High-temperature-oxidation-resistance-and-surface-electrical-conductivity-of-stainless-steels-with-f

High-temperature oxidation resistance and surface electrical conductivityof stainless steels with filtered arc Cr–Al–N multilayer and/orsuperlattice coatingsP.E.Gannon a,*,C.T.Tripp a ,A.K.Knospe a ,C.V .Ramana a ,M.Deibert a ,R.J.Smith a ,V .I.Gorokhovsky b ,V .Shutthanandan c ,D.Gelles caMontana State University-Bozeman,306Cobleigh Hall,PO Box 173920,Bozeman,MT 59717-3920,United StatesbArcomac Surface Engineering,LLC,Bozeman,MT 59715,United States cPacific Northwest National Laboratory,Richland,WA 99302,United StatesAvailable online 18September 2004AbstractThe requirements of low cost and high-temperature corrosion resistance for bipolar interconnect plates in solid oxide fuel cell (SOFC)stacks has directed attention to the use of metallic alloys with oxidation-resistant coatings.Candidate coatings must exhibit chemical and thermal-mechanical stability and high electrical conductivity during long-term (N 40,000h)exposure to SOFC operating conditions.The high-temperature oxidation resistance and surface electrical conductivity of 304,440A and Crofer-22APU steel coupons,with and without multilayer and/or superlattice coatings from a Cr–Al–N system,were investigated as a function of exposure in an oxidizing atmosphere at high temperatures.The coatings were deposited using large area filtered arc deposition (LAFAD)technology [V .I.Gorokhovsky,R.Bhattacharya,D.G.Bhat,Surf.Coat.Technol.140,(2001),82-92]and subsequently annealed in air at 8008C for varying times.Area-specific resistance (ASR)was measured in air as a function of time and temperature using a four-point technique with Pt paste for electrical contact between facing oxidized coupon surfaces.The surface composition,structure and morphology of the sample coupons were characterized using RBS,nuclear reaction analysis,XPS,SEM and AFM techniques.The structure of the CrN/(CrN/AlN)multilayered superlattice coatings was characterized by transmission electron microscopy (TEM).By altering the architecture of the coating layers,both surface electrical conductivity and oxidation resistance [R.J.Smith,C.Tripp,A.Knospe,C.V .Ramana,A.Kayani,Vladimir Gorokhovsky,V .Shutthanandan,D.S.Gelles,J.Mater.Eng.,(2004),in press]improved significantly for some of the coated samples tested up to ~100h.An order of magnitude decrease in the ASR parabolic growth rate constant was observed in two Crofer 22APU coupons coated with ~1.5A m CrN/AlN superlattice coatings with different nanometrical bilayer periods.D 2004Elsevier B.V .All rights reserved.Keywords:Filtered arc;Chromium;Aluminum;Nitrogen;Nanostructure;Fuel cell1.IntroductionPlanar solid oxide fuel cells (SOFCs)are increasingly promising candidates for future energy conversion due to their inherently high efficiencies and decreased emissions [3].A typical anode-supported planar SOFC design is diagramed in Fig.1.This cell unit is anticipated to produceelectric power at 0.5W/cm 2at an operating voltage of 0.7V with a fuel conversion efficiency of N 40%[4].A 5-kW modular SOFC stack design with 10Â10cm active area at each cell would require ~100individual cells in electrical series.Cell units are separated by hermetic,electrical current collecting and stack supporting interconnect components.During operation,the SOFC interconnect realizes simulta-neous dual atmosphere (wet reducing and oxidizing)exposure under thermally cyclic conditions up to 8008C.Previous ceramic SOFC interconnect materials (CrO 3)operated well at higher temperatures (~10008C),0257-8972/$-see front matter D 2004Elsevier B.V .All rights reserved.doi:10.1016/j.surfcoat.2004.08.067*Corresponding author.Tel.:+14065709449;fax:+14069945308.E-mail address:pgannon@ (P.E.Gannon).Surface &Coatings Technology 188–189(2004)55–61but suffered from high costs and difficulties in fabrication.Recent advancement in SOFC electrolyte materials has reduced operating temperatures to somewhat below 8008C,allowing for a greater selection of interconnect materials.High-temperature metallic alloys have received attention for use as intermediate-temperature (600–8008C)SOFC interconnects due to their higher relative toughness and formability and much lower costs com-pared to their ceramic counterparts.Of particular interest are bcc ferritic stainless steels,which exhibit compatible thermal expansion coefficients with other SOFC compo-nents,but form electrically resistive thermally grown oxide (TGO)scales when exposed to the complex SOFC operating conditions.TGO scales can also introduce adverse chemical and thermal-mechanical incompatibilities with adjoining SOFC components through deleterious species volatilization,interdiffusion and thermal-mechan-ical stresses.A thorough investigation of several heat-resistant alloys concluded that for improved oxidation resistance and electrical conductivity either new alloys will need to be developed or surface engineering of existing alloys will be required [5].Among the candidates in the former category is Crofer 22APU,an Fe-based ferritic stainless steel (22%Cr),with engineered additions of Al,Si,Mn,Ti and La (available from ThyssenKrupp VDM).The present work falls into the latter category of surface engineering,namely the use of multilayered and/or superlattice large area filtered arc deposition (LAFAD)coatings to improve the SOFC-interconnect performance characteristics of promising metallic alloys.Related research in this area includes deposition of different conductive oxide and nitride coatings using various techniques [6–8].This research is aimed at enabling the use of inexpensive alloys for SOFC interconnects by developing a coating material system and deposition process to ensure electrical,chemical and thermal-mechanical compatibility with other SOFC components,throughout the device’s life-time.The use of coatings to improve oxidation resistance on metal alloys has been known for many years.The Cr–Al–N system was selected for this study because it not only offers oxidation resistance at temper-atures up to 9008C [9,10],but also provides wear resistance typical of many metal nitrides [11].Further-more,multilayer structures consisting of alternating AlN and CrN layers have been investigated with the goal to reveal the affect of individual layer thickness on oxidationkinetics and surface electrical properties of the coated metals.It is known that exposing CrN to oxygen at elevated temperatures leads eventually to loss of nitrogen and formation of Cr 2O 3[12],a semiconductor with sufficiently low resistivity for the interconnect application at an operating temperature of 8008C [13].However,volatile Cr(IV)species from Cr 2O 3-based TGO scales electrochemically reduce at the cathode/electrolyte/gas triple phase boundary to form deleterious Cr(III)solid phases.The oxidation of AlN leads to the formation of Al 2O 3with unacceptably low conductivity for the interconnect,yet alumina is known to be a good oxidation-resistant diffusion barrier and may form Cr-retentive phases.Two primary coating architectures have been investigated:the first involves a variety of multilayer structures with nanometer-thick layers of CrN alternating with superlattice layers comprised of several individual layers of AlN and CrN (see Fig.3a);the second design consists entirely of several subnanometer thick superlattice layers of dielectric AlN and conductive CrN.It is expected that the thin alumina layers will be sufficiently discontinuous to have electron conduction pathways,or if continuous,be sufficiently thin to have significant electron transport via tunneling.2.ExperimentalThis coating investigation focuses on three substrates:stainless steels (SS)304and 440A and Crofer 22APU.Coatings composed from the Cr–Al–N system were deposited on circular SS coupons,0.75-in.diameter and 0.125-in.thick and square coupons of Crofer 22APU,0.5in.2and ~0.04-in.thick.The coupons were mechanically polished with 1500grit SiC paper prior to deposition.In the present work,findings are reported from three unique CrAlN coatings on the above mentioned substrates.CrAlN multilayer and/or superlattice coatings were deposited by Arcomac Surface Engineering,using patented LAFAD technology,which utilizes the patented rectangular large area filtered arc source (LAFAS)[1].The LAFAS uses a rectangular plasma-guide chamber with two rectangular deflecting coils installed on opposite sides,as shown in Fig.2.In this design,two primary cathodic arc sources utilizing Cr and Al targets are placed opposite each other on the sidewalls of the plasma-guide chamber,surrounded by rectangular deflecting coils,and separated by ananodicFig.1.Typical planar anode-supported SOFC repeated unit design.P .E.Gannon et al./Surface &Coatings Technology 188–189(2004)55–6156baffle.The LAFAS vapor plasma source uses a super-imposed deflecting magnetic field to turn the metal ion flow 908toward the deposition chamber and substrates.More massive droplets of material from the source follow straighter trajectories and are captured on baffles,resulting in droplet-free coatings.A set of scanning magnetic coils allows the ion plasma jet to be swept in the vertical direction,which can facilitate the coverage of large surface areas [1].The substrates were mounted on pedestals distributed about the outer rim of rotating carousel in the LAFAD chamber.With both sources on,the substrates were successively exposed to Cr then Al ions,in a nitrogen atmosphere,resulting in a bilayer of CrN/AlN.The multi-layer structure shown schematically in Fig.3a was achieved by periodically blocking the Al source,allowing for longer deposition periods of CrN only.Thickness of the individual bilayers in the superlattice substructure of the coating was determined by the rotation speed of the carousel,deposition time and total coating thickness,and averaged ~4.5and ~1.1nm for the 3and 9rpm modes,respectively,for the coatings considered here (Table 1).(The substrate temperature during deposition was about 5008C.)The substrates were first cleaned in an Ar ion plasma at 8Â10À2Pa for 20min,followed by 2min of high voltage Cr ion etching in Ar (2Â10À2Pa).Cr and Al ions were deposited in a nitrogen atmosphere at 4Â10À2Pa with an applied substrate biasvoltage of 50V at 40kHz low-frequency.Coating thickness varied depending on substrate,deposition time and rotationmode.Fig.3.(a)Schematic drawing of the multilayer structure of the LAFAD coatings consisting of repeated sections of CrN and a CrAlN superlattice.(b)TEM image of the multilayer coating showing the CrN (dark bands)and CrN/AlN superlattice (lightbands).Fig.2.Schematic drawing of the LAFAD deposition chamber (top view)showing the dual arc sources and the substrate carousel.P .E.Gannon et al./Surface &Coatings Technology 188–189(2004)55–6157A cross-section of one of the multilayer coatings (CrN/(CrN/AlN)in Table 1)prior to oxidation treatment is shown in the transmission electron microscope (TEM)image in Fig.3b [2].The light colored bands correspond to the lower density CrN/AlN superlattice and the dark bands are the CrN layers.The resolution is not sufficient to reveal the individual subnanometer thick layers of the superlattice within the light band.The wavy character of the bands can be attributed to polishing of the substrate prior to deposition.Total coating thicknesses (Table 1)were estimated using the Calo k test [14],in which a steel ball is rotated on the coating in the presence of submicron diamond slurry.Abrasive diamond particles wear away the coating and expose the underlying substrate.Two distinctive wear scars result:a smaller circular scar exposing the substrate and a larger concentric scar indicating penetration of the coating surface.An example of the Calo k wear scar test is shown in ing the difference in scar diameters and ball radius,an estimation of coating thickness is made through the relationship developed in Ref.[14].Coating adhesion was assessed using the Mercedes indentation test,using a Rockwell C indenter with ~400N load [15].Radial cracks surrounding the indentation indicate good coating adhesion,radial cracks with localized delami-nation indicate fair adhesion and concentric cracks around the indentation with large area delamination indicate poor coating adhesion.The SEM image in Fig.4b provides an example of characteristic good (g)adhesion,and Fig.4c,of poor adhesion (p).The coatings considered here exhibited good adhesion to all substrates.Oxidation of the sample coupons (in Bozeman,MT air)was carried out using a standard furnace operated with no control of humidity or air circulation.Measurements of area specific resistance (ASR)using standard procedures with Pt paste electrodes [5]on pre-oxidized samples vs.time and temperature have been made on coated and uncoated SS 440A,304and Crofer 22APU coupons.A schematic describing the ASR sample measurement apparatus is shown in Fig.5.Prior to ASR measurements,all coupons were oxidized in air at 8008C for 100h,with a furnace temperature ramp rate of 38C/min,to minimize TGO scale spallation.Subsequent to coupon preoxidation,Pt paste was applied to a ~0.5-cm 2contact area on two identical samples and cured for ~30min at 1108C in air.Two Pt wires were spot welded (local oxide scale/coating removed)to the alloys opposite the Pt paste contact area and the sample sandwich was assembled.To ensure sandwich electrical contact,the assembly was placed between two metal blocks (~0.5kg)with alumina spacers.The entire apparatus was then inserted into a bench-top furnace with a type K thermocouple near the apparatus and alumina tubes to insulate the Pt wires.The following experimental method was used.First,the system is heated to 8008C before applying a small and increasing DC voltage across the sandwich with the electrical current recorded;when the current density reaches 0.5A/cm 2,the power supply is set to constant current mode and ASR recorded.The ASR generally decreases for a period of time,reaches a minimum,then increases.Some-time after the ASR reaches a minimum,the furnace temperature is swept from 800to 6008C and backtoFig.4.(a)Calo k wear scar test to estimate coating thickness (photo 0.9Â1.2mm);(b)b Mercedes Q indentation indicating good coating adhesion (g)(on 440A)and poor adhesion (p)(on 304).Table 1Structure parameters of CrAlN coatings on tested substrates Coating Process time (h)Rotation speed,rpm (bilayer period,nm)Thickness a (A m)Performance evaluation presented (X)304440A Crofer 22APU CrN/(CrN/AlN)19(~1.1)~0.5X CrAlN 23(~4.5)~1.6X X X CrAlN 29(~1.1)~1.2X X X Uncoated–––XXXaDetermined either though RBS,TEM or the Calo k wear scar test using SS 304coupons.P .E.Gannon et al./Surface &Coatings Technology 188–189(2004)55–6158provide data for establishing activation energy for electronic conductivity.3.Results and discussionOxidation studies,using RBS,nuclear reaction analysis,XPS,SEM and AFM techniques have been previously performed on multilayered CrN/(CrN/AlN)coatings on SS 304substrates [2].Summarizing these results,some super-lattice coatings provided an order of magnitude increase in oxidation resistance as determined by RBS.In the multi-layered coatings,the authors observed the presence of Fe species on the surface of some oxidized coupons after as little as 1h of oxidation.Some superlattice coatings displayed no Fe presence after 4h.Two explanations were presented:the first was that Fe migrated from the bulk alloy substrate to the surface via cracks in the coatings,which may be attributed to the rapid heating and cooling during that study (room temp to 8008C in 30min),and the second was that the CrN/AlN superlattice structures may have provided an effective Fe diffusion barrier.The focus of the current investigation is to determine the effects on the surface electrical characteristics of the CrAlN coated sample coupons during high temperature oxidation.In this study,Pt paste is used as contact material due to its high temperature stability and conductivity.The SOFC electrode-interconnect contact differs from that of Pt paste;however,Pt paste accommodates surface morphology in a consistent manner and forms a powder-pack with sufficient porosity for continued surface oxidation.Electrical conductivity of metallic alloys is several orders of magnitude greater than TGO scales,which form on their surfaces,and therefore any significant surface electrical resistance is attributed to the TGO.Many TGO scales are known semiconductors (e.g.Cr 2O 3is a p-type semiconduc-tor)[19].Impurities and/or dopants act as either charge carrier donors or acceptors and are generally thermally activated.The relationship between a TGO scale’s ASR and temperature follows:ASR=AT exp(ÀE a /kT ),where A is a pre-exponential constant (includes parabolic growth rate constant and conductivity),T is absolute temperature,k is the Boltzmann constant and E a is the activation energy (including contributions from oxidation,conductivity through the TGO and charge transfers at the alloy/TGO and TGO/contact interfaces)[5].Logarithmic comparison of ASR/T vs.1/T allows extraction of the activation energy.No significant difference in activation energy was determined among all of the samples tested,and values ranged from 0.8to 0.9eV.Investigations are currently in progress to understand how changes in these values may indicate changing conductivity mechanisms.Due to mismatch in thermal expansion coefficients,the Austenitic 304and Martensitic 440A steel alloys are not under consideration for the SOFC interconnect application.However,these steels provide comparable substrate charac-teristics for the coatings studied here and are widely used for other high temperature applications.The ASR results for these samples are presented in Figs.6and 7,and may provide evidence for the efficacy of these,or similar coatings for high temperature oxidation resistance for conventional applications.Fig.6displays ASR vs.time for SS 440A coupons with and without coatings.Most samples reached a minimum ASR during the first hour above 8008C;however,the ASR of a thinner bilayer period (~1.1nm)CrN/AlN superlattice coated sample continued to decrease after 30h.Minimum ASR values of all samples were similar (~35F 10m V cm 2),with the exception of the CrN/(CrN/AlN)multilayer single rotation coating.This coating resulted in a minimum ASR value of b 10m V cm 2,which is significantly lowerthanFig.6.ASR stability on coated and uncoated SS 440A coupons at 8008C inair.Fig.5.Schematic drawing for ASR measurement system.P .E.Gannon et al./Surface &Coatings Technology 188–189(2004)55–6159uncoated coupons,and similar to that of uncoated alloys specifically designed for the SOFC interconnect application [5].This particular coating design also demonstrated superior oxidation resistance(when deposited on SS304) [2].Fig.7displays ASR vs.time for SS304coupons with and without coatings.Two CrN/AlN superlattice coatings with thicker and thinner individual layers are compared with two uncoated304coupons.The uncoated coupons display analogous behavior,with ASR increasing slowly with time. The uncoated coupons exhibited significant spallation of the oxide scales after oxidation.This was not observed on any of the coated coupons and may provide evidence of increased surface stability at high temperatures.The coated coupons require some time to reach a minimum ASR value,with the ASR of the thinner layered CrN/AlN superlattice coated coupon continuing to decrease after100h of testing. This is likely due to different surface chemistry and morphology of the coated samples and their evolution during oxidation,in addition to the changing nature of the electrical contact with the Pt paste.Results from ASR vs.time up to~1200h of coated and uncoated Crofer22APU(observed in this study)are shown in Fig.8.Minimum ASR values occur at various times for these samples,and the acquired values for the uncoated sample are in good agreement with those observed previously[5].After reaching a minimum value,ASR increases and seems to obey the parabolic law,which describes diffusion-controlled oxidation processes[17,18]. Although there is no definitive evidence to explaintheFig.8.ASR stability of coated and uncoated Crofer22APU coupons at8008C in air(SEM cross-sectionsinserted).Fig.7.ASR stability on coated and uncoated SS304coupons at8008C in air.P.E.Gannon et al./Surface&Coatings Technology188–189(2004)55–6160initial drop in ASR,this is likely due to the evolution of surface chemistry and microstructure and Pt powder-pack densification during testing.For example,a slow-forming, Cr-retentive MnCr2O4spinel phase on the surface of similar alloys during high-temperature oxidation has been reported [16].The growth of a TGO scale and therefore ASR (assuming constant scale conductivity),results from the inward diffusion of oxygen,and/or the outward diffusion of oxidizing species,e.g.Cr,Al,Fe and Mn.After minimum ASR values,parabolic growth rate constants for ASR were determined for Crofer22APU samples.Both coated samples display parabolic growth rates an order of magnitude less than the uncoated sample.The ASR parabolic growth rate constant for the coated Crofer22APU with the thinner bilayer period(~1.1nm)CrN/AlN superlattice coating is half that of the thicker bilayer period(~4.5nm)coated sample and data extrapolation shows promise for the SOFC interconnect application.SEM cross-section analysis of the thinner(1.1nm)bilayer period~1.6A m coating on Crofer 22APU after~1200h ASR testing yields a combined scale/ coating thickness of~1.7A m,while the TGO scale grown on an uncoated sample after~750h testing is~3A m.This confirms the dramatic decrease in scale growth of the CrN/ AlN superlattice coatings.A significant difference in ASR vs.time behavior was observed among the samples tested,especially SS304, when the electrical current was applied at room temperature vs.8008C.This may suggest the influence of high voltage fields on TGO morphology.Future work will investigate this further in addition to determining the electrical, chemical and thermal-mechanical compatibility of the interconnect/coating material system with adjoining SOFC components and life-time operation.4.ConclusionsEarly results are reported for the high-temperature surface electrical conductivity and oxidation resistance of 304,440A and Crofer22APU steel alloys with and without LAFAD multilayer and/or superlattice coatings from the Cr–Al–N system.Both short-term oxidation resistance and ASR growth rate were improved significantly for some of the coated coupons[2].Two superlattice CrN/AlN coatings deposited on Crofer 22APU samples resulted in an order of magnitude decrease in the ASR parabolic growth rate constant compared to the uncoated sample.ASR values and growth rates of the thinner bilayer(~1.1nm)CrN/AlN superlattice coating are favorable for the SOFC interconnect application.Remaining evaluation includes other coatings from the Cr–Al–N system as well as those with the addition of oxygen and/ or other elements.AcknowledgmentsWe acknowledge the technical assistance of Norm Williams,Lyman Fellows and John Getty at Montana State University.Coatings were skillfully prepared by Duane Jones and Oleg Popov.Microscopy(AFM)was performed under the supervision of Prof.Dick Smith at ICAL,located at Montana State University.Ion beam analysis was performed at the EMSL,a national scientific user facility located at PNNL and supported by the U.S.Department of Energy’s Office of Biological and Environmental Research. PNNL is a multi-program national laboratory operated for the U.S.DOE by Battelle Memorial Institute under contract no.DE-AC06-76RLO1830.Work at MSU was supported through the High-temperature Electrochemistry Center (HiTEC)supported by a DOI and DOE subcontract from PNNL,number3917(413060-A).References[1]V.I.Gorokhovsky,R.Bhattacharya,D.G.Bhat,Surf.Coat.Technol.140(2001)82.[2]R.J.Smith, C.Tripp, A.Knospe, C.V.Ramana, A.Kayani,V.Gorokhovsky,V.Shutthanandan,D.S.Gelles,J.Mater.Eng.13(3) (2004)295.[3]S.C.Singhal,K.Kendall,High-Temperature Solid Oxide Fuel Cells:Fundamentals,Design and Applications,Elsevier,2004.[4]M.Williams,in:S.C.Singhal,H.Yokakawa(Eds.),Proceedings ofthe7th International Symposium on Solid Oxide Fuel Cells,The Electrochemical Society,Pennington,NJ,2003,p.3.[5]Z.Yang,K.S.weil,D.M.Paxton,J.W.Stevenson,J.Electrochem.Soc.150(9)(2003)A1188.[6]See Proceedings of Materials Solutions2003,13–15October2003,Pittsburgh,PA.(Fuel3.5and Fuel5.5).[7]Y.Yoo,M.Dauga,in:S.C.Singhal,H.Yokakawa(Eds.),Proceedingsof the7th International Symposium on Solid Oxide Fuel Cells,The Electrochemical Society,Pennington,NJ,2001,p.837.[8]N.Oishi,et al.,Surf.Coat.Technol.132(2000)58.[9]M.Kawate,A.K.Hashimoto,T.Suzuki,Surf.Coat.Technol.165(2003)163.[10]O.Banakh,P.E.schmid,R.Sanjines,F.Levy,Surf.Coat.Technol.163–164(2003)57.[11]S.PalDey,S.C.Deevi,Mater.Sci.Eng.,A342(2003)58.[12]F.H.Lu,H.-Y.Chen,C.-H.Hung,J.Vac.Sci.Technol.,A21(3)(2003)671.[13]K.Huang,P.Y.Hou,J.B.Goodenough,Mater.Res.Bull.36(2001)81.[14]V.I.Gorokhovsky,D.G.Bhat,R.Bhattacharya,A.K.Rai,K.Kulkarni,R.Shivpuri,Surf.Coat.Technol.140(3)(2001)215(b CALO test Q).[15]W.-D.Munz,T.Hurkmans,G.Keiren,T.Trinh,J.Vac.Sci.Technol.,A11(5)(1993(Sep./Oct.))2583(HR c adhesion test).[16]D.M.England,PhD thesis,The University of Utah,2001.[17]C.Wagner,Atom Movements,American Society of Metals,Cleve-land,OH,1951.[18]K.Hauffe,Oxidation of Metals,Plenum Press,New York,1965.[19]K.Huang,P.Y.Hou,J.B.Goodenough,Solid State Ionics129(2000)237.P.E.Gannon et al./Surface&Coatings Technology188–189(2004)55–6161。