corrosion of intermetallics - tsapps at nist

Corrosion of IntermetallicsP.K. Datta, H.L. Du, and J.S. Burnell-Gray, Advanced Materials Research Institute, Northumbria University, United KingdomR.E. Ricker, National Institute of Standards and Technology

IN MANY APPLICATIONS of structuralmaterials—aerospace, automobiles, power gen-eration—increasing demands are being made formaterials with temperature capabilities greaterthan those of superalloys. Intermetallics withhigher melting points can replace superalloyswith inadequate melting points (Ref 1–3).Intermetallics, characterized by strong, pre-dominantly metallic bonding between unlikeatoms, are situated between superalloys andceramics. From bonding comes crystal structure,ordering, high strength at low and hightemperature, and low ductility. Low fracturestrain and poor fracture toughness (KIc) of theintermetallics stem from their complex crystalstructures, large Burgers vectors (a crystal vectorthat defines the amount and direction of atomicdisplacement associated with dislocation motionin a crystal), high lattice stresses, inadequate slipsystems, inability to promote cross slip, and lackof grain-boundary cohesion. The influence ofsuch factors on the corrosion behavior is notinsignificant. Stress generation during the scalegrowth, scale spallation during thermal cycling,stress corrosion and corrosion fatigue, andfinally, cationic and anionic transports influen-cing the corrosion kinetics are all likely to beaffected by these substructure defects. Thus, thecorrosion behavior of the intermetallics stemsfrom their inherent immunity or susceptibility tocorrosion and from the modifications in macro-scopic parameters, such as grain size, stoichi-ometry, grain-boundary design, microalloying,and second-phase incorporation, to increase thenumber of slip and to hence confer improvementof KIc and fracture strain.This article reviews the progress that has been

made in understanding the corrosion behavior ofintermetallics. Such understanding is essentialfor the modeling of the corrosion processes andfor devising a strategy to create corrosion pro-tective systems by alloy and coating design.The main emphasis is on the high-temperature

corrosion properties of intermetallics, but infor-mation on aqueous corrosion is also revieweddue to the realization that aqueous corrosion canseriously compromise intermetallics usefulness.In the area of high-temperature corrosion, thediscussion is centered on aluminides and sili-cides, while the aqueous corrosion review is

concentrated on fundamental factors that makethe aqueous corrosion of an intermetallic phasedifferent from that of a homogeneous alloy or ofthe constituents in pure elemental form. Ther-modynamic principles in the context of high-temperature corrosion, information on oxidation,sulfidation, hot corrosion of NiAl-, FeAl-, andTiAl-based intermetallics, and silicides areincluded. Aqueous corrosion is divided into twomain topics: thermodynamic consideration,ordering influencing kinetics, stress-crackingcorrosion, and hydrogen embrittlement; andpractical issues of dealing with the corrosionproblems.

High-TemperatureCorrosion of Intermetallics

Three major types of high-temperature cor-rosion are oxidation, sulfidation, and hot corro-sion. Oxidation involves solely the formation ofan oxide scale. Sulfidation is concerned with thedevelopment of scales consisting of sulfide orsulfides. Materials exposed to environmentscontaining other contaminants, in addition tooxygen and sulfur, can cause the developmentof complex scales containing oxides and sul-fides, termed hot corrosion. The article“High-Temperature Gaseous Corrosion” inASM Handbook, Volume 13A, 2003, will assistin understanding the fundamentals of high-temperature corrosion as well as the testingmethods cited in this article.

General Principles

Important aspects of high-temperature corro-sion are the processes of scale formation andscale degradation. Two additional modes ofdegradation that confer susceptibility of materi-als to high-temperature corrosion are inter-granular corrosion and scale vaporization.

The scaling process in high-temperature cor-rosion involves the formation of a thermo-dynamically stable corrosion product (a scale)that separates the surface of the material from theaggressive environment. Although the followingdiscussion begins with basic principles, these

principles are very important and useful tounderstanding the formation of oxide/sulfidescales and the high-temperature corrosion mech-anisms of intermetallics.The formation of a defect-free, coherent and

adherent scale containing lattice defects capableof sustaining only cationic and anionic transportallows progressive scale thickening and diffu-sion-controlled parabolic kinetics. Linear kinet-ics predominate in the case of an inherentlynonprotective scale caused by the presence in thescale of inappropriate defect structures, physicaldefects, or by stress-induced scale spallation.A complex scaling process characterizes alloycorrosion accompanying the formation of amultiphase, multilayered scale; each layer growsin a parabolic rate with different rate constants.This steady-state scale development is oftenproceeded by the competitive processes ofnucleation and growth of transient corrosionproducts dictating the mode and nature ofsubsequent scale growth.Equilibrium thermodynamics, although not

predictive, allow an assessment of the natureof the possible reaction products, whether ornot significant evaporation or condensation ofa given species is likely, and the conditionsunder which a given product can react with acondensed deposit. See the article “Thermo-dynamics of Gaseous Corrosion” in ASMHandbook, Volume 13A, 2003. The standardfree energies of formation (DG�) of oxidesand sulfides as a function of temperatures andthe corresponding dissociation pressures ofthe oxides and sulfides are conveniently sum-marized in the form of Ellingham/Richardsondiagrams, as illustrated in Fig. 1 and 2 (Ref 4).Along the ordinates are plotted values ofDG� forthe oxides and sulfides and of the partial molarfree energy of oxygen and sulfur, while thetemperature is plotted along the abscissa. Thevalues of DG� refer to the standard free energiesof formation of oxides and sulfides per moleof oxygen or sulfur, for example, 4/3AlþO2=2/3Al2O3.In an environment containing oxygen and

sulfur, the following reactions need to be con-sidered for a divalent metal, M:

M+1/2O2=MO (Eq 1)

© 2005 ASM International. All Rights Reserved.ASM Handbook, Volume 13B, Corrosion: Materials (#6508G)

www.asminternational.org

M+1/2S2=MS (Eq 2)

The equilibrium oxygen and sulfur partial pres-sures are defined by the following relations:

p1=2O2

=expDG�

MO

RT

� �(Eq 3)

p1=2S2

=expDG�

MS

RT

� �(Eq 4)

Equations 3 and 4 allow the establishment ofthe conditions necessary for oxidation or sulfi-dation; however, a further reaction must beconsidered, namely:

MS+1/2O2=MO+1/2S2 (Eq 5)

with the equilibrium condition:

p1=2S2

=p1=2O2=exp

DG�MS

RT7

DG�MO

RT

� �aMS

aMO

(Eq 6)

If unit activities are assumed for the phasesMS and MO, Eq 6 can be reduced to:

p1=2S2

=p1=2O2=exp

DG�MS

RT7

DG�MO

RT

� �(Eq 7)

Examination of Eq 3, 4, and 7 permits theidentification of various limiting situations con-cerning the type of surface corrosion productsthat may be formed:

� If (pO2)gas4(pO2

)eq and (pS2 )gas5(pS2 )eq, thenMO is the only stable surface phase.


)eq and (pS2 )gas4(pS2 )eq, thenMS is the only stable surface phase.


)eq and (pS2 )gas4(pS2 )eq, thenbothMO andMS should be stable and form assurface products.

However, reference to Eq 7 indicates that onlyone phase will form, depending on which of thefollowing conditions prevails:

� (pS2=pO2)gas4(pS2=pO2

)eq. This condition willcause Eq 5 to proceed to the left, and MS willbe the stable phase, where the metal is incontact with the gas.

� (pS2=pO2)gas5(pS2=pO2

)eq. In this case, MOwill be the stable phase, and Eq 5 will proceedto the right.

If the equilibrium partial pressures of theoxidants in the environment are known, a ther-modynamic stability diagram can be constructedfor a given temperature, as shown in Fig. 3(Ref 5). Such a thermodynamic diagram givesthe stability range for all relevant phases, in thiscase, metal, oxide, and sulfide, at a given tem-perature. The boundaries are calculated fromthermodynamic data for the relevant reactions.The line between the oxide and sulfide remainsunchanged by activity changes in the alloy. Thecorrosion conditions, that is, the oxygen andsulfur pressures in a given gas atmosphere,represent a point on the diagram.The fields in the schematic thermodynamic

stability diagram are indicated:

� A, metal is the only stable phase.� B, oxide is the only stable phase.� C, sulfide is the only stable phase.

It should be stated that although the gasequilibrium can be calculated readily for manygasmixtures at a given temperature and pressure,in many cases, even at temperatures as high as1000 �C (1830 �F), gas equilibrium is notestablished. Furthermore, it should be noted thatas soon as the metal surface is either partially orcompletely covered by the corrosion product,corrosion would then not be exclusively deter-mined by thermodynamics. The kinetic factors,such as diffusivity of the different alloying ele-ments and of the reactive species (oxygen, sul-fur), as well as the morphological features of thescales formed significantly influence the degra-dation mechanisms of high-temperature corro-sion. The analysis leading to this conclusion isbased on the Wagner model (Ref 6) for theselective oxidation of an active element in abinary alloy to form a continuous external scalein the absence of transient oxidation. See thearticle “Kinetics of Gaseous Corrosion Pro-cesses,” in ASM Handbook, Volume 13A, 2003.According to the Wagner theory, a continuousexternal layer of oxide should form on a binary(A-B) alloy when the solute concentration in thealloy exceeds a critical atom fraction (NB

crit), asexpressed as:

NcritB =

pg*

3N

(S)O

DOVm

DBVOx

� �1=2

(Eq 8)

–100

0–300

–280

–260

–240

–220

–200

–180

–160

–140

–120

–100

–80

–60

–40

–20

0 400 800 1200 1600 2000 2400

Absolutezero

400 800

Temperature, °C

Temperature, °C∆G° = RT ln pO2, kcal log pO2

, atm

1200 1600 2000

–80 –60 –50 –42 –38 –34 –30 –26

–24

–22

–20

–18

–16

–14

–12

–10

–8

–6

–4

–3

–2

–1

M – melting point

S – sublimation point

4 Al + O 2 =

Al 2O 3

3

2 3

4 V + O 2 =

M

S

V 2O 3

3

2 34 Cr + O 2

= Cr 2

O 3

3

2 3

Ti + O2= TiO2

2Ni + O 2= 2NiO

2Mn + O2= 2MnO

2Co + O2= 2CoO

2H2 + O2

= 2H2O6FeO + O 2

= 2Fe 3O 4

4Fe 3O 4

+ O 2= 6Fe 2

O 3

Si + O2= SiO2

M

Fig. 1 Standard free energies of formation for selected oxides as a function of temperature and oxygen partial pressure.Source: Ref 4

Corrosion of Intermetallics / 491



where NO(S) is the oxygen solubility in the alloy;

DO and DB are the diffusivities of oxygen andsolute in the alloy; Vm and VOx are the molarvolumes of alloy and oxide, respectively; and g*

is the critical volume fraction of oxide scale. It isevident that if NO

(S) and DO are very high, NBcrit

will also be high. If the (NO(S)) and/or (DO) of

oxygen in the alloy are reduced, the value ofNBcrit

can also be significantly reduced. Long-termstability of the protective scale requires that theflux of solute to the alloy/scale interface remainslarge enough to prevent oxides of A frombecoming stable. Pettit (Ref 7), for example,found that there are two critical concentrationsfor the formation of alumina scales on nickel-aluminum alloys: one value required for devel-opment of the alumina scale, and a larger valuerequired for maintaining its stability.

When high-temperature alloys (usually basedon nickel, cobalt, or iron) containing a number ofelements are exposed to oxygen at elevatedtemperatures, oxidation may be anticipated inaccordance with a design rationale. Thus, certainelements (such as chromium, aluminum, andsilicon) with high affinities for oxygen may beexpected to oxidize in preference to thosederived from the base metals with high dis-sociation pressures. This process of selectiveoxidation is the concept for developing oxida-tion-resistant alloys (Ref 8–10). In particular, thecomposition of the alloy is chosen such that thestable oxide is the one that provides the mosteffective protective barrier. The oxides Cr2O3,Al2O3, SiO2, and possibly BeO are of primaryinterest because they exhibit low diffusivities forboth cations and anions and are also highly

stable. Usually, alumina is an excellent barrierto oxygen at temperatures below 1300 �C(2370 �F), but at higher temperatures, oxygenpermeation through silica occurs at a slower rate.References 11 and 12 have a comprehensive

review of the process of sulfidation. Compared tooxidation, sulfidation is characterized by fasterkinetics and the formation of scales with com-plex defective morphologies. Sulfidation is par-ticularly severe in high-pS2 and low-pO2

environments. Table 1 indicates the problems ofdesigning sulfidation-resistant alloys and coat-ings. The parabolic rate constant (kp) for thesulfidation of three elements, iron, cobalt, andnickel, that form the basis of many high-tem-perature oxidation-resistant alloys is typically10�6 to 10�7 g2/cm4/s at 640 to 800 �C (1180 to1470 �F) and is several orders of magnitudehigher than their oxidation kp (~10�11 to10�8 g2/cm4/s) at 800 to 1000 �C (1470 to1830 �F). It is equally important to note that thekp value for the sulfidation of chromium(~10�8 g2/cm4/s) at 750 �C (1380 �F) is sig-nificantly higher than that (~10�13 g2/cm4/s) at800 �C (1470 �F) for the formation of Cr2O3, animportant component in the development ofmany oxidation-resistant alloys. In contrast, therefractory metals (vanadium, molybdenum,niobium, and tungsten), together with zirconiumand hafnium, display low rates of sulfidation.These elements, when incorporated in sulfida-tion-resistant materials, are likely to enhance theresistance to sulfidation, particularly under re-ducing conditions (Table 2).Enhanced sulfidation resistance can be

achieved by employing the same principles usedin increasing the oxidation resistance of materi-als. The incorporation of appropriate elementsinto the base materials undergoing selectivesulfidation leads to the development of a barrierlayer capable of sustaining lower ionic transportrates. While selective sulfidation to form a

H2/H2S ratio

1020

020

0

–40

–80

–120

–160

∆G° =

RT

In p

S2, k

cal

–200

–240

Absolutezero

pS2, atm

400

YS4

Co4S8

2NiSNi3S2

2FeS

TiS2

V2S3

Mo2S3

800

Si-SiS 2

2FeS + S 2= 2FeS 2

1200 1600 2000

1015 1010 108 106 105

Temperature, °C

2H2 + S2 = 2H2S

MoS2

NbS2

2CrS

2MnS

104 103 102 10

1

10–1

10–2

10–3

10–4

10–5

10–6

10–7

10–8

10–9

10–10

10–11

10–12

10–1310–1510–1710–2010–2510–3510–50

10–150 10–80 10–60 10–50 10–42 10–34 10–30 10–2610–24

10–22

10–20

10–18

10–16

10–14

10–12

10–10

10–8

10–6

10–4

10–3

10–2

10–1

102

103

104

pS2, atm

1

10

2 3

4 5

1 4

2 3

Fig. 2 Standard free energies of formation for selected sulfides as a function of temperature and sulfur partial pressure.Source: Ref 4

AMetal

CSulfide

BOxide

Log pO2

Log

p S2

Fig. 3 Schematic of a thermodynamic stability diagramshowing the stable ranges where oxide or sulfide

can be formed at a given temperature. Source: Ref 5

492 / Corrosion of Nonferrous Metals and Specialty Products



sulfide barrier layer will be governed by the freeenergy of formation, the kinetics of this selectiveprocess, and hence the overall rate of sulfidation,will be controlled by the defect structure of thesulfides (Table 3), their ability to support fast orslow diffusion rates as measured by self-diffu-sion coefficients (Table 4), and the meltingpoints and mechanical stability indicated by thePilling-Bedworth ratio. See the article “High-Temperature Gaseous Corrosion” in ASMHandbook, Volume 13A, 2003.

Ni3Al and NiAl

Figure 4 shows the binary NiAl phase diagram(Ref 13). NiAl possesses the ordered cubic B2(cP2) CsCl crystal structure (Ref 1, 13). Thisstructure exists over the composition range of 45to almost 60 at.% Ni. NiAl is strongly ordered,even above 0.65 Tm (where Tm is the meltingpoint), with an intrinsic disorder parameter of

less than 5 ·10�3. The B2 structure is stable forlarge deviation from stoichiometry, and sig-nificant long-range order has been reported. NiAlnot only has the highest melting point of anycompound in the NiAl system but also is themoststable. This high degree of thermodynamic sta-bility and the existence of a wide phase fieldmake NiAl relatively easy to fabricate in a rangeof forms, from fine homogeneous powders tosingle crystals.For four decades, NiAl has been extensively

studied as a potential structural material in theaerospace industry due to (Ref 1, 13):

� High melting point (1638 �C, or 2980 �F), asshown in Fig. 4, which is nearly 300 �C(540 �F) higher than the melting temperatureof conventional superalloys

� Low density (5.35 to 6.50 g/cm3); the densityfor the stoichiometric composition is 5.85g/cm3, roughly two-thirds that of typicalnickel-base superalloys

� Good environmental resistance; the parabolicrate constant is very low, even in composi-tions with up to 60 at.% Ni, and is typically 2orders of magnitude lower than for typicalnickel-base superalloys

� High thermal conductivity� Attractive modulus� Metal-like properties above a modest ductile-

to-brittle transition temperature� Low raw materials cost� Relatively easy processing (conventional

melting, powder, metal forming)

However, two principal drawbacks exist forunalloyed NiAl: poor toughness and damagetolerance at room temperature, and inadequatestrength and creep resistance at elevated tem-peratures.Ni3Al has an L12 ordered and face-centered

cubic crystal structure (Ref 1) that can bemaintained up to 1395 �C (2545 �F) (Fig. 4). Itsunit cell contains four atoms, with three nickel

Table 1 Sulfidation and oxidation parabolic rate constants, kp, of selected metals

Metal

Sulfidation Oxidation (pO2~105 Pa)

Temperature

kp, g2/cm4/s pS2 , Pa

Temparature

kp, g2/cm4/s�C �F �C �F

Al 500 930 1.0 · 10�12 105 1000 1830 1.0 · 10�14

Ti 750 1380 4.5 · 10�12 10�1 800 1470 3.3 · 10�10

Zr 750 1380 2.7 · 10�12 10�1 800 1470 6.6 · 10�11

Hf 750 1380 7.4 · 10�12 10�1 800 1470 3.3 · 10�11

V 750 1380 2.3 · 10�10 10�1 . . . . . . LinearNb 750 1380 8.2 · 10�13 10�1 . . . . . . . . .Cr 750 1380 1.9 · 10�8 10�1 800 1470 1.0 · 10�13

Mo 750 1380 2.6 · 10�12 10�1 . . . . . . LinearW 750 1380 2.0 · 10�9 10�1 . . . . . . LinearFe 800 1470 2.0 · 10�7 105 800 1470 5.5 · 10�8

Ni 640 1185 1.6 · 10�6 105 1000 1830 9.1 · 10�11

Co 800 1470 6.7 · 10�6 105 900 1650 2.0 · 10�8

Source: Ref 11

Table 2 Parabolic rate constants, kp, of selected metals and alloys

Metal or alloy

Temperature Atmosphere, Pa

kp, g2/cm4/s�C �F pS2 pO2

V 750 1380 10�1 10�18 2.3 · 10�10

Nb 750 1380 10�1 10�18 8.2 · 10�13

Mo 750 1380 10�1 10�18 2.6 · 10�12

Fe-20Nb 700 1290 103 . . . 8.4 · 10�8

Fe-30Nb 700 1290 103 . . . 3.5 · 10�8

Ni-20Nb 700 1290 103 . . . 7.7 · 10�7

Ni-30Nb 700 1290 103 . . . 1.3 · 10�7

Co-20Nb 700 1290 103 . . . 1.6 · 10�7

Co-30Nb 700 1290 103 . . . 5.9 · 10�8

Fe-20Mo 700 1290 103 . . . 8.4 · 10�8

Fe-30Mo 700 1290 103 . . . 3.5 · 10�8

Ni-20Mo 700 1290 103 . . . 1.0 · 10�7

Ni-30Mo 700 1290 103 . . . 1.5 · 10�8

Co-20Mo 700 1290 103 . . . 4.4 · 10�8

Co-30Mo 700 1290 103 . . . 2.0 · 10�9

Co-20Cr-3.5Al-1Y-5V 750 1380 10�1 10�18 3.4 · 10�9

Co-20Cr-3.5Al-1Y-5Nb 750 1380 10�1 10�18 3.1 · 10�9

Co-20Cr-3.5Al-1Y-5Mo 750 1380 10�1 10�18 4.3 · 10�9

Co-20Cr-3.5Al-1Y-5W 750 1380 10�1 10�18 8.9 · 10�9

Co-20Cr-3.5Al-1Y-10V 750 1380 10�1 10�18 7.2 · 10�10

Co-20Cr-3.5Al-1Y-10Nb 750 1380 10�1 10�18 1.8 · 10�9

Co-20Cr-3.5Al-1Y-10Mo 750 1380 10�1 10�18 6.2 · 10�9

Source: Ref 11

Table 3 Physico-chemical properties ofsulfides of certain metals

Sulfide

Pilling-Bedworth

ratioDefect

structure

Melting point

�C �F

Al2S3 2.60 n-type 1099 2010TiS2 1.11 n-type 1999–2099 3630–3810ZrS 1.91 n-type 1549 2820HfS . . . . . . 2100–2273

(estimated)3810–4123(estimated)

V2S3 . . . n-type 1799–1999 3270–3630NbS2 . . . n-type . . . . . .TaS2 2.42 n-type 999 1830Cr2S3 2.50

(CrS)n/p-type 1550 2820 (CrS)

MoS2 3.54 n-type 1457 2655WS2 3.47 n-type 41800 43270FeS 2.50 p-type 1189 2172Co9S8 2.37 p-type 1080 1975NiS 2.50 p-type 796 1465

Source: Ref 11

Table 4 Self-diffusion coefficients (DM) ofcations in some metal sulfides and oxides

Material

Temperature

DM, cm2/s�C �F

Sulfide

Cu2þyS 650 1200 5.15 · 10�5

Co1�yS 720 1330 7.0 · 10�7

Ni1�yS 800 1470 1.4 · 10�8

Fe1�yS 800 1470 3.5 · 10�7

Cr2S3 1000 1830 1.0 · 10�7

Al2S3 600 1110 1.0 · 10�13

Oxide

Cu2�yO 1000 1830 1.7 · 10�8

Co1�yO 1000 1830 1.9 · 10�9

Ni1�yO 1000 1830 1.0 · 10�11

Fe1�yO 800 1470 1.3 · 10�8

Cr2O3 1000 1830 1.0 · 10�12

Al2O3 1000 1830 1.0 · 10�16

Source: Ref 11




atoms occupying face-centered positions andwith one aluminum atom at corner positions.Ni3Al has a homogeneous phase field ofapproximately 3 at.% around its stoichiometriccomposition, which is very important for alloydesign. With off-stoichiometric compositions,antisite substitutions are formed on both nickel-rich and aluminum-rich sides of Ni3Al. This isquite different from B2-NiAl, where antisitedefects are stable in the nickel-rich side andvacancies are formed in the aluminum-richside.Oxidation. In the oxidation of nickel alumi-

nides, Al2O3 grows in different modifications,depending on the oxidation temperatures. Atlower temperatures and/or in the early stagesof oxidation, the metastable oxides c-Al2O3,d-Al2O3, and h-Al2O3 develop. These types ofalumina contain a high concentration of cationvacancies. In the early stages of oxidation,the cubic oxides c-Al2O3 and d-Al2O3 havebeen observed on zirconium-doped NiAl (Ref14). Later, they transform to the monoclinich-Al2O3 (Ref 15), and finally, the transformationto a-Al2O3 takes place (Ref 14–17). Thecubic alumina modifications and their transfor-mation to h-Al2O3 were observed at lowertemperatures (800 to 1000 �C, or 1470 to1830 �F). The transformation to a-Al2O3 wasobserved at higher temperatures (1000 �C, or1830 �F, and over).The oxidation of NiAl generally results in the

formation of a protective alumina surface scale.The formation of protective a-Al2O3 scale canbe controlled by favorable alloying additions.The formation of pores and voids beneath thescale leads to a decrease in the area of coherence.Especially in NiAl with nickel excess, aluminumdepletion by Al2O3 growth causes rapid diffu-sion of nickel into the material. Loss of nickeland aluminum beneath the scale leads to voidnucleation and growth at the oxide/metal inter-face. Void formation also most probably occurs

at other heterogeneities, for example, grainboundaries, which may allow oxygen ingress atgrain boundaries, and intergranular oxidation ofNiAl may occur. Intergranular oxidation can befollowed or accompanied by internal oxidationof NiAl. NiAl with high aluminum content andmultiphase alloys are much less susceptible tointragranular and internal oxidation attack(Ref 18).

It is found that the oxide scales forming onNi3Al at Ti1000 �C (1830 �F) in air consist oftransiently equiaxed NiO/NiAl2O4 or c-Al2O3

above a columnar a-Al2O3 protective layer(Ref 19–21). Thus, the steady-state oxidationmechanism is governed by transport througha-Al2O3 grain boundaries, although the resultsmay vary depending on the oxidizing conditions(Ref 22, 23). The steady-state oxide scalesformed on binary Ni3Al are usually not adherent.Steady-state oxide scales are typically columnar-grained a-Al2O3. Temperatures below 1000 �C(1830 �F), low oxygen partial pressures (lessthan the NiO/Ni3Al dissociation pressure), andhigh-humidity environments favor the formationof transition Al2O3 (Ref 19, 24).

The benefits of active element additions toNiAl are well documented in the literature(Ref 25–30). The inclusion of elements such asyttrium and hafnium will have a number ofbeneficial results, such as decreasing the oxidegrowth rate, decreasing voids at the oxide/sub-strate interface, increasing the mechanicalproperties of the oxide, and acting as sulfurgetters. The effects of the ion implantation ofyttrium on the oxidation behavior of b-NiAl afterisothermal and thermal cycling at temperaturesranging from 1000 to 1300 �C (1830 to 2370 �F)is demonstrated in Ref 31. The use of ionimplantation as a method of adding activeelements has been reported to have severaladvantages over other processes. These includea decrease in segregation of the active elementsto the grain boundaries of the substrate and

the deposition of a thin layer onto the samplesurface in the case of higher doses. At a levelof 2 · 1014 Yþ/cm2, the addition has been foundto have little effect. However, at a level of2 · 1016 Yþ/cm2, the yttrium has been shown topromote the adherence of the alumina scaleand to decrease the oxidation rate.The effects of noble metals on the oxidation

behavior of nickel aluminide coatings wereextensively reported (Ref 32–34). Whether usedas an environmental coating (Ref 35) or as a bondcoat for thermal barriers (Ref 36, 37), platinumaluminides offer a higher level of protection thanthat afforded by standard aluminide coatings.This is due to the beneficial interaction betweenplatinum and aluminum, which helps to promoteand maintain the protective scale (Ref 38). First,a stable platinum oxide is not formed at theoperating temperatures of the turbines (Ref 39),and second, aluminum is very mobile in plati-num-rich phases, whereas other elements have tocompetitively diffuse relatively slowly to theoxide/coating interface (Ref 40). In essence, theplatinum helps to create a reservoir of aluminumin the outer portion of the coating. This in turnpromotes a more slowly grown, compact, andadherent alumina scale, which enables platinumaluminides to offer a greater level of protectionin oxidizing and corrosive environments thanstandard aluminides (Ref 41). Palladium-mod-ified aluminides have been reported to havebetter mechanical properties than platinum alu-minides and yet still offer a comparable level ofcorrosion resistance (Ref 42), whereas the addi-tion of rhodium is thought to increase bothcoating stability and oxidation resistance (Ref43). However, the performance of iridium-modified aluminides has rarely been reported inthe open literature. Iridium exhibits relativelylow oxidation rates, compared to other refractorymetals, and is known to have a low oxygen dif-fusivity (Ref 44). Iridium is less expensive thanplatinum, and investigations of IrAl inter-metallics have found that they have potential asalumina formers (Ref 45, 46).Sulfidation. The high-temperature sulfida-

tion behavior of nickel aluminides has beenshown to be strongly influenced by the alloy andgas compositions. The sulfidation behavior offour nickel-aluminum alloys containing 25 to45 at.%Al studied (Ref 47) over the temperaturerange of 750 to 950 �C (1380 to 1740 �F) in a gasmixture of H2/H2S (0.1 to 10 vol%) has beenreported to follow parabolic kinetics. Double-layered scales consisting of an outer layer ofNi3S2 and an inner layer of NiAl3.5S5.5 form onall alloys regardless of the aluminum content,indicating insignificant influence of aluminumcontent on sulfidation rate. However, the lowoxygen partial pressure in H2/H2S/H2O mixtureshas been shown to significantly influence thesulfidation behavior; severe attack by rapidinternal oxidation destroys all the alloys exceptNi25Al (25 at.% Al), with the internal oxidationzone consisting of a mixture of c0-Ni3Al andAl2O3. On the alloys containing 36 and 45 at.%Al, local attack occurs, with fast-growing pocks

400

800

600

1200

1000

1600

1400

1800

0 10 504020 30 60 70 80 90 100

(Al)

639.9 °C

660.9 °C 854 °C

Al3Ni5

~700 °C

AlNi3

1133 °C

AlNi(Ni)

1455 °C

1638 °CL

Al 3

Ni

Al 3

Ni 2

Al NiNickel (at%)

Tem

pera

ture

(° C

)

Fig. 4 The binary nickel-aluminum phase diagram




forming after an incubation period. An inter-ruption of the H2S gas flow stops internal oxi-dation; virtually no internal oxidation occurs inflowing H2/H2O atmospheres. These facts indi-cate that H2S is essential for triggering andmaintaining the internal oxidation of the nickelaluminides.The effects of H2S concentration on the sul-

fidation behavior of Ni3Al at 875�C (1605 �F)

in several H2/H2S gas mixtures ranging from0.16 to 4.7 vol% H2S are demonstrated in Fig. 5(Ref 3). Also shown for comparison are weight-change data for oxidation of the alloy in an airenvironment. The weight-change data indicatethat at H2S levels up to 0.75 vol%, the reactionrate follows a parabolic behavior, and at H2Slevels of 1.5 vol% or higher, the weight-changedata exhibit accelerated corrosion. At low levelsof H2S (0.16 and 0.38 vol%) in the gas phase, thesulfidation rates become similar to the oxidationrates in an air environment. As the H2S levelincreases, the sulfidation rate increases, and at~1.5 vol% H2S, liquid Ni-Ni3S2 eutectic forms,and accelerated corrosion ensues. At a H2S levelof 4.7 vol%, the calculated rate constant forsulfidation is 8.7 · 10�10 g2/cm4/s, which is ~4orders of magnitude larger than the rate constantfor the oxidation of the alloy. At higher sulfurlevels, the weight-change curves show anaccelerated rate for longer times, due primarilyto a decrease in the reaction surface area pro-duced as the reaction front advances deeper intothe alloy.Hot corrosion studies of the NiAl inter-

metallics are very few. An investigation of hotcorrosion (Ref 48) on a Ni3Al alloy at 605, 800,and 1000 �C (1120, 1470, and 1830 �F)demonstrates the significant effect of tempera-ture. The specimens coated with 1+0.1 mg/cm2

of Na2SO4-Li2SO4 and then exposed to a 1%SO2/air gas mixture have indicated highestweight gain at 1000 �C (1830 �F) and leastweight gain at 800 �C (1470 �F). It is proposedthat NiO oxide formation consumes oxygen inthe molten salt. The consumption of oxygenlocally reduces the oxygen potential andincreases the sulfur partial pressure in the moltensalt. As the sulfur partial pressure reaches theequilibrium partial pressure region of NiSx and/

or AlSx, NiSx and/or AlSx forms at the salt/alloyinterface through sulfidation reaction. The con-sumption of sulfur decreases the sulfur potential,and oxygen partial pressure increases in themolten salt. This leads to the formation of NiOagain. This mechanism also suggests that NiOand NiSx and/or AlSx are produced simulta-neously. Because the produced sulfide is ther-modynamically unstable when the oxygenpotential increases, it is possible for the sulfidesto convert into oxides (NiO, Al2O3, andNiAl2O4) through the necessary reactions. Thereare two possibilities for the formation of spinelphase, which is produced either through thereaction of aluminum and nickel with oxygen inthe molten salt or through the evolution ofsulfides.

Fe3Al and FeAl

The iron aluminides, mainly Fe3Al and FeAl,are of interest for many land-based applicationsbecause of their appropriate mechanical proper-ties, ease of fabrication, excellent oxidation andcorrosion resistance, conservation of strategicelements, low density, and low cost. Moreover,FeAl is characterized to possess good resistanceto catalytic coking, carburization, sulfidation,and wear. Therefore, iron aluminides are beingdeveloped for use as structural materials and/oras cladding for conventional engineering alloys.Figure 6 presents the phase diagram of the iron-aluminum binary system (Ref 1). In the ironaluminide system, the alloys of interest are ofcomposition Fe3Al and FeAl. It is reported (Ref3) that the crystal structure of Fe3Al is D03 or B2,

while FeAl has the B2 structure. The meltingpoint temperatures of Fe3Al and FeAl are 1520and 1250 �C (2770 and 2280 �F), respectively,and densities are 6.72 and 5.56 g/cm3, respec-tively. However, the Young’smodulus values forFe3Al and FeAl are 140.6 and 260.4 GPa (20 and38 · 106 psi), respectively, and the stiffer inter-metallic has a tendency to be much more brittle.Their tensile strength also compares favorablywith many ferritic and austenitic steels. How-ever, Fe3Al and FeAl are susceptible to envir-onmental embrittlement in the presence of watervapor, although the degree of sensitivity of Fe3Alappears to be less than FeAl. Limited ductility atambient temperatures and a sharp drop instrength at above 600 �C (1110 �F) have beenmajor deterrents to their acceptance for manystructural applications.Oxidation of FeAl-base intermetallic alloys

usually shows different behavior in air andoxygen atmospheres (Ref 49); the oxidation ratesof Fe-37Al oxidized in ambient air are higherthan in oxygen at 1000 to 1200 �C (1830 to2190 �F), with a double-layered oxide scaleforming on the surface of Fe-37Al specimensduring oxidation. In the case of air oxidation, theouter layer consists of convoluted whiskers of a-Al2O3, while the inner layer is comprised ofAl2O3þAlN. A convoluted a-Al2O3 layer withthe inner layer consisting of Al2O3 is the result ofoxidation in pure oxygen. It is believed that thenitrogen ions in Al2O3 and the growth of whiskeroxides cause an increase of the oxidation ratein air.Doping with yttrium, zirconium, and yttri-

umþ zirconium has a strong influence on the

Aluminum, at.%

01600

1400

1200

1000

800

600

4000 10 20 30 40

FeAl

(γFe)

(αFe)

ε

1102 °C

L

1310 °C

1165 °C1169 °C ~1160 °C

1232

Tc

Fe3Al

FeAl2

Fe 2

Al 5

FeAl3655 °C

(Al)

50

Aluminum, wt%

Tem

pera

ture

,°C

Fe Al

60 70 80 90 100

660.452 °C

912 °C

770 °C

1538 °C

1394 °C

2010 30 40 50 60 70 80 90 100

Fig. 6 The binary iron-aluminum phase diagram. Source: Ref 1

00

0.4

Wei

ght g

ain,

mg/

cm2

0.8

1.2

1.6

40 80 120

Exposure time, h

4.7% H2S-H2

160

0.38% H2S-H2

0.5% H2S-H2

1.5% H2S-H2

200 240

Air

Fig. 5 Thermogravimetric test data for sulfidation ofNi3Al at 875

�C (1605 �F). Source: Ref 3




oxidation kinetics of Fe-37Al intermetallics.Yttrium additions to Fe-37Al decrease the oxi-dation rates at 1000 to 1200 �C (1830 to2190 �F). Fe-37Al-0.1Y-0.2Zr revealed loweroxidation rates at 1100 and 1200 �C (2010 and2190 �F) than Fe-37Al but not at 1000 �C(1830 �F). Fe-37Al with zirconium demon-strated higher oxidation rates than Fe-37Al. Fe-37Al-0.8Zr exhibits a lower oxidation rate at1100 and 1200 �C (2010 and 2190 �F) than Fe-37Al-0.3Zr. The effects of the reactive elementson the isothermal oxidation of Fe-37Al can beclassified into different groups. The compactoxide scale formed on the doped Fe-37Al redu-ces the oxidation rate. Reactive elements formspecial oxides that have faster growth rates thanthe aluminum oxides. Furthermore, reactive ele-ments, such as yttrium or yttriumþ zirconium,increase the transformation temperature ofh-Al2O3 to a-Al2O3. The scale growth onundoped FeAl is influenced by the cation andanion short-circuit diffusion. The scale growthon doped FeAl is controlled mainly by anionshort-circuit diffusion. The oxide grain sizehas been observed to have significant effectson the effective diffusion coefficient of thescales formed on the undoped Fe-37Al but noton the yttrium-doped Fe-37Al alloys (Ref 50,51). The results obtained from the studiesof oxidation kinetics of the intermetallic phaseb-FeAl in the temperature range of 800 to1100 �C (1470 to 2010 �F) (Ref 52) demon-strate that at temperatures 4900 �C(1650 �F), the initially formed metastablealuminum oxides are converted to adherentand slow-growing a-Al2O3 scales.Parabolic kinetics after a period of transient h-

Al2O3 scale formation characterizes the oxida-tion behavior of Fe-40Al-1Hf, Fe-40Al-1Hf-0.4B, and Fe-40Al-0.1Zr-0.4B (at.%) alloys at900, 1000, and 1100 �C (1650, 1830, 2010 �F)(Ref 53). The isothermally grown scales areassociated with a propensity toward massivescale spallation due to both extensive rumplingfrom growth stresses and an inner layer of HfO2.Spallation in cyclic oxidation for 200 one-hourcycles produced little degradation at 900 or1000 �C (1650 or 1830 �F), but caused sig-nificant spallation at 1100 �C (2010 �F) in theform of small segments of the outer scale. Themajor difference in the cyclic oxidation of thethree FeAl alloys mentioned previously isincreased initial spallation for FeAlþ zirconiumand boron. Although these FeAl alloys indicatemany similarities to NiAl alloys, they are gen-erally less oxidation resistant. It is believed thatthis stems from the presence of nonoptimallevels of dopants and larger thermal-expansionmismatch stresses.It is reported (Ref 54) that the addition of a

reactive element (RE)—yttrium (usually) and/orhafnium—significantly improves the oxideadherence of Fe3Al over the range of tempera-ture from 900 to 1100 �C (1650 to 2010 �F) forup to 240 h. Without RE additions, the Al2O3

scales developed on Fe3Al alloys becomeconvoluted, with the growth of the oxide scale

being controlled by a mixed-diffusion mode ofaluminum and oxygen transport. Reducedaluminum diffusion and the formation of a flatscale growing mainly by the inward transportof oxygen are associated with adding REs,especially yttrium. The reduction in the transportof aluminum in RE-doped Fe3Al alloys, reducingthe Fe3Al alloy rate of oxidation, probably stemsfrom the segregation of the RE to the scale/alloyinterface. Extensive intergranular oxidationof the yttrium-containing Fe3Al alloys has beenobserved, probably due to the segregationof yttrium (not hafnium) to the alloy grainboundaries. The detrimental effect of excessivehafnium content on the oxide growth is believedto be due to the formation of hafnium-richoxide particles facilitating inward scale growth,leading to the formation of localized oxide-thickening “pegs” and the eventual formation ofless protective oxide, especially under thermalcycling.

Sulfidation. The results obtained from thesulfidation experiments conducted on severalsets of iron aluminides at temperatures between400 and 1000 �C (750 and 1830 �F) (Ref 55) areshown in Fig. 7, giving thermogravimetric datafor ternary iron aluminide tested in a 1.35 vol%H2S/H2 gas mixture at 650, 875, and 1000 �C(1200, 1605, and 1830 �F). Also shown in thefigure are data for type 310 stainless steel oxi-dized in air at 1000 �C (1830 �F) and sulfidizedin the 1.35 vol% H2S/H2 gas mixture at 875 �C(1605 �F). Figure 8 shows some scanning elec-tron micrographs of surfaces of Fe3Al and type310 stainless steel specimens after sulfidation.The temperature dependence of the morpholo-gies of surface sulfides on the Fe3Al probablyresults from the variable thermodynamic activityof sulfur in the exposure environment, with theactivity being lowest at 1000 �C (1830 �F) andhighest at 650 �C (1200 �F) (H2S concentrationin the gas was kept constant). In the scales on theFe3Al consisting of (Fe,Al) sulfides and ironsulfides, the relative proportion of the former tothe latter decreases with decreased temperature.The scale on the type 310 stainless steel is pre-dominantly (Fe,Cr) sulfides with some nodulesof iron sulfide.

Comparative studies have been conducted onthe sulfidation resistance of Fe3Al and severalchromia- and alumina-forming alloys in oxygen/sulfur mixed-gas environments (Ref 55, 56).Thermogravimetric studies on oxidation of iron-base alloys with differing aluminum concentra-tions and Fe3Al alloys show that a minimumaluminum level of 12 wt% is needed to develop acontinuous alumina scale that is resistant tosulfur attack. A detailed comparison has beenmade of the corrosion performance of alumina-and chromia-forming alloys exposed to oxygen/sulfur mixed-gas environments.

The results obtained from the high-tempera-ture corrosion behavior of FeAl (42 at.% Al)intermetallics in a mixture gas (95% N2þ 5%H2) plus 1% H2S at 600 �C (1110 �F) showexcellent corrosion resistance for FeAl in thisatmosphere; the mass gain of FeAl remains

50.15 mg/cm2 after exposure to the atmospherefor 90 h at 600 �C (1110 �F). The depth profileof x-ray photoelectron spectroscopy analysisindicates the formation of the double-layeredscale: a layer of a mixture of FeS and Al2O3 onthe top surface and an Al2O3 layer on the bottom(Ref 57).The corrosion behavior of Fe-28Al and Fe-

18Al-10Nb (at.%) over the temperature range of700 to 900 �C (1290 to 1650 �F) in a H2/H2S/H2O gas mixture with varying sulfur partialpressures of 10�2 to 103 Pa and oxygen partialpressures of 10�19 to 10�15 Pa is characterized,in general, by parabolic kinetics, although two-stage kinetics are noted in some cases. Thesteady-state parabolic rate constants increasewith increasing temperature. Based on theequivalent addition (28 at.%) of alloying ele-ments, Fe-18Al-10Nb exhibits a better corrosionresistance, being approximately 2 to 4 orders ofmagnitude lower than Fe-28Al (depending ontemperature). The scales formed on Fe-28Alconsist of mostly FeS, a-Al2O3, and minorFeAl2S4, while the scales formed on Fe-18Al-10Nb consist of mostly a-Al2O3, Nb3S4, Nb2O5,and minor FexNbWS2 (FeNb2S4/FeNb3S6) atTj800 �C (1470 �F) and of a-Al2O3, Nb3S4,and FexNbS2 at 900

�C (1650 �F). Significantlyreduced corrosion rates follow the formation ofAl2O3 and Nb3S4 (Ref 58, 59).The minimum aluminum concentration need-

ed to resist sulfidation and oxidation in H2S/H2/H2O environments was established in Ref 60;alloys containing i18% Al are uniquely resis-tant to H2S-containing environments at 800 �C(1470 �F). Chromium adversely affects the cor-rosion resistance in this mixed-gas environment,although this effect is partially offset by theaddition of molybdenum. Zirconium and yttriumhave no significant influence on the corrosionrate under the same temperature and environ-mental conditions. Investigations (Ref 61) on theeffects of yttrium, hafnium, and yttriumþhafnium on the sulfidation behavior of Fe3Al at900 �C (1650 �F) in a H2S/H2/H2O environmentdemonstrate that additions of the RE confersignificantly increased sulfidation resistance

310 stainless steel, 875 °C, sulfidation

Fe3Al + 5 wt% Cr, 1000 °C, sulfidation

310 SS, 1000 °C, oxidation

Fe3Al + 5wt%Cr, 650 °C, sulfidation

Fe3Al + 5wt%Cr,875 °C, sulfidation

0 40 80

Exposure time, h

Wei

ght c

hang

e, m

g/cm

2

–2

0

2

4

120 160

Fig. 7 Thermogravimetric test data for chromium-containing Fe3Al exposed to 1.35 vol% H2S/H2

gas mixture at 650, 875, and 1000 �C (1200, 1605, and1830 �F). Also shown are data for oxidation (at 1000 �C, or1830 �F) and sulfidation (at 875 �C, or 1605 �F) for type310 stainless steel. Source: Ref 3




to Fe3Al. However, the combined addition ofyttriumþ hafnium compromises the benefit tothe sulfidation-resistance enhancement. It isinteresting to note that a superlattice of MnSformed on the sulfidized Fe3Al with the additionof yttrium and hafnium, as revealed in Fig. 9.Hot Corrosion. Important information on

the hot corrosion behavior of FeAl intermetallicscomes from the studies on binary iron aluminide,Fe-25Al (at.%), at 827, 952, and 1057 �C (1521,1746, and 1935 �F) (Ref 62). The results ob-tained from hot corrosion studies conducted bycoating the specimen surfaces with 2.5+0.2 mg/cm2 of Na2SO4 prior to exposure inpure oxygen show parabolic rate constants. Thefaster kinetics in the initial stages of oxidationare thought to be related to the formation of h-Al2O3 and the slower kinetics in the later stagesof oxidation to the formation of a-Al2O3, withthe overall rate of hot corrosion being higher thanthat of oxidation at all the temperatures. Thepresence of a-Fe2O3 in addition to alumina isindicated by x-ray diffraction (XRD) analysis ofthe scales present on the surface of the samplesafter hot corrosion. Fourier transform infraredspectra from the spalled scales in hot corrosiondivulge the presence of a-Al2O3, a-Fe2O3, andsulfate. Cross-sectional microscopy reveals pit-ted metal/scale interfaces in hot corrosion con-ditions, with the pits containing aluminumsulfide. Sulfides are also identified along the

grain boundaries in the intermetallic near thescale/metal interface.

Ti3Al and TiAl

Recent research and development efforts onTi3Al and TiAl ordered intermetallics, which arepotential candidates for the structural materialsin aerospace and automobile industries, haveresulted in considerable improvement in theirmechanical and metallurgical properties. Thebinary titanium-aluminum phase diagram isgiven in Fig. 10. Intermetallic materials withimproved properties—superior specific stiffnessand strength—over conventional titanium-basealloys in the intermediate temperature range of600 to 800 �C (1110 to 1470 �F) have beendeveloped. Among other properties, high-tem-perature strength (up to 800 MPa, or 116 ksi,ultimate tensile strength at 700 �C, or 1290 �F),creep resistance (typically primary ~1% strainafter 500 h), and fatigue resistance (typically480 MPa, or 70 ksi, runout following turning,and 730 MPa, or 106 ksi, after high-speed mil-ling) make c-titanium aluminides ideal materialsfor weight-critical components subject to med-ium temperatures (up to 800 �C, or 1470 �F)/fluctuating stress. Themain advantage for the useof c-TiAl is weight saving, because it is typicallyhalf the density of alternative nickel-base

superalloys. High stiffness (~130 GPa, or19 · 106 psi, at 700 �C, or 1290 �F) is a furtheradvantage (Ref 63, 64). The maximum usefulservice temperature for c-TiAl appears to belimited to 800 �C (1470 �F) today; however,future applications are targeting operatingtemperatures 4800 �C (1470 �F). At thesetemperatures, titanium aluminides will besubjected to severe environmental attack, parti-cularly oxidation during service. Clearly,to achieve such enhanced high-temperaturecapability, it is essential to devise strategiesbased on alloy development and/or protectivecoating design. Such strategies must be under-pinned by a thorough understanding of thecorrosion behavior of these intermetallics.Oxidation behavior of TiAl and TiAl-base

materials displays a complex pattern (Ref 65–68). The formation of a slow-growing, adherent,and protective Al2O3 scale is muchmore difficultin TiAl intermetallics than in disordered tita-nium-aluminum alloys and in NiAl (Ref 69, 70).

Fig. 8 SEM micrographs of surface of Fe3Al and type 310 stainless steel after sulfidation exposure to a H2/1.35 vol%H2S gas mixture. Fe3Al at (a) 650

�C (1200 �F), (b) 875 �C (1605 �F), and (c) 1000 �C (1830 �F); (d) type 310stainless steel at 875 �C (1605 �F). Source: Ref 3

Fig. 9 Superlattice layer formed on the sulfidizedFe3Alþ yttrium-hafnium alloy in a H2/H2S/H2O

environment at 900 �C (1650 �F) for 240 h. (a) Dark-fieldimage. (b) Electron diffraction pattern. Source: Ref 61




The difficulty in the development of Al2O3 scalein TiAl intermetallics is associated with theretardation in the selective oxidation of alumi-num. The thermodynamic imperative in pro-moting the formation of both titanium andaluminum oxides leads to the development ofscale layering comprised of TiO2/Al2O3/TiO2.The number of layers formed is influenced bytemperature, time, and the partial pressure ofoxygen.In order to understand the oxidation mechan-

isms from a fundamental point of view, theauthors (Ref 71) have carried out nanoscalestudies of the early stages of the oxidationbehavior of TiAl, allowing nanoscale informa-tion on the scaling processes to be obtained. Atwo-stage strategy was used. The first stageinvolved scanning tunneling microscopy (STM)/scanning tunneling spectroscopy (STS) investi-gations of TiO2(110)-(1 by 1) surface in order toestablish “fingerprints” for the identification ofthe products formed during oxidation of TiAl. Inthe second stage, a TiAl intermetallic alloy wasstudied by STM/STS after repeated sputteringand heating (to provide ideal surfaces) in bothlow- and high-oxygen potential environments.For TiO2, it was found that the oxygen vacanciesproduced created additional defect states in theband gap of stoichiometric TiO2 (Fig. 11b). Thisstate can be taken as the fingerprint for thereduced TiO2 surface, indicating the presence ofTi2O3. The energy for this state, estimated fromthe STS measurement, is in agreement withprevious results (Ref 71).Following oxidation at room temperature with

100 Langmuir oxygen, the TiAl surface, withwell-developed terraces, ledges, and kinks,

shows islands of oxides of nanometer size andmonolayer height (Fig. 11a). The tunnelingspectroscopy results recorded on these islandsshow the semiconducting character and peaks at~1 eV below the Fermi level (SS) on the nor-malized tunneling conductance dI/dV curves(Fig. 11b). This curve resembles the curveobtained for TiO2 where surface states (SS) werecreated due to the formation of Ti2O3. Thus, thecurves in Fig. 11(b) confirm the nucleation ofTi2O3 in a low-oxygen-pressure environmentand clearly show that oxidation of TiAl beginswith the formation of Ti2O3. Surface morphol-ogy of the TiAl alloy after exposure to a high-(atmospheric) oxygen-potential environment isshown in Fig. 12(a). The current/potential (I/V)curve recorded on this surface (Fig. 12b) revealsasymmetric shape and shows that the tunnelingcurrent is higher for the positive polarization ofthe sample than for the negative voltage of thesame value. Furthermore, a well-defined sup-pression of the tunneling current, that is, pres-ence of the I(V) � 0 region in the �1.5 to 0.8 eVenergy range, is observed. This type of asym-metry and suppression of the tunneling currentcan be explained by the presence of TiO2 mate-rial on the surface as amorphous TiO2 and can beregarded as a 3 eV band-gap binary oxide (sup-pression of the tunneling current), with the Fermilevel shifted toward the valence band (asym-metry of the tunneling current) (Ref 71). Suchfundamental information is essential to studyingthe initial stages of scale growth and explainingthe rapid diffusion observed during the onset ofoxide nucleation.

Exposure of Ti3Al alloys to oxygen at hightemperature leads to both oxidation and dis-

solution in the alloy. High oxidation resistancewould be expected if a protective Al2O3

layer could be formed by selective oxidation.However, Al2O3 is only slightly more stable thanTiO, and the activity of titanium is much higherthan that of aluminum in Ti3Al (Ref 72). Thus,TiO is the stable oxide in contact with Ti3Al, andfurther oxidation leads to nonprotective forma-tion of rutile, TiO2. The oxidation behavior iscomplex, leading to a layered oxide scale struc-ture with TiO2 on the outside and oxides withhigher metal content underside, including Al2O3

(Ref 73). Besides scale formation at the surface,oxygen diffuses into Ti3Al as a solute, becauseTi3Al has a comparatively high solubility foroxygen. This leads to embrittlement, increasedstrength, decreased ductility, and crack forma-tion at the surface (Ref 74).Molybdenum, tantalum, and niobium increase

the oxidation resistance to such an extent that the

0 10 20 30 40 50

Aluminum, at.%

1670 °C

(βTi)

Ti3Al

TiAl3

TiAl

L

αTiAl3TiAl2

δ

665 °C660.452 °C

882 °C

(αTi)

Ti Al

0500

600

700

800

900

1000

~11251100

1200

1300

1400

1500

1600

1700

1800

10 20 30 40

Aluminum, wt%

Tem

pera

ture

, °C

Sup

erla

ttice

str

uctu

res

50 60 70

(Al)

80 90 100

60 70 80 90 100

~1285

Fig. 10 The binary titanium-aluminum phase diagram

–2.5 –2.0 –1.5 –1.0 –0.5 0.0

Bias, V

(a)

(b)

(dl/

dV)/

(l/V

), a

rb.u

nits

0.5 1.0 1.5 2.0 2.5

SS

TiAl

d

TiO2 (110)-(1×1)

Fig. 11 (a) 175 nm · 175 nm scanning tunnelingmicroscopy topography and (b) (dI/dV)/(I/V)

curves for TiAl after exposure to environment with 100Langmuir oxygen (1 Langmuir is 10�6 torr . s) SS, surfacestate. Source: Ref 71




obtained rate constants are intermediate betweenthose of TiO2 and Al2O3 formers (Ref 75). Theobtained oxidation resistances remain too lowand limit the application of Ti3Al-base alloys athigh temperature. Therefore, coating of thesealloys is necessary.The oxidation resistance of c-TiAl is higher

than a2 due to the higher aluminum content butstill orders of magnitude lower than that of Al2O3

formers. The oxidation resistance of TiAl relieson the formation of a protective Al2O3 layer thatis only slightlymore stable than TiO, with furtheroxidation producing nonprotective rutile. Thus,oxidation is a problem for TiAl alloys that havebeen optimized with respect to mechanicalbehavior.The majority of the studies on oxidation

behavior of TiAl intermetallics involved ex-periments in air. In general, it is found that fasteroxidation kinetics are promoted in air than inoxygen (Ref 76–81). In Ref 82, the oxidationbehavior of Ti-52Al (at.%) at 900 �C (1650 �F)in various atmospheres, including air and pure

oxygen gas and also in oxygen mixed with dif-ferent (2, 10, 90, 98, and 100%) nitrogen-con-taining atmospheres, was studied, and it wasconcluded that while a protective alumina scalewas developed in pure O2, the nitrogen-con-taining atmospheres do not promote the forma-tion of a continuous alumina scale; instead,intermixed TiO2/Al2O3 nodules are detected.The density of these nodules has been found toincrease with increasing nitrogen concentrationin the reactive atmosphere. Reference 83attributes the fast oxidation of TiAl in air tonitridation. The mechanism by which nitrogenadversely affects the oxidation behavior is stillnot yet satisfactorily understood. In contrast, theoxidation results of titanium-aluminum in Ref 84in air and Ar/O2 atmospheres show beneficialeffects of nitrogen for Ti-48Al-5Nb at 900 �C(1650 �F). The enhancing effect of nitrogen wasattributed to the elimination of internal oxidationof the niobium-containing alloy, which has beenfound to occur during exposure in Ar-20%O2.However, in the same study at 900 �C (1650 �F),the detrimental effect of nitrogen was alsoobserved for Ti-50Al.Using commercial gases (O2, Ar-21%O2, Ar-

1%O2, He-1%O2, and argon containing impu-rities such as CO2, N2, H2O, and H2), the influ-ence of the partial pressure of oxygen on theoxidation behavior of c-TiAl was studied. Theresults show the lowest oxidation at 1000 �C(1830 �F) in pure O2 and then increasing withdecreasing oxygen partial pressure (Ref 85). Theobserved oxidation kinetic results in argonshowing the highest mass gain at the sameexposure temperature can be attributed to thepresence of other oxidants present in the argon,such as CO2, H2, and H2O. In fact, the rapidincrease of the weight gain of the experimentalmaterial in the argon atmosphere probably stemsfrom the presence of water vapor (H2O). Suchresults are not consistent with those obtainedfrom other studies on the effects of the partialpressures of oxygen (Ref 86). The oxidation rateof TiAl-V in pure oxygen has been found to behigher than in Ar-1%O2 at 900 �C (1650 �F).The investigation of transport processes indi-cates the formation of protective Al2O3 increas-ing with the increase of oxygen partial pressurein the reactive atmosphere.It is possible to predict the alloying element

concentration, the temperature, and the oxygenpartial pressure of the atmosphere in whichpreferential oxidation of the alloying elementcan initially occur, provided that appropriatethermodynamic data are available. An oxidation

mechanism has been proposed by the authors(Ref 87) for the oxidation of an intermetallicalloy, Ti-46.7Al-1.9W-0.5Si, in air at 750 to950 �C (1380 to 1740 �F). The minimum activ-ities of titanium and aluminum required to formthe relevant TiO2 and Al2O3 can be calculated ifthe oxygen partial pressure (0.21 atm) andexperimental temperatures (750, 850, and950 �C, or 1380, 1560, and 1740 �F) are known.The free energies of formation (joule/mole) forTiO2 and Al2O3 (Ref 88) are given by:

DG�T,TiO2

=7910,000 + 173T (Eq 9)

DG�T,Al2O3

=71,676,000 + 320 T (Eq 10)

where T is the experimental temperature inKelvin. The calculated results of minimumactivities of aluminum and titanium to formAl2O3 and TiO2 are listed in Table 5. Detailedcalculations for the titanium-aluminum binarysystem (Ref 89) show that the activity of titaniumis slightly higher than the aluminum activity inthe Ti-46.7Al-1.9W-0.5Si alloy. It appears thatthe addition of tungsten and silicon does notsignificantly alter the situation concerning theactivity of titanium and aluminum. Thus, thepreferential formation of TiO2 is predicted dur-ing exposure to both environments. The dis-continuous nature of the TiO2 layer at the earlystages of oxidation, observed in this study, maybe attributed to the two-phase nature of themicrostructure of the alloy, which implies thatthe activities of titanium and aluminum varieddepending on the location of the phases on thealloy surface, and it was difficult for TiO2 to formin certain areas.During the initial exposure of the Ti-46.7Al-

1.9W-0.5Si alloy to air at the experimentaltemperatures, the high oxygen partial pressurepromotes the formation of TiO2, according to thereaction:

TiðsÞ þ O2ðgÞ ¼ TiO2ðsÞ (Eq 11)

The formation of TiO2 changes the balance of theactivities of titanium and aluminum and theoxygen partial pressure between the TiO2 layerand the substrate. The reduction of titaniumactivity then leads to the development of anAl2O3 layer:

2Al(s)+3/2O2(g)=Al2O3(s) (Eq 12)

This development of an Al2O3 layer in turn leadsto the formation of a titanium-enriched zone

(a)

–2.5

–0.1

0.0

0.1

0.2

0.3

0.4

0.6

(b)

0.5

–2.0 –1.5 –1.0 –0.5 0.0

Bias voltage, V

I/V

, nA

/V

0.5 1.0 1.5 2.0 2.5

Fig. 12 (a) 300 nm · 300 nm scanning tunnelingmicroscopy topography and (b) I/V curve for

oxidized TiAl after exposure to atmospheric environment.Source: Ref 71

Table 5 The minimum activities of titanium and aluminum to form TiO2 and Al2O3

0.21 atm of oxygen partial pressure at temperatures given

Metal Symbol

Activity, minimum

750 �C (1380 �F) 850 �C (1560 �F) 950 �C (1740 �F)

Aluminum aAl 1.19 · 10�34 7.72 · 10�31 1.18 · 10�27

Titanium aTi 1.77 · 10�37 2.44 · 10�33 7.05 · 10�30




beneath the Al2O3 layer. At the same time,nitrogen migrates through the Al2O3 layer tothis titanium-enriched zone from the externalatmosphere, which creates a favorable circum-stance for the following reaction to take place:

Ti(s)+1/2N2(g)=TiN(s) (Eq 13)

Thus, a TiN layer is formed beneath the Al2O3

layer. In the meantime, oxygen species alsodiffuse inward to the interface between theAl2O3 layer and the TiN layer, allowing abuildup of the oxygen partial pressure. Whenthe oxygen partial pressure reaches a certainlevel, TiN became unstable, favoring thefollowing reaction:

TiN(s)+O2(g)=TiO2(s)+1/2N2(g) (Eq 14)

The released nitrogen migrates inward via theTiN layer. The nitrogen partial pressure gradu-ally increases at the TiN/substrate interface, andthe nitrogen species, encountering titanium fromthe substrate, allows the reaction (Eq 13) to takeplace again. This gives rise to the formation of atitanium-depleted zone, demonstrated by theexistence of a TiAl2 band beneath the TiN layer.The oxidation mechanisms of the Ti-46.7Al-1.9W-0.5Si alloy in air are schematicallydescribed in Fig. 13. It is apparent that thethickness of the TiO2 layer increases withexposure time, as does the thickness of the TiNlayer as nitrogen migrates inward from theexternal environment. However, it is not clearwhy AlN does not develop between the TiN andTiAl2 band, because the affinities of aluminumand titanium to nitrogen are very close (Ref 88).This can probably be attributed to the faster self-diffusion of titanium in the TiAl substrate thanthat of aluminum (Ref 90); TiN became thekinetically favored product.Sulfidation. Information on the high-tem-

perature sulfidation behavior of titanium-alumi-num intermetallics is not extensive. Some workhas been reported on the response of titanium-aluminum-base materials, such as Ti-54Al andTi-48Al-2Nb-2Mn, after exposure at 900 �C(1650 �F) to mixed-gas environments with afixed pO2

(10�14 Pa) and with pS2 values of0.0016, 0.11, and 1.6 Pa (Ref 91). The resultsfrom this study indicate parabolic kinetics at allpS2 values for sulfidation of the alloys. Aftersome uncertainties at the early stages of sulfi-dation, prolonged exposure (168 h at 900 �C, or1650 �F) indicates that the greatest resistance tosulfidation occurs at the highest pS2 (Fig. 14, 15).In the initial stages of exposure, the high

affinity of oxygen for titanium and aluminumleads to development of an outer layer of TiO2,beneath which an Al2O3 layer forms. Sulfurdiffuses through the TiO2 and Al2O3 layers andreaches the substrate/scale interface, where thepO2

is low enough to promote the formation ofTiS2 and Al2S3 (and NbS2 in the case of Ti-48Al-2Nb-2Mn). Clearly, the presence of high pS2 in

the bulk environment provides a higher drivingforce for sulfur migration that then supports ahigher sulfur flux and favors the formation ofsulfides of alloy constituents. Accordingly, amixed layer of sulfides and oxides develops inthe high-pS2 atmosphere, thicker than in the low-pS2 environment, thereby providing higherresistance to sulfidation.

In summary, the superior sulfidation resis-tance of Ti-54Al and Ti-48Al-2Nb-2Mn can beascribed to several aspects of the scaling pro-cesses. First, the development of an inner layer ofsulfides (TiS2, Al2S3, NbS2) provides an effec-tive barrier to cation transport. The presence ofrefractory metal sulfide, NbS2, is known to cause

significant improvement in the sulfidation resis-tance of the intermetallic alloy (Ref 92). Fur-thermore, sulfidation resistance of Ti-54Al stemsfrom the formation of a TiAl3 layer at the scale/substrate interface by the diffusion of titaniumafter the development of TiO2 and TiS2. Tita-nium released from the dissociation of TiS2 islikely to promote a thick TiAl3 layer, thusimparting further resistance to sulfidation.The oxidation/sulfidation behavior of another

intermetallic alloy, Ti-56.6Al-1.4Mn-2Mo, withduplex and lamellar microstructure has beeninvestigated by the authors (Ref 93) in H2/H2S/H2O (pS2=0:1 Pa; pO2

=10712 Pa) at 750 �C(1380 �F) and (pS2=0:1 Pa; pO2

=10714 Pa)

TiO2

TiO2 TiO2

TiO2

Substrate

(a)

(b) (d)

(c)

Substrate

SubstrateSubstrate

AI2O3

TiNAI-rich zone

AI2O3

Ti-rich zone

AI2O3

TiO2

TiN

TiAI2 zone

Fig. 13 Schematic illustration of the mechanism controlling oxidation of Ti-46.7Al-1.9W-0.5Si in air between 750and 950 �C (1380 and 1740 �F). (a) Islands of TiO2 form. (b) Al2O3 layer and titanium-rich zone form, and

TiO2 grows. (c) TiN layer develops in titanium-rich zone, and aluminum-rich zone forms. (d) TiN layer oxidizes and shiftsinward, and TiO2 and Al2O3 layers and TiAl2 zone grow. Source: Ref 87

20

18

16

14

12

10

8

Wei

ght g

ain,

mg/

cm2

6

4

2

0

0 10 20 30 40

Time, h

50 60 70 80

Fig. 14 Sulfidation kinetics for Ti-54Al at 900 �C (1650 �F) for three mixed gases. Partial pressure (pO2), 10�14 Pa.

Partial pressures (pS2 ): diamond, 0.0016 Pa; square, 0.11 Pa; triangle, 1.6 Pa. Source: Ref 91




at 900 �C (1650 �F). The results of weight-gain/time data show parabolic kinetics(kp=10�12 g2/cm�4/s) at 750 �C (1380 �F) andcubic kinetics at 900 �C (1650 �F) and theabsence of significant effect on the corrosionbehavior of increased exposure temperature,with scaling pattern development resemblingthat of the Ti-Al-Nb-Mn alloy, that is, withan outermost TiO2/inner Al2O3 layer and base-element sulfides forming between the oxidelayers and the substrate.The enhanced corrosion resistance observed at

900 �C (1650 �F) compared to that at 750 �C(1380 �F) follows from the differences in thedefect structures. Both oxygen vacancies andinterstitial titanium ions are important pointdefects (Ref 8), with interstitial titanium ionspredominating at low pO2

, and high temperature,whereas oxygen vacancies predominate at highpO2

and low temperature. At low pO2, many

interstitial titanium ions are expected in TiO2,and hence, an increase in pO2

would result in adecreased number of interstitial ions. Reducingthe interstitial titanium ions would decrease thetransport of titanium through TiO2. The tem-perature change between 750 and 900 �C (1380and 1650 �F) may not significantly alter thediffusion of titanium and oxygen ions in thesealloys. The increased oxygen pressure willincrease the oxygen vacancies in TiO2 andthereby increase the inward diffusion of oxygenand promote Al2O3, imparting a slow corrosionrate for both materials.Further work by the authors (Ref 94) on the

oxidation and sulfidation behavior of Ti-46.7Al-1.9W-0.5Si alloy in an H2/H2S/H2O atmosphereyielding high sulfur (pS2 ~ 1:2 · 1071 Pa) andlow oxygen (pO2

~ 1:2 · 10715 Pa) potentials at850 �C (1560 �F) shows protective kinetics witha parabolic rate constant of 6 · 10�11 g2/cm4/s.Morphological analysis reveals the developmentof a multilayered scale consisting of a top rutile(TiO2) layer, a continuous layer of a-Al2O3

beneath the rutile layer, and a TiS layer con-taining scattered pure tungsten particles. It issuggested that fast outward diffusion of titaniumwithin the substrate results in the formation of azone of high concentration of aluminum (TiAl3and TiAl2) between the scale and substrate, asillustrated in Fig. 16.Work by the same group (Ref 95) has been

successful in increasing the sulfidation resistanceof TiAl by the deposition of high-aluminum-concentration titanium aluminide (e.g., TiAl3).The oxidation/sulfidation kinetics data ofuncoated and TiAl3-coated Ti-46.7Al-1.9W-0.5Si in a H2/H2S/H2O atmosphere at 850 �C(1560 �F) show increased high-temperaturecorrosion resistance of the Ti-46.7Al-1.9W-0.5Si alloy by the TiAl3 coating. Both coated anduncoated samples underwent oxidation/sulfida-tion following a parabolic rate lawwith parabolicrate constants of 2.38 · 10�13 g2/cm4/s and7.36 · 10�11 g2/cm4/s, respectively. It is indi-cated that the deposition of the TiAl3 coatingincreases the corrosion resistance of Ti-46.7Al-1.9W-0.5Si by more than 2 orders of magnitude.

Hot corrosion behavior of TiAl-base mate-rials is discussed by considering the response ofthree titanium-aluminum intermetallics—Ti-48Al, Ti-48Al-2Cr, and Ti-52Al—to exposure to80 wt% Na2SO4þ 20 wt% NaCl-containingenvironments at 800 �C (1470 �F). Hot corro-sion and scale spallation took place during the200 h tests. Ti-52Al showed the best corrosionresistance among the three alloys. It is believedthat the high aluminum content and single c-TiAlphase played a role, forming an aluminum-richoxide scale with relatively good protectiveability. Ti-48Al-2Cr also showed better corro-sion resistance than Ti-48Al. The latter suffereda high corrosion rate and severe scale spallation.Three different exposure methods were usedto study hot corrosion in salt-containing envi-

ronments. These represent different serviceconditions. Among these methods, suspendingthe specimens in salt vapor appeared to bethe most reliable method for kinetic studies.The influential factors are relatively easy tocontrol. When using the salt-deposition method,the spallation of the salt film at the early stageneeds to be considered. The immersion methodshowed the severest hot corrosion. Cruciblematerials and corrosion products dissolve intothe molten salts, further complicating the reac-tions (Ref 96).Hot corrosion tests (Ref 97) on Ti3Al,

Ti-44Al, Ti-48Al, Ti-52Al, Ti-48Al-2Cr, andTi3Al-11Nb (at.%) intermetallics were con-ducted in a salt mixture of 80 wt% Na2SO4þ 20wt%NaCl with a melting point of approximately

TiO2

Al2O3

TiS+W

TiAl3

TiAl2

Substrate

∼50 µm

Fig. 16 SEM micrograph for exposed Ti-46.7Al-1.9W-0.5Si in H2/H2S/H2O atmosphere at 850 �C (1560 �F) for240 h. Source: Ref 94

4

3.5

3

2.5

2

1.5

Wei

ght g

ain,

mg/

cm2

1

0.5

0

0 10 20 30 40

Time, h

50 60 70 80

Fig. 15 Sulfidation kinetics for Ti-48Al-2Nb-2Mn at 900 �C (1650 �F) for three mixed gases. Partial pressure (pO2),

10�14 Pa. Partial pressure (pS2 ): diamond, 0.0016 Pa; square, 0.11 Pa; triangle, 1.6 Pa. Source: Ref 91




700 �C (1290 �F). The specimens weresuspended in the salt vapor at 800 �C (1470 �F).It was found that the hot corrosion resistancegenerally increased with increasing aluminumcontent. Ti-48Al-2Cr showed the best corrosionresistance due to its relatively high aluminumcontent and chromium addition. Ti3Al-11Nbunexpectedly showed poor corrosion resistance.It is suggested that the niobium addition mayaccelerate the formation of sulfide at the inter-face, playing a detrimental role in the corrosionprocess.The effect of reactively sputtered Al2O3 and

enamel coatings on oxidation and hot corrosionbehaviors of TiAl was investigated at 900 �C(1650 �F). Isothermal oxidation testing indicatedthat both coatings were very effective inreducing the oxidation rate of c-TiAl. Cyclicoxidation testing indicated that reactive-sput-tered Al2O3 coating had little effect in improvingthe cyclic oxidation resistance of TiAl due to thespallation of Al2O3 coating. However, enamelcoating exhibited an excellent positive effect oncyclic oxidation of TiAl. The enamel coatingwas very adherent to the substrate, and neithercracks nor spallation appeared in the coatingsduring cyclic oxidation. Furthermore, the hotcorrosion behavior of TiAl with enamel coatingwas evaluated in (Na,K)2SO4 melts at 900 �C(1650 �F). The results showed that the enamelcoating was very stable in (Na,K)2SO4 melts andwas very effective in protecting TiAl from hotcorrosion attack (Ref 98).The hot corrosion behavior of Ti-50Al, Ti-

48Al-2Cr-2Nb, and Ti-50Al-10Cr alloys wasinvestigated in (Na,K)2SO4 and Na2SO4þNaClmelts (Ref 99). TiAl intermetallics showedmuchbetter hot corrosion resistance in (Na,K)2SO4 at900 �C (1650 �F) than the nickel-base super-alloy K38G. Two types of corrosion productsformed on Ti-50Al: Some areas were coveredwith a continuous Al2O3 scale, whereas otherareas formed a mixed Al2O3þTiO2 scale; TiSexisted at the scale/alloy interface. A mixedAl2O3þTiO2 scale developed on Ti-48Al-2Cr-2Nb, and no sulfide was found beneaththe scale. An adherent Al2O3 scale, however,formed on Ti-50Al-10Cr, which providedexcellent hot corrosion resistance. All threealloys suffered severe hot corrosion in Na2SO4þNaCl melts at 850 �C (1560 �F). A mixedAl2O3þTiO2 scale was generated on all threealloys, and many voids and pits existed in thescale/alloy interface.

Silicides

Silicides form an important class of structuralintermetallics, so their high-temperature corro-sion response is of interest. High-temperaturecapabilities of silicides are much higher thanthose of aluminides, allowing working tem-peratures greater than 1400 �C (2550 �F). Suchoperating temperatures are needed for improvingthe thermal efficiency and reliability of energy-conversion systems and advanced engine sys-

tems (Ref 100). The use of currently availablealloys, such as nickel-base single-crystal super-alloys, is limited to temperatures of ~1100 �C(2010 �F). Superalloys derive their intrinsicstrength by reinforcement with c0-Ni3Al pre-cipitates, but these tend to coarsen and ultimatelydissolve as the temperature increases beyond1100 �C (2010 �F). Aluminide alloys based onNiAl, which are currently under development,have the potential for use up to 1200 �C(2190 �F). The melting temperature (Tm) of amaterial for structural applications at 1400 �C(2550 �F) needs to be 42000 �C (3630 �F), sothat, at most, 0.75 Tm is reached during service,and appreciable high-temperature strength ismaintained. Of the available potential systems,refractory silicides, such asMoSi2 andNbSi2, arewidely recognized as potential structural andcoating materials because of their high meltingpoints, good mechanical strength, high thermaland electrical conductivities, and promisingoxidation resistance at elevated temperature dueto the formation of a self-passivating, glassysilica (SiO2) layer. However, the use of silicidesfor high-temperature applications poses prob-lems. The problems associated with the oxida-tion of silicides are well known, and so far, therehave been limited studies on the sulfidation ofsilicides.

Oxidation. Molybdenum silicides (MoSi2and Mo5Si3) possess excellent oxidation resis-tance at 41000 �C (1830 �F). However, it wasobserved (Ref 101–103) that MoSi2 showedaccelerated oxidation behavior, resulting in so-called “pesting” in the temperature rangebetween 400 and 650 �C (750 and 1200 �F),whereby the MoSi2 disintegrates to powder. Thepest degradation is associated with dissociativeoxidation to form SiO2 and MoO3. The MoO3

vapor pressure begins to increase within the pestdegradation temperature regime such that muchof the MoO3 formed tends to sublime with timeas temperatures increase. The poor oxidationresistance of molybdenum is generally assumedto be due to the instability of its related oxidephase (MoO3). However, niobium appears to beless critically limited in this respect. Nb2O5 isnonvolatile below 1370 �C (2500 �F) (Ref 104);therefore, the latter would be expected to, andindeed does, perform relatively well under oxi-dizing conditions.

A great deal of research has been conductedworldwide to achieve protective oxidation ofmolybdenum silicides by the addition of a thirdelement, such as aluminum, boron, germanium,tungsten, tantalum, titanium, zirconium, oryttrium (Ref 105–108). All of these elementsform oxides that are more stable than SiO2, and itis thought that the effect of their scavenging foroxygen may accelerate the scaling process ofSiO2 and thereby prevent or minimize molyb-denum oxide formation. The addition of theseelements, aluminum in particular, reportedlyreduces pesting by the formation of an amor-phousMo-Si-Al-O phase in the initial cracks andvoids (Ref 107). However, the growth rates forthe oxides of any of these third element additions

(at low concentration) are fairly low and com-parable to that of MoO3 and may not fullyminimize pesting attack for molybdenum sili-cides at low temperature. The protectivity ofNbSi2 in oxidizing environments may be com-promised by the simultaneous formation of SiO2

andNb2O5 due to the high solubility of oxygen inNbSi2 and the close affinities of silicon andniobium for oxygen.Sulfidation. A comprehensive review of the

sulfidation behavior of intermetallics (Ref 100)found the Mo5Si3-type intermetallics in a1.5 vol% H2S/H2 gas mixture and compared thesulfidation resistance with relevant oxidationperformance in air. Figure 17 shows the ther-mogravimetric test data obtained at severaltemperatures for oxidation and sulfidation of aboron-containing Mo5Si3. The data imply aprotective scaling of the alloy at 500 �C (930 �F)under oxidizing conditions, not because of SiO2

formation, but because the volatilization rate ofmolybdenum oxide is negligible. At tempera-tures of 800 and 1200 �C (1470 and 2190 �F),the curves demonstrate a sharp drop in specimenweight for ~2 h, after which a plateau is reachedand the weight changes little during 50 h ofadditional exposure, indicating a protectivescale. The morphology of the scale after oxida-tion at 800 �C (1470 �F) has been observed toconsist of a light-colored MoO2 phase and a darkgray silicon-rich oxide; the scale after oxidationat 1200 �C (2190 �F) consists predominantly ofsilicon-rich oxide and almost pure molybdenumparticles. The surface layer shows significantcracking and peeling and appears to remainhighly plastic, as indicated by curling ratherthan spalling of the oxide layer. Such degrada-tion of the oxide layer can expose interiormolybdenum silicide to additional oxidation,and the sequential processes of oxidation andpeeling can continue without offering oxidationprotection for the alloy over long periods ofexposure. The thermogravimetric test dataobtained during sulfidation of the material revealan approximately one order ofmagnitude smallerdecline in specimen weight than in the dataobtained during oxidation. It is not clear as to thecause for the initial drop in weight, except thatthe possible presence of residual moisture in thegas mixture can lead, in the early stages ofexposure, to the formation of volatile oxides suchas MoO3 and/or SiO (especially in the reducingcondition used in the experiments) before thedevelopment of sulfide reaction products.Figure 18 provides scanning electron micro-graphs of surfaces of boron-containing Mo5Si3alloy after sulfidation at 500, 800 and 1100 �C(930, 1470, and 2010 �F) in a 1.5 vol% H2S/H2

gas mixture. After exposure at 500 �C (930 �F),the specimen surface shows isolated regions ofmolybdenum sulfide but no gross oxidation.After 800 �C (1470 �F) exposure, the specimenexhibits a greater coverage of the surface withmolybdenum sulfide, whereas after exposure at1100 �C (2010 �F), it shows almost completecoverage by molybdenum sulfide. These pre-liminary results indicate that the intermetallic




material based on molybdenum silicide candevelop protective sulfide scales during servicein reducing environments at elevated tempera-tures.Recently, the sulfidation behavior of CrSi2,

WSi2, NbSi2, Nb5Si3, MoSi2, and Mo5Si3 for upto 240 h in a H2/H2S/H2O environment thatyields an oxygen potential of 1.2 · 10�15 Pa anda sulfur potential of ~6.8 · 10�1 Pa at 850 �C(1560 �F) was investigated (Ref 109). The sul-fidation kinetics of all silicides (Fig. 19) showthat all performed extremely well in thisaggressive atmosphere. CrSi2 shows the leastsulfidation resistance, which is consistent withpoor sulfidation resistance of chromium (Ref11). On the other hand, the silicon content insilicides also plays an important role to resistsulfidation attack; the high silicon concentrationsignificantly increases the sulfidation resistance.For example, the sulfidation rates of MoSi2 andNbSi2 are 1 order of magnitude slower thanMo5Si3 and Nb5Si3, respectively, as shown inTable 6.A protective SiO2 scale observed on all

exposed silicides is responsible for the preven-tion of further severe environmental attack, aspresented in Fig. 20. However, Cr2S3 and NbS2

were also formed on the exposed CrSi2 andNb5Si3, indicated by XRD and electron dis-persive x-ray data. It is interesting to note thatonly nodular Cr3S3 is formed on exposed CrSi2,and a continuous layer of Cr2S3 failed to formeven after 240 h exposure. Cross-sectionalscanning electron microscopy observationshowed that the Cr2S3 nodules virtuallydestroyed the underlying SiO2. In contrast, acontinuous and uniform NbS2 layer developed.The SiO2 layer still existed and effectively pro-tected the substrate.Hot Corrosion. Only very limited research

on hot corrosion of silicides has been conducted.In the mid-1950s, it was shown that silicon-containing high-alloy steels had the best resis-tance against V2O5 melts at 925 �C (1700 �F);silicon and chromium appear to be the mostpromising combination of elements for the besthot corrosion resistance of heat-resistant alloys(Ref 110). It also is reported (Ref 111–113) thatsilicon-containing overlay coatings on nickel-and iron-base alloys show improved resistanceagainst high-temperature oxidation and hot cor-rosion in the temperature range of 700 to1000 �C (1290 to 1830 �F). Siliconized turbineblades coated by chemical vapor deposition

processes as well as plasma spray tested (Ref114–116) at 710 �C (1310 �F) for 30,000 h in aburner gas containing 0.4% S, 15 ppm sodium,and 5 ppm vanadium, reveal the least attack.These materials have been demonstrated to beextremely resistant against hot corrosion andshow promise of higher lifetimes for turbineblades.Silicon-containing coatings with a matrix

consisting of the saturated c solid-solutionNi6Cr2Si2 and that also contain intermetallics ofcomplex tantalum silicides as precipitates showgreat promise as protective layers for nickel-basealloys, with improved oxidation and hot corro-sion resistance from the hot corrosion tests at1000 �C (1830 �F) for up to 4000 h in burner gaswith intermediate dipping in Na2SO4 and V2O5

melts. From silicon diffusion data, lifetimesgreater than 10,000 h at 1000 �C (1830 �F) canbe realized (Ref 117).However, it is reported that there is evidence

that silica or silicon-rich scales may be attackedunder basic fluxing conditions. MoSi2 has beenshown to be very resistant to attack by NaCland V2O5 but does dissolve in Na2CO3 andNa2SO4 (Ref 118). Burner rig test specimensmade of hot-pressed MoSi2, heated for 4 min

0.05

500 °C

800 °C

1200 °C

500 °C

800 °C

1000 °C

0

–0.15

–0.1Wei

ght c

hang

e, m

g/cm

2

Wei

ght c

hang

e, m

g/cm

2

–0.15

1

0

0.5

–0.5

–1

–1.5

–20 50 100

Exposure time, h

150 2000 50 100

Exposure time, h(a) (b)

150 200

Fig. 17 Weight-change data during (a) oxidation in air and (b) sulfidation in 1.5 vol% H2S/H2 gas mixture for boron-containing Mo5Si3 after exposure at several temperatures.Source: Ref 100

Fig. 18 SEM micrographs of surface of boron-containing Mo5Si3 after sulfidation in 1.5 vol% H2/H2S gas mixture at (a) 500 �C (930 �F), (b) 800 �C (1470 �F), and (c) 1100 �C(2010 �F). Source: Ref 100




intervals to 1000 to 1200 �C (1830 to 2190 �F)and then sprayed with 1 to 10% salt solutionsof NaCl, Na2CO3, and Na2SO4, remain relativelyunaffected after 20 h. The addition of acidicspecies, such as HCl, to the burner rig gas hasbeen shown to significantly reduce dissolution inalkaline salts.

Other Intermetallics

The following intermetallics show interestingresponse to oxidation and may be useful ascoating materials.Noble-Metal-Containing Intermetallics:

Platinum-Aluminum, Iridium-Aluminum,and Ruthenium-Aluminum Systems. Platinum-aluminum intermetallics favor the exclusiveformation of Al2O3 and show stoichiometry-dependent oxidation kinetics and scale mor-phology; the overall growth rate is extremelysmall at 1450 �C (2640 �F) (Ref 119, 120).Pt2Al, Pt5Al3, and PtAl2 all display ridged oxidemorphology.Iridium-aluminum intermetallics require ~55%

Al to form a protective Al2O3 scale at 1300 to1800 �C (2370 to 3270 �F) (Ref 121, 122).Ruthenium-aluminum intermetallics and their

alloys containing chromium and ytttriumdevelop a protective Al2O3 scale at 1250 �C(2280 �F); high oxide vapor pressures are alimiting factor (Ref 123, 124). CrAl and Cr2Alform a protective Al2O3 layer at 1000 to 1200

�C(1830 to 2190 �F) (Ref 125), with yttrium addi-tion improving the scale adhesion. Aluminum-deficient Cr2Al develops a chromium-rich scale,allowing N2 ingress and leading to the formationof a nonprotective nitride-containing subscale.Beryllide intermetallics with low densities

and melting temperatures owe their good high-temperature oxidation resistance to the forma-tion of a stable compact BeO scale (Ref 126,127). In water-vapor-containing atmosphere, theprotectiveness is lost due to the formation ofvolatile Be(OH). The lack of ductility and tox-icity of BeO are additional problems.

Aqueous Corrosion

High-temperature structural applications arethe main driving forces for the development ofintermetallic compounds, and corrosion researchhas focused on degradation in the high-temperature gases of these environments. How-ever, low-temperature aqueous corrosion ofintermetallic compounds is of interest for fourreasons. First, high-temperature materials willnot always be at high-temperatures. Damagecaused by low-temperature corrosion duringfabrication or shutdown could result in cata-strophic failures in service. This includes expo-sure to maintenance fluids or fire-extinguishingcompounds as well as water or humid air andcan be of particular concern for compoundssusceptible to hydrogen-induced cracking inhumid air. Second, there may be low- or inter-mediate-temperature applications where inter-metallic compounds have superior propertiesto the alloys currently used. In many cases,intermetallic compounds will have lowerdensities and higher strengths than competingalloys and may have better resistance to fatigue,wear, corrosion, or a combination of these.Third, there are many fabrication processes, heattreatments, and joining methods that result in theformation of intermetallic compounds. Becauseheterogeneous nucleation is preferred, thesephases tend to nucleate and grow at criticallocations, such as grain boundaries and inter-phase bondings, and may limit performance ifthey have poor corrosion resistance. Fourth,because intermetallic compounds are frequentlyvery hard or brittle at ambient temperatures,understanding corrosion could enable the crea-tion of inexpensive electrochemical fabrica-tion methods. Electropolishing, acid saw cutting,and electrochemical machining are examplesof fabrication processes that use corrosionreactions to produce finished parts or surfacefinishes.

Aqueous corrosion of intermetallic com-pounds is influenced by the same factors thatdetermine the corrosion performance of normal

metallic alloys. In addition, there is a vast num-ber of intermetallic compounds and environ-ments that may cause corrosion of thesecompounds, most of which have not been stu-died. However, aqueous corrosion research hasbeen conducted on a number of intermetalliccompounds for low-temperature structuralapplications, but most of this research has beenconducted on nickel, iron, and titanium alumi-nides. More detailed reviews of this research canbe found in the literature (Ref 128–132). Someresearch has been conducted on other inter-metallic phases that form during heat treatmentor joining of metals, but this section does notattempt to cover this literature (Ref 133, 134).Here, the fundamental factors that make the

1.2

1

0.8

0.6

0.4

Wei

ght c

hang

e, m

g/c

m2

0.2

00 50

Mo5Si3

MoSi2

WSi2

NbSi2

Nb5Si3

CrSi2

100 150

Exposure time, h

200 250

Fig. 19 Oxidation and sulfidation of silicides at 850 �C (1560 �F). Source: Ref 109

Fig. 20 SEMmicrographs of exposed CrSi2 and Nb5Si3in H2/H2S/H2O environment at 850 �C

(1560 �F). (a) CrSi2. (b) Nb5Si3. Source: Ref 109

Table 6 Parabolic rate constants (kp) forsulfidation of silicides at 850 �C (1560 �F)Material kp, g

2/cm4/s

CrSi2 1.1 · 10�12

WSi2 2.7 · 10�14

NbSi2 8.7 · 10�14

Nb5Si3 5.1 · 10�13

MoSi2 2.4 · 10�14

Mo5Si3 1.0 · 10�13




aqueous corrosion of an intermetallic phase dif-ferent from that of a homogeneous alloy or of theconstituents in pure elemental form are empha-sized.

Basic Considerations

When a homogeneous solid solution of two ormore species order on a lattice to form an inter-metallic compound, three factors are altered thatmay influence the corrosion rates and stress-corrosion cracking behavior of the resultingphase:

� Chemical activity (thermodynamics)� Physical arrangement of atoms (ordering-

influenced kinetics)� Deformation (slip character, stress-corrosion

cracking, and hydrogen embrittlement)

Thermodynamic Considerations. When twoor more elements are mixed to form a solidsolution, the chemical activity of the species inthis solution differs from that of the same speciesin pure elemental form. The literature onmetallurgical thermodynamics has measuredthese activities and developed methods for esti-mating them and calculating phase diagrams.Currently, commercial software packages areavailable for making these calculations. If theenergy of dissimilar interatomic bonds is lower,on average, than similar bonds, then the solidsolution will tend to lower its energy bybecoming ordered, with specific species on spe-cific lattice sites (typically, when the elementshave significantly different electronegativities).Of course, this ordering reaction also altersthe chemical activity of the various species

in the intermetallic phase. Taking these ther-modynamic changes into consideration andremembering that when etching metallurgicalstructures, the intermetallic phases becomeobservable due to a difference in the dissolutionrates, it is not uncommon for investigators toassume that intermetallic compounds will havesignificantly better corrosion properties thansimilar alloys. While this may be true in somecases, it should be kept in mind that most cor-rosion-resistant engineering alloys rely on pas-sivity for resistance to corrosion and not nobility.That is, they usually have a large overpotential(excess free energy) driving corrosion and relyon the formation of a thin, continuous, and pro-tective film of corrosion products to protect theunderlying microstructure from corrosion. Infact, a large overpotential can be beneficialbecause it may speed the reformation and sealingof the passive film when it becomes ruptured bychemical or mechanical means. Fortunately,while the energy of ordering is sufficient to drivethe ordering reaction at heat treating tempera-tures, the change in the chemical potentials ofthe species during ordering is usually smallcompared to the free-energy changes that drivecorrosion reactions. During ordering, thermo-dynamic changes occur and alter the chemicalpotentials of the species, but the magnitude ofthese changes is usually insufficient to producefundamentally different corrosion behavior(such as the transition from active to noble orpassive).While this may seem disappointing, it is

actually a very important deduction in terms ofmaking first-order approximations of the corro-sion behavior of intermetallic compounds in

aqueous environments. First-order approxima-tions of corrosion behavior are usually donethrough the use of electrochemical equilibriumdiagrams of the type first proposed and madepopular by Pourbaix and co-workers (Ref 135).These diagrams are essentially plots that identifythe lowest energy phase that can form betweenan element and water as a function of electrodepotential and pH. See the article “Potential ver-sus pH (Pourbaix) Diagrams” inASMHandbook,Volume 13A, 2003. Because the electrodepotential is determined by the free energy for thechemical reactions between the elemental spe-cies and water, one should calculate a new dia-gram for each element in the intermetallic phasethat takes into account chemical potential chang-es and the possibility of forming corrosionproduct phases that incorporate two or more ofthe species in the intermetallic phase (forexample, NiAl2O4 for NiAl or Ni3Al). Figure 21provides examples of electrochemical equili-brium diagrams calculated for copper and tin thatshow the changes in the equilibrium boundariesfor the various intermetallic phases that can formin the copper-tin system (Ref 136). By examin-ing these diagrams, it can be seen that the mag-nitude of the changes in the equilibriumboundary for the various phases is of the order of10 to 100 mV,which could be a small percentageof the corrosion driving force. Therefore, whilecalculation of diagrams is preferred, first-orderapproximations of behavior for monolithicstructural applications can be made from theelectrochemical equilibrium diagrams of thepure elemental species, which are readily avail-able in published atlases (Ref 135). On the otherhand, identifying phase-etching behavior in a

+1.0

Cu ++

[10–6]

Cu(OH)2

HCuO–2

CuO–2

O2Cu2O

[10–6]

Cu lmmune

H2Pure Cu (0 at.% Sn)

Cu-Cu3Sn (0.7 to 25.0 at.% Sn)

Cu3Sn-Cu6Sn5 (25.0 to 45.4 at.% Sn)

Cu6Sn5-Sn (45.4 to 99.99 at.% Sn)

+0.5

0.0

–0.5

–1.0

Pot

entia

l to

hydr

ogen

ele

ctro

de, V

–1.5–2 0 2 4 6 8

pH

10 12 14 16

O2

H2

HSnO–2

[10–6][10–6]

Pure Sn (0 at.%Cu)

Cu6Sn5-Cu3Sn (54.6 to 75.0 at.% Cu)

Sn-Cu6Sn5 (0 to 54.6 at.% Cu)

Cu3Sn-Sn (75.0 to 99.2 at.% Cu)

Pot

entia

l to

hydr

ogen

ele

ctro

de, V

+1.0

+0.5 Sn+4

Sn+2

Sn(OH)4

Sn(OH)2SnO–

30.0

–0.5Sn Immune

–1.0

–1.5–2 0 2 4 6 8

pH

10 12 14 16

Fig. 21 Electrochemical equilibrium diagrams for copper and tin comparing equilibria for pure elemental form to that in various intermetallic compounds. Lines are drawn for thechemical potentials for equilibrium between two phases. The equilibria between ions and insoluble oxides or hydroxides are unaltered by substrate ordering.




microstructure containing different intermetallicphases or the phase that corrodes first in a weld orsolder joint contaminated with flux residue maybenefit greatly from calculations of this type (Ref133, 134). Figure 22 shows the equilibrium dia-grams for nickel and aluminum; the shift in theequilibrium lines to correct for the chemicalpotentials of nickel and aluminum in nickelaluminide are added as dashed lines (Ref 137).By examining these diagrams, it can be seen thatthe equilibrium line for nickel has shifted so littlethat the change is not resolvable as a separate linefrom the pure nickel line. On the other hand, thealuminum equilibrium line has shifted dramati-cally toward more noble potentials, but it is stillsignificantly active compared to either thehydrogen or oxygen reduction reactions, so thatboth will still occur spontaneously and causecorrosion or passivity. Because corrosion reac-tions are usually mass transport rate limited atthese large overpotentials (driving forces), thesechanges in the driving force probably produce nosignificant changes in corrosion behavior underservice conditions. However, the kinetics ofdissolution or passivation could be influenced bythe imposition of particular species as nearestneighbors in the ordered structure.Ordering-Influenced Kinetics. While the

influence of ordering on thermodynamics is wellknown and can be calculated, the influence ofordering on the kinetics of dissolution and pas-sivation is less well understood. In the case of abinary solid-solution alloy with a completelyrandom distribution of the two different specieson the lattice sites, one may assume that the moreactive species will tend to be selectively removed

from the surface, leaving a surface enriched withthe more noble species. In some cases, the sur-face layer becomes porous, which enables theremoval of the active species to much greaterdepths into the alloy, leaving behind a dealloyedlayer that is a mechanically weak spongelikenetwork of pores. Dealloying corrosion, as this isknown, is covered in the article “Effects ofMetallurgical Variables on Dealloying Corro-sion” in ASM Handbook, Volume 13A, 2003.However, dealloying typically arises above aspecific concentration known as the parting limitand occurs most commonly in alloys of two ormore species with significantly different elec-tronegativities (gold-copper, copper-zinc) (Ref138, 139). Based on the assumption that theparting limit is the atomic concentration whererandom distribution allows for the formation of acontinuous percolation path of the active speciesthrough the solid, it is estimated that partingshould occur at concentrations above approxi-mately 18 mol fraction, which corresponds withthe experimental results (Ref 140). Intermetalliccompounds frequently exceed these concentra-tion and electronegativity difference require-ments without exhibiting dealloying (decom-position and dealloying), although dealloyingdoes appear to occur in some cases, such as thecopper-bearing grain-boundary precipitates inAl-Li-Cu alloys (Ref 141, 142). Because order-ing does not usually alter thermodynamic con-ditions enough to modify corrosion behavior, asdiscussed previously, then this observationindicates that atomic arrangement influencesthe kinetics of dissolution and lends supportto the percolation analysis of Ref 120. Of course,

this would mean that the type and stoichio-metry of the intermetallic structure will alsobe important and that the copper deposits foundoutside pits and cracks in Al-Li-Cu alloys maybe pieces of the dealloyed intermetallic phasebroken off by the hydrogen gas bubbles gener-ated in the cathodically active copper-rich spongeand not dissolved and reprecipitated copper.The L12 structure of A3B-type intermetallic

compounds frequently forms in binary face-centered cubic solid solutions (such as Ni3Aland Ni3Fe). In this structure, the B-type atomsare surrounded with only A-type nearest neigh-bors. As a result, these systems are excellentcandidates for investigating the ability of order-ing to induce stoichiometric limitations on dis-solution rates. If the A-type atoms are the morenoble in the structure, as is the case in the com-pounds identified previously, then the dissolu-tion rate of the more active species is limited tothe rate it becomes exposed by dissolution of themore noble species. Electrochemical measure-ments were conducted on Ni3Al in a wide rangeof solutions without observing any evidence ofdealloying (Ref 143). Figure 23 is an exampleof the current measured as this compound isdissolved anodically with closed-loop control ofpotential as the potential is increased in sulfuricacid (Ref 137, 143, 144). At this pH, this com-pound corrodes actively at open circuit but pas-sivates when anodically polarized. Included inthis figure is the current measured on pure nickel.By examining this figure, it can be seen thatordering in Ni3Al appears to block aluminumdissolution until nickel dissolution starts, atwhich point the current created by dissolution

+2.0

+1.5 [10–4]

+1.0

+0.5

0.0

– 0.5

–1.0

Pot

entia

l to

hydr

ogen

ele

ctro

de, V

–1.5

– 2.0

– 2.5

– 3.0– 2 0 2 4 6 8

pH

Ni Immune

Ni2+

NiO2

Ni3O4

Ni(OH)2

HNiO2–

H2

O2

NiEquilibria lines

Raised 10 to 100 mV

10 12 14 16

+2.0

+1.5

+1.0

+0.5

0.0

– 0.5

–1.0

Pot

entia

l to

hydr

ogen

ele

ctro

de, V

–1.5

– 2.0

– 2.5

– 3.0– 2 0 2 4 6 8

pH

Al Immune

Al3+

[10–4][10–4]

AIO2–

[10–4]

H2

O2

AlEquilibria lines

Raised 260 to 530 mV

Al oxidesand/or

hydroxides

10 12 14 16

Fig. 22 Electrochemical equilibrium diagram for nickel and aluminum showing the influence of L12 ordering (Ni3Al) on nickel and aluminum dissolution and passivation reactions.The potential ranges for the aluminide correspond to the aluminum concentration range from 0.2 to 0.3 mol fraction. The dashed lines illustrate the range of water stability

from oxygen to hydrogen evolution, and it can be seen that ordering has not significantly altered the position of the nickel and aluminum equilibria relative to these reactions.




increases slightly more rapidly on the inter-metallic until the onset of passivity, after whichlower currents are observed on the aluminide.Clearly, aluminum dissolution, which couldcause hydrogen evolution, absorption, andembrittlement, is limited to the stoichiometricratio and determined by the rate of nickel dis-solution (Ref 144).Another way to examine this issue is to sud-

denly expose the bare and reactive surface of anintermetallic compound to a solution whilemeasuring the potential with a reference elec-trode that has a potential fixed on the hydrogenelectrode scale and therefore allows for directmeasurement of the thermodynamic drivingforces. When a bare surface is exposed rapidly toa solution in this manner, anodic reactions startimmediately, while the mass transport requiredby cathodic reactions delays their initiation for afew milliseconds. Theoretically, this will resultin the potential of the sample becoming morenegative until the potential for equilibrium withthe chemical potential of the species dissolvingfrom the intermetallic compound is reached (Ref145–148). When an experiment of this type wasperformed on Ni3Al, the potential did not gobelow that predicted for nickel in Fig. 23, asshown in Fig. 24 (Ref 144). This does not indi-

cate that aluminum is not dissolving, only that atpotentials near the equilibrium potential fornickel dissolution, where there is still a largeoverpotential driving aluminum dissolution andthe net nickel dissolution rate is approaching 0,the charge created by aluminum dissolution isnot enough to overwhelm the nickel oxidation-and-reduction-exchange current density.Deformation, Slip Character, Stress-

Corrosion Cracking, and Hydrogen Embrit-tlement. When a random solid solutionbecomes ordered, it creates a specific arrange-ment of atomic nearest neighbors that lowers theenergy of the solid. When a dislocation passesthrough this solid, it breaks this arrangement andcreates a higher-energy region on the slip planebehind it, where the nearest neighbors of theatoms are not of the preferred type, until a seconddislocation passes and returns the region to theoriginal configuration. This antiphase regionbetween the two dislocations is analogous to astacking fault, where, instead of a dislocationdecomposing into two partial dislocations con-nected by a stacking fault, there is now a super-dislocation that is two dislocations connected byan antiphase domain boundary. As in the case ofstacking faults, the formation of antiphasedomain boundaries limits cross slip and pro-motes less homogeneous planar slip. Limitingcross slip reduces the ability of the solid toaccommodate slip at grain boundaries where theslip planes of the adjacent grains are not aligned

and promotes the formation of dislocationpileups and cracks (Ref 149). The limited duc-tility of many intermetallic phases has beenattributed to this phenomenon, and the beneficialinfluence of boron on the ductility of manyintermetallic compounds has been attributed tothe influence of this addition on order and slip inthe grain-boundary region as well as to a bene-ficial influence on grain-boundary cohesive for-ces (Ref 150, 151). Of course, hydrogen andstress-corrosion cracking are both known topromote intergranular fracture. As a result, theinfluence of hydrogen produced by cathodiccharging on the ductility of Ni3Al (þ200 ppmboron) was studied (Ref 152), and it was foundthat cathodic hydrogen lowered the ductility ofthis material dramatically and changed the frac-ture mode from ductile transgranular to inter-granular. Similarly, the influence of potential andpH on the ductility of Ni3Al (þ200 ppm boron)was studied (Ref 153, 154), with the conclusionthat in any solution or pH, intergranular fractureand low ductility resulted whenever the potentialof the sample became low enough to producehydrogen, as shown in Fig. 25. Because mostintermetallic compounds have at least one com-ponent that has a chemical potential that is activewith respect to the potential required to sponta-neously produce hydrogen by decomposition ofwater, investigators have found that most inter-metallic compounds can be embrittled by expo-sure to humid air. This has created questions50

40

30

20

Cur

rent

den

sity

, A/m

2

Cur

rent

den

sity

, A/ft

2

10

0

–10

4.6

3.7

2.8

1.9

0.9

0

–0.9–300 –200 –100

Deaerated

NiNi3Al

0.5 mol/L H2SO4

0Potential, mV vs. SCE

100 200 300

Fig. 23 Comparison of the current density producedby dissolution of nickel and nickel aluminide

in deaerated sulfuric acid. SCE, saturated calomel electrode

0.0

– 0.4

– 0.8

–1.2

Pot

entia

l vs.

SC

E, V

–1.6

0 0.5 1.0 1.5

Time, ms

2.0

105 Pa(1 atm)H2 Equil. Fugacity

0.5 mol/L NaClPure AlNi3Al Alloy IC50Pure Ni

2.5 3.0

Fig. 24 Potential transients during and following thegeneration of fresh bare surface by scratching

of nickel, aluminum, and nickel aluminide in 0.5 mol/LNaCl, with the hydrogen evolution potential for thisenvironment indicated SCE, saturated calomel electrode

(d)

Ni3Al IC50

P(H2 ) = 10 5 Pa (1 atm)

–1.0

–0.8

–0.6

–0.4

–0.2

0.0

Cold workedAnnealed

0 4 8pH

(a)

Pot

entia

l vs.

SC

E, V

12

Strain ratio

0

0.2

0.4

0.6

0.8

1.0D

uctil

ity r

atio

, env

ironm

ent/

N2

Strength ratio

(b)

0 4 8

pH

12

(c)

Fig. 25 Slow strain-rate tensile test results for free corrosion (open-circuit) conditions in different pH environments,showing that when (a) the free corrosion potentials reach the point where hydrogen evolution is thermo-

dynamically possible, then (b) the ductility drops, and (c) the fracturemode changes from ductile microvoid coalescence to(d) intergranular. (c) 0.5 mol/L Na2So4, 6.5 pH. (d) 0.5 mol/L H2SO4, 0.7 pH. SCE, saturated calomel electrode




about whether the low ductility usually observedin room-temperature laboratory experiments onthese types of materials is an inherent property,

an environmental effect, or a combination ofboth (Ref 154, 155–157).

Practical Understanding

The objective of corrosion research is to con-vert scientific understanding into practical alloyselection and design guidelines. The under-standing of aqueous corrosion of intermetalliccompounds is not to the point where one can dothis unequivocally, and thorough testing toevaluate performance should be conducted forany application. This testing should include out-of-service conditions as well as in-service con-ditions. References 158, and 159 have recom-mended the use of corrosion-mode diagrams toguide alloy selection for corrosive applicationsin a manner similar to the deformation mechan-ism maps of Ref 160 to 162. These diagrams areessentially electrochemical equilibrium dia-grams where the modes of corrosion failure

observed are plotted instead of the calculatedequilibrium phase. An example of one of thesediagrams is shown in Fig. 26 for Ni3Al (Ref 163).By examining this figure, it can be seen that thisintermetallic compound is passive over most ofthe range of potentials and pH that water isthermodynamically stable. It is interesting tonote that while the chemical potential of alumi-num in this solid is sufficient to cause hydrogenevolution, no evidence of hydrogen embrittle-ment was observed in any solution until thepotential of the bulk sample reached thatrequired for hydrogen evolution. Also interestingis the ability of Ni3Al to exhibit passivity at amuch lower pH than that predicted by the equi-librium diagram of Fig. 23, but nickel also showspassivity at anomalously low pH (Ref 145).Table 7 shows the corrosion rates determined fornickel aluminide alloys in a variety of differentsolutions, and Table 8 lists similar data for ironaluminide as determined by Ref 164 and 165.The corrosion rates in this table reflect the

Table 7 Average corrosion penetration rates for nickel aluminide alloys in various aqueous solutionsCalculated from continuous 200 h exposure mass-loss data

Penetration, mm/d (mils/d)

Concentration Exposure at 25 �C (75 �F) Exposure at 95 �C (205 �F)

Aqueous solution mol/L Mass% IC-50(a) IC-218(b) IC-221(c) IC-50(a) IC-218(b) IC-221(c)

Sulfuric acid 1 9 0.14 (0.006) 0.14 (0.006) 0.20 (0.008) 434 (1.3)(d) 21 (0.8) 460 (2.4)(d)18 96 0.49 (0.02) 0.42 (0.02) 0.42 (0.02) 4.7 (0.19) 1.3 (0.05) 1.4 (0.06)

Hydrochloric acid 1 4 0.35 (0.01) 0.20 (0.008) 0.20 (0.008) 35 (1.4) 437 (1.5)(d) 77 (3.0)12 37 8.4 (0.03) 13 (0.51) 26 (1.02) 435 (1.4)(d) 435 (1.4)(d) 93 (3.7)(d)

Nitric acid 1 6 30 (1.2) 4111 (4.4)(e) 37 (1.5) 436 (1.4)(d) 437 (1.5)(d) 21 (0.83)8 40 4850 (33.5)(f) 0.07 (0.003) 0.20 (0.008) 433 (1.3)(d) 3.1 (0.12) 3.9 (0.15)

16 70 4.4 (0.17) 0.07 (0.003) 0.20 (0.008) 436 (1.4)(d) 4.1 (0.16) 14 (0.55)Hydrofluoric acid 1 2 0.20 (0.008) 0.20 (0.008) 0.20 (0.008) 437 (1.5)(d) 434 (1.3)(d) 28 (1.1)Phosphoric acid 1 10 0.14 (0.006) 0.14 (0.006) 0.14 (0.006) 433 (1.3)(d) 436 (1.4)(d) 52 (2.0)Oxalic acid 1 10 0.07 (0.003) 0.07 (0.003) 0.07 (0.003) (g) 0.20 (0.008) 0.28 (0.011)Acetic acid 1 6 0.20 (0.008) 0.014 (0.0006) 0 (0) 0.28 (0.01) 0.28 (0.011) 0.28 (0.011)Sodium chloride 0.6 3.5 0.014 (0.0006) 0.014 (0.0006) 0.007 (0.0003) 0.014 (0.0006) (g) (g)Sodium hydroxide 10 30 0.028 (0.001) 0.014 (0.0006) 0.014 (0.0006) (g) (g) (g)Ammoniumhydroxide

10 18 0.42 (0.02) 0.020 (0.0008) 0.014 (0.0006) (g) (g) (g)

Ferric chloride 0.2 20 4410 (16.1)(h) 4320 (12.6)(i) 62 (2.5) 436 (1.4)(d) 434 (1.3)(d) 479 (3.1)(d)

Alloy compositions in mass fraction: (a) IC-50, 11.3%Al, 0.6% Zr, 0.02%B, bal Ni. (b) IC-218, 8.5%Al, 7.8%Cr, 0.8% Zr, 0.02%B, bal Ni. (c) IC-221, 8.5%Al, 7.8%Cr, 1.7%Zr, 0.02%B, bal Ni. (d) Total dissolution in5200 h.(e) Total dissolution in565 h. (f) Total dissolution in58 h. (g) Very small mass gain. (h) Total dissolution in517 h. (i) Total dissolution in524 h. Source: Ref 164, 165

1.0

0.5

Pitting

Corrosion

0 2 4 6 8

pH

10 12 14

Passivity andlow corrosion rates W

ater

sta

ble

Stress-corrosion cracking(hydrogen embrittlement)

Pitting may occurif halide ions are present

(line based on [Cl–] ≈0.1 mol/L)

0

–0.5Pot

entia

l vs.

hyd

roge

n el

ectr

ode,

V

[H2] = 1 atm

[O2] = 1 atm

Fig. 26 Corrosion-mode diagram for nickel aluminide(Ni3Al) in aqueous solutions, as estimated

from electrochemical experiments

Table 8 Cavitating water-jet erosion test fornickel and iron aluminide compared to othermaterials

Material

Erosionrate,mg/h

Relativeerosionrate(a)

Nickel aluminide (IC-50, cold worked) 1.4 0.04Nickel aluminide (IC-221) 2.5 0.08Nickel aluminide (IC-218) 2.9 0.09Nickel aluminide (IC-50) 4.1 0.1Nickel aluminide (IC-50, plasma weldoverlay)

7.0 0.2

Stellite 21 (stick) 7.3 0.2Iron aluminide 11.6 0.4Nitronic 60 bar stock 12.8 0.4Aluminum oxide ceramic plate 18.4 0.6308 stainless steel overlay (stick) 31.6 1.0Carbon steel (cold-rolled bar) 32.3 1.0304 stainless steel plate 33.2 1.1Aluminum bronze 84.4 2.7

Water-jet pressure, 41 MPa (6000 psi). (a) Erosion rate/rate for 308stainless steel. Source: Ref 164, 165

Table 9 Average corrosion penetration rates for iron aluminide (Fe3Al) alloys in aqueoussolutions at room temperature

Solution mol/L

Concentration, penetration, mm/d (mils/d)

FA-84(a) FA- 129(b) FAL-Mo(c) 304L(d)

HCl 1 430 (17) 52 (2.5) 14 (0.55) 9.7 (0.38)H2SO4 0.5 400 (15.7) 72 (2.8) 120 (4.7) 0.014 (0.0006)HNO3 1 160 (6.3) 3.5 (0.14) 1.4 (0.06) 0.007 (0.0003)NaOH 1 0.042 (0.002) 0.021 (0.0008) 0.056 (0.002) 0.007 (0.0003)Na2S2O3 1 7.8 (0.31) 9.9 (0.39) 2.6 (0.10) 0.004 (0.0002)Na2S4O6 1 21 (0.83) 6.5 (0.26) 6.5 (0.26) 0.020 (0.0008)Acetic acid 1 0.20 (0.008) 0.014 (0.0006) 0 (0) 0.28 (0.011)NaCl 0.6 0.19 (0.008)(e) 0.13 (0.005)(e) 0.11 (0.004) 50.007 (0.0003)Seawater(synthetic)

. . . 0.07 (0.003)(e) 0.07 (0.003)(e) 0.06 (0.002) 50.007 (0.0003)

Alloy compositions in mass fraction: (a) FA-84, 28% Al, 2% Cr, 0.05% B, bal Fe. (b) FA-129, 28% Al, 5% Cr, 0.5% Nb, 0.2% C, bal Fe. (c) FAL-Mo,28% Al, 5% Cr, 1% Mo, 0.08% Zr, 0.04% B, bal Fe. (d) 304L, 18–22% Cr, 8–12% Ni, 2% Mo (max), 0.03% S (max), 1% Si (max), 0.045% P (max),0.03% C (max), bal Fe. (e) Localized corrosion initiated within 24 h




behavior indicated by the corrosion-mode dia-gram for Fig. 25. Perhaps the most interestinglow-temperature applications for intermetalliccompounds require resistance to wear and cor-rosion combined. These applications require acombination of strength and corrosion resis-tance, and as shown in Table 9, intermetalliccompounds have this combination and canexhibit many times better resistance to thesecombined failure mechanisms than existingalloys.This article was not written to provide a cat-

alog of the corrosion response of the numerousvarieties of intermetallics. Instead, the emphasisis placed on generic principles governing thecorrosion phenomena of intermetallics and thenfocusing on the corrosion response of selectedstructural intermetallics. Some of the informa-tion presented here originated from the authors’own research in this area. Lack of space hasforced the authors to be selective in choosingthe topics. Consequently, such areas as inter-diffusion modeling in closed and open corrosionsystems and modeling of scale spallation havebeen excluded.

REFERENCES

1. N.S. Stoloff and V.K. Sikka, Ed., PhysicalMetallurgy and Processing of IntermetallicCompounds, Chapman & Hall, 1996

2. T.A. Kircher and E.L. Courtright, Mater.Sci. Eng. A, Vol 155 (No. 1–2), 1992,p 67–74

3. P.K. Datta, K. Natesan, and J.S. Burnell-Gray, Coatings Technology, IntermetallicCoatings and Coatings for Intermetallics,Intermetallic Compounds: Principles andPractice, Vol 3, Progress, J.H. Westbrookand R.L. Fleischer, Ed., JohnWiley&SonsLtd., 2002, p 561–588

4. F.D. Richardson and J.H. Jeffes, JISI, Vol171, 1952, p 165

5. H.J. Grabke, High Temperature Corrosionin Multi-Reactant Gaseous Environments,High Temperature Materials Corrosion inCoal Gasification Atmospheres, J.F. Nor-ton, Ed., Elsevier Applied Science Pub-lishers, 1984, p 59–82

6. C. Wagner, Z. Elektrochem., Vol 63, 1959,p 772

7. F.S. Pettit, Trans. Metall. Soc. AIME, Vol239, 1967, p 1296–1305

8. P. Kofstad, High Temperature Corrosion,Elsevier Applied Science, London andNew York, 1988

9. N. Birks, G.H. Meier, and F.S. Pettit, JOM,Vol 39 (No. 12), 1987, p 28–31

10. D.L. Douglass, P. Kofstad, andG.C.Wood,Oxid. Met.,Vol. 45 (No. 5–6), 1996, p 529–620

11. K.N. Strafford and P.K. Datta, Mater. Sci.Technol., Vol 5 (No. 8), 1989, p 765–779

12. P.K. Datta, H.L. Du, D. Jenkinson, J.S.Burnell-Gray, and K.N. Strafford, TheDesign of Sulphidation Resistant Materi-

als: State-of-the-Art and Future Prospects,Elsevier Science BV (Netherlands), 1998,p 176–187

13. J.H. Westbrook and R.L. Fleischer, Ed.,Intermetallic Compounds: StructuralApplications and Intermetallic Com-pounds, John Wiley & Sons Ltd., 2000

14. G.C. Rybicki and J.L. Smialek, Oxid. Met.,Vol 31 (No. 3–4), 1989, p 275–304

15. A. Steiner and K.L. Komarek, Trans.Metall. Soc. AIME, Vol 230 1964, p 786

16. F.H. Stott and G.C.Wood,Mater. Sci. Eng.A, Vol 87 (No. 1–2), 1987, p 267–274

17. M.W. Brumm and H.J. Grabke, Corros.Sci., Vol 33 (No. 11), 1992, p 1677–1690

18. H.J. Grabke, M.W. Brumm, and B.Wagemann, Mater. Corros., Vol 47 (No.12), 1996, p 675–677

19. E. Schumann, G. Schnotz, U. Salzberger,and M. Ruhle, Oxidation of Single-Crystalc-Ni3Al at Low Oxygen Partial Pressure,High-Temperature Ordered IntermetallicsIV, Vol 213, L.A. Johnson, D.P. Pope, andJ.O. Stiegler, Ed., Materials ResearchSociety, 1991, p 951–956

20. J. Doychak and M. Ruhle, Oxid. Met., Vol31 (No. 5–6), 1989, p 431–452

21. K. Natesan, Oxid. Met., Vol 30 (No. 1–2),1988, p 53–83

22. J.L. Smialek and G.H. Meier, SuperalloysII, C.T. Sims, N.S. Stoloff, and W.C.Hagel, Ed.,Wiley, NewYork, 1989, p 293–326

23. R.T. Haasch, A.M. Venezia, and C.M.Loxton, J. Mater. Res.,Vol 7 (No. 6), 1992,p 1341–1349

24. J.A. Horton, J.V. Cathcart, and C.T. Liu,Oxid. Met.,Vol 29 (No. 3–4), 1988, p 347–365

25. G. Fisher, P.K. Datta, J.S. Burnell-Gray,W.Y. Chan, and J.C. Soares, Surf. Coat.Technol., Vol 110 (No. 1–2), 1998,p 24–30

26. E. Schumann, J.C. Young, M. Ruhle,and M.J. Graham, Oxid. Met., Vol 46(No. 1–2),1996, p 37–49

27. J. Jedlinski and G. Borchardt, Oxid. Met.,Vol 36 (No. 3–4),1991, p 317–337

28. B.A. Pint, J.R. Martin, and L.W. Hobbs,Oxid. Met.,Vol 39 (No. 3–4), 1993, p 167–195

29. B.A. Pint, M. Treska, and L.W. Hobbs,Oxid. Met.,Vol 47 (No. 1–2), 1997, p 1–20

30. I. Rommerskirchen and V. Kolarik,Mater.Corros., Vol 47 (No. 11), 1996, p 625–630

31. J. Jedlinski and S. Mrowec, Mater. Sci.Eng., Vol 87 (No. 1–2),1987, p 281–287

32. H. Todd, U.S. Patent 3,494,748, Feb 10,1970

33. R. Strieff and D.H. Boone, J. Mater. Eng.,Vol 10 (No. 1) 1988, p 15–26

34. R.G. Wing and I.R. McGill, Platin. Met.Rev., Vol 25 (No. 3), 1981, p 94–105

35. J.R. Nicholls and D.J. Stephenson, MetalsandMater.,Vol 7 (No. 3), 1991, p 156–163

36. T.N. Rhys-Jones, Corros. Sci., Vol 29(No. 6), 1989, p 623–646

37. B.M. Warnes and D.C. Punola, PowerGeneration Technology, International Re-view of Primary Power Production, Ster-ling Publishing, 1995, p 207

38. H.M. Tawancey, N.M. Abbas, and T.N.Rhys-Jones, Surf. Coat. Technol., Vol 49,1991, p 1–7

39. H. Jehn, J. Less Common Met., Vol 100,1984, p 321–339

40. J. Schaffer, G.M. Kim, G.H. Meier, andF.S. Pettit, The Role of Active Elements inthe Oxidation Behaviour of High Tem-perature Metals and Alloys, Elsevier, Am-sterdam, 1991, p 231

41. H.M. Tawancey, N. Sridar, B.S. Tawabin,and N.M. Abbas, J. Mater. Sci., Vol 27(No. 23), 1992, p 6463–6474

42. P.C. Patnaik, Advances in High Tempera-ture Structural Materials and ProtectiveCoatings, NRC Canada 1994, p 164

43. T.E. Strangman and P.A. Solfest, U.S.Patent 4,880,614, Nov 14, 1989

44. H. Hosada, T. Takahashi, M. Takehara,T. Kingetsu, and H. Masumoto, Mater.Trans., Vol 38 (No. 10), 1997, p 871–878

45. W.M. Clift, K.F. McCarty, and D.R.Boehme, Surf. Coat. Technol., Vol 47 (No.1), 1990, p 29–40

46. K.N. Lee and W.L. Worrell, Oxid. Met.,Vol 32 (No. 5–6), 1989, p 357–369

47. B. Schramm and W. Auer,Mater. Corros.,Vol 47 (No. 12), 1996, p 678–684

48. W.H. Lee and R.Y. Lin, Mater. Chem.Phys., Vol 77 (No. 1), 2002, p 86–96

49. W. Gao and C.H. Xu, High Temp. Mater.Process., Vol 19, 2000, p 379–387

50. C.H. Xu, W. Gao, and Y.D. He, Scr.Mater., Vol 42, 2000, p 975–980

51. W. Gao, C.H. Xu, and H. Gong, Inter-metallics, Vol 8, 2000, p 769–779

52. I. Rommerskirchen, B. Eltester, and H.J.Grabke, Mater. Corros., Vol 47, 1996,p 646–649

53. J.L. Smialek, J. Doychak, and D.J. Gay-dosh, Oxid. Met., Vol 34, 1990, p 259–275

54. A.A. Aljarany, “Studies of the DegradationBehaviour of Fe3Al and c-TiAl Inter-metallics,” Ph.D. thesis, NorthumbriaUniversity, 2002

55. K. Natesan, Proc. Seventh Annual Conf. onFossil Energy Materials, ORNL/FMP-93,Oak Ridge National Laboratory, 1993,p 249

56. K. Natesan and R.N. Johnson, Proc. Sec-ond Int. Conf. Heat-Resistant Materials,K.Natesan, P. Ganesan, and G. Lai, Ed., ASMInternational, 1995, p 591

57. W. Gao and C.H. Xu, High Temp. Mater.Process., Vol 18, 1999, p 351–362

58. W. Kai, S.H. Lee, D.L. Chiang, and J.P.Chu, Mater. Sci. Eng. A, Vol 258, 1998,p 146–152

59. W. Kai, M.T. Chang, C.D. Liu, D.L.Chiang, and J.P. Chu, Mater. Sci. Eng. A,Vol 329–331, 2002, p 734–744

60. J.H. DeVan and P.F. Tortorelli, Mater.High Temp., Vol 11, 1993, p 30–35




61. H.L. Du, P.K. Datta, J.S. Burnell-Gray, andJ. Bernardi, Microscopy of SulphidisedFe3Al, J. Mater. Sci., in press

62. D. Das, R. Balasubramaniam, and M.N.Mungole, J. Mater. Sci., Vol 37, 2002,p 1135–1142

63. C.M. Austin, T.J. Kelly, K.G. McAllister,and J.C. Chesnutt, Struct. Intermetallics,1997, p 413–425

64. T. Tetsui, Curr. Opin. Solid State Mater.Sci., Vol 4 (No. 3), 1999, p 243–248

65. R.A. Perkins and G.H. Meier, Proc.Industry-University Advanced MaterialsConference, F.W. Smith, Ed., Denver,1991, p 92

66. J.L. Smialek and D.L. Humphrey, Scr.Metall. Mater., Vol 26 (No. 11), 1992,p 1763–1768

67. G. Welsh and A.I. Kaveci, Oxidation ofHigh Temperature Intermetallic, T. Grob-stein and J. Doychak, Ed., The Minerals,Metals and Materials Society, 1989

68. G. Welsh, S. Freidman, and A. Kahveci,Microscopy of Oxidation,M.J. Bennett andG.W. Lorimer, Ed., Institute of Materials,London, 1991

69. H.L. Du, P.K. Datta, D.B. Lewis, and J.S.Burnell-Gray, Corros. Sci., Vol 36 (No. 4),1994, p 631–642

70. G. Fisher, P.K. Datta, and J.S. Burnell-Gray, Surf. Coat. Technol., Vol 113 (No.3), 1999, p 259–267

71. H.L. Du, P.K. Datta, Z. Klusek, and J.S.Burnell-Gray, Nanoscale Studies of theEarly Stages of Oxidation of TiAl-BasedAlloy, Oxid. Met., Vol 62 (No. 3–4), 2004,p 175–193

72. G.H.Meier and F.S. Pettit,Mater. Sci. Eng.A, Vol 153, 1992, p 548–560

73. Y. Shida and H. Anada, Corros. Sci., Vol35, 1993, p 945–953

74. D.W. McKee, Oxidation and Protection ofTi3Al-Based Intermetallic Alloys, High-Temperature Ordered Intermetallic AlloysV, Materials Research Society, 1993, p953–958

75. Z. Liu and G. Welsch, Metall. Trans. A,Vol 19, 1988, p 1121–1125

76. W.B. Retallick, M.P. Brady, and D.L.Humphrey, Intermetallics, Vol 6 (No. 4),1998, p 335–337

77. S. Frangini, A. Mignone, and F. DeRicca-dis, J. Mater. Sci., Vol 29, 1994, p 714–720

78. S.A. Kekare and P.B. Aswath, J. Mater.Sci., Vol 32, 1997, p 2485–2499

79. V. Shemet, H. Hoven, and W.J. Qua-dakkers, Intermetallics, Vol 5 (No. 5),1997, p 311–320

80. J. Geng, G. Gantner, P. Oelhafen, and P.K.Datta, Appl. Surf. Sci., Vol 158, 2000,p 64–74

81. S.K. Varma, A. Chan, and R.N.Mahapatra,Oxid. Met., Vol 55 (No. 5–6), 2001, p423–435

82. G.H. Meier, F.H. Pettit, and S. Hu, HighTemperature Corrosion, Les Embiez,

France, 1992; Proc. J. Phys. IV, Vol 3,1993, p 395

83. N.S. Choudhury, H.C. Graham, and J.W.Hinze, Proceeding of High TemperatureAlloys, Z.A. Foroulis and F.S. Pettit, Ed.,The Electrochemical Society, 1976, p 99

84. N. Zheng, W.J. Quadakkers, A. Gil, andH. Nickel, Oxid. Met., Vol 44, 1995,p 477–499

85. S. Taniguchi, Y. Tachikawa, and T. Shi-bata, Mater. Sci. Eng. A, Vol 232 1997,p 47–54

86. S. Becker, A. Rahmel, M. Schorr, andM. Schurtz, Oxid. Met., Vol 38 (No. 5–6),1992, p 425–464

87. H.L. Du, A. Aljarany, K. Klusek, P.K.Datta, and J.S. Burnell-Gray, Effects ofOxygen Partial Pressures on the OxidationBehaviour of Ti-46.7Al-1.9W-0.5Si be-tween 750 and 950 �C,Corros. Sci.,Vol 47(No. 7), 2005, p 1706–1723

88. D.R. Gaskell, Introduction to Metallurgi-cal Thermodynamics, Hemisphere Pub-lishing Corporation, 1981

89. H.L Du, P.K. Datta, A.L. Dowson, andM. Jacobs, Effects of Sulfur Partial Pres-sures on Degradation Behaviour of TiAlin H2/H2S/H2O Environments at 900 �C,Oxid. Met., in press

90. Defects and Diffusion in Metals—AnnualRetrospective II, D.J. Fisher, Ed., ScitecPublications, 1999

91. P.K. Datta, J.S. Burnell-Gray, H.L. Du,A.L. Dowson, and M.H. Jacobs, Proc.World Ceramics Congress and Forum onNew Materials, (Florence, Italy), WorldScientific Publishing Co., 1998

92. H.L. Du, P.K. Datta, J.S. Gray, and K.N.Strafford, Corros. Sci., Vol 36, 1994,p 99–112

93. H.L. Du, P.K. Datta, J. Leggett, J.R.Nicholls, J.C. Bryar, and M.H. Jacobs,The Influence of Processing Route onthe Oxidation/Sulphidation of a Ti-48Al-2Nb-2Mn Intermetallic Alloy at 750and 900 �C, Advances in Surface Engi-neering, Vol I, Proc. Conf. on Advances inCoatings and Surface Engineering forCorrosion and Wear Resistance and OtherApplications, (Newcastle upon Tyne,U.K.), Royal Society of Chemistry 1996,p 53–66

94. H.L. Du, P.K. Datta, D. Griffin, A. Al-jarany, and J.S. Burnell-Gray, Oxid. Met.,Vol 60, 2003, p 29–46

95. H.L. Du, S.R. Rose, Z.D. Xiang, and P.K.Datta, Mater.wiss. Werkst.tech., Vol 34,2003, p 421–426

96. K. Zhang, Z. Li, and W. Gao,Mater. Lett.,Vol 57, 2002, p 834–843

97. K. Zhang, Z. Li, M.A. Hodgson, and W.Gao, High Temp. Mater. Process., Vol 21,2002, p 167–175

98. Z. Tang, F. Wang, and W. Wu,Mater. Sci.Eng. A, Vol 276, 2000, p 70–75

99. Z. Tang, F. Wang, and W. Wu, Oxid. Met.,Vol 51, 1999, p 235–250

100. K. Natesan and P.K. Datta, SulphidationBehaviour, Intermetallic Coatings andCoatings for Intermetallics, IntermetallicCompounds: Principles and Practice,Vol 3, Progress, J.H. Westbrook and R.L.Fleischer, Ed., John Wiley & Sons Ltd.,2002, p 707–719

101. P.J. Meschter, Metall. Trans. A, Vol 23,(No. 6), 1992, p 1763–1772

102. Y.Q. Liu, G. Shao, and P. Tsakiropoulos,Intermetallics, Vol 9 (No. 2), 2001, p125–136

103. T.E. Tietz and J.W.Wilson, Behaviour andProperties of Refractory Metals, StanfordUniversity Press, 1965

104. R. Mitra and V.V. Rama Rao, Mater. Sci.Eng. A, Vol 260, 1999, p 146–160

105. B.V. Cockeram, G. Wang, and R.A. Rapp,Mater. Corros., Vol 46, 1995, p 207–217

106. K. Yanagihara, K. Przybylski, and T.Maruyama, Oxid. Met. Vol 47 (No. 3–4),1997, p 277

107. T.C. Chou and T.G. Nieh, JOM, Vol 45(No. 12), 1993, p 15–21

108. R.E. Regan, W.A. Baginski, and C.A.Krier, Ceram. Bull., Vol 46, 1967, p502–508

109. H.L. Du, P.K. Datta, R. Cutting, and J.S.Burnell-Gray, “Sulphidation Behaviour ofSiliconised Mo and Nb by Pack Cementa-tion at 850 �C,” presented at the Interna-tional Conference on Metallurgical Coat-ings and Thin Films, ICMCTF 2004, April19–23, 2004 (San Diego, CA), AmericanVacuum Society

110. E. Fitzer and J. Schwab,Corrosion,Vol 12,1956, p 455–464

111. E. Fitzer, W. Nowak, and H. Maurer, Arch.Eisenhuttenwes., Vol 49, 1978, p 211–216

112. E. Fitzer, W. Nowak, and H.J. Maurer,Materials and Coatings to Resist HighTemperature Corrosion, D.R. Holmes andA. Rahmel, Ed., Applied Science Publish-ers, London, 1978, p 313–331

113. E. Fitzer, H.J. Maurer, W. Nowak, and J.Schlichting, Thin Solid Films, Vol 64,1979, p 305–319

114. P. Felix and E. Erdos,Werkst. Korros.,Vol23, 1972, p 627–636

115. E. Erdos, M. Semlitsch, and P. Felix,Z. Werkstofftech., Vol 3, 1972, p 193–197

116. M. Villat and P. Felix, Tech. Rundsch.Sulzer, Vol 58, 1976

117. I. Daimer, E. Fitzer, and I. Schlichting,Thin Solid Films, Vol 84, 1981, p 119–125

118. E. Fitzer and I. Schlichting, High Tem-perature Corrosion, R.A. Rapp, Ed.,NACE-6, National Association of Corro-sion Engineers, 1981, p 604

119. H.J. Felten and F.S. Pettit, Oxid. Met., Vol10, 1976, p 189–223

120. J.S. Sheasby andD.B. Jory,Oxid. Met.,Vol12, 1978, p 527–539

121. K.N. Lee and W.L. Worrell, Oxid. Met.,Vol 32, 1989, p 357–369

122. T.C. Chou, J. Mater. Res., Vol 5, 1990,p 378–384




123. R.L. Fleischer, Platin., Met. Rev., Vol 36,1992, p 139–145

124. D.W. McKee and R.L. Fleischer, HighTemperature Ordered Intermetallics IV,Vol 213, L.A. Johnson, D.P. Pope, and J.O.Stiegler, Ed., Materials Research Society,1991, p 969–974

125. E.J. Felten, Trans. Metall. Soc. AIME, Vol245, 1969, p 1349–1355

126. R.M. Paine, A.J. Stonehouse, and W.W.Beaver, Corrosion, Vol 20, 1964, p307–313

127. E.A. Aitken, Intermetallic Compounds,J.H. Westbrook, Ed., Wiley, 1967, p491–516

128. R.A. Buchanan, J.G. Kim, R.E. Ricker, andL.A. Heldt, Oxidation and Corrosion ofIntermetallic Alloys, G.E. Welsch and P.D.Desai, Ed., Purdue University, HTMIAC/CINDAS, 1996, p 351–419

129. R.A. Buchanan, J.G. Kim, R.E. Ricker, andL.A. Heldt, Physical Metallurgy and Pro-cessing of Intermetallic Compounds, N.S.Stoloff and V.K. Sikka, Ed., Chapman andHall, 1996

130. D.J. Duquette,Mater. Sci. Eng. A, Vol 198(No. 1–2), 1995, p 205–211

131. N.S. Stoloff, M. Shea, and A. Castagna,Environmental Effects on AdvancedMaterials,R.H. Jones and R.E. Ricker, Ed.,The Minerals, Metals and MaterialsSociety, 1991, p 3–19

132. N.S. Stoloff, C.T. Liu, and S.C. Deevi,Intermetallics, Vol 8, 2000, p 1313–1320

133. R.G. Buchheit, R.K. Boger, M.C. Carroll,R.M. Leard, C. Paglia, and J.L. Searles,JOM, Vol 53, 2001, p 29–33, 36

134. U. Bertocci, W.J. Boettinger, C.A. Hand-werker, and S.M. Hues, Corrosion andReliability of Electronic Materials andDevices, R.B. Comizzoli and J.D. Sinclair,Ed., Vol. Proc. of the Second Int.Sym., The Electrochemical Society, 1992,p 99–118

135. M. Pourbaix, Atlas of ElectrochemicalEquilibria in Aqueous Solutions, NACEInternational, 1974

136. R.E. Ricker and U. Kattner, NIST, Gai-thersburg, MD, communication to D.Wang, Lehigh Univ., 1997

137. R.E. Ricker, International Symposium onNickel and Iron Aluminides: Processing,Properties and Applications, S.C.Deevi, V.K. Sikka, P.J. Maziasz, and R.W.Cahn, Ed., ASM International, 1997, p253–263

138. H.H. Uhlig, Corrosion and CorrosionControl, 2nd ed., JohnWiley & Sons, 1971

139. M.G. Fontana and N.D. Greene, CorrosionEngineering, McGraw-Hill Book Co.,1978

140. K. Sieradzki, R.R. Corderman, K. Shukla,and R.C. Newman, Philos. Mag. A,Vol 59,1989, p 713–746

141. R.G. Buchheit, M.A. Martinez, and L.P.Montes, J. Electrochem. Soc., Vol 147,2000, p 119–124

142. R.G. Buchheit, L.P. Montes, M.A. Marti-nez, J. Michael, and P.F. Hlava, J. Elec-trochem. Soc.,Vol 146, 1999, p 4424–4428

143. U. Bertocci, R.E. Ricker, D.E. Hall, P.V.Madsen, and J.L. Fink, Corros. Sci., Vol31, 1990, p 471–478

144. R.E. Ricker, Mater. Sci. Eng. A, Vol 198,1995, p 231–238

145. W. Latimer, Oxidation Potentials, 2nd ed.,Prentice Hall, 1952

146. T. Hagyard and J.R. Williams, Trans.Faraday Soc., Vol 57, 1961, p 2288–2294

147. A.A. Adams and R.T. Foley, Corrosion,Vol 31, 1975, p 84

148. M.R. Stoudt, A.K. Vasudevan, and R.E.Ricker, Corrosion Testing of AluminumAlloys, STP 1134, V.S. Agarwala and G.Ugiansky, Ed., ASTM, 1992

149. R.G. Davies andN.S. Stoloff,Philos.Mag.,Vol 12, 1965, p 297–304

150. R.G. Davies and N.S. Stoloff, Trans.AIME, Vol 230, 1964, p 391–395

151. G.M. Camus, N.S. Stoloff, and D.J.Duquette, Acta Metall., Vol 37, 1989,p 1497–1501

152. A.K. Kuruvilla and N.S. Stoloff, Scr.Metall., Vol 19, 1985, p 83–87

153. R.E. Ricker, D.E. Hall, and J.L. Fink, Scr.Metall. Mater., Vol 24, 1990, p 291–296

154. R.E. Ricker, M.R. Stoudt, and J.L. Fink,Hydrogen Embrittlement of Ductile NickelAluminide during Corrosion in AqueousSolutions, TMS/AIME, 1990

155. C.T. Liu, E.H. Lee, and C.G. McKamey,Scr. Metall., Vol 23, 1989, p 875–880

156. C.T. Liu and W.C. Oliver, Scr. Metall.Mater., Vol 25, 1991, p 1933–1937

157. C.T. Liu and N.S. Stoloff, Diffusionin Ordered Alloys, The Minerals, Metalsand Materials Society, 1993, p 223–246

158. R.W. Staehle,Mechanisms of EnvironmentSensitive Cracking of Materials, P.R.Swann, F.P. Ford, and A.R.C. Westwood,Ed., The Metals Society, London, U.K.,1977, p 574–602

159. R.W. Staehle, Environment-InducedCracking of Metals, Vol 10, R.P. Gangloffand M.B. Ives, Ed., NACE International,1989, p 561–612

160. M.F. Ashby, C. Gandhi, and D.M.R.Taplin, Acta Metall., Vol 27, 1979, p699–729

161. R.J. Fields, T. Weerasooriya, and M.F.Ashby, Metall. Mater. Trans. A, Vol 11,1980, p 333–347

162. C. Gandhi and M.F. Ashby, Acta Metall.,Vol 27, 1979, p 1565–1602

163. R.E. Ricker, JOM, Vol 47, 1995, p 32–35164. R.A. Buchanan and J.G. Kim, “Aqueous

Corrosion Characteristics of Nickel Alu-minides: Final Report,” Univ. of Tennes-see, final report to Nickel DevelopmentInstitute, 1988

165. R.A. Buchanan and J.G. Kim, “AqueousCorrosion of Iron Aluminides,” Univ. ofTennessee, report to Oak Ridge NationalLaboratory, 1989

SELECTED REFERENCES

� M. Bachorczyk, M. Danielewski, P.K. Datta,and R. Filipek, Model of HeterogeneousReaction Controlled by Diffusion, J. Mole-cular Liq., Vol 86, 2000, p 61

� M. Bachorczyk, M. Danielewski, P.K. Datta,R. Filipek, and G. Fisher, Computer Simula-tions of Heterogeneous Reactions Controlledby Diffusion in Modified Aluminide Coatingson a Nickel-Based Superalloy, Solid StatePhenom., Vol 72, 2000, p 53–58

� M. Berlanga, J.L. Gonzalez-Carrasco, M.A.Montealegre, and M.A. Munoz-Morris, Oxi-dation Behaviour of Yttria DispersionStrengthened Fe40Al Alloy Foils, Inter-metallics, Vol 12, 2004, p 205–212

� P.K. Datta, R. Filipek, and M. Danielewski,Interdiffusion Issues in Pt-Modified NiAlCoatings,Defect Diffus. Forum,Vol 203–295,2002, p 47–60

� V. Gauthier, F. Dettenwanger, and M.Schutze, Oxidation Behavior of c-TiAl Coat-ed with Zirconia Thermal Barriers, Interme-tallics, Vol 10, 2002, p 667–674

� D. Griffin, A. Daadbin, and P.K. Datta,Deformation and Fracture, during Cooling, ofthe Alumina Scale Developed on Fe3Al, Surf.Coat. Technol., Vol 126, 2000, p 142–151

� J. Kumpfert, Intermetallic Alloys Based onOrthorhombic TitaniumAluminide, Adv. Eng.Mater., Vol 3, 2001, p 851–864

� F. Lang, Z. Yu, S. Gedevanishvili, S.C. Deevi,and T. Narita, Effect of Pre-Oxidation on theCorrosion Behaviour of Fe-40Al Sheet ina N2-11.2O2-7.5CO2-500 ppm SO2 Atmo-sphere at 1273 K, Intermetallics, Vol 11,2003, p 129–134

� M.K. Lei, X.P. Xhu, andX.J.Wang, OxidationResistance of Ion-Implanted c-TiAl-Base In-termetallics, Oxid. Met., Vol 58, 2002,p 361–374

� Z. Li, W. Gao, J. Liang, and D.L. Zhang,Oxidation Behaviour of Ti3Al-TiC Compos-ites, Mater. Lett., Vol 57, 2003, p 1970–1976

� Z. Li, W. Gao, M. Yoshihara, and Y. He,Improving Oxidation Resistance of Ti3Al andTiAl Intermetallic Compounds with Electro-Spark Deposit Coatings, Mater. Sci. Eng. A,Vol 347, 2003, p 243–252.

� Z. Liu and T. Narita, The Effect of WaterVapor on the Oxidation Behaviour of c-TiAl-Ag Coatings at 1073 K in Air, Intermetallics,Vol 11, 2003, p 795–805.

� X.K. Meng, H. Shen, H. Vehoff, S. Mathur,and A.H.W. Ngan, Fractography, ElasticModulus, and Oxidation Resistance of NovelMetal-Intermetallic Ni/Ni3Al MultilayerFilms, J. Mater. Res.,Vol 17, 2002, p 790–796

� S. Naka, M. Thomas, and T. Khan, Potentialand Prospects of Some Intermetallic Com-pounds for Structural Applications, Mater.Sci. Technol., Vol 8, 1992, p 291–298

� E.C.T. Ramos, G. Silva, A.S. Ramos, C.A.Nunes, and C.A.R.P. Baptista, Microstructureand Oxidation Behaviour of Ti-Si-B Alloys,Mater. Sci. Eng. A, Vol 363, 2003, p 297–306




� M. Schutze, Z. Schumacher, F. Dettenwanger,U. Hornauer, E. Richter, E. Wieser, and W.Moller, The Halogen Effect in the Oxidationof Intermetallic Titanium Aluminides, Cor-ros. Sci., Vol 44, 2002, p 303–318

� J.L. Smialek, J.A. Nesbitt, C.A. Barrett, andC.E. Lowell, Cyclic Oxidation Testing andModeling, a NASA Lewis Perspective, Eur.Fed. Corros. Pub., Vol 27, 1999, p 148–168

� V. Supatarawanich, D.R. Johnson, and C.T.Liu, Effects of Microstructure on the Oxida-tion Behaviour of Multiphase Mo-Si-BAlloys, Mater. Sci. Eng. A, Vol 344, 2003,p 328–339

� Third International Symposium on StructuralIntermetallics, Sept 23–27, 2001 (JacksonHole, WY), The Minerals, Metals and Mate-rials Society

� F. Wu, H. Murakami, and H. Harada, CyclicOxidation Behavior of Iridium-ModifiedAluminide Coatings for Nickel-Base SingleCrystal Superalloy TMS-75, Mater. Trans.,Vol 44, 2003, p 1675–1678

� Z.D. Xiang and P.K. Datta, Deposition ofSi Modified Aluminide Coatings on NickelBase Superalloys by Pack Cementation Pro-cess, Mater. Sci. Technol., Vol 19, 2003,p 935–942

� Z.D. Xiang, S.R. Rose, and P.K. Datta, Oxi-dation Resistance of Diffusion CoatingsFormed by Pack-Codeposition of Al and Si onc-TiAl, J. Mater. Sci., Vol 39, 2004, p2099–2106

� H. Yokota, T. Kudoh, and T. Suzuki, Oxida-tion Resistance of Boronized MoSi2, Surf.Coat. Technol., Vol 169–170, 2003, p171–173

� L. Zhao and E. Lugscheider, High Velo-city Oxy-Fuel Spraying of a NiCoCrAlYand an Intermetallic NiAl-TaCr Alloy,Surf. Coat. Technol., Vol 149, 2002, p231–235




corrosion of intermetallics - tsapps at nist

Documents