`0}
`O3
`00
`Ln
`
`
`
`ArbitraryUnits
`
`
`
`BadFairGoodExcellent
`
`A2
`
`B2
`
`C2
`
`D2
`
`E2
`
`F2 G2
`
`H2
`
`12
`
`J2 K2
`
`L2
`
`Coating Variations
`
`Fig.8-13
`
`Substrate
`
`Coating & Conditions
`
`*
`"'
`*
`Co‘:7Cr12AIO.8Y
`Co25Cr6AIO.5Y
`*
`*
`"
`Ni17Cr11AfO.5Y
`*
`
`CrAl
`CrTiSi
`CoNiCrAlY
`
`C026Cr11AlY
`CoCrAlTaY
`NiCrAl
`
`NiCrAISiY
`
`"
`*
`
`NiCrA§Y + Pt(PVD-spuitered)
`NiCrA!Y + A} (PVD +pack}
`
`240
`
`
`
`COAT INGS
`
`: CHEMICAL PRGPERTIES
`
`
`
`ES):50:98co:mExo
`
`
`
`.ozoxo
`
`co_mo.:oo~05Ooow
`
`—
`
`M8ADU.
`
`-11?
`
`3U.9U59
`
`
`
`
`
`co_wo:oo5cozmmmaoa_._o_wo:oo5co:m:E_
`
`_mE_mEow_
`
`cozmufio
`
`5%+
`
`e6ueuo—1q6geM
`
`2:0:EmE:.
`
`E3
`
`
`
`
`
`mE9m>wco_mo:ooEx8_mo_9:.macs.co_wE._oo“:-m.m_u.
`
`
`
`
`
`241
`
`
`
`.—:a::. W~74
`
`.-7‘
`
`
`
`
`
`weightchange,mg/cm?
`
`Effect of cooiing rate
`
`,7
`
`1/
`
`16 .
`
`. 20
`
`Time, h
`
`_ MO
`—~"*
`
`Fig.8—15: Cyclic: Oxidation of CoCrA! & NiCrA| Alloys. Cooiing Rate Effect.
`(Nicoli 1934)
`
`8O
`D?'
`T‘
`
`O(
`
`0-8
`
`. Co15Cr6Al
`. Co15Cr8AI
`. Co1OCr4AI
`
`. Co1DCr5AI
`. Co10Cr1‘éAI
`. Co25Cr6A|
`. Co25Cr6AI1Y
`
`EU
`
`}>(
`
`U
`"U
`C)
`O1“
`L-
`
`E£
`
`1
`
`01
`E(J"\
`U)
`E
`
`.9.“CU
`D’:
`
`wt.% Hf or Y
`
`Fig.8-‘£6: Co-10Cr—11Alat1000°C for 100 h in Air. Effects of Hf & Y.
`
`242
`
`
`
`
`
`o-m-_oE%m.9:5.Emammémocmwzxw.PC.-m.mE
`>9.82-ooow
`
`
`
`....o-m-_zE$m>w9:._8Efimmfimocmfiiw“mm_.-w.m_n_
`
`243
`
`
`
`COATINGS: CHEMICAL PROPERTIES
`
`8.5. OXIDATION & EIGB TEMPERATURE CORROSION
`
`in
`Oxidation is often used to refer to corrosion in general,
`addition to expressing degradation as a result of
`reaction with
`oxygen. Many of the early concepts, em;
`the classic Wagner model
`for diffusion and the Pilling and Bedworth ratio of scale to
`substrate molecular/atom volume ratio, evolved using oxides as
`examples. However, high temperature reactions have become quite
`complex and the simple framework indicated in the previous sec—
`tion has to be considered with respect to many systems undergoing
`corrosion simultaneously; The outcome of a reappraisal may indi-
`cate the several features which need to be addressed for scale
`growth, composition and stability (Rahmel et al l985h
`
`1. Adsorption, nucleation and initial stages of oxidation (Grabke
`1985).
`2. Complex defects in metal oxides and sulphides; their nature
`and transport mechanism (Hobbs l985L
`3. Lattice, line defect and grain boundary transport mechanism
`(Atkinson l985L
`4. Transport of gaseous species in growing oxide scales (Mrowec
`l98S).
`5. Atomistics of scale growth at the scale/gas interface (Rapp
`l985).
`l987L
`6. Lateral growth in oxide scales (Smeltzer 1985;
`‘L Adhesion mechanisms ~ the reactive element effect.(Stringer
`1985; Moon & Bennett
`l987L
`8. Key mechanical properties of oxidising components in maintai—
`ning scale integrity (Manning l985L
`9. interaction between oxidation, sulphidation, carburization
`and creep (Ilschner & Scutze l985L
`Internal and intergranular oxidation of alloys (Yurek l985L
`10.
`ll.Influence of impurities
`such as sulphides on the transport
`properties and nature of oxide scales (Wagner
`l985L
`l2. Faults,fissures and voids in scales (Graham l985L
`l3.Role of chlorides in high temperature corrosion (Hancock
`1985).
`14. Hot corrosion (Pettit l985L
`l5. Role of oxides in abrasion, erosion and wear
`l6. Oxidation of compounds
`(Schmalzreid l985L
`l7. New methods for studying high temperature corrosion, particu~
`larly in situ methods (Rahmel 1985M
`
`(Stott l985L
`
`It is beyond the scope of this review even to summarise the
`outcome of the above appraisal; coatings are an integral part of
`a high temperature system as more and more high temperature
`structural components are coated, Many of
`the available studies
`are conducted on individual candidate materials with a view to
`their utility as coatings, and later followed up by both corro—
`sion and mechanical property studies in a substrate/coating mode.
`Information condensed herein must be viewed as such.
`
`244
`
`
`
`COATINGS : CHEMICAL PROPERTI ES
`
`Oxide layers are not always protective although they predominate
`scale morphology in most respects. They are liable to damage,
`non-stoichiometry, mixed phases and stresses and present a com~
`plex diffusion pattern depending on scale morphology. Thus oxida-
`tion has to be viewed in many aspects w.rA; scale growth, where
`plasticity is affected (Douglass 1969), where stresses generated
`in columnar scales lead to pit formation (Louat
`& Sadananda
`1987), or lamellar stresses which occur
`in plasma spray coatings
`affect its adhesion and transport properties (Knotek & Elsing
`1987)
`. Defects in oxide scales influence transport properties,
`and at intermediate temperatures short-circuit diffusion paths
`can take over and become the predominant mode (Gesmundo 1987;
`Moon & Bennett 1987; Moon l987; Atkinson 1985; Gesmundo et al
`l985; Chadwick & Taylor 1984, l982).The marked influence which
`elements like Y and Ce have has been argued on the basis of
`pegging (Stringer 1987; Pendse & Stringer 1985). But increasing
`evidence appears to come from effects of grain boundary diffusion
`and nucleation (Moon & Bennett 1987; Smeggil 1987). Oxides of Ce,
`La and Y exert a strong influence on oxide composition, adhesion
`and the scale growth direction of Cr2O3—formation;
`the effect is
`more pronounced on Ni—25Cr
`than the Co-25Cr
`(wt.%) alloys at
`l00O° and ll00OC, and in this case the ‘peg’
`theory does not
`apply (Eou & Stringer 1987). A similar effect was observed on
`stainless steels doped with Y, Ce and La enhancing oxidation
`resistance while Hf and Zr offered no benefit
`(Landkof et al'
`1984).
`
`Dispersed oxides of Mg, La and Y and element Y in plasma.sprayed
`NiCrAl coatings at ll50OC and 122S°C showed that yttrium oxide
`gave the best oxidation resistance which was influenced by the
`size and distribution of
`the oxide particles (Luthra & Hall
`l9§§J. La was ascribed to provide a diffusion barrier layer as
`Cr
`diffused 40 times slower in LaCrO3 than in Cr2O3 in Cr-La
`alloys (Tavadze et al
`l986). YZO3 dispersoids were found to
`influence the microstructure rather thannprovidinfi oxide nuclea-
`tion sites for Ni-20Cr
`in low p02 of 10
`to 10
`at 1000°c. It
`reduced Cr—evap0ration,
`the oxide grain size and porosity {Braski
`et al 1986). A small amount of
`impurities enriched at grain
`boundaries may greatly affect the deformation characteristics and
`influence the mechanical and transport properties of the growing
`scales (Kofstad 1985).
`Implanted Y in chromia-forming Co—45 wt.%
`Cr alloys exhibited a x100 reduction in oxidation rate at 1000°C
`in pure 02, with the mechanisnxof chromia growth changing from
`cation to anion diffusion, and Y-segregation at Cr2O3 fine—grain
`boundaries. A solute-drag effect is proposed as the mechanism
`(Przyhylski et al l988a,b; Przybylski & Mrowec 1984L
`
`internal oxidation are a few
`Embrittlement, break-away oxidation,
`other detrimental effects. Hydrogen embrittlement is a well-
`recognised form of corrosion in aqueous systems» Water vapour can
`play a large part in high temperature embrittlement as also H2
`and 02. Thus pre-oxidation, envisaged as a means of prolonging
`the initiation stage of hot corrosion may not always be benefi~
`cial. Ni undergoes oxygen-embrittlement at high temperature;
`the
`
`416
`
`245
`
`
`
`COATINGS 2 CHEMICAL PROPERTIES
`
`onset of embrittlement is dependent on the p02. Results on Inco~
`nel 718 seemed to indicate that the first stage oxidation process
`occurs at the grain boundaries before chromia formation influen~
`ced embrittlement
`(Andrieu & Henon 1987). An inverse phenomenon
`can occur in alloys prone to break—away oxidation. In 9Cr-lMo
`steels, general oxidation occurs at 500-600°C and above 7000C
`Cr2O3 is the dominant oxide (Khanna et al 1986). Such preferen-
`tial temperaturendependent scale development
`is not uncommon in
`Ni— and Co-base superalloys.
`
`in
`A side—step development of a 20Cr/25Ni/Nb steel precluding Al
`the Fe—base alloy is reported. The wt.% steel composition was
`l9.9Cr , 24.6Ni, O.7Nb, 0.6Mn, 0.5681, 0.04C, balance Fe, which
`was first selectively oxidised in a 20% cold worked condition in
`a 50:1 H2:H O for 2 hours at 800°C or lhour at 93OOC.The scales
`which deve oped were Fe- and Ni“ free, with an average of 0.4 -
`0.8 microns of Cr2O3. The alloy showed improved resistance to
`oxidation and carburization. Its sulphidation resistance is not
`reported (see Fig.8—19)
`(Bennett et al 1984).
`Implantation of Ce
`and Y reduced the oxidation rate consistently by more than 50% at
`750°—950°C (Bennett & Tuson 1988/89). Alloying Ce
`(0.00l—l.00
`wt.%) and CeO had a similar beneficial effect on Fe~(lO~20)Cr
`alloys at 100 DC, p02 0.13 bar (Rhys—Jones l987).Ceria disper-
`sion added by jet injection to carbon steel was beneficial to
`oxidation of carbon steel
`(Tiefan et al l984L
`
`Ni- and Co—base superalloys generally are formulated to be prefe-
`rential chromia— or alumina~forming variety. The degradation
`modes of MCrA1~coatings will be summarized later in this chapter.
`There are many studies on their oxidization behaviour
`(Wood &
`Whittle 1967; Wood & Bobby 1969; Wood et al 1970, 1971; Benard
`1964; Kubaschewski & Hopkins 1961). C0 effect on the oxidation of
`Ni-base alloys was found to lower the cyclic oxidation resistance
`on high—Cr alloys and was to be an optimum 5% to be effective in
`Al—containing alloys when tested in static air at 10000, 11000
`and ll50OC. The ratio Cr/Al decides the Cr2O3/chromite spinel
`(Cr/A1>3.5) or
`the A1203/aluminate spinel
`(Cr/Al <3Jn, with the
`latter having a better resistance to cyclic oxidation. Refractory
`metal additions(Ta, Nb, W and M0) were beneficial, with Ta the
`most effective.
`In all cases, any factor which promoted NiO
`formation resulted in scale breakdown (Barrett 1986)
`
`Ti, as a light , high temperature metal is a prime candidate for
`aerospace vehicle systems. The metal oxidizes readily and also is
`prone to embrittlement
`(Shenoy et al 1986; Strafford et al l983;
`Datta et al 1983; Strafford 1983; Datta et al 1984). It is found
`to manifest a moving boundary parallel to the interface as the
`two oxides change in proportion (Unnam et al 1986). Coatings of
`Al by EB, sputtered SiO2, and CVD silicate via silane and borane,
`and mixed coatings of A1 with the latter two were tested on Ti-
`6Al—2Sn-4Zr—2Mo foils and oxidi2ed.Al+SiO2 presented the best
`barrier layer performance with no effect on its mechanical prop~
`erties (Clark et al l988)..Amorphous metals are not widely stu~
`died for their high temperature characteristics. Ti, Zr and HE
`
`417
`
`246
`
`
`
`COATINGS: CHEMICAL PROPERTIES
`
`affected void formation at the interface ine1Ni3Al~O.lB alloy.
`Hf addition was the most effective in promoting a protective
`oxide layer
`(Taniguchi & Shibata l986). Studies on a rapidly
`solidified Ta-Ir alloy at 500 and 700°C are also reported. Ta was
`selectively oxidized, while the Ir coalesced into platelets of
`Irerich crystalline alloy oriented roughly parallel to the oxide-
`alloy interface. Although the unoxidized core remained in the
`glassy state, dissolved oxygen and the oxidation process had
`embrittled it {Cotell
`& Yurek 1986). S ectacular stratified oxi-
`dation layers develop on Ti
`(7600, 960
`, Ta (9000) and Nb
`(4500,
`600°C),
`in pure oxygen as demonstrated by Rousselet et al
`(l987L
`
`8.6. HOT CORRDSIO
`
`Hot corrosion has been variously defined but always involves the
`presence of sulphur species in the environment. Hot corrosion
`attack is recognized by the fact that a protective layer loses
`its resistance characteristics to shield the system it covers and
`this may lead to catastrophic failure. In general,
`the useful
`life of a coating in a corrosive atmosphere is very much a factor
`of time. There is a period over which it develops a protective
`scale; once the protective coating forms, its endurance to stress
`and temperature changes decides the subsequent breakdown of the
`scale. Coating degradation can be rapid or restrained, depending
`on the coherence, plasticity and shock proof qualities of the
`outer scale,
`the stability and compactness, free of voids of the
`intermediate scale, and,
`the stress and embrittlement accommoda-
`tion and adherence of the scale/metal interface. Hot corrosion
`aggravates the scale breakdown mode and occurs in two recogni-
`zable stages, and it is particularly severe where a liquid phase
`is involved, irrespective of whether it is a fluid product or a
`fluid reactant.
`
`Monitoring coating life thus can be considered in two stages:
`
`Stage 1 — Scale initiation, retention and repair;
`Stage 2 ~ Scale damage, weak repair, and rapid deterioration.
`
`Stage 1 is termed the initiation stage and stage 2 is the propa-
`gation stage. Fig.8—l4 (iii)
`(p.412) shows the two stages in
`environments ranging from mild to severe.
`
`8.6.1. FACTORS AT THE INITIATION STAGE OF HOT CORROSION:
`
`The overall factors which govern the onset of hot corrosion are
`(Giggins & Pettit 1979; Condé et al 1982M
`
`l. Coating condition and its composition;
`2. Gas composition and velocity;
`3. Reactive particle/salt
`composition, its deposition rate and
`
`418
`
`247
`
`
`
`COATINGS: CHEMICAL PROPERTIES
`
`existence state;
`4. Temperature ~
`isothermal/cyclic;
`5. Non~reactive particle inclusion, deposition and impact;
`6. Component geometry.
`
`Hot corrosion is always a secondary reaction in the degradation
`process. It originates at the initiation stage,
`reduces the scale
`coherence and thus its protective life time there itself and
`follows it up with a catastrophic attack at
`the propagation
`stage. There have been several discussions on it which are avai-
`lable in the various references cited. Several contributory fac~
`tors appear at the initiation stage; opinions differ about the
`definition,
`the factors of
`importance and the mechanism, One of
`the very common featurs is that a liquid phase is involved.
`Although in carburization this does not occur a liquid phase
`involvement
`is common with reactants where sulphur— or chloride
`media are present and also certain oxides, erg. oxides of V.
`Gaseous oxide formation such as CrO3, SiO, oxides of W , Mo etc“
`are also to be considered as promoters of catastrophic corrosion.
`The role of a coating then is to prolong the initiation stage,
`under
`isothermal and cyclic hot corrosion conditions by develop—
`ing stable scales with adequate creep and rupture strength.
`
`8.7.. CARBDRIZATION/OXIDATION DEX’.-1RADATI%l
`
`A typical situation of carburization occurs in coal gasification
`and fluidized—bed combustion processes. Fe—base alloys are the
`most widely used in this field and hence documented (Hsu 1987;
`Tachikart et al l987; Debruyn et al 1987; Ramanarayanan 1987;
`John 1986; Kofstad 1984; Terry et al 1987).
`'
`
`The 4—step kinetics involved in carburization may be seen as
`follows:
`
`1. Transport in the gaseous environment by flow or diffusionr
`Uncombusted hydrocarbons, C0, C02 or CH4 can induce carbide
`formation in the matrix;
`
`2. Transfer of carbon to the metal matrix by phase boundary
`and/or reduction reactions which result in carbon atoms;
`
`3.
`
`The dissolved carbon diffuses inwards;
`
`4. Reaction of carbon with any or all of the alloy constituents
`which have the free energy for carbide-forming at the available
`carbon activity,
`is accompanied by diffusion of
`these constitu~
`ents to the precipitate.
`
`Oxide layers can be destroyed if graphite (or coke) deposits or
`gets trapped in the growing scale, or if reducing conditions
`prevail. High carbon activities are possible as the pCO/pCO2
`ratio links to the metal/oxide equilibrium. The metal matrix
`
`419
`
`248
`
`
`
`COATINGS: CHEMICAL PROPERTIES
`
`itself and the oxide that grows preferentially on it determine
`the growth of the graphite reductant. For
`instance, graphite
`grows much faster on Fe than on Ni, but in the presence of H28,
`the growth is accelerated on Ni, while on Fe it is retarded.
`Internal carbide formation in alloys such as Fe-Ni-Cr
`is called
`‘metal dustingfi Often,
`the formation of the same oxide can allow
`or arrest carburization. Naturally formed oxide on a Fe—l2Ni—2OCr
`alloy caused local carburization of the alloy by impurities
`present
`in a N2-H2 atmosphere, while the material was fully
`resistant under the same conditions once the oxide was sandblast—
`ed. At higher than 1l000C,
`the protective Cr2O3 can undergo
`reduction by CO to form Cr3C2 and/or Cr7C , especially if coke
`deposition is in significant amounts, an
`the oxide is buried
`under it. Internal carbide precipitation can also occur in CO—H2~
`H20 atmosphere.
`
`In the absence of a protective scale, carbon ingress into a Fe-
`Ni-Cr alloy is by phase boundary reaction and diffusion control-
`led. The presence of sulphur retards the transfer of carbon, but
`to curtail the solubility and diffusion, a high Ni/Fe ratio is
`needed and additives like Si. If the oxide is coherent, dense and
`adherent
`carbon cannot penetrate since it has no solid solubili-
`ty in oxides, but if there are fissures and cracks or pores,
`then
`carburization is possible. The integrity of the scale could be
`hampered by creep fatigue or thermal cycling.Any factor which
`induces stress is thus conducive to carburization. Additions of
`Nb, Ce, Si etc., would control in this case. Formation.of higher
`oxides, spinels or mixed oxides lower
`the resistance to carburi—
`zation. Fig.8—l8 shows the oxide failure modes by carburization
`(Grabke & Wolf 1987; Ramanarayanan 1987; Hsu l987L
`
`The following carbon pick-up was recorded for various Fe-Base
`alloys in Argon-5%H —5%CO—5%CH4 environment
`(Rothman et al 1984):
`(ranking in order of
`increase of carbon pick-up)
`
`925%, 215 h:
`Cabot 214 < Cabot SOOH < Multimet < Cabot 600 < Hastelloy S <
`Hastelloy X < Inconel 610 < Haynes 230 < Inconel 617 < 310
`stainless steel;
`
`980°C, 55 h:
`Cabot 214 < Cabot BOOH < Multimet < Hastelloy S < Haynes 230 <
`Bastelloy X < Cabot 600 < Inconel 601 < Inconel 617
`
`The influence of selective oxidation on 20/25/Nb stainless steel
`at 650° and 700°C is given in Fig.8—l9a,b. On electropolished
`surfaces of chromia-forming Fe—Cr-Ni and Cr-Ni alloys,
`the oxida-
`tion layer appears to form non—unif0rmly, while cold worked
`Samp%es undergo uniform oxidation in H2—CH4 with very low p02 of
`10
`at 825°C with carbon activity at 0.8. Removal of alpha-
`chromia during nucleation of
`the carbide M7C3 occurred followed
`by internal carbide precipitation.
`(Smith et a1 l985a,b). At
`low
`p02 heat resistant steels are completely under
`the influence of
`carbides wth M7C3 developing beneath Cr3C2 and are not affected
`
`420
`
`249
`
`
`
`COATINGS: CHEMICAL PROPERTIES
`
`b. carburizaiion via
`
`cracked oxide layer
`
`no protective scale
`
`
`
`Va»
`
`4
`
`d: carburization with
`graphite deposition
`
`e:V_surface—com/erted
`oxide-carbide
`
`/
`
`//1
`
`f. siiica barrier layer
`
`Schematic diagram of carburization with oxidation &
`
`barrier
`
`layers
`
`FiG.8—‘18
`
`by brief periods of oxidation (Kinniard et al 1986). Breakaway
`oxidation and laminated structure morphology were observed at
`600°C on Fe-9Cr—lMo steel in a high pressure CO2/CO atmosphere
`(Newcombe & Stobbs l986L
`
`S102 coatings by PAVD on IN BOOH provided excellent resistance to
`Carburization in H2~CE4 mixtures at 825°C, but were totally in—
`effective at l000°C where the Sioz was converted to Sic by a gas
`phase reduction. Partial degradation was found to occur even at
`8250 as Ti, Mn and A1 additives in the alloy reduced the silica
`(Lang et al 1987). Silicide coatings are not protective on steels
`and if they are produced via a methyl-silane, carburization was
`found to set in at the coating stage itself (Southwell et a1
`1987). Ferritic steels with 6 wt.% Al showed good resistance to
`cyclic oxidation~carburization in H2—H2O—CO—CO2 atmospheres with
`carbon activity 0.2, and Ti and Zr had very favourable effects in
`catalyzing the alumina phase transformation from the early theta-
`to the a1pha—phase. They also increased the sintering rate and
`fracture toughness, while suppressing grain growth. None of these
`benefits were realized in the austenitic series tested with 4%Al,
`and Y also did not improve oxide adherence. Instead it caused
`grain boundary embrittlement. In creep tests the oxide layer
`cracked with subsequent intergranular oxidation and carburization
`(Wambach et al l987L
`
`421
`
`250
`
`
`
`:3...22m9,N
`
`
`
`>EEmmmmcmE_omamE:oEm.oa5mo?.m,E:.
`
`1-‘-'
`
`V-‘
`
`T
`
`2
`
`ZLUC3/5LlJ QUORBJIUQQUOO UOQJEQ
`
`awe/Sm ‘ugefilm
`
`fimwfixo3m>:om_mw.ofimmfixo>_m>:om_mm.2coo?Hm69wgzamamBmfixo>_m>:o9mm.3.
`
`
`
`
`
`Emczmmbo:.m”_m2mmmo_c_£mnz\mm\om.m“gm¢m.m_u_oooonE6%gz\mm\om.m.8Eoi
`
`
`
`
`
`
`
`
`
`ooommE5%nzmmamumafixo>_m>_5m_mm.n
`
`
`
`
`ooommE_m.9mQzamsm.0
`
`
`
`
`
`Empm.m.,.a_c_smgz\m.o.5m.:o:wU_xOm>_59ow.5seamHmTmdi
`
`
`
`
`
`fo\oN:\N:\Noo\ooc_
`
`251
`
`
`
`COATINGS: CHEMICAL PROPERTIES
`
`In the above case the alloys were pre—oXidized prior to exposure
`to reactive atmosphere. Pre—oxidation has been found to be a
`deterrent, but not with sustained efectiveness. A coating of
`63%Al~33%Cr—4%Hf on Incoloy 800 performed well at 980°C in coal
`char environment after pre-oxidation (Douglass & Bhide l981L
`MA956,
`the mechnically alloyed product, yielded good mechanical
`stability of the alumina scale (Sheybany & Douglass l988L
`
`8.7.1. CARBURIZATION IN THE PRESENCE OF SULPHIDATION:
`
`Carburization conditions are not encountered on their own in
`coal~derived atmospheres, but
`along with gases such as H2—H2S,
`and H2-H20. The oxide layer which is perfectly resistant other"
`wise,
`is then exposed to conditions which render it unstable. In
`the previous section the ‘inhibiting’ effect of sulphur to Carbu-
`rization was briefly mentioned. A number of environmental varia-
`tions have been studied: CO2, CO, H2 (introduced as forming gas),
`H28, with and without CH4, chloride etc., with p02 variation
`brought in from the equilibrium CO/CO and/or H /H20. Temperature
`appears to exercise a critical contro
`on the
`egree of carburi—
`zation. At ffrbon acti%:ty 0.8 in H2—CH4, H28 at 100 ppm at
`ps2=2.2x1o"
`to 5.5x1o‘
`, at l0OO°C, carbide M7c3 on Fe-Ni-Cr
`alloys exhibited a preferred growth orientation in the [001]
`direction, and in commercial alloys surface carbides of M7C3 and
`M23C6 nucleated with Mns buried underneath,
`in contrast to sub-
`scale M2 C6 and surface M7C3 embedded in surface alpha*Cr2O3 in
`sulphur- ree atmospheres. An apparent reductfiop of 75% in weight
`gain occurs when p82 is introduced at l.4xlO
`, with significant
`decrease of
`internal carburization. However, sulphide scales are
`formed and the overall corrosion increases with corrosion at 950°
`greater than at 10000 and l0S0°C (Barnes et al
`l985,1986L
`
`The inference from the above would be that although ‘metal dus~
`ting‘ by internal carbide precipitation and external carbide
`formation can be arrested in the presence of sulphur, no ultimate
`benefit is obtained as one form of corrosion is exchanged for
`another.
`Internal precipitation reflects in material degradation
`by physical and mechanical factors such as creep and stress, and
`a marked loss in constituents will also result in chemical degra-
`dation. Surface scales formed by chemical reaction, on the other
`hand need to be plastic, stable and coherent; adverse factors
`will destroy thenn Attack by sulphur manifests in catastrophic
`corrosion in the majority of cases, and such corrosion is often
`associated with a liquid phase. The effect of H2-H28 has been
`extensively studied and reported in literature. A few papers are
`listed here:— Weber & Hocking 1985; Majid & Lambertin 1985;
`Strafford et al 1983,1985; Floreen et al 1981; Gibb l983; Grabke
`1984; Norton 1984; Grabke et al 1980; Hemmings & Perkins 1977;
`Mrowec 1976; Mrowec & Werber 1975. The principal aspect of attack
`by sulphur—media is that one or more liquid phases are often
`involved.
`
`252
`
`
`
`COATINGS: CHEMICAL PROPERTIES
`
`8.8.
`
`SE LIQUID PHASE EFFECT
`
`8.8.13 LIQUID PHASE FROM THE ENVIRONMENT:
`
`In the marine gas turbine, sodium is picked up as NaCl, which
`reacts with sulphur and oxygen irom the fuel + air mixture to
`form Na2SO4:
`
`4l\laCl + S2 + 402 = 2Na2sO4 + 2Cl2.
`
`The melting points of Nacl and Na2SO4 are 8000 and 884°C respec~
`tivelyr The two salts however,
`form low melting eutectics and can
`exist in a solid + liquid state over a range of composition and
`temperature (Fig.8—20a). The deposits which arrive at the hot
`turbine blade are composed of Nacl, Na2SO4, carbon (from the
`fuel) and perhaps dust/sand .. They do not cover the entire blade
`but are deposited at discrete areas in the leading edge or at a
`middle concave (suction) region. The brunt of the deposition is
`taken by the first stage vanes and blades operating at 900 -
`ll50°C., In coal gasifiers CaSO4+CaO provides a deposit and sul-
`phidizing conditions and vanadium participates in the formation
`of gxolten oxides (Fig.8-20b). The eutectic 'NaSO4+Na\i'O3 forms at
`610 C.
`
`I N i. Llq.+ea it
`
`Temperaiure,°C
`
`N39‘ + beta 3UiPh3i9
`
`Kfacl lhaupte ‘
`
`I‘
`
`10 O O
`
`CO C) O
`
`0::
`
`‘*4 C) 0
`
`Temperature, 03 O O
`
`eutectic (14 mole%
`sulphate meits at 610 °C)
`
`‘V
`
`‘ H H ‘
`
`H 0
`
`100
`
`i 20
`
`40
`
`50 80
`
`NECI
`
`MOI."/o
`
`N62504:,
`
`Na\/O3
`
`.
`
`MOI.”/o
`
`Na2SO4,
`
`The Sodium Chloride — sulphate
`system
`
`The Sodium Vanaclate — Chloride
`system
`
`Fig.8-20a
`
`Fig.8~20b
`
`8.8.,2.. LIQUID PHASE FROM ALLOY + GAS REACTIONS:
`
`The gas enviroment in a turbine is 02 predominant with a trace of
`sulphur which arrives as 802. The presence of metals can catalyse
`the reaction to give a mixture of 02, S02 and S03. The metal /
`S02, S03, 02 reaction mechanisms have been discussed elsewhere
`(Alcock et al 1969; Birks 1975; Kofstad & Akesson l9'79; Seybolt &
`Beltran 1967). it suffices here to say that oxides, sulphides and
`
`42%
`
`253
`
`
`
`COATINGS: CHEMICAL PROPEERTI ES
`
`sulphates form as reaction products under specific conditions.
`A few examples of the melting points of metal + metal sulphides
`of turbine and coal gasifier alloys are given below:~
`
`Eutectic Ni + Ni3S2, nnp. = 6450C; Ni3S2, nnp. = 8100C;
`
`Eutectic Co + Cos, nnp. = 879°C; COS, m4; = 1O70OC;
`
`Eutectic Fe + Fes, nnp. = 965°C
`
`in coal gasifier environments pro-
`The more reducing environment
`motes Fe sulphide eutectics and in Ni—additive steels Ni—sulphi~
`des can appear as a liquid phase.
`
`Liquid Phases from Alloy & Environment Reaction Product Inter-
`action; a few examples :-
`
`coo or C0304 + s03 = coso4; coso4 + Na2SO4 ; mg» 2 565°C
`
`Nio + so3 = NiSO4; NiSO4 + Na2SO4 ; Hnp. = 670°C
`
`2NaVO3 + 2NiO = Ni2V2O7 + Na2O ; m.p. = 850°C,
`
`Ni2V2O7 + NiO = NiV2O8 ; m.p. >10oo°c,
`
`2NaVO3 + N10 = Ni(VO3)2 + Na2O ; m.p. 750°C,
`
`2NaVO3 + 3NiO = Ni3(VO4)2 + Na2O ; m.p. = 1210°c,
`
`A1203 + 2NaVO3 = 2AlVO4 + Na2O; AlVO4 decomposes at 6250Ce
`
`Cr 0
`2 3
`
`+ 2NaVO = 2CrVO + Na 0; CrVO melts between 8lO°
`3
`4
`2
`an
`900°C
`
`8.8.3. THE EFFECT or THE LIQUID PHASE;
`
`The emergence of a liquid phase during a corrosion reaction will
`result in:
`'
`
`(i) drastic effects on diffusion control parameters, transport
`and mobility,
`(ii)create electrochemical conditions — bimetallic cell situa-
`tion between different metals and different phases,
`(iii)dissolve protective reaction product oxides via acidic
`and/or basic fluxing,
`(iv)facilitate fast transport of reactant species to the alloy
`/scale interface hitherto barred by coherent scales,
`(V) physically undermine scale coherence by mass flow.
`
`Products which vaporize, or react to form a vapour product and
`deposit in a more stable form on the cooler parts are well known.
`For
`instance in a chloride—containing environment reactive spe-
`cies can form chlorides which later oxidize. Mass spectrometric
`
`425
`
`254
`
`
`
`lug;
`
`«$5
`
`
`
` aw\mfitsdM.WSW:32¢
`
`NokomEE:ShowEEzm
`
`ooommHm.,,_ms_:88m9889m.MW?2.8m_n_xo-®i.8mam
`
`_.<.2
`
`1....
`
`25cm
`
`255
`
`
`
`COATINGS: CHEMICAL PROPERTIES
`
`studies by the authors (Hocking & Vasantasree 1976) showed forma-
`tion of volatile T132; needles of Tio growing out of the Cr2O3
`barrier layer may be expected to have
`ormed via a vapour deposi-
`tion mode
`(see Fig.8—2l).
`Fig.8—22a,b show another needle—form—
`ing system where liquid sulphide eutectic pipes up a column of
`Ni—Ni3S2 to be rapidly covered by NiO (Hocking & Vasantasree
`l976,l982L
`
`The catastrophic effect of Na2SO4 would have been confined to
`temperatures just around 900°C except for the chloride effect
`first recognised in UK laboratories in the sixties. Much of
`turbine hot corrosion was shown to fall into two categories w the
`low temperature degradation in the region of 620—750°C and the
`high temperature corrosion over 850 — 950°C.
`sulphides and sul-
`phates largely responsible for hot corrosion of Ni — and Co—base
`alloys are unstable or do not
`form beyond these temperature
`regions. Aggravated corrosion occurs well above 900°C if a float-
`ing potential of S02 is allowed to prevail
`in the reactant
`(Vasantasree & Hocking l976L
`
`Nacl itself is completly converted to Na2SO4 within 3 minutes
`(Conde et al 1977), but its continuous arrival and its stability
`in solution with Na2S04 at lower temperatures well below its
`melting point mostly in a solid + liquid state and sometimes
`liquid state is an important factor in low temperature hot corro-
`sion. Co—base alloys are vulnerable to Na2SQ4 itself as the C00 -
`COSO4 - Na2SO4 reaction and solution progresses.Volatile chlo-
`rides are often intermediate products in hot corrosion, which
`enhance the corrosion rate to result in more stable products and
`also to form more stablegreao antsr Ch oridation of Fe, Niland
`Fe—§i alloys
`in pCl2 10
`, 10
`and 10 with p02 range 10'
`to
`10" 3 indicates that the aggressive effect of the chloride meddum
`on Fe is greater between 800 and 100000, except for pCl2 10- . A
`limiting content of 50%Ni
`in the Fe—Ni system provided excellen
`chloridation resistance;
`above l00O°C Ni was
`inert
`in pC12 10"
`(Strafford et al 1987b
`
`Vanadic melts form the molten salt media encountered in low grade
`fuel combustion and residual fuel oil ash hot corrosion (Johnson
`& Littler 1963). The deposit—forming reaction may be of the type
`
`wNa + X02 + yv + zS + aCl ~~~~~-> deposit
`
`(Halstead 1973)
`
`in which Na, 02, V; S and C1 are in the vapour phase, and can be
`involved in several reaction systems.
`
`3.9. FLUXIE§G MECHANISMS IN HOT CQRROSION
`
`Thermochemical diagrams give a direction on corrosion in melts
`(Pourbaix 1987) and the eletrochemistry of molten salt corrosion
`is reviewed by several workers: Rapp (1987); Rahmel
`(1987);
`Pourbaix (1987); Hocking & Sequeira (l982L
`
`427
`
`256
`
`
`
`
`
`.95EEN.%$zHm938955m.mOmc_2802M.new.2.UooonHag7%
`
`
`>o__<
`
`Loo\i.o-MZNmm._2“mg.E£520new.m_mfimmz222.092Eammom5&5»98m
`
`257
`
`
`
`COATINGS: CHEMICAL PROPERTIES
`
`869.1. THE Na2SO4 MODEL:
`
`The oxy—anionic Na2SO4 melt has been a model on which most of the
`acidic and basic—fluxing concepts have been developed. The con-
`cept can be extended to all molten salt metal product reactions
`as long as ion—exchange reactions can be applied validly to the
`reaction system. Thus sulphates, carbonates, nitrates and vana-
`dates can be included in oxy—anionic reactions at high tempera-
`ture (Johnson & Laitinen l963; Cutler l97l).The primary factor
`in viewing molten salts with respect to hot corrosion is that of
`the availability of the melt and not its mass. An attack will be
`self—sustaining as long as the melt can participate in the excha—
`nge and remain as the intermediate means by which alloy component
`elements will eventually react to solid corrosion products.In
`the turbine operating conditions melt can fornxas and when the
`component particles are deposited. Products can remain in solu~
`tion with the melt, form a eutectic, precipitate out or form a
`solid complex with the total mass of available melt.
`
`The following reactions will clarify the various points noted
`above:
`
`Na2SO4 “““““““““"‘> N320 4“ S03
`
`so:
`
`Acidic
`
`0: +
`
`5303
`
`Basic Acidic
`
`l/252 + (3/2)O2
`
`8.9.2. ACIDIC FLUXING:
`
`Hot corrosion reactions occur where 80: participates with the 82,
`S02 and SO
`species from the gas or dissociated melt
`in convert-
`ing the al oy to corrosion products either by chemical thermo-
`dynamic reaction or electrochemically transported as an ion for a
`subsequent reaction with the gaseous mediunu Thus for an alloy
`AB,
`
`(i) A (alloy) + s03 + (1/2)o2 = A
`
`+ s04:
`
`++
`
`For a continuous solution of ASO4 in Na2SO4, S03 and 02 must be
`available, e.g. COSO4 + Na2SO4,
`
`(ii) A2+ + s04‘ + (1/2)o2 = A0 + so3
`alloy melt
`solid
`
`(iii) A0 can remain in solution with Na2SO4 melt if there is a
`negative solubility gradient
`[note this cannot happen in a small
`mass or thin layer of melt].
`
`429
`
`258
`
`
`
`COATINGS: CHEMICAL PROPERTIES
`
`(iv) 13 (alloy) + sof + (3/2)o2 = B04: + 503
`Of
`
`A(alloy)B{alloy) + 202 = A
`
`2+
`
`+ B04: (solution in melt)
`
`A2+ +so4:=Ao-+303
`
`The melt remains as a via media; very small amounts of Na2SO4 can
`permit a substantial alloy~to—alloy metal oxides conversion.
`
`In practical systems virtually all alloys are susceptible to
`acidic fluxing depending on the level of p803 or the amount of
`V705 formed or deposited. The condition is prevalent when chlo-
`ride is present and induces alloy depletion in gas turbine envi-
`ronment and in carburizing conditions when oxygen starvation
`occurs. Alloys containing Mo, W or V are very Vulnerable because
`they can be auto-generative to acidic oxides, e.g. B-1800,
`IN
`100, Mar M