Cyclic oxidation behavior of β-NiAlDy alloys containing varying aluminum content at 1200 °C

Progress in Natural Science: Materials International 2012;22(4):311–317 Chinese Materials Research Society

Progress in Natural Science: Materials International www.elsevier.com/locate/pnsmi www.sciencedirect.com

ORIGINAL RESEARCH

Cyclic oxidation behavior of b-NiAlDy alloys containing varying aluminum content at 1200 1C Dongqing Lia, Lu Wanga,b, Hui Penga, Xiaoyu Zhaoa, Hongbo Guoa,n, Shengkai Gonga a

School of Materials Science and Engineering, Beihang University, No. 37, Xueyuan Road, Beijing 100191, China Shenyang Liming Aero-engine Group Corporation Ltd., AVIC, Shenyang 110043, China

b

Received 7 March 2012; accepted 6 May 2012 Available online 1 September 2012

KEYWORDS b-NiAl; Reactive elements; Dysprosium; Aluminum content; High-temperature oxidation

Abstract b-NiAlDy cast alloys containing varying aluminum content were prepared by arcmelting. The microstructures and cyclic oxidation behavior of the alloys at 1200 1C were investigated. Grain refinement was achieved by increasing aluminum content in the alloy, which is beneficial to selective oxidation. The Ni–55Al–0.1Dy alloy showed excellent cyclic oxidation resistance due to the formation of a continuous, dense and slow-growing oxide scale. In contrast to this, severe internal oxidation as well as large void formation at the scale/alloy interface occurred in the Ni–45Al–0.1Dy alloy. The aluminum content dependence of the reactive element effects in bNiAlDy was established that Dy doping strengthened the scale/alloy interface by pegging mechanism in high-aluminum alloys but accelerated internal oxidation in low-aluminum alloys during high-temperature exposure. & 2012 Chinese Materials Research Society. Production and hosting by Elsevier Ltd. All rights reserved.

1. n

Corresponding author. Tel.: þ86 10 8231 7117; fax: þ86 10 8233 8200. E-mail address: [email protected] (H. Guo). Peer review under responsibility of Chinese Materials Research Society.

Introduction

Thermal barrier coatings (TBCs) consisting of insulating ceramic top coat and oxidation-resistant metalic bond coat are one of the most promising high-temperature protective coatings [1–3]. More recently, b-NiAl is considered as a potential candidate material for high-temperature protection of underlying superalloy substrates and suitable bond coat in TBC system due to its high melting point, low density and good isothermal oxidation resistance [4–12]. During high-temperature oxidation, a stable Al2O3 scale is formed on NiAl which protects the alloy substrate

1002-0071 & 2012 Chinese Materials Research Society. Production and hosting by Elsevier Ltd. All rights reserved. http://dx.doi.org/10.1016/j.pnsc.2012.06.003

312

D. Li et al.

from further damage. However, the poor cyclic oxidation performance of NiAl associated with severe scale spallation is an obstacle to its application under extreme high-temperature conditions, especially above 1200 1C. Extensive efforts have been made to solve this problem. The addition of Pt group metals such as Pt, Pd and Rh has been demonstrated to significantly improve the cyclic oxidation resistance of NiAl [13–16]. Pt addition to the NiAl coating, typically about 6–8 at%, was substantially contributed to its excellent scale adhesion, while having no effect on scale growth rate and excessive cost of coating preparation limit the development of NiPtAl coatings [16]. Recently, much attention has been focused on adding reactive elements such as Hf, Zr and Y as well as their oxides dispersions into NiAl, which not only enhances the oxide scale adhesion but also slows down the scale growth rate [10,13,17–22]. Dysprosium, as a reactive element, was also doped in NiAl. Previous work [23–26] has shown that NiAl alloys and coatings with minor Dy exhibited improved cyclic oxidation resistance, which can be attributed to three aspects: (i) the addition of Dy prevented the surface rumpling of the oxide scale, (ii) the presence of Dy suppressed sulfur segregation and void growth at the scale/alloy interface and (iii) the pegs consisting of AlDyO3 core and outer alumina sheath developed deeply in the alloy or coating and improved the oxide scale adhesion. Note that the above mechanisms can only be applicable to stoichiometric NiAl. In engine service, continuous oxidation and interdiffusion lead to drastic depletion of aluminum from the bond coat in TBCs system, which will result in stoichiometry deviation in NiAl. However, whether the effect mechanisms of Dy will change correspondingly with the aluminum content still remain unclear. Therefore, it is necessary to clarify the aluminum content dependence of the effect of Dy and the effect of aluminum content on the oxidation behavior of the NiAlDy alloy. Model alloys containing various levels of Al, Pt and Hf have been studied previously [27]. Resulting from low aluminum content, fastgrowing Ni-rich oxides were formed in hypostoichiometric bNiAl, depriving the oxide scale continuity. Fortunately, this problem can be effectively solved by using higher amounts of Pt and Hf. In this work, a series of b-NiAlDy alloys were produced by arc-melting and casting and subsequently oxidized in air at 1200 1C. The microstructures and cyclic oxidation behavior of the alloys were investigated, to understand the effect of aluminum content on the oxide scale growth rate and oxide scale adhesion of the NiAlDy alloy. Efforts were also made to elucidate the aluminum content dependence of the effect of Dy on the cyclic oxidation resistance of NiAl.

2.

Cyclic oxidation tests were conducted in an air furnace equipped with an automation system that allows the samples moving in and out automatically. Each sample was hung in a pre-annealed Al2O3 crucible by Pt wire to obtain the mass gains and soaked at 1200 1C for 50 min in the furnace and cooled for 10 min by compressed air. The mass gains were measured by an electronic balance (Sartorious CPA 225D, Germany) with a precision of 104 g. Microstructures of the specimens were characterized by scanning electron microscopy (SEM, CS3400) and field emission-scanning electron microscopy (FE-SEM, Apollo 300) equipped with energy dispersive X-ray spectrum (EDXS) and back scattering electron (BSE) detector. The chemical compositions of the alloys and the oxides were determined by an electron probe micro-analyzer (EPMA, JXA-8100). For cross-sectional observation, the specimens after oxidation were embedded in resin, ground and finely polished.

3.

Results and discussion

3.1.

Microstructural observation of b-NiAlDy alloys

The chemical compositions of the as-annealed alloys are shown in Table 1, determined by EPMA. The contents of Al in the alloys were slightly different from those designed. Fig. 1(a)–(d) shows the cross-sectional optical micrographs of the as-annealed NiAlDy alloys with different Al content, respectively. The Ni–55Al–0.1Dy alloy revealed finer grains compared to Ni–45Al–0.1Dy, implying that the increase of Al content resulted in grain refinement. In our previous work, it has been shown that the solubility of Dy in b-NiAl showed a reduction with the increase of Al content [28]. More Dy-rich phases precipitated at the grain boundaries in high-aluminum alloys during alloy preparation. These Dy-rich phases inhibited grain boundary migration or acting as crystal nuclei, thus resulting in grain refinement of the alloys. Pint et al. [29] showed that a fine alloy grain size promoted protective scale formation by increasing the diffusivity of Al in the substrate as smaller grains provided more diffusion paths. A back-scattered electron cross-sectional image of grain boundaries in the as-annealed Ni–52Al–0.1Dy alloy is given in Fig. 2. The reactive element Dy mainly distributed along grain boundaries and within the grains in the form of DyNi2Al3 phase, which has been reported elsewhere [23]. Dy dissolved in the alloy substrate forming supersaturated solid solution just after arc-melting and casting, while most precipitated at the grain boundaries and within the grains during subsequent heat-treatment due to its extremely low solubility in b-NiAl.

Experimental Procedures

b-NiAlDy alloys with different Al content were used in this work. The nominal compositions were Ni–(55, 52, 50, 45)Al– 0.1Dy (in at%). All the alloys were prepared from Ni (purity 99.99 wt%), Al (purity 99.99 wt%) and Dy (purity 99.99 wt%) by arc-melting and casting in argon atmosphere and annealed in vacuum at 1150 1C for 8 h. Specimens with dimensions of 10  8  3 mm3 were cut from the center of the button-shaped alloys, polished to an 800-grit SiC finish and ultrasonically cleaned in alcohol and acetone prior to oxidation.

Table 1 Chemical compositions of the as-annealed alloys (in at%, EPMA data). Alloy

Ni (at%)

Al (at%)

Dy (at%)

Ni–55AlþDy Ni–52AlþDy Ni–50AlþDy Ni–45AlþDy

45.99 49.38 51.13 56.38

53.92 50.49 48.77 43.50

0.09 0.13 0.10 0.12

Cyclic oxidation behavior of b-NiAlDy alloys containing varying aluminum content at 12001C

313

Fig. 1 Optical micrographs of cross-sections of as-annealed NiAlDy alloys: (a) Ni–55Al–0.1Dy; (b) Ni–52Al–0.1Dy; (c) Ni–50Al–0.1Dy; and (d) Ni–45Al–0.1Dy.

To further analyze the cyclic oxidation kinetics, the mass gain data were fitted to a form of exponential equation as Dw ¼ ktn

Fig. 2 Back-scattered electron cross-sectional image of grain boundaries in the as-annealed Ni–52Al–0.1Dy alloy.

3.2.

Cyclic oxidation of b-NiAlDy alloys at 1200 1C

Fig. 3 shows the mass gains of the NiAlDy alloys during 1-h cyclic oxidation at 1200 1C. All the samples revealed rapid mass gain in the first 10 h oxidation and subsequently stepped into a steady-state oxidation stage except for the Ni–45Al– 0.1Dy alloy which kept a high oxidation rate even after 30 h oxidation and obtained a mass gain of 5.0 mg/cm2 after 145 thermal cycles. For the Ni–55Al–0.1Dy alloy, significant mass gain did not occur in cyclic testing. It exhibited the lowest mass gain of 1.5 mg/cm2 after 145, 1 h cycles. The Ni–55Al– 0.1Dy alloy surface was completely covered with Al2O3 scale in the shortest time during thermal exposure as finer alloy grains can improve selective oxidation [29], which reduced the initial oxidation stage and finally the mass gain (Fig. 3b).

where Dw is the mass gain per unit area, k is the oxidation rate constant, t is the thermal exposure time and n is the oxidation rate exponent. Results of the numerical fitting are listed in Table 2. The cyclic oxidation kinetics of the Ni–55Al–0.1Dy alloy followed a cubic relation, while the other alloys showed a square relation. It has been reported that the cubic relation represents better oxidation resistance [30]. The parabolic rate constants reflect a decreased tendency in oxide scale growth rate with the increase of aluminum content in NiAlDy alloys except for Ni–52Al–0.1Dy, which exhibited a slightly higher scale growth rate compared to Ni–50Al–0.1Dy. One possibility was that the surface layer of the alloy was poor in Al due to compositional heterogeneity induced by alloy preparation, thereby increasing the final mass gain of the alloy. In addition, the experimentally determined content of Dy in Ni–52Al– 0.1Dy was 0.13 at%, slightly higher than those in the other alloys. As Dy over-doping can lead to accelerated oxidation [23], the little more amount of Dy might be also deleterious to the oxidation behavior of the alloy. The top views of surface morphologies of the NiAlDy alloys after 2-h cyclic oxidation at 1200 1C are given in Fig. 4. All the oxide grains exhibited a granular structure, indicating that the addition of Dy eliminated the ridge structure of Al2O3 scale, which is a typical morphology for the oxides grown on the Al2O3-forming alloy surface [31]. Pint et al. [31–33] proposed that the formation of ridge structure was attributed to the outward circuit diffusion of Al and y–a phase transformation during thermal exposure. In our previous work, it has been shown that the segregation of Dy at the oxide grain

314

D. Li et al.

Fig. 3 Mass gain curves of the NiAlDy alloys during 1-h cyclic oxidation at 1200 1C.

Table 2 Results of the numerical fitting using an exponential equation form. Alloy

n

k

Ni–55Al–0.1Dy Ni–52Al–0.1Dy Ni–50Al–0.1Dy Ni–45Al–0.1Dy

0.35 0.55 0.48 0.56

2.96 6.34 4.82 6.43

boundaries suppressed the outward diffusion of Al and allowed the inward diffusion of oxygen to become predominant. Dy doping also affected the nucleation and growth of Al2O3, thus inhibiting the formation of ridge structure [34]. An even and dense oxide scale with the grain size of 0.5 mm was observed on the Ni–55Al–0.1Dy alloy after 2-h cyclic oxidation (Fig. 4a). Several cracks existed in the Al2O3 scale, which might be originated from volume contraction of the oxide film caused by the transformation of y- to a-Al2O3. The oxide grain size of Ni–50Al–0.1Dy was larger than that of Ni–55Al– 0.1Dy and squeezing deformation of the oxide grains induced by growth stress can be seen in Fig. 4b, which reveals a higher oxide growth rate compared to Ni–55Al–0.1Dy. This is consistent with the result of mass gain of Ni–50Al–0.1Dy in Fig. 3. For the Ni–45Al–0.1Dy alloy, a porous oxide scale was formed (Fig. 4c). The visible pores between the oxide grains allowed oxygen diffusing through them rapidly and attacking the alloy substrate directly, thus resulting in accelerated oxidation. In addition, no cracks were observed in Fig. 4(b)– (c), indicating a slower phase transformation rate of the Al2O3 scale compared to Ni–55Al–0.1Dy. It has been known that the tensile stress resulting from the volume contraction might be large enough to arouse microcracks in the oxide scale if the y–a phase transformation developed quickly and new oxides would grow fast along the cracking surface [35]. Despite of this, as y-Al2O3 had a higher growth rate than a, its rapid transformation to a guaranteed a slow growth rate for subsequently formed oxides on the NiAlDy alloy. The cyclic oxidation kinetics suggest that the latter was prevalent. The surface morphologies of the alloys after 100 thermal cycles at 1200 1C were examined. A small amount of spalls of the oxide scale was observed on Ni–55Al–0.1Dy and Ni–50Al– 0.1Dy, as shown in Fig. 5(a) and (b). The oxide scale on the Ni–45Al–0.1Dy alloy revealed a mutilated surface (Fig. 5c). Severe spallation of the oxide scale occurred and the substrate

alloy was directly exposed to the air after 100-h cyclic oxidation. Only some chippings remained on the alloy surface. The formation of voids at the scale/alloy interface weakens the interfacial bonding between the alloy and the oxide scale, which is mainly responsible for the poor scale adhesion. Although the interfacial void growth is a considerable problem for the alumina-forming alloys, its mechanistic explanation is still not clarified. One possibility is that these voids are triggered by Kirkendall effect and grow up during thermal exposure [36]. At higher magnification (Fig. 5(d)–(f)), several voids can be seen in the areas of each alloy surface where the oxide scale has already spalled and their average size on each alloy surface was measured. Results revealed an increased tendency in void size with the decrease of Al content in the NiAlDy alloys. Previous studies have shown that the diffusivity of Ni and Al in NiAl is a function of alloy composition [37]. In stoichiometric NiAl, Al atoms diffuse to the alloy surface and react with oxygen to form alumina scale during initial oxidation, leaving Al vacancies at the scale/alloy interface. With the consumption of Al, the interface is poor in Al and rich in Ni compared to the bulk. As Al atoms diffuse slower than Ni in the bulk due to DAloDNi (DAl/ DNio1), Ni vacancies are also formed at the scale/alloy interface. These two kinds of atomic vacancies nucleate and develop into voids when the alloy is thermally exposed for extended periods. In nickel-rich NiAl, Al diffuses even slower than Ni (DAl/ DNio1/3), which means that more Ni vacancies can be left at the scale/alloy interface and larger voids can be formed consequently. In contrast to this, if aluminum-rich NiAl is oxidized, diffusion of Al will become predominant (DAl/ DNi41). There will be only Al vacancies existing at the interface and void nucleation and growth will be greatly suppressed. Fig. 6 shows the secondary-electron cross-sectional images of the specimens after desired oxidation time at 1200 1C. A dense alumina scale was intimately bonded to the Ni–52Al– 0.1Dy alloy substrate. Similar cases also occurred on Ni– 55Al–0.1Dy and Ni–50Al–0.1Dy. Compared to Ni–52Al– 0.1Dy, a much thinner Al2O3 scale was formed on the Ni– 45Al–0.1Dy alloy and in some areas even when no scale was observed (the Ni-layer was electroplated before SEM observation for oxide scale protection). This implies an excess spallation for the oxide scale. In addition, severe internal oxidation was observed in the Ni–45Al–0.1Dy alloy which did not occur in Ni–52Al–0.1Dy. It seems that the Ni–45Al–0.1Dy alloy exhibited a much higher scale growth rate and weaker scale adhesion relative to

Cyclic oxidation behavior of b-NiAlDy alloys containing varying aluminum content at 12001C

315

Fig. 4 Secondary electron plan-view images of the NiAlDy alloys after 2-h cyclic oxidation at 1200 1C: (a) Ni–55Al–0.1Dy; (b) Ni–50Al–0.1Dy; and (c) Ni–45Al–0.1Dy.

Fig. 5 Secondary electron plan-view images of the NiAlDy alloys after 100-h cyclic oxidation at 1200 1C: (a,d) Ni–55Al–0.1Dy; (b,e) Ni–50Al–0.1Dy; and (c,f) Ni–45Al–0.1Dy.

the other alloys. However, only the decrease of Al content cannot be responsible for the abnormally poor oxidation performance of Ni–45Al–0.1Dy. There must exist some other

factors affecting its oxidation resistance. One of them is that the effect mechanisms of Dy on the oxidation resistance of NiAl might change with the decrease of Al content. At

316

D. Li et al.

Fig. 6 Back-scattered electron cross-sectional images of the NiAlDy alloys after long-term cyclic oxidation at 1200 1C: (a) Ni–52Al–0.1Dy, 200 h; and (b) Ni–45Al–0.1Dy, 100 h.

high-temperature, Al ions in the alloy surface layer were rapidly consumed due to oxide scale formation and those in the bulk diffused to the alloy surface layer constantly. In highaluminum alloys (Al content is more than 50 at%), the reactive element Dy which distributed at the grain boundaries near the scale/alloy interface were oxidized to form AlDyO3 phases. These AlDyO3 phases worked as ‘‘oxide pegs’’ to improve scale adhesion which consisted of AlDyO3 core and an outer alumina sheath and were defined as protrusions of secondary oxide inclusions, as shown in Fig. 6a. Note that internal oxidation did not occur in this alloy even after 200-h cyclic oxidation. The sufficient Al and fine alloy grains guaranteed the rapid formation of dense alumina scale and oxygen cannot diffuse into the bulk. While in low-aluminum alloys (Al content is less than 50 at%), the situation may be different. As the oxide scale on the alloy was loose and nonprotective (Fig. 4c), oxygen can pass through the scale and diffuse into the bulk along grain boundaries. Meanwhile, since the oxygen affinity of Dy is much higher than that of Al, Dyrich compounds at grain boundaries and precipitates within the grains were oxidized preferentially and rapidly, which served to the aggravation of internal oxidation. A conclusion might be drawn as to the content of Al in the b-NiAlDy alloy should be further enhanced to acquire better cyclic oxidation resistance, since the Ni–55Al–0.1Dy alloy gave an optimal performance in the present work. The refutation may lie on two aspects. On the one hand, the phase constitution of the b-NiAlDy alloy and associated oxidation mechanism will be changed with the composition range of Al in b-NiAl is 4055 at% according to the Ni–Al phase diagram. On the other hand, the stoichiometric NiAl has the highest melting point (1638 1C) in b-NiAl alloys, which is crucial to its application as high-temperature protective coatings, and the deviation of Al content will give rise to a sharp drop of the melting point of b-NiAl. Consequently, the further enhancement of Al content in the b-NiAlDy alloy is not acceptable. 4.

Conclusions

(1) Increasing Al content, a finer alloy grain size was obtained due to the formation of more Dy-rich precipitates. As a result, protective scale formation was promoted in Ni–55Al–0.1Dy. (2) The Ni–55Al–0.1Dy alloy exhibited the best cyclic oxidation resistance due to the formation of a

continuous, dense and slow-growing oxide scale, while Ni–45Al–0.1Dy revealed an abnormally high scale growth rate and weak scale adhesion relative to the other alloys. (3) The interfacial void size was reduced by increasing the Al content in NiAlDy, which significantly improved the interfacial bonding between the oxide scale and the alloy substrate. (4) The effect mechanisms of Dy on the cyclic oxidation resistance of NiAl were greatly affected by the concentration of Al in the alloy. The addition of minor Dy improved the oxide scale adhesion by pegging mechanism in high-aluminum alloys but accelerated internal oxidation in low-aluminum alloys during long-term oxidation. Acknowledgments This research is sponsored by National Nature Science Foundations of China (NSFC, Grant Nos.50731001 and 51071013) and National Basic Research Program (973 Program) of China under Grant No.2010CB631200.

References [1] N. Czech, H. Fietzek, M. Juez-Lorenzo, V. Kolarik, W. Stamm, Studies of the bond-coat oxidation and phase structure of TBCs, Surface & Coatings Technology 113 (1999) 147–164. [2] N.P. Padture, M. Gell, E.H. Jordan, Thermal barrier coatings for gas-turbine engine applications, Science 296 (2002) 280–284. [3] M.J. Pomeroy, Coatings for gas turbine materials and long term stability issues, Materials & Design 26 (2005) 223–231. [4] R. Sivakumar, B.L. Mordike, High temperature coatings for gas turbine blades: a review, Surface & Coatings Technology 37 (1989) 139–160. [5] D.B. Miracle, The physics and mechanical properties of NiAl, Acta Metallurgica et Materialia 41 (1993) 649–684. [6] I. Baker, A review of the mechanical properties of B2 compounds, Materials Science and Engineering A 192/193 (1995) 1–13. [7] Z. Yu, D.D. Hass, H.N.G. Wadley, NiAl bond coats made by a directed vapour deposition approach, Materials Science and Engineering A 394 (2005) 43–52. [8] H.F. Li, S.F. Tao, Z.H. Zhou, S.K. Gong, Element diffusion during fabrication of EB-PVD NiAl coating and its 1100 1C isothermal oxidation behaviour (II), Surface & Coatings Technology 201 (2007) 6589–6592.

Cyclic oxidation behavior of b-NiAlDy alloys containing varying aluminum content at 12001C [9] H.F. Li, H.B. Guo, S.K. Gong, Failure mechanism of EB-PVD thermal barrier coatings on NiAl substrate, Transactions of Nonferrous Metals Society of China 4 (2007) 811–815. [10] H.B. Guo, L.D. Sun, H.F. Li, S.K. Gong, High temperature oxidation behavior of hafnium modified NiAl bond coat in EBPVD thermal barrier coating system, Thin Solid Films 516 (2008) 5732–5735. [11] H.B. Guo, Y.J. Cui, H. Peng, S.K. Gong, Improved cyclic oxidation resistance of electron beam physical vapor deposited nano-oxide dispersed b-NiAl coatings for Hf-containing superalloy, Corrosion Science 52 (4) (2010) 1440–1446. [12] Q. Wu, R.B. Yang, Y.X. Wu, S.S. Li, Y. Ma, S.K. Gong, A comparative study of four modified Al coatings on Ni3Al-based single crystal superalloy, Progress in Natural Science: Materials International 21 (2011) 496–505. [13] C. Leyens, B.A. Pint, I.G. Wright, Effect of composition on the oxidation and hot corrosion resistance of NiAl doped with precious metals, Surface & Coatings Technology 133/134 (2000) 15–22. [14] P.Y. Hou, K.F. McCarty, Surface and interface segregation in b-NiAl with and without Pt addition, Scripta Materialia 54 (2006) 937–941. [15] J.A. Haynes, B.A. Pint, Y. Zhang, I.G. Wright, Comparison of the cyclic oxidation behavior of b-NiAl, b-NiPtAl and g-g0 NiPtAl coatings on various superalloys, Surface & Coatings Technology 202 (2007) 730–734. [16] P.Y. Hou, V.K. Tolpygo, Examination of the platinum effect on the oxidation behavior of nickel–aluminide coatings, Surface & Coatings Technology 202 (2007) 623–627. [17] C.A. Barrett, Effect of 0.1 at% zirconium on the cyclic oxidation resistance of b-NiAl, Oxidation of Metals 30 (5/6) (1988) 361–390. [18] B.A. Pint, M. Treska, L.W. Hobbst, The effect of various oxide dispersions on the phase composition and morphology of Al2O3 scales grown on b-NiAl, Oxidation of Metals 47 (1/2) (1997) 1–20. [19] B.A. Pint, The oxidation behavior of oxide-dispersed b-NiAl: I. short-term performance at 1200 1C, Oxidation of Metals 49 (5/6) (1998) 531–559. [20] B.A. Pint, I.G. Wright, W.Y. Lee, Y. Zhang, K. PruO ¨ ˆ ner, K.B. Alexander, Substrate and bond coat compositions: factor affecting alumina scale adhesion, Materials Science and Engineering A 245 (1998) 201–211. [21] J.A. Haynes, B.A. Pint, K.L. More, Y. Zhang, I.G. Wright, Influence of sulfur, platinum, and hafnium on the oxidation behavior of CVD NiAl bond coatings, Oxidation of Metals 58 (5/ 6) (2002) 513–544. [22] X. Song, L. Wang, Y. Liu, H.P. Ma, Effects of temperature and rare earth content on oxidation resistance of Ni-based superalloy,

[23]

[24]

[25]

[26]

[27]

[28]

[29]

[30] [31]

[32] [33]

[34]

[35]

[36]

[37]

317

Progress in Natural Science: Materials International 21 (2011) 227–235. H.B. Guo, X.Y. Wang, J. Li, S.X. Wang, S.K. Gong, Effects of Dy on cyclic oxidation resistance of NiAl alloy, Transactions of Nonferrous Metals Society of China 19 (5) (2009) 1185–1189. H.B. Guo, X.Y. Wang, S.K. Gong, Effects of Dy on the microstructure and spallation failure of the alumina scales grown on NiAl, International Journal of Modern Physics B 24 (15/16) (2010) 3149–3154. H.B. Guo, S.X. Wang, X.Y. Wang, S.K. Gong, Cyclic oxidation behaviors of EB-PVD Dy-doped b-NiAl coatings at 1100 1C, International Journal of Modern Physics B 24 (15/16) (2010) 3143–3148. G.Y. Zhang, H. Zhang, J.T. Guo, Improvement of cyclic oxidation resistance of a NiAl-based alloy modified by Dy, Surface & Coatings Technology 201 (2006) 2270–2275. B.A. Pint, The role of chemical composition on the oxidation performance of aluminide coatings, Surface & Coatings Technology (188/189) (2004) 71–78. X.Y. Zhao, The effect of chemical composition on the oxidation behavior of b-NiAlDy alloys at elevated temperature, Master Thesis, Beihang University, 2011. B.A. Pint, J. Leibowitz, J.H. Devan, The effect of an oxide dispersion on the critical Al content in Fe-Al alloys, Oxidation of Metals 51 (1/2) (1999) 181–197. J.M. Xiao, C.N. Cao, Principles of materials corrosion, Chemical Engineering Press, Beijing, 2002. B.A. Pint, J.R. Martin, L.W. Hobbs, 18O/SIMS characterization of the growth mechanism of doped and undoped a-Al2O3, Oxidation of Metals 39 (3/4) (1993) 167–195. J.L. Smialek, Oxide morphology and spalling model for NiAl, Metallurgical and Materials Transactions A 9 (1978) 309–320. J. Doychak, J.L. Smialek, T.E. Mitchell, Transient oxidation of single crystal b-NiAl, Metallurgical and Materials Transactions A 20 (1989) 499–518. H.B. Guo, T. Zhang, S.X. Wang, S.K. Gong, Effect of Dy on oxide scale adhesion of NiAl coatings at 1200 1C, Corrosion Science 53 (2011) 2228–2232. D.M. Lipkin, D.R. Clarke, M. Hollatz, M. Bobeth, W. Pompe, Stress development in alumina scales formed upon oxidation of (111) NiAl single crystals, Corrosion Science 39 (1997) 231–242. H. Singh, D. Puri, S. Prakash, Some studies on hot corrosion performance of plasma sprayed coatings on a Fe-based superalloy, Surface & Coatings Technology 192 (2005) 27–38. M.W. Brumm, H.J. Grabke, Oxidation behavior of NiAl-b. Cavity formation beneath the oxide scale on NiAl of different stoichiometries, Corrosion Science 34 (1993) 547–561.