Effect of preferential dissolution on erosion-corrosion for chromium steel in alkali slurry

Vol.12 No. 5Trans. Nonferrous Met. Soc. ChinaOct. 2002[ Article ID ] 1003 - 6326( 2002 )05 -0931 -05Effect of preferential dissolution on erosion-corrosionfor chromium steel in alkali slurry"YUAN Qing-long( 袁庆龙)' ,M. M. Stack2( 1. Surface Engineering Research Institute,Taiyuan University of Technology , Taiyuan 030024 , China ;2. Department of Mechanical Engineering , University of Strathclyde , Glasgow G11XJ ,UK )[ Abstract ] An investigation was carried out concerming the effect of preferential dissolution on the erosion-corrosion for achromium steel in 1 mol/L NaOH.Preliminary tests using a potentiodynamic technique were performed in order to establishthe presence of preferential dissolution in the alkali solution with and without the alumina particles at different rotationspeeds. For purposes of quantifying the observed phenomena a potentiostatic mass loss method was also used. The resultsshow that the active peaks occur at potential between +0.4 and +0.5 V on the polarization curves , which indicates thatthere is a preferential dissolution for chromium steel under erosion-corrosion conditions and the ferrite phase acts as a sacri-ficial anode in favor of( Fe Cr )C; phase. Addition of particles can promote the preferential dissolution at different rotationspeeds. The combined effects of erosion-corrosion results in total mass loss rates to be greater than the sum effects of eachprocess taken alone , thus showing a strong synergism between erosion and corrosion due to preferential dissolution.[ Key words ] chromium steel ; erosion-corrosion ; preferential dissolution ; synergism[ CLC number ] TC172. 85[ Document code ]A1 INTRODUCTIONprecise determination was not possible.Yue et al7] detailed a mechanism for describingIn the Bayer process ，in hydroelectric powerthe interaction between erosion and corrosion in aque-plants and in offshore piping systems , flowing corrosionous condition :media often contain solid particles. These solid parti-K。c=K。+K。+△K。+△K。(1)cles suspended in liquid form slurries , and the processof slurry erosion-corrosion can be particularly destruc-where Ke is the erosion-corrosion rate ,K。is the puretive. Under these environments , the surfaces of compo-corrosion rate ,K。 is the pure erosion rate ,AK。is corro-nents are damaged not only by the scouring action ofsion rate increment , and△K。is erosion rate increment.the solid particles , but also by the electrochemical cor-If Kee =K。+K。+△K。, then the erosion-corrosion isrosion. Both erosion and corrosion processes may actconsidered to be " additive" ( erosion enhanced corro-synergistically to produce an overall effect that is grea-sion ). If this is not the case ,then a synergistic effect is .ter or less than the sum of the damage caused by eachobserved , that is of corrosion enhancing the erosion.process acting independently. There was a great deal ofHowever , few have determined the separate contribu-interest in this phenomenon'MutsumuraSbl de-tions of erosion and corrosion to the overall rate.termined that synergy existed during the erosion-corro-For the stainless steels , preferential dissolution ission of stainless steel and iron in both acids and alkali ,well known , which occurs when different componentswith corrosion accelerating the erosion. This was attrib-of an alloy , or when different phases of a multiphaseuted to dissolution of the work hardened layer and anmaterial，corrode at different rates.However，theincrease in surface roughness causing more destructiveeffect of preferential dissolution on erosion was not al-erosion impacts. Corrosion was also reported to be in-ways taken into consideration. In present work , the po-hibited by erosion due to the production of impact cra-larization curves are determined and used in order toters having crystal planes on which dissolution was slo-define the behavior of a chromium steel within a certainwer. It was considered , however , that the conclusionsmade may not be accurate due to an inability to resolverange of potentials ( active corrosion ) in alkali slurriesthe various variables of erosion and corrosion , a prob-of different particle sizes with respect to preferentiallem that was acknowledged. Li et al31 also found a .dissalution and. the reason for nreferential dissolution tosynergism to be present in the erosion-corrosion of alu-acce中国煤化Ised.minum over a range of pH value , with the main effectYHCNMHGarising from corrosion enhanced erosion. The specific2EXPERIMENTALmechanism of enhancement was concluded to be en-hanced crack growth rate in the surface oxide though2.1 Apparatus①[ Re据2002 - 03 -05 ;[ Acepled date 12002 -05 -16932.Trans. Nonferrous Met. Soc. ChinaOct.2002A rotation cylinder electrode( RCE ) apparatus50 μm , 100 μm and 150 μm alumina particles withwas used for the erosion-corrosion tests , which is sche-nominal concentration of 275 g/L respectively. In-situmatically shown in Fig. 1. The RCE rig consists of apurging of nitrogen into the cell was further made dur-hollow cylindrical working electrode ( specimen ring ) ,ing the wear test through an upper inlet to the cellan auxiliary counter electrode ( platinum foil )，bothwhich can avoid possible cavitation erosion associatedcontained in the electrode cell , and an outside stand-with nitrogen bubbling via a lower inlet. The test wasard reference electrode ( Hg/HgO ) connected to thecarried out at ambient temperatures and a water bathsolution by a salt bridge. Eletrical connection to thewas used for the high-speed rotation test to prevent sig-cell , made through the three electrodes , is controllednificant slurry temperature rises.by an ACM potentiostat. Signal inputs to and data ac-quisition from the potentiostat were carried out by a2.3 Potentiodynamic sweepscontrol computer via a high velocity digital-to-analoguePolarization curves in 1 mol/L Na0H solution with( D/A ) and analogue-to-digital ( A/D ) conversion in-and without alumina particles were measured at the re-terface board interfaced by an Easyest LX software.quired rotation speeds from -1.0V to +0.7V vs Hg/HgO at 1 mV/s with a pre-treatment at - 1.5 V for 60s2.2 Material and experimental conditionsThe material tested was a chromium steel whichprior to the initiation of potential sweep to remove anyhas a chemical composition of 0. 13% ~0. 18%C ,residual air-formed oxide film. The potential ramp was0.51%Mn ,0. 44%Si ,0. 46% Al , 12.0% Cr and bal-supplied to the controlling potentiostat by a linearance Fe. The specimens were machined to a final di-sweep generator ; whilst both potential and currentmension of 38 mm( outer diameter ) x2 ~4 mm( wallmeasurements were taken at the data acquisition boardthickness) x 10 mm( height ). Prior to testing , theand stored on the host computer.specimen surface was finished with silicon carbide ab-rasive paper , rinsed in de-ionized water , degreased in2.4 Determinations of K。,K。and Kecacetone ,dried in air and weighted. After testing , theThe pure erosion component ,K。in the slurry wasspecimen was cleaned , removed and re-weighted.determined under protection condition so that only ercThe slurries , based on the 1 mol/L NaOH solu-sion can occur. The mass loss under effective cathodictions , were prepared from de-ionized water and suffi-protection( -0.9~ -1.0V ) and at the potentals in-cient deaerated using pressurized nitrogen ，containedsufficiently negative to cause redeposition of dissolvedConnected. to rotationdriverGas, VortexoutetbaffleStainlesssteelcoreR GainletNylon-BafleR.E.PIFEwasher+ Sample中国煤化工. ThrendW.EYCNMHG. Solution而方数据schematic diagam of rotating cylinder electrode system used for erosonocorrosiono testsVol. 12 No. 5Effect of preferential dissolution on erosion-corrosion for chromium steel933-iron'was taken to be the pure erosion rate , K。Determination of the pure corrosion rate , K。wascarried out under the same conditions as for erosion-corrosion test and at the same potential to allow com-parison. However , as the corrosion of chromium steelin sodium hydroxide is comparatively low , significantand measurable mass losses could only be made between +0.2V and +0. 6 V. The polarization potentialand rotation were applied at the same time and data ac-quisition at a rate of 1 Hz started. The current was re-corded for a period of 1 h until the test was stopped.The current data was integrated with respect to time toprovide the quantity of change passed during the exper-iment , which in turn was used to provide a value forthe corrosive mass loss using Faraday's second lawt1o].: 6828.日kU10yum4B. U CORR0055The total erosion-corrosion rate( K。。 ) at a givenB 58BXpotential was measured as the total mass loss experi-Fig.2 Micrograph of chromium steel surface afterenced by a sample during 1 h experiment in the pres-erosion-corrosion testence of the required particulate concentration.( rotation speed = 6m/s )3 RESULTS AND DISCUSSION1013.1 MicrostructureIn the annealed condition , the microstructure ofthe chromium steel consists of a matrix of fine ferrite10grains with chromium carbides( Fe ,Cr )C3 dispersed52V31in the grain boundaries. These precipitates have a min-imum of 36% Cr and cause chromium content in re-i 10-1 2-gions adjacent to the phase boundaries to be loweredRotation speedbelow the critical level 12% needed for corrosion re-sistance 1.Therefore , the chromium steel is suscepti-2- -2 m/s10~3一4 m/sble to intergranular and interphase corrosion. Fig. 21-shows that the surface morphology of the chromiumsteel was extensively pitted with evidence of general10-3Lcorrosion in 1 mol/L deaerated sodium hydroxide at a0.4 0.8potential of +0.4 V. The pits ranged in size from 2 μm .Potential (vs Hg/HgO)/Vto sub-micron with interior of the larger pits showingloose fibrous structures remaining after the preferentialFig.3 Polarization behaviors for chromiumdissolution of ferrite phases and the original precipitatesteel in 1 mol/L NaOH solution withfalling off.different rotation speeds3.2 Polarization curveallowing passivation to occur more easily and enhance3.2.1 Efect of rotation speedthe stability of the film formed. There was a large ac-The measurement results of polarization curves fortive peak developed at +0.4 V due to the preferentialchromium steel in 1 mol/L sodium hydroxide solutiondissolution of ferrite phase in chromium steel.with different rotation speeds are presented in Fig. 3.With cylinder rotation , the size of the active peakCompared with that of static state , the slight increase( at -0.6 V ) reduced with increasing rotation speedof passive current density in rotative state is the resultpossibly related to the transport of Fe ions , while theof the joint action of acceleration of the transportlevel of passivation current density( J , .。)increased sprocess of the residual oxygen. An apparent positivequen中国煤化工s a result of the flow af-shift of the free-corrosion potential occurred becausefectiEve film. A stepwise in.the depolarization of oxygen became one of the maincreafYHC N M H Galso noted over the po-cathodic reactions in rotating state'tential range of the preferential dissolution( at +0.4 ~Without flow ，the active peaks at -0.8 V and+0.5 V ) with the peak moving to slight more negative-0.6 V were decreased indicating less active dissolu-potentials at higher speeds.tion while_ the ensyring passive region also resided at a3.2.2 Effect of particle sizelower current部护, as a result of the chromium contentPotentiodynamic scans were run to determine the934.Trans. Nonferrous Met. Soc. ChinaOct.2002 .effect of particle size on the polarization data. Fig. 4shows the current density vs the polarization potential0.25S(afor chromium steel under test conditions of 6 m/s flow0.20}speed and 1 mol/L NaOH with 273 g/L alumina parti-0. 154cles with different particle sizes.0.10Fo-k:。-K+k.1040.05-?0.2--0:30.4050.6S(b)目0.4首0.310-4Particle sixe号0.。一K;。-K;50 pmg 0.1-100 pm3-150 μm0- =10-2-44- -No erodent0.20.0c)_0.4$三多10-3-1.2 -0.8t0.4 0.80.3|Potentinl (vs Hg/HgO)/Vo-K.◆-Koe+K.0.1Fig.4 Effect of paricle size onpolarization behavior of chromium steeloPotentialNThe addition of alumina particles increased thecurrent density throughout the anodic potential region.And the presence of the small active peak was removedFig.5 Variation of erosion-corrosion masswhen no erodent particles were present ， shifting theloss with potential and erodent particlecorrosion potential negatively by 0. 10 ~0.15 V. Thesize for chromium steelincrease at the end of the passive region near 0 V was(a)- -50μm;( b)-100 um;( c)-150 μmalmost one order of magnitude regardless of the particlesamples had a pronounced effect on the pure corrosionsize added. It can be seen from Fig. 4 that the variationrate ,K。and the erosion-corrosion rate ,Ke。. Especial-of particle size does not greatly affect the polarizationly ,as the potential of the test was moved anodicallybehavior. The similarity despite differing particle sizessuggested that the effect on the electrochemical re-the mass loss due to erosion-corrosion significant in-sponse produced by particle impacts was relatively con-creased at the potential between +0.4V and +0.5 Vstant. Thus the greater number of expected impacts forwhere preferential dissolution was expected with refer-smaller particle sizes seems to compensate for the grea-ence to their respective polarization curves. In addi-ter expected damage resulting from larger particle im-tion , a distinct variation in erosion-corrosion responsepacts. In the passive region equality in the current re-was found with the three particle sizes tested. K.. wassponse occurred between 50 and 100 μm particles withsimilar for both 100 and 150 μ.m erodent particles whilethe response for 150 μum being slightly lower. Howev-for 50 μm particles it was approximately a third of thater , where active preferential dissolution of chromiumfor the other size. This indicated that particle size en-steel at +0.4 V occurred the responses for 100 andhanced the corrosion rate along with increased mass150 μum particles were comparable with that for 50 μmtransfer. Moreover , erosion-corrosion rate ( Ke。) wasbeing lower. It should be noted that the additions of a-apparently higher than the sum( K。+K。) of the purelumina particles ，compared with the solution withouterosion rate( K。) and the pure corrosion rate( K。).particles , accelerated the preferential dissolution of theAs discussed above , the total synergism consistsferrite phase.of twincrement" (△K。) and中国煤化工E In present testing , the3.3 Synergistic effect of erosion-corrosioneffecMYHC N M H Gcludes the fllowing twoThe variation of erosion-corrosion mass loss withaspects'the first is that erosion ( flow ) can en-potential and erodent particle size for chromium steel inhance the transport process of both reactants( such asde-aerated 1 mol/L NaOH at a rotation speed of 6 m/sthe residual oxygen ) reaching the metal surface whichwith 273 g/L of, alumina is shown in Fig. 5.may improve the ability of passivation and repassivationThe et8a插g of electrochemical potential ofof chromium steel and corrosion products leaving theVol. 12 No. 5Effect of preferential dissolution on erosion-corrosion for chromium steel935-metal surface which generally accelerate the corrosionthe preferential dissolution of the chromium steel. Theprocess ; the second is that erosion can apply a me-variation of particle size does not greatly affect the po-chanical force on the metal surface which weakens andlarization behavior. The responses for 100 μm and 150breaks the passive film on metal surface and results inμm particles occurred at +0.4 V are comparable withan increase of internal energy which changes the sur-that for 50 μm being lower.face activity of the metal and the formation of" straindifference cell"[I31.[ REFERENCES ]The main effect of corrosion on erosion is atribu-ted to the preferential dissolution occurred in chromium[1 ] Stack M M ,ZhouS , Newman R C. Identification of tran-steel. The preferential dissolution rate of ferrite phasesition in erosion-corrosion regimes in aqueous environ-ments[J] Wear ,1995 , 186- 187 :523 -532.at potential +0.4 ~ +0. 5 V is large ( as shown inFig.2 ) because of the depletion of main corrosion re-[2] Zhou S , Stack M M , Newman R C. Electrochemicalstudies of anodic dissolution of mild steel in a carbonate-sistant element Cr in matrix which was caused by thebicarbonate buffer under erosion-corrosion conditionsdeposition of( Fe ,Cr )C3. There is a potential differ-[J ] Corrosion Science , 1996 ,38( 7 ) :1071 - 1084.ence between the ferrite phase and of( Fe ,Cr )C3.[3] Li Y , Burstein G T , Hutchings I M. The influence ofTheir anodic curves show that in the vicinity of the cor-corrosion on the erosion of aluminium by aqueous silicarosion potential for chromium steel , the( Fe ,Cr )C3slurries[ J ] Wear ,1995 ,186- 187 :515 -522.phase is more noble than the ferrite phase. Hence , this[4] PittC H ,Chang Y M. Jet slurry corrosive wear of high-chromium cast iron and high carbon steel grinding ball al-results in an accelerated corrosion rate for the ferriteloys[J] Corrosion ,1995 ,42( 6):312 -317.phase and a protection of the( Fe ,Cr )C3 phase. The[5 ] Matsumura M , Oka Y. Slurry erosion-corrosion of com-result of preferential dissolution enhances the erosionmercially pure Iron in fountain-jet testing facility [ A ]by the increase of surface roughness and also by the7th Intermational Conference on Erosion by Liquid andcorrosion removal of the work hardened layer formed bySolid Impac[ C ]. Cambridge ,UK , 1987 ,40/1 -40/8.erosion impacts. The synergism effect of erosion-corro-[ 6] Matsumura M. Erosion-corrosion of metallic materials inslurries[ J ] Corrosion Reviews ,1994 ,12( 3-4) :321sion decreases the bonding strength between the ferrite- 340.matrix and chromium carbide. The latter is gradually[7] Yue Z , Zhou P. Some factors influeneing erosion- corro-exposed and fall off under the frequent impact of abra-sion performance of materials[ A ] Wear of Materials ,sion. This is one of the reason why pitting holesASME C] New York , USA ,1987 ,763 - 768. .orm and total mass loss increases at the potential be-[8] Zhou s , Stack M M , Newman R C. Characterization oftween 0. 4 V and 0.5 V. Meanwhile , it can be seensynergistic effects between erosion and corrosion in an a-from Fig. 5 that the addition of particle enhances thequeous environment using electrochemical techniqud[ J ]Corrosion , 1996 ,52( 12 ):934 - 946.synergism effect.[9] Armstrong R D , Baurhoo I. The dissolution of iron inconcentrated alkali[ J ] Journal of Electroanalytical and4 CONCLUSIONSInterfacial Electrochemistry ,1972 ,40 :325 -338.[ 10] James J S. Investigations of aqueous erosion-corrosion u-1 ) In the polarization curves , the position of thesing rotating cylinder electrodes[ D ]. UMIST 1997.active peaks obtained by potentiodynamic measure-[11 ] Smith W F. Structure and properties of engineering al-ments shows that preferential dissolution for the chromi-loys[ M ]. Materials Science and Engineering Series ，um steel containing 12% Cr occurs in 1 mol/L NaOHUK , 1991.solution with and without particles. The ferrite phase is[12] Zheng Y G ,YaoZM ,WeiX Y ,et al. The synergistieeffct between erosion and corrosion in acidic slurry me-preferentially dissolved at the potential between +0.4dium[J]. Wear ,1995 , 186- 187 :555 -561.V and +0.5V.[13 ] Heidemeyer J. Influence of the plastic deformation of2 ) As the rotation speed increased , the potentialmetals during mixed friction on their chemical reactionpeak of the preferential dissolution( at +0.4~ +0.5rate[J ] Wear ,1981 ,66 :379 - 387.V ) moves to light more negative potential. It shows[14] Wang R , Wang W C ,WangSZ ,et al. A study of sand .that the rotation speed is beneficial to preferential dis-slurry erosion of W-alloy white cast irons[ J ] Wear,1993 ,162 :259 -254.solution for chromium steel.( Edited by YANG Bing )3 ) The additions of alumina particles , comparedwith the solution in the absence of particles , accelerate中国煤化工MHCNMHG.