UBC Theses and Dissertations

UBC Theses Logo

UBC Theses and Dissertations

Post deposition treatment of thermal sprayed coatings John, George 1996

You are currently on our download blacklist and unable to view media. You will be unbanned within an hour.
To un-ban yourself please visit the following link and solve the reCAPTCHA, we will then redirect you back here.

Item Metadata


831-ubc_1996-0442.pdf [ 8.61MB ]
JSON: 831-1.0078445.json
JSON-LD: 831-1.0078445-ld.json
RDF/XML (Pretty): 831-1.0078445-rdf.xml
RDF/JSON: 831-1.0078445-rdf.json
Turtle: 831-1.0078445-turtle.txt
N-Triples: 831-1.0078445-rdf-ntriples.txt
Original Record: 831-1.0078445-source.json
Full Text

Full Text

POST DEPOSITION TREATMENT OF THERMAL SPRAYED COATINGS by GEORGE JOHN B. Tech., Banaras Hindu University, 1994 A THESIS SUBMITTED IN PARTIAL F U L F I L L M E N T OF THE REQUIREMENTS FOR THE DEGREE OF M A S T E R OF APPLIED SCIENCE in THE F A C U L T Y OF G R A D U A T E STUDIES (Department of Metals and Materials Engineering) We accept this thesis as conforming to the recnjjred standard THE UNIVERSITY OF BRITISH COLUMBIA August, 1996 © George John, 1996 In presenting this thesis in partial fulfilment of the requirements for an advanced degree at the University of British Columbia, I agree that the Library shall make it freely available for reference and study. I further agree that permission for extensive copying of this thesis for scholariy purposes may be granted by the head of my department or by his or her representatives. It is understood that copying or publication of this thesis for financial gain shall not be allowed without my written permission. Department of The University of British Columbia Vancouver, Canada Date 2 2 . / Q f t / 3 f c DE-6 (2/88) 11 ABSTRACT Thermal sprayed coatings are frequently used in corrosive environments, even when their major purpose is to provide wear or thermal resistance, rather than corrosion resistance. This includes Thermal Barrier Coatings (TBC), where high porosity is a desired feature to give good thermal protection. Porosity also provides increased strain tolerance of the coating. However, as this proves to be a limiting factor in the corrosion protection, a trade off is involved. This is because the interconnected porosity in TBCs allows the corrosive media to reach the coating-substrate interface, which eventually leads to delamination of the coatings. The current techniques to reduce coating permeability include polymer impregnation and laser glazing. Polymer impregnation into the pores is difficult and it severely limits the high temperature properties of the parent coating. Laser glazing produces segmented cracks in the coating which are useful in providing strain tolerance, but accelerate the rate of bond coat oxidation. This work addresses the problem of permeability of TBCs which is responsible for the poor corrosion protection offered by these coatings. The coatings studied included yttria-stabilized zirconia and alumina TBCs. A simple infiltration technique has been proposed using sol-gel ceramic precursors. The precursors included metal alkoxides which decompose at relatively low temperatures to their respective oxides. It has been proven that these precursors effectively penetrate into the pores and are beneficial in drastically reducing the coating permeability to gases. Electrochemical tests were carried out in 3.0 wt.% NaCl solution to study the effectiveness of the sealant in reducing the coating permeability. These potentiodynamic tests Ill as well as gas permeability tests show a considerable decrease in interconnected porosity with sol-gel modification of the coatings. Burner rig tests show an increase in sealed coating life time under thermo-mechanical fatigue conditions. The peel adhesion tests confirm an increase in the adhesion values with such a post deposition treatment. It is concluded that the advantages of sol-gel route to ceramics have been successfully utilized to post-treat porous thermal sprayed coatings in an attempt to reduce their permeability and to increase the life expectancy under corrosive and/or thermo-mechanical fatigue conditions. iv Table of Contents A B S T R A C T ii T A B L E OF CONTENTS iv LIST OF FIGURES vi LIST OF FIGURES IN APPENDIX I ix A C K N O W L E D G M E N T S x 1. INTRODUCTION 1 2. LITERATURE REVIEW 2.1 Thermal Sprayed Coatings 4 2.1.1 Thermal Barrier Coatings 5 2.1.2 Thermal Spray Methods 7 2.1.3 Plasma Spraying 10 2.1.4 Properties of Plasma Sprayed Coatings 15 2.2 Techniques for Sealing Porous Coatings 19 2.2.1 Polymer Impregnation 19 2.2.2 Laser Sealing of Thermal Sprayed Coatings 20 2.2.3 Sol-Gel Sealing of Thermal Sprayed Coatings 22 2.3 Other Coating Methods ' 24 2.3.1 Chemical Vapor Deposition 24 2.3.2 Physical Vapor Deposition 25 2.4 Methods of Testing Coatings 26 2.4.1 Electrochemical Testing of Coatings 26 2.4.2 Mechanical Tests for Thermal Sprayed Coatings 29 2.4.3 Gas Permeability Tests for Thermal Sprayed Coatings 34 2.4.4 Burner Rig Tests 35 2.5 Sol-Gel Chemistry 36 3. SCOPE A N D OBJECTIVES 41 4. EXPERIMENTAL 4.1 Processing and Characterization of Thermal Sprayed Coatings 44 4.2 Post Deposition Treatment of Thermal Sprayed Coatings 45 4.3 Electrochemical Tests for Permeability 48 4.4 Gas Permeability Tests 50 4.5 Hardness Tests 51 4.6 Peel Adhesion Tests 53 4.7 Burner Rig Tests 54 V Table of Contents (cont.) 5. RESULTS AND DISCUSSION 5.1 Precursor Properties 55 5.2 Electrochemical Tests 61 5.3 Gas Permeability Tests 70 5.4 Hardness Tests 73 5.5 Peel Adhesion Tests 74 5.6 Burner Rig Tests 78 5.7 Microstructural Analysis 85 6. CONCLUSIONS 91 7. R E C O M M E N D E D FUTURE WORK 93 REFERENCES 95 APPENDIX I 101 vi List of Figures Figure 1. Summary chart showing the high performance ceramic coating 2 service market according to applications for the year 2000 Figure 2. Sequence of events in thermal spraying 4 Figure 3. Potential turbine temperature benefits offered by thermal 6 barrier coatings Figure 4. Cross-section of a wire flame torch 8 Figure 5. Cross-section of a high velocity oxy-fuel torch 8 Figure 6. Cross-section of a detonation gun 10 Figure 7. Cross-section of a radial spray plasma torch 12 Figure 8. Effect of particle velocity in a plasma flame on coating quality 13 Figure 9. Schematic of an axial plasma torch 14 Figure 10. Cross-section of a splat 16 Figure 11. Low yttria region of zirconia-yttria phase diagram 17 Figure 12. Phases present as a function of composition in zirconia coatings 18 Figure 13. Coating life as a function of composition in zirconia coatings 18 Figure 14. Schematic of a typical polarization curve 28 Figure 15 Load vs. indentation depth from a depth sensing indentation test 31 Figure 16. Steps involved in the preparation of the peel adhesion test samples 33 Figure 17 Infiltration process involved in post-deposition treatment 48 Figure 18. Electrochemical corrosion measurement console 49 Figure 19. Cross-section of the permeability jig 51 Figure 20. Schematic of the depth sensing indentation apparatus 52 vii List of Figures (contd.) Figure 21. Schematic of the burner rig apparatus 54 Figure 22. Viscosity vs. concentration of alumina precursor 55 Figure 23. Viscosity vs. concentration of spinel precursor 57 Figure 24. Weight gained vs. infiltration time in coatings 59 Figure 25. Weight gained vs. no. of infiltrations 60 Figure 26. Polarization curves of alumina infiltrated specimens, 62 heat treated in a flame Figure 27. Limiting current of the infiltrated coatings 63 Figure 28. Polarization curves of alumina infiltrated specimens, 64 heat treated on a hot plate Figure 29. Limiting current of the infiltrated coatings 65 Figure 30. Polarization curves of spinel infiltrated specimens 66 Figure 31. Polarization curves of silica infiltrated specimens on Rokide 67 alumina coatings Figure 32. Gas permeability at room temperature on sprayed formed 71 alumina coating Figure 33. Gas permeability on sprayed formed alumina coatings, heat 72 treated at 1100°C for 24 hours Figure 34. Depth sensing indentation studies on sprayed formed zirconia 73 Figure 35. Peel adhesion tests on steel foils 76 Figure 36. Peel adhesion tests on nickel foils 77 Figure 37. Temperature profile of burner rig test 78 Figure 38. SEM micrograph of the coating cross section 80 Figure 39. Failure in burner rig test I 81 viii List of Figures (contd.) Figure 40. SEM micrograph of the interface after 500 cycles 82 Figure 41. Failure in burner rig test II 83 Figure 42. SEM micrograph of the bond coat fracture surface 87 Figure 43. SEM micrograph of the microcracked coating 88 Figure 44. SEM micrograph of the infiltrated coating 89 Figure 45. X-ray map of aluminum in the sealed zirconia-yttria coating 90 ix List of Figures for Appendix I Figure I-1. Thermo-gravimetric analysis of the wet alumina gel 101 Figure 1-2. Thermo-gravimetric analysis of the dry alumina gel 102 Figure 1-3. Thermo-gravimetric analysis of the dry silica gel 103 Figure 1-4. Thermo-gravimetric analysis of the dry magnesia-alumina gel 104 Figure 1-5. X-ray diffraction patterns of the phase transformation in alumina gel 105 Figure 1-6. X-ray diffraction patterns of phase transformation in silica gel 107 Figure 1-7. X-ray diffraction patterns of phase transformation in spinel gel 108 Figure 1-8. Energy dispersive spectroscopy of the corroded surface 109 Figure 1-9. Pourbaix diagram for the Fe-Cl"-H20 system at 25°C 110 Figure I-10. Energy dispersive spectroscopic analysis of the interface 111 Figure I-11. Failure criteria in the burner rig test 112 Figure 1-12. X-ray diffraction pattern of spalled interface showing presence of silica 113 ACKNOWLEDGMENTS The author wishes to acknowledge the guidance and encouragement received from his research supervisor Dr. Tom Troczynski. The assistance of Northwest Mettech, Pratt & Whitney and Waploc Ceramics in providing the thermal sprayed coatings, and the National Research Council, Vancouver, for the use of their facilities is greatly appreciated. Thanks are also extended to other faculty members, staff and fellow graduate students for their helpful advice. The financial support of The Science Council of British Columbia is gratefully acknowledged. 1 1. INTRODUCTION Ceramic coatings have come to play an increasingly important role in applications where high temperature, corrosion, oxidation and wear related problems are to be addressed. Their role is likely to increase in the future as well because of the diminishing resources of exotic materials, such as titanium and superalloys, currently employed in all high technology applications. Ceramic coatings are vital for numerous other applications including electrical, optical, lubrication, and as biocompatible coatings for prostheses and dental implants. The versatility of ceramic coatings make them prime candidates for certain material designs. Their diverse properties and their ability to be deposited on a variety of substrates make them a leading technology for fabricating materials previously considered unacceptable for certain demanding applications, such as in gas turbine engines. Although ceramics have many advantageous properties such as good wear and corrosion resistance, high strength at elevated temperatures and low thermal conductivity, application of monolithic ceramics in heat engines has not always been successful. This is because of the large variation in mechanical strength and the difficulty in fabricating shapes without inherent defects. This is enhanced by the well known brittle behavior of ceramics. Ceramic coatings on metal parts combine the reliability of the metal and the inherent advantageous properties of the ceramic. For example Thermal Barrier Coatings (TBC) are used in heat engines to increase their efficiency, reduce fuel consumption and to allow the use of lower cost fuels. It is postulated that an increase in the engine operating temperatures from 1000°C to 1400°C would cause an increase in the efficiency by 6-10%. 2 Because of these advantages, thermal sprayed coatings have found applications in many industries which can be broadly divided into three categories: i) manufacturing (e.g. automotive, land-based turbine engines), ii) infrastructure and repair (e.g. petrochemical plants, cutting tool inserts) and iii) high technology (e.g. aircraft and aerospace engines). In the advanced ceramic segment, ceramic coatings are the fastest growing market. An estimate made by the Freedonia Group (Cleveland, USA) predicts that the ceramic coating service market in the United States alone will grow close to $940 million by the year 2000 [1]. It is also predicted that the aerospace and the cutting tools markets will absorb the majority of the ceramic coating service market, Fig. 1. Wear parts & industrial Auto, diesel and land based Heat C U t t ' n 9 1 0 0 ' i P S e r t S t u r b i n e e n 3 ' n e s 60% Aircraft/Aerospace engines Figure 1. Summary chart showing the high performance ceramic coating service market according to predicted applications for the year 2000 [1] There are a wide variety of processes used to deposit ceramic coatings. Most of these fall under the category of physical vapor deposition (PVD), thermal spraying (TS) and chemical vapor deposition (CVD). There are also wet chemical techniques using liquid precursors such as painting, slurry spraying and sol-gel. A range of hybrid coating techniques 3 such as ion assisted deposition have been developed to meet certain specific application needs [2]. This thesis focuses on coatings that are thermally sprayed. However, the porous and microcracked structure of these coatings has a negative influence on their corrosion resistance. The present work addresses the problem of gas permeability of TBCs which can lead to premature delamination due to interfacial corrosion. The coatings studied included yttria-stabilized zirconia and alumina TBCs. A simple infiltration technique has been proposed using sol-gel ceramic precursors. Furthermore, the beneficial aspects of such a post deposition treatment on the mechanical, thermal and corrosion properties of the coating have been determined. 4 2. L I T E R A T U R E R E V I E W 2.1 Thermal Sprayed Coatings Thermal Spraying represents a cost effective ($0.02/inch for a typical thickness of 0.1-0.3 mm), application oriented alternative to conventional surface protection measures [3]. Thermal sprayed coatings (TSC) are produced by deposition of molten or near molten particles which are heated in a thermal source and accelerated towards the surface to be coated, Fig. 2. These droplets solidify on impact into thin lamellar "splats" which build up on each other to form the coating. In general the powder particles used are in the size range of 40-60 um. THERMAL SPRAYING SEQUENCE Solid or powder material Molten particles Flatten Melts in accelerate on flame or in gas substrate arc ^ stream ^ <# <P Figure 2. Sequence of events in thermal spraying [4] The outstanding advantage of thermal sprayed coatings is that they can be applied without significantly heating the surface to be coated (substrate temperatures can be maintained at ~ 250-300°C). Secondly the process offers the flexibility to spray a large range of materials. Almost any ceramic, metal or cermet that melts without decomposing can be sprayed by this method. However, the bond strength and the cohesive strength of the coatings 5 varies considerably depending on the process, the substrate preparation and also the materials being sprayed. Substrates are generally grit blasted to increase coating adhesion through mechanical interlocking of the rough surfaces. 2.1.1 Thermal Barrier Coatings (TEC) The gas turbine engines of commercial aircrafts typically utilize superalloys with melting point in the range of 1230°C to 1315°C [4]. However, the temperature in the combustion gas chamber often exceeds 1370°C [4]. To avoid structural failure, for example through melting, creep, oxidation or thermal fatigue, combustors and turbines components have to be cooled with compressor discharge air. This is achieved in turbine blades by the use of intricate cooling passages through the components to bleed the air. However, to maximize efficiency it is desirable to minimize the use of cooling air. In contrast to metals, ceramics exhibit better structural and thermal capabilities at high temperatures. The heat conductivities of most ceramics lie between 3-15 W/mK (2-3 W/mK for zirconia) [5]. These are very low values in comparison to the thermal conductivities of commonly used structural metals which have values as high as 60 W/mK [5]. In general ceramics are more resistant to creep, oxidation, corrosion, erosion and wear. Protecting components by application of a thin layer of ceramic, which acts as a thermal barrier facing the elevated temperatures, is a cost effective method. Thermal Barriers Coatings (TBCs) reduce metal temperatures by ~ 170°C, provide improvement of up to 1% in thrust-specific fuel consumption and reduce thermal fatigue cracking [4]. In this way the advantageous properties 6 of the ceramic are utilized, while the metal provides mechanical and structural integrity to the system. TBCs offer a number of benefits .such as a reduction in substrate temperature (to a degree depending on the thickness of the ceramic layer), fuel savings and improved durability. The typical temperature reduction of approximately 170°C achieved by the utilization of 0.2 -0.3 mm thick TBCs is greater than the cumulative gains made in the temperature capability of superalloys over the past 25 years, Fig. 3. It is estimated that the use of TBCs in all high temperature turbine airfoils, in a typical modern gas turbine engine, would result in an annual fuel saving of as high as 10 million gallons for a 250 fleet aircraft [4]. 1965 1975 Year 1985 Figure 3. Potential turbine temperature benefits offered by thermal barrier coatings [4] The search for an insulating material with thermal expansion closely matching that of the metal substrate has focused the attention of TBCs on zirconia. However pure zirconia undergoes polymorphic transformations on thermal cycling accompanied by large volume changes and inevitable cracking [6]. To overcome this problem, zirconia has to be stabilized with yttria, magnesia or calcia (refer to Section 2.1.4). The current consensus is that zirconia 7 with an yttria content in the range of 6-9 wt.% forms the best thermal barrier coatings [6]. Coatings within this composition range consists of the non transformable tetragonal (T) phase [6]. The reluctance of this phase to undergo the potentially damaging tetragonal to monoclinic transformation is believed to be an important factor affecting the durability of coatings. The phase stability of zirconia TBCs is further discussed in Section 2.1.4. The TBC system consists of a ceramic "top coat", typically stabilized zirconia, and an underlying metallic " bond coat". Typical values of thermal expansions of top coat (zirconia) and substrate (mild steel) are 10"5/°C and 17xlO" 6/°C respectively [5]. The composition of the bond coat is tailored to buffer the mismatch in thermal expansion of the substrate and the ceramic. It also provides a physical tie, in terms of better adhesion, between the ceramic and the underlying structure. Typical thermal barrier coatings sprayed are zirconia stabilized with 6-12% Y 2 0 3 deposited on a MCrAlY (where M = Ni, Mo or Fe) bond coat. 2.1.2 Thermal Spray Methods Wire Flame Spraying Figure 4 shows the cross section of a typical wire flame spray gun [7]. The feed stock material, in the form of a wire or a rod, is fed into the source of thermal energy obtained by the combustion of fuel gases. A stream of compressed air atomizes the molten material and propels it onto the grit blasted substrate. The maximum temperature that can be achieved in this process is limited to 1500-1600°C, depending on the fuel gases [7]. Likewise, the particle velocity is restricted to around 180-190 m/s [7]. 8 The HVOF process is an emerging technology. The first development of this spraying technique was only about 10 years ago. In contrast to the previous wire flame spray process, combustion takes place in an external water cooled chamber, Fig. 5. Gas Feed Port Cooling Passage Nozzle Combustion Chamber Oxygen - F u e | Figure 5. Cross section of a high velocity oxy-fuel torch [7] The maximum temperature that can be reached in this process is limited to ~ 2900°C and this is insufficient to melt certain refractory oxides [7]. However, the high particle 9 velocity (500-1000 m/s) ensures dense coatings [7]. Thus this technique is well suited for spraying dense metals, alloys and cermets. Coatings sprayed using HVOF possess good bonding and mechanical strength in addition to high density [7]. Detonation Gun (D-Gun) Detonation gun is a promising technique used to deposit high quality oxide ceramic coatings. Under detonation conditions the flame temperature is ~ 4000°C, about 1000°C higher than in a free burning gas [7]. The fuel mixture is generally a mixture of acetylene, propane or propylene, and oxygen gas. The process is similar to that of an internal combustion engine. As shown in Fig. 6, the fuel mixture is fed into the chamber and a spark detonates the mixture. The powder is fed into this ignited mixture and due to the explosive nature of the detonation, the molten particles are accelerated towards the substrate reaching velocities in excess of 1 km/s. The spark plug produces around 3-7 detonations per second and after each detonation a pulse of nitrogen purges the barrel to prevent choking [7]. D-Gun sprayed coatings exhibit low porosity and higher microhardness in comparison to other TSCs, including plasma sprayed coatings [8]. The pulsed and localized powder feeding into the barrel guarantees high uniformity in the particle spray and the ability to work with several feeders simultaneously. The coatings produced by a D-Gun are comparable to monolithic and sintered materials. Very dense coatings can be sprayed by D-Gun because of the high particle velocity [8]. 10 Spark Plug Work Piece / Powder I Nitrogen Gas Barrel 4 Acetylene Gas Oxygen Gas Figure 6. Cross section of a detonation gun [7] 2.1.3 Plasma Spraying Plasma, often referred to as the fourth state of matter, is essentially a partially ionized gas comprised of molecules, atoms, ions and electrons in a state of electrical neutrality. The presence of free electrons is responsible for its high electrical conductivity (8 kA/Vm for Ar at 15,000°C) which distinguishes it from ordinary gases at ambient temperatures [7, 9]. In general there are two different types of plasmas: thermal and cold plasmas. Thermal plasmas are characterized by their high specific enthalpy (350x10 kJ/kg for Ar at 15,000°C) and the equality in electron and heavy particle temperatures (Te=Th) [7]. These plasmas are usually generated at atmospheric or high pressures. Cold plasmas are characterized by their strong deviation from kinetic equilibrium (Te»Th) and are generated at low pressures [7]. Thermal plasmas are widely adapted for the deposition of protective coatings and forming of near net-shaped bodies through thermal spraying. 11 The plasma gun, Fig. 7, begins operation when a pulse of current creates an arc across the gap of the electrodes. As the arc forms, the electrons are torn away from the atoms. The electrons and the positive ions they leave behind are accelerated towards the anode and cathode respectively. Frequent collisions transfer energy from the electrons to the positive ions, accelerating them until the plasma reaches a thermal equilibrium. In addition to the enthalpy reflected in its high temperature (~ 15,000°C), the plasma contains enthalpy associated with the ionization of the gas atoms and the dissociation of molecules. A plasma of hydrogen, whose molecules must be dissociated into two atoms before they can be ionized, has a higher enthalpy at a given temperature than a plasma of argon [7, 9]. Normally the plasma gases used are H 2 , Ar, He and N 2 . Hydrogen as the plasma gas is beneficial because it increases the heat content of the plasma flame (enthalpy of Ar is 350x10 kJ/kg while that of H 2 is 750x10 kJTkg at 15,000°C) [7]. Though most suited, hydrogen is usually mixed with Ar to reduce the cost and explosion risks. Progress in the use of thermal barrier coatings has been facilitated by the introduction of plasma spraying process which superseded the earlier flame spraying techniques. Plasmas produce a far higher particle stream temperature, thus allowing the efficient spray of refractory materials like zirconia (Tm= 2750°C). Radial Plasma Spray Torch The plasma spray system consists of a spray gun, a vibratory powder feed mechanism, a water cooling system for the anode, a power source and a control panel. The basic design of 12 the plasma torch has not changed much since its inception in the early 1970's. It consists of a cone shaped cathode inside a cylindrical anode, which forms the nozzle, Fig. 7. An inert gas (usually argon, argon/hydrogen or argon/helium) flows through the space between the electrodes where it is ionized to form a plasma. The powdered coating material is injected into the flame either within the nozzle or as it emerges from the outer face of the anode. The flame accelerated molten particles are propelled towards the target where they solidify and accumulate over a number of passes to form a coating [9]. For this process to succeed, a number of delicately balanced criteria must be satisfied. The particles must take enough heat from the flame to melt thoroughly, but not excessively so as to overheat and vaporize. At the same time, the particles must travel fast enough to flatten and spread out when they strike the target, flowing into crevices and strongly adhering to the surface. The pressure of the carrier gas must be adjusted so as to blow the particles into the plasma stream and not through it, Fig. 8 [10]. Figure 7. Cross-section of a radial spray plasma torch [9] 13 SOME POWDER PASSES THROUGH PLASMA UNHEATED. POWDER RIDES ON TOP OF PLASMA. PARTICLES TRAVEL IN CENTER OF PLASMA. Figure 8. Powder injected at different velocities into a plasma flame [10] Until recently, manipulating these variables and many others to produce high quality coatings was at best a matter of guesswork. Now fundamental studies are underway at various laboratories to lay theoretical foundations for designing plasma systems [11-14]. Much work also has been done to improve coating life expectancy. For example using Laser Doppler techniques the velocity of the particles can be analyzed (~ 600-800 m/s) and the effects of average velocity distribution on the coating properties can be studied [11]. Axial Plasma Spray Torch In a conventional plasma gun, Fig. 7, the powder is fed radially into the plasma stream outside or inside the torch nozzle. With this mode of injection, it is difficult to get all the TOO MUCH CARRIER GAS (BAD) CORRECT CARRIER GAS (GOOD) 14 particles into the central region of the plasma flame. To increase the dwell time of the powder within the plasma jet, the most elementary idea is to inject the powder along the axis of the plasma jet [15], Fig 9. Plasma Axial Powder Feed Figure 9. Axial Plasma Torch1 The mode of generation of the plasma in the axial torch remains the same as in the radial torch. But in this case the powder is injected axially into the arc. This injection mode has many advantages. First of all, it ensures that all the powder is exposed to the plasma. The longer heating zone assures that the particles are more likely to melt. This gives the process more flexibility in terms of the injection parameters. Axial spray also ensures that all the particles undergo the same thermal history. Thus it is reasonable to expect that the consistency of such a coating would surpass that of a radial spray. Axial spray torch eliminates the negative effect of radial powder injection on plasma laminarity [16]. Reduced turbulence substantially increases the plasma loading capacity and allows spray rates two to three times higher than that possible with conventional (radial injection) systems [16]. The nearly full entrainment of the powder stream within the plasma plume increases the deposition efficiency by 50% to 80% compared with other thermal spray systems [16]. Consequently powder costs ' Axial III Plasma Torch manufactured by Northwest Mettech, B.C., Canada 15 are reduced by 50 to 70% and the gas costs per unit area of coating are reduced by 50 to 60% [16]. 2.1.4 Properties of Plasma Sprayed Coatings When fast moving droplets of the coating material hit the target, they flow into the microscopically tortuous shape of the roughened surface locking into its irregularities. T E M of single splats shows their internal structure as a mosaic of crystalline grains containing many flaws. The fine network of microcracks suggests an extremely fast quench, in perhaps 10"6 s. On impact a low mound of solidified material forms, while the remaining melt spills off it, finally hardening into a raised rim, Fig. 10 [9]. The substrate acts as a heat sink and as more splats flatten out on the surface, the solidification front moves upward through the splat. The result of this process repeated at a rate of 106 /s is a coating that can typically range in 5 3 thickness from 10" mtolO" m. Thermal sprayed ceramic coatings in particular reveal a multitude of flaws, e.g. cracks formed as the ceramic cools, and voids. Such flaws can prove disastrous for a coating exposed to mechanical stress. In addition, such a coating would fail to provide protection against corrosion, if the flaws extend all the way to the substrate. At the same time however, it is the porosity that suits plasma sprayed ceramic coatings to one of their most important applications: as Thermal Barriers Coatings (TBC). Porosity increases the coating's thermal insulating capability (for YSZ thermal conductivity & d e n s e = 2-3 W/mK while £ p o r o u s = 0.6-1.2 W/mK) [17]. Secondly, pores and microcracks decrease the 16 stiffness of the material (for YSZ £ d e n s e = 170 GPa while Eporous = 7-48 GPa) [17]. This enables the coatings to survive differential strains that result from repeated exposures to high temperatures of the coating/substrate assembly. n SPRAYING Figure 10. Cross-section of a splat [9] Yttria Stabilized Zirconia (YSZ) TBCs Pure zirconia cannot be used in coatings because of its polymorphic transformations accompanied by large volume changes (5% volume expansion upon tetragonal —»monoclinic transformation) [18]. The high temperature phases can be stabilized to room temperature by the addition of yttria (or magnesia) in sufficient quantities, Fig. 11 [18]. During spraying the molten zirconia-yttria particles impinge the cold substrate and solidify into the high temperature fluorite cubic phase. In conditions of moderate cooling rate, a phase separation normally occurs into high yttria cubic (F) phase and low yttria tetragonal (T) phase. Depending on the temperature and composition, the tetragonal phase is capable of further undergoing the potentially destructive transformation to monoclinic phase during 17 cooling, Fig. 11. Under certain conditions, as achieved in plasma spraying, the tetragonal phase can be retained metastably to room temperatures. In this process, the cooling rate is fast enough (105-106 °C/s) to allow much of the cubic (F) phase formed at high temperatures to transform to non-transformable tetragonal (T') phase by a diffusionless transformation. It is postulated that the fast quench may have suppressed the transformation of low yttria tetragonal phase [19]. (L) Cubic 1600"C Tic Line 1400°C Tic Line * Monwclinic • INon Transformable Tetragonal (T') M O L E % YO, s Figure 11. Low yttria region of zirconia-yttria phase diagram [18] This retained phase has the same chemical composition as the cubic phase, but is stable with respect to monoclinic transformation [18]. It has been shown that the non-transformable tetragonal phase decomposes during annealing at elevated temperatures into the equilibrium high temperature phases [19]. The rate of transformation is dependent on temperature (relatively sluggish at temperatures below 1400°C) and composition. It is observed that the 18 coating performance begins to deteriorate when the level of yttria drops below 4.4 wt.% and powdering of the ceramic due to the chemical instability is noted [6]. From Fig. 12 it is clear that the maximum tetragonal content (also a region of small monoclinic phase content) coincides with 6-9% YOj 5 region [6, 19]. This is also the region of maximum coating durability, Fig. 13. %Y0j 5 *Y0 j 5 Figure 12. Calculated mole fractions of Figure 13. Coating life as a function of (A) monoclinic, (o) tetragonal, (V) cubic composition [6] phases as a function of composition [6] Wu et al [20] reported that the T' phase does not transform on heating for 655 hours at 1000°C or 70 hours at 1150°C, but the transformation begins after 30 hours annealing at 1300°C. Jasim et al [21] reported that transformation of the T' phase to give a mixture of (F) and (M) phases was complete in 120 hours at 1300°C. The essential difference between the low-yttria, metastable (transformable) tetragonal (T) phase and the high-yttria, quenched in (non-transformable) tetragonal (T') phase lies in their thermodynamic stability. At room temperature, the chemical free energy of the T phase is greater than that of the M phase [6]. Thus the metastable low-yttria tetragonal phase is chemically unstable, but can be retained if it 19 is stabilized by surface or strain energies [6]. The high-yttria tetragonal phase (T') has a lower free energy than the corresponding high-yttria monoclinic phase. Minor amounts (2-3%) of monoclinic phase can have significant influence on coating properties [6]. For example, the martensitic transformation that accompanies cooling produces a favorable network of cracks in the coating. A fine network of cracks can serve to dissipate the energy associated with the extension of large cracks in brittle materials. By reducing coating stiffness, it also helps the coating to survive thermal strains due to difference in thermal expansion coefficients of the coating and the substrate. 2.2 Techniques for Sealing Porous Coatings The problem of interconnected porosity, which impairs the corrosion resistance, and surface roughness, which affects the aerodynamic performance of sprayed coatings in aerospace applications [22, 23], can be overcome by modifying the characteristics of the surface. Post treatments are designed to densify the surface and thus make it impermeable to corrosive attack and also to improve the wear properties of the coating. Mentioned below are few of the many methods that can be applied to coatings to achieve this purpose. 2.2.1 Polymer Impregnation Impregnation of coatings with organic sealants can be done under a vacuum or under pressure. The major limitation of this method is the difficulty of penetration of the resins into 20 the coating. These resins have viscosities in the range of 200 mPa s and so the impregnated depth is small (10-100 urn) even under external pressures of 5 MPa [24]. As a result, the corrosion resistance of the impregnated coating decreases when the polymer layer on the coating is ground or worn away [24]. The corrosion resistance in aqueous solutions is enhanced as a result of polymer sealing, but the high temperature and wear properties of the coating are limited by the properties of the sealant. Thus such a sealing technique is not useful for coatings intended to withstand high temperature and wear. 2.2.2 Laser Sealing of Thermal Sprayed Coatings A few investigations have been carried out to explore the feasibility of using high power energy beams to optimize the properties of plasma sprayed layers by decreasing porosity and improving the surface finish [2, 25-27]. High power beams, such as ion beams, and lasers, such as continuous wave C 0 2 lasers (0.5-1 kW), have been studied. Lasers have many advantages compared to ion beams, in particular being more economical and not requiring a vacuum. The efficacy of melting ceramic coatings with lasers has been studied [25-27]. The problem of cracking when ceramic coatings are melted by the laser has been recognized. The phenomenon of delayed fracture after laser melting was also reported [27], although no satisfactory explanation has been given for this unusual occurrence. In some cases controlled cracking introduced to the coating can be beneficial. For instance it has been suggested that cracks perpendicular to the coating surface can help in accommodating the strains during thermal cycling [23, 28]. However, longitudinal cracks that run parallel to the 21 coating surface are detrimental since they can lead to premature spallation of the coating [28]. It has been reported that a power density of 50 W/ram2 is sufficient to melt plasma sprayed zirconia to a depth of 0.01 cm at a scanning speed of 80 cm/min [25]. In the pulse mode, it is possible to obtain a desired melt thickness by adjusting the pulse parameters such as the pulse width, frequency and scan speed. Apart from sealing the pores, laser melting can provide a smoother surface as compared to thermal spraying. Such a smooth surface would reduce aerodynamic losses and the problem of foreign matter such as slag sticking to the surface of the coating, which can be responsible for localized loss of the coating [27]. The variation in the cracking behavior of various ceramic coatings can be rationalized by estimating the elastic stresses that develop during cooling of a melted spot. At high temperatures (> 0.6 Tm) these stresses can be relieved by plastic deformation. But below this temperature tensile stresses build up in the coating during cooling. This can thus result in thermal shock fracture of the coating when the fracture strength of the fully melted area is exceeded. It has been shown that for an infinitely fast quench the critical temperature change, ATC, to produce cracking due to thermal stresses is [29] : a , ( l - v ) ATC = J- (1) Ea where oy is the fracture stress, v is Poisson's ratio, E is the Young's Modulus and a is the coefficient of thermal expansion. Substituting standard values for yttria stabilized zirconia (YSZ) in Eq. 1 (a ~ 650 MPa, E ~ 170 GPa, a ~ 10"5 /°C and v ~ 0.29) the critical temperature change (A7/c) is 270°K. Thus under an infinitely fast quench, cracking can be prevented if the difference between the lowest temperature for plastic flow and the processing temperature is less than 270°K. Assuming that 22 plastic flow occurs approximately above 60% of its melting point, i.e. 0.6 x (2973°K) = 1785°K, an unusually high processing temperature of about 1240°C is required to prevent thermal shock cracking [29]. Additional thermal stresses would build up upon cooling of the system due to the differences of thermal expansion coefficients of the coating (a ~ 10"5 /°C) and the metal substrate (a ~ 1.7x10"5 /°C). In addition to these demands for very high processing temperatures, there are other limitations of laser post-treatment that have been reported. For example, laser glazing does not have any significant effect on reducing the oxidation rate of the bond coat [30]. However, laser glazing produces segmented cracks which are beneficial for accommodating the strains which can result from either thermal cycling or from oxidation of the bond coat [30]. This leads to an enhancement in the TBC life. The improvement in service life of these glazed coatings can be thus related to better accommodation of thermal stress, rather than better corrosion resistance [30]. As it can be expected from the high power levels required, laser sealing is an energy intensive process which can upset the overall cost benefits of thermal sprayed coatings. 2.2.3 Sol-Gel Method of Sealing Thermal Sprayed Coatings Infiltration of liquids into a porous medium consisting of particulate materials can be described by the Washburn's model which considers the porous medium as a constant cross-section capillary [31]. For simplicity of treatment thermal sprayed coatings are assumed to be 23 a porous body behaving as an assemblage of very small capillaries. For such a system of capillaries the following holds true [32]: e = l ^ r t (2) 2r\ In this model, the infiltrated depth (/) is parabolic in time (r) with its dependence on fluid viscosity (Y|), surface tension (y), pore radius (r) and the wetting angle (0). In the above calculation it is assumed that the external pressure is small as compared to the capillary pressure, 2ycos9/r, which is the main driving force for infiltration. Therefore, the physical properties required for good penetration of the sealant material into the porous coating are i) low viscosity, ii) high surface tension and iii) good wetting. In addition, the sealant must be resistant to the corrosive media present, temperature and must have neutral or beneficial influence on the mechanical properties of the coating. The use of sol-gel techniques to densify the coatings with materials similar to that of the parent coating has been explored [33, 34]. This method involves infiltrating the porous coatings with ceramic precursors. On heat treatment at relatively low temperatures these precursors decompose in situ to their respective oxides. This post-treatment serves to reduce the permeability of the coatings by filling up the pores. The depth of penetration of the precursor in the coating is controlled by the immersion time in the sol. This method of sealing offers several advantages. First, the sealant being a ceramic ensures that the high temperature capabilities of the coating are not impaired. Second, the sol-gel technique produces fine oxide powder (2-5 nm) which is capable of filling the micropores and microcracks in the coatings. Third, the low temperature synthesis of ceramics via the sol-gel route ensures that the 24 structural integrity of the component is not deteriorated as a result of such a post treatment. However, not much work has been done in this area to characterize the effects of such a treatment on the thermal insulation properties, adhesion, durability and hardness of the coatings. This thesis presents a detailed discussion on the use of sol-gel method to modify thermal sprayed coatings. Refer to Section 2.5 for further discussion of the sol-gel technique. 2.3 Other Routes to Ceramic Coatings 2.3.1 Chemical Vapor Deposition (CVD) C V D coatings are fabricated by the thermal reaction of gaseous compounds on the surface of the substrate [3]. The major types of reactions used are : Pyrolytic: A X ( g ) > A ( s ) + X ( g ) Reduction: 2 A X ( g ) + H 2 ( g ) > 2A ( s ) + 2HX ( g ) Exchange: A X ( g ) + E ( g ) > A E ( s ) + X ( g ) One of the major advantages of C V D is its ability to allow deposition in all directions, i.e. it is not a line of sight method like thermal spraying. Thus complex shapes can be coated if uniform heating and gas flows are maintained. Elaborate gas distribution systems are sometimes required to achieve the latter [3]. C V D method is more suited for batch processes where hundreds of components are to be coated at a time in a furnace. A major limitation of C V D process is the high substrate temperature requirement, usually in the range of 800°C to 1000°C, in order to achieve high deposition rate. The high temperature requirement limits the 25 choice of substrate materials that can be coated without thermal distortion or other detrimental changes in the microstructure and mechanical properties. Few modifications of the CVD process, such as plasma, microwave and laser assisted C V D have been developed, which allow the deposition of coatings at lower temperatures. The toxicity and corrosivity of the reactants (such as SiCl 4 , C H 4 , NH 3 ) as well as the reaction by-products are the limitations of this process. The thickness of the coatings produced are usually less than 50 um. C V D coatings replicate the surface of the substrate and in many cases thin coatings < 10 um provide sufficient wear resistance. 2.3.2 Physical Vapor Deposition (PVD) PVD coatings are produced by a variety of techniques including thermal evaporation, electron beam evaporation, ion beam plating and sputtering. In thermal evaporation, the coating material is simply evaporated or sublimed using electric, resistive or inductive heating. Electron beam evaporation is a variant of thermal evaporation in which an electron beam is used to evaporate the target [2]. In reactive evaporation, a second element (gaseous) is introduced into the chamber to react with the evaporated material to form the coating [3]. In sputtering, the coating material is electrically charged and vaporized by the impact of high energy ions. Similarly in ion implantation, gaseous and metallic ions generated in an ion source are accelerated towards the charged substrate. The kinetic energy of the ions causes them to penetrate into the surface or deposit on the surface depending on the bias of the substrate. The coating thickness usually range from 10 to 15 pm for most wear resistant 26 applications, 25-75 um for corrosion resistant coatings and greater than 75 um for use as thermal barriers and clearance control coatings. PVD coatings require clean substrate surfaces and are used for a wide variety of purposes including electronics/semiconductors, wear resistance, corrosion resistance, thermal barriers and optical films. Deposition is usually carried out in batches, in moderately sized vacuum systems. PVD coatings have a typical columnar microstructure which is beneficial for accommodating the strains resulting from thermal cycles or from oxidation of the bond coat [28, 30]. 2.4 Methods of Testing Coatings 2.4.1 Electrochemical Testing of Coatings Electrochemical techniques offer a potentially easy and quick way to qualitatively determine the permeability of coatings in aqueous solutions. The basis of this approach lies in the implicit understanding of the chemical reactions that are responsible for the corrosion of the substrate covered by a permeable coating. For example, polarization technique measures the current associated with metal corrosion as a function of the change in applied potentials. This test indicates how the coated substrate responds to the shifts in the oxidizing power of the electrolyte. The specimen is made one of the electrodes and is anodically polarized in the corrosive medium. The potential of the specimen is measured using a reference electrode. The corrosion current is a measure of the permeability of the coating [35, 36]. In addition, 27 electrochemical methods can be used to observe and understand substrate passivity in specific environments. Figure 14 shows a schematic diagram of a typical polarization curve in aqueous electrolyte. The curve can be divided into three regions. The part "A" of the curve corresponding to an increasing current density with increase in electrode potential represents an activation controlled process. In this region corrosion is controlled by the movement of species between the Inner Helmoltz Plane (IHP) and the Outer Helmoltz plane (OHP). The IHP corresponds to a layer of ions closest to the surface of the metal being corroded, while the OHP is the closest layer of fully solvated ions in solution. Thus this part "A" of the polarization curve refers to an electrochemical process that is controlled by the reaction sequence at the metal-electrolyte interface [37]. Increasing the anodic potential on the specimen provides increasing driving force for the cations to diffuse across the activation barrier to the OHP. This leads to an increasing anodic current. The region "B" of Fig. 14, corresponding to a limiting current density, represents a mass transfer controlled process. Mass transfer controlled process, in contrast to activation controlled process, is controlled by the diffusion of species in the electrolyte. There is a possibility of film (corrosion product) formation depending on the pH of the solution. The formation of the film slows down the corrosion process because of limited mass transfer across this film. Mass transfer controlled processes are frequently encountered in freely corroding bodies, where the concentration of reducible species is small and can be attributed to i) slow transport of soluble species from the corroded surface to the bulk solution and ii) slow transport of oxidant to the corroding surface. Thus, the resulting current is approximately 28 constant in region "B", although the electrode potential increases. The magnitude of the average current in region "B" is used in this work to qualitatively estimate the effectiveness of sealing of the porous coatings. c t • i i B I T Current Density (i) Figure 14. Schematic of a typical polarization curve The region "C" of Fig. 14 corresponds to a situation where localized breakdown of the passive film causes pitting, leading to enhanced corrosion. Pitting is aggravated in the presence of chlorides. At potentials higher than the potential for the decomposition of water, oxygen evolution results in large increases in the anodic current. The discussion of the exact nature of the film breakdown and mechanism of pitting is beyond the scope of this work. As the corrosion reactions in aqueous solutions are electrochemical in nature, the corrosion rates can be expressed in terms of the equilibrium corrosion current. Thus for species going into solution according to the reaction : 29 A = An+ + ne" the corrosion rate (CR) can be expressed in pg/s or in the commonly used units of mils per year (mpy) by [36]: C R J - A ( E W ) F C R = o->29yEW) m p y < 3 ) where Icorr is the corrosion current density in pA/cm 2, d is the density in g/cm3, A is the area of the corroding surface in cm and EW is the equivalent weight of the species A. The corrosion current density Icorr is measured by extrapolation from the anodic and cathodic Tafel slopes. In the past few years this testing procedure of coatings has attracted considerable amount of interest. Further details of this technique, as applied, to the coatings industry, can be found in [24, 35-40]. 2.4.2 Mechanical Tests for Thermal Sprayed (TS) Coatings Hardness Tests for TS Coatings The evaluation of mechanical properties of TS coatings by microhardness is difficult due to the large variations in hardness, elastic modulus and fracture toughness across the thickness of the coating. Moreover TS coatings are anisotropic because of their lamellar structure. Hardness testing involves pressing an indentor of known geometry into the surface of the material under a given load. Upon unloading there is some elastic recovery of the 30 impression. For a Vickers indentor, the hardness can be measured from the indentation diagonals (d) according to the following relationship : Depth Sensing Indentation (DSI) is a technique where the load and the depth of penetration of the diamond indentor are continuously monitored and recorded. Consequently information such as elastic recovery, hardness and the work of indentation can be obtained, without having to measure the diagonals of impressions. Since the measured depth has both elastic and plastic contributions, the elastic penetration depth has to be subtracted to obtain the plastic penetration depth. Extrapolation of the initial slope of the unloading curve to zero load gives the plastic depth of indentation (hp), Fig. 15. The Young's Modulus can be determined from the slope of the unloading portion (dP/dh) of the indentor load vs. depth curves, Fig. 15 On the assumption that for the Vickers indentor the ratio of the diagonal to depth is 7, the hardness (H) can be calculated from the following equation [41]: #=1854-^-d2 (4) [41]. (5) where P is the load and hp is the plastic depth of indentation. 31 20 40 60 80 100 120 140 160 180 Indent depth (nm) hp S l o p e d P / d h 0 Figure 15. Load vs. indentation depth data from a depth sensing indentation test [42] The uniqueness of DSI technique is that the hardness can be directly read without any of the inaccuracies introduced by subjective measurements by the operator [41, 42]. The load vs. depth plots are fingerprints relating the microstructure to the micro-mechanical properties of the surface. This measurement technique is however rather sensitive to the surface roughness of the sample and to external vibration. Adhesion Tests for TS Coatings The adhesion of coatings to substrates is an important criterion for the structural integrity of coatings. The effects of infiltration and heat treatment on the interfacial adhesion strength have been assessed in this work. There are several adhesion tests for thermal sprayed coatings : 32 i) Stress Based Methods: pull test (ASTM C 633-79), peel tests, indentation, and scratch techniques [43]. ii) Linear Elastic Fracture Mechanics (LEFM): cantilever beam and notched beam tests [44]. Because of the previously documented difficulties with adhesion measurement techniques [43, 44], a modified version of A S T M D-3167 peel test for coatings was developed at the University of British Columbia [45-49]. Peel testing was traditionally used to measure adhesion of tapes and glues [50]. This method allows the simultaneous measurement of crack tip location and the opening force. There are several advantages of using this technique. It allows testing of thermal sprayed coatings with a high degree of resolution and repeatability. Apart from being simple to perform, the results are easy to interpret. Because the crack position is directly controlled, the test shows the variation in adhesion along the sample and can indicate the presence and location of flaws. This test is stable, flaw insensitive and thus produces less scatter than the A S T M C 633-79 bond test. Also as the crack front is controlled, this test provides information about the entire interface rather than the largest flaw in the coating. The results of a peel adhesion test are commonly reported as the force per unit width (N/m). Figure 16 shows the main steps involved in the peel sample preparation as developed for thermal sprayed coatings [46]. The metal foil is soldered or glued to a large copper block (heat sink) with a silicone adhesive. The foil edges are masked and the test area is grit blasted, Fig. 16a. The coating is then deposited onto the sample, Fig. 16b. An epoxy is used to bond the coated assembly to an aluminum plate. After curing, the assembly is heated to melt the solder. The copper block is removed and the excess solder is brushed from the foil. The 33 sample is placed in the peeling jig, Fig. 16c. A dead weight is attached to the rear of the sample, providing a constant force which holds the foil against the mandrel. The force measured (Wj) represents a combination of the plastic work in the foil (Wp), the friction in the jig (WJ), the background weight (£>), and the peeling force (We) according to the equation : Wj= We+ Wf+ Wp + D (6) Grit Blasting 1 1 1 1 1 Tape 1 Solder Copper T Foil (a) / Solder Copper r Foil (b) Aluminum Plate Epoxy / Mandrel -Foil | Peel Force (c) Figure 16. Steps involved in the preparation of a peel test sample. A foil is soldered to a copper block and grit blasted, (a). The assembly is coated, (b). The copper and solder are removed and the foil is peeled from the coating, (c) [46] In order to obtain the effect of friction and plastic work, a calibration is performed. For this purpose a foil is placed in the jig and pulled in the same way as the test sample. The measured force represents the plastic work required to bend the foil and the friction in the jig. 34 The peel test represents an alternative to the available adhesion measurement techniques like pull and scratch tests. Because the test has a long history of use in the adhesives industry, a large amount of theory has been developed to understand its mechanics [50]. The peel test is also capable of detecting differences in bonding between different portions of the particle stream. 2.4.3 Gas Permeability Tests for Thermal Sprayed (TS) Coatings The A S T M standard (ASTM C-577-78) test method for gas permeability of refractories can be easily extended to porous bodies like coatings. This test however requires that the coating be detached from the substrate. The specimen is placed between two rubber gaskets, enclosed between the two halves of an airtight holder. Dry nitrogen or air is passed through the specimen. The permeability is calculated from the through flow rate of the gas under a fixed pressure difference across the specimen coating, according to the following equation [51]: K=(T}QL/AAP)X100 (7) where K is the permeability, expressed in centidarcys, r| [mPa s] is the viscosity of the gas, Q 3 2 [cm /s] is the flow rate of the gas, L [cm] is the sample thickness, A [cm ] is the area of the sample and AP [atm] is the absolute pressure drop across the sample. One centidarcy is defined as a flow of 0.01 cm /s of a fluid of 1 mPa s viscosity through a 1 cm cube under a pressure difference of 1 atmosphere. The viscosity of air is taken as 182.7 x 10"3 mPa s [51]. If nitrogen is used as the permeating medium the formula has to be decreased by 5%, since the viscosity of nitrogen at room temperature is approximately 5% lower than that of air. 35 2.4.4 Burner Rig Tests One of the major restraints for the widespread use of TBCs is the lack of long term durability. For example in the aircraft gas turbines, the temperature of the exhaust gas is typically around 1325°K at takeoff conditions, while during cruise it is decreases to around 1126°K [22]. Such temperatures along with thermal fluctuations experienced during operation, and large centrifugal forces, make the thermo-mechanical fatigue conditions in the turbines extremely severe. As an assessment of the endurance of these coatings in an actual engine is rather expensive, the long term durability of these coatings can be estimated in terms of the cyclic life in a simple burner rig test. The purpose of the thermal fatigue rig tests is to simulate engine operating conditions and to establish endurance limits of the coatings in as deposited and modified states. The rig is intended to provide accelerated testing which allows a relative ranking of the available coatings based on various materials, coating and sealing techniques. In such a test the coating is heated in a flame to the desired temperature, held at the temperature for a specified time and subsequently quenched (in air or water) [52]. This cycle is continued till spallation is observed. In general a decrease is seen in the cyclic lifetime of the coatings as the maximum hold temperature increases. This is believed to result from the three main effects: i) thermal shock stress, ii) coefficient of thermal expansion (CTE) mismatch and iii) stress generated by the oxidation of the bond coat. Miller [53] proposed that rapid cooling produces plane tensile stress state in the coating surface and thus segmented cracks. This is in contrast to the compressive stress state produced by slow cooling. On rapid cooling the top coat tends to 36 contract more than the bond coat because of the large thermal gradient and this results in a plane tensile stress state in the coating. This tensile stress causes the top coat to form radial cracks which relieves the stress induced by the C T E mismatch [20]. 2.5 Sol-Gel Chemistry Sol-Gel (SG) route to ceramics offers several advantages over conventional methods of powder preparation [5, 54]. Sol-Gel (SG) route to ceramics, e.g. alkoxide route, involves the hydrolytic decomposition of the metal alkoxide and subsequent thermal dehydration (decomposition) of the resulting gel to form the respective ceramic oxide. The advantage of SG route is that the starting precursor is a low viscosity fluid yielding a fine powder on decomposition. Furthermore, the transformations of sol —>• gel —> amorphous powder —> crystalline powder are achieved under conditions of relatively low temperature (< 800°C). The low temperature SG synthesis of oxides ensures that the structural and mechanical integrity of the coated metallic component is not deteriorated. The necessary condition for the formation of a gel is a sufficient concentration of the sol. As the sol concentrates through hydration or electrolyte addition, the repulsion between the sol particles decreases. Clustering and cross linking of the individual particles takes place and finally the sol ceases to behave as a population of individual particles suspended in solution. As the reaction progresses, a stage is reached where clustering and cross linking of these clusters spans the whole volume of the sample [55]. This corresponds to a state where the sol has set into a gel. A solution can be considered as a truly single-phase liquid, while the 37 sol is a stable suspension of colloidal particles. At the gel transition point the sol becomes a rigid porous mass. The sol-gel transition proceeds without discontinuity in properties. Nevertheless, the viscosity of the system changes strikingly in such a short period of time that this change is often defined as the point of gelling [55]. The simplest and most economical method for the large scale production of alkoxides involves the addition of anhydrous metal halides to a mixture of anhydrous alcohol in the presence of anhydrous ammonia [54]. MC1 4 + 4NH 3 + 4ROH > M(OR) 4 + 4NH4C1 where M represents the transition metals and R is the organic group, such as C H 3 , C 2 H 5 etc. The removal of NH 4C1 by filtration is troublesome and time consuming. The filtration step can be avoided by carrying out the process in the presence of amides and nitriles, whereby the metal alkoxide separates out as the upper layer, while the ammonium chloride remains in solution in the amide or nitrile in the lower layer [54]. The alkoxide of some metals can also be prepared by the reaction of the metal with alcohol in the presence of a catalyst such as HgCl 2 or Hgl 2: M + ROH > M(OR) n + ^ H 2 where n is the valence of the metal M and R is the organic group Most alkali and alkaline earth metal alkoxides can be prepared by this reaction [56]. 38 2.5.1 Alkoxide Chemistry The decomposition of metal alkoxides is initiated via a hydrolytic reaction with water followed by thermal dehydration of the resulting precipitates. The general decomposition can be described in two steps [57]. M(OR) n + nH 2 0 > M(OH) n + nR(OH) hydrolysis M(OH) n > M O n / 2 + ^ H 2 0 dehydration The hydrolysis usually occurs at room temperature and the dehydration occurs below 600°C resulting in the formation of powders with particle size in the range of 2-5 nm. These super-fine ceramic powders are sinterable at temperatures in the range of 600-1000°C, significantly below that for classical micron-size ceramics. Mullite by Sol-Gel Aluminum tri-isopropoxide and silicon tetra-isopropoxide are usually the starting precursors. The mixed alkoxides are dissolved in excess isopropyl alcohol before hydrolysis to ensure proper mixing. The hydroxy-aluminosilicate is then prepared by slowly adding the alkoxide solution to ammoniated distilled deionized water according to the reaction [57]. 6A1(0C 3H 7) 3 + 2Si(OC 3H 7) 4 + xH 2 0 • 2Al 3Si(OH) 1 3. xH 2 0 + 26C 3 H 7 OH 39 The resulting hydroxy-aluminosilicate is washed repeatedly and calcined at 600°C to produce mullite. Yttria Stabilized Zirconia (YSZ) by Sol-Gel Typically a solution of high purity zirconium isopropoxide (Zr(OC 3H 7) 4) and yttrium isopropoxide (Y(OC 3H 7) 3) in a mutual solvent such as n-hexane or benzene is used. The mixed oxide is precipitated from the solution according to the following reaction [58]: 2Y(OC 3 H 7 ) 3 + 3H 2 0 > Y 2 0 3 + 6 C 3 H 7 O H Zr(OC 3 H 7 ) 4 + 2H 2 0 • Z r 0 2 + 4 C 3 H 7 O H A small amount of mineral or organic acid can be added to catalyze the reaction. The above sol-gel chemistry is suitable for fabricating bulk ceramics. However infiltration of the coatings with precursors in solvents not soluble in water (e.g. hexane, benzene) will not work for the post treatment as proposed in this work. Spinel by Sol-Gel The precursors usually selected for the synthesis of spinel (MgAl 2 0 4 ) are aluminum isopropoxide Al(OC 3 H 7 ) 3 and magnesium ethoxide Mg(OC 2 H 5 ) 2 . The two metal alkoxides are mixed together and gently heated in isopropyl alcohol until the Mg(OC 2 H 5 ) 2 is completely 40 dissolved producing a clear solution. During the refluxing process magnesium aluminum double alkoxide (MgAl 2(OR)) 8 is formed [57, 59], which is readily hydrolyzed by water. According to X-ray diffraction analysis, powders obtained after calcination at 550°C are amorphous. X-ray diffraction of the powder calcined at 1000°C shows broad peaks of spinel which are characteristic of diffraction patterns for fine particles (< lum). Alumina by Sol-Gel The starting precursor to prepare high purity alumina is usually aluminum isopropoxide dissolved in isopropyl alcohol. Aluminum isopropoxide is readily hydrolyzed by water to produce mono and tri-hydroxides. Thermal decomposition of these hydroxides yields alumina of high purity. The influence of electrolytes upon the sol-gel transition is characterized by their peptizing effect. The amount of electrolyte present in the sol, such as H N 0 3 and HC1, determines the gelling point of the system as well as the properties of the oxide resulting from the gel on heat treatment. It has been reported that the gelling volume in alumina goes through a minimum when the acid (HC1) concentration is about 0.07 mole per mole of hydroxide [59]. The initial decrease in the gelling volume is caused by the effect of acids on the electrical charge of the particles. When acids are introduced, positively charged micelles strongly absorb the anions of the acids and shift towards a more neutral state. The interparticle repulsion is decreased as a result and this leads to a reduction in the gelling volume. 41 3. SCOPE AND O B J E C T I V E S Plasma sprayed ceramic coatings are widely used in various industrial fields because of their excellent wear resistance, thermal stability and mechanical integrity when exposed to elevated temperatures [1, 4, 18]. However they frequently fail to meet the requirements of a corrosion resistant coating. This is due to the nature of the plasma spraying process which gives coatings exhibiting residual porosity. The high surface tension of the molten particles and the rapid rate of solidification on impact prevents the particles from completely filling the voids. Additionally air inclusions associated with the turbulence and the sublimation of the material components can appear. Porosity is also influenced by the non molten residual powder. Thermal sprayed (TS) coatings display a submicroscopic system of cracks in addition to the microscopic porosity [60]. Consequently, many TS coatings and in particular Thermal Barrier Coatings (TBC) fail to provide sufficient corrosion protection to the coated substrates. A few techniques have been developed to reduce coating permeability by artificially reducing the pore diameters within the coatings [60]. However, the thermal properties of thermal barrier coatings are dependent on the existence of a highly microcracked and porous structure. Thus, a compromise must be reached in terms of the levels of porosity and permeability desired from TBCs. Although porosity is an invariant characteristic of TBCs, it is desired to seal the surface of TBCs preferentially in an attempt to make it impervious to corrosive attack, yet retain the highly porous structure beneath it. The porous structure of TS coatings is essential to retain its low thermal conductivity and high tolerance to strain. The performance of plasma sprayed coatings has been attempted to be improved by sealing the 42 coating with polymers, but these commercial sealants suffer from many disadvantages which have been discussed in Section 2.2.3. The method proposed in this work involves sealing the pores by infiltration with liquid precursors of ceramics. Sealing by infiltration with sol-gel precursors is a unique post-deposition treatment method to enhance the performance of thermal sprayed coatings. In addition to reducing the coating permeability, it is expected that such a sealing treatment would enhance the aerodynamic performance of the coating by reducing surface roughness [22] and also introduce some degree of compressive stress on the surface. Similar sol-gel infiltration techniques have been used to fabricate multi-phase ceramic bodies with improved properties [61-63]. The coatings studied included yttria stabilized zirconia and alumina TBCs. The sealants included alumina, silica and magnesia-alumina spinel. The starting precursor is a relatively low viscosity fluid which aids in the infiltration process. Thermal decomposition of the precursors yields fine powders (2-5 nm) at relatively low temperatures (< 800°C). The low temperature synthesis of the oxide ensures that the structural integrity of the coating is not deteriorated because of such a post-treatment. It is believed that these precursors would effectively penetrate into the pores and on subsequent heat treatment decompose in situ. This would result in reducing the pore diameter and considerably reducing the coating permeability. 43 Objectives The objectives of this work were : 1) to modify the surface and volume of thermal sprayed coatings (TBCs in particular) in an attempt to decrease the permeability to gases and therefore to increase the high temperature corrosion resistance of the coated substrates. 2) to design a sandwich structure consisting of a dense top layer (10-300 u.m thick) with a porous layer beneath it, so as to have an optimum balance of corrosion resistance and insulating properties. 3) to develop a simple, cost-effective post-deposition treatment to enhance the mechanical properties such as adhesion strength and hardness without deteriorating the thermal properties of the coating. 4) to characterize the effects of such an infiltration on the permeability and mechanical properties of the coating. 44 4. E X P E R I M E N T A L P R O C E D U R E 4.1 Processing and Characterization of Thermal Sprayed Coatings The Thermal Sprayed coatings studied included' yttria stabilized zirconia (YSZ) and alumina. The coatings were supplied from three vendors: i) Northwest Mettech, B.C., Canada, ii) Waploc Ceramics, Osaka, Japan and iii) Pratt & Whitney, Quebec, Canada. The processing parameters for the spray formed ceramics supplied by Waploc Ceramics and YSZ coatings from Pratt & Whitney were not provided. The processing parameters of the coatings sprayed at Northwest Mettech and used for the electrochemical tests are listed below: Test I Test II Powder Z r 0 2 - 8 Y 2 0 3 Z r 0 2 - 20Y 2 O 3 Powder Feed Rate (g/min) 60 40 Total gas flow (1/min) 240 240 Carrier gas flow (1/min) 6 6 Standoff distance (mm) 75 15 Power (kW) 78 75 The electrochemical tests were carried out on mild steel coupons with YSZ coatings sprayed at Northwest Mettech using the Axial III plasma torch. One set of electrochemical tests was carried out on alumina coatings (thickness 0.3 mm) sprayed with Rokide Flame Torch (Norton, Worcester, MA) and infiltrated with silica. The spraying conditions for these coatings were not available. The gas permeability tests were carried out on spray formed alumina (thickness 4 mm) supplied by Waploc Ceramics. This test requires that the coatings be detached from their substrates and thus thick spray formed ceramics was chosen. The hardness tests were carried out on spray formed YSZ coating (thickness 4 mm) supplied by Waploc Ceramics. 45 The specimens for the peel adhesion tests were also sprayed at Northwest Mettech. These coatings consisted of 0.2 mm thick YSZ deposited on 0.17 mm thick steel (low carbon 1010 steel) and nickel (cold rolled) foils with the following spraying conditions: Powder Z r 0 2 - 8 Y 2 0 3 (Metco 204-NS powder) Powder feed rate (g/min) 38 Total gas flow (1/min) 120 Carrier gas flow (1/min) 8 Power (kW) 79 Burner rig tests were carried out on coatings supplied by Pratt & Whitney. These coatings consisted of a YSZ top layer (cubic + tetragonal Z r 0 2 > 95%, monoclinic Z r 0 2 < 5%, thickness 0.38 mm) and MoCrAlY alloy bond coat (thickness 0.22 mm) deposited on a Ni based alloy. The spraying conditions were not provided, but these were coatings actually used in the combustors of the turbine engines manufactured by Pratt & Whitney. X-ray diffraction (Siemens Diffraktometer D-5000) analysis was carried out on the as-received coatings to analyze their phase compositions. The porosity of these coatings was measured by Archimedes water displacement method. 4.2 Post-Deposition Treatment of Thermal Sprayed Coatings The precursors for infiltration were prepared through the alkoxide route. The inorganic route to sol-gel preparation was abandoned because it involved the use of chloride and sulfate aqueous solutions which were corrosive to the substrates. For example, the pH of magnesia-alumina spinel (MgAl 2 0 4 ) inorganic precursor, i.e. aqueous solution of MgS0 4 and Al 2 (S0 4 ) 3 j was in the range of 3 to 4. Subsequent gelation of this precursor with N H 4 O H involved 46 subjecting the substrate to alternate acidic and basic conditions. Infiltration with these precursors was thus not feasible. With the alkoxide route longer infiltration times in the sols (i.e. before the gelation stage) was possible as the precursors were alcohol based. Since these precursors gel spontaneously in contact with water, special attention was given to avoid any contact with moisture. This was achieved by maintaining a temperature of ~ 40-50°C during the preparation of the precursors and in the infiltration stage. The post-deposition treatment involved infiltration of the coatings with liquid precursors of oxides. The precursor for alumina was aluminum isopropoxide (Al(OC 3 H 7 ) 3 , (Aldrich Chemical Company, 98% purity), dissolved in isopropyl alcohol (C 3 H 7 OH) at a concentration of two moles (2M) per liter. The solution contained 25 wt.% A1 2 0 3 . The optimum concentration of 2M was chosen based on viscosity measurements discussed in the next section. For silica deposition the precursor was prehydrolysed ethyl silicate (Si(OC 2H 5) 4, R E M E T R-25, Remet Chemical Corporation) containing 25 wt.% of Si0 2 . For sealing with magnesia-alumina spinel (MgAl 20 4), the precursor chosen was a mixture of aluminum isopropoxide and magnesium ethoxide (Mg(OC 2H 5) 2 , Alfa AESAR) dissolved in isopropyl alcohol at a concentration of IM, containing 27 wt.% MgAl 2 0 4 . In this case the two oxides were refluxed in isopropyl alcohol until the solution turned clear, i.e. till all the magnesium ethoxide completely dissolved into solution. Subsequent hydration and heat treatment of the precursor produced magnesia-alumina spinel. The sprayed specimens were infiltrated by immersing them in their respective precursor solutions for various lengths of time (4 minutes to 1 hour) in the setup schematically 47 shown in Fig. 17. The process of infiltration was found to improve under the action of an ultrasound, in terms of increased infiltration depths of the precursors. It is believed that the ultrasound increases the rate of penetration of the liquid precursors into the capillaries by adding an additional pressure term to the capillary pressure 2ycos9/r, Eq. 2. The infiltrated specimens were then immersed for 20 seconds in deionized water containing 10"2 M HC1, which acts as a catalyst in the gellation of hydrates from the infiltrated sol, and then dried at 65°C for 12 hours. The use of acid catalysis increases the rate of hydrolysis and ensures a uniform particle and pore size distribution in the resulting gel [64]. Some of the infiltrated and dried coatings were flame heat treated at 600°C or 700°C for 15 or 30 seconds. The heat treatment for the spray formed ceramics was carried out in the furnace at 600°C for 15 minutes. The heat treatment temperature was chosen based on thermo-gravimetric analysis (TGA) and X-ray diffraction (XRD) results which show complete decomposition of alumina, silica and magnesia-alumina spinel at temperatures of 650, 800 and 650°C respectively, Fig. I-1 -1-7, Appendix I. SEM/EDS (Hitachi S-2300) studies were carried out on the infiltrated samples to assess the depth and nature of the infiltrant in the parent coating material. T G A (Du Pont 950, Delaware, USA) and X-ray diffraction analysis were done to characterize the phase transformations in the sealant at elevated temperatures. 48 Figure 17. The infiltration process involved in the post treatment SEM/EDS analyses were also done on the samples assessed in the potentiodynamic tests to evaluate the corrosion products. X-ray mapping of the sealant material in the parent coating was done to evaluate the effectiveness of such an infiltration. Viscosity (Brookfield Model DVIII Viscometer) of the precursors was measured at room temperature. The weight gained with infiltration was measured to evaluate the process yield on heat treatment. 4.3 Electrochemical Tests for Permeability The effectiveness of the sealing process was verified using gas permeability tests and the permeability to electrolyte was determined by potentiodynamic polarization tests. The potentiodynamic apparatus consists of a standard corrosion cell with high purity platinum as an auxiliary electrode and a saturated calomel electrode (SCE) as the reference electrode, Fig. 18a. The equipment consists of an electrochemical interface (EG & G Princeton Applied Research, Model 350A, Corrosion Measurement Console) and a data analysis module. The 49 materials which were tested include mild steel coupons sprayed with YSZ using an Axial III plasma torch. The processing parameters for these coatings are given in Section 4.1. For electrical connections copper wires were riveted to the steel substrates without damaging the coatings. The sprayed coupons were then mounted in an epoxy (Lecoset 7007) such that only the sprayed surface was exposed as the working electrode, Fig. 18b. Electochemical Testing Solution Figure 18. Potentiodynamic polarization apparatus consisting of a potentiostat interfaced with a data acquisition module, (a). Schematic of a sample mounting, (b). The electrochemical tests were carried out in deaerated 3 wt.% aqueous NaCl solution. The specimen was immersed in the solution for an hour before the scan was initiated. The potential of the coated substrate was scanned from -1.0 V to 1.0 V at a rate of 1.0 mV/s with the current being continuously recorded. The reason for using a chloride solution as the 50 corrosive medium was that chlorides, being powerful ligands form complexes with the metal, stabilizing the ions in solution. This serves to aggravate the substrate corrosion process. Many designs require thermal sprayed coatings to face saline solutions as in marine applications and thus the choice of aqueous NaCI solution was typical for the electrochemical tests. The criterion used for comparing the permeability of the coatings in aqueous solutions at room temperature was the limiting current value in the passive region [23, 24], refer also to Fig. 14, p. 28. A comparison of the different polarization curves obtained was used to assess the reduction in coating permeability with varying infiltration treatments. 4.4 Gas Permeability Tests Gas permeability was evaluated on free standing spray formed yttria stabilized zirconia (YSZ) ceramics (Waploc Ceramics, Osaka, Japan). This test was in conformance to the A S T M standard C 577-78 for refractories, with slight modifications in the test jig, Fig. 19. The specimens were infiltrated with alumina precursors for various lengths of time (4, 10,15 and 20 minutes) and subsequently hydrated by immersing in water. The dried samples were heat treated at 600°C in a furnace for 15 minutes. The sample was then placed between two rubber gaskets and enclosed in an air tight holder as shown in Fig. 19. Dry nitrogen was passed through the specimen at a pressure of 0.21 MPa (30 psi). The gas permeability was calculated from the through flow rate of the gases according to Eq. 7. Since nitrogen was used as the permeating medium, the permeability calculated from the above equation had to be decreased by 5% to account for the lower viscosity of nitrogen, as discussed in Section 2.4.4. 51 Pressure Q Control 1 ][ Gas Flow I — Gas Through Flow Figure 19. Cross-section of the permeability jig 4.5 Hardness Tests Hardness tests were performed using Depth Sensing Indentation (DSI) at loads of 300 mN (30g) on the cross section of both the treated and untreated specimens. The schematic of the apparatus used is shown in Fig. 20. It consists of a conventional Vickers indentor (Fischerscope Nano-indentor H-100, FischerTech. Inc., CT, USA) that is coupled to a displacement measuring device. The load is generated by an electro-magnetically controlled system. The specimen is mounted on a table, displacement of which is recorded digitally by a micrometer. A microprocessor is used to control the complete testing cycle. This technique uses the depth of penetration of the indentor to calculate the hardness of the coating according to the Eq. 5. The results of DSI are comparable with those obtained from Vickers microhardness, but allow large spatial resolution and do not require 52 measurement of indentation size. These tests were carried out on zirconia coatings, polished to a surface finish of one micron. Such a fine polish is essential because the DSI measurement is sensitive to surface topography. Current source Digitize Figure 20. Schematic of the depth sensing indentation (DSI) apparatus 4.6 Peel Adhesion Tests Peel Adhesion Tests (PAT) were performed to determine the effect of infiltration on adhesion. The foil (low carbon 1010 steel or nickel, thickness 0.17 mm) was soldered to a copper block which acts as a heat sink according to the procedure described in Section 2.4.2. Some of the foils were glued to the copper block using a silicone adhesive (Dow Corning 732 Scellant RTV) mixed with three parts by weight of Cu powder. The test samples were grit blasted with 60 grit (177 pm) particles of A1 2 0 3 at a pressure of 0.56 MPa, with stand-off 53 distance of ~ 80 mm at 45 degrees to the surface. The samples were then cleaned, degreased and the edges masked. The PAT test samples were then sprayed with Z r 0 2 - 8 Y 2 0 3 powder (Praxair, Indianapolis, USA) using the Axial III plasma torch. The processing parameters are given in Section 4.1. Substrate cooling during spraying of the PAT samples was provided using an air knife. The air knife consists of a jet of compressed air directed onto the substrate of the foil. This was necessary to avoid overheating and debonding of the foil/copper block interface. The sprayed sample was glued to an aluminum plate using a thermoset epoxy (Struers Epoxyfix hardener and resin), Fig. 16c. After curing, the copper block was removed from the foil by melting the solder. The sample was then mounted in a jig and the foil was peeled from the coating at a constant rate of 2.5 mm/min in an Instron having a 5 kN load cell. The load and the cross head displacement were monitored and recorded continuously. A calibration was performed by pulling the foil in the jig around the mandrel. The calibration is necessary to measure the effects of friction in the mandrel and the plastic work in the foil. From the load-displacement curve the peel strength as a function of crack position was calculated using the calibration. 4.7 Burner Rig Tests To assess the durability and thermal fatigue resistance of thermal sprayed coatings, a simple burner rig apparatus was built. The rig consists of a high temperature natural gas/oxygen flame, a turntable to hold the coated specimens, and air cooling, Fig. 21. Severe 54 heating and cooling schedules were selected to accelerate the test. This could be achieved by having large swings in temperature in relatively short periods of time [25]. The endurance test consisted of heating the coated specimens in a flame to 1170°C ± 20°C for 54 seconds or 1270°C ± 20°C for 56 seconds and then quenching to room temperature with a jet of cold air, in about 20 seconds. Six specimens could be tested simultaneously in the setup, as shown in Fig. 21. Coating failure was considered to have occurred when 75% or more of the test zone had spalled. This criterion for failure was set based on previous reports [25]. The specimens tested in the burner rig were YSZ coatings (Zr0 2-18Y 20 3) supplied by Pratt & Whitney, Quebec, Canada. Substrate Figure 21. Schematic of the burner rig apparatus 55 5. RESULTS AND DISCUSSION 5.1 Precursor Properties The plot of viscosity (cP) vs. molar concentration of the alumina precursor dissolved in isopropyl alcohol is shown in Fig. 22. The viscosity was found to increase sharply beyond 2M concentration. 250 -t in > 100 -~ 150 -q. . 200 H 50 H o * 0 2 3 4 Concentration (M) Figure 22. Viscosity vs. concentration of aluminum isopropoxide dissolved in isopropyl alcohol 56 As discussed earlier, good sealing requires an optimum balance between low viscosity (for greater infiltration depth) and high concentration (for better yield of the infiltrant). Higher concentration of the precursor would increase the amount of oxide deposited in the pores. Accordingly, an optimum concentration of 2M, having viscosity of 26 cP, was chosen for all subsequent tests with precursor of alumina. The viscosity of prehydrolyzed ethyl silicate was 3.5 cP, seven times lower than that of alumina at 2M concentration. Accordingly, the infiltration depth of silica is expected to be significantly higher than that of alumina precursor for similar time and wetting conditions (refer to Eq. 2, p. 23). The viscosity of the magnesia-alumina spinel precursor, i.e. aluminum isopropoxide and magnesium ethoxide dissolved in isopropyl alcohol, was high even for low concentrations, Fig. 23. The viscosity of spinel precursor at IM concentration was found to be 48 cP. Thus the infiltration depth expected with a given immersion time in the spinel precursor would be lower than that of the alumina and silica precursors. All the precursors show Newtonian flow characteristics, i.e. viscosity independent of shear rate. The decomposition characteristics of the alumina, spinel and silica gels were analyzed and their respective phase transformations were studied. T G A showed complete decomposition of alumina and spinel gels at 650°C, while decomposition of the silica gel was complete at 800°C. In the case of alumina, the decomposition yield from the dry gel was 78% by weight, whereas it was 60.2% for spinel and 71% for silica, refer to Fig. 1-1,1-2, 1-3, 1-4, Appendix I. 57 100 80 A 0.00 0.25 0.50 0.75 1.00 1.25 1.50 Concentration (M) Figure 23. Viscosity vs. concentration of magnesia-alumina precursor Hydrolysis of aluminum isopropoxide dissolved in isopropyl alcohol produced mono-hydroxide (AlOOH). Similarly hydrolysis of the other precursors produced their respective hydroxides. These hydroxides decomposed upon heat treatment to form oxides according to the following reactions: Al(OC 3 H 7 ) 3 + 2H 2 0 > AlOOH + 3 C 3 H 7 O H 2A100H '—> A1 2 0 3 + H 2 0 Si(OC 2 H 5 ) 4 + 4H 2 0 > Si(OH)4+ 4 C 2 H 5 0 H Si(OH)4 > Si0 2 + 2H 2 0 Mg(OC 2 H 5 ) 2 + 2H 2 0 > Mg(OH) 2 + 2 C 2 H 5 O H 2A100H + Mg(OH) 2 + 2H 2 0 > MgAl 2 (OH) 8 MgAl 2 (OH) 8 > M g A l 2 0 4 + 4H 2 0 58 X-ray analysis of the gels showed the following phase transformation sequence, Fig. I-5,1-6,1-7, Appendix I (the broad peaks observed are typical for fine particles < 0.1 urn): AlOOH 650"c >yAl203 850°c ,oAl2Q3 ">50°c >QA1203 "50"c ,aAl2Q3 Si(OH)4 800°c > amorphous Si02 1 2 0 0 ° c > crystobalite MgAl2 (OH)8 MgAl204 The phase transformation from YA1 2 0 3 (d = 3.85 g/cc) to a A l 2 0 3 (d= 3.98 g/cc) results in a small shrinkage of ~ 3%. However, the phase transformation from amorphous silica (d = 1.4 g/cc) to crystobalite (d = 2.32 g/cc) is accompanied by a substantial shrinkage of ~ 40%. The transformations in silica with their large volume changes suggest that significant stresses could be built up in the coating during thermal cycling. Weight gain measurements were carried out to evaluate the amount of infiltrant deposited in the coating before and after heat treatment, Fig. 24 and 25. These tests were carried out on spray formed zirconia yttria coatings, infiltrated with 2M alumina precursor and heat treated at 600°C in a furnace for 15 minutes. The open porosity of these coatings, measured by Archimedes water displacement method, was found to be 12%. The weight gained, i.e. the weight of dry infiltrant (gel) in the coating, with 10 minutes of infiltration (Fig. 24a) was 5.5 mg. On heat treatment at 600°C in a furnace for 15 minutes, the weight of the infiltrant (y alumina) in the coating was found to be 4.2 mg. This corresponds to a yield of 76%o and is in agreement with T G A results (Fig. 1-2, Appendix I) which show a decomposition yield of 78% from the dry alumina gel. 59 The weight gained and hence the infiltrated depth (weight = density x area of the pore 1/2 x depth) followed a t relationship (Fig. 24b) in the initial part of the infiltration process, according to Eq. 2. This proves the earlier discussion on the validity of using Washburn's model for porous ceramic coatings. The deviation from the above model at infiltration times greater than 15 minutes can be attributed to the complete filling of most pores. 0 5 10 15 20 25 1 4 8 10 15 20 Time of Infiltration (min) Time of Infiltration (min) Figure 24. Weight gained with infiltration of 2M alumina precursor, (a) and log/log plot of the 1/2 same showing a linear t relationship upto 15 min of infiltration, (b). Figure 25 shows the weight gained in the coating after multiple infiltrations. The yield on decomposition from the dry alumina gel is in conformance with T G A results, Fig. 1-2, Appendix I. The amount of alumina deposited in the coating with ten minutes of infiltration repeated twice, was more than what was achieved with 20 minutes of infiltration. Thus multiple infiltrations are beneficial to density the surface preferentially with shorter infiltration times. 60 16 14 12 10 D) E. •D 1 8 n 10.0 min. infiltration • Wt. gained (dry) • Wt. gained (heat treated) No. of Infiltrations (min) Figure 25. Weight gained vs. no. of infiltrations (10 minutes each) The volume yield of 2M alumina precursor was 4.5%. Assuming that the liquid precursor fills up all the pores in the coating, the maximum decrease in porosity would also be 4.5%, e.g. porosity would decrease from 12% to 11.46%. Similarly the maximum porosity decrease that can be expected with complete filling of the pores with precursors of silica and spinel (IM) would be 4% and 7% respectively. The conclusions is that the single infiltration process does not substantially change the overall porosity of the coatings. Further tests indicate however that the porosity and the microcrack morphology is modified, e.g. permeability to gas and liquid is decreased and resistance to thermo-mechanical fatigue is increased. 61 5.2 Potentiodynamic Measurements The polarization curves shown in Fig. 26 are that of mild steel samples with YSZ zirconia coatings, infiltrated with 2M alumina precursor for 0, 6, 10 and 20 minutes, and subsequently heat treated at 600°C in a flame for 30 seconds. The spraying parameters for these coatings are given in Section 4.1, Test I. Such a short heat treatment schedule was chosen to explore the possibility of decomposing the gel in situ during service of these coatings. With increasing infiltration time, which corresponds to greater penetration depth, the average limiting current decreased from 9.8 uA in the untreated sample to 4.3 uA in the sample infiltrated for 20 minutes, Fig. 27. Thus, there was over 56% reduction in the current with 20 minutes of infiltration. The limiting current was calculated as the average of the current values in region B of the polarization curve (refer to Fig. 14, p. 28). 1.5 Infiltration time: 1.00E+02 1.00E+04 1.00E+06 1.00E+08 Current (nA) Figure 26. Polarization curves of zirconia coatings infiltrated with 2M alumina precursor for 0, 6, 10 and 20 minutes, heat treated in a 600°C flame for 30 seconds. Potentials referred to SCE. 63 1.00E+05 < r 7.00E+04 £ 0 c 4.00E+04 1 1.00E+04 6 10 Time of Infiltration (min.) 20 Figure 27. Limiting current as a function of the infiltration time in 3 wt.% NaCl solution Similar potentiodynamic tests were carried out on zirconia-yttria coatings (Test I, p. 44) infiltrated with 2M alumina precursor, heat treated on a hot plate (20-600°C in 25 minutes and 5 minutes hold at 600°C). The heat treatment schedule was varied to explore the effect of heating rate and time on the polarization plots, as compared to the shorter and more severe heat treatment followed earlier (30 sec in a 600°C flame). The polarization plots obtained from these experiments are shown in Fig. 28. As in the previous case there was an overall shift in the curves (in particular the limiting current) towards lower current values. The limiting current decreased by 60% from 20 mA to 8 mA with 30 minutes of infiltration, as shown in Fig. 29. Thus there was no significant change observed in the permeability of the coatings for fast (flame) and slow (hot plate) heat treatment. Figure 28. Polarization curves of zirconia coatings infiltrated with alumina precursor for 0, 10, 20 and 30 minutes, heat treated on a hot plate (0-600°C in 25 min and held at 600°C for 5 min). 65 2.10E+O5 < 1.60E+05 c c s 3 o c 1.10E+05 6.00E+O4 1.00E+04 10 20 Time of Infiltration (min) 30 Figure 29. Limiting current as a function of infiltration time for various treated coatings in 3 wt.% NaCl solution Figure 30 shows the polarization curves of zirconia coatings infiltrated with I M spinel precursor for various infiltration times. The heat treatment was carried out on a hot plate heated to 600°C in 25 minutes and held at that temperature for 5 minutes. The spraying parameters for the coatings tested are given in Section 4.1, Test II (p. 44). Figure 31 shows the polarization curves of alumina coatings (sprayed with Rokide Flame Torch) which were infiltrated with prehydrolyzed ethyl silicate. The coatings were deposited on an intermediate Ni-Cr alloy as the bond coat. The heat treatment was carried out at 700°C for 30 seconds in a flame. The trend in the curves, i.e. shift towards lower current values, remained the same, but in this case the effect of lowered current was seen only after 10 minutes of infiltration. The lower level of permeability decrease with silica and spinel infiltration can be attributed to the lower yield from their gels (71% and 60% from silica and spinel gels), in comparison to that of alumina (78%). The higher viscosity (48 cP) and the lower concentration of the spinel precursor (IM) comparison to that of alumina precursor (viscosity = 26 cP at 2M 66 Figure 30. Polarization curves of zirconia coatings infiltrated with IM spinel precursor for 0, 4, 6, and 15 minutes, heat treated on a hot plate (0-600°C in 25 minutes and held at 600°C for 5 minutes). Figure 31. Polarization curves of alumina coatings infiltrated with ethyl silicate for 0, 6 and 10 minutes, heat treated in a flame at 600°C for 15 seconds. Potentials referred to SCE 68 concentration) were contributing factors as well. Thus a lesser number of pores would be sealed in comparison to alumina. Examination of the corroded samples after the test showed predominantly crevice corrosion at the coated substrate. SEM/EDS analysis of the corroded surface at the end of the test showed high signals of Fe and CI" (probably FeCl 2) , Fig. 1-8, Appendix I. The polarization curves can be explained with the help of a Fe-Cl"-H20 E-pH diagram, Fig. 1-9, Appendix I [65]. Corrosion in aqueous solutions, being electrochemical in nature, is controlled by the migration of species (anions and cations) and their stability in solution. For a mild steel substrate immersed in an aqueous salt solution, the anodic reaction involves dissolution of the metal ions into solution. The cathodic reaction involves the reduction of 0 2 at the metal (substrate) surface : 2+ Fe —» Fe + 2e" Anodic reaction 0 2 + 2H 2 0 + 4e" -> 40H" Cathodic reaction As the corrosion progresses, 0 2 is depleted inside the pores due to the limited mass transfer within the pores as compared to the bulk solution (pH = 7). This decreases the rate of the cathodic reaction and hence the (OH)" ion concentration. The hydrolysis of Fe results in proton formation according to the reaction : Fe 2 + + H 2 0 -> Fe(OH) ++ H + Thus there is a possibility of a transient pH decrease inside the crevice. The pH within the pore could be as low as 2-3, depending on the pore geometry. The anodic reaction however continues, building up an excess positive charge inside the crevice. To compensate for the charge imbalance, CI" anions migrate into the pores. From the E-pH diagram of a Fe-Cl"-H20 69 system (Fig 1-9, Appendix I), it can be seen that at low pH conditions the precipitation of ferrous chloride at potentials in the range of -0.2 V S H E - 0 V S H E is favorable. The reaction of the metal ions in solution with the chloride ions could thus result in the precipitation of ferrous chloride, depending on the local solution chemistry, according to the reaction: Fe 2 + + 2CT-> FeCl 2 Precipitation of this corrosion product in turn slows down the corrosion process and the current approaches a limiting value. At higher potentials which correspond to higher concentration of Cl" ions in the pores, pitting starts due to local instability in the film leading to accelerated corrosion. At even higher potentials (>1 V S H E ) there is also the possibility of oxygen evolution which further increases the current flow (Fig. 1-9, Appendix I). With infiltration, the size of the pores is reduced which also restricts mass transfer processes. It can be expected therefore that 0 2 depletion in a smaller pore would be faster than in a larger pore. Chloride ion migration results in the precipitation of ferrous chloride (FeCl2). In general a shift in the limiting current is seen with infiltration, Fig 26, 28, 30, 31. Due to local changes in solution chemistry, film formation occurs at varying current levels. In general a smaller pore would accelerate film formation due to a more aggressive (low pH) solution chemistry. During the cathodic polarization H 2 evolution results in an increasing pH inside the crevice according to the reaction : 2H + + 2e" -> H 2 The coating system is thus quite complicated because of the changing solution chemistry inside the crevice, i.e. increasing pH during cathodic polarization and decreasing pH during the anodic polarization. Based on the above discussion it can be said that the limiting 70 corrosion current can be used as a qualitative tool to ascertain the effectiveness of the sealing treatment. Even though the current decreases with increasing infiltration times, it does not necessarily mean that the current densities are decreasing likewise. Instead it is possible that the sealing treatment could affect the aspect ratio (height/width) of the crevice and thus accelerate pitting corrosion. With sealing, the solution chemistry (pH) within the crevice is made more severe due to limited mass transfer within a constricted pore. The dissolution of the metal ions and the attack of chlorides make conditions favorable for the precipitation of ferrous chloride. This in turn reduces the current flow. This technique cannot be used to quantify corrosion rate because of the uncertainty in the relationship between current and surface area of the metal exposed. 5.3 Gas Permeability Measurements The results of the gas permeability tests carried out on the spray formed alumina coatings (Waploc Ceramics), infiltrated with alumina precursor (2M) at room temperature, are shown in Fig. 32. These coatings were heat treated in a furnace at 600°C for 15 minutes. The open porosity of these coatings measured by water displacement method was 5.2%. As expected there was a steady decrease in permeability observed with increasing infiltration times, due to the increasing number of pores being sealed. The permeability decreased by 61% with 4 minutes of infiltration (KQ = 32.21 x 10"4 centi-darcys, K4 = 12.57 x 10"4 centi-darcys where the subscript indicates infiltration time), 71 over 85% with 10 minutes of infiltration (Kl0 = 4.62 x 10"4 centi-darcys) and over 90% with 20 minutes of infiltration (K20 = 3.09 x 10"4 centi-darcys). The significant reduction in gas permeability confirms that the sealant is effective in reducing the pore and capillary diameters in the coating. 70 60 50 H 5 40 o o 30 20 10 H Infiltration time: • 0 min. ^ 4 min. 10 min — A 20 min t 1 r 2 4 6 8 10 Time of Flow (min) 12 14 16 Figure 32. Gas permeability of the alumina coatings, infiltrated with alumina precursors for 0, 4, 10 and 20 minutes respectively. Test carried out at room temperature To study the high temperature stability of the sealant, these samples were heat treated in a furnace at 1100°C for 24 hours and retested for gas permeability at room temperature, Fig. 33. An overall reduction in the permeability was observed in both the infiltrated and uninfiltrated samples, due to the sintering of the whole coating. Compared to the unsealed 72 samples however, there was a considerable reduction in the through flow of the treated samples as before the heat treatment. 0 2 4 6 8 10 12 14 16 18 Time of Flow (min) Figure 33. Permeability of the coatings after heat treatment at 1100°C for 24 hours There was a decrease in permeability by 49% with 4 minutes of infiltration (K0 = 11.55 x 10"4 centi-darcys, K4 = 6.0 x 10"4 centi-darcys), 83% with 10 minutes of infiltration (K10 = 3.0 x 10"4 centi-darcys) and over 85% with 15 minutes of infiltration (Kl5 = 2.22 x 10"4 centi-darcys). The test results showed that there was no deterioration in the sealant properties at elevated temperatures. This test was carried out to establish that the sealing treatment would enhance the high temperature corrosion resistance of the coatings in corrosive gaseous environments by reducing coating permeability. 73 5.4 Hardness Measurements Results of the Depth Sensing Indentation (DSI) tests, at loads of 300 mN (30g), evaluated on the cross section of sprayed coatings are shown in Fig. 34. The spray formed zirconia coatings were infiltrated with alumina precursor (2M) for 0 and 10 minutes and then heat treated in a furnace at 600°C for 15 minutes. 2500 . Infiltration time: 2000 . -A / \ A i \ — 0 min. (MPa) 1500 -1 \ 1 \ JL 1 0 R M N rdness rdness 1000 -re X 500 -0 0 200 400 600 800 Distance (|im) Figure 34. Depth sensing indentation tests on spray formed zirconia-yttria coating infiltrated with alumina precursor (2M) for 0 and 10 minutes, heat treated in a furnace at 600°C for 15 minutes In the untreated sample (0 min. infiltrated) the hardness fluctuated between 9-15 GPa . This large scatter is due to the fact that this hardness test is localized and thus sensitive to inter and intra splat characteristics. The hardness at the surface with 10 minutes infiltration increased to 18 GPa on an average and was consistent to a depth of ~ 250 um, which was approximately the penetration depth expected with that amount of immersion time. The 74 consistency along with higher hardness values of the surface up to the infiltrated depth can be explained by the presence of alumina (hardness 18-23 GPa) in the pores of the zirconia coating (hardness 10-15 GPa) . The consistency in the hardness values up to the infiltrated depth, in comparison to the large variations in the hardness of the parent coating shows that the treated coating has a smoother surface than the untreated coating. Increased hardness of the surface can be translated into improved wear resistance of the treated coatings. This has been confirmed experimentally by other sources [68]. 5.5 Peel Adhesion Measurements The results of the peel adhesion tests (PAT) of zirconia coatings of thickness 0.2 mm on steel foils (thickness 0.17 mm) are shown in Fig. 35a-c. These coatings were infiltrated with prehydrolyzed ethyl silicate and subsequently flame heat treated at 700°C for 15 seconds. Such a short heat treatment was necessary to avoid damaging the foils. A calibration was performed to measure the load consumed by friction and plastic work in the foil (3.75 N) and this was subtracted from the peel load to give the adhesion of the interface. The peel strength of the coatings on the steel foil increased from about 430 N/m in the untreated coating, (Fig. 35a) to 650 N/m with 15 minutes of infiltration, (Fig. 35b) and to over 750 N/m with 30 minutes of infiltration (Fig. 35c). These values were calculated as the average of the load/width values obtained over the length of the coating. The increase in adhesion by 220 N/m and 320 N/m with 15 and 20 minutes of infiltration are significant within the experimental error involved in calibration [48, 49]. With 15 and 30 minutes of infiltration it 75 was expected that the precursor would have infiltrated all the way through the coating to the interface. This was confirmed by the high Si count at the interface, in an EDS analysis for the 30 minute infiltrated coating (Fig. I-10, Appendix I). The increase in the adhesion can be attributed to the high reactivity of the colloidal particles and presence of surface hydroxyl groups. It is postulated that the hydroxyl group of metal alkoxides combines with the surface of the metal oxide and the ceramic at the interface to form M - O - M bonds on heat treatment (M = metal, O = oxygen). The improved anchoring of the coating to the substrate due to the formation of the chemical bonds may explain the increase in adhesion seen with silica infiltration. The increase in adhesion strength with 30 minutes of infiltration as compared to the 15 minutes of infiltration is due to the formation of more M - O - M bonds at the interface. Peel adhesion tests were also carried out on zirconia coatings (thickness 0.23 mm) deposited on Ni foils (thickness 0.17 mm), Fig. 36. These coatings were infiltrated with alumina precursors for various lengths of time and subsequently heat treated in a 600°C flame for 15 seconds. A constant load of 5.3 N was subtracted from the peel load to account for the friction and the plastic work in the foil. As in the previous case, an increase in the adhesion strengths was observed. The overall high peel strengths for nickel as compared to that of steel was expected because nickel, being more ductile, can deform to accommodate the stresses developed during spraying. The peel strength increased from approximately 1540 N/m in the untreated specimen (Fig. 36a) to about 1850 N/m for the coatings infiltrated with alumina precursor for an hour (Fig. 36c). With 4 minutes of infiltration (Fig. 36b) there was no change seen in the adhesion. 76 1000 (a) 25 35 45 55 P o s i t i o n ( m m ) 65 75 30 40 P o s i t i o n (mm) (b) 40 P o s i t i o n ( m m ) (c) 70 Figure 35. Peel strength of zirconia coatings on steel foils infiltrated with prehydrolysed ethyl silicate for 0 minutes (a), 15 minutes (b) and 30 minutes (c) respectively, heat treated in a flame (700°C)for 15 seconds 77 2500 co £ 500 -30 40 50 60 70 80 P o s i t i o n (mm) 2500 _ 2000 -E 2 1000 -V ) "3 a! 500 -10 20 30 40 50 60 P o s i t i o n (mm) 2500 £ 1000 CO o S. 500 10 15 (C) 20 25 30 35 40 P o s i t i o n (mm) 55^ Figure 36. Peel strength of zirconia coatings on nickel foils infiltrated with 2M alumina precursor for 0 minutes (a), 4 minutes (b) and lhour (c) respectively, and heat treated in a flame (600°C) for 15 seconds 78 5.6 Burner Rig Tests Figure 37 shows the temperature profile of the coated specimens, held in the flame for 54 seconds and subsequently quenched in an air jet. The maximum test temperature was 1170°C + 20°C for the first 200 cycles and was measured using a thermocouple, positioned just below the surface (in the coating), connected to a chart recorder. o o 1400 1200 H 1000 800 2 a. E a> 600 400 200 H Time (seconds) Figure 37. Temperature profile of the cycles followed in the burner rig test The specimens of 18% Y 2 0 3 stabilized Zr0 2 TBCs used in the test were supplied by Pratt & Whitney, Quebec, Canada. SEM/EDS analysis showed that the bond coat was a MoNiCrAlY alloy and the substrate was a Ni based alloy. The thickness of the coating and the bond coat were measured using a SEM and was found to be 0.38 mm and 0.22 mm 79 respectively, Fig. 38. Four of the five specimens tested were infiltrated with alumina precursor (2M) for 5, 10, 20, and 30 minutes. The infiltrated specimens were dried and subsequently heat treated at 600°C in a flame for 15 seconds. During the test, the cracking behavior of the coatings was visually examined and recorded in terms of the total area exposed. The definition of coating failure in these types of experiments is rather arbitrary. In some investigations failure is considered to have occurred when the ceramic coating spalled from 50% or more of the test zone, whereas in others complete spallation is used as the criterion [25]. Appearance of cracks is used in some investigations as a sign of failure [25]. Hence it is difficult to compare the results from two different investigations. In our setup failure was considered to have taken place when 75% or more of the test zone (area of the coating exposed) spalled, Fig. I-11, Appendix I. After the specimens had been through 200 cycles in the burner rig test the maximum test temperature was increased to 1270°C + 20°C, in an attempt to accelerate coating failure. Fig. 39 summarizes the results obtained in the cyclic burner rig tests. There was a significant increase in the cyclic life, from about 560 cycles in the untreated coating to about 780 cycles in the coating treated with 30 minutes of infiltration. Figure 40 shows a SEM photograph of the bond coat-coating interface of the untreated sample after 500 cycles, showing an evidence of delamination prior to spallation (compare with as received coating in Fig. 38). It is believed that the stresses generated by the mismatch in thermal expansion coefficients of the coating and bond coat materials and bond coat oxidation leads to spallation at the interface. With increasing infiltration time, the permeability of the coating is reduced and this reduces the rate of bond coat oxidation. 80 Figure 38. SEM micrograph showing the coating morphology 81 30 rrin 500 550 600 650 700 No. of cycles to failure 750 800 Figure 39. Number of cycles to failure in the burner rig test for YSZ TBCs infiltrated for 0, 5, 10, 20 and 30 minutes with 2M alumina precursor A similar test was carried out to assess the possibility of improvement in the cyclic life of the coated specimens in service. In this test the samples provided by Pratt & Whitney were heated to 1270°C ± 20°C in the flame for 56 seconds and then cooled with a stream of compressed air for 56 seconds. About 15 seconds in the flame was required for the hot zone of the samples to reach the test temperature. As in the previous test six specimens could be analyzed simultaneously. Some samples were additionally infiltrated with the precursors for 30 minutes after being cycled through 300 cycles in the burner rig. There were two untreated coatings at the start of the test, Fig. 41. After 300 cycles one of them was infiltrated with alumina precursor for 30 minutes. There were also two coatings infiltrated with alumina precursor for 30 minutes, one of them treated again after 300 cycles. Figure 41 summarizes the results obtained from this test, in terms of the number of cycles to failure. 82 Figure 40. SEM micrograph of the interface showing onset of delamination after 500 cycles 83 P r i m a r y Infiltration: after 0 c y c l e s S e c o n d a r y infiltration: after 3 0 0 c y c l e s Mate r ia l : 1 8 Y S Z - T B C o n M o N i C r A l Y B C / Ni a l loy 1. C o n t r o l N o t infiltrated 2. C o n t r o l Infil. @ 3 0 0 els 3. T e s t Infil @ 0 els 4 . T e s t Infil @ 0 & 3 0 0 e ls F a i l u r e 5 4 0 c y c l e s F a i l u r e 7 4 0 c y c l e s T i F a i l u r e 8 7 0 c y c l e s F a i l u r e 1 0 5 0 cycles) 1 2 0 0 4 0 0 6 0 0 8 0 0 1 0 0 0 1 2 0 0 Number of burner rig cycles to failure Figure 41. Number of cycles to failure for 18% Y 2 0 3 - Z r 0 2 TBCs in the burner rig test, infiltrated with 2M alumina precursor at various stages of the test. The cyclic life of the untreated coating increased from 540 cycles to 740 cycles with 30 minutes of infiltration (after 300 cycles) with 2M alumina precursor. The life of the initially infiltrated coatings increased from 870 cycles to 1050 cycles with 30 minutes of infiltration (after 300 cycles) with alumina precursor. The infiltration treatment was carried out after 300 cycles to establish the feasibility and usefulness of this sealing technique to improve the life of coatings that are already in service. Coating failure in the burner rig tests occurred in the ceramic layer parallel and close to the bond coat-coating interface. This is in agreement with predictions of coating failure close to the interface, as this is the weakest part of the coating/substrate system. In the burner rig, 84 the test specimens are subject to severe thermo-mechanical fatigue conditions, leading to gradual accumulation of damage at the interface. It is believed that the infiltration treatment by active ceramic powders affects these microcracks, e.g. through bridging of the crack. Other possible modes of coating degradation are due to thermally activated processes such as bond coat oxidation, phase transformation and shrinkage due to ceramic sintering. Bond coat oxidation is slower in the sealed samples due to the reduced permeability to air. In all the cases, failure is influenced by the initial residual stress state of the coating. It has been shown that thermal compressive stresses due to cooling of the coating from processing temperature are the highest at the interface and diminish towards the coating surface [66]. Thermal compressive stresses due to heating on the other hand are the highest at the surface of the coating. Bond coat oxidation affects the coating durability in several ways. The oxides formed usually occupy a higher volume than the parent metal and thus oxidation generates tensile stresses at the interface. The oxides formed at the interface may also crack, severely restricting the coating durability. Oxidation could also reduce bond coat ductility, leading to fracture at lower mismatch strains. Similar burner rig tests with intermittent infiltrations were performed on YSZ TBCs modified using silica precursor (prehydrolyzed ethyl silicate). In this case spallation occurred very early, in about 30 cycles. With 30 minutes of infiltration the precursor penetrates through the porous coating to the bond coat-coating interface, as indicated by the presence of crystalline silica at the interface (Fig. 1-12, Appendix I). It is believed that tensile stresses are generated at the interface due to the polymorphic transformations in silica on thermal cycling to 1270°C. It has been shown that amorphous silica transforms to crystobalite at 1200°C [67]. 85 On repetitive heating/cooling, crystoballite undergoes polymorphic transformation to quartz with a volume expansion of 13% (d = 2.66 g/cc for quartz and d = 2.32 g/cc for crystobalite). It was observed that in the silica treated coatings, uniform delamination occurred during the heating cycle of the test. This confirms our hypothesis that the tensile stresses developed at the interface during heating to 1270°C were responsible for the catastrophic loss of coating so early during the test. Figure 42 shows the brittle fracture surface of the bond coat-coating interface. The decrease in coating durability was in contrast to the significant increase in adhesion (from 400 N/m to 750N/m with 30 minutes of infiltration) seen with silica infiltration. However due to the polymorphic transformation in silica during thermal cycling and its accompanying volume expansions, silica infiltration is not beneficial for applications subject to thermal fatigue. 5.7 Microstructural Analysis SEM studies were done on the infiltrated coatings to determine the distribution of the sealant material within the parent coating. Figure 43 shows the surface of an unsealed spray formed YSZ specimen which was highly microcracked. The microcracked structure is typical of thermal sprayed coatings. Figure 44 shows the surface of a coating which was infiltrated with alumina precursor for 10 minutes and heat treated at 600°C in a furnace for 15 minutes. The white fluffy regions (due to higher intensity of reflected secondary electrons) are representative of alumina as determined by EDS. It is believed that the precursors are effective in penetrating into the pores and cracks, and decompose in situ on heat treatment to 86 reduce the pore diameters and bridge the cracks. This was confirmed by the gas permeability tests which show a significant decrease in coating permeability with infiltration, and burner rig tests which show increased lifetime. However, it is difficult to see the distribution of the sealant material inside the pores and microcracks of the coating in a SEM because of the absorption of X-rays inside the pores by the surrounding material. The distribution of the sealant material (alumina) in the coating was also studied by mapping the element (aluminum) in the precursor across the cross section of the coating. Figure 45a shows a SEM photograph of the cross section of the a YSZ coating infiltrated with alumina precursor for 10 minutes and Figure 45b shows a X-ray map of aluminum (from alumina) distributed in the parent zirconia, spray-formed by Waploc Ceramics. The high concentration of aluminum on the surface and in the pores of the coating shows that the precursor is effective in sealing the surface and in penetrating into the pores of the coating. This serves to reduce coating permeability and heal the microcracks. In addition to improving the above mentioned properties, it is postulated that such a sealing treatment would potentially improve the aerodynamic performance of these coatings by reducing surface roughness [22]. Preliminary results of the tests carried out by NASA suggest a substantial reduction of flame radiation on smoothening the coating surface with alumina powder. This can be applied to this type of sealing treatment as well [22]. Figure 42. SEM micrograph of the bond coat fracture surface Figure 43. SEM micrograph of the microcracked YSZ coating before sealing 89 Figure 44. SEM micrograph of the YSZ coating infiltrated with 2M alumina precursor for 10 minutes and heat treated at 600°C in a furnace for 15 minutes 90 (b) x 2 . 0 k 7 9 9 0 1 0 k V 2 0 J J m Figure 45. SEM image of the cross-section of a YSZ coating, (a) and X-ray map of aluminum (from the infiltrated alumina) in the above coating (b). Sample was infiltrated with 2M alumina precursor for 10 minutes and heat treated at 600°C in a furnace for 15 minutes 91 6. C O N C L U S I O N S Sealing treatments were carried out to improve the protective characteristics of porous plasma sprayed ceramic coatings in aqueous and gaseous environments. The sealing was performed using liquid precursors of alumina and silica. The effects of such a sealing treatment on the permeability and mechanical properties of the parent coating have been characterized. The results of the investigations on post treated coating permeability and durability indicate a potential for the application of this method for high temperature environments subject to corrosive gaseous attack, such as in the gas turbine engines. X-ray map of the sealant material in the parent coating shows that the infiltrant is effective in penetrating into the pores and capillaries of the coating. Coating permeability is effectively retarded as a result of the pore and capillary diameters being reduced. The suitability of sealed thermal sprayed coatings as improved thermal barriers has been demonstrated by the various tests performed. The shift in the polarization curves towards lower current values is indicative of the permeability decrease to electrolyte, achieved with infiltration. The decrease in gas permeability was confirmed by tests measured at room temperature. Tests carried out on samples heat treated at 1100°C for 24 hours show a likewise decrease in gas permeability. The peel adhesion test shows a significant increase in the peel adhesion strengths (from 400 N/m to 750 N/m) of the coatings deposited on steel foils, with silica.infiltration. In the case of YSZ deposited on nickel foils the adhesion increased by 300 92 N/m with alumina infiltration. Porosity of the modified coatings is decreased minimally (~ 4-5%) and thus this sealing technique preserves the overall strain tolerance of porous coatings. Surface hardness of the YSZ coating was increased (from ~12 GPa to 18 GPa) with alumina infiltration. The consistency in the higher hardness values (18 GPa) as compared to the large fluctuations in the untreated coating (9-15 GPa) confirms the effectiveness of this sealing technique. It is expected therefore that the modified coatings would exhibit improved wear resistance. Finally, burner rig cyclic tests show an improvement in the cyclic life of the YSZ coating from 560 to 780 cycles with alumina infiltration. Burner rig tests on coatings additionally sealed after 300 cycles show a further improvement in the cyclic life of these treated coatings to 1050 cycles. It appears therefore that this technique is equally suited for sealing fresh coatings as well as to enhance the life expectancy of coatings already in service. In contrast to alumina, silica was found to have a deleterious effect on the cyclic life. This is attributed to the polymorphic transformations in silica when cycled through its transition temperatures. 93 7. R E C O M M E N D E D F U T U R E W O R K This work has an exploratory character and a number of experimental methods was introduced for testing of thermal sprayed coatings for the first time. The future work should focus on the following topics: 1) The use of other alkoxide precursors with identical chemistry as the parent coating should be explored. 2) The burner rig test should be modified to allow the use of impure fuels with the flexibility to introduce controlled amounts of impurities such as vanadium and sodium sulfate. The test conditions can be chosen so as to simulate actual engine operating conditions 3) Further work is required to characterize the pitting behavior of the coating/substrate systems and to study the efficacy of the post-deposition treatment in other aqueous solutions. For example the effect of the sealing treatment on the pitting behavior can be studied on coatings deposited on thin foils by measuring the time required to perforate the foils. 4) The depth sensing indentation technique should be utilized to further characterize the variations in Young's Modulus across the sealed and unsealed layers of the coating. Similar tests can be used to study the variations in fracture toughness across the cross section of the coating with and without sealing treatments. 94 5) The accumulation of fatigue cracks due to the burner rig test as a function of the infiltration conditions should be studied quantitatively, using for example the attenuation of ultrasonic wave at the interface. 6) Field tests with modified Thermal Barrier Coatings should be carried out in the combustors of real turbine engines to evaluate any deleterious effects of such a sol-gel post deposition treatment on the other parts of the engine. 95 R E F E R E N C E S 1. G. Geiger, "Ceramic Coatings", American Ceramic Society Bulletin. Vol. 71, No. 10, 1992, pp. 1470-1481. 2. S. B. Quadri, E. F. Skelton, C. Kim, M . Z. Harford, "Electron-Beam Deposition of Zirconia Films Stabilized in High Temperature Phases by Different Oxides", Surface and Coatings Technology. Vol. 49, pp. 67-70. 3. R. C. Tucker, "An Overview of Thermal Spray Coatings Compared to Physical and Chemical Vapor Deposited Coatings", Proceedings ofiTSC'95. Kobe, pp. 253-258. 4. S. M . Meier, D. K. Gupta, K. D. Sheffler, "Ceramic Thermal Barrier Coatings for Commercial Gas Turbine Engines", Journal of Materials. March 1991, pp. 50-53. 5. W. J. Lackey, D. P. Stinton, G. A. Cerny, A. C. Schaffhauser, "Ceramic Coatings for Advanced Heat Engine- A Review and Projection", Advanced Ceramic Materials. Vol. 2, No. 1, 1987, pp. 24-30. 6. R. A. Miller, R. G. Garlick, J. L. Smialek, "Phase Distribution in Plasma Sprayed Zirconia Yttria", Ceramic Bulletin. Vol. 62, No. 12, 1983, pp. 1355-1358. 7. M . I. Boulos, A short course on "Fundamentals of Materials Processing Using Thermal Plasma Technology", October 17-18, 1989. 8. T. P. Gavrilenko, Y. A. Nikolaev, V. Y. Ulianitsky, "D-Gun "OB" Detonation Spraying", Proceedings ofITSC'95. May, pp. 425-429. 9. H. Herman, "Plasma Sprayed Coatings", Scientific American. September 1988, pp. 112-117. 10. M . L. Thorpe, W. R. Kratochvil, "Inductive Heating for a New Type of Plasma", Thermal Spray Technology-New Ideas and Processes. NTSC'88. Ohio, pp. 9-18. 11. A. C. Ackbreth, W. C. Roman, "Laser Diagnostics for Thermal Spray Processing of Materials", Thermal Spray Technology-New Ideas and Processes. NTSC'88. Ohio, pp. 47-54. 12. S Das, V. K. Suri, U. Chandra, K. Sampath, "A One Dimensional Mathematical Model for Selecting Plasma Spray Process Parameters", Proceedings of the 7th NTSC'94. Boston, pp. 1-11. 13. D. K. Das, R. Sivakumar, "Modelling of the Temperature and the Velocity of ceramics Powder in the Plasma Flame", Acta Metall. Mater.. Vol. 38, No. 11, pp. 2193-98. 96 14. R. C . Dykhuizen, M . F. Smith, "Investigations into the Plasma Spray Process", Thermal Spray Technology-New Ideas and Processes. NTSC'88 . Ohio, pp. 61-66. 15. M . Vardelle, D . Vardelle, P. Fauchias, I. Sarry, "Characterization of a D . C . Plasma Torch with Axial Injection of Powders", Proceedings of the Third N T S C ' 9 0 . California, pp. 65-69. 16. "Plasma Spray System", Thermal Spray Society, Advanced Materials and Processes. Vo l . 148, No. 6, D e c , 1995, pp. 22-23. 17. E . H . Lutz, "Microstructure and Properties of Plasma Ceramics", Journal of American Ceramic Society. Vo l . 77, No. 5, pp. 1274-80. 18. P. Vincenzini, "Zirconia Thermal Barrier Coatings for Engine Applications", Industrial Ceramics. Vo l . 10, No. 3, 1990, pp. 113-126. 19. J. R. Brandon, R. Taylor, "Phase Stability of Zirconia based Thermal Barrier Coatings-Part I. Zirconia Yttria Alloys" Surface and Coatings Technology. Vo l . 46, 1991, pp. 75-90 20. B . C . Wu, E . Chang, D. Tu, "Degradation Mechanisms of Z r 0 2 - 8 wt.% Y 2 0 3 / Ni-22Cr-1 0 A M Y Thermal Barrier Coatings", Journal of American Ceramic Society. V o l . 72, No. 2, 1989, pp. 212-218. 21. K . M . Jasim, R. D . Rawlings, D . R. F. West, "Stability of Z r 0 2 - Y 2 0 3 t' Phase Formed During Laser Sealing", Material Science and Technology. Vol . 8, January, 1992, pp. 83-91. 22. F . S. Stepka, "Summary of N A S A research on Thermal Barrier Coatings", Vol./Iss 770343. 23. T. Sitonen, T. Konos, P. O. Kettunen, "Corrosion Properties of Stainless Steel Coatings made by Different Methods of Thermal Spraying", Proceedings of the 7th N T S C 9 4 . June, Boston, pp. 105-110. 24. T. Kinos, P. Sitonen, P. Kettunen, "Effect of Post Treatments on the Corrosion Resistance of Plasma Sprayed Duplex Stainless Coating in Salt Water", Proceedings of the 7th N T S C ' 9 4 . June, Boston, pp. 537-542. 25. I. Zaplatiynsky, "Performance of Lazer-Glazed Zirconia Thermal Barrier Coatings in Cyclic Oxidation and Corrosion Burner Rig Tests", Thin Solid Films. Vo l . 95, No. 2, 1982, pp. 275-283. 26. K . M . Jasim, R. D. Rawlings, D. R. F. West, "Pulsed Laser Processing of Thermal Barrier Coatings", Journal of Material Science. Vol . 26, 1991, pp. 909-916. 97 27. R. Sivakumar, B. L. Mordike, "Laser Melting of Plasma Sprayed Ceramic Coatings", Surface Engineering. Vol. 4, No. 2, 1988, pp. 127-139. 28. T. E. Strangman, "Thermal Barrier Coatings for Turbine Airfoils", Metallurgical and Protective Coatings. 1984, pp. 93-105. 29. K. M . Jasim, R. D. Rawlings, D. R. F. West, "Pulsed Laser Sealing of Plasma-Sprayed Layers of 8 wt.% Yttria Stabilized Zirconia", Journal of Materials Science. Vol. 27, 1992, pp. 3903-3910. 30. H. L. Tsai, P. C. Tsai, "Thermal Cycle Response of Laser-Glazed Plasma-Sprayed Thermal Barrier Coatings", Material Science and Engineering. A: Structural Materials, Vol. 177, 1994, pp. 227-232. 31. Edward W. Washburn, " The Dynamics of Capillary Flow", The Physical Review. Second Series, Vol. 27, No. 3, March, 1921, pp. 273-284. 32. Erik O. Einset, "Capillary Infiltration Rates into Porous Media with Application to Silicomp Processing", Journal of American Ceramic Society. Vol. 79, No. 2, 1996, pp. 333-338. 33. K. Moriya, H. Tomino, Y. Kandaka, T. Hara, A. Ohmori, "Sealing of Plasma-Sprayed Ceramic Coatings by Sol-Gel Process", Proceedings of the NTSC'94. Boston, pp. 549-553. 34. K. Moriya, W. Zhao, A. Ohmori, "Improvement of Plasma-Sprayed Ceramic Coatings Treated by Sol-Gel Process", ITSC'95. Kobe, pp. 1017-1021. 35. Y. Arata, A. Ohmori, C-J Li , "Electrochemical Methods to Evaluate the Connected Porosity in Ceramic Coatings", Thin Solid Films. Vol. 156, 1988, pp. 315-325. 36. M . Inaba, K. Tani, Y. Harada, A. Naakahira, "Corrosion Behavior of Carbide Cermet Sprayed Coating in Alkaline aqueous Solution", Proceedings of the ITSC'95. May, Kobe, pp. 791-796. 37. M . G. Fontanna, Corrosion Engineering. McGraw-Hill Book Company, 1986. 38. R. Divakar, S. G. Seshadri, M . Srinivasan, "Electrochemical Techniques for Corrosion Rate Determination in Ceramics", Journal of American Ceramic Society. Vol. 72, No. 5, 1989, pp. 780-784. 39. A. A. Ashary, R. C. Tucker, "Electrochemical Corrosion Studies of Alloys Plasma Sprayed with Cr 2 0 3 ", Surface and Coatings Technology. Vol. 39, No. 40, 1989, pp. 701-709. 98 40. M . R. Dorfman, J. A. DeBarro, "Development and Applications of Corrosion Resistant Thermal Sprayed Coatings", Proceedings of the .TSC'95. May, Kobe, pp. 567-572. 41. L. C. Erickson, T. Troczynski, H. W. Hawthorne, "A Micromechanical Evaluation of Plasma Sprayed Coatings using Depth Sensing Indentation", Proceeding of ITSC'95. Kobe, May, pp. 743-74. 42. D. S. Rickerby, A. Mathews, "Advanced Surface Coatings: A Handbook of Surface Engineering", 1991, pp. 315-324. 43. S. D. Brown, B. A. Chapman, G. P. Wirtz, "Fracture Kinetics and the Mechanical Measurement of Adherence", Thermal Spray Technology, Proceedings of the NTSC'88. October, Ohio, pp. 153-157. 44. M . J. Filiaggi, R. M . Pilliar, "Mechanical Testing of Plasma Sprayed Ceramic Coatings on Metal Substrates: Interfacial, Fracture Toughness and Tensile Bond Strength", Journal of Materials Science. 1991, pp. 5383-5395. 45. M . Sexsmith, T. Troczynski, "Development of Peel Adhesion Test for Thermal Spray Coatings", Proceedings of the 14th ITSC'95. May, Kobe, pp. 897-902. 46. M . Sexsmith, T. Troczynski, "Variation in Coating Properties Across a Spray Pattern, Proceedings of the 7th NTSC'94. Boston, pp. 751-757. 47. M . Sexsmith, T. Troczynski, "Peel Adhesion Test for Thermal Spray Coatings", Journal of Thermal Spray Technology, to be published (May 1996). 48. E. Breslauer & T. Troczynski, submitted for publication in J. Adh. Sci. & Tech.. (1996). 49. M . Sexsmith, T. Troczynski, E. Bresslauer, accepted for publication in J. Adh. Sci. & Tech.. (1996) 50. A S T M D903-49 (1983), Standard Test for Peel or Stripping Strength of Adhesive Bonds. 51. A S T M C-577 (1978), Standard Test Method for Permeability of Refractories. 52. Melvin E. Woods, "Thermal Fatigue Rig Testing of Thermal Barrier Coatings for Internal Combustion Engines", Thermal Spray Technology-New Ideas and Processes". 1989, 245-253. 53. R. A. Miller, "Oxidation Based Model for Thermal Barrier Coating Life", Journal of American Ceramic Society. Vol. 67. No. 8. 1984. pp. 517-21. 99 54. K. S. Mazdiyasni, "Chemical Synthesis of Single and Mixed Phase Oxide Ceramics", Chemical Processing of Ceramics, edited by Burtand I. Lee, Edward J. A. Pope, pp. 176-186. 55. J. Zarzycki, "Physical-Chemical Factors in Sol-Gel Processes", Science of Ceramic Chemical Processing, edited by L. L. Hench, D. R. Ulrich, 1985, pp. 21-36. 56. K. S. Mazdiyasni, C. T. Lynch, J. S. Smith, "Preparation of High-Purity Submicron Refractory Oxides", Journal of American Ceramic Society. Vol. 48, No. 7, 1965, pp. 372-375. 57. T. Mah, E. E. Hermes, K. S. Mazdiyansi, "Multicomponent Ceramic Powders", Chemical Processing of Ceramics, edited by Burtand I. Lee, Edward J. A. Pope, pp. 75-101. 58. K. S. Mazdiyansi, C. T. Lynch, J. S. Smith, "Cubic Phase Stabilization of Translucent Yttria-Zirconia at very Low Temperatures", Journal of the American Ceramic Society. Vol. 50, No. 10, 1967, pp. 532-537. 59. B. E. Yoldas, "Alumina Gels that form Porous Transparent A1 20 3", Journal of Material Science. Vol. 10, 1975, pp. 1856-1860. 60. E. Lugsceider, P. Jokiel, V. Messeschmidt, G. Beckschulte, "Subsequent Sealing of Thermally Sprayed Coatings to increase Corrosion Resistance", The Institute of Materials. August, 1994, pp. 46-51. 61. S. J. Glass, D. J. Green, "Surface Modification of Ceramics by Infiltration", Advanced Ceramic Materials. Vol. 2, No. 2, 1987, pp. 129-131. 62. S.J. Glass, D. J. Green, "A Novel Technique for the Surface Modification of Y-TZP", Advances in Ceramics. Vol. 24A. 1988. pp. 311-318. 63. S. J. Glass, D. J. Green, "Fabrication of Multiphase Particulate Ceramics by infiltration into Powder Compacts", Ceramic Transactions. Vol. IB, Ceramic Powder Science II, 1988, pp. 784-791. 64. L. L. Hench, "Use of Drying Control Chemical Additives in Controlling Sol-Gel Processing", Science of Ceramic Chemical Processing. 1985, edited by L. L. Hench, D. R. Ulrich, pp. 52-64. 100 65. S. Kesavan, T. A. Mozhi, B. E. Wilde, "Potential-pH Diagrams for the Fe-Cl"-H20 System at 25 to 150°C", Corrosion Science. Vol. 45, No. 3, 1989, pp. 213-214. 66. R. A. Miller, C. E. Lowell, "Failure Mechanisms of Thermal Barrier Coatings exposed to Elevated Temperatures" Thin Solid Films. Vol. 95, 1982, pp. 265-273. 67. W. C. Wei, J. W. Halloran, "Phase Transformation of Diphasic Aluminosilicate Gels", Journal of American Ceramic Society. Vol. 71, No. 3, 1988, pp. 166-172. 68. L. C. Erickson, H. M . Hawthorne, T. Troczynski, S. Hogmark, M . Olssom, "Wear and Micromechanical Properties of Microstructurally Improved Ceramic Plasma Sprayed Coatings", Proceedings of the International Symposium on Advanced Ceramics for Structural and Tribological Applications. 1995, pp. 557-70. 101 Appendix I TGA /XRD /SEM Analysis 10 8 4 2 SAMPLE: Al(OC,H 7) 3 wet gel X-AXIS Y-AXIS RUN NO. OPERATOR DATE George John -TEMP. SCALE 100 °c SCALE (SCALE 2 n "9: ch SHIFT TIME Incl fl. . inch ir SETTING X 2) HEA ATM TIMi TINfi RATF (. mm. Ar SIZE 9.3 mg. SCALE (ALT.) SUPPRESSION mg. E CONSTANT sec. . -- - -- - - -• •> : *l • - - - - - -v - - -- - - - - - -1 -s > --z -- 1 -- ----- - - --7 -- - - -- - -- 1 -- - - - - - --- -- - - -- - -- - -200 300 400 TEMPERATURE*, °C 600 700 800 900 Figure 1-1. Thermo-Gravimetric Analysis of the wet alumina gel 102 20 16 E 8 4 S A M P L E : A L / X , ( d r y gel) X - A X I S Y - A X I S R U N NO. DATE O P E R A T O R O e o r g e J o h n TEMP. S C A L E loo °c S C A L E {SCALE 4 mg. S H I F T T I M E 0 incl inch i n c h SETING x 21 H E A A T M TIME riwr, BATF W 1-mm. A i r S I Z E 21.2 mg. S C A L E (ALT.) S U P P R E S S I O N o Tig. • C O N S T A N T 1 s e c . . - - --- : - 1 - - -- - - -- - - -- - --- - -- -- : .: : : ; -'--- : : : - z : - ----- -- - ; - - -- - --- -- ; : ; - - z -- • - -= - - - - - - _ -- -- - - - - - - - -100 200 300 400 T E M P E R A T U R E * *C 6 0 0 700 800 900 Figure 1-2. Thermo-Gravimetric Analysis of the dry alumina gel 103 42 38 30 26 SAMPLE: SiOj X-AXIS Y-AXIS RUN NO. OPERAT< DATE eorge DR c John TEMP. SCALE 100 °C SCALE (SCALE 4 mg. SHIFT TIME inch 0 inrh inch S E T T I N G x 2) HEA ATM TIM T I N f i R A T F 4U "C min. Air S I 7 F 44.6 ma. SCALE (ALT.) SUPPRESSION 22 mg. i C O N S T A N T sec. : -- - 1 -1 - - -- • ; - - - -: : - - - -• -- N -; -- - - -- - -- ; ; -_ L. - - - : - - - 1. -T -- - -- - -; :. - 'J - - --_ - --; - : z -- -200 300 400 TEMPERATURE*, °C 000 700 800 900 Figure 1-3. Thermo-Gravimetric Analysis of the dry silica gel Figure 1-4. Thermo-Gravimetric Analysis of the dry magnesia-alumina gel 105 20 30 40 50 60 70 80 90 Angle (2 theta) Figure 1-5. (a) X-ray diffraction of the dry alumina gel (major phase AlOOH) (b) X-ray diffraction of the alumina powder calcined at 600°C in a furnace for 1 hour (c) X-ray diffraction of the alumina powder calcined at 850°C in a furnace for 1 hour 106 Fig. 1-5 contd. (d) X-ray diffraction of the alumina powder calcined at 1050°C in a furnace for 1 hour (e) X-ray diffraction of the alumina powder calcined at 1150°C in a furnace for 1 hour 107 80 Angle (2-theta) Figure 1-6. (a) X-ray diffraction of the silica gel calcined at 700°C for lhour (b) X-ray diffraction of the silica gel calcined at 1200°C for 1 hour 100 80 Angle (2 theta) Figure 1-7. (a) X-Ray diffraction of the dry magnesia alumina gel (amorphous) (b) X-Ray diffraction of the gel heat treated in a furnace at 650°C for 1 hour 1 0 - J u l - 1 9 9 6 1 1 : 4 1 : 4 3 ppt on s u r f a c e V e r t = 2 0 0 0 0 coun t ; D i s p = 1 P r e s e t = E l a p s e d * 200 sees 100 sees EC 1 :Fe; 4- 0 . 0 0 0 Range : 10 .230 keV I n t e g r a l 0 10 .110 453456 1 0 ~ J u l - 1 9 9 6 U i 4 3 i 3 1 pp t on s u r f a c e A c c e l e r a t i n g v o l t a g e £0. 0 KeV B e a n - s a m p l e i n c i d e n c e a n g l e 9 0 . 0 d e g r e e s X r a y e m e r g e n c e a n g l e £0.0 d e g r e e s X r a y - w indow i n c i d e n c e a n g l e 0 . 0 d e g r e e s Window t h i c k n e s s 7 . 5 m i c r o n s STflNDBRDUESS EDS A N A L Y S I S (ZAF CORRECTIONS V I P MAGIC V) ELEMENT & L I N E WEIGHT PERCENT ATOMIC P R E C I S I O N PERCENT* £ SIGMA K - R A T I O * * ITER C l KA Mn KA Fe KA B. £0 3 . 40 S B . 40 12 . 33 3 . 30 8 4 . 37 0 . 06 0 . 0 5 0 . S0 0 . 0 6 0 0 0 . 0349 0 . 9051 TOTAL 1 0 0 . 0 0 * N O T E : ATOMIC PERCENT i s n o r n a l i z e d t o 100 * * N O T E : K - R A T I O w h e r e R K - R A T I O x R r e f e r e n c e ( s t a n d a r d ) / r e f e r e n c e ( s a m p l e ) NORMALIZATION FACTOR: 0 . 964 E l e m e n t C l Mn Fe ZAF C o r r e c t i o n F a c t o r s Z f a c t o r A f a c t o r 0 . 7 8 S 0 . 9 5 7 0 . 946 1. 8SS 1 . 0 9 2 1 .071 F f a c t o r 0 . 9 9 4 0 . 966 1. 0 0 0 Figure 1-8. EDS of the corroded surface Figure 1-9. Pourbaix diagram for the Fe-CT-H 20 system at 25°C constructed for CT activity of 5.6 [58] 23-Jan-1996 13:42:42 Sample Depos i t Sur f a c e ( C o a t i n g ) Vert= 5080 counts Disp= 1 Pr eset = E l a p s e d ' 200 sees 200 sees ;Fe! 10.230 keV I n t e g r a l 0 10.110 224022 2 3 - J a n ~ 1 9 9 6 1 3 : 5 3 : 3 8 S a n p i e D e p o t i t S u r f a A c c e l e r a t i n g v o l t a g e 2 0 . 0 KeV B e a n — s a r a p l e i n c i d e n c e a n g l e 9 0 . 0 d e g r e e s X r a y e m e r g e n c e a n g l e 3 0 . 0 d e g r e e s X r a y - w i n d o w i n c i d e n c e a n g l e 0 . 0 d e g r e e * W i n d o w t h i c k n e s s 2 0 . 0 m i c r o n s STAN DA R D L E S S E D S A N A L Y S I S CZAF C O R R E C T I O N S V I B MAGIC V) E L E M E N T WEIGHT P R E C I S I O N O X I D E O X I D E NO. OF C A T I O N S S L I N E K - R A T I O * * P E R C E N T 2 S IGMA FORMULA P E R C E N T IN FORMULA fll KA 0 . 0 3 6 0 S . 7 5 0 . I E R 1 E 0 3 10 . 0 7 0 . 6 4 2 B S i KB 0 . 0 2 9 3 3 . 74 0 . 0 S S i D 2 7 . 9 9 0 . 4 0 0 9 F e KR 0 . 5 9 7 7 £ 9 . as 0 . IE* F o £ 0 3 4 £ . te 1 . 6 1 1 1 Y L B 0 . etsti 5 . 6.8 0 . 13 V S 0 3 7 . 2 2 0 . 1 9 2 7 Ir L B 0 . 2 7 5 0 2 3 . 13 0 . 18 Z r 0 2 3 1 . 24 0 . 7 6 4 1 • a 3 1 . S S 9 D E T E R M I N E D BY S T O I C H I Q M E T R Y NUMBER O F C A T I O N S C A L C U L A T E D ON B A S I S O F 6 O X Y G E N ATOMS. * * N O T E : K - R A T I O w h e r e R 1 K - R A T I O x R r e f e r e n c e ( s t a n d a r d ) / r e f e r e n c e ( s a n p l e ) N O R M A L I Z A T I O N F A C T O R t 0 . 8 5 3 Z A F C o r r e c t i o n F a c t o r s E l e m e n t Z f a c t o r A f a c t o r F f a c t o r A l 0 . £ .35 5 . 6 2 8 0 . 991 S i 0 . 6 5 3 4 . 3 7 8 0 . 9 S 7 F e 0 . 926 1. 195 1. 0 0 0 Y 0 . 53fc 3 . 7 9 0 0 . 9 9 9 Z r 0 . 561 3 . 324 0 . 999 Figure I-10. EDS of the peeled foil showing presence of silica at the interface 112 Figure 1-11. Failure criteria for the burner rig test 113 1800 _ 1600 .. 1400 1200 1000 W c 800 c 600 --400 -. 200 0 50 Angle (2-theta) Figure 1-12. X-ray diffraction of the spalled interface showing presence of crystalline silica at the interface of a yttria stabilized zirconia coating 


Citation Scheme:


Citations by CSL (citeproc-js)

Usage Statistics



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"
                            async >
IIIF logo Our image viewer uses the IIIF 2.0 standard. To load this item in other compatible viewers, use this url:


Related Items