UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

MHigh temperature oxidation behaviour of the single crystal superalloy CMSX-10 Hegde, Subray R. 2006

Your browser doesn't seem to have a PDF viewer, please download the PDF to view this item.

Item Metadata

Download

Media
831-ubc_2006-0049.pdf [ 8.15MB ]
Metadata
JSON: 831-1.0078654.json
JSON-LD: 831-1.0078654-ld.json
RDF/XML (Pretty): 831-1.0078654-rdf.xml
RDF/JSON: 831-1.0078654-rdf.json
Turtle: 831-1.0078654-turtle.txt
N-Triples: 831-1.0078654-rdf-ntriples.txt
Original Record: 831-1.0078654-source.json
Full Text
831-1.0078654-fulltext.txt
Citation
831-1.0078654.ris

Full Text

HIGH T E M P E R A T U R E OXIDATION BEHAVIOUR OF T H E SINGLE C R S Y T A L SUPERALLOY CMSX-10 B y SUBRAY R HEGDE M.Sc.(Engg), Indian Institute of Science, 2003 A THESIS SUBMITTED IN PARTIAL FULFILMENT OF THE REQUIREMENT FOR THE DEGREE OF MASTER OF APPLIED SCIENCE in THE F A C U L T Y OF G R A D U A T E STUDIES ( M A T E R I A L S E N G I N E E R I N G ) T H E U N I V E R I S I T Y OF B R I T I S H C O L U M B I A D E C E M B E R 2005 © Subray R Hegde, 2005 Abstract The gas turbine industry has sought to increase the operating temperature of turbines so as to achieve higher efficiency. The compositions o f superalloys used in the critical components of engines are changing to improve their high temperature capabilities. Historically, efforts in improving the performance of superalloys, in general, have been focused only on the mechanical properties; predominantly, the creep resistance. Surface protection, though an important issue, has received less attention. The recent trend in the superalloy design is to reduce the chromium content and to increase additions of refractory metals such as Ta and Re. A s a result, mechanical strength o f the alloy has been considerably improved with little attention to corrosion resistance. The present work studies the high temperature oxidation behaviour of a 3 r d generation single crystal superalloy, C M S X - 1 0 . In this work, a series of isothermal oxidation tests of the alloy were conducted in air using a thermo gravimetric analyser at temperatures ranging from 800 °C to 1250 °C. The mass gain versus time curve shows parabolic growth at short exposure, followed by sub-parabolic growth at longer times. The alloy exhibits a transition from parabolic oxidation to sub-parabolic oxidation behaviour at all temperatures. The mass gain versus temperature curve shows anomalous behaviour. Characterisation of the oxidised specimens revealed that this anomaly is attributable to the diffusion driven morphological changes of spinel and alumina phases in the subsurface. Chromium oxide (C^Os) , which is considered to be a diffusion barrier, does not form in the oxidation of C M S X - 1 0 . However, the alloy forms continuous layers of spinel, NiAl204 and alumina when oxidised at temperatures above 1000 °C. Theses layers may act as a diffusion barrier at a later stage. i i Table of Contents Page Abstract i i Table of Contents i i i List of Tables v List of Figures v i Acknowledgement ix Chapter 1 - Introduction 1 1.1 Scope of the present work 5 Chapter 2 - Literature review 6 Chapter 3 - Experimental methods 10 3.1 Materials characterisation 10 3.2 Isothermal oxidation 14 3.3 Characterisation of the oxidised specimens 15 Chapter 4 - Results 17 4.1 Thermogravimetry 17 4.2 Microstructure 23 4.3 Energy dispersive spectroscopy 27 4.4 Elemental X - R a y mapping 33 4.5 X - R a y diffraction 38 i i i Chapter 5 - Discussion 42 5.1 The mechanism of oxidation in C M S X - 1 0 46 5.2 Loss of rhenium 54 5.3 Uneven oxidation 54 5.4 The poor oxidation resistance of C M S X - 1 0 55 Chapter 6 — Conclusions 56 6.1 Suggestions for future work 56 References 57 iv List of Tables Page Table I. Chemical compositions (in wt %) of a few single crystal 3 nickel base superalloys. Table II. Results of the isothermal oxidation experiments on 17 C M S X - 1 0 for 100 h. Table III. A list of the phases involved in the oxidation of C M S X - 1 0 . 41 Table IV. A list of the rate constant and the critical time at various temperatures. 43 Table V . A comparison of the parabolic rate constants obtained from 45 the present work and those found in the literature List of Figures Page Figure 1. A diffractometer trace showing the major reflections from 11 a C M S X - 1 0 specimen. Figure 2. A Scanning Electron Micrograph of a specimen showing the uniformly 12 distributed cuboids of y ' in the matrix of y phase. Figure 3. A Bright Field T E M image showing the dislocations and the 12 cuboids of y' aligned along a particular [001] direction. Figure 4. A selected area diffraction pattern of the C M S X - 1 0 showing the 13 regular F C C pattern from both y and y ' phases with superlattice reflections from y ' . Figure 5. A schematic of the experimental set up on a T G A . 15 Figure 6. A plot of mass gain versus temperature after 100 h of 18 isothermal oxidation. Figure 7. A plot of mass gain versus temperature of specimens with a thickness 19 of approximately 0.85 mm. Figure 8. The plot of mass gain versus temperature of specimens with the thickness 19 of approximately 1.4 mm. Figure 9. The mass gain versus time curves for specimens with the 21 thickness of about 0.85 mm at temperatures ranging from 800 °C to 1100 °C. Figure 10. The mass gain versus time curves for specimens with the 21 thickness of about 0.85 mm at temperatures ranging from 1100 °C to 1250 °C. Figure 11. The mass gain versus time curves for specimens with the thickness of 22 about 1.4mm at temperatures ranging from 850 °C to 1125 °C. Figure 12. The mass gain versus time curves for specimens with the thickness of 22 about 1.4 mm at temperatures ranging from 1125 °C to 1185 °C. Figure 13. A plot of mass gain rate versus temperature after 50 h of isothermal 23 oxidation. v i Figure 14. A scanning electron micrograph o f the transverse section o f 24 a specimen oxidised at 800 °C for 100 h. Figure 15. A sectional scanning electron micrograph of the transverse section 24 of a specimen oxidised at 900 °C for 100 h. Figure 16. A scanning electron micrograph of the transverse section of 25 a specimen oxidised at 1000 °C for 100 h. Figure 17. A scanning electron micrograph of the transverse section of 25 a specimen oxidised at 1100 °C for 100 h. Figure 18. A scanning electron micrograph of the transverse section of 26 a specimen oxidised at 1155 °C for 100 h. Figure 19. A scanning electron micrograph of the transverse section of 26 a specimen oxidised at 1250 °C for 100 h. Figure 20. The chemical composition profile of a specimen oxidised 29 at 900 °Cfor 100 h. Figure 21. The chemical composition profile of a specimen oxidised 30 at 1100 °C for 100 h. Figure 22. The chemical composition profile of a specimen oxidised 31 at 1155 °C for 100 h. Figure 23. The chemical composition profile of a specimen oxidised 32 at 1250 °C for 100 h. Figure 24. The elemental X-ray maps of N i , A l and O of a specimen 33 oxidised at 800 °C. Figure 25. The elemental X-ray maps of N i , A l and O of a specimen 34 oxidised at 900 °C. Figure 26. The elemental X-ray maps of N i , A l and O of a specimen 34 oxidised at 1000 °C. Figure 27. The elemental X-ray maps of N i , A l and O of a specimen 35 oxidised at 1100 °C. Figure 28. The elemental X-ray maps o f N i , A l , O, Cr, Co, Ta, W , Re and H f 36 of a specimen oxidised at 1155 °C. v i i Figure 29. The X R D profiles of a specimen oxidised at 900 °C for 100 h; 39 (A) as oxidised specimen, (B) after the 1 s t round of grinding and (C) after the 2 n d round of grinding. Figure 30. The X R D profile of a specimen oxidised at 1000 °C for 100 h. 40 Figure 31. The X R D profile of a specimen oxidised at 1250 °C for 100 h. 40 Figure 32. A plot o f unit mass gain versus t ime 1 7 2 at various temperatures 42 Figure 33. A plot of mass gain versus temperature of C M S X - 1 0 , after 3 h 44 of isothermal oxidation. Figure 34. A plot o f mass gain versus temperature after 20 h o f isothermal oxidation 44 Figure 35. Binary phase diagram of N i - A l system showing the shifts in the 48 equilibrium positions of the advancing front of the O R D at various temperatures. Figure 36. A plot of Z N i / (ENi + EA1) versus depth of the specimens 48 oxidised for 100 h. Figure 37. A plot of the thicknesses of N i O and F S Z versus temperature 50 after 100 h of isothermal oxidation. Figure 38. A plot of the thicknesses of the spinel and the alumina layers 50 versus temperature after 100 h of isothermal oxidation. Figure 39. A plot of unit mass gain versus time of (a) a fresh sample oxidised at 51 800 °C for 100 h (b) a sample preoxidised at 1000 °C for 100 h followed by isothermal oxidation at 800 °C for 100 h. Figure 40. The mass gain curve of a high temperature - low temperature 53 oxidation experiment along with the time-temperature profile. Figure 41. The mass gain curve of a low temperature-high temperature 53 oxidation experiment along with the time-temperature profile. Figure 42. A scanning electron micrograph showing uneven oxidation 55 at the curved surface of a specimen oxidised at 850 °C for 100 h. v i i i Acknowledgement I thank my supervisors Dr. A inu l Akhtar and Dr. Roger C. Reed for their support and guidance in completing this work successfully. The T G A used in this work belongs to Dr. Olivera Kesler, my sincere thanks to her. I gratefully acknowledge the financial support received from N S E R C . Thanks are due to Rudy Cardeno and Mary Mager for the technical assistance. The support from my family and friends during my arrival and stay in Vancouver was invaluable, I thank them all. The cooperation and the support received from my wife, Raji was extraordinary, I dedicate this thesis to her. ix Chapter 1 Introduction The quest of gas turbine designers and manufacturers to raise the turbine inlet temperature for achieving higher efficiencies is ever increasing. This is the driving force for the development of and subsequent improvement of superalloys as the materials for the high temperature structures. The increase in turbine inlet temperature for commercial aircraft has the advantage of higher fuel efficiency. In military aircraft, increased inlet temperature leads to higher thrust to weight ratios. The increase in inlet temperature o f the gas turbine engines is limited by the material capabilities of parts such as combustor liners, vanes and turbine blades which are in contact with hot fluids. While ceramics have some o f the desired properties, such as high melting point, excellent resistance to corrosion and high temperature oxidation, they are not as reliable in terms o f mechanical properties when compared to metals. Also , it is often difficult to fabricate ceramic materials into intricate engine components having complicated cooling channels. Though some advantages of ceramics, such as low thermal conductivity and high thermal stability can be accessed with thermal barrier coatings (TBCs) on hot section components, the base materials on which these coatings are deposited are metallic superalloys. Hence, in spite of their high cost, superalloys continue to be an inevitable material system in the gas turbine industry. Superalloys are defined as the alloys based on Group VIIIA-base elements (of the periodic table) for elevated temperature service, which demonstrate combined mechanical strength and surface stability (1). They are classified into three categories; nickel-base, cobalt-base and iron-base superalloys. The nickel base alloys are the most developed and are widely used. In addition to aircraft, marine and land based gas turbines, superalloys find applications in space vehicles, rocket engines, nuclear reactors, submarines and steam power plants. 1 A separate category of superalloys called 'corrosion resistant superalloys' are used in chemical industries in place of stainless steel. The nickel base superalloys are sometimes called the daughters of stainless steel as their primary face centered cubic austenitic structure, which can dissolve large quantities of a wide range of alloying additions, were developed from stainless steel. Superalloys possess high strength and corrosion resistance even at elevated temperatures and can be used above 85% of their melting temperatures. Though solid solution strengthening of the austenitic (y) matrix is an important contributor, the coherent precipitate Ni3(Al,Ti) is the main strengthening phase (y'). There are minor strengthening phases such as Ni3(Al,Ti ,Nb) (y"), metal carbides (M23C6, M&C) and borides. They act as grain boundary strengtheners in polycrystalline superalloys and often impart brittleness and hence are avoided in modern single crystal nickel base superalloys. The present trend is to have a coherent y-y' dual phase microstructure with about 75% by volume of y' precipitates in the form of cuboids embedded uniformly in the matrix of y. Modern superalloys consist of more than 12 critical alloying additions. Table I gives the chemical compositions of selected nickel base single crystal superalloys. There is a trend o f an increased rhenium content in new alloys for improving the creep and fatigue properties. However, there is a trend to decrease the chromium and titanium contents o f the alloys in order to avoid the precipitation of undesirable phases called T C P s (Topologically close packed phases). 2 Table I. Chemical compositions (in wt %) of a few single crystal nickel base superalloys. Cr Co Mo W Ta V Nb Al Ti Hf Re Ru Ni 1 s t - Generation PWA1480 10 5 0 4 12 5 1.5 Bal Rene N4 9 8 2 6 4 0.5 3.7 4 .2 Bal SRR99 8 5 0 10 3 5.5 2.2 Bal RR2000 10 15 3 0 0 1 5.5 4 Bal CMSX2 8 5 0.6 8 6 5.6 1 Bal CMSX6 10 5 3 0 2 4.8 4.7 0 .1 Bal 2 n d - Generation 0 Bal PWA1484 5 10 2 6 9 5.6 0 0.1 3 Bal Rene N5 7 8 2 5 7 6.2 0 0.2 3 Bal CMSX4 7 9 0.6 6 7 5.6 1 0.1 3 Bal 3 r d - Generation Rene N6 4 .2 12.5 1.4 6 7.2 0.1 5.75 0 0.15 5.4 Bal CMSX10 2 3 0.4 5 8 5.7 0.2 0.03 6 Bal TMS75 3 12 2 6 6 6 0 0.1 5 Bal 4 t h - Generation TMS-162 2.9 5.8 3.9 5.8 5.6 5.8 0.09 4 .9 6 Bal MC-NG 4 0.2 1 5 5 6 0.5 0.1 4 4 Bal Note: 1. P W A 1480 and PWA1484 are the trade marks of Pratt and Whitney, U S A . 2. Rene N4, Rene N5 and Rene N6 are the trade marks of General Electric, U S A . 3. SRR99 and RR2000 are the trade marks of Roll-Royce, U K . 4. C M S X 2 , C M S X 4 , C M S X 6 and C M S X 10 are the trade marks of Cannon-Muskegon Corporation, U S A . 5. TMS-75 and TMS-162 are the trade marks of the National Institute o f Materials Science, Japan. 6. M C - N G is the trade mark of Snecma and Turbomeca, France. 3 It is understood that alloy design and microstructural engineering are the main contributors to the high temperature capabilities of superalloys. However, production o f these alloys with accurate compositional control was made possible by processes such as Electroslag Refining (ESR) and Vacuum Arc Melting ( V A M ) . The ambition of the turbine designers to achieve greater weight reduction by reducing the wall thickness and by incorporating complicated 'serpentine' cooling tunnels was materialised by the investment casting technique. Also , the improvements in the casting technology o f the turbine blades from the polycrystalline casting via directional solidification to the single crystal growth helped to increase the creep resistance of superalloy. Coating technology has also contributed to the use of superalloy. The pack-cementation process or the chemical vapor deposition technique ( C V D ) is generally used to develop a nickel-aluminide (NiAl ) layer on the superalloy blades for the surface protection. The latest addition o f coatings on superalloys is the ceramic thermal barrier coatings (TBCs) . There has been a tremendous improvement in the high temperature capability o f superalloys over several decades. However, these developments were focused on one particular aspect, i.e. high temperature mechanical properties, predominantly creep resistance. The oxidation/corrosion study o f superalloys has received little attention though it is very important. A s a consequence, a few modern superalloys, in spite of possessing much better creep properties failed to replace their previous generation alloys because of their poor corrosion properties. Hence, the study of high temperature corrosion/oxidation is gaining importance in the design of modern superalloys. Oxidation of a commercial superalloy is the subject of this thesis. 4 1.1 Scope of the present work: This thesis aims to study the high temperature oxidation behaviour of a 3 r d generation superalloy, C M S X - 1 0 . The alloy is characterised by its higher rhenium content and lower chromium content when compared to its 2 n d generation counterpart, C M S X - 4. A s a consequence, when compared with its previous generation alloy, C M S X - 1 0 exhibits poor oxidation/corrosion resistance. The reason for the poor oxidation/corrosion properties of the alloys with reduced chromium content is not well understood so far. The present work conducts a comprehensive isothermal oxidation study of C M S X - 1 0 over a wide temperature range of 800 °C to 1250 °C. 5 Chapter 2 Literature Review A s the oxidation o f N i is industrially relevant, its mechanism has been investigated over almost a century (2). The study o f transition metal-oxide systems including N i - O became a subject of intense research in the 1950s during which Wagner derived the famous parabolic oxide growth kinetics (3). Subsequently, he found the phenomenon of internal oxidation in alloys and modelled the transition from internal to external oxidation (4). The investigations on high temperature oxidation of nickel conducted by K o i et al (5) revealed that N i O is a metal deficient p-type semiconductor. Atkinson & Taylor (6,7) conducted a comprehensive investigation on the diffusion mechanisms in N i O . They distinguished between the bulk diffusion, grain boundary diffusion & dislocation diffusion mechanisms and deduced empirical expressions for the diffusion coefficient in each case. K i m and Hobbs (8) found that bulk diffusion is predominant during high temperature oxidation and short circuit diffusion at low temperatures. Recently, in a comprehensive investigation, Haugsrud (9) confirmed that pure N i follows Wagner's parabolic diffusion law at low (T < 600 °C) and high temperatures (T > 1100 °C). He found that the oxidation at intermediate temperature (600 °C < T < 1100 °C) is sub-parabolic. Also , he confirmed that the thermo-mechanical history o f the material has an important role in the oxidation rate, i.e., the work hardened material oxidises faster than the annealed one. In a separate investigation, Haugsrud (10) studied the effects of coatings o f Si02, Ce02 and La203 on the oxidation of N i , and found that these coatings reduce the rate of oxidation of N i by an order of magnitude at 800 - 900 °C. However, at temperatures above 1000 °C, these coatings have no beneficial effects. Czerwinski et al (11-13) investigated the effects of Ce02 particles on the oxidation mechanisms of N i , and reported an overall reduction in the oxidation of surface modified N i . Moosa et al (14, 15) studied the effects of alloying of N i with Y on the oxidation of N i in the temperature range o f 500 °C to 900 °C. They found that the oxidation rate increases with temperature at the lower temperatures (500 °C to 700 °C) and decreases at the higher temperatures (700 °C to 900 °C). Their microstructural investigation revealed that the formation of 6 coarse grains o f the N i O scale on the N i - Y alloy surface is responsible for the decrease in the oxidation rate. Peng et al (16-18) studied the effect o f the dispersion of La203 particles on the oxidation of N i at 900-1000 °C and found that the oxidation rate decreases by an order of magnitude. The associated T E M investigation revealed that the La2C>3 segregates to the grain boundary regions of the N i O scale which hampers the outward diffusion of N i . Also , it was found that the La203 particles improve the fracture toughness of the N i O scale by forming fine grains which prevent scale detachment. Strawbridge et al (19, 20) studied the effects o f the coatings of alkaline earth metals on the oxidation of N i and found that C a gives a good protection for N i . Green & Bestow (21) studied the effect of the addition of small quantities (0.5%) of Y , Ce, Zr and H f on the oxidation of N i . They found that all those additions reduce the oxide scale growth, Y and Ce being the most and least effective respectively. Giggins and Pettit (22) studied the high temperature oxidation o f N i - A l - C r ternary alloys and found that the alloys undergo a rapid transient oxidation followed by a steady state condition. They found that three different oxidation mechanisms occur depending on the composition o f the alloy. The alloys containing low A l (<5wt %) and low C r (<15wt %) fall under Type-I and form a continuous scale of N i O externally with the formation of discrete phases of Cr 203, Ni (Cr , Al)204 and AI2O3 internally. Type-2 alloys which contain low A l (<5wt %) and high Cr (>15wt %) form a continuous external scale o f Cr 203 with a discontinuous AI2O3 subscale. Type-Ill alloys containing high A l (>5wt %) form only an external scale of AI2O3. Moniruzzaman et al (23) studied the effect o f the variation of Re on the oxidation of two groups of Ni-base alloys (Ni-8Cr-2Ti-10Al- xRe and Ni -8Cr -2Ti -15Al - xRe; x = 0-2 M o l %). The isothermal and the cyclic oxidation tests conducted by them indicate that Re deteriorates the oxidation resistance of the alloy. The low oxidation resistance is attributed to the formation of discontinuous AI2O3 due to the evaporation of Re207. Also , they found that the R e / A l ratio is a good indicator of the oxidation resistance of a Ni-base alloy, i.e., R e / A l > 0.1 is detrimental to the oxidation resistance of the alloy. Jayne and Smialek (24) studied the segregation o f sulphur in a few Ni-base alloys including P W A 1480. They found out that the segregation o f sulphur to the oxide-metal interface improves the adhesion of the oxide scale, and hence improves 7 the cyclic oxidation behaviour o f the alloy. The effect of water vapour on the oxidation o f various Ni-base superalloys was investigated by Onal et al (25). They found that water vapour adversely affects the selective oxidation of aluminium in the AI2O3 formers and it affects the adherence of the oxide scale. Also , they reported the loss of chromia in the oxidation tests of chromia forming superalloys in the presence of water vapour. Dent et al (26) studied the oxidation of a 2 n d generation superalloy, C M S X - 4 at 750 °C and 950 °C. The T E M characterisation of the oxidised specimens revealed that the oxidation mechanism o f Re containing super alloys is temperature sensitive. They reported that C M S X - 4 predominantly forms an outer scale made of N i O when oxidized at 750 °C, while at 950 °C it forms a thin outer scale o f c t -A^Os/C^Os. Also , they concluded that T i is detrimental to the formation of a- AI2O3 and forms less protective oxide scales. Hook (27) studied the isothermal oxidation behaviour of SRR99 (1 s t generation superalloy) C M S X - 4 (2 n d generation superalloy) and RR3000 (3 r d generation superalloy) in the temperature range of 800°C to 1000 °C for 100 hrs. He found that the oxidation rate o f RR3000, which is also known as C M S X - 1 0 , is at least an order of magnitude higher than those o f SRR99 and C M S X - 4 . He attributed the increase in the oxidation rate of the 3 r d generation alloy to the lower chromium content in it. Akhtar et al (28-30) studied the isothermal oxidation behaviour of C M S X - 1 0 . The key finding of their study was that C M S X - 1 0 does not form alumina at low temperatures, up to 1000 °C. Instead, the alloy undergoes a diffusion controlled phase transformation at the sub surface region leading to the formation of discrete [3-NiAl phase. They have also argued that on further oxidation P-NiAl transforms through 8-M2AI3 to Spinel (MAI2O4) and that, alumina forms thereafter at high temperatures (above 1000 °C). The development of computational tools and thermodynamic database o f materials is helping in developing computational models for diffusion phenomenon during the oxidation of alloys. These models in general do not provide microstructural information. However, they give a fair idea about the diffusion path. One such model is the ' local equilibrium model for internal oxidation' developed by L i & Morral (31). The model provides the concentration profiles o f the solutes in the internal oxidation zone of a multi component system. It is based on the local equilibrium condition which is 8 illustrated with an error function model. The model predicts the average concentration o f the solutes in the internal oxidation region with the help of 'zig-zag diffusion model' developed by Hopfe & Morral (32). The local equilibrium model does not provide information regarding the development of internal oxide phases. Recently, Nijdam et al (33) developed a coupled thermodynamic-kinetic oxidation model which was tested using the experimental data on the oxidation of a single phase ternary alloy (Ni -Cr -Al ) . The model computes the composition-depth profiles and the amount of each oxide phase as a function o f oxidation time. Engstrom et al (34) simulated multiphase diffusion couples in the N i - C r - A l system using D I C T R A software. They considered a bond coated ( N i C r A l Y ) N i base superalloy as the diffusion couple between two Ni-base alloys. 9 Chapter 3 Experimental methods 3.1 Materials characterisation: The single crystal bars (13.85 mm diameter) o f C M S X - 1 0 received from the University of Cambridge were grown and heat treated commercially to produce the y-y' microstructure. The chemical composition of the alloy was confirmed by the E D S analysis. X - R a y Diffraction was used to establish that the alloy was in single crystal form. Scanning Electron Microscopy was used to confirm the microstructure. Transmission Electron Microscopy was employed to confirm the orientation relationship between the y and y' phases. The back reflection Laue technique was used for studying the orientation of single crystal. A Phillips X-Ray Diffractometer with Laue back-reflection accessories was used for the present work. A polychromatic radiation from a copper target was used for the experiment. A C M S X - 1 0 bar was ground at one end in order to produce a flat surface perpendicular to the cylinder axis. The flat surface was fine polished using a series of abrasive papers and finally with a diamond paste (6um size). The fine polished surface was then etched in aquaregia (HNO3: HC1 = 3:1) for 30 minutes for obtaining a stress free surface for the exposure. The Laue back reflection camera loaded with a photographic film was mounted on the diffractometer. The specimen was loaded in to the specimen holder with a distance o f 30mm from the film. A tube voltage o f 35 K V and a tube current o f 20 m A were attained before opening the X - R a y window for the exposure. It was ensured that the radiation was unfiltered white X-rays from Cu-target and the beam size was 1mm. The film was exposed to the back reflected X-rays for about 30 minutes and was then developed. A stereogram of the single crystal was constructed by trial and error method (36) using the Laue pattern, Greniger chart, Wolf f net and a standard [001] stereogram. The stereogram indicates that a [001] direction of the single crystal is oriented at 8° to the axis of the bar. 10 In a separate experiment, a standard X-ray diffraction profile (28 = 10°-100°) of the sample was obtained using a Yokagawa X-Ray Diffractometer. The flat transverse surface of the CMSX-10 bar ground to lum finish was exposed to Cu-K a radiation with a tube current of 20 mA and voltage of 40 kV. The profile (Figure 1) was then indexed using the JCPDS files and was used as the reference/base material profile during the X R D analysis of the oxidised specimens. ro c 20 40 60 80 100 2Theta Figure 1. A diffractometer trace showing the major reflections from a CMSX-10 specimen. A Scanning Electron Microscope (Hitachi S-3000N) was employed to confirm the microstructure of the alloy. The CMSX-10 bar was sectioned in a low speed diamond saw to obtain a disc of about 1 mm thickness and 13.85 mm diameter. One of the flat surfaces was ground to lum finish and loaded into the Scanning Electron Microscope. The microstructure of the sample (Figure 2) was observed and recorded in secondary electron (SE) modes at an accelerating voltage of 20 kV. 11 Figure 2. A Scanning Electron Micrograph (in Secondary Electron mode) showing the uniformly distributed cuboids of y' (grey region) in the matrix of y phase (bright region). Dislocations Figure 3. A Bright Field TEM image showing the dislocations and the cuboids of y' aligned along a [001] direction. 1 2 Transmission Electron Microscopy was employed to obtain the crystallographic details and the coherency of the precipitate phase (y') with the matrix (y). The C M S X - 1 0 bar was sliced in an abrasive cutter to obtain a disc of 13.85 mm diameter and 1 mm thickness. The disc was then brazed onto a copper block using a filler material made o f a silver alloy. A few discs o f 3 mm diameter and 1mm thickness were cut by Electron Discharge Machining. The discs were removed from the copper block by melting the filler material using an oxy-acetylene flame. The discs were then glued onto a holder for polishing and were thinned in steps to 100 \i thick. Each disc was then electro-polished with a solution o f 30% perchloric acid-70% acetic acid at 5 °C and 30 V for about 2 minutes to get a concave specimen with a small hole at the centre. A Hitachi H-800 Scanning Transmission Electron Microscope (STEM) was used at an accelerating voltage of 200 K e V for the present work. The bright field (BF) images and the selected area diffraction ( S A D ) patterns are as shown in the Figures 3 and 4 respectively. 7 Figure 4. A selected area diffraction pattern of the CMSX-10 showing the regul; FCC pattern from both y and f phases with superlattice reflections from f 13 3.2 Isothermal Oxidation: A high temperature thermogravimetric analyser (Linseis L81/1750) with an accuracy of ±1.5 °C and +0.001 mg was used for the isothermal oxidation tests. C M S X - 1 0 discs o f 13.85 mm diameter and about 1 mm thickness were used for the oxidation experiments. The flat surfaces were polished to l u m finish while the curved surface was left unpolished. In order to accommodate the C M S X - 1 0 discs of a larger size (13.85 mm diameter), two longitudinal slots were made in the sample crucible (7 mm diameter) by using a low speed diamond saw. Before doing the actual oxidation experiments, several base line thermo gravimetric experiments were run at 1000 °C for 100 hrs with empty crucibles in still air. A heating rate of 20 °C/minute was used to raise the temperature and it took about 50 minutes to reach 1000 °C from room temperature. It was found that the background noise level was within + 0.2 mg. A C M S X - 1 0 disc was then placed in the slot of the crucible as shown in Figure 5. The isothermal oxidation experiments were carried out in the still air at 1000 °C for 100 hrs with the same heating rate (20 °C/min). The mass gain (mg) with respect to time (h) was recorded. The above experiment was repeated with different samples at several temperatures from 800 °C to 1250 °C. 14 T G A furnace TGA specimen Crucible To microbalance Figure 5. A schematic of the experimental set up on a TGA. 3.3 Characterisation of the Oxidised Specimens: The X R D patterns of the as oxidised samples were obtained from an X-Ray Diffractometer (Rigaku) by using C U - K Q radiation at 40 k V and 20 m A . The specimens oxidised at 900 ° C and 1000 ° C displayed peaks belonging to the external scale only, no peaks were found from the phases of the subsurface. It was assumed that the external scale was too thick for X-rays to penetrate up to the subsurface. It was decided to enable X - Rays to reach the subsurface by reducing the thickness of the external scale. Hence, as oxidised specimens were ground with (a 600 grit abrasive paper) in steps and an X R D pattern was obtained in each step. The above steps were repeated until the X R D pattern displayed peaks from y/y' base material only. 15 For the cross sectional microscopy, the oxidised samples were sectioned at the centre using a low speed diamond saw. The transverse sections were then mounted using G - l epoxy and allowed to set for 24 hrs at room temperature. The samples were then polished as per the standard metallographic procedure to 1 um diamond finish. The ground surface was carbon coated before loading into the scanning electron microscope in order to avoid the charging. The back scattered electron (BSE) imaging mode was employed at an accelerating voltage o f 20kV. Chemical compositions of the various phases were obtained by energy dispersive x-ray spectroscopy (EDS). The estimated accuracy for the light elements, O, A l , N i , T i , Cr and Co is within 2 at% and within 5 at% for heavy elements, M o , Hf, Ta, W and Re. A series of area (3u x 3(a) E D S analysis was performed on the transverse sections of the specimens in order to obtain the variation o f the chemical composition from the surface through the subsurface to the bulk o f the base material. A lower accelerating voltage of 5kV was employed for the elemental mapping of fine phases. 16 Chapter 4 Results 4.1 Thermogravimetry: The results of isothermal oxidation experiments conducted in a T G A for 100 hrs at temperatures ranging from 800 °C to 1250 °C are shown in Table II. Figure 6 is a plot of the variation of mass gain with oxidation temperature. Table II. Results of the experiments after isothermal oxidation for lOOh. Expt. No. Temperature (°C) Thickness (mm) Surface area (cm2) Mass gain (mg) Mass gain/Area (mg/cm2) 1 800 0.8 3.36 11.2 a i l l L 3.33 2 850 1.34 3.60 16.75 4.66 3 900 0.8 3.36 12.9 3.84 4 950 1.49 3.66 18.2 4.97 5 1000 0.87 3.39 13.43 3.96 6 1050 1.32 3.59 12.8 3.57 7 1075 1.32 3.59 11.53 3.21 8 1100 0.83 3.37 9.5 2.82 9 1125 1.44 3.64 10.1 2.77 10 1155 1.32 3.59 11.15 3.11 11 1165 1.43 3.64 13.73 3.78 12 1175 0.93 3.42 11.9 3.48 13 1185 1.46 3.65 14.4 3.95 14 1200 0.84 3.38 10.95 3.24 16 1250 0.86 3.39 12.54 3.7 17 5.5 5.0 4.5-E .o O ) E 4 .0 -co °> 3.5 H to 1 — 3.0-c 2.5-2.0-800 900 1000 1100 Temperature (°C) 1200 1300 Figure 6. A plot of mass gain versus temperature after 100 h of isothermal oxidation. Figure 6. indicates that the mass gain due to the oxidation does not increase monotonically with temperature. There is a trend of a decrease in the mass gain of the alloy with an increase in the oxidation temperature between 1000 °C and about 1100 °C. Beyond 1100 °C however, the mass gain increases with temperature. However, it is possible to sort the results of the oxidation experiments from the Table II. into two sets based on the thicknesses of the specimens. The first set would comprise specimens with the thickness 0.85 + 0.5 mm and a second set would be those with the thickness 1.4 + 0.1 mm. It is appropriate to draw plots of the mass gain versus temperature of these two sets separately, and they would appear as shown in Figure 7 and Figure 8. 18 5.5-r 5.0-4 - 5 -- i • I ' I • 1 1 1 900 1000 1100 1200 1300 Temperature (°C) Figure 7. A plot of mass gain versus temperature of the specimens with a thickness of approximately 0.85 mm. 5.5 2.5 H - I ' I ' 1 •— 1 1 1 . 1 800 900 1000 1100 1200 1300 Temperature (°C) Figure 8. A plot of mass gain versus temperature of specimens with a thickness of approximately 1.4 mm. 19 The results of isothermal oxidation experiments obtained from the T G A can be expressed directly as the plots of mass gain per unit surface area of the specimen (mg/cm 2) versus oxidation time (h) over 100 h at different temperatures. The mass gain versus time curves of C M S X - 1 0 drawn separately for the two sets of experiments are shown in the Figures 9-12. It is clear from Figures 12, and 14 that the mass gain due to the oxidation of the alloy increases with an increase in temperature between 800 °C and 950 °C. The mass pick-up after 100 hrs of isothermal oxidation at 800 °C and 900 °C are 3.33 mg/cm 2 and 3.84 mg/cm 2 respectively which are close to the values, 3.6 mg/cm 2 and 4.5 mg/cm 2 respectively reported by Hook (27). However, as shown in Figurel2, at 1000 °C the curve saturates after about 25 hrs of oxidation and remains so even after 100 hrs indicating the inhibition of oxidation. Though this behaviour was reported by Hook earlier (27), the present work indicates a higher mass pick-up (4.1 mg/cm 2), approximately twice that reported earlier (2.3 mg/cm ). From Figures 12 and 14 it is convincing that the mass gain due to the isothermal oxidation decreases with an increase in temperature between 950 °C and 1100 °C. And , there is a trend of an increase in the mass gain with temperature from 1100 °C to 1250 °C (Figures 13, 15). Hence, at this stage, based on the T G A results, it may be stated that the high temperature oxidation of the single crystal superalloy, C M S X -10 is not a simple monotonous process and that it exhibits complex transition oxidation behaviour. 20 0 20 40 60 80 Time (h) 100 120 Figure 9. The mass gain versus time curves for specimens with a thickness of about 0.85 mm at temperatures ranging from 800 °C to 1000 °C. HI 20 40 60 80 Time (h) — i — 100 120 Figure 10. The mass gain versus time curves for specimens with a thickness of about 0.85 mm at temperatures ranging from 1100 °C to 1250 °C. 21 —I— 20 40 I— 6 0 3 0 Time (h) 1050°C 1075°C — i — 100 120 Figure 11. The mass gain versus time curves for specimens with a thickness of about 1.4 mm at temperatures ranging from 850 °C to 1075 °C. 4.5-4.0 H _ 3.5 E o 3.0 H £ 2.5 co 8 2.o H to E « 1.5 1 0.5-1 0.0 —T— 2 0 —I— 6 0 8 0 Time (h) 1185°C 1165°C — i — 1 0 0 1 2 0 Figurel2. The mass gain versus time curves for specimens with a thickness of about 1.4 mm at temperatures ranging from 1125 °C to 1185 °C. 22 However, as shown in Figure 12, at 1000 °C the curve saturates after about 25 hrs o f oxidation and remains so even after 100 hrs indicating the inhibition o f oxidation. Though this behaviour was reported by Hook earlier (27), the present work indicates a higher mass pick-up (4.1 mg/cm 2), approximately twice that reported earlier (2.3 mg/cm ). From Figures 9 and 11 it is convincing that the mass gain due to the isothermal oxidation decreases with an increase in temperature between 950 °C and 1100 °C. And , there is a trend of an increase in the mass gain with temperature from 1100 °C to 1250 °C (Figures 10, 12). The plot of oxidation rate of the alloy after 50 hrs of oxidation versus temperature shows the anomalous behaviour (Figure 13). Hence, at this stage, based on the T G A results, it may be stated that the high temperature oxidation of the single crystal superalloy, C M S X - 1 0 is not a simple monotonous process and that it exhibits complex transition oxidation behaviour. Figure 13. A plot of mass gain rate versus temperature after 50 h of isothermal oxidation. 4.2 Microstructure: The scanning electron micrographs of the transverse sections are as shown in the F i 14-19. gures 23 Externa] Scale Zone I Zone 2 Base S E WD15.6mi^  20.0kV xl .°5k °30°um Figure 14. A scanning electron micrograph of the transverse section of a specimen oxidised at 800 °C for 100 h. Figure 15. A scanning electron micrograph of the transverse section of a specimen oxidised at 900 °C for 100 h. 24 « t e . i s ! & &i . B g! ! B§$s£ . | Figure 16. A scanning electron micrograph of the transverse section of a specimen oxidised at 1000 °C for 100 h. Figure 17. A scanning electron micrograph of the transverse section of a specimen oxidised at 1100 °C for 100 h. 25 Figure 18. A scanning electron micrograph of the transverse section of a specimen oxidised at 1155 °C for 100 h. J Figure 19. A scanning electron micrograph of the transverse section of a specimen oxidised at 1250 °C for 100 h. 26 From the micrographs (Figures 14-18) it is clear that on exposure to high temperatures, C M S X - 1 0 develops both internal and external oxides. It has been confirmed by the visual examination that the external scale is olive green in colour and forms at all temperatures. It has been observed that the external scale formed at lower temperatures (up to 1155 °C) is adherent and remains intact even after cooling. However, the scale formed at a high temperature (above 1155 °C) spalls during cooling. The micrographs clearly indicate that there are various internal oxide phases which appear in the form of zones in the subsurface region while there is only one external scale. The first internal zone (adjacent to the external scale) in the subsurface has a fine microstructure when compared to the other phases and it w i l l henceforth be called the 'fine structure zone' (FSZ). The second internal zone, which is adjacent to the F S Z appears as a discrete phase at low temperatures (Figures 14, 15) However, at temperatures above 1000°C, it forms a continuous layer (Figures 17 - 19). The third internal zone appears as a thin dark layer at temperatures above 1000° C (Figures 17, 18). The fourth internal zone forms between the third zone and the base material especially at high temperatures is called the depleted zone. The depleted zone is distinctly visible as the uniform single phase region adjacent to the dual phase (y + y') base material (Figure 18). 4.3 Energy Dispersive Spectroscopy: The change in the chemical composition with the depth of the specimen as obtained from the E D S analysis. The results are shown in the Figures 20-23. In general, the profiles o f O, N i depict that the external scale is made of N i O at all temperatures. The profile of C o indicates its presence in the external scale. The profiles of Cr, Ta and W in comparison with that of O indicate that they form oxides in FSZ. The profile of Re at a low temperature (Figure 20) indicates its presence in the F S Z . However, at a temperature above 1000 °C, there is no Re in the F S Z which indicates the loss of Re (Figures 21, 22). The profiles of M o , T i and H f do not show any significant trend. 27 The composition profiles do not clearly depict the presence of alumina. This may possibly be due to the alumina layer being too thin for the window size (3um x 3 urn) used for the area E D S . However, the trends of O, N i and the A l profiles in comparison with the microstructure (Figure 23) indicate that the ratio, N i : A l : 0 approaches 0:2:3 in the vicinity of third internal zone suggesting alumina formation. The fourth zone which is adjacent to the y/y' base material is depleted o f A l (Figures 21, 22) and is called Depleted Zone. This zone has about 80 % N i and less than 5% A l . It can also be observed from Figure 18 that this zone is a single phase region and has the same contrast as that of y in the y/ y' base material. This suggests that the depleted zone is possibly made of a y (single phase solid solution of nickel). The X R D analysis and the elemental maps discussed in the later sections confirm the presence o f N i O in the external scale, presence of compound oxides such as NiTa204, M W O 4 , spinel (N1AI2O4), AI2O3, reflections from a depleted zone and y'-Ni3Al. 28 29 6- - • Co Mo I 1 1 ' 1 1 r 10 20 30 40 Depth (micron) Figure 21. The chemical composition profile of a specimen oxidised at 1100 °C for 100 h. Depth is from the surface of the external scale. 30 20 30 40 Depth (micron) Figure 22. The chemical composition profile of a specimen oxidised at 1155 °C for 100 h. Depth is from the surface of the external scale. 31 10 20 30 40 50 Depth (micron) Figure 23. The chemical composition profile of a specimen oxidised at 1250 °C for 100 h. Depth is from the surface of the external scale. 32 4.4 Elemental X-Ray Mapping: The elemental maps of selected specimens oxidised for 100 hrs, obtained from the E D S , are shown in the Figures 24-28. Figure 25. The elemental X-ray maps of Ni, Al and O of specimen oxidised at 900 °C. Figure 27. The elemental X-ray maps of Ni, Al and O of a specimen oxidised at 1100°C. 3 5 36 The discrete regions in the specimens oxidised at 800 °C (Figure 24) show depletion o f nickel, enrichment of aluminium and the presence of oxygen suggesting that they could be spinel. The map of the specimen oxidised at 900 °C (Figure 25) indicates that the discrete phase is elongated. The elemental map of the specimen oxidised at 1000 °C (Figure 26) depicts that the F S Z is enriched with oxygen and aluminum and is depleted of nickel. It possibly indicates that the spinel is the continuous phase which forms the matrix o f the F S Z , and the bright patches found with in the F S Z are precipitates of oxides. The results o f X-ray diffraction discussed in the next section confirm that they are N iTa2C> 4 , NiWC>4. The edge of the F S Z towards the base material is even more enriched with both oxygen and aluminum. And , nickel is almost absent. This suggests that the region is possibly alumina and is continuous. The same observation can be made with the map of the 1100 °C specimen (Figure 27). It can be observed that a distinct spinel layer is formed adjacent to the F S Z . However, it is the map of the specimen oxidised at 1155 °C which gives the fine details (Figure 28). The O, N i and A l maps confirm the observations made above. The map indicates the presence of Cr in the spinel. C o shows its presence in the external scale and is depleted in F S Z . 37 4.5 X-ray diffraction: The X R D patterns of the specimens, especially those oxidised at lower temperatures (below 1155 °C) showed the peaks belonging to the external scale o f N i O only as X-rays could not penetrate down to the F S Z . However, the specimens revealed the subsurface information as they were ground using 600 grit abrasive papers in steps and X R D patterns were obtained by a trial and error method. One such effort is shown in the Figure 29. It shows that the specimen oxidised at 900 °C has an external oxide of N i O (Figure 29A). Numerous peaks obtained after the first round of grinding (Figure 29B) indicate that the F S Z is a composite layer made of several phases. When the pattern was indexed with the help of the J C P D S files (Table III.), it was found that F S Z was made o f several oxides, predominantly, NiTa206 and NiW04. Also , at this stage a distinct peak of K-alumina at 14.1° was observed. The spinel was not detected at this stage possibly because its peaks were overlapped by the peaks of other phases predominantly NiTa204. A s the specimen was ground further and X - R a y diffraction profile was obtained repetitively, at one stage the profile changed significantly. The peaks were indexed and the presence of spinel, depleted zone (y-Ni) and y' - N i 3 A l was confirmed (Figure 29C). A similar trend as above was observed in the case of the specimen oxidised at 1000 °C. (The specimen was ground t i l l the peaks from the subsurface appeared) However, the data from the specimen oxidised at 1000 °C reveals the presence of both K-alumina and a-alumina (Figure 30). The intermetallic compound [3-NiAl was also observed in the specimen. The elemental map of the specimen (Figure 26) indicates that P-NiAl appears as a discrete phase within the depleted zone. In the case o f the specimens exposed to high temperatures (above 1155°C), the external N i O scale and the F S Z spalled during cooling, and this enabled X-rays to penetrate up to the base material. And , hence the presence o f the spinel, a-alumina, y -N i and y ' - M 3 A I was clearly revealed as shown in the Figure 31. Table III. lists all the phases that are involved in the oxidation of C M S X - 1 0 . 38 >• (TO* w e as "i 0 K» 1 i 2 re c V D 8 o i ' l n o © w o C M I 8 9 O . <6 M © "1 " O Bi M g in * g , b i 2 8, a are) B 2. g. D C L M — • a v o HQ © • © a O 88 s ffi © OTQ Intensity (Arbitrary units) Intensity (Arbitrary units) lntensity(Arbitrary units) CD sr g 0 < » I T I NiO T NiTa,0 ( f NiWO, | ic-Al 0 ? (J-NiAl3 T Depleted zone F NiAI;Ot T u -AlO »t . . .» ».» —II— 20 — I — 40 60 2Theta 80 100 Figure 30. The XRD profile of a specimen oxidised at 1000 °C for 100 h. » t y'-NijAI T Depeieted zone 1 NiAI,0 4 1 ' 1 J i j l i ^ ^ ^ * - 1 1 1 1 1 1 1 40 60 80 100 2Theta Figure 31. The XRD profile of a specimen oxidised at 1250 °C for 100 h. 40 Table III. A list of phases involved in the oxidation of CMSX-10. No. Species Symbol JCPDS file No. 1 Nickel oxide NiO 04-835 2 Nickel Tantalate N iTa 2 0 6 32-702 3 Nickel Tungstate N iW0 4 15-755 4 Spinel N iAI 2 0 4 10-339 5 Kappa Alumina K - A l 2 0 3 04-878 6 Alpha Alumina a - A l 2 0 3 10-173 7 Y - Nickel Ni 04-850 8 P - Nickel Aluminide NiAl 20-019 9 Y' - Nickel Aluminide NisAI 09-097 41 Chapter 5 Discussion 1 D A plot o f mass gain versus time for CMSX-10 is shown in Figure 32 for a set of isothermal tests carried out in the present work. It is seen that, the mass gain increases linearly with t l / 2 initially. Deviation from linearity occurs after a critical time (tc) at each temperature. A t times in excess of tf, the mass gain occurs at a lower rate than would be dictated by a t 1 / 2 relationship. Figure 32. A plot of unit mass gain versus time1/2at various temperatures 42 In the linear regime mentioned above, the unit mass gain, AM may be related to time, t through a constant, k as follows. AM = k[t]1/2 (1) The rate constant (k) and the critical time (tc) for the parabolic to subparabolic transition o f C M S X - 1 0 for temperatures 850, 950 and 1050C were obtained from Figure 32 and are listed in Table IV . Due to lack of data points in Figure 32, for the linear region o f the sample oxidised atl 155 °C, these data are excluded from Table IV . Table IV. The rate constant for initial oxidation and the critical time for various temperatures Temperature (°C) Parabolic rate constant k ( mg /cm 2 / h 1 ' 2) Critical Time M h ) 850 0.53 25 950 0.88 20 1050 1.5 3 From Figure 32 and Table IV it is clear that the oxidation rate of the alloy in the initial region, which is parabolic with respect to time, increases with the oxidation temperature. A t the same time, the critical time required for the transition from the parabolic to the sub-parabolic behaviour decreases with an increase in temperature. Moreover, it is clear from Figure 32 that the rate constant is higher and the critical temperature lower at 1155 °C than the corresponding figures for 1050 °C as would be expected from the trend in their variation with temperature (Table IV) although the data is not adequate to quantify these parameters. This transition from the initial parabolic regime to the post transition sub-parabolic is reflected in the mass gain after 100 h versus temperature plot (Figure 6). The point is made clearer i f one considers mass gain versus temperature plots at shorter isothermal exposures. Figure 33 and 34 are the plots for times 3 h and 20 h respectively. 43 3.0 4 (mg/cm' 2.5-c co O) 2.0-mass 1.5-1.0- a m •* i ' 1 ' 1 •- 1 . 1— 800 900 1000 1100 1200 Temperature (°C) Figure 33. A plot of mass gain versus temperature after 3 h of isothermal oxidation Figure 34. A plot of mass gain versus temperature after 20 h of isothermal oxidation 44 From Figure 33 it is seen that there is a monotonic increase in the unit mass gain o f the alloy with an increase in temperature up to about 1050 °C at an exposure time of 3 h, whereas, at 20 h (Figure 34) the monotonic increase occurs up to a temperature of around 950 °C. The reason for the difference is as follows. A s seen from Table IV, the critical time for the parabolic to sub-parabolic transition is 3 h at 1050 °C, and, in excess of that value at lower temperatures. Therefore, the gain in the mass at 3 h at any temperature up to 1050 °C is a result of oxidation in the parabolic regime alone. Moreover, at temperatures in excess o f 1050 °C it is a combination o f the mass gain through oxidation in the parabolic and that in the sub-parabolic regimes. Similarly, Table I V shows that the critical time is 20 h at 950 °C. Consequently, a monotonic rise in the mass gain occurs up to around 950 °C resulting from oxidation in the parabolic regime (Figure 34). A t temperatures in excess of 950 °C and an isothermal exposure time of 20 h, the total mass gain is a result of oxidation in the parabolic regime combined with that in the sub-parabolic domain. Table V compares the rate constants in the initial parabolic oxidation o f C M S X - 1 0 , calculated from the present work and the ones calculated from the mass gain versus time curves found in the literature (27). It can be observed in general that the rate constant increases with an increase in the temperature. However, the present work shows a higher rate constant than the earlier work did at each temperature. Table V. A comparison of parabolic rate constants obtained from the present work and that found in the literature Temperature (°C) Parabolic rate constant k ( ma /cm 2 / h V 2 ) Present work Literature 800 0.49 0.36 850 0.53 900 0.75 0.54 950 0.88 1000 1.13 0.64 1050 1.5 -45 5.1 The mechanism of oxidation in CMSX-10: Earlier investigations at temperatures below 1000 °C have shown that an oxidation induced phase transformation occurs in the 2 n d generation single crystal superalloy C M S X - 4 (30), and, in the 3 r d generation single crystal superalloy C M S X - 1 0 (28). The process starts with the formation of a N i O scale through external oxidation, i.e. N i diffuses out o f the superalloy lattice to react with O at the surface and form N i O . The associated depletion of N i affects the phase stability of y/y' in the subsurface. A s a result, the N i rich y' (NiaAl) transforms into P ( N i A l ) in the subsurface. Internal oxidation takes place only after the formation o f the (3 phase. From the above information it appears that the initial parabolic oxidation behaviour of C M S X - 1 0 observed in the present work (Figures 35 and Table IV) is due to the growth of N i O . The transition from the parabolic oxidation behaviour after a critical time (tc) is possibly due to the formation of oxides in the internal oxidation zones. A s the internal oxides precipitate in the subsurface, the outward diffusion of N i is hindered leading to a reduced growth o f N i O . Hence, it may be said that the initial mass gain is due to the growth o f N i O external scale, whereas at times in excess of tc it is due to a combination of external and internal oxidation. It is seen from the microstructures (Figures 14 - 19) that, upon oxidation, C M S X - 1 0 forms various phases dispersed in the internal oxidation zone. A t the lower temperatures (Figures 14, 15), these phases remain fine and dispersed. However, at higher temperatures (Figures 16 - 19) some o f the phases become continuous. The X R D and E D S analyses (Figures 30, 22) indicate that Zone 1 (the FSZ) , is predominantly made of compound oxides such as NiTa206, N iW04 . It can be observed that Zone 2 is a single phase region predominantly made of N i , A l and O. The N i : A l : 0 ratio o f Zone 2 is close to 1:2:4 (Figure 22) indicating that this region is possibly made of N1AI2O4. The X R D pattern as shown in the Figure 31 confirms the presence of MAI2O4. The E D S analysis (Figure 23) o f Zone 3 indicates that this region is possibly made of AI2O3 as the A1:0 ratio is close to 2:3. The X R D pattern confirms the presence of alumina (Figure 31). It can be observed from the microstructure that Zone 4 is also a single phase region. 46 The E D S analysis of the region indicate that it has about 85 at% N i which is above the solvus composition (82 at% N i ) in the binary phase diagram at 1250 °C (Figure 35). The X R D pattern (Figure 31) confirms that the structure o f the depleted zone is y. It appears from the X R D analysis (Figure 29) that C M S X - 1 0 forms K-A1 2 0 3 when oxidised at a temperature below 1000 °C and forms 01-AI2O3 at a temperature above 1000 °C (Figure 31). It is interesting to note that at 1000 °C it forms both K-AI2O3 and a - A l 2 0 3 (Figure 30). It is possible that, K-AI2O3 which is a discontinuous phase at the lower temperatures gets transformed at about 1000 °C, into a continuous layer of (X-AI2O3. C M S X - 1 0 contains 10 alloying elements alloying elements (Table I). Therefore it is appropriate to consider the effect of other elements, along with that of N i and A l , on the oxidation behaviour of the alloy. It has been discussed in the literature (28) that some of the elements in the alloy occupy N i sites, and, are called N i like elements (Co, Re), while a few elements occupy A l sites, and, are called A l like elements (Ti , M o , Ta, W , and Hf). To assess the cumulative effects of N i and N i like elements a parameter called N i equivalent LNi has been used (28). Similarly A l equivalent EA1 is used to sum up the concentrations of A l like elements. Such a procedure enables one to copare the change in composition brought about through oxidation with respect to the binary N i - A l phase diagram. •5) It has been shown earlier (28) that a plot of E N i / (ENi + EA1) versus the depth o f the specimen gives an insight into the phase transformations that may occur in the subsurface. The plots for specimens oxidised at 900 °C for 100 h indicate the formation of P phase in the Oxidation Reaction Domain (ORD) o f the subsurface. It has also been shown that at 1000 °C, the profile of I N i / ( I N i + IA1) in the advancing front of the O R D falls into the 8 phase region. Due to the depletion of N i , the phase equilibrium of the advancing front of the O R D shifts from the initial y/y' region through P into 8. It has been shown that, on isothermal oxidation at 900 °C, this shift remains confined to the P even after prolonged oxidation. However, at 1000 °C, the equilibrium of the advancing front of the O R D shifts further so much so that after 100 h it falls into the 8 region. 47 Depth (nm) Figure 35. A plot of LNi / (ZNi + LAI) versus depth of specimens oxidised for 100 h. Figure 36. The binary phase diagram of Ni-Al system showing the shifts in the equilibrium positions of the advancing front of ORD at various temperatures. The composition reached at 900 °C and 1000 °C are taken from reference 28. 48 In the present work, a similar exercise was done on the specimens oxidised at temperatures above 1000 °C (Figure 35) namely, 1155 °C , 1200 °C and 1250 °C. The dotted horizontal lines shown in Figure 35 denote the phase boundaries at an intermediate temperature, 1200 °C. It can be observed from the phase diagram (Figure 36) that these phase boundaries do not shift much going from 1155 °C to 1250 °C. Hence they approximately hold good for the entire range (1155 °C to 1250 °C). It can be seen that at the advancing front o f the O R D , the profile of S N i / ( S N i + SA1) dips into the liquid phase region at 1155 °C. And , at higher temperatures, 1200 °C and 1250 °C, the composition becomes A l enriched in the liquid phase. A n alternative representation o f the phase shift may be seen more clearly on the binary phase diagram (Figure 36). It can be seen from the phase diagram that the 8 phase is stable up to about 1133 °C. Therefore the specimens oxidised at higher temperatures of 1155 °C and above (Figure 35) are not expected to form the 8 phase. However, a trend of increased depletion of N i at the advancing front o f the O R D with an increase in the oxidation temperature is clear from the Figure 36. It has been suggested (29) that continuous layer of 8-M2AI3 may be responsible for the saturation of the mass gain curve at 1000 °C which occurs following the initial rapid mass gain. Whereas, the specimen oxidised at 1000 °C in the present work indicates the presence of a dark continuous layer at the edge of F S Z towards the base material. The elemental X - R a y map (Figure 23) indicates that this layer is possibly alumina as the region is depleted of N i and enriched with O and A l . The X R D pattern of the specimen oxidised at 1000 °C (Figure 30) in the present work confirms the presence of alumina. Hence it may be said that the formation of a continuous layer o f alumina is responsible for the saturation of the mass gain in the isothermal oxidation of the alloy at 1000 °C. A t a higher temperature, possibly, this diffusion barrier of alumina forms quickly leading to the reduced mass gain. The impact of the formation of a continuous barrier to oxidation is reflected in the relative values of the thickness of the N i O layer and the F S Z as seen from the F S Z as seen from Figure 37. 49 - * - NiO -i ' 1 ' 1 1 1— 900 1000 1100 1200 Temperature(°C) Figure 37. A plot of the thicknesses of NiO and FSZ versus temperature after 100 h of isothermal oxidation. 1 2 -10H E CD C 1000 1050 1100 1150 Temperature( °C) 1200 1250 Figure 38. A plot of the thicknesses of the spinel and the alumina versus temperature after 100 h of isothermal oxidation. 50 It can be observed that the thicknesses o f the oxide layers after 100 h o f oxidation increase with an increase in the temperature up to 1000 °C. After 1000 °C, the formation o f the continuous diffusion barrier results in a decrease in the thickness o f these layers with an increase in temperature. A plot of the thicknesses of the spinel and the alumina versus temperature is shown in the Figure 38. It is clear from the figure that the thicknesses of these two layers increase with an increase in temperature. Further experimentation was done to explore the role of the continuous layers in the oxidation behaviour o f C M S X - 1 0 . A specimen prepared as per the standard metallographic procedure was oxidised at 1000 °C for 100 h. The same specimen was cooled to room temperature and then exposed to a lower temperature o f 800 °C for 100 h. Figure 39 shows the mass gain curve of the pre-oxidised specimen in comparison with that o f the fresh specimen isothermally oxidised at 800 °C for 100 h. The mass gain o f the pre-oxidised specimen is reduced by an order of magnitude. This lower mass gain suggests that the continuous layers formed at 1000 °C act as a diffusion barrier. Time (h) Figure^. A plot of unit mass gain versus time of (a) a fresh sample oxidised at 800 °C for 100 h (b) a sample preoxidised at 1000 °C for 100 h followed by isothermal oxidation at 800 °C for 100 h. 51 In another experiment to gain further insight, a specimen was isothermally oxidised at 1250 °C for 30 h and cooled to a lower temperature of 1000 °C. The specimen was then held at the latter temperature for nearly 3 Oh. The corresponding mass gain curve shows a sudden change in the slope at 30 h indicating a reduction of the oxidation rate (Figure 40). In yet another experiment, a specimen was isothermally oxidised at 1000 °C for 30 h and then heated to 1250 °C. The specimen was then held at the latter temperature for 30 h. The mass gain curve of this experiment shows an increase in the oxidation rate when the exposure temperature was raised (Figure 41). The two experiments mentioned above indicate the following. If the continuous layers formed are an effective diffusion barrier then the mass gain in the steady state is due to the growth of these layers. However, should they not be an effective diffusion barrier then the mass gain in the steady state could be due to the growth o f the external N i O brought about by the outward diffusion of N i through these layers. The combination o f the above two processes is also possible. In any case, these experiments indicate that, above 1100 °C, the monotonic increase o f mass gain with temperature (Figure 6) is due to the thermal factor which enables the kinetics o f diffusion across the barrier. 52 E o CO E cn 7-6 -5HI 4-31-2-H o-•1 -~ i — • 1 ' r 10 20 ' ' i 1 — i — 1 r —r -30 Temperature Mass gain -1— 40 50 Time (hrs) -r— 60 •1400 •1200 •1000 CD 3 S ET S o -400 « O -800 -600 -200 70 Figure 40. The mass gain curve of a high temperature - low temperature oxidation experiment along with the time-temperature profile. Figure 41. The mass gain curve of a low temperature-high temperature oxidation experiment along with the time-temperature profile. 53 5.2 Loss of rhenium: The mass gain versus time curves at a temperature above 1000 °C shows a negative growth over a small region in the graph (Figures 10 and 12). This is attributed to the loss o f volatile rhenium oxide from the F S Z . In fact, chemical composition profiles of the specimens exposed to high temperatures (Figures 21, 22) clearly show that Re is lost in the F S Z , and is not accumulated elsewhere in the system. Whereas the profiles show that Ta, W and H f are accumulated in the F S Z . However, the composition profile at 900 °C does not show a loss of rhenium in the F S Z (Figure 20). Hence the isothermal mass gain curve at 900 °C does not show a negative mass gain at any instance. 5.3 Uneven oxidation: Figures 6 - 8 depict that the specimens with 1.4 mm thickness registered a higher mass gain when compared to the thinner (0.8 mm) specimens. The reason for this behaviour is not clear. However, the microstructures of the specimen reveal that the curved surface of the specimen disc has a high roughness as it was not ground before conducting the experiments. Hence, it shows a rapid oxidation at the circumference of the disc. The total thickness of the oxides at certain regions of the curved surface exceeds 100 um (Figure 42) as against a scale thickness of about 40um on the flat surface. This difference in the oxidation rates at the flat surface and the curved surface o f the discs could be responsible for the higher values of unit mass gain for the thicker specimens. 54 Figure 42. A scanning electron micrograph showing uneven oxidation at the curved surface of a specimen oxidised at 850 °C for 100 h. 5.4 The poor oxidation resistance of CMSX-10: The reason for the rapid oxidation o f C M S X - 1 0 is not clear at this stage. The X-ray diffraction analysis o f an earlier work (27) has shown that C r 2 0 3 forms in the 2 n d generation alloy, C M S X - 4 and the 1 s t generation alloy, SRR99 (Table I) at temperatures up to 1000 °C but not in C M S X - 1 0 . Although it was not demonstrated in that study the formation o f a continuous layer o f C ^ C h , it was suggested that the presence of Cr203 is possibly essential for the oxidation resistance at temperatures below 1000 °C. The present work confirms that C M S X - 1 0 does not form Cr 2 03 at any temperature. However, the alloy forms continuous layers o f the spinel, NiAl 2 C«4 and alumina at a later stage in the subsurface. This enables the rapid oxidation of the alloy until a continuous layers of the diffusion barrier is formed in the subsurface. 55 Chapter 6 Conclusions The 3 r d generation single crystal superalloy, C M S X - 1 0 exhibits anomalous oxidation behaviour. This anomaly is attributed to the diffusion driven morphological changes in the subsurface. C M S X - 1 0 exhibits a transition from parabolic to sub-parabolic oxidation behaviour at all temperatures. The initial parabolic behaviour is attributed to the growth o f an external scale N i O . The post transition sub-parabolic behaviour is due to a combination o f both external and internal oxidation. C M S X - 1 0 does not form Cr203 at any temperature. However, it forms continuous layers of the spinel, MAI2O4 and alumina when exposed to a temperature above 1000 °C. One or both of these layers may act as the diffusion barrier. 6.1 Suggestions for future work: In the present work, specimens were characterised after 100 h of the isothermal oxidation tests. It is important to know the change in microstructure with the oxidation time. Hence, intermittent oxidation tests followed by the microstructural characterisation would shed further light into the oxidation of C M S X - 1 0 . 56 References 1. Superalloys II, Edited by C.T. Sims, N . S. Stoloff, W . C . Hagel, Jhon Wiley & Sons, 1987. 2. R. Haugsrud, Corrosion Science, 45 (2003) p211. 3. C . Wagner, J. Electrochem. Soc , 99 (1952) p369. 4. C. Wagner, Z Elektrochem., 63 (1959) p772. 5. N . N . K o i , W. W . Smeltzer, J. D . Embury, J. Elctrochem. S o c , 122 (1975) p l495 . 6. R. Atkinson, S. Taylor, Phi l . Mag. , A 39 (1979) p581. 7. R. Atkinson, S. Taylor, Phi l . Mag. , A 43 (1981) p979. 8. C. K . K i m , L . W . Hobbs., Oxid . Met., 45 (1996) p247. 9. R. Haugsrud, Corrosion Science, 44 (2002) p 1569. 10. R. Haugsrud, Corrosion Science, 45 (2003) p i289. 11. F. Czerwinski, J. A . Szpunar, W . W . Smeltzer, J. Electrochem. S o c , 143(1996)p3000. 12. F. Czerwinski, J. A . Szpunar, Corrosion Science, 39 (1997) p l47 . 13. F. Czerwinski, J. A . Szpunar, Thin Solid Films, 289 (1996) p213. 14. A . A . Moosa, S. J. Rothman, L . J. Nowick i , Oxid . M e t , 24 (1985) p i 15. 15. A . A Moosa, S. J. Rothman, Oxid. Met., 24 (1985) p33. 16. X . Peng, D . Ping, T. L i , W . Wu, J. Electrochem. Soc , 145 (1998) p389. 17. X . Peng, T. L i , W . Wu , Oxid. Met. 51(1999) p291. 18. X . Peng, T. L i , W . Wu , W . P. Pan, Mater. Sci . Engg. A . , 298 (2001) plOO. 19. A . Strawbridge, R. A . Rapp, J. Electrochem. S o c , 141 (1994) p l905 . 20. A . Strawbridge, R. A . Rapp, G . I. Sproule, R.J . Hussey, M . J. Graham, J. Electrochem. Soc , 142 (1995) p2301. 21. M . L . H . Green, Bestow, Proceedings, Microscopy of Oxidation-2, University of Cambridge, March 1993, p l37 . 22. C. S. Giggins, F. S. Pettit, J. Electochem. S o c , 118 (1971) p l782. 23. M d . Moniruzzaman, M . Maeda, Y . Murata and M . Morinaga, ISIJ International, 43, 3 (2003) p386. 57 24. D . T. Jayne, J .L. Smialek, Proceedings, Microscopy of Oxidation-2, University o f Cambridge, March 1993, p i83 . 25. K . Onal, M . C. Maris-Sida, G . H . Meier, F. S. Pettit, Proceedings o f Superalloys 2004, T M S , p607. 26. A . H . Dent, S. B . Newcomb, W . M . Stobbs, Proceedings, Microscopy of Oxidation-2, University of Cambridge, March 1993, p i 50. 27. M . S. Hook, Ph.D. thesis, University of Cambridge, 2003. 28. A . Akhtar, M . S. Hook, R. C. Reed, Met.Trans. A . , 36A (2005) p3001. 29. A . Akthar, S. Hegde, R. C. Reed, T M S Annual Technical Meeting, 13-17 February, 2005, San Francisco. 30. A . Akhtar, S. Hegde, R. C . Reed, J O M , 58 (2006) p37. 31. Y . L i , J . E . Mortal , Acta Mater., 50 (2002) p3683. 32. W . D . Hopfe, J. E . Morral , Acta Mater., 42 (1994) p3887. 33. T. J. Nijdam, L . P. H . Jeurgens, W . G . Sloof, Acta Mater., 51 (2003) p5295. 34. A . Engstrom, J. E . Morral , J. Agren, Acta Mater., 45 (1997) p i 189. 35. J. L . Smialek, G . H . Meier, Chapter 11, Superalloys II, Edited by C T . Sims, N . S. Stoloff, W . C . Hagel, Jhon Wiley & Sons, 1987, p293 . 36. B . D . Culli ty, Elements of X-ray Diffraction, Edison-Wesley Publishing Company, Inc., M . A . 58 

Cite

Citation Scheme:

        

Citations by CSL (citeproc-js)

Usage Statistics

Share

Embed

Customize your widget with the following options, then copy and paste the code below into the HTML of your page to embed this item in your website.
                        
                            <div id="ubcOpenCollectionsWidgetDisplay">
                            <script id="ubcOpenCollectionsWidget"
                            src="{[{embed.src}]}"
                            data-item="{[{embed.item}]}"
                            data-collection="{[{embed.collection}]}"
                            data-metadata="{[{embed.showMetadata}]}"
                            data-width="{[{embed.width}]}"
                            async >
                            </script>
                            </div>
                        
                    
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:
http://iiif.library.ubc.ca/presentation/dsp.831.1-0078654/manifest

Comment

Related Items