Coatings for the Protection of Low-Alloy Carbon Steels Against Corrosive Attack in Waste Heat Recovery Systems of Coal Gasifiers

7/27/2019 Coatings for the Protection of Low-Alloy Carbon Steels Against Corrosive Attack in Waste Heat Recovery Systems

1/16

1157

COATINGS FOR THE PROTECTION OF LOW-ALLOY AND CARBON STEELS AGAINSTCORROSIVE ATTACK lN WASTE REAT RECOVERY SYSTEMS OF COAL GASIFIERS

The presence of sulfur, oxygen, and carbon in the gaseous environmentof a coa1 gasifier provides conditions in which rapid degradation ofstructural materials may occur. Corrosive attack almost invariablyresults in a 10ss of load-bearing capacity of structural components, andthe combined effects of corrosion and applied stress will promote premature failure. The physical and mechanical propert y requirements ofmany heat exchanger components used interna11y or downstream from thegasifier vesse1 demand that the materials of construction be metallic.High-strength, corrosion-resistant a110ys such as austenitic stainlesssteels meet the requirements for applications at temperatures up toabout 950C in aIl but the most severe gaseous environments. At theoperating temperatures of steam generators and superheaters. usually notexceeding 350C, 10w-al1oy steels such as Cr-Mo steels and carbon steels

D. J. BaxterMaterials Science and Technology DivisionArgonne National Laboratory9700 South Cass AvenueArgonne, Illinois 60439 USA

1. INTRODUCTION

ABSTRACT. The poor corrosion resistance of low-alloy, boiler-gradesteels at the temperatures of future coal gasifier waste heat recoverysystems (typically exceeding 400C) and the relatively high cost ofaustenitic stainless steels necessitate the development and evaluationof protective coatings for low-alloy steels. Only coatings that couldbe applied to large components were used for the experimental investigation. Pack-diffusion aluminized, chromized, and aluminized/chromizedcoatings with various chemical compositions and microstructures andfurnace-fused FeCrAl and CoCrAl coatings were exposed to simulatedmedium-Btu coal gasifier environments under both isothermal and thermalcycling conditions. Four welding a110ys on aluminized substrates werealso evaluated. Minimum Al and Cr contents of 14-20 wt % and 18-22 wt %, respective1y,.were required to support protective oxidationbehavior under isotherma1 conditions, although the presence of defects,particularly cracks, in aluminized and furnace-fused FeCrAl and CoCrAlcoatings can promote severe internaI degradation. ln sorne cases,thermal cyc1ing promoted sulfidation attack. AlI welding a1loysexhibited high rates of corrosive degradation.-


2/16

1158

possess the required properties for prolonged use. A clear economicincentive exists to operate steam generators and superheaters at higher

temperatures, typically up to 650C. UnfortunatelY,at the higher tem- l!eratures, the low-alloy and carbon steels exhibit poor resistance to 'corrosion and the high-strength, corrosion-resistant alloys impose a ~1signlflcant cost penalty. Thus, a program was initiated to develop andevaluate coatings and claddings to upgrade the corrosion resistance of aboller-grade, low-alloy steel, 2 1/4Cr-lMo, and Al06 carbon steel inenvironments typical of those expected in a downstream coal gasifierheat recovery system. The desired service life of components ln such asystem ls on the order of 100,000 h.

Heat exchanger tubing and paneling in heat recovery systems of coalgasifiers may be as much as 12-15 m in length. Owing to technical limitations and cost, this size clearly precludes the use of more exoticcoating techniques such as low-pressure plasma spraylng, physical andchemical vapor deposition, and processes requiring elaborate postcoati~j application treatments such as laser densification. Previousworkl has shown that two processes, pack diffusion and furnace fusionof sprayed metallic powders, may be used to produce metallic coatings onlarge components at reasonable cost. Pack diffusion involves soakingthe substrate material at elevated temperature in a mixture of the oxideof the metal to be applied as a coating, the metal itself, and a halideactivator. Good furnace-fused coatings are obtained with coating alloysthat fuse in a narrow temperature range and easily wet and flow on thesubstrate. Various compositions of MerAl alloys ha)e been successfullyapplied as coatings on austenitic stainless steels. These twotechniques were thus employed to obtain coatings rich in the"protectiveoxide-formlng elements Cr and/or Al on low-alloy and carbon steelsubstrates.

A number of coated low-alloy 2 1/4Cr-lMo and Al06 carbon steelspecimens were obtained from a range of commercial coating suppliera.ln order ta evaluate the capacity ta join coated components, a number ofprecoated specimens were welded using four different welding alloys.Coated and coated/welded specimens were exposed to simulated coalgasification environments in order to determine the influence of coatingtechnique, thickness, and composition as weIl as of the presence ofdefects on the corrosion resistance imparted.

2. MATERIALSTen different coatings on each of the two steel substrates were appliedby a number of coating suppliers using commercial or close to regularcommercial processes. (The proprietary nature of each of the processesprecluded the release of de ta Ils of the conditions for processing bythe individual suppliers.) The steels were aluminized using threeprocesses. chromized using two processes, simultaneously aluminized/chromized using two processes. sequentially chromized/aluminized usingone process, and coated by furnace fusion of alloy powders using oneprocess. Two different alloy compositions were used for the latter, onean FeCrAl alloy containing 20 wt % Cr and 45 wt % Al, and the other a


3/16

t

, clear economic\ hea ters at higher,y~t the higher tem-poor resistance to

,nt alloys impose atiated to develop androsion resistance of a06 carbon steel inream coal gasifiercomponents in such a

covery systems of coal'ing to technical limiuse of more exoticaying, physical andelaborate post

fication. Previouson and furnace fusione metallic coatings on~n involves soakinga mixture of the oxideitself, and a halide

ed with coating alloyswet and flow on thea~e been successfully These twoich in the protectivend carbon steelAl06 carbon steelcoating suppliers.

omoonents, a number ofr ~lding alloys.__ ~ulated coal

e influence of coatingf the presence of

bstrates were appliedDr close to regulareach of the processesfor processing by

ized using threeeously aluminizedlzed/aluminized usingpowders using one

ed for the latter, one~l, and the other a

1159

CoCrAl alloy containing Z4 wt % Cr and 46 wt % Al. The bulk chemicalcompositions and thicknesses of aIl coatings are given in Table 1. Thepoor adhesion of MCrAl coatings to ferritic steels necessitated theapplication of a Ni flash to the surfaces of the steels prior to coatingwith the a1loys. As a resu1t, the coatings contained Ni. AlI coatedspecimens were examined visua1ly and in a very explora tory manner byeddy current and u1trasonic nondestructive evaluation techniques forevidence of defects prior to corrosion testing. The limited effortplaced on the use of these techniques, however, resulted in failure todetect defects, such as cracks which were revealed by metallographiccross sectioning.

Specimens of steels aluminized utilizing one of the three processeshad V-notches cut through the coating and into the substrate and werethen welded using each or the four welding alloys whose nominal compositions are given in Table II. Owing to the disparity in chemicalcompositions among the welding alloys, coatings, and substrates,considerable dilution of weld alloy elements in weld beads occurred(Table Il). A variation in the concentration of each element of up to8% in the weld deposits from specimen to specimen was observed.

3. EXPERIMENTAL DETAILS

Tubular specimens lZ.7 mm long with an outer diameter of Zl mm and awall thickness of 2.5 mm were coated on the outer surfaces. Theincorporation of male and female joints at the opposite euds of eachspecimen permitted a stack of specimens to be axia11y ioterlocked toforro a column. The column configuration permitted the inner specimensurfaces to be cooled, ,using steam, while the outer surfaces wereexposed to a desired gas mixture using a test rig described elsewhere.4Tests of 1000 h io duration were carried out under the heat fluxconditions with temperatures for the mixed gas of 871C and for thespecimen surfaces of 600C. Two simulated medium-Btu product gasmixtures with the components CO, COZ' CH4, HZS, H~O. and HZ were used_giving oxygen and ~~lfur partial pressures at 600 C of POZ = 8.1 x 10 19and PSZ = 5.4 x 10 Pa for gas mixture A (the 10w-PS2 gas mixture),and PO = 4.Z x 10-20 and PSZ = 4.9 x 10-4 Pa for gas mixture Be ! - gas mixture). Gas mixture B was also used for thelcycling tests in which specimens were exposed at the above tempe raturesfor lOO-fi pe lods between thermal cycles to a metal temperature oflSOC. Each test consisted of 10 cycles. The upper liroit of HZS usedin the investigation (gas mixture B) was 1.5 vol %. ln both gasmixtures, CrZ03' AlZ03, FeS, and C09S8 were thermodynamically stablehases. Uncoated Incoloy 800 and Z 1/4Cr-lMo specimens were included lneach test as reference materials.

4. RESULTSDetailed characterizations of the as-received coated materials beforethe corrosion tests and after exposure to the gaseous environments were


4/16

il:1i

'!IL~

-ij~.ti1:

-1.~~ilJ~1

..~~*.~J.....

1160

TABLE I. Chemica1 compositions and thicknesses of coatings used in theinvestigation

Coating andDiffusiononeing Composition (wt %)

ThicknessCoating Typelreoi (~) Aluminized0.9.76.4-00-210

128.9.39.8-20-230 21.4.76.9 70-80 25.24.8-90-100 24.5.24.3 300-350 26.53.5-00-350Chromized-3.176.960-65 !8.321.710-12 -7.482.670-75 -2.517.515-20Simultaneous1y0.9.04.1-BO-2002-6 ~31.7.B2.6-00-220 13.6.14.2-50-270 14.6.45.950-270Sequentially3.38.97.890-1001-9 Furnace-fused FeCrAl3.6.06.3-5.1 60-70urnace-fused FeCrAl4.2.52.0-4.3 60-70urnace-fused CoCrAl2.4.1l.2.16.230-60urnace-fused CoCrAl4.84.5.80.1.90-BOts the steel substrate: 1 = 2 1/4Cr-lMo; 2 = A106 carbon steel. The second number is the coating

performed with the aim of deriving a specification for a coating whichmay eventua11y be used in a commercial gasifier system. ln view of thevolume of information produced from the microstructural characterizationof a large number of specimens5 and the space limitations of this paper,on1y a short summary of the results obtained is included here.


5/16

1161..;~:(~;..r

"--'Jsed in the

:::?;"

TABLE II. Nominal compositions of welding a110ys and typica1 compositions of weld deposits

:rate:Jer is the coating

for a coating which:tem. ln view of the:ural characterization.ations of this paper,luded here.

_(wt %)Co Ni

1. ~ovJ-QK.y ,J('lDA(... . -I,.Lt \VS~ -r( ~J~- ~SC~Vbvvf

Composition (wt %)Weld Metala

Feriol Other 2.52.0al.9.0 Mo, 3.6 Nb(W)

61.3.02.9.32.4 Mo, 0.7 Nb(R)4.0Bal..7 Ti(W)

56.29.62.5 1.5.3 Ti (R).00.00.0al.5.0 W, 1.5 Mn(W)

48.21.4.14.6.0.9 W, 0.7 MnR)al.3.03.02.0 Mn, 1.0 Si(W)

88.3.2.5-.91.1 Mn, 0.8 Sicomposition of welding rod.

W = typical composition of weldment.4.1.

Characterization of As-received Coated/Welded Steels

The chemica1 compositions of aIl coatings, determined by an x-ray energydispersive analysis technique, are given in Table 1. ln each case, thechemical composition represents the volume of material just beneath theouter specimen surface. ln general, aluminized material had a maximumAl content inthe range 20-29 wt %. The Al concentration was usuallyhlgher in the carbon steel than in the low-alloy steel, although thecoatings on the low-alloy steel also contained Cr. The Al concentrationln this range was maintained for approximately one-third of the totalcoatlng plus the diffusion zone thickness given in Table 1. A graduaIdecrease in Al concentration and microhardness from the layer of maximumAl content to the substrate characterized the diffusion zone. DiscreteAl oxide particles and oxide colonies were present in the outer reglonof coatings, and fine cracks extended perpendicularly from the surfaceto approximately one-third of the coating thicknesses. An additionalphase, a mixed Fe,Al carbide with an acicular morphology, was alsopresent in the diffusion zone of the aluminized carbon steel. The grainstructure of aIl aluminized material was columnar with individual grainsextending through at least half. the coatlng thickness

ln thickness and structure, the simultaneously aluminized/chromizedsteels were similar to the aluminized steels, but the chemical compositions varied depending on the supplier, and thus, the process. Coatingscontaining 31% Al, 5.5% Cr, and 14% Al, 0.5% Cr were obtained. Thesequential chromizing/aluroinizing process could not be achieved for theAI06 carbon steel; however, a two-layer coating comprised of an inner

:.:."

.:.:~::.:

~.:\".

;';:.:'

~~.

:.;:"

".~

::1-1.Jl

60-7060-7030--6030-80

60-6510-1270-7515-20

200-210220-23070-8090-100

300-350300-350

180-200200-220250-270250-27090-100

Coating andDiffusion

ZOneThickness

(IJID)

25.114.356.27.97. _).1

"----


6/16

.'11:1.~J1

i'~li11~~:~t'~fihd,"11:', il.~*~,~~,'~li'~.'

,~~ ~,.~,!{~~il:l.:'$il~~~:f

'1~Il

1

1L

1162

layer rich in Cr and an outer 1ayer containing both Al and Cr wasobtained on the low-a1loy steel. Cracks were present in both layers.

The characteristics of coatings obtained by chromizing weredependent upon the composition of the substrate on which they wereapplied. On carbon steel, coatings were 10-ZO ~m thick, with a highvolume fraction of MZ3C6 and M7C3 carbides giving a net Cr content of78-83%. A steep Cr concentration gradient from the 78-83% level to thesubstrate reflected a very narrow diffusion zone, while carbon depletionin the substrate extended to a depth of 1Z0 ~m. The coatings wereapparently defect-free. Coatings 60-75 ~m thick and containing amaximum Cr concentration in the range 16-Z5% were obtained on the2 1/4Cr-lMo steel. ln some cases, an outer carbide layer, up to8 ~m thick, was accompanied by numerous volds at the carbide/Fe,Crinterface. The grain structure of the coatings on the 10w-alloy steelwas columnar and an abrupt change in Cr concentration between thecoating and the substrate was observed.

Costings produced by furnace fusion of sprayed metal powders weremultiphased with the bulk chemical compositions given in Table 1. BothCoCrAl and FeCrAl coatings contained (Ni,Fe)AI precipitates in an FeCrAIor FeCoCrAl solid solution. Cracks from the outer surfaces traversed90% of the coating thickness and numerous discrete voids were presentmainly in the broad diffusion zone.

The main feature of the weld deposits was their chemical composition (Table II). Reductions of the nominal Cr contents of the weld rodsby approximately 50% occurred upon welding to the low-alloy or carbonsteel substrates, although some pickup of Al from coatings occurred.A very narrow diffusion zone existed at the weld metal/coating interfaceand no cracking was observed.4.2. Microstructural Examination of Corrosion-tested Materials

AlI specimens were sectioned axially, mounted in epoxy resin, andpolished prior to microstructural examination. The corrosion scalethickness and depths of interna! attack are summarized in Table III.The uncoated 2 l/4Cr-lMo steel formed Fe sulfide Bcales (>500 ~m thick)in both gas mixtures A and B. Combined with internaI attack of thesubstrate ()ZOO ~m deep), the rate of sound metal loss clearly precludedpractical application of the steel in the uncoated condition. Incoloy800 exhibited protective oxidation behavior in gas mixture A, but atthe higher-PSZ and 10wer-pZ levels of gas mixture B, localized breakaway corrosion behavior, characterized by the presence of Fe-richsulfide blisters both on the surface and in the substrate, was observed.Thermal cycling resulted in similar breakaway behavior.4.2.1. Aluminized substrates. Isothermal exposure to gas mixture Aresulted in protectlve oxidation behavior of aIl aluminized materialswith the exception of those obtained from one supplier where localizedsuIfidation occurred, both at the surface of the scale and internally.Sulfide nodules rich ln Fe were formed over areas where internaIoxidation and sulfidation appeared to have been facilitated by thepresence of localized multibranched cracks. Single fine cracks


7/16

1163

TABLE III. Summary of corrosion behavior under the three sets ofconditions

contributed litt1e to the corrosion process in gas mixture A, whereas ingas mixture B, small su1fide particles were found in almost aIl cracksin a1uminized steels. ln gas mixture B, the surface scales werecomposed predominantly of oxide, and intragranular oxidation of thecoating occurred to a depth not exceeding 40 Ilm ( Fig. la).

Exposure to gas mixture B under thermal cycling conditions resu1tedin sulfide scale formation and enhanced internaI attack. The scalesranging in thickness from 50-300 Ilm (Fig. lb) were composed mainly ofFe sulfide with sorne Al enrichment close to the sea1e/metal interface.The Iayered structure of the scales provided evidenee supporting the

32580

2503002020

10015018040

6080

10010010o

120702530

30080

PenetrationMixturesB TIc(Ilm) (Ilm)


8/16

"(~-.',

.",. .... , ..~4' ~-~..: :..

>';'20 fU!l, f-!-f

1164

Figure 1. Cross sections of aluminized carbon steel.after exposure tothe high-PS2 gas mixture B under (a) isothermai conditions showing oxidescaie and intragranular oxidation, and (b) thermal cycling conditionsshowing sulfide scale and Iocalized intergranular oxidation/sulfidationattack.

occurrence of mechanicai damage during thermal cycling, but goodadhesion to the substrates prevented severe scaie spallation. InternaIoxidation/suifidation of the coatings, both intragranularly and via finepreexisting cracks, was very light, but more severe Iocalized intergranular attack occurred to a depth of approximately 80% of thethickness of ~oatings (Fig. Ib). Both oxide and sulfide coexisted inthe narrow confines of the grain boundaries, which became a path of easytransport of oxygen and sulfur to greater depths in the coatings whereprogressively lower concentrations of Al were present and where lateralgrowth of the corrosion products occurred. ln Fig. Ib, Iateral growthof the corrosion products coincides with the region where the acicularcarbide phase is present in the aluminized carbon steel. Within theduration of the 1000-h test, complete penetration of the coatings due tointernaI attaek was not observed.

4.2.2. Aluminized/chromized substrates. The simultaneously aluminized/ehromized steels with eoatings eontaining 14% Al and 0.4-2% Cr exhibitedbreakaway corrosion behavior upon exposure to gas mixture A. Seaiesrich in Fe sulfide and up to 100 ~ thiek formed over the coatings whichsuffered severe internal degradation to a maximum depth equai to theirthickness. Both oxides and sulfides of Fe and Al formed internally.Coatings of the above composition were not tested further.

Proteetive oxidation behavior was exhibited by the simultaneous1yaluminized/chromized steels eontaining 31% Al and 5% Cr upon exposure toboth gas mixtures A and B. The seales were Al oxide eontaining some Feand no more than 10 ~m thick. InternaI attack of the coatings resulted

:t:.

'--


9/16

~el .after exposure tolnditions showing oxide_ cycling conditionsoxidation/sulfldation

:Iing, but goodspallation. InternaI:ranularly and via fine'e localized inter-,ly 80% of the:uIfide coexisted in1 became a path of easy.r ~ coatings where:e._ and where lateral;. Ib, Iaterai growth,n where the acicularsteel. Within theof the coatings due to

;ltaneously aluminized/.nd 0.4-2% Cr exhibitedmixture A. Scales,ver the coatings whichdepth equal to theirformed internally.further.'Y the simuitaneously5% Cr upon exposure to.de containing some Fethe coatings resuited

1165

in intragranular oxidation close to the outer surface and light oxidation via preexisting cracks to a depth tiot exceeding 30 ~m in coatingsapproximately 200 ~m thick. Thermal cycling ln gas mixture B resultedin the formation of discrete Fe-rich sulfide nodules on the outersurfaces of thin oxide scales. Localized internaI attack by a modesimilar to that observed in aluminized material (Fig. lb) reached amaximum depth of 180 ~m in the coatlng on the 2 1/4Cr-1Mo steel, butonly 50 ~ in the coatlng on the carbon steel.

The sequentially chromized/aluminized 2 1/4Cr-lMo steel underwentsevere localized internaI attack to 90% of the coating thickness in gasmixture A although thln surface oxide scales were formed. Both oxideand sulfide particies were detectedin grain boundary regions in thecoatlng where Cr carbides were located in the as-received materiai. lnaddition, corrosion products were found within the void region markingthe interface between the two Iayers of the coating. The poor resistance to corrosive degradation in gas mixture A precluded furtherevaluation of the sequentially chromized/aluminized steel in moreaggressive environments.4.2.3. Chromized substrates. ln gas mixture A, aIl chromized materialexhibited protective oxidation behavior, characterized by the formationof thin Fe,Cr surface oxide scales. InternaI oxidation occurred in thevoid regions beneath the outer thin carbide layer on the chromized2 1/4Cr-lMo steel, but did not penetrate the body of the coatings.Evidence of cracking in the carbide-rich coating on the carbon steel wasfound after corrosion testing. As a result of the possible cracking,internaI oxides formed across the thickness of the coating (Fig. 2).

Exposure to gas mixture B produced contrasting behavior dependingon the Cr content of th~ coatings. Coatings on the 2 1/4Cr-lMo steelcontaining 22-25% Cr oxidized in a protective manner although small,discrete nodules of Cr sulfide were present on the outer surface of theCr-rich oxide. Light internaI oxidation, but not sulfidation, occurred.Coatings containing 16-18% Cr exhibited breakaway corrosion behaviorwith the formation of Fe-rich suifide seales up to 120 ~m thick andinternaI attack extending beyond the coating thickness into thesubstrate (Flg. 3). Despite the high Cr content of a carblde-richcoating on the carbon steel (coating 2-5), total breakdown of theprotection afforded by the coating occurred. ln vlew of the crackingbehavior of the same coating upon exposure'to gas mixture A (Fig. 2),the breakdown of protection in gas mixture B is tentatively attributedto the same cause. The similar carbide-rich coating on the carbon steelproduced by a different supplier (coating 2-4) resisted cracking and,therefore, breakaway corrosion was inhibited.

Exposure of the surviving chromized 2 1/4Cr-lMo and carbon steels(coatings 1-4 and 2-4) to gas mixture B under thermal-cycling conditionsresulted in growth of protective oxide scales over which sma11, discreteCr-rich sulfide particles were present. InternaI oxidation, with only atrace of sulfur, occurred intergranularly and around voids in thecoating on the 2 1/4Cr-1Mo steel (Fig. 4a), while internaI oxide formedin thermal cycling induced cracks in the eoating on the carbon steel


10/16

.:~I:!

oxide

-Substrate

33fLm~

8 !-Lmf---!---t

1166

Figure 2. Cross section of chromized carbon steel showing oxide formation locally across the entire carbide-rich coating thickness afterisothermal exposure to the low-PS2 environment A.

Figure 3. Cross section of the chromized (containing 17% Cr) 2 1/4CrlMo steel with sulfur concentration profile after isothermal exposure tothe high-PS2 gas mixture B.

~.~.l.~j~;Il

,tll[[1"1\:1

'.11.~~:..\~tt

-,ii1~r


11/16

'---

ternaloxide

ubstrate

l showing oxide formag thickness after

ing 17% Gr) 2 1/4Grisothermal exposure to

1167

- Cr sulfidescale.. ',.~- .~t -Cr oxide scale

.., r' 1.. 1"~.' . '.f't# ---Void. ". ~"-,-""' 1i r /; . t Intergranularxide~

-Coating

Figure 4. Gross sections of (a) 23% Gr-containing chromized 2 1/4Gr-lMosteel showing surface oxide and sulfide layers and internaI attack, and(b) the chromized carbon steel in which cracking of the carbide-richcoating has occurred, both after thermal cycling exposures in thehigh-PS2 environment B.

(Fig. 4b). ln neither case was the coating completely breached bycorrosive attack.

4.2.4. Furnace-fused coated substrates. The FeGrAl and GoGrAI coatingson both steel substrates underwent marked degradation upon exposure tothe low-PS2 environment of gas mixture A. Thin oxide scales formed onthe FeGrAl coatings, but internaI oxide and sulfide formation via preexisting cracks occurred to depths approximately 90% of the coatingthickness. Scales on GoGrAI coatings were nonuniformly thick andcontained large oxide-rich nodules accompanied by severe internaI attack(Fig. 5). ln the outer region of the coatings, oxide and sulfideparticles were present on grain boundaries. Transport of oxygen andsulfur through the coating was evident by the presence mainly ofsulfides in the void region of the coating adjacent to the coating/substrate interface. ln the substrate, sulfide rich in Fe was the majorcorrosion product. Although similar coatings exhibited slight differences in corrosion behavior on the same steel substrate, the severity ofcorrosive degradation of any one of the coatings in gas mixture Aprecluded consideration in further corrosion tests.


12/16

Figure 5. Cross section of the CoCrAl-coated carbon steel showlnginternaI oxidation and su1fidation attack of the coating and the substrate, resulting from isothermal exposure to the 10w-PS2 environment A.

4.2.5. Coated/we1ded substrates. Protective oxlde scales formed on theIncone1 625 and Inconel 72 we1d deposits upon exposure to gas mixture A,whereas the L605 and 309 SS weld deposits exhibited breakaway corrosionbehavior, characterized by the formation of thick multilayer scales andthe occurrence of deep internaI attack. Protective scales on the Nibase al10ys were composed of Cr-rich Fe,Cr oxides, with internaI, intergranular oxldation extending to 'a depth not exceeding 25 Ilm. PorousFe-rich oxide- and su1fide-containing scaies enabled both intergranularand intragranular oxldatlon and suifidation of the Co- and Fe-base weldmetals to depths of 90 and 40 ~m, respectively.

Upon exposure to the higher-psZ environment of gas mixture B, boththe lN 625 and lN 72 weld metals fa11ed to form protective surfacescales. Scales composed mainly of porous Fe sulfide with Fe,Ni sulfideouter reglon (up to 500 ~ thick) formed on the lN 625 weld metal.InternaI attack produced Cr-rich Fe,Cr suifide to depths of 100-300 ~m. Attack of the higher Cr-containing lN 72 weld metal was comparatively less severe with scales 25-100 ~m thick and internaI penetration to a depth of 80 ~m. Corrosive attack of coatlngs immediatelyadjacent to the weld metals, in the heat-affected zones of the welds,was no more severe than that occurrlng in areas of coating not affectedby the welding process. Degradation of the weld metals was slightlymore advanced under thermal cycling conditions in gas mixture B.

1168

5 DlSCUSS ION

The severe corrosive degradation of the uncoated 2 1/4Cr-lMo steel inboth environments used in the investigation clearly precludes consideration of the alloy ln the uncoated state as a structural materiai


13/16

bon steel showingcoating and the sublow-PS2 environment A.

de scales formed on theosure to gas mixture A,ed breakaway corrosionmultilayer scales and

ve scales on the Ni-, with internaI, interding 25 ~m. Porousled both intergranulare Co- and Fe-base weld

o ,s mixture B, bothr~ctive surfaceide with Fe,Ni sulfide~ 62S weld metal.depths of 100-weld metal was com

k and internaI penecoatings immediatelyzones of the welds,

f coating not affectednetals was slightlygas mixture B.

~ 1/4Cr-lMo steel in,y precludes con-1 structural material

p.0.'

1169

in advanced steam generator or superheater systems. Metallic coatingspromote resistance to corrosive degradation, but the degree to whicheach coating investigated was effective ln minimlzing corrosive attackwas dependent upon a number of variables. For the 1000-h tests carriedout under both isothermal and thermal cycling conditions, coating composition and defects played important roles in the degradation process.5.1. The Corrosion Behavior of CoatingsCoating composition is an important factor influencing resistance tocorrosion. While a minimum of 10-12 wt % Al may be necessary to ensureprotec6ive oxidation kinetics for Fe-Al alloys under oxidizing conditions, the minimum under oxidizing/sulfidizing conditions would appearto exceed this level. Coatings containing 20-29% Al generally exhibitedprotective oxidation behavior under isothermal test conditions, but at14% Al, even with 0.4-2% Cr, breakaway corrosion occurred. Up to 29% Alwas Inadequate in preventing sulfidation from taking place under thermalcycling conditions, but the combined presence of 31% Al and 5% Crlimited the severity of breakaway attack. Increasing the Al contentbeyond 31%, however, may not automatically induce improvements incorrosion resistance since the brittle phases7 FeAl) and Fe2AIS becomestable and the disparity in the thermal expansion coefficients betweencoating and substrate increases, both of which increase the likelihoodof cracking.

InternaI oxidation was a notable feature of aIl aluminizedmaterial, although the additional presence of Cr reduced the effect.Formation of internaI Al oxide results in depletion of Al from the Fe-Alcoating material, which leaves less Al available for protective surfacescale growth and increases the likelihood of FeS participating in thecorrosion reaction. The kinetics of internaI oxidation were not determined, but the degree of Al deplet10n would appear small. The presenceof a fine dispersion of integnal oxide particles could ev en be beneficial to corrosion resistance by promoting easy transport of Al to thecoating surface and by reducing the thermal expansion coefficientmismatch between the coating and the surface scale. The latter would beparticu1arly advantageous in inhibiting mechanical breakdown of protective surface scales during thermal cycling. Preexisting cracks mayserve more than one function. Large, open cracks permit the inwardtransport of gases into the coating where oxygen and sulfur activitiesmay diverge substantially from those of the bulk gas mixture and rapidinternaI corrosion may ensue. It was observed that very fine crackshad a minor influence on the severity of internaI attack under bothisothermal and thermal cycling conditions. Fine cracks may evenfacilitate the accommodation of stress generated during thermal transients since the y did not appear to be associated with deep localizedattack in aluminized specimens during the thermal cycling tests.

The corrosion resistance of chromized steels was criticallydependent upon the coating composition and the presence of defects.Coatings with a range of Cr contents were tested. The range of Crcontents arose from the different application techniques adopted bydifferent suppliers. ln the region of 17% Cr on the 2 1/4Cr-lMo

----- __'w- mu u uo_

'-----


14/16

1170

substrate, breakaway corrosion was observed and coatings 60-70 ]Jm thickwere consumed within 1000 h of exposure. Incoloy 800, containingZO% Cr, also exhibited breakaway corrosion behavior in the high-pszenvironment of gas mixture B. However, with Cr contents in the range22-25%, protective oxidation behavior was sustained, even under thermalcycling conditions. Coatings on the 2 1/4Cr-1Mo steel were essentiallyfree from life-limiting defects, whereas in the carbide-rich coatings onthe carbon steel, the occurrence of severe corrosive attack was attributed to the presence of cracks. Despite the chemical composition andmorphological similarities of the carbide-rich coatings provided bytwo different suppliers, one coating promoted protective oxidationbehavior, whereas the other underwent substantial sulfidation attack.The reason for the latter form of breakdown behavior, during exposureto low-PS2 conditions in Fig. 2, is probably related to the coatingprocess parameters. The mismatch in the coefficients of thermalexpansion between Fe-Cr coatings and an Fe substrate i6 relativelysmall, but the carbides are known to be brittle and thus certain thermaltreatments May induce cracking. The very narrow zone between chromizedlayers and the carbon steel will increase the possibility of mechanicaldamage occurring. ln view of the limited coating thicknesses that wereobtained by chromizing the carbon steel, it is questionable whetherprotective oxidation behavior could be sustained for the required lifeof a heat recovery system component, even if the coatings weredefect-free.

ln spite of the high levels of Cr (20-24%) and Al (45-46%) ln thefurnace-fused coatings, the presence of defects produced very lowresistance to corrosive degradation. The particular compositions of theFeCrA! and CoCrA! alloys had a suitably narrow wetting range andadequately wetted the substrate, both necessary prerequisites for thefurnace-fusion coating process. However, as mentioned previously, bothhigh Cr and Al contents can cause brittleness. ln the case of eachalloy, costing cracks in the as-received material enhanced internaIattack of the coatings and permitted sulfide of the base metal substrateto form on the outer coating surface. Unfortunately, the base metal ofeach coating, Fe and Co, formed stable sulfides FeS and CogSS' respectively, in both gas mixtures.5.2. The Corrosion Behavior of Welds

The high degree of dilution of the weld filler metal during welding to amaterial of dissimilar composition was a major contributor to the lowcorrosion resistance of the weld metals. Chromium contents in the range6-11% in the 309 SS and 1605 weld metals.were clearly Inadequate forprotective Cr-rich surface oxide scales to be formed. The protectiveoxidation behavlor ln the 10w-PS2 gas mixture of lN 625 with 6-11% Cran be attributed to the compararively higher level of A! and theadditional presence of Nb. Given that A! and Nb also contribut~ tocorrosion resistance, a Cr equivalent (EC) has been established where

EC = % Cr + 1.1% (Ti + Nb) + 0.7% Al.

~:::::::::.i~:..I~~\r IIlr'\

;f:i


15/16

1171

6. CONCLUS I ONSThe examination of coated low-alloy and carbon steels in the as-receivedcondition and after testing in low- and high-PS2 coal gasifierenvironments has provided an understanding of a number of factorsimportant in determining the level of protection that can be achieved bysurface coatings. These can be summarized as follows:

2) Defect-free coatings no more than 10 ~m thick can provideprotection to the underlying substrate for 1000-h exposure ina high-PS envirorunent under thermal cycling conditions.Sufficient data are not available for a determination of theminimum coating thickness required to provide protection forthe typical component life of 100,000 h.ln the absence of Al, a minimum Cr concentration in the rangelS-22 wt % is required to promo te protective oxidation behaviorof pack-diffusion coated materials. The minimum Al concentration to produce the same effect lies in the range 14-20 wt %.

4) Defects such as cracks, present in aIl pack-diffusionaluminized and furnace-fused coated material, permit easyaccess of reactive gaseous species to inner regions ofcoatings, while preexisting internaI oxide particles and voidsexert a minor influence on the corrosion reaction. The cursorynondestructive evaluation effort, as applied, failed to detect

1) Pack-diffusion chromizing, aluminizing, and simultaneousaluminizing/chromizing processes can produce coatings with goodresistance to accelerated rates of degradation. Sequentiallychromizing/aluminizing by pack-diffusion and furnace-fusionprocesses produces coatlngs that are prone to rapid degradation, due largely to the presence of defects.

Therefore, in lN 625 the value of EC is approximately two percentagepoints above that of the Cr content (see Table II). ln addition, thepresence yS ~i can also be beneficial in promoting protective oxidationbehavior. ,1 An equlvalent Cr content of 20% was adequate in preventing breakaway corrosion of the lN 72 weld metal ln the low-PS2 gasmixture, but not in the high-PS2 gas mixture. While Ni was an activeparticipant in the corrosion reaction in the high-PS? environment andontributed ta the formation of sulfide sc ales on bo~h lN 625 and lN 72weld metals, evidence12 suggests that Ni is not responsible for initiating the breakaway process. Given that dilution of the weld alloyoccurs ta such a great extent during welding ta the low-alloy and carbonsteel substrates, future effort should be made to counter the dilutioneffect by modifying the welding process ta reduce heat input, by using ahigher Cr content weld alloy, or by adding an additional overlayer ontop of the original filler metal.

~d Al (45-46%) in theroduced very lowlar compositions of thetting range andrerequisites for theloned previously, bothl the case of eachen~anced internaI

se metal substrateeL~, the base metal of~S and C09SS' respec-

tal during welding to altributor to the low~ contents in the range}rly Inadequate for~ed. The protective[N 625 with 6-11% Cr~l of Al and thellso contribut~ to~n established where

'ngs 60-70 ~m thick, containing

or in the high-PS2ontents in the rangeed, even under thermalsteel were essentiallyarbide-rich coatings onive attack was attrinical composition andatings provided bytective oxidationsulfidation attack.

ior, during exposureted to the coatingents of thermalate is relativelynd thus certain thermalzone between chromizedsibllity of mechanicalthicknesses that were

~stionable whetherfor the required life:oatings were


16/16

~~

..

J1,

:i;:

.1'

~t:

The author wishes to thank A. Sather for assistance in conducting thecorrosion tests and R. W. Puccetti and T. M. Galvin for metallographicwork. This work was supported by the Surface Gasification Materia1sProgram of the U. S. Department of Energy under contraet W-31-109-Eng-38.

1172

7. ACKNOWLEDGMENTS

6) Widely dissimi1ar .chemica1 compositions of weld al10ys and 10wal10y steels resu1t in substantial dilution of elements in theweld metal deposits. Dilution of Cr by around 50% was theprincipal cause for the breakaway corrosion observed for aIlweld alloys.

5) InternaI oxidation/sulfidation is a major mode of degradationof pack-diffusion alumlnized material. Chromium reduces theseverity of internaI oxidation.

the fine cracks capable of promoting substantia1 internaIcorrosive degradation of coatings.

8. REFERENCES

1. 'A1uminizing Steel to Slow High-Su1fur Corrosion,' Chem. Week131(2), 25 (Ju1y 14, 1982).

2. M. J. Weinbaum and W. A. MeGill, in Proc. of Symp. on Corrosion inFossi1 Fuel Systems, Detroit, MI, October 1982, The ElectrochemicalSociety, Vol. P83-5, pp. 233-246 (1983).

3. R. A. Perkins, 'Su1fidation-resistant Coatings for CoalGasification Process Equipment,' U. S. DOE Report LMSC-D-879299(DE85011922), May 1985.

4. D. J. Baxter, 'The Corrosion Behavior of Coated 2 1/4Cr-1Ho andMi1d Steel Substrates in a Simulated Waste Heat Recovery SystemEnvironment of a Coa1 Gasifier,' Argonne National LaboratoryReport, ANL/FE-85-5, July 1985.

S. D. J. Baxter, in Surface Gasification Materials Program Semiannua1Pro ress Report for the Period Ending September 30, 1985, Oak RidgeNational Laboratory Report, ORNL SGMP-85/2 (December 1985).

6. G. C. Wood and F. H. Stott, in Proc. of NACE Conf. on HighTemperature Corrosion, March 1981, San Diego, CA, ed. R. A. Rapp,pp. 227-250 (1983).

7. R. Sivaknmar and E. J. Rao, Oxid. Met. 17, 391 (1982).8. J. Stringer, M. E. El-Dahshan, and I. G. Wright, Oxid. Met. 8, 361

(1974).9. H. Lewis and R. A. Smith, in Proc. of First Intl. Congress on

Meta1lie Corrosion, Butterworths, London, p. 202 (1961).la. M. G. Hobby and G. C. Wood, Oxid. Met. 1, 23 (1969).Il. W. B. A. Sharp, Corros. Sei. 10, 283 (1970).12. D. J. Baxter, 'The Corrosion Behavior of Coated 2 1/4Cr-lMo and

Carbon Steels in a Simulated High-PS2 Waste Reat Recovery SystemEnvironment of a Coa1 Gasif1er,' Argonne National LaboratoryReport, ANL/FE-86-1, 1986.

Documents

Coatings for the Protection of Low-Alloy Carbon Steels Against Corrosive Attack in Waste Heat Recovery Systems of Coal Gasifiers