasing exposure time,which tends to increase the concentration of OH-ions strengthening the surface from corrosion causing further.This is attributed to corrosion occurring over increasing fraction of the surface was observed, which is the insoluble corrosion products.The insoluble corrosion products on the surface of the alloy could slow down the corrosion rate.

As shown in the Fig.10,at lower exposure times,trench like cavities appear on the surface of AZ31 specimen.At the higher exposure times,a little amount of corrosion pits was observed on the surface of the material.When the exposure time is more than 4.5 h,the barrier of the β phase and the adhesion of the corrosion products cause the corrosion weight loss rate to decreases.As a result,the weight loss is greater, and the corrosion weight loss rate is faster.With the prolongation of the exposure time,the corrosion weight loss rate decreases due to the adhesion of the corrosion products to the specimen surface.The f i lm on the specimen surface in the NaCl solution is considerably compacted,which can then effectively prevent furtherreactionsfrom taking place. Consequently,the growth rate of the f i lm remains slow and forms shallower corrosion dents.At the lower exposure time, the pit became deepens but in higher pH,the pit seems little widened possessing corroded products in large.It is clear from the Fig.11 localized corrosion associated with dense pitted areas showing lot of cracks on the surface of corrosion f i lm for the all specimens tested[28,29].

Fig.8.Effect of chloride ion concentration on corrosion morphology and pit depth of potentiodynamic polarization test specimen.

Fig.12 shows the effect of exposure time on Tafel plots, the corrosion potential shifted to a more positive direction with the increase in exposure time,the anodic branches are shifted to more positive potential direction compared,indicating the anodic dissolution is retarded with the increase of exposure time.As a result,the corrosion current also decreases with increasing exposure time.It was suggested that the existence of the corrosion f i lm in NaCl solution with the increase of exposure time.In the early stages of the corrosion process,the anodic dissolution of AZ31B magnesium alloy is balanced by hydrogen evolution in the cathodic areas in NaCl solution.

Fig.9.Effect of chloride ion concentration on Tafel plots.

Therefore,countless electrochemical cells begin to form on the surface.Then the cations(Mg2+)will migrate towards the cathodic areas,at the same time anions(Cl-,OH-)will migrate towardsto the dissolution sites.Accordingly, Mg(OH)2produces close to the anodic sites.

Fig.10.Effect of exposure time on corrosion morphology of immersion test specimen.

Nevertheless,the Mg(OH)2f i lm can be destroyed by Cl-in the solution,dissolution of the f i lm results in magnesium alloy exposure to the solution,providing the active centre for further electrochemical reactions.Thus,the localized corrosion takes place on the surface.The formation of the insoluble corrosion products may further retard the migration of ions on the surface of specimens,so it will reduce the electrochemical corrosion rate.That is,with exposure time increased,the corrosion rate will decrease[30].

On comparative evaluation,the corrosion rate of immersion tests was quite higher than the corrosion rate obtained from the potentiodynamic polarization test.This was due to the exposure surface of the specimen.For immersion testing,larger the surface area while in potentiodynamic polarization tests,only 1 cm2was exposed to the solution.Nevertheless,the general corrosion was evident in immersion test which tends to increase the corrosion rate,but only specif i c pitting was occurred in the polarization test.

5.XRD results

Fig.13(a)shows the XRD for AZ31B magnesium alloy, which exhibits on the α-Mg matrix phase and the results are conf i rms the presence of Al12Mg17precipitates along with the traces of Mg2Zn11.XRD study of the corrosion layer produced after immersion and potentiodynamic polarization test in NaCl solution revealed brucite(Mg(OH)2)as the main corrosion product,and its peaks exhibited higher intensity for AZ31B magnesiumalloyduetotheformationofathickercorrosionlayer during the severe attack that the material suffered Fig.13(b).

6.Conclusions

From this investigation,the following important conclusions are derived:

Fig.11.Effect of exposure time on corrosion morphology and pit depth of potentiodynamic polarization test specimen.

1.Empirical relationship was established to predict the corrosion rate of AZ31B magnesium alloy,incorporating the pH value,chloride ion concentration and exposure time for both immersion and potentiodynamic polarization tests.The developed relationship can be effectively used to predict the corrosion rate of AZ31B magnesium alloy at 95%conf i dence level.

Fig.12.Effect of exposure time on Tafel plots.

2.During immersion testing,the corrosion rate was quite higher than the corrosion rate obtained from the potentiodynamic polarization test.This was due to the exposure surface of the specimen.For immersion testing,entire surface area while in potentiodynamic polarization tests, only 1 cm2was exposed to the solution.

3.Nevertheless,the general corrosion was evident in immersion test which tends to increase the corrosion rate,but only specif i c pitting was occurred in the polarization test. Hence low corrosion rate was obvious in the potentiodynamic polarization tests.

4.In AZ31B magnesium alloy,the highest corrosion rate was observed at pH3.The corrosion rate was higher in the acidic media than in alkaline and neutral media for both corrosion tests conditions.

Fig.13.XRD results.

5.AZ31B magnesium alloys corroded more seriously with the increase in chloride ion concentrations in both the conditions.The increase in corrosion rate with increasing chloride ion concentration may be attributed to the participation of chloride ions in the dissolution reaction.

6.The corrosion resistance was formed in the AZ31B magnesium alloy with the increased exposure period.A corrosion resistivity prevails with the increase of corrosion time,resulting with the formation of hydroxide layer as a dominant factor to avoid the corrosion further.

7.The AZ31B alloy specimens were lower corrosion rate in the electrochemical environments than the immersion environment.Thus,the specimens proved to give a long life from corrosion in electrochemical environments.

Acknowledgements

The authors wish to record their sincere thanks to Mr.R. Selvendiran,Technical Assistant,Corrosion testing laboratory,Department of Manufacturing Engineering,Annamalai University for his help in carrying out this investigation. Authors also wish to thank Mr.N.Sairaman and Mr.John of project assistants,CEMAJOR for their help and support.The technical support of Dr A.K.Lakshminarayanan,Associate Professor,SSN College of Engineering,Chennai during the course of this work is gratefully acknowledged.

[1]B.L.Mordike,T.Ebert,Mater.Sci.Eng.A 302(2001)37.

[2]R.Ambat,N.Aung,W.Zhou,J.Appl.Electrochem.30(2000)865.

[3]L.Gao,C.Zhang,M.Zhang,X.Huang,N.Sheng,J.Alloys Compd.468 (2009)285.

[4]Holly J.Martin,M.F.Horstemeyer,Paul T.Wang,Corros.Sci.52(2010) 3624.

[5]G.Song,A.Atrens,Adv.Eng.Mater.9(2007)177.

[6]Rong-chang Zeng, Jin Zhang,Wei-jiu Huang,W.Dietzel, K.U.Kainer,C.Blawert,Wei Ke,Trans.Nonferrous Met.Soc.China 16(2006)s763.

[7]Feng Liu,Ying-Wei Song,Da-Yong Shan,En-Hou Han,Trans. Nonferrous Met.Soc.China 20(2010)638.

[8]Lei Wang,Tadashi Shinohara,Bo-Ping Zhang,J.Alloys Compd.496 (2010)500-507.

[9]Yingwei Song,Dayong Shan,Rongshi Chen,En-Hou Han,Corros.Sci. 52(2010)1830.

[10]R.Ambat,N.Aung,W.Zhou,Corros.Sci.42(2008)1433.

[11]A.Pardo,M.C.Merino,A.E.Coy,F.Viejo,R.Arrabal,S.Feliu Jr., Electrochim.Acta 53(2008)7890.

[12]L.Wang,T.Shinohara,B.P.Zhang,J.Solid State Electrochem.14(2010) 1897.

[13]A.Dhanapal,S.R.Boopathy,V.Balasubramanian,Mater.Des.32(2011) 5066.

[14]D.C.Montgomery,Design and Analysis of Experiments,John Wiley& Sons Ltd,New Delhi,2007.

[15]ASTM G31-72,Standard Practice for Laboratory Immersion Corrosion Testing of Metals,2002.

[16]ASTM G5-94,Standard Reference Test Method for Making PotentiostaticandPotentiodynamicAnodicPolarization Measurements, 2011.

[17]S.Kumar,P.Kumar,H.S.Shan,J.Mater.Process Tech.182(2007) 615.

[18]Z.M.Zhang,H.Y.Xu,B.C.Li,Trans.Nonferrous Met.Soc.China 20 (2010)697.

[19]G.Song,A.Atrens,M.Dargusch,Corros.Sci.41(1998)249.

[20]H.Altun,S.Sen,Mater.Des.25(2004)637.

[21]D.Thirumalaikumarasamy,K.Shanmugam,V.Balasubramanian,Trans. Indian Inst.Met.67(2014)19.

[22]W.Xu,J.Liu,H.Zhu,Electrochim.Acta 55(2010)2918.

[23]M.Jonsson,D.Persson,D.Thierry,Corros.Sci.49(2007)1540.

[24]Ming-Chun Zhao,Ming Liu,Guang-Ling Song,Andrej Atrens,Corros. Sci.50(2008)3168.

[25]M.C.Merino,A.Pardo,R.Arrabal,S.Merino,P.Casajus,M.Mohedano, Corros.Sci.52(2010)1696.

[26]N.Liu,J.Wang,Y.Wu,L.Wang,J.Mater.Sci.43(2008)2550.

[27]K.H.Goh,T.T.Lim,P.C.Chui,Corros.Sci.50(2008)918.

[28]M.C.Zhao,M.Liu,G.Song,A.Atrens,Corros.Sci.50(2008) 1939.

[29]G.Song,A.Atrens,Z.Wu,B.Zhang,Corros.Sci.40(1998)1769.

[30]R.Lindstrom, L.G.Johansson, G.E.Thompson,P.Skeldon, J.E.Svensson,Corros.Sci.46(2004)1141.

Received 2 December 2013;revised 27 January 2014;accepted 28 January 2014 Available online 22 March 2014

*Corresponding author.Tel.:+91 09894319865(mobile);fax:+91 4144 238080x238275.

E-mailaddresses:tkumarasamy412@gmail.com,kumarthirumalai68@ ymail.com(D.Thirumalaikumarasamy).

Peer review under responsibility of National Engineering Research Center for Magnesium Alloys of China,Chongqing University

Production and hosting by Elsevier

http://dx.doi.org/10.1016/j.jma.2014.01.004.

2213-9567/Copyright 2014,National Engineering Research Center for Magnesium Alloys of China,Chongqing University.Production and hosting by Elsevier B.V.All rights reserved.

Copyright 2014,National Engineering Research Center for Magnesium Alloys of China,Chongqing University.Production and hosting by Elsevier B.V.All rights reserved.