`N. Bornstein
`
`To cite this version:
`
`N. Bornstein. Oxidation of advanced intermetallic compounds. Journal de Physique IV, 1993,
`03 (C9), pp.C9-367-C9-373. <10.1051/jp4:1993937>. <jpa-00252376>
`
`HAL Id: jpa-00252376
`
`https://hal.archives-ouvertes.fr/jpa-00252376
`
`Submitted on 1 Jan 1993
`
`HAL is a multi-disciplinary open access
`archive for the deposit and dissemination of sci-
`entific research documents, whether they are pub-
`lished or not. The documents may come from
`teaching and research institutions in France or
`abroad, or from public or private research centers.
`
`L’archive ouverte pluridisciplinaire HAL, est
`destin´ee au d´epˆot et `a la diffusion de documents
`scientifiques de niveau recherche, publi´es ou non,
`´emanant des ´etablissements d’enseignement et de
`recherche fran¸cais ou ´etrangers, des laboratoires
`publics ou priv´es.
`
`GE-1012.001
`
`
`
`Article published online by EDP Sciences and available at http://dx.doi.org/10.1051/jp4:1993937
`
`JOURNAL DE PHYSIQUE IV
`Colloque C9, suppltment au Journal de Physique 111, Volume 3, decembre 1993
`
`Oxidation of advanced intermetallic compounds
`
`N.S. Bornstein
`
`United Technologies Research Center, East Hartford, CT 06108, U.S.A.
`
`Abstract. - Intermetallic compounds with low densities and high melting temperatures are
`prime structural material candidates for advanced heat engines. The close link between crystal
`structure and mechanical properties has led the search for new cubic pseudo-binary transition
`metal aluminides. Compositing leads the way for improved mechanical properties of refractory
`metal silicides. In all cases there is a strong competition for oxide scale frequency between the
`energetic oxide formers of the intermetallic compounds. In the case of titanium aluminides,
`efforts are directed towards modification of the concentration of oxygen vacancies to reduce
`growth rates and influence scale dominance. Glass formers are added to the disilicides to pro-
`mote scale fluidity to seal cracks which are pathways for oxidation.
`
`1. Introduction.
`
`The most successful intermetallics are coatings. The most successful coatings are the inter-
`metallics based upon the aluminides of the transition metals notably nickel, cobalt and iron.
`The remarkable growth of the gas turbine superalloys is based in part upon the ability of
`the material engineer to concentrate efforts to improve mechanical properties and employ
`coatings to achieve acceptable oxidation and corrosion resistance.
`Further increases in efficiency and performance of gas turbine engines is possible through
`the use of materials that can operate at higher temperatures and possess greater strength
`to weight ratios. Intermetallic compounds are candidate materials. The nickel aluminides
`are more refractory than the current transition element alloys. The aluminides of titanium
`promise higher strengths with lower densities, while composites based upon molybdenum
`disilicide offer high temperature strength and oxidation resistance. Thus the objective of
`this address is to identify current strengths and weaknesses in order to further promote the
`state-of-the-art.
`
`2. Oxidation resistance.
`
`Oxidation resistance is often the limiting factor that determines the life of high temperature
`structural metals and alloys. The kinetics, morphology and composition of the oxidation
`products are obtained from thermogravimetric analytical studies (TGA), light and electron
`microscopy studies, and various analytical techniques. All of the above are used to determine
`the rate of growth of the oxidation products. The rate of growth of the oxidation products are
`often described by a parabolic function and the relative merit of performance is commonly
`presented in pictorial form as a plot of the log of the parabolic scaling constant versus the
`reciprocal of the absolute temperature.
`
`GE-1012.002
`
`
`
`368
`
`JOURNAL DE PHYSIQUE IV
`
`Although the rate of growth of the oxide scale, determined from isothermal studies is a
`measure of performance; it is, at best, a poor yardstick. Oxidation resistance is not only a
`measure of the rate at which an oxide scale thickens, but includes the ability of the scale to
`resist the thermally induced stresses associated with cyclic behavior. A typical example is the
`relative behavior of the materials, Ni-15Cr-6Al and Ni-15 Cr-6Al-O.1Y. In isothermal tests
`both materials form alumina scales. However in a cyclic test the yttrium-containing alloy
`maintains its ability to form a protective adherent oxide scale. However the alumina scale
`on the yttrium free alloy spalls at the completion of each cycle (Fig. 1). Thus although both
`alloys form alumina scales, the alloy containing yttrium exhibits "oxidation resistance". If in-
`termetallic compounds are to be structural components operating in oxidizing environments,
`they must exhibit both the necessary mechanical properties as well as oxidation resistance. It
`is therefore important to understand the factors that influence oxide scale adherence.
`
`OXIDATION OF AS-CAST NtCrAI AN0 NSrNY ALLOYS AT 11504:
`
`0
`
`-40
`
`E
`'O
`
`e W
`
`0 z a
`1 -80
`0
`f
`!2
`W
`3 -120
`
`ISOTHERWL
`
`CYCLIC
`ALLOY
`0 N i 1 5 C r 6 U O . l Y
`
`0
`
`100
`
`200
`
`3M)
`TIME. hours
`
`400
`
`500
`
`Fig. 1. - Yttrium greatly enhances oxide scale adherence.
`
`3. Alumina formers.
`
`3.1 NICKEL BASE ALLOYS. - Tien et al. [I] describing cyclic oxidation says "violence is done
`on oxide scales during cyclic oxidation by growth and thermal stresses". It is well established
`that the addition of small quantities of certain elements and compounds greatly improves
`oxide scale adherence. The phenomena is called the "reactive element effect" (REE). A num-
`ber of mechanisms have been proposed and reviewed [2,3]. Yttrium is the most commonly
`used reactive element to impart scale spallation resistance to alumina forming alloys. Re-
`cently it has been proposed that the alumina scale that forms on the transition metal alloys
`and compounds is strongly bonded, but impurities present in the substrate promote oxide
`scale spallation. The principal impurity is sulfur. It is important to establish if the removal of
`sulfur below some critical level imparts oxidation resistance since it is not always commercially
`feasible to add reactive elements without incurring other liabilities.
`
`GE-1012.003
`
`
`
`OXIDATION OF ADVANCED INTERMETALLIC COMPOUNDS
`
`369
`
`The sulfur model is the product of ONR-sponsored research to explain the improvement
`observed in oxide scale adherence when surfaces are laser processed [4]. The strength of the
`model is its simplicity. The model assumes the existence of a strong chemical bond between
`the alumina scale and the alloy substrate. The bond is weakened by sulfur, a tramp impurity,
`which segregates to the oxide scale interface. The mechanism by which sulfur weakens the
`bond is not described, however it is proposed that superior oxide scale adherence is achieved
`by reducing the activity of sulfur in the substrate either by removal of sulfur or uniting the
`sulfur with a potent sulfide former. The role of the REE is to unite with and thereby reduce
`the activity of sulfur in the substrate.
`
`I
`
`NiCrAlY
`
`I
`
`Fig. 2. -Auger spectra showing S, Zr or Y segregation.
`
`ELECTRON ENERGY. r V
`
`Auger studies are used to demonstrate that the reactive elements reduce the activity of
`sulfur in Ni-Cr-A1 alloys, figure 2 [5]. Other studies have shown that the addition of sulfur
`to NiCrAlY alloys prevents the alloy from forming spallation resistant scales. However the
`
`GE-1012.004
`
`
`
`370
`
`JOURNAL DE PHYSIQUE IV
`
`addition of more yttrium restores resistance to oxide scale spallation (Fig. 3). This data sup-
`ports the sulfur-yttrium interaction. However the recent work of Smialek et al. [6] strongly
`indicates that the simple reduction in the concentration of sulfur in the matrix is sufficient
`to impart resistance to oxide scale spallation. The heat treatment of nickel base alloys in
`hydrogen environments was used to remove sulfur from the alloy, and the subsequent high
`temperature cyclic oxidation tests verified the formation of a spallation resistant oxide scale.
`
`(1 100.C Cyclic Test)
`
`- 8
`
`-I6
`- 2 4
`
`- 3 2
`
`Welghl change,
`rng/crn2
`
`pprn
`S
`
`pprn
`Y
`
`- -
`
`(a)
`(b)
`( c )
`(d)
`( e )
`
`40
`3 0 0
`40
`3 0 0
`3 0 0
`
`0 - 4 0
`O
`0
`890
`900
`4 0 0 0
`
`2 0 0
`
`4 0 0
`Time hours,
`cyclic oxldation
`
`6 0 0
`
`8 0 0
`
`Fig. 3. - Effect of yttrium and sulfur content on the cyclic oxidation resistance of NiCrAl.
`
`The intermetallic compound NlAl is quite brittle at low temperatures and rather weak at
`elevated temperatures. Nevertheless it is suggested for use in advanced applications as the
`matrix phase for metallic based composites. The face centered cubic intermetallic compound
`Ni3Al, when alloyed with small quantities of boron, exhibits attractive mechanical proper-
`ties over a wide range of temperatures. Although the oxidation behavior of NiAl has been
`extensively studied, further improvements, taking into account the role of sulfur, should be
`explored. The oxidation behavior of boron modified Ni3Al with respect to sulfur and oxide
`scale adherence should be investigated.
`
`3.2 TITANIUM ALUMINIDES. - Alloy designers are concerned with finding materials with
`superior properties. In the aircraft industry new light weight strong oxidation resistant ma-
`terials are sought. Some of the most attractive aluminides from the view point of low density
`and high temperature potential are those in the Ti-A1 system such as Ti3A1 and TiAl. Their
`densities of 3.76 to 4.50 g/cm3 are only half the densities of nickel and cobalt superalloys.
`Large weight savings are possible, however the use of titanium aluminides at high tempera-
`ture is limited by relatively poor oxidation resistance.
`Titanium and aluminium are both energetic oxide formers. It is generally agreed that
`alumina is the scale that forms on the surface of TiA13. However this very brittle intermetallic
`compound draws little interest.
`Meier et al. [7] using the criteria established by Wagner [8] and data available in the litera-
`ture conclude that the solubility of aluminum in titanium in both the alpha and beta phases
`is insufficient for the alloy to develope an alumina protective external scale. The aluminum
`in the alloy is preferentially internally oxidized. Welsch et al. [9] as well as Choudhury et
`
`GE-1012.005
`
`
`
`OXIDATION OF ADVANCED INTERMETALLIC COMPOUNDS
`
`37 1
`
`al. [lo] studied the oxidation behavior of Ti3AI and found that the alloy formed primarily
`titania-rich scales. Moreover Welsch reports that even in the isothermal tests, the weight gain
`curves exhibit discontinuities due to cracking or spalling of the oxide layer. Nevertheless they
`were able to determine parabolic scaling constants and concluded that the principal scale is
`titania.
`Meier reports that TiAl exposed in oxygen in the temperature range 800 to 1000 "C forms
`alumina scales. Moreover the scaling constant is comparable to that of other alumina formers
`such as NiAl. However when the temperature is increases above 1000 OC, the oxidation rate
`increases and the alloy forms titanium rich scales. Moreover the authors clearly identify an
`effect related to nitrogen. The same alloys exposed in air does not form alumina scales and
`the authors note that the effect of nitrogen on the selective oxidation of aluminum from TiAl
`is very much in need of further study.
`Hanamura et al. [ l l ] takes a different approach to improve the oxidation resistance of
`TiAl. They report that the alloy is a titania former. According to the authors the growth of
`titanium oxide occurs by the inward diffusion of oxygen ions through Tion. Therefore the
`polyvalent elements of group Vb and VIb having higher valencies than Ti (4+), introduced
`into the titania scale, should reduce the concentration of oxide ion vacancies thereby reducing
`the oxidation rate. The authors report that the introduction of about 100 ppm of selenium
`or tellurium is effective. It decreased the weight gain of TlAl exposed at 800 "C by more
`than five-fold. Ikematsu et al. [12] also studied the effect of additions of phosphorus on the
`oxidation behavior of TiAl. They report that the addition of 500 ppm P also decreases the
`weight gain of the alloy by five-fold when exposed 500 hours at 900 "C. Again the authors
`relate the improvement in performance to the reduction in the concentration of oxygen va-
`cancies in Tion. It is quite clear that the area of oxidation of titanium base intermetallics is
`quite complex and poorly understood.
`
`4. Silica formers.
`
`The refractory metals typified by niobium, tantalum, and molybdenum exhibit desirable high
`temperature properties. They are strong and ductile and can be formed and joined. Thus
`the refractory metals offer distinct advantages over ceramics. However the refractory metals
`and alloys lack oxidation resistance. The coatings developed to protect the refi-actory metals
`are based upon the oxidation performance of silicide intermetallic compounds.
`The silicide coated refractory metals, exposed at elevated temperatures in highly oxidizing
`environments, do not offer the degree of reproducibility generally required for man rated
`applications. The issue is not the rate of oxidation of the coating. Unfortunately, the current
`family of silicide coatings contain numerous cracks which permit oxygen to enter into and
`react with the substrate alloy. The source of the cracks is generally believed to be stresses asso-
`ciated with the differences in coefficients of thermal expansion between the substrate and the
`coating. The changes in mechanical properties of the substrate limit the usefulness of the sys-
`tem long before the coating has been compromised. Nevertheless efforts should be directed
`toward improving the oxidation resistance of silicide coatings and manufacturing methods
`to produce defect - free coatings. Molybdenum disilicide is unique in that it is the principal
`constituent of a widely - used heating element for high temperature furnace applications and
`can be considered as a matrix material for high temperature composites. Coatings based
`upon the compound MoSi2 differ from other silicides in that scale is composed of only silica.
`According to Lavendel et al. [13] the formation of essentially pure silica scales on MoSi2 is
`the result of the selective oxidation due primarily to the large differences in thermodynamic
`
`GE-1012.006
`
`
`
`372
`
`JOURNAL DE PHYSIQUE IV
`
`stabilities of silica and molybdenum trioxide (Tab. I). They argue that since the free energy
`of formation of silica is much greater than the oxides of molybdenum, the available oxygen
`will partition itself strongly in favor of the formation of silica. The situation however is not
`true for other refractory silicides such as niobium or tantalum. The standard free energies of
`formation of the oxides of tantalum and niobium are nearly the same as that of silica. Thus
`according to the authors, the available oxygen will tend to partition approximately equally
`between the metal and silicon and there will be no selective oxidation of either component.
`Of course kinetics are overlooked as well as the fact that of the refractory metal oxides, only
`molybdenum trioxide is volatile. Molybdenum trioxide melts at 795 "C and boils at 1280 "C.
`Laslty Grabke et al. [14] studied the oxidation behavior of chromium disilicide. They report
`that the intermetallic compound forms a protective chromia scale in the temperature range
`from 900 to 1050 "C. They note the absence of selective oxidation of silicon even though
`the activity of chromium in the substrate is low and the silicon activity and concentration are
`high.
`
`Table I. - Free energy of formation of various oxides at 1500 K.
`
`Oxide
`Si02
`Tan05
`Nb205
`Moo3
`Cr203
`
`-AF kCal/g. atom 0
`73.15
`66.9
`59.98
`31.96
`59.53
`
`"Pest" oxidation first described by Fitzer [15] involves the disintegration of the material into
`small disilicide particles, covered by a loose, voluminous scale. The pest phenomenon occurs
`in many intermetallic compounds over a wide temperature range that is specific for each
`material. The reaction has been studied and the review by Krier [16] presents a number of
`observations. It is suggested that the reaction is caused by the contamination of the Si02 scale
`by Moo3 for the MoSi2 system and W03 for WSi2. Pest is found to occur in the temperature
`range in which the vapor pressure of the volatile oxide is insufficient for its vaporization from
`the SO2. The upper temperature limit for pest is approximately that at which the oxide has
`a vapor pressure of exp(-3) atm; 780 "C for Moo3 and 1288 "C for W03.
`The most extensive study of the pest reaction of MoSi2 was conducted by Berkowitz et al.
`[17, 181. Their investigation covered the temperature range of 450-650 " C . Failure due to
`pest is found in oxygen, but not in nitrogen, carbon dioxide, carbon monoxide or argon.
`Moreover the rate of disintegration is strongly dependent upon the oxygen partial pressure.
`The pest reaction has been eliminated in several disilicides by alloy additions. Zeitch [19]
`found that the addition of 5-15 pct MoGe2 stopped pest in MoSi2. In addition, high tempera-
`ture oxidation produced a Ge02-Si02 glass surface layer with a higher coefficient of thermal
`expansion and a lower viscosity than that of pure Si02. This significantly improved the cyclic
`oxidation behavior.
`Schlichting [20] reported that pest could be eliminated by keeping the porosity of sintered
`material to very low levels.
`
`GE-1012.007
`
`
`
`OXIDATION OF ADVANCED INTERMETALLIC COMPOUNDS
`
`373
`
`I n conclusion, intermetallic compounds offer unique and desirable properties and could
`very well replace many alloys in the near future provided that the promise of improved re-
`sistance to oxidation and corrosion can be realized.
`
`References
`
`[l] TIEN J.K. and DAVIDSON J.M., Stress Effects and the Oxidation of Metals. J. Cathcart
`Ed. (Metallurgical Society AIME, 1974) p.200.
`[2] WHITTLE D.F! and STRINGER J., Phil. Trans. R. Soc. London A 295 (1980) 309.
`[3] STRINGER J., Matex Sci. Eng. A 120 (1989) 129.
`[4] SMEGGIL J., FUNKENBUSH A. and BORNSTEIN N.S., Studies ofAdherent Oxide Scales,
`ONR Contracts N00014-82-C-0618 and N00014-85-C-0421 (1982-1987).
`[5] SMIALEK J., Proceedings of the Symposium on High Temperature Materials
`Chemistry-IV, Z. Munie, D. Cubicciotti and H. Tagawa Eds. (Electrochem. Soc. 1988)
`p. 241.
`[6] SMIALEK J., TUBBS B.K. and SMIALEK J., Corrosion and Particle Erosion at High
`Temperature Proceedings (TMS-AIME, 1989).
`[7] MEIER G.H., APPALONIA D., PERKINS R. and CHIANC K., Oxidation of High Temper-
`ature, Intermetallics, T. Grobstein and J. Doychak Eds. (TMS, 1988) 185.
`[8] WAGNER C., 2. Electrochem 63 (1 959) 772.
`[9] WELSCH G. and KAHVECI A.I., Oxidation High Temperature Intermetallics, T Grob-
`stein and J. Doychak Eds. (TMS, 1988) p.207.
`[lo] CHOUDHURY N.S., GRAHAM H.C. and HINZE J.W., Properties of High Temperature
`Alloys, Z.A. Foroulis and F.S. Petit Eds. (Electrochem. Soc., 1976) p.668.
`[l 11 HANAMURA T., IKEMATSU Y., MORIKAWA H., TANINO M. and TAKAMURA
`J., Pro-
`ceedings of the International Symposium on Intermetallic Compounds, (JIMIS-6)
`0. Izumi Ed. (Japan Institute of Metals, 1991) p.179.
`1121 IKEMATSU, HANAMURA Y.T., MORIKAWA H., TANINO M. and TAKAMURA J., Proceed-
`
`ings of the International Symposium on Intermetallic Compounds, 0. Izumi Ed.
`(Japan Institute of Metals, 1991) 191.
`[13] LAVENDEL H.W. and GRANT A., Elliot Trans A i m 239 (1967) 143.
`[14] GRABKE H.J. and BRUMM M., Oxidation of High Temperature Intermetallics, T.
`Grobstein and J. Doychak Eds. (TMS, 1988) p.245.
`1151 FITZER E., Plansee Proc. 2nd Seminar, ReutteITyrol (Permagon Press, London, 1955)
`p.56.
`[16] KRIER C.A., Coatings for the Protection of Refractory Metals from Oxidation, DMIC
`report 162 (Nov. 1961).
`
`(171 BERKOWITZ-MATI'UCK J.B., BLACKBURN P.E. and LEE D., Refractory Metals and Al-
`loys Part IV (Gorden and Breach London, 1969) p. 1062.
`J.B., ROSSETI M. and LEE D., Met. Trans. 1 (1970) 479.
`[18] BERKOWITZ-MA~TUCK
`[19] ZEITCH K., PhD Thesis Universitat Karlsruhle (1970).
`[20] SCHLICHTINC J., High Temp. High Pressure 10 (1978) 241.
`
`GE-1012.008