Investigation on the corrosion process of carbon steel coated by HVOF WC-Co cermets neutral

Investigation on the corrosion process ofcarbon steel coated by HVOF WC/Cocermets in neutral solutionC.Monticelli *,A.Frignani,F.ZucchiDepartment of Chemistry,Corrosion Study Centre ‘‘A.Dacc o ’’,University of Ferrara,Via L.Borsari,46-44100Ferrara,ItalyReceived 5June 2003;accepted 26September 2003AbstractThe corrosion protection afforded to a carbon steel substrate by two cermet coatings (WC/12wt.%Co and WC/17wt.%Co;0.05,0.01and 0.2mm coating thickness),applied by highvelocity oxygen fuel (HVOF)technique,has been studied in 3.5%NaCl solutions.Potentio-dynamic polarization curves of cermet constituents,substrate and coated samples,iron andcobalt dissolution kinetics under potentiostatic conditions and galvanic coupling tests havebeen carried out.Cermet layer hinders the anodic process and WC/17%Co is more protective than WC/12%Co,particularly at high coating thickness.I t is likely that the increase in the matrix cobaltcontent changes the pore morphology,from interconnected to isolated pores,with enhancedprotective efficiency.Ó2003Elsevier Ltd.All rights reserved.Keywords:A.Cermet coatings;A.Cobalt;C.Corrosion;C.Sodium chloride;A.WC1.IntroductionCermet coatings on metals are useful when high surface hardness and resistance towear are required.Higher densities and improved quality of the coating can be*Corresponding author.Tel.:+39-532-291136;fax:+39-532-240709.E-mail address:mtc@unife.it (C.Monticelli).0010-938X/$-see front matter Ó2003Elsevier Ltd.All rights reserved.doi:10.1016/j.corsci.2003.09.013/locate/corsciCorrosion Science 46(2004)1225–12371226 C.Monticelli et al./Corrosion Science46(2004)1225–1237achieved by high velocity oxygen fuel(HVOF)technique,that is a thermal spray process characterized by increased sprayed particle velocity and lower temperatures than those employed by other techniques,such as plasma spraying[1,2].The corrosion process of cermet coated metallic substrates is a very complex phenomenon.In WC base cermets both the corrosion protection efficiency afforded to the substrate(which is usually anodic to the carbide)and the corrosion of the coating constituents themselves have been investigated[3–8].It has been found that the corrosion process on the matrix depends on the type of metal alloy.For instance, in sea water86WC–10Co–4Cr undergoes a general attack which dissolves the matrix near the hard particles due to micro-galvanic effects,while50WC–50Ni–Cr–B–Si is subjected to crevice and pitting[3–5].The dissolution of the metallic substrate strictly depends on the presence of in-terconnected porosity in the coating[1,2,9],which is very difficult to reduce to very low or zero levels.Then the development of microgalvanic effects between the sub-strate and the coating can become important[3,4].I n the case of WC–Co cermets,it is found that the cobalt matrix corrodes just at the WC/Co interface[3,6]and in erosion–corrosion tests this cermet presents a very high corrosion rate,because of the active behaviour of cobalt[10].This research concerns the study of the corrosion behaviour of WC/12wt.%Co and WC/17wt.%Co coated steel specimens in neutral chloride solutions,at different coating thickness and constant and low porosity values,by both electrochemical and chemical techniques.Polarization curves recorded on the base materials(carbon steel substrate,massive or electrodeposited pure cobalt,pure WC and WC/5%Co)as well as on coated samples,have been carried out.Moreover,iron and cobalt dissolution kinetics in potentiostatic conditions have been determined by solution analysis and the presence of galvanic effects has been investigated by short-circuiting steel or cobalt to the cermets.The influence exerted by the coating thickness and matrix amount in the cermets has been taken into consideration.2.ExperimentalWC/12wt.%Co and WC/17wt.%Co cermet coatings were sprayed on shot-blasted steel cylinders(Uni EN10011steel type;nominal composition:0.37–0.44% C;0.5–0.8%Mn;0.15–0.4%Si;balance Fe)by high velocity oxygen fuel(HVOF) technique.These cylinders(10mm height;15mm diameter;4.7cm2lateral surface area exposed to the aggressive solution)had a central threaded hole for their mounting on a rotating shaft for the corrosion tests.The constancy of the experimental conditions in HVOF technique maintained an almost constant porosity in the coatings(measured by image analysis on polished sections of specimens,Table1),while(1or2or4times)repetition of the coating process produced0.05;0.1or0.2mm coating thickness.This procedure induced a decrease in the surface roughness as measured by a Talysurf Plus Taylor Hobson Profilometer(Table1).C.Monticelli et al./Corrosion Science46(2004)1225–12371227 Table1Porosity and surface roughness for the various WC/12%Co and WC/17%Co coatingsCoating type WC/12%Co WC/17%CoThickness/mm0.050.10.20.050.10.2 Porosity/% 2.4±1.0 2.3±0.9 2.0±0.7 1.8±0.9 1.8±1.0 1.7±0.6 Roughness/l m7.5±0.6 5.4±0.4 5.4±0.5 6.9±0.3 4.8±0.5 3.8±0.2 SEM and EPMA analysis revealed that in WC/12%Co coating the hard phase is partially and uniformly dissolved within the matrix,while in WC/17%Co coating WC dissolution in the matrix is inhomogeneous,as locally the matrix shows different tungsten to cobalt content ratios[11].Beside the steel substrate,also pure WC and WC/5wt.%Co discs,99.995wt.% pure cobalt wire and electrodeposited cobalt were studied as reference materials. Massive WC and WC/5%Co were obtained by powder hot pressing;cobalt wire was purchased from Aldrich Chemical Company,I nc.,while electrodeposited cobalt was prepared according to the procedure described in[12].Electrodeposited cobalt was used as prepared,while the massive materials werefinished by grinding with up to n.600emery paper(average surface roughness¼0.093±0.006l m)and degreased with acetone.Most tests were carried out in3.5%NaCl aqueous solution,at room temperature, but some tests were carried out in the same solution with0.25%suspended silica (Silica Gel60,0.2–0.5mm particle size,Fluka product).Potentiodynamic(1mV sÀ1)polarization curves were recorded on coated and uncoated steel electrodes,after1and4h of immersion in the solution at a rotating speed of3000rpm.Polarization curves were also recorded on reference materials under stagnant conditions,after0.5h of immersion.Potentiostatic tests were carried out on coated specimens at+0.030V with respect to their corrosion potential(E cor), 300s after the immersion in the aggressive solution.Iron and cobalt concentrations in the aggressive solution were analysed at intervals during900s by inductive cou-pling plasma–atomic emission spectrometry(ICP–AES)technique.Galvanic coupling effects were checked by recording the galvanic currents and potentials of equal area short-circuited electrodes.All the potentials quoted in the text are referred to the saturated Ag/AgCl ref-erence electrode.3.Results and discussion3.1.Polarization curves of the reference materials(coating constituents and substrate)The polarization curves recorded on steel,cobalt wire,WC and WC/5%Co under stagnant conditions are reported in Fig.1.The corrosion potential(E cor)of the steel is about–0.52V and its anodic polar-ization curve shows a tafelian behaviour with a slope(b a)of0.068V decadeÀ1,closeto the expected value of 0.060V decade À1for iron active dissolution.Cobalt wire(E cor ¼À0:31V)shows an anodic behaviour similar to that of the steel.On the contrary,the anodic polarization curve of WC (E cor ¼þ0:2V),a nobleelectroconducting compound [13,14]is typical for a quasi-passive electrode,withcurrents lower than 50l A cm À2in the investigated potential range.The cathodic polarization curves of these specimens,which refer to the oxygenreduction process,coincide and reach a common limit current (i l ).In the case ofcobalt this curve shows a current peak at about )0.7V,that may be related to thereduction of Co þþions to the metallic state.WC/5%Co has a more active (E cor ¼À0:34V)corrosion potential than pure WC.I n this case the oxidation of cobalt,instead of WC,occurs.For limited polarizations,the slope of the anodic curve of WC/5%Co is low and close to that exhibited bycobalt wire,suggesting the dissolution of surface cobalt regions in the cermet.On thecontrary,at nobler potentials,increasing anodic overpotentials are found,because ahigh ohmic drop is involved in cobalt dissolution through the micro-channels in theWC skeleton.WC/5%Co shows limit cathodic current densities close to those re-corded on the materials previously described and again the cathodic jump at )0.7Vis found.E cor of WC/5%Co is slightly more active than that of cobalt wire and this suggeststhat cobalt dissolution rate in the cermet is faster than that in the massive state,probably as a consequence of the alloying with the carbide [9,15]and/or a moredisjointed metal structure in the cermet [15].I n fact,the effective anodic polarizationcurrent densities of cobalt in WC/5%Co,which are referred to the effective area ofthe cobalt surface region (calculated by assuming that in the cermet the cobaltvolume fraction,V Co ,is equal to the cobalt surface area fraction,A Co [16,17])aremuch higher than those of the cobalt wire (Fig.2).Similar results are obtained withelectrodeposited cobalt,which is more active than WC/5%Co (E corr ¼À0:46V)andits anodic curve is not far from the effective anodic curve of cobalt inWC/5%Co;Fig.1.Polarization curves of massive materials:steel (1),thick line;cobalt wire (2);WC (3);WC/5%Co(4),in 3.5%NaCl solution under stagnant conditions.1228 C.Monticelli et al./Corrosion Science 46(2004)1225–1237therefore the electrodeposited metal can probably represent more closely the cobaltdissolution behaviour in the cermet,owing to its less inert nature.3.2.Polarization curves on coated and uncoated specimensThe comparison between the polarization curves recorded on quiet and rotatingsteel electrodes shows that rotation stimulates the cathodic reaction by one order ofmagnitude,while it exerts much smaller effects on the anodic one (Fig.3).parison of the anodic polarization curves of:cobalt wire (2);electrodeposited cobalt (5);WC/5%Co,referred to the effective area of the cobalt surface in the cermet (40),in 3.5%NaCl solution understagnantconditions.Fig.3.Polarization curves in 3.5%NaCl solution under flowing (3000rpm)conditions:uncoated (thickline)or WC/12%Co coated steel at increasing coating thickness values (0.05;0.1;0.2mm,thin lines);uncoated steel (thick line)under stagnant conditions (0rpm).C.Monticelli et al./Corrosion Science 46(2004)1225–12371229The anodic and cathodic polarization curves recorded on rotating WC/12%Cocoated specimens are close to those recorded on rotating bare steel specimens (Fig.3).The cermet layer exerts a negligible influence on the oxygen reduction reaction,while it can retard the anodic reaction only at high anodic overvoltages.On WC/17%Co coated steel electrodes the action afforded by the cermet is much more evi-dent,particularly on the anodic process;by increasing the coating thickness,lowerand lower anodic polarization currents (with higher anodic slopes)are recorded,while the cathodic polarization currents are scarcely affected (Fig.4).The comparison between the polarization curves of the two coating types at thesame thickness values (0.05and 0.2mm,Fig.5)clearly shows that WC/17%Co ismore protective than WC/12%Co,particularly at high coating thickness.During anodic polarization,underlying iron dissolution through the coating poresis certainly stimulated and,at increasing anodic polarization,concomitant cobaltdissolution is expected.As already evidenced in the case of cobalt dissolution fromWC/5%Co,high ohmic drops are involved in steel dissolution through the coatingpores and cobalt dissolution through WC skeleton pores in the coating,thus in-ducing the high observed anodic slopes.The pores have an interconnected morphology in WC/12%Co coating,as thegradual formation of red insoluble corrosion products (mainly hydrated ferric hy-droxide)is noticed after only 1h of immersion.The same occurs in the case of WC/17%Co at thickness values of 0.05and 0.1mm,while only at the end of 4h im-mersion red corrosion products form in 0.2mm coating.As both WC/12%Co andWC/17%Co coatings exhibit more or less the same porosity,it is likely that thehigher protectiveness offered by the latter,particularly at the highest thickness,islinked to a different pore morphology,i.e.isolated pores in WC/17%Co replaceinterconnected pores inWC/12%Co.Fig.4.Polarization curves in 3.5%NaCl solution,3000rpm:uncoated (thick line)or WC/17%Co coatedsteel at increasing coating thickness values (0.05;0.1;0.2mm,thin lines).1230 C.Monticelli et al./Corrosion Science 46(2004)1225–1237The polarization curves recorded after 4h immersion of the steel specimen show adecrease in the cathodic currents,with a consequent shift of the corrosion potentialin the negative direction (Fig.6).This is attributable to the growth of the ironcorrosion products on the specimen surface,hindering oxygen diffusion from thesolution to the electrode surface.The polarization curves carried out under the sameconditions,in the presence of suspended silica,are close to those recorded on a baresteel surface after 1h immersion,as the suspended particles remove the corrosionproducts and stimulate the cathodicreaction.parison between the polarization curves of WC/17%Co (thick lines)and WC/12%Co (thinlines)coated steel in 3.5%NaCl solution,3000rpm.Coating thickness:0.2mm or 0.05mm.Fig.6.Polarization curves of uncoated steel in 3.5%NaCl solution,3000rpm:after 1or 4h immersion(thin lines);after 4h immersion in the presence of 0.25%suspended silica (thick line).C.Monticelli et al./Corrosion Science 46(2004)1225–12371231Fig.7shows that suspended silica can markedly stimulate the cathodic reactionalso in the case of coated specimens with a low coating thickness,suggesting thepresence of bulky red iron corrosion products on the specimen surface.On thecontrary,no or limited stimulation is detected on coated specimens at the highestcoating thickness,where the formation of iron corrosion products is confined tosmall regions at the pore mouths.This confirms that these thermal spray coatingsoffer an hindrance to the steel dissolution only at high thickness values as the ag-gressive solution reaches the steel substrate with increasing difficulty.3.3.Potentiostatic tests on coated specimensThe concomitant oxidation of steel (iron)and coating matrix (cobalt)is expectedduring corrosion of coated specimen.Therefore,an accelerated test was carried outto evaluate the relative amount of metals dissolved.It consisted in a 900s polari-zation of the coated specimens at +0.030V vs E cor ,followed by chemical analysis ofplanned-interval solution aliquots.Table 2collects the iron ion amounts,the dissolved Fe/Co ratio,measured atincreasing polarization times and dissolved from a single coated specimen,withdifferent coating thickness values.Fig.8shows some examples of dissolved iron andcobalt trends as a function of time.The iron ion amounts are usually greater and often much greater than cobalt ionamounts,even at short polarization times.These observations confirm the presenceof bare steel areas in contact with the solution,at the bottom of interconnected poresin the coatings.Smaller areas are present in thicker coatings,particularly in the caseof WC/17%Co coated specimens.Fig.8also shows that dissolved Fe þþions quickly accumulate in the solution atincreasing polarization time.This trend indicates that the bare steel areasincreaseFig.7.Corrosion rates,extrapolated from the cathodic polarization curves of uncoated and 0.05mm and0.2mm WC/12%Co,WC/17%Co coated steel,after 1or 4h immersion in 3.5%NaCl solution or 4himmersion in 3.5%NaCl +0.25%suspended silica,3000rpm.1232 C.Monticelli et al./Corrosion Science 46(2004)1225–1237Table2Average amounts of dissolved iron ions and iron/cobalt ratio of a coated specimen polarized at+0.030V vs E corCoating type WC/12%Co WC/17%CoThickness/mm0.050.10.20.0500.10.2E appl a/V)0.40)0.37)0.32)0.40)0.36)0.32 Polarization time/s Fe/l g Fe/Co Fe/l g Fe/Co Fe/l g Fe/Co Fe/l g Fe/Co Fe/l g Fe/Co Fe/l g Fe/Co 20025410.872037.0221 5.7743013.92133 3.00360.72 40058913.2458611.126919.26138034.41296 4.17580.81 600147028.11112014.25128013.38270056.49662 6.651080.97 800317048.77171016.93200016.26451084.6211609.21182 1.46 900409061.04202018.36243017.365760103.23145010.43258 1.80a The applied potential vs saturated Ag/AgCl reference electrode.C. Monticelli et al. / Corrosion Science 46(2004) 1225–1237 1233with time,as a consequence of the progressive detachment of the coating due to the corrosion attack on steel at the pore bottom.The same probably occurs under free corrosion conditions,because of the acidification of the solution inside the coating pores,due to the hydrolysis of metal ions.The coated specimen E cor values,measured 300s after the immersion,become nobler as the coatings become thicker and are more or less independent from the coating type.As a consequence,also the applied potentials become nobler at in-creasing coating thickness.Notwithstanding,the amount of dissolved iron ions de-creases as thickness increases.This supports the hypothesis of a decrease in the interconnected porosity at increasing coating thickness.At variance with iron,dissolved Co þþamounts always increase linearly as a function of the time (Fig.8)i.e.cobalt has a constant dissolution rate.The cermet with a higher matrix percentage produces higher amounts of dissolved cobalt ions,which also increase as thickness increases,particularly in the case of 12%Co cermet.This last observation could be linked both to the presence of longer channels through thicker coatings and to the quoted increase in the applied potential at increasing coating thicknessvalues.Fig.8.Dissolution kinetics of iron (solid symbols)and cobalt (open symbols)in WC/12%Co or WC/17%Co coated steel,under potentiostatic conditions (+0.030V vs E cor ).Coating thickness:0.05mm (circles);0.2mm (triangles).1234 C.Monticelli et al./Corrosion Science 46(2004)1225–12373.4.Galvanic coupling testsCobalt–tungsten carbide galvanic coupling,simulated by the couple cobalt-WC/5%Co,can always affect the corrosion process,while steel–cermet coupling can be found only when the solution comes in contact with the steel substrate at the pore bottom in the coatings.Table 3shows both the stable potentials (usually reached within 12h of test)and the charge amounts passed in 24h (Q,referred to a unity electrode area),obtained by short circuiting either steel or cobalt to the cermets,which always result cathodic to the metals.Fig.9shows some examples of the evolution of the short-circuit currents (scC)and short-circuit potentials (scP)recorded during these tests.In the case of coupling steel to tungsten carbide (either WC/5%Co or WC/12%Co or WC/17%Co)the scP shifts from about )0.45V towards more negative values,first quickly and then more slowly.The stable potential values are always close to the steel corrosion potential measured 4h after the immersion in the aggressive solution (Fig.6).This suggests that galvanic corrosion is under cathodic control.By coupling cobalt to WC,a rather similar trend is found,but in this latter case nobler potential values are recorded.Table 3Parameters deduced from the galvanic coupling testsGalvanic couple aQ /C cm 2Stable coupling potential/V Steel––WC/17%Co (coating)2.69)0.58Steel––WC/12%Co (coating)3.25)0.57Steel––WC/5%Co1.58)0.58Co––WC/5%Co1.39)0.35a The first element is theanode.Fig.9.Short circuit potentials and short circuit currents of Co-WC/5%Co coupling (thin lines)or Steel-WC/12‚17%Co coupling (thick lines)in 3.5%NaCl,0rpm.C.Monticelli et al./Corrosion Science 46(2004)1225–123712351236 C.Monticelli et al./Corrosion Science46(2004)1225–1237The scCs of steel coupled either to WC–12%Co or WC–17%Co increase at the beginning of the test and reach almost constant and common values(around35 l A cmÀ2after6–8h),while with steel or cobalt coupled to WC/5%Co scCs reach lower values(around15l A cmÀ2).The galvanic currents measure the stimulation of the corrosion process on the anode due to short circuiting.The Q values obtained for steel or cobalt coupled to WC/5%Co are quite close to each other,probably because the galvanic current density corresponds to the limit current density on the cathode,which is the same in both cases.Higher Q values are measured in the case of steel coupled to WC/12%Co or WC/17%Co coatings,as these coatings have a higher surface roughness and higher effective surface areas than WC/5%Co and induce higher galvanic current densities on the anode.4.Conclusions•The corrosion protection afforded by WC/Co coatings to the underlying steel sub-strate depends on the morphology of the cermet layer porosity.In the case of 17%Co cermet,by increasing the layer thickness(from0.05to0.2mm)it is likely that the interconnected porosity turns into the more protective isolated one.Therefore such coating offers the highest hindrance to the anodic oxidation of the steel,as the access of the aggressive solution to the steel surface is more diffi-cult.•The iron oxidation takes place at the bottom of the cermet pores and by increas-ing the testing time an enlargement of the substrate corroding surface area occurs.•Cobalt and iron dissolution is enhanced by WC coupling.A galvanic current is present on steel or cobalt(anodic)coupled to a WC coating(cathodic)and cor-responds to the limit current of oxygen reduction on the coating itself.The couple activity strictly depends on cermet to exposed substrate area ratios and therefore on the amount of interconnected porosity in the coatings. AcknowledgementsThe Authors wish to thank Mr.D.Contri and Mr.A.Volta(PMC s.r.l.,Pieve di Cento,Bologna,I taly)for their valuable assistance in the preparation of the speci-mens.References[1]A.Neville,T.Hodgkiess,Br.Corros.J.34(4)(1999)262–266.[2]P.L.Cabot,J.Fernadez,J.M.Guilemany,Mater.Sci.Forum289–292(2)(1998)679–687.[3]A.Neville,T.Hodgkiess,Surf.Eng.12(4)(1996)303–312.[4]A.Neville,T.Hodgkiess,J.Am.Ceram.Soc.62(8)(1999)2138–2144.C.Monticelli et al./Corrosion Science46(2004)1225–12371237[5]J.M.Perry,A.Neville,V.A.Wilson,T.Hodgkiess,Surf.Coating Technol.137(2001)43–51.[6]J.M.Guilemany,J.M.de Paco,Surf.Eng.14(2)(1998)129–132.[7]J.M.Guilemany,P.L.Cabot,J.Fernandez,J.M.de Paco,J.S a nchez,pos.Mater.7(3)(1998)205–208.[8]A.Collazo,X.R.Novoa,C.P e rez,Electrochim.Acta44(1999)4289–4296.[9]A.M.Human,H.E.Exner,I nt.J.Refr.Metals Hard Mater.15(1997)65–71.[10]E.Bardal,M.Bjordal,T.G.Eggen,T.Rogne,T.Solem,Prevention of erosion and corrosion inslurries using various inorganic coatings,in:J.M.Costa,A.D.Mercer(Eds.),Progress in the Understanding and Prevention of Corrosion,The I nstitute of Materials,The University Press, Cambridge,1993,pp.158–164.[11]C.Monticelli,A.Frignani,F.Zucchi,Corrosion Behaviour of Thermal Spray Coated Carbon Steel ina Saline Environment,in:Proc.15th CC,Granada(Spain),22–27September2002,paper no.222,CD-Rom.[12]P.Pedeferri,Corrosione e Protezione dei Materiali Metallici,CLUP(Milano)1978.[13]J.D.Voorhies,J.Electrochem.Soc.119(1972)219–222.[14]D.zur Megede,E.Heitz,Werkst.Korros.37(1986)207–214.[15]S.P.Pednekar,Corrosion53(10)(1997)813–819.[16]E.Underwood,Quantitative Stereology,Addison Wesley,1970.[17]Annual Book of ASTM Standards,Standard Practice for Determining the I nclusion or Second PhaseConstituents of Metals by Automatic mage Analysis,E1245,04.14,ASTM,Philadelphia,1995.。