缝隙腐蚀试验1

Temperature and potential dependence of crevicecorrosion of AISI 316stainless steelP.T.Jakobsen *,1,E.MaahnMaterials Technology,Department of Manufacturing Engineering,Technical University of Denmark,Building 204,DK-2800Lyngby,DenmarkReceived 16March 2000;accepted 9October 2000AbstractCrevice corrosion of AISI 316steel has been investigated in a modi®ed ÔAvesta Cell Õ.Po-tentiodynamic scans have been made in sodium chloride solutions of various pH values at di erent temperatures with and without crevice.The breakthrough potential changes dis-continuously with temperature.At all temperatures the attack is localised to the crevice,and the breakthrough potential is lower than for experiments without the crevice.The experiments in acidi®ed environments indicate that crevice corrosion at low temperatures results from acidi®cation in the crevice.At higher temperatures crevice corrosion is believed to be the result of metastable pitting stabilised by the crevice.Ó2001Elsevier Science Ltd.All rights re-served.Keywords:Stainless steel;Polarisation;Crevice corrosion1.IntroductionCrevice corrosion is one of the most common forms of corrosion of stainless steels.Because it cannot easily be discovered before it is fatal for the construction,crevice corrosion is a very detrimental form of corrosion.Crevice corrosion of stainless steel shares many similarities with pitting corro-sion of stainless steel [1].Both phenomena are in¯uenced by pH,bulk chloride/locate/corsciCorrosion Science 43(2001)1693±1709*Corresponding author.Tel.:+46-8-674-17-00;fax:+46-8-674-17-80.E-mail address:pia.jakobsen@corr-institute.se (P.T.Jakobsen).1Present address.Swedish Corrosion Institute,Roslagsv a gen 101,hus 23A,SE-10405Stockholm,Sweden.0010-938X/01/$-see front matter Ó2001Elsevier Science Ltd.All rights reserved.PII:S 0010-938X (00)00167-0concentration and temperature,and in much the same way.Crevice corrosion is induced by less severe conditions than pitting corrosion.The resistance against these forms of corrosion is often expressed as a critical temperature [1,2].For pitting corrosion it is found that there exists a potential independent critical pitting corrosion temperature (CPT)[3,4].Below this temperature pitting is not possible and only transpassive corrosion occurs.Above this temperature pitting is possible,but it depends on the potential whether it will actually occur.This is illus-trated in Fig.1,which shows the temperature dependence of the breakthrough po-tential.The potential independent CPT is the temperature at which a sudden change in breakthrough potential is observed.A curve of the shape shown in Fig.1is found if the breakthrough potential is found from potentiodynamic scanning experiments on a specimen free from crevices as shown by Qvarfort [5,6].By potentiostatic testing with increasing temperature it is possible to determine the potential independent CPT if the potential is chosen between the upper end of the pitting potential range and the lower end of the transpassive corrosion range [4].On the other hand if the chosen potential is within the pitting potential range a potential dependent CPT is deter-mined [4].Arnvig and Bisg ard [4]determined a potential independent CPT whereasMellstr om and Bernhardsson [7]determined a potential dependent CPT.Crolet and Defranoux [8]and Old®eld and Sutton [9]have described a theory for initiation of crevice corrosion of stainless steels by a gradual change of the en-vironment in the crevice.Following Old®eld and Sutton [9],there is a neutral chloride solution containing dissolved oxygen inside and outside the crevice at the outset.Two reactions take place within as well as outside the crevice:Metal dissolution :Me 3Me z z e À1 Oxygen reduction :O 2 2H 2O 4e À34OH À21694P.T.Jakobsen,E.Maahn /Corrosion Science 43(2001)1693±1709Reaction2consumes oxygen which inside the crevice is only replenished by di usion. If the crevice is su ciently deep and tight a part of the crevice will be depleted from oxygen.Reaction1can still take place within the crevice balanced by reaction2 occurring on surfaces external to the crevice.The metal ions thus generated in the crevice will have a positive charge,and to maintain charge neutrality metal ions will migrate out and chloride ions will migrate into the crevice.The main process will be ingress of chloride ions into the crevice.At the same time the metal ions(particularly chromium)will hydrolyse thereby forming acid,e.g.:Cr3 H2O3Cr OH 2 H 3 These processes lead to a more aggressive environment in the crevice(higher chloride content and more acidic).This will accelerate the metal dissolution process(reac-tion1)leading to further migration of chloride ions into the crevice and to fur-ther acidi®cation due to hydrolysis.This continues until the environment formed is su ciently aggressive to break down the passive®lm permanently within the crevice. Once this happens crevice corrosion has initiated[9].Stockert and B o hni suggested that crevice corrosion initiates by a geometrical stabilisation of metastable pitting corrosion[10].A metastable pit is a small pit which initiates and grows for a while under a cover,which presumably is a remnant of the passive®lm.The growth of the pit leads to the formation of a concentrated environment within the pit,and an osmotic pressure builds up over the cover.When the osmotic pressure exceeds the strength of the cover the cover ruptures[10,11].It is found that the concentration of metal ions on the corroding surface of a cavity must exceed about75%of saturation concentration if rapid dissolution are to continue [12].Based on this®gure and a consideration of the steady state di usion out of an open hemispherical cavity Pistorius and Burstein[11]®nd that the product of the pit radius r and the current density i must ful®ll the criteria given in Eq.(4)for con-tinued pit growth to follow the rupture of the cover:ir P 3zFD D C2p; 4where z is the average charge of the metal ions,F is the Faraday number,D is the di usion coe cient and D C is the di erence in metal ion concentration from the corroding pit surface to the bulk solution.The product ir is called the pit stability product.If the pit stability product does not ful®ll the criteria given by Eq.(4)the metastable pit will repassivate when the cover ruptures.The current density by which a metastable pit grows is constant during the metastable growth period[11].A metastable pit initiated at a higher potential on average will grow with a higher current density.Pistorius and Burstein[13]further found that the radius of the metastable pit when the cover ruptures is almost in-dependent of the current density by which the metastable pit grows.This means that the probability that a metastable pit will be able to continue to grow as a stable pit after the cover ruptures will increase with increasing potential[11,13]. Metastable pitting will occur in a crevice as well as on open surfaces.If the metastable pit occurs in a particularly tight point in the crevice the opposite crevice P.T.Jakobsen,E.Maahn/Corrosion Science43(2001)1693±170916951696P.T.Jakobsen,E.Maahn/Corrosion Science43(2001)1693±1709wall will act as a di usion barrier.If the di usion out of the pit is limited su ciently the metastable pit will be stabilised even if it would not have survived on an open surface[10].This is geometrical stabilisation.The radius when the cover ruptures presumably will be una ected by the crevice,and thus a metastable pit growing at a smaller current density can be stabilised in a crevice.A metastable pit growing at a smaller current density will require a smaller crevice width to become stabilised. It is found that the number of metastable pits per unit area of the crevice must exceed a certain number before it is sure that initiation of crevice corrosion has happened[10].This is linked to the property that only su ciently tight points in the crevice are e ective as initiation sites.According to Laycock et al.(1998)[14]a metastable pit at temperatures below the potential independent CPT will repassivate by precipitation of a salt layer in a manner similar to that of iron passivating in sulphuric acid.The critical current density for passivation will increase with temperature.The current density required for stabilisation of a metastable pit will also increase with temperature but slower than the critical current density for passivation.The temperature where these two current densities are equal is the CPT[14].The current density required to stabilise the pit in the crevice is smaller than the current density required to stabilise the metastable pit on an open surface.The critical current density for passivation presumably will not depend on whether the pit is inside or outside the crevice.Following the argument of Laycock et al.(1998)[14] this will mean that the critical crevice corrosion temperature(CCT)will be lower than the CPT.However,the CCT will depend on the tightness of the point in crevice where the metastable pit happens.The CPT can be determined by electrochemical methods,for instance in Refs. [4,7].It may also be determined by simple immersion in solutions with increasing temperature[1].The CCT can also be determined by immersion tests[1,2]or by electrochemical tests[15].It appears from the literature that the CCT most often is determined by immersion tests.Some authors have determined breakthrough po-tentials for crevice corrosion by potentiodynamic scanning[16±18].These investi-gations only used one ycock et al.(1997)[19]measured breakthrough potentials as a function of temperature on rods of904L stainless steel with two o-rings as crevice formers.This paper presents a determination of the breakthrough potential for crevice corrosion as a function of temperature and aims at answering the question whether a potential independent crevice corrosion temperature exists in analogy with the po-tential independent CPT.2.ExperimentalTheÔAvesta CellÕwas developed to test pitting corrosion without the in¯uence of crevice corrosion[5].This is achieved by placing a®lter-paper gasket between the test specimen and the cell and having a small¯ow of distilled water through this gasket. This eliminates the possibility of crevice corrosion.TheÔAvesta CellÕhas later beenmodi®ed to test a larger exposed area of the test specimen [4].This larger area makes it possible to apply a crevice onto the exposed area of the test specimen in the ÔAvesta Cell Õ.In this way the e ect of the test specimen edges is eliminated from the test for crevice corrosion resistance.Fig.2is a schematic drawing of the Ôcrevice corrosion ÔAvesta Cell Õ(CCAC)Õ.The crevice in the CCAC is formed between the test specimen surface and a ring-shaped crevice former of polyvinylidene ¯uoride (PVDF).The crevice former is ground on silicon carbide paper to a #1000®nish.The crevice former is attached to a silicone rubber ring which in turn is attached to a connection piece.The silicone rubber ring ensures that the crevice forming surfaces are aligned.The connection piece is extended by a glass tube which allows the application of pressure onto the crevice.Pressure is applied by compressing the disc springs with the screw.The heating coil allows the cell temperature to be controlled.The small ¯ow of distilled water increases the electrolyte volume.To counteract this the CCAC has an over-¯ow.The dilution caused by the ¯ow of distilled water is compensated for by adding solution of the same composition but higher concentration than the bulk electrolyte.To assure an even ¯ow of distilled water,the distilled water is allowed to ®ll the exposed area before the electrolyte is poured into the cell.An even ¯ow is necessary if crevice corrosion between test specimen and cell is to be avoided.The electrolyte is diluted less than 2&.The exposed area of the test specimen is 10cm 2including the crevice area.The crevice area is 2.4cm 2.The ¯ow rate of distilled waterisFig.2.Schematic drawing of the CCAC.P.T.Jakobsen,E.Maahn /Corrosion Science 43(2001)1693±17091697approximately 40ml/h.The crevice was applied after the electrolyte had been poured into the cell,which means that the electrolyte will ®ll the volumes inside and outside the crevice former.The crevice will also be ®lled with electrolyte from the start of the experiment.The springs were compressed with a force of 220Æ10N.The test specimens were made from cold rolled,as delivered AISI 316plate.The surface is designated 2B according to EN 10088.The composition of the speci®c material used is given in Table 1.The test specimens were washed with hot water and soap,degreased by rinsing with acetone and ethanol,and following this dried in hot air.The test solution was 1MNaCl made from distilled water and analytical grade NaCl.A few experiments were done in 1MNaCl acidi®ed with HCl.Polarisation scans in anodic direction,starting from 0mV/SCE,were made at di erent temperatures.The scan rate was 1mV/min.The scan was stopped when the current density reached a predetermined value.All current densities are referred to the exposed area of 10cm 2.The test solution was held thermostatically at the desired temperature Æ0:5°C.3.ResultsFig.3shows breakthrough potentials as a function of temperature for pitting corrosion measurements.The breakthrough potential is taken as the potential at which the current density continuously exceeds 10l A/cm 2.The data used in Fig.3were found without applying the crevice.The test specimen surface has thus been openly exposed to the electrolyte in these experiments.In Fig.3linear regressions for the data are included.These are included to show the overall tendency of the data and thus assist the interpretation of the ®gure.The data for pitting corrosion show the potential independent CPT to be 15.5°C.Fig.3only includes few points,but more points would not have changed the conclusion signi®cantly,as the CPT is described by other authors [4]to be well de®ned.Fig.4shows breakthrough potentials for crevice corrosion experiments as a function of temperature.Linear regressions of the data are included in this ®gure for the same reasons as in Fig.3.The data for crevice corrosion in some respects re-semble and in some respects di er from the data for pitting corrosion.The data for crevice corrosion does not have a sharp transition from high to low breakthrough potentials.The transition happens over a temperature interval,and in this tem-perature interval also intermediate values of the breakthrough potential are observed.The breakthrough potential for crevice corrosion is lower than the breakthrough potential found on an open surface both at temperatures above and below the transition temperature interval (tti)(hereafter transition interval).TheTable 1Composition of the AISI 316steel used for the investigation Element Cr Ni Mo Si Mn C N P S Others Fe %17.0412.702.620.451.690.0150.0540.0280.0010.96Balance1698P.T.Jakobsen,E.Maahn /Corrosion Science 43(2001)1693±1709limits of the breakthrough potentials for crevice corrosion at temperatures in the transition interval are seen to be set by the behaviour at temperatures aboveandP.T.Jakobsen,E.Maahn /Corrosion Science 43(2001)1693±17091699below the transition interval.It is found that the attack is localised to the crevice at all temperatures (even at temperatures below the transition interval)for the crevice corrosion experiments.Fig.5shows a typical curve from potentiodynamic scanning with the crevice at a temperature above the transition interval.The passive current density increases very slightly as the potential increases.When the breakthrough is reached the current rises quite fast.This behaviour is typical for localised corrosion of stainless steel.The increase in current from 1to 100l A/cm 2is typically observed to happen over 100±150mV at temperatures above the transition interval.This is somewhat slower than seen for pitting corrosion,but the propagation of crevice corrosion in the early stages is slowed down by the electrolyte resistance in the crevice cavity [20].The attacks seen above the transition interval are in a few distinct positions in the crevice,most commonly just one or two.The attack is a patch of uniform corrosion,and it is deeper at the outer circumference of the crevice area.The attacks observed on specimens from experiments in the transition interval are not signi®cantly di erent from those observed above the transition interval.This implies that the attacks are initiated by the same mechanism in and above the transition interval.Fig.6shows typical curves from potentiodynamic scanning with the crevice in 1M NaCl solutions of various pH values at a temperature below the transition interval.Considering the curve in 1MNaCl solution it is seen that this curve di ers from the curve in Fig.5in several ways.Firstly by the higher breakthrough potential.However,this is the characteristic by which the 1MNaCl curve in Fig.6isde®ned1700P.T.Jakobsen,E.Maahn /Corrosion Science 43(2001)1693±1709as being below the transition interval,and the curve in Fig.5is above the transition interval.At a potential around 750mV/SCE the current density in the 1MNaCl curve in Fig.6becomes almost constant until the breakthrough starts.If the ex-periment is stopped at low current density,a ring of yellowish brown discoloration can be seen approximately halfway between the inner and outer peripheries of the crevice area.The attack seen when the experiment is stopped at a higher current density will be a large number of small patches of uniform attack evenly distributed along the circle and with discoloration between the patches.There will be more patches of attack near the outer periphery of the crevice area than near the inner periphery.Fig.7shows typical curves from potentiodynamic scanning in 1MNaCl solutions of various pH values in experiments where the crevice was not used.The experiments were made at a temperature below the potential independent CPT.This means that the resulting corrosion was transpassive corrosion.The 1MNaCl curve in Fig.7is characterised by having a local maximum in the passive current density at a potential around 750mV/SCE.At a higher potential a local minimum in the current density is observed.The curves shown in Fig.7for experiments done without crevice in 1MNaCl acidi®ed to pH 4and pH 2show that the current density in the maximum becomes higher the lower the pH is.The change is small at pH 4.At pH 2the maximum current density is near 10l A/cm 2.The current density measured at the minimum is also higher the lower the pH is.At pH 2the current density decreases to about 2l A/cm 2.At pH 1the current simply increases without passing through amaximum.P.T.Jakobsen,E.Maahn /Corrosion Science 43(2001)1693±17091701The rise of the current starts where the increase towards the maximum starts when the pH value is higher.This behaviour indicates primary transpassive corrosion followed by a secondary passivation and ®nally secondary transpassive corrosion [6].This kind of behaviour has been observed by other authors for stainless steels [6,21].After terminating experiments done without crevice below the CPT the solutions smelled as chlorinated water does.This indicates that the current rise after the minimum was due to oxidation of chloride,although it is possible that some oxygen evolution took place simultaneously.The secondary passivity is in¯uenced by pH and is not possible if the pH decreases too much.The curves shown in Fig.6for experiments done with crevice in 1MNaCl acidi®ed to pH 4and pH 2show that the behaviour in the pH 4solution is very similar to the behaviour in 1MNaCl.At pH 2the current passed 1l A/cm 2at a lower potential than in 1MNaCl.Between 10and 15l A/cm 2the current rose very slowly in the pH 2solution.4.DiscussionAt temperatures below the transition interval the initiation of crevice corrosion appears to happen quite uniformly in the part of the crevice where the distance to the edges of the crevice area is largest.The experiments done in acidic bulk solutions help to explain the mechanism of initiation of crevice corrosion at temperatures below the transition paring the curves done at the same pH value from Figs.6and 7it is seen that the overall behaviour of the two curves is the sameuntil1702P.T.Jakobsen,E.Maahn /Corrosion Science 43(2001)1693±1709the potential at which the maximum in current density occurs for the experiment done without crevice.As the potential increases above the value at the maximum the current density decreases for the experiment without the crevice,but it remains almost constant for the experiment with the crevice.Thus no secondary passivation is observed in the experiments with crevice although it is probable that secondary passivation will occur outside the crevice in these experiments.The current density within the crevice will therefore increase during the interval in which almost constant current density is observed.A rising current density inside the crevice can be rationalised if the pH within the crevice is lower,cf.Fig.7.The occurrence of sec-ondary passivation in the experiments without crevice show that primary transpas-sive corrosion has taken place[22,23].By primary transpassive corrosion soluble Cr(VI)-species are formed,e.g.[24]:Cr 4H2O3HCrOÀ4 7H 6eÀ 5 It is seen that acid is formed simultaneously with the Cr(VI)-species.Convection will make the concentration of Cr(VI)-species and acid low at the surface of the metal outside the crevice.Inside the crevice there is no convection,and the concentration of Cr(VI)-species and acid will increase.In Appendix A[26]the pH in the crevice is calculated based on assumptions as indicated in the appendix.Depending on the crevice width used in the calculation the pH is found to be between0.57and1.57. However,the curve found in Appendix A for the current within the crevice indicates that secondary passivation does take place in the crevice before breakthrough.Thus it is proposed that crevice corrosion below the transition interval initiates due to a gradual acidi®cation of the environment in the crevice,but initiation is only possible once secondary passivation has occurred in the crevice.When secondary passivation occurs the passive®lm becomes more iron-rich.It appears that the environment formed in the crevice is only able to break down the more iron-rich®lm. Experiments atÀ0.5°C in which the potential was scanned by1mV/min to 600mV/SCE and then held potentiostatically at600mV/SCE for48h showed de-creasing current density during the potentiostatic period.Similar experiments but with the potential held at800mV/SCE also showed decreasing current density during the potentiostatic period.Thus it appears that an over-potential is necessary for stable propagation of crevice corrosion at temperatures below the transition interval.This mechanism can be considered a modi®cation of the mechanism pro-posed by Crolet and Defranoux[8]and by Old®eld and Sutton[9].However,it is important to note that the mechanism proposed here depends critically on the high potential to produce an acidic environment and that probably initiation is only possible once secondary passivation has occurred within the crevice.In and above the transition interval the corrosion attack is seen as one or perhaps a few patches of uniform corrosion in the crevice.This indicates that the initiation is pointwise.The geometrical stabilisation of metastable pitting theory of Stockert and B o hni[10]will lead to pointwise initiation.Under the experimental conditions used in this investigation the number of metastable pits as a function of temperature and potential during potentiodynamicscanning becomes important.Pistorius and Burstein[11,13]found that the meta-stable pitting rate(number of metastable pits per unit area per unit time)has a maximum at intermediate potentials(which actually are quite low).The number of possible metastable pitting sites increases with potential at®rst and then becomes constant[11,13].Gar®as-Mesias and Sykes[25]did potentiodynamic scanning on 25Cr duplex stainless steel.Their curves show some transients but each of these probably includes several metastable transients.They also®nd a maximum in metastable pitting rate at intermediate potentials.At higher potentials the transients are few and occur randomly.The size and number of observed transients increase with temperature in the measurements of Gar®as-Mesias and Sykes[25].This pre-sumably signi®es that the number of transients increase with temperature and that the current density at which the individual metastable pit is growing also increases with temperature.It is assumed that at temperatures above the transition interval the number of metastable events at low potentials exceeds the number necessary to initiate crevice corrosion.The breakthrough potential therefore is low.As the metastable events are distributed randomly over the surface,the actual number of events happening before the initiation will vary.This explains the variation in the breakthrough potential at one temperature.If the temperature is decreased the number and size of the meta-stable events decreases.Both of these e ects will shift the breakthrough potential to higher values with decreasing temperature.In the transition interval it is believed that the number of metastable events at low potentials will not exceed the number necessary to initiate crevice corrosion.Thus the initiation might happen at low potentials,or it might happen at higher potentials. If crevice corrosion does not initiate at low potentials the breakthrough will appear to happen coincidentally,because the metastable events occur randomly.This is the reason for the large variation of the breakthrough potential in the transition interval. When decreasing the temperature further,the size of the metastable pits becomes too small to initiate crevice corrosion,because the crevice width is not su ciently small. At this temperature the mechanism of initiation shifts from geometrical stabilisation of metastable pitting to gradual acidi®cation(moving towards lower temperatures). The lowest temperature at which a metastable pit can be stabilised in the crevice is a CCT in the sense that initiation of crevice corrosion will not(normally)be possible at lower temperatures where the redox-potential of the bulk solution will have to be unusually high to give crevice corrosion.The transition interval arises because of the variations in crevice geometry be-tween the experiments and because the metastable pits occur in random positions on the surface.The tightness of the tightest points probably will not be the same in each experiment and a metastable pit might not occur in the tightest points.Both these factors might lead to a scatter in temperature as well as potential.Thus the ap-pearance of a transition interval is presumably unavoidable for experimental de-termination of crevice corrosion behaviour.At the lowest temperature in the transition interval it will be possible to initiate crevice corrosion by geometrical stabilisation of metastable pitting at a low potential although this situation will not be realised in every experiment due to the dependence。