Design of Suspension Plasma Sprayed Thermal Barrier Coatings

ASHISH GANVIR PhD Thesis Production Technology 2018 No. 20

Design of Suspension Plasma Sprayed Thermal Barrier Coatings Thermal barrier coatings (TBCs) are widely used to provide thermal insulation to com- Design of Suspension Plasma ponents in both power generation and aero engine gas-turbines. Improvement in performance of TBCs will lead to more efficient and durable engines that is why there is a con- DESIGN OF SUSPENSION PLASMA SPRAYED THERMAL BARRIER COATINGS Sprayed Thermal Barrier Coatings stant need to ﬁnd new TBC solutions. Suspension plasma spraying is a relatively newer deposition technique and gaining increasing research interest since it has been shown to be capable of producing TBCs with superior performance. This work was focused on identifying the relationships between process parameters, coating microstructure, thermal conductivity and lifetime in suspension plasma sprayed TBCs. A further objective was to utilize these relationships to enable tailoring of the TBC Ashish Ganvir microstructure for superior performance compared to state-of-the-art TBC used in industry today, i.e. atmospheric plasma sprayed TBCs using solid feedstock. A variety of microstructures, such as highly porous, vertically cracked and columnar, were produced and investigated. Progressive evolution of the much sought-after columnar microstructure produced by SPS was experimentally demonstrated. It was shown that there are strong relationships between microstructure, thermo-mechanical properties and performance of the coatings. Finally, based on the experimental results, a tailored columnar microstructure design for a superior TBC performance is also proposed.

Ashish Ganvir Ashish comes from a small town ‘Amravati’, Maharashtra, India and received his bachelor’s degree in materials and metallurgical engineering from Indian Insti- tute of Technology Kanpur, India in 2013. Before starting his master’s study, he did an apprenticeship in summer 2013 at International Advanced Research Centre for Powder Metallurgy and New Materials (ARCI), Hyderabad, India. He received his master’s degree in manufacturing engineering from University West, Trollhättan, Sweden in 2014. In 2016, he worked in industry as a visiting researcher at Innovnano materials, Coimbra, Portugal. He was also a visiting research fellow at the Materials Synthesis and Processing (IEK-1) research centre of Forschungszentrum Jülich, Germany in summer 2016. His main interests are 2018 NO.20 in the areas of materials science, surface & coatings engineering, powder metallurgy & defect analysis.

ISBN 978-91-87531-92-7 (Printed version) ISBN 978-91-87531-91-0 (Electronic version)

Tryck: BrandFactory, april 2018. PhD Thesis Production Technology 2018 No. 20

Design of Suspension Plasma Sprayed Thermal Barrier Coatings

Ashish Ganvir

University West SE-46186 Trollhättan Sweden +46 52022 30 00 www.hv.se

Trollhättan, Sweden, 2018

Dedicated to all those who financially supported my education!

III

Acknowledgements

It is with immense gratitude that I acknowledge the financial support provided by the Västra Götalandsregionen, Sweden, to accomplish this research work.

First and foremost, I wish to thank my supervisors Prof. Nicolaie Markocsan, Dr. Mohit Gupta and Prof. Per Nylen for the continuous support in my PhD study, for their patience and immense knowledge. Their guidance helped me in doing a great research and writing of this thesis. I could not have imagined having better supervisors for my PhD study. It’s a privilege to be your student and I wish I keep learning from you all. Thank you for always motivating and offering me an immense freedom during this research!

My sincere thanks also go to Dr. Nicholas Curry, Mr. Jonas Olsson and Mr. Stefan Björklund for their immense help with spraying and sharing their vital knowledge to enhance my practical skills. It’s been a pleasure to work with you all! Thanks to Prof. Shrikant Joshi, Prof. Robert Vassen and Prof. Uta Klement for their valuable collaboration and feedback during my research. Also, thanks to Ms. Johanna Ekberg, a former PhD student at Chalmers University of Technology for making this collaborative project ‘PROSAM’ between Chalmers University of Technology and University West more successful.

I must also acknowledge the support from various partners associated with my research: GKN Aerospace, Trollhättan, Sweden; Innovnano Coimbra, Portugal; IPP, Prague, Czech Republic; Swerea IVF, Gothenburg, Sweden; and Chalmers University of Technology, Gothenburg, Sweden.

I would like to thank all my friends and colleagues at PTC for making these days so enjoyable. It’s been a great fun to be around you.

I am grateful to my family: my parents, brother and sisters who have provided me moral and emotional support so far in my life. I am also lucky and grateful to have very special people in my life who always loved and believed in me.

I must thank to all my seniors and teachers right from the school times till now for constantly believing in me and guiding me.

And finally, to all those individuals/institutions/government agencies back in India who supported me financially when it was needed the most. Thank you!

Ashish Ganvir 15th of June 2018, Trollhättan

Populärvetenskaplig Sammanfattning

Nyckelord: Mikrostruktur; Termiska barriärbeläggningar; Axialmatning; Suspensionsplasmasprutning; Värmeledningsförmåga; Brotthållfastheten; Livstid

Titel: Design av suspensionsplasmasprutade termiska värmebarriärbeläggningar

Termiska värmebarriärbeläggningar (TBC) används i stor utsträckning på gasturbinkomponenter för att åstadkomma termisk isolering och oxidationsskydd. TBCs, i kombination med avancerad kylning, möjliggör högre förbränningstemperaturer i gasturbinen även över smälttemperaturen för metalliska material. Det finns ett ständigt behov, främst av miljöskäl, för att både minska bränsleförbrukning och emissioner och att öka förbränningstemperaturen vilket kräver nya typer av TBC-lösningar.

Genom att använda en suspension vid termisk sprutning, kan nya typer av TBC framställas. Suspensionsplasmasprutning och lösningsbaserad plasmasprutning är exempel på tekniker som kan användas. Dessa tillvägagångssätt, som är alternativ till den konventionella tekniken med tillsatsmaterial i fast form, röner ett allt större forskningsintresse: Anledningen till det stora forskningsintresset är möjligheten att producera beläggningar med överlägsna funktionella egenskaper.

Syftet med detta avhandlingsarbete var att undersöka samband mellan processparametrar, beläggningarnas mikrostruktur, beläggningarnas värmeledningsförmåga och beläggningarnas livstid för termisk sprutade värmebarriärbeläggningar där tillsatsmaterialet införs i lågan i flytande form. Ett ytterligare syfte var att utnyttja denna kunskap för att producera en värmebarriärbeläggning med lägre värmeledningsförmåga jämfört med state-of- the-art inom industrin idag, det vill säga då tillsatsmaterialet tillförs i fast form. Olika spruttekniker såsom suspensionssprutning med plasma, lösningsbaserad sprutning med plasma, suspension med höghastighetsflamsprutning undersöktes och jämfördes med plasmasprutning där tillsatsmaterialet införts i fast form.

En mängd olika mikrostrukturer, såsom mycket porösa, strukturer med vertikal sprickor och kolumnära strukturer erhölls. Det visades att det finns starka samband mellan mikrostrukturerna, de termo-mekaniska egenskaperna och de funktionella egenskaperna hos beläggningarna. Specifikt axiell suspensionsplasmasprutning visades som en mycket lovande teknik för att producera olika mikrostrukturer samt beläggningar med lång hållbarhet. Baserat på experimentella resultat föreslås också en skräddarsydd kolumnär mikrostruktur för en överlägsen TBC-prestanda.

VII

Abstract

Title: Design of Suspension Plasma Sprayed Thermal Barrier Coatings

Thermal barrier coatings (TBCs) are widely used on gas turbine components to provide thermal insulation, which in combination with advanced cooling, can enable the gas turbine to operate at significantly higher temperatures even above the melting temperature of the metallic components. There is a permanent need, mainly due to environmental reasons, to increase the combustion temperature in turbines, hence new TBC solutions are needed.

By using a liquid feedstock in thermal spraying, new types of TBCs can be produced. Suspension plasma/flame or solution precursor plasma spraying are examples of techniques that can be utilized for liquid feedstock thermal spraying. This approach of using suspension and solution feedstock, which is an alternative to the conventional solid powder feedstock spraying, is gaining increasing research interest since it has been shown to be capable of producing coatings with superior performance.

The objective of this research work was to identify relationships between process parameters, coating microstructure, thermal conductivity and lifetime in suspension plasma sprayed TBCs. A further objective was to utilize these relationships to enable tailoring of the TBC microstructure for superior performance compared to state-of-the-art TBC used in industry today, i.e. solid feedstock plasma sprayed TBCs. Different spraying techniques, namely suspension high velocity oxy fuel, solution precursor plasma and suspension plasma spraying (with axial and radial feeding) were explored and compared to solid feedstock plasma spraying.

A variety of microstructures, such as highly porous, vertically cracked and columnar, were produced and investigated. It was shown that there are strong relationships between microstructure, thermo-mechanical properties and performance of the coatings. Specifically, axial suspension plasma spraying was shown as a very promising technique to produce various microstructures as well as highly durable coatings. Based on the experimental results, a tailored columnar microstructure design for a superior TBC performance is also proposed.

Keywords: Microstructure; Thermal Barrier Coatings; Axial Injection; Suspension Plasma Spraying; Porosity; Thermal Conductivity; Fracture Toughness; Lifetime.

ISBN: 978-91-87531-92-7

Preface

The experimental research performed in this thesis work was primarily conducted at the Production Technology Centre within the Thermal Spray research group at University West, Trollhättan, Sweden. Additional experiments required for this work were performed at several research partners namely Chalmers University of Technology, Gothenburg, Sweden; Jönköping University, Jönköping, Sweden; GKN Aerospace, Trollhättan, Sweden; Swerea IVF, Gothenburg, Sweden; Forschungszentrum Jülich, Germany; Institute of Plasma Physics (IPP), Prague, Czech Republic and Innovnano Materials, Coimbra, Portugal.

The research conducted in this thesis work was performed in two parts:

1. The first part was dedicated towards screening of several existing liquid feedstock thermal spray techniques and comparing the different resultant microstructures. Techniques such as Solution Precursor Plasma Spraying (SPPS), Suspension High Velocity Oxy Fuel Spraying (S-HVOF), Suspension Plasma Spraying with radial injection (SPS) and axial injection (ASPS) were employed to produce TBCs with different microstructures. In addition, conventional powder feedstock plasma spraying (APS) was utilized as a reference process.

2. The second part of this work was primarily dedicated towards establishing the relationships between process characteristics– microstructure-properties and performance of ASPS produced TBCs to enable designing highly durable TBCs for gas turbine applications. Additional efforts were made during this part of the work to establish a reliable and easy way of porosity measurement procedure for such in- homogeneously porous new microstructures produced by ASPS. Moreover, the much sought after column formation shadowing theory that is widely used in SPS was also experimentally proved with a systematic progressive microstructure evolution study.

List of Appended Publications

Paper A. Characterization of thermal barrier coatings produced by various thermal spray techniques using solid powder, suspension and solution precursor feedstock material Ashish Ganvir, Nicholas Curry, Sivakumar Govindarajan, Nicolaie Markocsan International Journal of Applied Ceramic Technology, vol. 13, no. 7, pp. 324-332, September 2015

Ashish Ganvir has performed all experimental characterization, analyzed all results, designed the structure of the article and had the main responsibility in writing the article. Co-authors contributed in formulating concepts and ideas, planning the project, spraying and article editing. In addition, Nicholas Curry also contributed in writing the article.

Paper B. Comparative study of suspension plasma sprayed and suspension high velocity oxy-fuel sprayed YSZ thermal barrier coatings Ashish Ganvir, Nicholas Curry, Nicolaie Markocsan, Per Nylen, Filofteia-Laura Toma Journal of Surface and Coating Technology, vol. 268, pp. 70–76, April 2015

Paper C. Characterization of microstructure and thermal properties of YSZ coatings obtained by axial suspension plasma spraying (ASPS) Ashish Ganvir, Nicholas Curry, Stefan Björklund, Nicolaie Markocsan, Per Nylen Journal of Thermal Spray Technology, vol. 24, no. 7, pp. 1195-1204, June 2015

XIII

Paper D. Influence of microstructure on thermal properties of axial suspension plasma sprayed YSZ thermal barrier coatings Ashish Ganvir, Nicholas Curry, Nicolaie Markocsan, Per Nylen, Shrikant Joshi, Monika Vilemova, Zdenek Pala Journal of Thermal Spray Technology, vol. 25, no. 1, pp. 202-212, November 2015

Ashish Ganvir has performed all experimental characterization except mercury intrusion porosimetry (MIP) and x-ray diffraction (XRD), analyzed all results, designed the structure of the article and had the main responsibility in writing the article. Co-authors contributed in formulating concepts and ideas, planning the project, spraying and article editing. MIP and XRD were carried out by co-authors Monika Vilemova and Zdenek Pala respectively.

Paper E. Influence of isothermal heat treatment on porosity and crystallite size in axial suspension plasma sprayed thermal barrier coatings for gas turbine applications Ashish Ganvir, Nicolaie Markocsan, Shrikant Joshi Journal of Coatings, vol. 7, no. 1, pp. 1-14, December 2016

Paper F. Experimental visualization of microstructure evolution during suspension plasma spraying of thermal barrier coatings Ashish Ganvir, Rosa Calinas, Nicolaie Markocsan, Nicholas Curry, Shrikant Joshi Submitted to Journal of Materials and Design, in Apr. 2018

Ashish Ganvir has performed all experimental characterization, analyzed all results, designed the structure of the article and had the main responsibility in writing the article. Co-authors contributed in planning the project, suspension fabrication, spraying of suspensions, article structuring and editing.

XIV

LIST OF APPENDED PUBLICATIONS

Paper G. Failure analysis of thermally cycled columnar thermal barrier coatings produced by high-velocity-air fuel and axial-suspension- plasma spraying: A design perspective Ashish Ganvir, Venkateswaran Vaidhyanathan, Nicolaie Markocsan, Mohit Gupta, Zdenek Pala, Frantisek Lukac Journal of Ceramics International, vol. 44, no. 3, pp. 3161-3172, February 2018

Ashish Ganvir performed all experimental characterization except image analysis (IA) and X-ray diffraction (XRD), analyzed all results, designed the structure of the article and had the main responsibility in writing the article. Co-authors contributed in formulating concepts and ideas, planning the project, and article editing. IA and XRD were carried out by co-authors Venkateswaran Vaidhyanathan and Frantisek Lukac and Zdenek Pala respectively.

Paper H. Tailoring columnar microstructure of axial suspension plasma sprayed TBCs for superior thermal shock performance Ashish Ganvir, Shrikant Joshi, Nicolaie Markocsan, Robert Vassen Journal of Materials and Design, vol. 144, no. 15, pp. 192-208, April 2018

Ashish Ganvir has performed all experimental characterization except mercury intrusion porosimetry (MIP) and X-ray diffraction (XRD), analyzed all results, designed the structure of the article and had the main responsibility in writing the article. Co-authors contributed in formulating concepts and ideas, planning the project, and article structuring and editing.

Relevant Non-Appended Publications

1. A. Ganvir, C. Kumara, M. Gupta, P. Nylen, “Thermal Conductivity in Suspension Sprayed Thermal Barrier Coatings: Modeling and Experiments,” Journal of Thermal Spray Technology, vol. 26, no. 1-2, pp. 71–82, January 2017

2. U. Klement, J. Ekberg, A. Ganvir, “EBSD Analysis and Assessment of Porosity in Thermal Barrier Coatings Produced by Axial Suspension Plasma Spraying (ASPS),” Material Science Forum, vol. 879, pp. 972– 977, 2017

3. J. Ekberg, A. Ganvir, U. Klement, S. Creci, L. Nordstierna, “The influence of heat treatments on the porosity of suspension plasma sprayed yttria-stabilized zirconia coatings” Journal of Thermal Spray Technology, vol. 27, no. 3, pp. 391–401, February 2018

XVII

Table of Content

Acknowledgements ...... V Populärvetenskaplig Sammanfattning ...... VII Abstract ...... IX Preface ...... XI List of Appended Publications ...... XIII Relevant Non-Appended Publications...... XVII Table of Content ...... XIX Abbreviations / Nomenclature ...... XXIII 1 Introduction ...... 1 1.1 Thesis outline ...... 1 1.2 Objective and research questions ...... 2 1.3 Scope ...... 2 1.4 Limitations ...... 3 2 Background ...... 5 2.1 Thermal Barrier Coating (TBC) ...... 7 2.1.1 Bond Coat (BC) ...... 7 2.1.2 Thermally grown oxide (TGO) ...... 8 2.1.3 Top Coat (TC) ...... 8 2.2 Thermal spraying of a TBC ...... 10 2.2.1 High velocity oxy fuel spraying (HVOF) ...... 12 2.2.2 High velocity air fuel spraying (HVAF) ...... 12 2.2.3 Atmospheric Plasma spraying (APS) ...... 13 2.3 Liquid feedstock spraying ...... 15 2.3.1 Suspension and Solution ...... 16 2.3.2 Solution Precursor Plasma Spraying (SPPS)...... 17 2.3.3 Suspension High Velocity Oxy Fuel (S-HVOF) ...... 17 2.3.4 Suspension Plasma Spraying (SPS) ...... 18 3 Process characteristics of SPS ...... 19 3.1 Type of solvent ...... 19

XIX

3.2 Initial size of the solute powder particles ...... 19 3.3 Solid load in the suspension...... 20 3.4 Injection of suspension...... 20 3.4.1 Axial injection versus radial injection ...... 21 3.5 Sequence of events in SPS coating formation ...... 22 Stage 1: Injection ...... 23 Stage 2: Atomization ...... 24 Stage 3: Evaporation of the solvent ...... 25 Stage 4: Agglomeration of the solute particles ...... 26 Stage 5: Melting of the solute particles ...... 26 Stage 6: Deposition of the molten droplets ...... 27 3.6 Coating formation in SPS ...... 27 4 Characteristics of TBC ...... 31 4.1 Microstructure ...... 31 4.1.1 Solid powder feedstock APS sprayed coatings ...... 31 4.1.2 Suspension plasma sprayed coatings ...... 32 4.2 Porosity ...... 35 4.3 Thermal conductivity ...... 37 4.3.1 Heat transfer in conventional APS & SPS TCs ...... 37 4.4 Mechanical properties ...... 39 4.5 Sintering of ceramic TC in TBCs ...... 40 4.6 Lifetime of TBCs ...... 41 4.6.1 Lifetime and failure associated with the BC and TGO ..... 42 4.6.2 Lifetime and failure associated with the ceramic TC ...... 44 5 Experimental techniques used to characterize TBCs .. 47 5.1 Microstructure evaluation ...... 47 5.2 Porosity measurement ...... 48 5.3 Evaluation of thermal conductivity ...... 49 5.4 Evaluation of mechanical properties ...... 51 5.5 Evaluation of sintering behaviour ...... 53

TABLE OF CONTENT

5.6 Evaluation of thermal cyclic fatigue (TCF) lifetime ...... 53 5.7 Evaluation of thermal-shock (TS) lifetime ...... 54 6 Microstructure selection and coating design for TBC applications using SPS ...... 55 6.1 Processing and microstructure ...... 55 6.1.1 Influence of process parameters on TC microstructure ... 57 6.1.2 Tailoring of porosity in a SPS TC microstructure ...... 59 6.2 Effect of SPS TC microstructure on thermal conductivity ... 63 6.3 Sintering behaviour of SPS TCs in TBCs ...... 66 6.4 Designing SPS TBCs for superior TCF performance ...... 68 6.5 Designing SPS TBCs for superior TS performance ...... 69 7 Conclusions ...... 73 8 Future work ...... 75 9 Summary of appended publications ...... 77 References ...... 83 APPENDED PUBLICATIONS ...... 95

XXI

Abbreviations / Nomenclature

APS: Atmospheric Plasma Spraying

ASPS: Axial Suspension Plasma Spraying

BC: Bond Coat

C: Columns

CMAS (Calcia-Magnesia-Alumina-Silica or molten silicate)

CSN: Chromia, Spinel and Nickel oxide

CTE: Coefficient of Thermal Expansion

D50: Median size (diameter) of a particle in a certain particle size distribution.

DSC: Differential Scanning Calorimetry

DVC: Dense Vertically Cracked

EBPVD: Electron Beam Physical Vapor Deposition

FC: Fine cracks

HVAF: High Velocity Air Fuel

HVOF: High Velocity Oxy Fuel

HVSFS: High Velocity Suspension Flame Spraying

IA: Image analysis

IC: Inter-columnar spacing

IGTs: Industrial Gas Turbines

LFA: Laser Flash Analysis

LPPS: Low Pressure Plasma Spray

MIP: Mercury Intrusion Porosimetry

MP: Micron Pores

XXIII

NP: Nano pores

OM: Optical Microscopy

S: Spherical particles

SEM: Scanning Electron Microscopy

S-HVOF: Suspension-High Velocity Oxy Fuel

SP: Sub-micron Pores

SPPS: Solution Precursor Plasma Spraying

SPS: Suspension Plasma Spraying

TBC: Thermal Barrier Coating

TC: Top Coat

TCF: Thermal Cyclic Fatigue

TGO: Thermally Grown Oxide

TS: Thermal shock

VPS: Vacuum Plasma Spray

XRD: X-ray Diffraction

YSZ: Yttria Stabilized Zirconia

XXIV

1 Introduction

A brief introduction to the outline of this thesis, objective of this work and research questions are presented in this chapter. Additionally, the scope and limitations of this thesis work are also described in brief. 1.1 Thesis outline

This thesis is organized in nine chapters. The coverage in each chapter is briefly summarized below:

Chapter 1 provides a brief outline of the thesis, discusses the objective, research questions and also the scope and limitations of the research work. Chapter 2 provides a detailed background of the overall motivation behind this research and why the research work is important to the scientific society, specifically to the thermal spray community. Additionally, this chapter also provides a brief description of a thermal barrier coating (TBC) and its deposition, and furthermore, outlines the advantages of spraying a fine structured powder feedstock instead of a conventional coarse solid powder feedstock. In addition, the chapter also talks about the necessity of using liquid feedstock instead of solid powder feedstock. Finally, this chapter also provides details about different liquid feedstock processes used in this work. Chapter 3 outlines the process characteristics of suspension plasma spraying (SPS) and the theory behind coating formation in SPS.

Chapter 4 presents the main characteristics of the TBCs in general, and especially highlights the microstructure, porosity, thermal conductivity, mechanical properties, sintering behaviour and lifetime. Chapter 5 describes the characterization techniques utilized to investigate the TBCs produced in this work.

Chapter 6 is the core of this work. The chapter starts with briefly describing the several critical challenges faced so far in understanding the SPS microstructures as pointed out in the literature. Building upon those challenges, the chapter then describes the microstructural selection and design route for obtaining highly durable SPS TBC for various gas turbine applications. A complete design cycle right from the processing-microstructure-properties to TBCs performance is discussed in brief summarizing the major findings from all the appended articles in this work. Chapter 7 provides the conclusions of this thesis work.

Finally, the future work for this research and a summary of all the journal papers included in this thesis are given in Chapter 8 and 9 respectively. 1.2 Objective and research questions

The objective of this research work was to explore relationships between process parameters, coating microstructure, thermal properties, mechanical properties and performance in suspension plasma sprayed TBCs. A further objective was to utilize this knowledge to produce a SPS TBC with lower thermal conductivity compared to state-of-the-art thermal spray TBCs used in industry and propose a tailored microstructure design of SPS TBC for achieving a superior performance.

The work was accomplished by trying to answer the following research questions:

1. How is the TBC microstructure produced by liquid feedstock thermal spraying affected by changing spraying techniques and process parameters? 2. What are the microstructural features present in SPS TBCs and how can these microstructural features be characterized and quantified? 3. How can thermal conductivity of plasma sprayed TBCs be reduced by using liquid feedstock instead of solid feedstock? 4. How do different microstructural features influence the lifetime and sintering behaviour of SPS TBCs? 1.3 Scope

The work was mainly focused on TBCs for industrial gas turbines (IGTs) and aero engine turbines applications. However, the screening study also revealed that varying microstructural features and porosity could result in a wide variety of coating microstructures, which could be applicable for various other applications ranging from solid oxide fuel cells to rocket nozzles.

The work was mainly focused on studying the columnar type microstructures and understanding the influence of various microstructural features present in those microstructures on the performance of suspension plasma sprayed 8 wt. % yttria partially stabilised zirconia (8YSZ) TBC. However, the results and knowledge gained from this study can also be utilized for studying such a relationship between microstructure and performance for any other possible types of microstructures, produced by any liquid feedstock thermal spray techniques using 8YSZ or any other ceramic materials such as gadolinium zirconate or dysprosia stabilized zirconia etc.

INTRODUCTION

1.4 Limitations

The major limitations of this work can be briefly summarized as follows:

i. Material Only one ceramic top coat (TC) material, namely 8YSZ, was used in various forms (powder/suspension/solution precursor). For bond coat (BC) a material of a composition NiCoCrAlY/CoNiCrAlY was used, and for substrate Hastelloy® X (Ni-based super alloy) was used. However, use of alternative ceramic TC materials needs to be investigated in order to understand their role in further improving the performance of TBCs.

ii. Spraying techniques and process parameters Various liquid feedstock thermal spray techniques were utilized in this work however only axial suspension plasma spraying (ASPS) was explored in more detail. Techniques such as solution precursor plasma spraying (SPPS), suspension high velocity oxy fuel spraying (S-HVOF) and SPS that were used in this work can also be further explored in greater detail.

While exploring the ASPS process, systematic experimental design was used to understand the influence of individual feedstock (suspension) properties in microstructure formation and coating performance. However, no such systematic study was conducted in studying the role of individual plasma spray parameters, and plasma gun hardware aspects which can also potentially affect the microstructure and performance in SPS TBCs.

iii. Characterization techniques Several characterization techniques were utilized in this work to investigate coating microstructure, properties and performance. The limitations of those techniques are described here:

Microstructure was analysed using low and medium high resolution microscopes. Use of high resolution microscopy to observe very fine scale pores and/or grains and grain boundaries would provide even better insights to further refine the correlations between coating microstructure and performance.

Water impregnation and mercury intrusion porosimetry (MIP) used for porosity analysis can only measure the open pores (pores which can be accessible by water or mercury). MIP can over-predict the porosity at low pressure regime due to the significant influence of TC surface roughness. In addition, MIP can also introduce errors in measuring the pore size of pores/cracks that have a narrow opening to the coating surface but then widens-up later through the coating thickness which can under-predict the mean pore size. Image analysis can reveal both open and

closed pores, but is sensitive to sample preparation. The above constraints should be borne in mind while utilizing the results of the present study, although the broad qualitative trends involving porosity are expected to remain valid.

Laser flash analysis (LFA), which was utilised for thermal diffusivity measurements, is prone to uncertainty in measurements due to the transparent nature of zirconia to the wavelengths of light produced in the LFA laser. This issue was minimized by coating a very thin layer of graphite in order to allow the laser light to be absorbed.

Instrumented micro-indentation was used for measuring fracture toughness. The indents were made at half of the TC thickness and in the centre of the column (away from the inter-columnar spacing (IC)). The measured toughness can only provide the local toughness and can be considered to be the upper bounds of effective toughness in coatings investigated.

Lifetime was investigated using isothermal thermal cyclic fatigue (TCF) test and thermal shock (TS) using burner rig test. The failure in the TCF test is mainly driven by the nature of the BC (oxidation of the BC) whereas in the latter one it is mainly driven by the TC nature. However, combination of both BC oxidation as well as TC effect should be considered simultaneously as it can be the case in real life (e.g. in commercial aircrafts). Hence, a modified lifetime testing procedure combining these two factors could also provide more information about coatings lifetime, especially for commercial aircraft applications.

iv. TBC performance The performance of a TBC is driven by its low thermal conductivity and high lifetime. Thermal conductivity is discussed in more depth however in case of lifetime many more aspects were not considered which could have further strengthened this work if considered. For instance, the influence of BC preparation, that is, by varying BC spraying techniques, varying the BC roughness, post treatment of the BC by shot peening or by heat treating etc. Additionally, measurement of residual stresses in the ceramic TC could have also given more information about the lifetime of SPS TBCs.

Other performance related aspects such as sintering resistance, erosion, wear and high temperature corrosion could have further strengthened this work by adding more information regarding TBCs’ overall performance.

2 Background

Over the past 40 years, the gas turbine manufacturing industry has extensively used TBCs in various sectors, such as power generation and aero engines for high temperature applications [1], [2]. The hot sections in gas turbines include the combustor, turbine blades and vanes, and the afterburner which can be seen in Figure 1 (the parts in red and orange).

Figure 1: Schematic of gas turbine RM-12 aero engine [courtesy: GKN Aerospace]

In today’s aero engines, the hot gas temperature exceeds by more than 150 ᵒC the softening point of the metallic structure of the turbine material, which usually is a Ni-based super-alloy [1], [2]. This requires extensive cooling, which reduces the efficiency of the engine [3]–[5]. Thermodynamics suggest that the efficiency of gas turbine engines can be increased by increasing the combustion temperature [3]. However increase in combustion temperature is limited by the inherent temperature capability of the Ni-based super-alloys. There is a continuous desire to increase combustion temperature to achieve higher gas turbine engine efficiencies [3]. TBCs play a key role in enabling this [1], [2]. Enhancing their insulating capability and augmenting their durability is a continuous challenge.

Due to its thermal insulation properties, there is a transient temperature drop across the TBC in-service conditions. This drop allows higher turbine entry temperatures and hence, higher engine efficiency [1]–[3]. The extent of temperature drop can be increased either by lowering the thermal conductivity of the insulating layer of ceramic TC or by increasing the TC thickness (see section 2.1.3 for TC definition). This temperature drop also prolongs the component life in extreme operating conditions. The thickness of the TC is mainly limited by the

weight penalties, coating durability and cost issues. Though there are some regions such as combustor where thicker coatings are applied; thicker coatings may result in lower lifetime due to the increased residual stresses in the coating [6], especially if they are used on turbine blades and vanes where coatings can experience excessive thermo-mechanical loading. Hence, reduction in thermal conductivity with optimal coating thickness remains the most feasible option.

In general, thermal conductivity of a TBC can be reduced if a low thermal conductivity TC material is used. However, this is limited by the availability of the materials that fulfil the necessary requirements of a TC, e.g. low thermal conductivity, high thermal expansion matching with the metallic BC, thermal and phase stability at elevated temperatures, high fracture toughness etc. Although there has been significant progress in developing new TC materials, the most used material so far is YSZ, a ceramic [7] that was introduced in the late ‘70s [8]. Another possible route of reducing the thermal conductivity is by changing the microstructure of the deposited TC [9], [10], since the microstructure, i.e. microstructural features such as pores and cracks strongly affect the thermal properties of the TC.

Presence of various microstructural features in different microstructures of TCs, such as pores and cracks of different shape, size and orientation [5], [11]–[13], make these microstructures significantly different from each other. It’s these microstructural features that, as previously stated, can significantly influence the heat transfer through the coating [5], [10], [13]. Higher total porosity can for instance significantly reduce the thermal conductivity of the TBC [13], [14].

Other important characteristics of a TBC than having a low thermal conductivity are, excellent thermal cyclic durability and thermal shock resistance, phase stability at higher temperatures for longer exposure, and low cost (this is a relative factor and it is not as big driver as the performance). Depending on the application, a balance between these properties is needed. For example, aero engine turbines operate at peak power for shorter periods during take-off and landing, hence experience relatively frequent thermal-cycling and in some cases (e.g. military aircrafts) also a severe thermal shock, whereas industrial gas turbines (IGTs) run typically for longer duration at constant temperature and hence experience significantly less frequent thermal-cycling and negligible thermal shock. This would require for instance, an excellent thermal shock resistance along with optimum balance between other properties for TBCs used on military aero- turbine engines.

BACKGROUND

2.1 Thermal Barrier Coating (TBC)

TBC is a multilayer coating system and there are three primary constituents in a system as shown in Figure 2: (1) a metallic bond coat (BC) typically 150 µm to 300 µm thick [15]; (2) a thermally grown oxide (TGO) which grows from 0 µm to 10 µm (or even higher) between the TC and the BC [16], [17] during the in service life; and (3) the ceramic top coat (TC) of typically 300 µm to 1 mm [18], which acts as a main thermal insulation layer. The BC and the TC layers can be deposited using various coating technologies. The important thing to note here is that the TGO is a layer which is generated due to the oxidation of the BC during the in service life of the TBC. The different layers are described further in the following subsections.

Figure 2: Schematic of TBC system showing each layer, its material and applications 2.1.1 Bond Coat (BC)

The BC is the first layer that is deposited on the Ni-based superalloy substrate. The purpose of this coating is to protect the substrate from oxidation and high temperature corrosion and to provide the necessary bonding of the ceramic to the substrate material. Also, it helps in reducing the mismatch in coefficients of thermal expansion (CTE) between the TC and the substrate.

The BC can be divided into two categories, diffusion (Pt-modified Aluminides) and overlay (MCrAlX-type) coatings [19]. Overlay coatings are typically used in thermal sprayed TBCs (TBCs can also be deposited by other techniques than

thermal spray, such as physical vapor deposition etc.). In overlay coatings, M is commonly Ni, Co, Fe or a combination of these and X is an oxygen-active element, for example Y, Si, Ta, Hf or a combination of these [20]. Usually a mixture of Nickel and Cobalt based coatings are used. 2.1.2 Thermally grown oxide (TGO)

The second layer in a TBC system is the TGO layer (see Figure 2). When a TBC is under operating conditions (during the in-service life), a TGO layer is formed between the TC and the BC at the metal – ceramic interface [16], [21]. This layer is formed due to the oxidation of the BC. The porous ceramic TC allows oxygen to diffuse to the metallic BC, which results in the oxidation of the BC. Even if the ceramic TCs were fully dense, the high ionic diffusivity of oxygen in the ZrO2-based ceramic TC renders it ‘oxygen transparent’ [22], that’s why oxidation still is an issue. The TGO plays an important role in TBCs performance. Failure in TBCs is initiated, in or near the TGO, mostly at the TGO–BC and TGO-TC interface [23]–[25]. The growth of a TGO layer causes stresses during in service in the TBC, due to the difference in CTEs of different layers in the TBC. These stresses will initiate cracks or cause crack growth in the TBC, which finally causes failure of the TBC due to spallation of the TC. Thus, controlling the growth of this oxide is a critical issue to increase the lifetime of the TBC.

It should be noted that several oxides are formed due to the oxidation of the BC, which typically are Cobalt, Chrome, Aluminium alloys, such as Chromia ((Cr, Al)2O3), Spinel (Ni(Cr, Al)2O4), Nickel oxide (NiO), Silica, and Alumina [23]. Chromia, Spinel and Nickel oxides are usually referred together as CSN in the literature. 2.1.3 Top Coat (TC)

The last layer is the ceramic TC (see Figure 2). The main purpose of this layer is to provide the thermal insulation to the substrate.

The choice of a material for the TC layer is determined by some basic requirements such as: high melting point, no phase transformation between room temperature and the operating temperature, low thermal conductivity, chemical inertness, high thermal expansion coefficient, good mechanical properties and good adherence to the metallic BC [10], [26]. Over the years of development of TBCs, YSZ has become the most widely used commercial material because of its superior thermal and functional performance compared to other ceramics [1], [2]. It has been shown [8] that by stabilising zirconia with 3-4 mol. % or 6-8 wt. % of

BACKGROUND yttria, a superior coating can be created; 3-4 mol.% YSZ was shown to have high thermal expansion coefficient, low thermal conductivity ((2.25 푊 푚−1퐾−1) in bulk form), good mechanical properties and good phase stability up to 1200 ᵒC [27], [28]. Other ceramics which can be used as TC materials are mullite, Al2O3, TiO2, CeO2+YSZ, pyrochlores, perovskites, etc. [28].

Zirconia (ZrO2) has three allotropic crystal structures, i.e. monoclinic, tetragonal and cubic. The monoclinic phase is stable below 1197 ᵒC but transforms to tetragonal above this temperature; the tetragonal phase is stable between 1197 ᵒC and 2300 ᵒC; above 2300 ᵒC tetragonal transforms to cubic, which then is stable up to 2700 ᵒC, the melting point of zirconia [29]. This phase transformation is illustrated below:

1197 ᵒ퐶 2300 ᵒ퐶 2700 ᵒ퐶 푀표푛표푐푙𝑖푛𝑖푐 푡푒푡푟푎𝑔표푛푎푙 푐푢푏𝑖푐 푙𝑖푞푢𝑖푑 ↔ ↔ ↔ As the structural phases have different volumes, transformations from one to another are detrimental for the performance of a TBC as it induces cracks due to the corresponding volume change and thus promotes failure of the TC [29]. The largest volume change is from tetragonal to monoclinic (4.5% volume expansion) and moreover it happens at temperatures which are in the range of the in-service temperature of a gas turbine. Thus, an external stabilizer e.g. yttria (Y2O3), ceria (Ce2O3), magnesium oxide (MgO), calcium oxide (CaO) is added to stabilize a desired phase in the material.

Figure 3 explains the effect of yttria content on the stability of the phases present within the YSZ. The addition of yttria in zirconia is critical and it is important to have the content of Y2O3 approximately between 3-4 mol.% in order to obtain a non-transformable tetragonal (T’) phase (in Figure 3, it is shown as 6-11 mol. % since it is 0.5 (Y2O3)), which is very resistant to the transformation. Below 3-4 mol. % Y2O3 content a transformable tetragonal (T) phase is formed, which may undergo phase transformation as explained above.

As it can be seen from the phase diagram (see Figure 3), at higher temperatures, even the metastable non-transformable tetragonal (T’) phase, undergoes a phase separation by diffusion, when aged at temperatures greater than 1200 °C. This can allow the tetragonal to monoclinic (T → M) transformation upon cooling. This transformation, as explained earlier, induces cracks, which then may lead to the failure of the coating. Hence, YSZ becomes an ideal material for applications only up to operating temperature less than 1200 °C [7]. Addition of higher contents of yttria could completely stabilize the high temperature phases, but the mechanical properties (toughness, erosion resistance) are also altered, so that the in-service requirements and lifetime are negatively affected [2], [29].

Figure 3: Phase diagram of zirconia showing the influence of yttria content at respective temperatures on different phases (tetragonal (T), cubic (F) and monoclinic (M)) of zirconia [29] 2.2 Thermal spraying of a TBC

Thermal spray [18] is a technique used to deposit a coating on various structures; in which metallic or non-metallic coating feedstock material, in the form of rod, powder, wire, suspension or solution, is fed into a spray ‘gun’ by a controlled feed system. This feedstock material is heated by electrical (plasma or arc) or chemical (combustion flame) means and then accelerated by the hot gas/plasma plume

BACKGROUND towards the substrate. Accelerated feedstock material then impinges the substrate in the form of molten or semi-molten state droplets, which upon impact form a pancake shaped so called ‘splats’ and adheres to the substrate by rapid solidification and quenching.

This coating technology, which is a branch of surface engineering is considered as one of the production technology methods. This method can be used not only for repairing, rebuilding, and retrofitting machine components but also for restoring original dimensions or applying protective metal layers to various infrastructures, such as bridges, turbines etc. [18], [30].

A major advantage of thermal spraying is that it can be used to produce a coating without a significant heat transfer to the substrate. This avoids thermal distortion and possible damage to the substrate. The major disadvantage of thermal spraying is that it is a ‘line of sight’ process, which makes complex geometry components difficult to spray. Plasma spray physical vapor deposition is another process by which shadowed areas can also be coated [31].

For spraying a TBC, several thermal spraying techniques are available. The BC is usually sprayed by plasma spraying or high velocity oxy-fuel (HVOF) and more recently by high velocity air-fuel (HVAF) spraying [32].

BC needs to be dense (negligible porosity) to serve its purpose in the TBC, that is to minimize the oxygen penetration to the substrate and also to avoid internal oxidation. Due to the high jet velocity involved in the HVAF process (typically over 1000 m/s), a very dense coating can be produced [33].

Feedstock material in HVAF is mostly in the form of powder, and is injected axially into the flame at the nozzle exit. The feedstock material is then partially or fully melted and accelerated towards the substrate. With a very high impact the particles then form a splat. Subsequent splat deposition over each other at high velocity lead to the dense coating formation [33].

The TC is usually sprayed by plasma spraying since this process inherently has enough thermal energy to enable melting of the ceramic powder during the short- dwell time typical of thermal spray processes. EBPVD process is also commercially used for TC deposition as it can produce strain tolerant TCs. In the following subsections, deposition of the TC and BC layers using various thermal spray processes and feedstock materials used in this study are described.

2.2.1 High velocity oxy fuel spraying (HVOF)

HVOF spraying is one of the variant of conventional flame spraying where a warm flame is produced due to the combustion process which acts as a heat source to melt the powder feedstock. In this process a mixture of fuel gases (typically Hydrogen, Propane and Propylene) and Oxygen are introduced into the combustion chamber along with the powder feedstock. The combustion process produces high temperature and pressure in the chamber which in combination with the specifically designed long convergent-divergent nozzle (through which these gases are accelerated) produces a supersonic jet resulting in high particle speed. The feedstock powder particles can fully or partially melt (depending on the fuel/Oxygen) mixture in the chamber as well as during the flight through the nozzle.

The advantages with this process, due to its reasonable process temperature and very high particle velocity on impact, are very dense and adherent coatings [34]. The process is suitable for spraying metallic, cermet as well as ceramic coatings for various applications. This makes it more suitable technique for spraying especially the BCs [23] but also TCs [35] in TBCs. 2.2.2 High velocity air fuel spraying (HVAF)

Another variant and more recent addition to the flame spraying techniques is HVAF spraying. Unlike HVOF, a mixture of compressed air and fuel gas is used in the combustion process which makes HAVF process relatively colder than HVOF [36]. Hence, HVAF is more suitable for spraying metallic powders especially those which are sensitive to high temperature oxidation such as metallic BCs in TBCs.

Advantage with this process over HVOF is the extremely high particle speed resulting in coatings with negligible porosity and good adhesion strength. In addition, the low heat input to the feedstock allows good control of oxide content in some materials which are sensitive to oxidation at high temperature [33], [36]. The low heat input also ensures minimal in-situ particle oxidation and reduced phase transformation or elemental depletion/decomposition of in-flight particles due to the lower in-flight particle temperature [37]. Moreover, higher impact velocity in HVAF can also produce relatively smoother coating surfaces which is considered to be beneficial, especially for the production of BCs in case of SPS TBCs to enhance their strain tolerance by increasing the column density (number of columns per unit length) [38]. In this study, HVAF was used to deposit BCs because of its capability to produce extremely dense, smooth, oxidation resistant and adhesive coatings.

BACKGROUND

2.2.3 Atmospheric Plasma spraying (APS)

Plasma spraying is the conventional technique to deposit TBCs (both TCs as well as BCs) where plasma is used as a heat source to melt the feedstock material. If the process is carried out in inert/controlled atmosphere then it is referred as low pressure plasma spray (LPPS) or vacuum plasma spray (VPS) process. If the process is carried out in a normal atmosphere then it is called as atmospheric plasma spray (APS) which is discussed in more detail in this section. Due to the high temperature in the plasma, which can exceed 20000 °C [30], and thus the high thermal energy, the plasma plume can easily melt any type of feedstock material.

Figure 4: Schematic of conventional powder atmospheric plasma spraying process

The basic principle of this technique consist of melting a consumable (most often powder of size 10 µm to 100 µm [39]) and projecting it as droplets of molten or semi molten particles onto the substrate as can be seen in Figure 4. The modern plasma spraying process uses a direct current electric arc, to generate a stream of high temperature ionized plasma from one or a mixture of inert gases (Ar, He, H2, N2), which act as the spraying heat source. The coating material, in powder form, is conveyed by a carrier gas (typically Argon) stream into the plasma plume, where it is heated and propelled with a typical particle velocity of 200-300 m/s towards the substrate [30]. The accelerated droplet of molten or semi molten particles strike the substrate and under its impact to the substrate forms a splat. Subsequent deposition of splats over each other leads to the formation of a coating, see Figure 5. The generated splat can be in a size range of few tens to few hundreds of micrometre in diameter and few micrometre thick [39], depending on the spraying conditions or particle size used during spraying.

The formed splats, shrink and solidify upon impact, and can form a void or pore in between the splats. Additionally, a splat may also have gases trapped in it, which can also lead to the formation of a pore within the splat when the splat is solidified upon cooling.

To generate a desired coating microstructure by plasma spraying, several process parameters need to be controlled, which are explained through the schematic shown in Figure 5. This figure can be generalized for almost all types of plasma spraying techniques with certain changes depending on the specification of the technique. In general, these process parameters can be categorized as: injector and feedstock parameters, spray gun parameters, robot fixture parameters, substrate and plasma plume parameters. The desired microstructure that forms during powder spraying is controlled by the complete history of powder material used; from production, to in-flight conditions to final impact on the substrate leading to the formation of splats.

The size of the powder particles is very important as it is a major factor influencing the microstructure of the coating and hence its functional properties [40], [41]. Many advantages have been found when the coatings were deposited using fine (submicron and nano-size) powder particles, which are explained below in detail.

Figure 5: Schematic of plasma spray process characteristics showing the various process parameters

BACKGROUND

2.2.3.1 Fine powder feedstock

Spraying fine powder particles compared to conventional coarse powder particles result in coatings having fine microstructure. These fine structured coatings are characterized by their fine (submicron or nano) microstructural features present in the coating microstructure such as fine pores/cracks, fine grains etc. Any nanostructured (or nano-crystalline) material is characterized by a microstructural length or a grain size of 1 nm – 100 nm, whereas submicron structured material is characterized by 0.1 µm to 0.3 µm of its grain size [42].

Such fine structured coatings have shown to have advantages over conventional coatings such as lower thermal diffusivity and thermal conductivity at room temperature [24] as well as at higher temperatures up to 1200 ᵒC [43], [44]. The sub-micrometric and nanometric microstructural features have been shown to lower the thermal conductivity [45]. Furthermore, these fine structured coatings have also been shown to be better in thermal shock resistance, thermal-cyclic fatigue lifetime and a better CTE match [43], [46].

It is difficult to spray fine solid powder feedstock (20 nm to 5 µm) using a standard APS equipment. This is because, these fine particles do not have a good flow- ability and also can’t achieve enough momentum to penetrate the high velocity plasma stream [47]. Spraying fine powder particles may also be an environmental issue i.e. a health problem.

For both health reasons as well as to avoid particle agglomeration during storage and feeding into the spray equipment, a liquid feedstock can be used [48]. The liquid injection method can increase the momentum of the feedstock particles, aiding penetration of fine particles into the thermal jet core. Details about liquid feedstock spraying are provided in the next section. 2.3 Liquid feedstock spraying

In liquid feedstock spraying, fine powder particles are mixed with a fluid to form a liquid (suspension or solution), which then is injected into the plasma/flame either by atomization or by a liquid stream [49], [50], [50]. In Figure 4, instead of solid powder particles one can imagine a stream of liquid or atomized (break-up of liquid stream into smaller droplets) liquid droplets to be injected into the plasma plume.

Examples of thermal spray processes that use liquid feedstock are SPPS, S- HVOF, SPS and ASPS [35], [49]–[53], where a liquid in the form of suspension or solution, instead of direct solid powder, is used as a feedstock material.

2.3.1 Suspension and Solution

Figure 6: Schematic of a suspension and a solution precursor droplet

Suspensions are prepared by mixing solute powder particles of size a few microns, submicron or nano-meter in range in a solvent (water or alcohol (mainly ethanol)). A serious issue with suspensions is that they settle down with time, which if not properly re-dispersed before spraying, can influence the coating microstructure and properties. To overcome this problem, some additional elements can be added such as dispersant agents, which can prevent agglomeration of the particles or settling down of the suspension [49]. Proper stirring is required for the suspensions during the spray in order to minimize the settling problem. Also, pumping of suspensions from the storage tank to the plasma torch should be made in a way to generate as stable flow as possible. Any fluctuations may cause a non-homogenized coating structure.

Solutions are a homogeneous mixture of a solute (precursor material) and a solvent, and unlike suspensions no solid particles exist in it. In Figure 6, a droplet of suspension with many fine solid particles and a droplet of homogenized solution are shown. Typically the solutions are salts of respective ceramic powders, such as nitrates, chlorides, acetates, iso-prop oxides and other combinations [49].

BACKGROUND

2.3.2 Solution Precursor Plasma Spraying (SPPS)

A recent development in liquid feedstock spraying is the SPPS process, where a liquid in the form of a salt solution of a respective ceramic powder is injected into the plasma. The solution is fed into the plasma plume, where, the salt is oxidized in-flight forming an oxide particle. Under rapid heating, the solution droplet evaporates and the solid oxide particles in contact with the plasma, are melted and subsequently deposited onto the substrate [54]. Deposition of fine, melted particles then leads to a fine structured coating.

An advantage of SPPS over conventional APS technique in case of TBCs is that it can produce a strain tolerant vertically cracked microstructure with lower thermal conductivity than APS, due to vertical cracks (cracks normal to the substrate) and the submicron and nano-sized interconnected porosity [55]. Another advantage of SPPS is that new chemistries can be sprayed that may not be possible with other techniques due to unavailability of feedstock. However, a limitation of this technique is currently the deposition efficiency due to the difficulty in spraying high molarity solutions. 2.3.3 Suspension High Velocity Oxy Fuel (S-HVOF)

Another possibility in liquid feedstock spraying is the high velocity suspension flame spraying (HVSFS), which is a modification of the conventional HVOF thermal spraying process. This is also referred as S-HVOF spraying [52]. Here, the liquid is in the form of a suspension of the respective ceramic powder.

S-HVOF was developed with the aim of spraying submicron or nanoparticles suspensions with hypersonic speed to deposit thin and very dense coatings. Suspension injection in this process usually is done axially i.e. in same direction as the spray direction [35].

The process parameters which can influence the final coating structures in this process are different from plasma spraying. Examples of important parameters that can be controlled in order to obtain a desired microstructure are [35]:

 Suspension injection nozzle shape and geometry: In HVSFS, shape and exit nozzle diameter plays a crucial role as it decides the shape of the exiting suspension spray jet.  Combustion fuel type: Different fuels have different burning characteristics, which can influence the flame temperature and hence the particle temperature. Combustion fuels such as propane, ethane, acetylene etc. can be used.

 Total gas flow: Gas flow can influence the flame velocity and hence the particle velocity. Higher total gas flow rate can result in higher particle velocity.

Due to the high particle velocity involved in this process (typically around 400 m/s to 600 m/s [30]) the coatings are expected to be denser than the conventional APS and the SPPS coatings. 2.3.4 Suspension Plasma Spraying (SPS)

Another spraying process by which a liquid feedstock can be sprayed is SPS [49], [50], [56] which is also the process mostly used in this study. Like in S-HVOF, submicron or nano-sized powder particles are first dispersed in a solvent, typically water or alcohol, in a certain ratio (typically referred as solid loading) to form a liquid (in this case a suspension). This suspension then is injected into the plasma. The suspension is then undergoes atomization to form very fine suspension droplets, where each suspension droplet consists of extremely fine solid powder particles individually or in agglomerated form. The liquid from the droplet evaporates quickly and the fine powder particles can undergo sintering, partially or fully melting and may form agglomerates. Then, they get deposited with an impact on the substrate to form a fine structured coating [49].

The major drive for SPS in case of TBCs comes from the possibility to develop coatings with a unique strain tolerant columnar microstructure similar to EBPVD. SPS coatings have shown to have a similar columnar structure as of EBPVD, but with a lower thermal conductivity [57]. Columnar coatings with a lower thermal conductivity are demanding in gas turbine industry due to their inherent strain tolerant nature, which enhances the lifetime of the coating.

Even though, SPS has shown a great interest of research for producing TBCs for gas turbine applications, the complexity involved in the process as such makes it a very difficult process. This process can be described by understanding the detailed process characteristics, which are explained in detail in the next chapter.

3 Process characteristics of SPS

Similar to the conventional plasma spray process, coating microstructure and thus coating properties in SPS depend on various process parameters. In fact, in this case the process becomes even more complex due to the chemistry of the suspension and very fine powder particles involved. All process parameters schematically shown in Figure 5 are also important in SPS. However, some extra parameters, which are also important in this process and can have significant influence on the coating microstructure and properties, are discussed below. 3.1 Type of solvent

As explained in section 2.3.1, suspensions can either be water based, alcohol (mainly ethanol) based or a mixture of water and alcohol based. Both water and alcohol have some advantages over each other, but also have limitations. The water based suspension is cheaper and easy to transport/handle than the ethanol based suspension which requires more safety control. But, the heat required to vaporize ethanol is one-third of that of water, which leads to higher usage of power needed during spraying in case of water based suspensions [58]. Also, it was found that the deposition efficiency and the ratio of the coating mass to the powder mass sprayed towards the substrate doubled, when switching from a water-based YSZ suspension to an ethanol-based YSZ suspension [59]. Properties such as viscosity and surface tension of the suspension are also important and depend on the type of solvent used, which can play crucial role in SPS coating formation, as further discussed in 3.5 and 3.6. Water based suspensions has higher surface tension than ethanol based suspensions, due to the higher surface tension of water than ethanol [49], [60]. However, viscosity of the suspension is not significantly altered by changing the solvent but by changing the solid load (amount of powder in suspension) and solute particle size [49], [60], [61]. 3.2 Initial size of the solute powder particles

In conventional solid powder feedstock APS coatings, the scale of microstructural features present in a coating microstructure directly depends upon the initial powder particle size. However, unlike APS, initial powder particle size of the respective solute ceramic powder in a suspension may not necessarily be directly reflected on the microstructure and the microstructural features in SPS TCs [62].

Initial powder particle size is usually denoted as the median size or D50, which is the median diameter of a powder particle in a certain particle size distribution and it signifies that half of the particles lies below this size. As discussed earlier, typical fine powder particles used in SPS are shielded by the solvent fluid around them, which forms a suspension droplet as shown in Figure 6. Unlike in APS, upon injection in the plasma plume, it is this droplet which is in direct contact with the plasma and not the powder particles. This droplet consisting several of those fine powder particles undergoes a thermal treatment in the plasma, which is discussed in the section 3.5. Due to this thermal treatment the initial powder particle size may indirectly influence the microstructure, which is also discussed in section 3.5 (in-flight conditions).

3.3 Solid load in the suspension

Solid load is defined as the relative amount of solute suspended or mixed with the solvent (water, alcohol or mixture). Typically solid load in suspensions are measured in units of weight percent. For example 25 wt. % solid load refers to 25 grams of solute and 75 grams of solvent in 100 grams of suspension. Theoretically, higher the solid load in the suspension more the potential deposition rate of the SPS process [63] which is of course a major industrial drive. However, there are several challenges associated with the high solid load content in the suspension. For instance, the stability of the suspension (suspensions with higher solid loading may settle down quickly than lower solid loading); clogging of the gun injector/nozzle; sufficient power/energy required for proper melting (more power is needed to properly melt the highly loaded suspensions than low solid load suspension); but more importantly the challenge in droplet formation or atomization (discussed in more details in section 3.5) which is necessary for column formation in the columnar structured TBCs. All these challenges can lead to an undesired coating microstructure which must be countered in order to use highly solid loaded suspensions in SPS process.

3.4 Injection of suspension

Proper injection of a suspension is as important as other process parameters discussed so far to obtain the desired coating microstructure. This is because in SPS, droplet formation has a great influence in the coating formation which will be discussed in section 3.6.

Injection of the suspension can be influenced by following two factors:

1. Injector type and injector size

PROCESS CHARACTERISTICS OF SPS

a. Mechanical injectors b. Atomizing injectors 2. Injection type a. Radial injection b. Axial injection

Injectors can be of two types, either a mechanical injector or an atomizing injector. In a mechanical injector, which is also called a stream injector, a stream of suspension is injected into the plasma plume. In this case, the plasma plume itself performs the step of stream break-up into small droplets. However, in an atomizing injector, the suspension stream is first atomized (broke-up) into smaller droplets and then injected into the plasma plume. Apart from the type of injector, the size or the opening of the injector (orifice) can also significantly affect the droplet formation. Narrow orifice can shear the suspension stream more rigorously augmenting in fine droplet formation and can also provide higher droplet exit velocity into the plasma plume.

The liquid feedstock can be injected into the plasma plume either by radial injection or axial injection which are explained below. 3.4.1 Axial injection versus radial injection

The term ‘radial injection’ means injecting the feedstock material perpendicular to the plasma flow. ‘Axial injection’ means injecting the feedstock material in the same direction as the flow of plasma. This is illustrated schematically in Figure 7.

If the liquid feedstock material is injected radially, it may have a strong effect on the thermal treatment and the trajectory of the droplets during the in-flight. Different droplet sizes result in different momentum of the droplets when they enter the plasma stream and hence their trajectories will be different. For example, very small droplets may not penetrate the plasma at all or pass only through the periphery. Optimum sized droplets may enter the core of the plasma and get fully treated. However, larger droplets may completely pass through the plasma resulting on the substrates as un-molten particles (overspray) due to insufficient thermal treatment. Entering different zones of the plasma plume as well as different dwell times (the total time, a droplet spends in a plasma plume) lead to different heat treatment and acceleration of the droplets towards the substrate which leads to a more heterogeneous microstructure. This is explained schematically in Figure 7 (a).

Figure 7: Schematic of radial injection (a) versus axial injection (b) of liquid feedstock

Axial injection on the other hand can overcome this problem, as there is no issue of trajectory involved due to the parallel injection of feedstock material to the spray direction. As can be seen from Figure 7 (b) that, instead of an individual droplet trajectory, distribution of droplets undergo a uniform treatment during the in-flight conditions in the plasma. This work mostly explored an axial injection and hence more emphasize is given on axial injection from hereon. 3.5 Sequence of events in SPS coating formation

Assuming a stable flow of suspension upon injection in the plasma plume, suspension droplets may undergo various stages before impacting the substrate (see Figure 8). These are, injection of suspension, atomization of suspension, vaporization of solvent from the suspension droplet, agglomeration of the fine solid particles in the suspension droplet, melting of solid particles and finally impacting the substrate in the form of molten/semi-molten droplets to form fine splats leading to a coating build-up. The treatment of the suspension droplet during each stage will depend on the characteristics of the solute particles in the suspension droplet e.g. particle size (D50) and amount/mass (solid load) etc. as shown for three different cases i.e. medium particle size (I & II), very fine particle size (III) and coarse particle size (IV). All these different stages are explained in great detail below.

PROCESS CHARACTERISTICS OF SPS

Figure 8: Schematic of various stages of a suspension droplet (having varied particle

size (D50) i.e. medium (I & II), very fine (III) and coarse (IV)) during the in-flight, from injection of the suspension to the final deposition on the substrate Stage 1: Injection

Figure 9: Schematic illustration of an axial injection utilized in this work for axial III plasma system

The first stage is the injection of the suspension. As stated above, in radial injection the suspension droplets can be treated non-uniformly in the plasma depending on whether the suspension droplets can enter in the core of the plasma plume, which is extremely hot, or at the periphery which is colder [64]. However, there is no such problem in an axial injection as the suspension is injected co- axially with the plasma plume so that most of the suspension droplets are introduced into the core of the plasma. The axial injection of suspension as utilized in this work is schematically shown in Figure 9. An important aspect while axially injecting the suspension is the position of the injector exit orifice (see Figure 9) as it can significantly influence the suspension stream breakup which is the second stage as explained below in more detail. Stage 2: Atomization

The second stage is the atomization of the suspension, which occurs in two steps that is primary atomization and secondary atomization. Primary atomization is mainly caused by the atomizing gas which occurs just after the suspension stream comes out of the injector exit (see Figure 8 and Figure 9). As explained earlier in Stage 1, the position of the injector exit can also affect the primary atomization i.e. the primary atomization effect can be made stronger by retracting the injector as shown in Figure 9. Secondary atomization occurs due to the plasma drag once the suspension stream or suspension droplets come in contact with the plasma as shown in Figure 8. The final droplet size is the combination of both primary as well as secondary atomization effects. However, in high power axial injection system such as Mettech III axial gun as used in this work, the primary atomization effect can be considered not as significant as the secondary atomization effect which is due to the plasma. This is because unlike in conventional plasma spraying with radial injection, in an Axial III plasma torch the plasma plume forms very close to the injector exit, thereby negating the primary atomization effect. With this assumption i.e. neglecting the primary atomization effect, the focus from now onwards is mainly on secondary atomization effect.

The following factors can influence the secondary atomization:

1. Plasma drag force 2. Surface tension of the suspension 3. Viscosity of the suspension

Fazilleau et al. have derived a relation between these parameters and the atomized suspension droplet size [65], as shown below:

8 휎 퐷 = 2 (1) 퐶퐷 휌 푢

PROCESS CHARACTERISTICS OF SPS

Where D (m) is the suspension droplet diameter produced due to atomization, 휎 (N/m) is the surface tension of the suspension, 퐶퐷 (unit less) is the drag coefficient, whereas 휌 (kg/m3) and 푢 (m/s) are density and stream velocity of the plasma respectively.

From the above equation, it can be seen that, to obtain a smaller suspension droplet, a lower surface tension and a higher plasma drag force, which depends on plasma density and velocity, is needed. Also, the viscosity of a suspension, which is not discussed in this equation influences the atomization. Lower viscosity favours higher atomization of the suspension [60]. More detailed relation between atomized droplet size and various factors in general for aerodynamic breakup of a liquid droplet (in this case suspension) by a certain media (in this case plasma) is given by Wolfe et al. [66] as shown below:

1/3 1 3 136ɳ푑2휎2 퐷 = ( 1 ) (2) 4 2 2 푢 휌푠 휌푝 where D (m) is the final droplet diameter produced after secondary atomization, ɳ (Pa.s) is the viscosity of the suspension, d (m) is the initial diameter due to primary atomization, 휎 (N/m) is the surface tension of the suspension, 푢 (m/s) 3 3 is the stream velocity of the plasma, whereas 휌푠 (kg/m ) & 휌푝 (kg/m ) are the densities of the suspension and plasma respectively. The utility of this relation for SPS coatings is discussed by Curry et al. [62].

Plasma stream velocity and plasma density (and hence plasma drag force) can be controlled directly from the plasma gun hardware. Similarly, they can also be controlled by the plasma gas compositions, flow rate and the arc current. Surface tension depends on the solvent used. Water has much higher surface tension than ethanol, which then makes difficult to atomize the suspension. Viscosity, on the other hand can be modified by altering the suspension solid load and solute particle size [49], [60]. In general, the initial suspension droplet size (d) depends on the atomizing gas composition, atomizing gas flow, injector size, injection type (radial or axial) and also injector retraction distance if it is axial injection. Stage 3: Evaporation of the solvent

The atomized suspension droplets are rapidly heated up in the plasma plume and hence the solvent from the droplets vaporizes. Vaporization of the suspension during the in-flight stage undergoes with thermal energy absorption from the plasma plume which thus gets cooler. Cooler plasma may result in insufficient melting of the solid particles. Hence, suspension plasma spraying needs more

energy or power during spraying than conventional powder spraying. Also, the droplets are accelerated concurrently with the vaporization stage. Stage 4: Agglomeration of the solute particles

The solute particles (initial powder particles of certain D50 mixed with the solvent as discussed in 3.2) of respective ceramic powder used in the suspension are usually very fine. After the solvent evaporation, these fine particles have a tendency to undergo agglomeration or sintering. The extent of agglomeration can depend upon the size of the particles as can be seen from the Figure 8 in case IV where there are coarse powder particles the agglomeration stage is not valid. Whereas in case of extremely fine powder particle size as in case III the agglomeration can be larger. Anyway, such agglomeration may certainly lose the original characteristics of fine particles, since they are no more independent fine particles but agglomerated and often sintered or fused into larger particles. This is also one of the reason that the initial powder particle size (D50) may not directly influence the microstructure. However, larger initial particle size (D50) means that the atomized droplet (containing some agglomerated fine particles) contains larger solid particles’ mass, which undergoes melting in the next stage. Stage 5: Melting of the solute particles

Complete vaporization of solvent leads the agglomerated fine solid powder particles in direct contact with the plasma, thus they are heated up and melted and from molten/semi-molten droplets. The molten droplets may consists of few of the initial powder particles of certain D50, which were agglomerated in the previous stage and which may result in slightly larger solid mass in the molten droplet than the initial powder particle mass. This then explains, why the fine microstructural features may not necessarily be a direct result from the fine initial powder particles in the suspension started with. Instead, it is this molten droplet (of different size and mass (typically larger and heavier) than the initial powder particle), which decides the splat size and hence the other microstructural features in the coating. The degree of heating so as melting depends again on the overall power level of the spray process and also the plasma gas composition [67]. The heating of a solid particle/suspension droplet also depends on the spray distance, usually in SPS the spray distance is kept shorter than the conventional APS. This is because of the very fine size of the powder particles involved in SPS; which needs less time to be melted. The small size also means that the particles are cooled faster, hence a shorter spray distance is necessary. Longer spray distance in SPS thus can result in re-solidification of the molten droplets before they arrive at the substrate resulting in round spherical particles in the coating microstructure.

PROCESS CHARACTERISTICS OF SPS

Stage 6: Deposition of the molten droplets

In the final stage, the molten or semimolten agglomerated solid particles in the form of a molten droplets (larger and heavier than the initial powder particle) end up on the substrates with a heavy impact to generate fine splats (the region is highlighted in Figure 8 by red dotted lines close to the substrate). The way the molten droplet impact the substrate is presented in next section 3.6. Subsequent deposition of these splats then leads to a coating of a certain thickness depending on the spraying conditions and the application. 3.6 Coating formation in SPS

Figure 10: Schematic showing the shadowing effect on the substrate asperity of molten droplets with different sizes

Coating build up in suspension plasma sprayed coatings is completely different to that in conventional APS powder sprayed coatings. Coating formation in SPS is understood to be related to the generation of very fine suspension droplets due to the atomization of suspension after injection and resulting fine molten droplets [68]. The trajectory of the droplets smaller than 5 µm can be affected by plasma stream in the boundary layer close to the substrate and deposit at shallow angles on surface asperities leading to shadowing effect [68], which is shown schematically in Figure 10. This is because of the low momentum of these smaller droplets, which can be influenced by the drag of the plasma stream in the boundary layer close to the substrate. This can be seen in Figure 11 where,

Berghaus et al. simulated the influence of plasma direction changes on droplet/particle velocities [68]. Note that Figure 10 and Figure 11 are describing the region highlighted by red dotted lines close to the substrate in Figure 8.

Figure 11: Schematic showing the effect of plasma flow on a trajectory of the droplets just before the final impact as a function of droplet momentum, Adopted from Berghaus et al. [68]

Based on the above concepts VanEvery et al. has proposed three major possible types of coating microstructures in SPS [69]. If the suspension droplet size after the fragmentation is extremely small (< 1 µm), the shadowing effect is larger, which can generate a columnar type structure. Suspension droplets having higher size (> 1 µm) but still small enough (< 5 µm) to be affected by the plasma flow can form a structure with some porosity bands within the columns which can be termed as feathery columnar structure. Such a column formation is also demonstrated in Figure 12 where it can be clearly seen the columns building-up pass-by-pass on BC asperities due to the shadowing effect.

Finally, droplets with significantly larger size (> 5 µm) can have a direct impact on the substrate with a very little influence of a plasma drag. This may result in to a lamellar (similar to the conventional powder sprayed APS process) or vertically cracked structure. The mechanism of vertical crack formation in coatings produced by conventional APS process is known to be due to the cooling and shrinkage of the deposited splats which result in high tensile stresses and subsequent intra-splat cracking during cooling [53]. The presence of such cracks

PROCESS CHARACTERISTICS OF SPS can eventually form vertical cracks as the splats overlap subsequently during coating build-up. Similarly, increased tensile stress level can be relieved in SPS by forming vertical cracks as observed in this work during coating build-up [53].

Figure 12: Demonstration of column build-up on an as-sprayed bond coat

A detailed description about the microstructure evolution and coating formation in SPS is thoroughly discussed in Paper F appended at the end of this thesis.

4 Characteristics of TBC

The most important characteristics of a TBC are the microstructure, porosity, thermal conductivity and lifetime under cyclic thermal loads. The significance of all of them is discussed in detail in this chapter. 4.1 Microstructure

If there is something which can explain most about the final properties and performance of a thermal spray coating, then it is the coatings’ ‘microstructure’. Changes in coating microstructure result in changes in coating properties [10], [11], [70], thus it is important to understand its formation and how various microstructural features influence the coating performance. In this section a typical APS microstructure and a SPS coating microstructure are discussed. 4.1.1 Solid powder feedstock APS sprayed coatings

Typical solid feedstock APS coatings’ microstructure have a lamellar structure as shown in Figure 13. The coating is built-up by subsequent deposition of splats with sizes in the range of tens to hundreds of micrometer in diameter and few tens of micrometer in height, which finally form lamellar structured coating. The lamellar structured coating is heterogeneous and has lot of microstructural features such as fine pores, globular pores, cracks and delaminations of various sizes. They are important characteristic of the coating, since they strongly influence the coating properties.

Delaminations and cracks are formed in the coating due to the insufficient bonding of splats to the surface and cooling of splats respectively. Fine pores are formed due to the gas bubbles between splats or gas voids within powder particles. Globular pores on the other hand, are formed due to low temperature or velocity parameters resulting in partially molten particles and voids between particles [18].

All these features contribute to the overall porosity of the coating which influence the thermal insulation as well as lifetime of the coating. The role of these various microstructural features on thermal properties will be discussed in the later sections in detail. Mechanical and functional properties such as coatings’ stiffness,

thermal cyclic fatigue as well as thermal shock lifetime etc. are also influenced by these features and discussed in the later sections as well.

Figure 13: SEM micrograph of a cross-section of a typical APS powder sprayed lamellar TC microstructure highlighting the important microstructural features 4.1.2 Suspension plasma sprayed coatings

Suspension plasma spray, as explained in Chapter 2, can give a wide variety of coating microstructures. On a macro-scale these coatings exhibit different microstructural features such as: vertical cracks, spacing between columns (inter- columnar spacing (IC)), inter-pass porosity (IP) bands, branching cracks etc.; whereas at a micro-scale, coatings show features such as fine pores and cracks (interconnected or unconnected). Based on the microstructure at a macro-scale these coating structures can be mainly categorized as follows (Figure 14):

1. Vertically cracked structure 2. Highly porous structure 3. Columnar structure

CHARACTERISTICS OF TBC

Figure 14: SEM micrographs of a cross-section of the coating, showing the variety of microstructures of TC which can be produced by suspension plasma spraying. (Vertically cracked (a), Highly Porous (b) and Columnar (c))

Depending on a specific combination of numerous factors such as feedstock material, spray technique and spray parameters etc., which are discussed in Chapter 3, any of the particular structure or a structure similar to these can be obtained.

As explained above, these coatings also look different at micro-scale, as can be seen in Figure 15. It can be observed from the figure that the coating also has features in submicron and nano range. Various possible typically observed microstructural features in SPS polished and fractured microstructures are highlighted in Figure 15 and Figure 16 respectively, where they are indicated with arrows of different colours and notations as follows: columns (C in blue), inter- columnar spacing (IC in violet), spherical particles (S in orange), well-molten splats creating dense zones (black), fine cracks (FC in yellow), nano pores (NP in red), submicron pores (SP in green) and micron pores (MP in white).

Figure 15: SEM micrographs of a cross-section of a SPS TC within a column (a) and near inter-columnar spacing (b) showing various microstructural features: columns (C in blue), inter-columnar spacing (IC in violet), spherical particles (S in orange), fine cracks (FC in yellow), nano pores (NP in red), submicron pores (SP in green) and micron pores (MP in white) [Ganvir et al. Materials and Design, 2017]

CHARACTERISTICS OF TBC

Figure 16: SEM micrographs of a fractured top-view (a) and its magnified picture within a column (b) of a SPS TC showing various microstructural features: columns (C in blue), inter-columnar spacing (IC in violet), spherical particles (S in orange), fine cracks (FC in yellow), nano pores (NP in red), submicron pores (SP in green), micron pores (MP in white) and well molten splats (in black) [Ganvir et al. Materials and Design, 2017] 4.2 Porosity

Porosity is one of the key characteristics of the ceramic TC. In SPS coatings, several of the above mentioned different microstructural features, both at large

scale and small scale, contribute to the total porosity content of the coating. These pores are nothing but empty voids mostly filled with air at ambient conditions. Porosity can be open when the pores are interconnected (accessible by a fluid if injected externally) or closed when they are isolated from each other (non- accessible by a fluid from outside the coating). The definition of open porosity and closed porosity may vary depending on the usage of the porosity term in various applications, but in case of TBCs it is dealt with a fluid accessibility in pores.

 Open porosity: Vertical cracks, inter-columnar spacing, inter-pass porosity bands which are typically connected with the vertical cracks or inter-columnar spacing. Also, in some cases there can be branching cracks, cracks which originate at the vertical cracks or inter-columnar spacing and grow parallel to the substrate.  Closed porosity: Smaller and larger scaled globular pores which are independently present in the coating, independent clustered pores which are connected with each other but not with any of the features shown in the open porosity so that an external fluid cannot reach them. Also, in some cases, fine cracks which are present in the coating but not connected with any of the features shown in the open porosity.

Figure 17: Simplified schematic of typical suspension plasma sprayed TC microstructure, showing all possible features which can be counted in as total porosity

CHARACTERISTICS OF TBC

All features shown in Figure 17 can significantly affect both the thermal insulation nature as well as the lifetime of the coating. Vertical cracks or inter-columnar spacing are through the thickness of the TC. These cracks. if present in large number (higher vertical crack density in the coating), can increase the overall thermal conductivity of the coatings [70], [71]. However, branching cracks and inter-pass porosity bands are perpendicular (or close to perpendicular) to the direction of heat flow within the coating and hence can act as a significant thermal barrier within the coating [70], [72]. Other fine features such as cracks and pores also help in decreasing the overall thermal conductivity of the coating. This is because of the much lower thermal conductivity of these features (containing air) compared to the bulk material. 4.3 Thermal conductivity

Thermal insulation (defined by thermal conductivity) is an important functionality of a TC, which drives the overall performance of a TBC on a gas turbine. The lower the thermal conductivity of a TC, the better the thermal insulation. Hence, it is important to understand possible modes of heat transfer in a TC and henceforth to understand how to lower the thermal conductivity of a TBC. Since, the work was focused on 8YSZ TCs, in brief, different possible modes of heat transfer in 8YSZ TCs and their thermal conductivity, both for conventional and suspension plasma sprayed TCs are discussed. 4.3.1 Heat transfer in conventional APS & SPS TCs

Golosnoy et al. [73] explained several possible contributing modes of heat transfer in conventional APS 8YSZ TCs, which influences the overall thermal conductivity of the TBC. These are

 conduction through the solid YSZ,  conduction through the gases in pores,  radiative heat transfer, and  some contribution from convection if the segmented cracks are present.

At room temperature and pressure, conductive heat transfer through solid YSZ is largely predominant followed by conduction through the gases in pores [73]. Radiative heat transfer is significant only at higher temperatures (greater than 1000K) [73]. Convection plays partial role if ceramic TC have thick or wide and through vertical cracks or segmented cracks [73]. This is because these segmented cracks can allow the hot gases to pass through the coating easily hence increasing the overall thermal conductivity of the TBC [70], [71].

For bulk 8YSZ, the thermal conductivity at ambient conditions is ~2.25 (푊 푚−1퐾−1), which is due to the phonon conduction [74]. In reality, since the coating is not fully dense and consists of many microstructural features such as pores and cracks, the overall thermal conductivity is lower. This is because the microstructural features interrupt the phonon conduction by scattering at these microstructural features’ boundaries. Hence, more scattering interfaces are preferred in the form of microstructural features in TCs to get a lower thermal conductivity.

It was found that the pore size can have a significant effect on the pore conductivity [73]. Conduction through a typical gas in a defect such as pores can be significant, if the defect size is comparable to the mean free path (μ) of the gas molecules at atmospheric pressure (μ ~ 60 to 100 nm at ambient conditions) [73]. However, the μ is dependent on the temperature and pressure, where higher pressure can decrease the μ and higher temperature can increase it [74]. Moreover, the pore conductivity is nearly close to the free gas conductivity, if the pore size (d) is much larger than the mean free path (d > ~10 μ). If d is less than 1 µm its conductivity can be significantly below that of a free gas (air) [74]. In general, in conventional APS TCs the average size of pores is larger than 1 µm. Hence, typically at high gas pressures, similar to the gas turbine operating conditions, in conventional APS TCs, gas conduction can play a significant role and can increase the overall thermal conductivity of the TBC [74]. As explained earlier in this section, radiation and convection can be neglected in such coatings under certain conditions.

4.3.1.1 Heat transfer in SPS TCs

Similar to conventional APS TCs, all the modes of heat transfer, which contribute for thermal conductivity, are also valid for 8YSZ SPS TCs. However, because of the presence of different microstructural features, these coatings can conduct heat through both solid YSZ as well as gases in pores compare to APS coatings in a significantly different manner. As in conventional APS, radiation heat transfer can be neglected at ambient conditions but conduction through solid YSZ and gases in pores can have significant different influence on thermal conductivity due to the different microstructure of these coatings compared to conventional APS coatings. As presented in the sections 4.1 and 4.2, SPS coatings can have a microstructure with considerable higher content of very fine (<1 µm) submicron or nano-sized porosity.

Due to the presence of fine (<1 µm) submicron or nano-sized microstructural features in these coatings [75], [76], phonon scattering enhances, which then helps in lowering the overall thermal conductivity [73], [74]. Also, since the pores

CHARACTERISTICS OF TBC present in such coatings are in submicron or nano range, conductivity in the gas within the pores can also be significantly lower than in conventional APS coatings (see the discussion in the previous section) [73], [74]. However, presence of segmented cracks or inter-columnar spacing can increase the thermal conductivity, if they are large or wide enough to allow convection through them [70], [71], [73], [74]. 4.4 Mechanical properties

As shown in the previous section, the presence of various microstructural features such as pores and cracks are necessary for providing the required thermal insulation in a TBC. These features can also affect the coatings mechanical properties such hardness, E-modulus and fracture toughness [77]. E-modulus and toughness are considered as the two most crucial mechanical properties for TCs from the design point of view. In fact most of the studies on mechanical properties of conventional APS TCs were based on evaluating the E-modulus [78]–[80]. Similarly, in case of SPS TCs, the mechanical properties were mostly evaluated in terms of coatings’ E-modulus. However, the literature is limited on SPS in comparison to APS [81]. E-modulus is a measure of coatings’ compliance or strain tolerance whereas toughness is measure of coatings’ ability to resist crack formation and propagation. The lower the E-modulus, higher the strain tolerance and the higher the toughness, higher the crack initiation and propagation resistance. Hence from a design perspective, a coating with low E-modulus and high fracture toughness would be preferred.

High porosity is reported to be beneficial for reducing the E-modulus [82], however the toughness of the coating was found to deteriorate if coating possesses high porosity [82]. Also, different microstructural features in the ceramic TC microstructure can affect these properties in a different way. In fact, E-modulus is found to be more sensitive to the presence of various microstructural features due to its anisotropic nature [80] since different features in the ceramic TC microstructures can be present along different directions. For instance, the presence of delaminations or inter-splat cracks in conventional APS TC microstructure can lower the E-modulus in the spraying direction whereas the presence of vertical intra-splat micro cracks can result in lowering the in-plane E- modulus which is crucial during thermal cycling [80]. In fact, such understating was the motivation for developing the dense vertically cracked (DVC) APS TCs to lower the in-plane E-modulus which can improve the thermal cyclic performance by improving the coatings compliance as it is in the case of EBPVD columnar TCs [83]. This has further motivated the thermal spray community to develop DVC or columnar coatings by SPS process [84]. Higher porosity and

unmolten zones in SPS TCs than APS was found to lower the E-modulus whereas the presence of dense zones were found to increase the E-modulus [81], [85].

Similar to the E-modulus, the toughness of the TC can also be reduced due to the presence of delaminations or horizontal cracks in the APS sprayed conventional microstructures as the crack propagation along these features can be easier. Similarly, in SPS sprayed microstructures the presence of branching cracks, interpass-porosity bands or high porosity etc. can also substantially reduce the toughness due to easier crack propagation. As reported by Zhou et al., toughness was identified to be more important mechanical property than E-modulus for lifetime and has been considered as the crucial thermal cycling life-determining factor in case of SPS TCs [86]. 4.5 Sintering of ceramic TC in TBCs

The exposure of TBCs at high temperatures can lead to significant alterations in the 8YSZ ceramic TC microstructure due to sintering [87]–[90] of ceramic TC which can affect TBCs performance significantly [91], [92]. Sintering is a term most often used in the field of powder metallurgy which is defined as a process by which a metal/ceramic powder compact (loose powders) is transformed into dense solids at a temperature lower than the melting point. Loosely bonded compact powders contain significant porosity which is reduced due to sintering. Similar to the compact powders, ceramic TCs in the TBCs also contain porosity in the form of pores, delaminations and cracks, which at higher temperature can be altered due to sintering. The major driving forces for sintering in general are the curvature of the pore (surface energy), an externally applied pressure and or chemical reaction [93]. However, only the first two are reported to be the major driving forces to cause sintering related changes in 8YSZ ceramic TCs in TBCs [94], [95].

For conventional APS as well as EBPVD TCs, sintering has been reported to mainly decrease the porosity due to the closure of pores or healing of cracks (also referred as densification) [96]–[98]. Mechanism of densification is reported to be due to the combination of surface and grain boundary diffusion [89].

In addition to densification, coalescence of pores and or widening/opening of column gaps (also referred as pore coarsening) was also reported in case of EBPVD TCs [99], [100]. The mechanism of pore coarsening as explained by Exner et al. in case of several powder compacts during solid state sintering is caused by localized transport of atoms/molecules due to diffusion and/or bulk particle rearrangement [101]. Zotov et al. reported that the large pores formed after annealing in EBPVD TCs are due to the coalescence of initially (in as-

CHARACTERISTICS OF TBC produced state) isolated pores or by nucleation and growth during the recrystallization of the ceramic TC [99]. Widening of column gaps in case of EBPVD TBCs was reported to be occurring due to the thermal expansion mismatch between the TC and the substrate (especially when higher thermal expansion coefficient for substrate than the ceramic TC was noticed) [102].

Another important microstructural change is ‘grain growth’ reported in conventional APS and EBPVD TCs [98], [99], [103]. Grain growth is the term that describes the increase in the grain size of a solid which occurs in both dense and porous polycrystalline solids at higher temperatures [93]. Due to the conservation of matter, an increase in the average grain size is accompanied by the disappearance of smaller grains. The driving force for grain growth is the decrease in free energy that accompanies reduction in the total grain boundary surface area [93].

These sintering related microstructural changes such as densification, pore coarsening and grain growth can affect coatings’ thermal and mechanical properties. For instance it was shown in case of both APS as well as EBPVD TCs that reduction in porosity due to densification can increase the thermal conductivity significantly [5], [98], [104], [105]. Similarly, the mechanical properties such as hardness, E-modulus and fracture toughness were also found to be increased due to the reduction in the porosity because of densification [77], [92], [97], [99], [106].

Like conventional APS and EBPVD TCs, SPS sprayed TCs also showed similar sintering related microstructural changes (densification, pore coarsening and grain growth) [107] [108]. As a result, both thermal and mechanical properties of SPS TBCs are also reported to be affected. For instance, as reported by Zhao et al. and Guignard et al. the hardness and E-modulus were found to increase significantly as a function of temperature and time respectively [107] [108]. However, the literature is limited when it comes to the understanding of sintering and the evolution of properties due to sintering related changes in SPS TCs, unlike for/about conventional APS and EBPVD TCs. 4.6 Lifetime of TBCs

All the gas turbine components protected by TBCs experience severe thermo- mechanical loading due to the extreme operating conditions. The extent of the thermo-mechanical load can depend on: 1) the gas turbine application (for example military aero-engine or land based power generation IGTs or commercial civilian aero-engines) and 2) the type of component in the gas turbine (for example: combustor or turbine blades and vanes or after burner). For instance,

land based power generation IGTs are run over a long period of time (long running cycles) under isothermal conditions without frequent start-up and shut down of the engines; whereas the civilian aero-engines experience more frequent thermal cycling due to the frequent start-up and shut-down of an engine during the take-off and landing; and the situation can be extreme when it comes to the military aero-engines where the engines can experience most frequent start-up and shut down (very rigorous thermal cycling with severe thermal shock). The TBCs applied on the gas turbine components for any of the above mentioned applications can experience high stress loads which are relaxed by forming new cracks or propagating the pre-existing ones ultimately leading to the delamination and failure of the TBC. As mentioned earlier, TBC is a multilayer coating system containing BC, TC and the in-service grown TGO layer and all the three layers are considered to be equally important from the failure analysis and lifetime point of view [16], [17], [92], [109]–[113]. The significance of these layers is discussed in detail in the following sections. 4.6.1 Lifetime and failure associated with the BC and TGO

The significance of metallic BC during the lifetime of the TBC is considered as the most influencing factor where the failure in the TBC is typically associated with BC oxidation [109], [110], [113]. During TBCs’ exposure to high temperature operating environment, the metallic BC undergoes oxidation. This is because the 8YSZ ceramic material is transparent to oxygen under those conditions. Typically used BC material as mentioned in section 2.1.1 is MCrAlY alloy where M is Ni, Co or a mixture of Ni and Co. While MCrAlY allows formation of several oxides in the BC the most common, primary and stable oxide is alumina [114]. This is because alumina has the least formation energy among other possible oxides of constituent metals such as Ni, Co and Cr present in the BC [115]. In fact, formation of alumina is considered as beneficial for the TBC as it is dense and acts as a protective oxide layer for further oxygen diffusion and prevents formation of other detrimental mixed oxides [116] collectively named as CSN (Chromia, Spinel and Nickel oxide) [117]. The oxides formed at the interface of the bond coat-ceramic TC due to the oxidation BC are collectively referred as TGO as mentioned earlier in section 2.1.2.

MCrAlY bond coats in the as-sprayed conditions typically show a two phase matrix structure which as reported in the literature consists of β-(Ni(Co)Al) (beta phase) precipitates in a continuous γ-(Co/Ni/Cr) (gamma phase) matrix [17]. Beta phase is an aluminium rich phase which acts as an aluminium reservoir in the formation of α-alumina [17], [26]. After exposing these TBCs at high temperature, the aluminium starts to diffuse towards both the BC –TC interface as well as the BC-substrate interface consuming the aluminium reservoir (beta phase). After

CHARACTERISTICS OF TBC some time the beta phase depleted region close to both the interfaces become rather clear and this region is referred as ‘depletion zone’. These depletion zones are formed in the BC due to the incomplete diffusion of the aluminium from the BC.

After certain time during the lifetime of the TBC, the aluminium level within the BC is reduced due to its continuous depletion. Thermodynamically stable α- alumina slowly grows first as a TGO layer and continues to grow to a certain critical thickness. The α-alumina layer grown on a highly wavy BC profile cracks on cooling as a result of high local curvatures and thermal expansion mismatch between BC and TGO (α-alumina). These cracks can now provide an easy ingress of oxygen to the underlying BC. This may then further continue the α-alumina formation if there is still some aluminium left, otherwise start forming mixed oxides of other underlying BC elements which in this case are Ni, Co, Cr etc. It has been shown that when the aluminium is depleted to below 2.5 wt. % Al in the BC, no more formation of the protective α-alumina occurs [118].

Formation of other mixed oxides such as Ni/Co/Cr spinels [16], [17], [26] in the TGO is considered as detrimental for the TBC life due to their faster growth rate which leads to the rapid local volume change in TGO that generates excessive stresses [110]. It has been identified as impossible to completely eliminate the formation of such detrimental spinels [26], [119]. Spinels are reported to have lower fracture toughness than the alumina which then eases the cracking and crack propagation through TGO [113].

The major cause of failure in TBCs associated to BC oxidation is related to the stress state at the BC/TGO/ TC interfaces. These stresses are introduced in the TBC due to mismatch in CTE (mismatch stress) between different layers (BC- TGO-TC) and/or due to the growth of a new layer in the form of TGO in between BC and TC (growth stress) [118], [120]. These stresses are relaxed by forming cracks close to the BC-TGO-TC interface ultimately delaminating the ceramic TC and resulting in failure of the TBC.

Conventional plasma sprayed BCs are reported to be very rough (about 11 µm - 12 µm Ra [38]). BC roughness is crucial in a thermally sprayed coating as it provides mechanical anchoring of the TC layer on the underlying BC. The roughness however creates a wavy profile with hills and valleys on the BC. The TGO is typically formed following nearly the BC profile. As reported by Evans et al. [26] and Schilichting et al. [16], in case of conventional TBCs this wavy profile creates a different stress distribution along the profile in the TGO. Due to the wavy profile of the TGO geometry, out-of-plane stresses are produced resulting in tensile stresses at the hills and compressive stresses at the valleys [121]. These tensile stresses at the hills act as a driving force for the crack propagation

through the BC-TGO interface [121]. These cracks are formed because of the volume change near the TC-BC interface due to the formation of an extra layer (TGO). As the TGO layer is formed due to BC oxidation in this case, such a crack formation cannot be easily avoided.

The major influencing factors controlling the lifetime and failure associated with the BC and TGO are BC chemistry, BC roughness and TGO composition. In order to minimize the failure associated with BC and TGO and increase the lifetime of the TBC, BC with sufficiently high aluminium reservoir is necessary. In addition, smoother bond coats can also minimize the cracking issue at the BC- TGO interface and hence increase the lifetime of the TBC. In fact the relatively new techniques such as HVOF and HVAF were reported to produce smoother BCs than the conventional plasma sprayed BCs (in the range of about 3 µm - 8 µm [122]). Hence selection of both the BC chemistry as well as the BC spraying technique becomes more important while controlling the lifetime of the TBC. 4.6.2 Lifetime and failure associated with the ceramic TC

Significance of the ceramic TC in controlling lifetime of the TBC is as crucial as BC and TGO layers. Most common failure mode observed in the conventional APS as well as EBPVD TBCs is cracking in the ceramic TC close to the TGO- TC interface [113]. Such crack formations in the TC are also associated with the TGO formation. The crack initiates at the region where the curvature of the BC roughness profile changes from concave to convex as reported in case of conventional TBCs [123]. The interfacial fracture toughness of the TC/spinels interface is lower than that of the TC/α- Al2O3 interface which can increase the possibility of fracture between the TC and TGO [113]. Such failure can be minimized by increasing the fracture toughness of the TC at the TGO-TC interface. This was achieved by several researchers by adding a thin dense YSZ layer close to the TGO-TC interface [124], [125].

In several cases where TBC experiences a severe thermal gradient or thermal shock (for example in case of military aero-engine) the significance of ceramic TC becomes even more crucial in the failure of TBC. Failure associated with severe thermal gradients through the ceramic TC in TBCs is typically analysed by studying TBCs’ thermal shock behaviour. As reported in the literature, under thermal shock conditions, TBC experiences a severe thermal gradient across the ceramic TC layer compared to the isothermal cyclic conditions [126]. Higher thermal gradient across the ceramic TC layer can affect the BC temperature which in turn can affect the TGO formation and thermal mismatch stresses in the ceramic TC near the BC-TC interface affecting the lifetime [127].

CHARACTERISTICS OF TBC

The thermal gradient is of course dependent on coatings’ thermal conductivity which in turn depends on TC microstructure and porosity. Higher the thermal conductivity of the ceramic TC lower the thermal gradient across it, meaning the BC can experience higher temperature which as reported by Vassen et al. can result in lower lifetime [128]. In case of conventional APS TBCs, porosity and various microstructural features of the TC were found to have a significant role on their thermal shock behaviour. For example, the presence of inhomogeneous distribution of pre-existing micro cracks and pores were found to deteriorate the thermal shock life [129]. Whereas the presence of segmented vertical cracks and nanostructured features were found to improve the thermal shock life [129]– [131].

As mentioned in section 4.4, the mechanical properties of the ceramic TC are very much dependent on its microstructure and porosity which in turn can affect the lifetime. For instance, higher strain tolerance, which is a major characteristic of the columnar EBPVD as well DVC TCs, can accommodate the strain introduced in the TC during the frequent thermal cycling in both isothermal as well as thermal gradient conditions [132], [133]. Such an increased strain tolerance in those types of TBCs is reported to be beneficial for improving their thermal cyclic lifetime [131], [133]. In addition to the strain tolerance, the fracture toughness of the ceramic TC is also widely reported to be one of the crucial factors in determining the lifetime of the TBC [134], [135]. Higher fracture toughness is needed not only close to the TC-TGO interface but also throughout the ceramic TC, as cracks can initiate anywhere in the coating especially under sever thermal gradient conditions. For conventional TBCs, the influence of fracture toughness on TBCs’ lifetime is extensively studied where an improvement in the fracture toughness of the ceramic TC was shown to be beneficial for the lifetime [124], [134], [135]. Similarly, in SPS TBCs an improvement in fracture toughness of the ceramic TC also showed to enhance the lifetime [86], however unlike in conventional APS TBCs the literature on SPS TBCs is limited.

5 Experimental techniques used to characterize TBCs

Many characterization techniques are used to investigate TBCs [11], [12], [136]– [139], however, only those which are within the scope of this work are presented in this chapter. Figure 18 describes in brief different types of characterization techniques used in this work.

Figure 18: Different characterization techniques used in this work to study TBCs 5.1 Microstructure evaluation

Microstructure of the TBCs was studied in this work in general using scanning electron (SEM) and optical microscopy (OM). Phase analysis was performed using X-ray diffraction (XRD). In case of SPS coatings, due to the fine scaled features present in the microstructure, a high resolution SEM is necessary to get a thorough understanding of all the features.

Prior to the microstructure study in a microscope, it is very important to prepare the samples using proper metallographic techniques. Typically, sample

preparation of TBCs involves cutting, mounting in a low viscosity epoxy resin, grinding and polishing the samples manually or using semi-automated machine. The standard procedure for conventional TBCs is presented in detail elsewhere [136]. For SPS TBCs, the samples are first mounted in a low-viscosity epoxy resin prior to the steps involved in a standard procedure as mentioned above, to avoid any damage to the coating during cutting. Details can be found in any of the paper appended in this thesis. 5.2 Porosity measurement

Measuring the porosity in a SPS coating is a big challenge due to the presence of different length scaled porosities, starting from micrometer-sized vertical cracks/inter-columnar spacing to submicron and nano-sized pores and cracks. Hence, porosity analysis was done using three different techniques. Open porosity measurement was done using water impregnation (also known as Archimedean porosimetry) and mercury intrusion porosimetry (MIP). Also, image analysis (IA) was carried out to determine the content of both closed and open porosity in the coating.

In Archimedean porosimetry, a free standing ceramic coating is immersed in water and kept there under vacuum for certain time (around 5 minutes) to force the water into the pores. Both wet (immersed coating) weight and dry (as-sprayed coating) weight are measured by an electronic weighing machine. Weight difference of wet and dry coating is then calculated, which gives a measure of the water residue in the coating and hence the coating porosity. Detailed explanation of the experimental procedure can be found in the appended Papers A and Paper B.

Unlike water as in Archimedean porosimetry technique, mercury is intruded into the coating in MIP. Similar to water impregnation, a free standing ceramic coating is kept under vacuum in a capillary tube and the mercury is forced by using an external pressure, which can then fill the coating pores. MIP is based on the premise that a non-wetting liquid (one having a contact angle greater than 90°) will only intrude capillaries under pressure [140]. A detailed explanation of the MIP technique and the equipment is provided in Paper C and Paper H. The pore size distribution is determined from the volume intruded at each pressure increment; whereas the total porosity content is determined from the total volume intruded.

The conventional way of analysing porosity of TBCs is using IA on a determined number of SEM micrographs of coatings’ cross-section. The analysis involves converting a SEM greyscale image into a binary image (black and white) using any

EXPERIMENTAL TECHNIQUES USED TO CHARACTERIZE TBCS image analysis software (ImageJ, Aphelion etc.) with proper thresholding for determining the boundary between bulk and porosity. Finally, the fraction of white (pores) portion relative to the black (bulk material) portion of the binary image of a given SEM greyscale micrograph gives a porosity content in that particular image. Average of the porosity of all the images then gives the net porosity content of the coating.

Since SPS sprayed coatings showed porosity at varied length scale (micron, submicron, and nano pores); two or three different magnifications were utilized for analysis, e.g low (x500 or x1000), high (x 7000 or x10000) and sometimes even very high (x15000) in this work. These multi magnifications were selected such that low magnification should capture features like inter-columnar spacing, big vertical cracks, and micron-sized pores, whereas high and very high should capture features like submicron and nano-sized pores. While capturing the higher magnification micrographs, it was made sure that these micrographs should not contain any big features (which were already captured at lower resolution) in order to avoid repetition. Such a repetition was avoided by using a count mask in a software (i.e. the software will consider porosity under certain pore size range only). This technique is explained in detail in Paper C & Paper H. The image analysis technique adopted for this work is time-consuming but can provide an estimate of all types of pores (connected, non-connected, vertical cracks, branching cracks etc.) present in the coating. This is not possible using techniques like Archimedean porosimetry or MIP which can only measure open porosity. However, image analysis is still sensitive for sample preparation and may introduce errors while measuring fine pores. For this reason a combined porosity measurement technique that is coarse porosity measurement by image analysis and fine porosity measurement by MIP was adapted in this study. More details about such a procedure can be found in Paper H. 5.3 Evaluation of thermal conductivity

Thermal conductivity is an indicator of the amount of heat flow through a material (in this case a ceramic TC). The most common method to evaluate the thermal conductivity of a TC in TBC is to derive it from its diffusivity which can be measured experimentally. Thermal conductivity is calculated using the following equation:

휆 = 훼 휌 퐶푝 (3)

2 −1 −1 −1 where 훼 (푚 푠 ) is the thermal diffusivity, 퐶푝 (퐽 푘𝑔 퐾 ) is the specific heat capacity, 휌 ( 푘𝑔 푚−3) is the coating density and 휆 (푊 푚−1 퐾−1) is the thermal conductivity.

The most widely used experimental technique for thermal diffusivity measurement of a TC is the laser flash analysis (LFA) technique [141]–[143]. Several types of LFA equipment are available today so that thermal diffusivity can be measured either at room temperature or higher temperatures (up to 1500 ᵒC) as well as in ambient atmosphere and controlled atmosphere (Argon). The schematic of a typical laser flash equipment during the LFA experiment is shown in Figure 19.

During the laser flash experiment a laser pulse is fired at the rear face of the TBC sample (at the substrate). The heat pulse travels through the TBC sample and the resulting temperature increase at the surface of the ceramic TC is measured with an InSb infrared detector at a high sampling rate. The thermal diffusivity is calculated using the following equation [144]:

2 훼 = 0.1388퐿 /푡(0.5) (4)

Where 훼 (푚2 푠−1) is a thermal diffusivity, 퐿 (푚) is the thickness of the sample and

푡(0.5) (푠) is the time taken for the rear face temperature of the TC to reach half of its maximum rise. Detailed information about the origin and usage of equation 4 for diffusivity measurement in TBC can be found elsewhere [144]. An average of five laser shots is made for each measurement when the sample is at a steady state condition.

As zirconia is transparent to the laser wavelength used for measurements, the samples need to be coated with a layer of graphite or gold to prevent the laser pulse from traveling through the ceramic TC layer and allow energy from laser light to absorb [145]. The wavelength used in this work was 1064 nm (LFA 427, Netzsch Gerätebau GmbH, Germany), which is transparent to zirconia, hence, the samples were coated prior the diffusivity measurements.

Thickness measurement in this work was performed by optical microscopy by taking around 200 measurements all across the coating cross-section. Thickness for each layer in a multilayerd TBC system (퐿 in the equation 4) is an important factor for the accuracy of the results in a LFA experiment. It can be seen from the equation 4 that thickness has a square dependence on the thermal diffusivity. An error of 20 % in the total thickness (퐿) of the TBC can give almost 50% change in thermal conductivity [144]. It is important to realize that SPS deposited ceramic TCs show significant variation in the coating thickness due to their rough nature because of the presence of columnar features.

The specific heat capacity can be measured using differential scanning calorimetry (DSC). For this work, the specific heat capacity measurement was not performed,

EXPERIMENTAL TECHNIQUES USED TO CHARACTERIZE TBCS instead the values from existing databases on the same material (8YSZ) for specific heat capacity were used [15], [25], [25].

Figure 19: Schematic of the laser flash analysis equipment, note that the sample is placed in a specimen holder with a substrate facing the laser beam 5.4 Evaluation of mechanical properties

Hardness, Young’s modulus (E-modulus) and fracture toughness are the key mechanical properties of TCs in TBCs as mentioned in section 4.4. Several measurement techniques are employed in literature for measuring these properties of ceramic TCs in the TBC. For instance, E-modulus is reported to be measured by three or four-point bending [146], [147], indentation method [79] etc. Similarly, for fracture toughness measurement, several techniques such as double torsion method [82], double cantilever bending [148], indentation method [149] etc. are reported. Despite several techniques can be used there are no standardised techniques available for TBC characterization. Moreover, it has been shown that the absolute values are highly dependent on the techniques used to evaluate them and the results could vary even if the test specimens are identical between different techniques [150]. Hence, these properties, especially E-modulus (and hence toughness), need to be considered as an engineering value (used for design purpose) for quantifying and ranking of TBCs rather than a fundamental property. Indentation method was used in this work to measure the three

mechanical properties and hence only indentation method is explained in detail below.

Indentation method was employed in this work to measure hardness, E-modulus and also fracture toughness. Instrumented micro-indentation method was employed to measure hardness and E-modulus. The instrumented hardness (HIT) (convertible into Vickers Hardness HV) and instrumented E-modulus (EIT) (obtained from the unloading slope) values were determined by the commercial WIN-HCU® software (Helmut Fischer GmbH) using equations (5) and (6):

푃푚푎푥 푃푚푎푥 퐻퐼푇 = = 2 퐴푝 퐶표푛푠푡푎푛푡 × ℎ푐 (5)

1 − 휈2 퐸 = 퐼푇 푑ℎ 2√퐴 ( ) 푃 푑푃 1 − 휈2 (6) {( 푃푚푎푥 ) − ( 푖 )} √휋 퐸푖

2 where 푃푚푎푥 (N) is the maximum applied indentation load which was 1 N; 퐴푝 (mm ) is the projected (cross-sectional) area of contact between the indenter and the test specimen determined from the force-displacement curve and a knowledge of the area function of the indenter; ℎ푐 (mm) is the contact depth; 퐶표푛푠푡푎푛푡 is constant for a given indenter, here 24.5; 휈 is the Poisson’s ratio for 8YSZ, which was considered to be 0.25 [151]; 휈푖 and 퐸푖 are the Poisson’s ratio and E-modulus of the indenter, here for diamond 0.07 and 1140 GPa, respectively. A total of 25 indents were made on as-sprayed, polished cross-sections with a maximum load of 1 N (being equivalent to HV0.1) yielding average values of both hardness and E-modulus.

Cracks were generated for fracture toughness measurements using Vickers indentation technique. Cross-section polished TBCs were indented at a maximum load of 9.807 N (HV1) in the ceramic TC using a Vickers indenter for 13 seconds. The chosen load was identified as the optimum one for such coatings to produce Palmqvist type cracks without significantly damaging the coating. The following equation was used to calculate the fracture toughness, as the cracking pattern was identified as Palmqvist cracks (0.25≤l/a≤ 2.5) type in these coatings [152]:

2 1 퐸 5 1 푎 2 퐾퐼퐶 = 0.018 ( ) 퐻퐼푇푎2 ( ) (7) 퐻퐼푇 푙

1/2 Here 퐾퐼퐶 is the mode I indentation fracture toughness (MPa.m ), E is the elastic 2 modulus (MPa) of the coating determined from 퐸퐼푇 (퐸 = 퐸퐼푇 ∗ (1 − 휈 )), a is the indentation half-diagonal length (m) and l is the crack length (m).

EXPERIMENTAL TECHNIQUES USED TO CHARACTERIZE TBCS

Detailed procedure adapted for measuring the mechanical properties using the above mentioned indentation methods can also be found in Paper H. 5.5 Evaluation of sintering behaviour

In order to assess the sintering behaviour of TBCs isothermal heat treatment in an inert atmosphere was employed in this work. The reason for using an inert atmosphere was to avoid any cracking in the ceramic TC associated with the TGO formation. The oxidation of the BC and hence the TGO formation in an inert atmosphere can be minimized which can be helpful while studying sintering related changes in the ceramic TC microstructure. It must be borne in mind that in reality the operating conditions would not be in an inert atmosphere. However the purpose in this case was to study independently the TC microstructural changes due to sintering. This was only possible by using inert atmosphere to avoid the BC oxidation and TGO formation.

Thin square plate TBC samples were first water jet cut into ϕ10 mm coupons and subjected to isothermal heat treatment. Samples were placed into a furnace at room temperature, and heated to a temperature of 1100 ᵒC in about 4 hour. Prior to adding the samples, the furnace was evacuated and flushed with Argon/Hydrogen mixture with a flow rate of 200 L/hour. The heat treatment was done for 1, 3, 10 and 200 hours. Finally, all the samples were cooled down to room temperature in about 4 hours and removed from the furnace for further characterization. 5.6 Evaluation of thermal cyclic fatigue (TCF) lifetime

Lifetime assessment of TBCs is all time challenging, since it is hard to provide the realistic conditions as it is in-service on real components. However, accelerated and simplified tests are used to estimate the lifetime for comparative purposes [2], [132].

In TCF testing, samples are heated in a furnace (in air and normal atmospheric conditions) up to 1100 °C for a period of 1 hour followed by rapid cooling, using compressed air to approximately 100 °C within 10 minutes. This heating and cooling is termed as one cycle. After each cycle, a visual record of the samples’ surface is made and samples are repeatedly cycled in this fashion until failure. This allows crack evolution to be observed and cycles to failure to be estimated. The criterion for failure in TCF testing is when the TCs show more than 20% spallation of the whole coating surface. A more extensive description of this technique is provided in Paper A, Paper B and Paper G.

5.7 Evaluation of thermal-shock (TS) lifetime

In order to replicate the rapid heating and cooling experienced by the TBCs during civil or military jet engine operation, thermal shock (TS) tests are employed. Mostly such a thermal shock testing is done in a so called burner rig and the testing is called burner rig testing [153].

In burner rig testing, samples are heated to a set TC surface temperature (in this study it was about 1300 °C) using combustion burners for a set duration (in this study it was 75 seconds) and then forced cooled by compressed air to a set TC surface temperature (about 600 °C in this study) in a set duration (75 seconds in this study). Prior to the start of the cycling, the samples are preheated by hot air guns from the back side (from the substrate side) to a set substrate surface temperature (about 700 °C in this study) and maintained at that temperature throughout the cycling. This allows the TCs in TBC samples to experience a severe temperature gradient during each cycle as it is the case in reality. Hence, burner rig tests can be considered as the most relevant test to assess the failure and thermal shock behaviour of ceramic TCs simulating the conditions close to the real life jet engine operating conditions. However, there are still some important differences to real life conditions that must be considered. For instance, the burners are normally run on gas fuel such as propane and does not have the same characteristics as it is in the case of jet fuels. Also, most of the times the burner rig tests are performed at atmospheric conditions and subsonic speed which are also not close to the real life conditions.

More detailed description of the burner rig test to evaluate the thermal shock behaviour for the TBCs studied in this work is provided in Paper H.

6 Microstructure selection and coating design for TBC applications using SPS

Like all other engineering applications, no single microstructure type and coating system exists that can be suitable for all the TBC applications. Unlike conventional APS and EBPVD which are mainly capable of producing porous lamellar and dense columnar microstructures respectively, SPS has larger process window giving rise to various types of microstructures as shown in Figure 14 in section 4.1. In addition, the major advantage of SPS over both APS as well as EBPVD is that it can produce a porous columnar microstructure combining both the coating attributes which are provided by the above mentioned conventional techniques as shown in Figure 20. Despite of this unique advantage of SPS over APS and EBPVD, the large process window, especially in case of ASPS, makes the microstructure selection and tailoring process for a given TBC application more challenging.

Figure 20: Comparison of typical microstructures produced by EBPVD, APS and SPS i.e. dense columnar, porous lamellar and porous columnar respectively

This chapter provides a general thought process towards microstructure selection and coating design for different TBC applications in gas turbine industry. Moreover the chapter also discusses in detail a methodology for microstructure tailoring and designing an optimized columnar TBC with enhanced durability sprayed by SPS for both industrial as well as aero gas turbine applications. 6.1 Processing and microstructure

Figure 21 shows various possible routes which can be utilized to produce a specific type of microstructure as presented in section 4.1. The routes shown in Figure 21 summarizes the results from an extensive experimental study that was

performed with various liquid as well as solid feedstock thermal spray techniques and discussed in more detail in appended Papers (A, B, C and D).

Solution precursor plasma spraying (investigated in Paper A) and suspension high velocity oxy fuel spraying (investigated in Paper B) have produced TBCs with vertically cracked type microstructure. Conventional solid powder feedstock plasma spraying as expected has resulted in lamellar microstructure (investigated in Paper A). However, conventional solid powder feedstock plasma spraying in addition to its typical lamellar structure can also produce vertically cracked structure under certain spraying conditions, which has been shown elsewhere [133].

Suspension plasma spraying on the contrary has shown to be able to produce all three types microstructures, i.e. vertically cracked, highly porous and columnar. Investigations in Paper A and Paper B provided the first hint of a microstructure with features resembling like columns while using suspension plasma spraying. This was further investigated in Paper C and Paper D where axial suspension plasma spraying (ASPS) was utilized and it was found that it is possible to obtain, not only different types of columnar microstructures i.e. with different porosity, but also other types of microstructures as discussed in section 4.1.2 and shown in Figure 21 below using the same ASPS technique.

Figure 21: Flow chart showing the route of production of a specific type of microstructure for a TBC application using various feedstocks and spray processes

MICROSTRUCTURE SELECTION AND COATING DESIGN FOR TBC APPLICATIONS USING SPS

6.1.1 Influence of process parameters on TC microstructure

As discussed earlier in Chapter 2 & Chapter 3, the microstructure of coatings sprayed by both APS as well as SPS can be significantly influenced by several process parameters. A few of those parameters that can significantly affect the coating microstructure (and relevant to this work) while deposition by SPS process are listed in Figure 22. For simplicity, these process parameters can be grouped as shown in Figure 22 i.e. plasma spray parameters, suspension parameters, plasma gun hardware related parameters and others such as surface roughness of bond coat/substrate etc. These parameters and their influence on coating microstructure in case of SPS are discussed in detail in the appended papers. For instance, influence of plasma spray parameters (without changing other parameters) is discussed in Papers (B, C, D, G and H); whereas the influence of suspension parameters as well as surface roughness (without changing other parameters) on the microstructure is discussed in detail in Paper F. Influence of gun hardware (especially for Axial plasma gun) such as nozzle, injector size etc. was not discussed as it was not in the scope of this work. However, these can also significantly influence the microstructure and hence need to be investigated in future.

Figure 22: Flow chart listing major processing parameters that can significantly influence the microstructure produced by SPS process

The parameters mentioned in Figure 22 can be difficult to study individually as most of them have a strong interdependence among each other. Hence, in case of SPS, the coating formation (columnar or vertically cracked) as mentioned in Chapter 3 should be discussed in terms of droplet momentum which is dependent on most of the parameters mentioned above. Higher droplet momentum can result in normal deposition giving rise to lamellar and or vertically cracked type coatings whereas lower momentum can result in shadowing effect giving rise to columnar structures (see section 3.6).

Figure 23: Simplified flow chart showing the necessary process parameters and their target in the formation of columnar type microstructures and for vertically cracked type the target of these process parameters can be considered as vice-versa (note: there is an interdependence between droplet size and speed, however droplet size can be considered more significant)

The relationships between various important process parameters and droplet momentum and hence columnar microstructure type are given in Figure 23 where it is shown that low droplet size and droplet speed can produce columnar microstructures. To achieve a vertically cracked type microstructure, it can be considered as vice-versa i.e. high droplet size and high droplet speed. Again, the figure here is just a simplified representation of the relative magnitude of different process parameters needed for forming either columnar or vertically cracked microstructures using SPS. Also, due to the interdependence of droplet size and droplet speed, it’s more difficult to assess the individual role of different parameters as listed in Figure 23. However, the significance of droplet size is

MICROSTRUCTURE SELECTION AND COATING DESIGN FOR TBC APPLICATIONS USING SPS considered to be more important than the droplet speed due to the third-power momentum dependence of droplet size as shown in equation 8 below (see Paper F for more details).

4 퐷 3 M표푚푒푛푡푢푚 = ( 휋 (( ) )) 휌 푣 (8) 3 2 푠

Here, 퐷 is the droplet size after secondary atomization, 휌푠 is the suspension density and 푣 is the mean droplet velocity. Hence, while designing columnar structures using SPS process, it’s important to prioritize the fabrication of finer droplets. More detailed discussion regarding the influence of process parameters on microstructure (columnar or vertically cracked) is performed in the appended Papers (F and H). 6.1.2 Tailoring of porosity in a SPS TC microstructure

As mentioned in Chapter 2, TBCs are extensively used in the gas turbine industry for various high temperature applications. The performance-based criteria for ensuring superior performance of TBCs can differ, depending upon the different environments and exposure conditions experienced by the TBCs. This particular work has mainly focused on achieving high thermal insulation, high TCF and high TS life. However, several other criteria such as high CMAS (Calcia-Magnesia- Alumina-Silica or molten silicate), erosion and corrosion resistance could also be crucial under certain exposure conditions. For example, as explained in Chapter 2, high TS life is more important along with high thermal insulation for TBCs deposited on military aircrafts’ aero-engines. Similarly, for TBCs that are deposited on commercial/military aircrafts’ aero engines which regularly fly through desert areas will require in addition to high thermal insulation and high life, high erosion and CMAS resistance. Also, TBCs deposited on IGTs’ engines which typically encounter corrosive elements such as Sulphur due to their exposure to low grade fuels, require in addition to high thermal insulation and high TCF life, a superior high-temperature corrosion resistance.

The performance and durability of TBCs in any of the above-mentioned environments is intimately related to the porosity and pore structure/size. Therefore, tailoring of porosity in SPS microstructures to achieve a desired TBC is very important. The significance of various process parameters on total porosity content and pore size distribution in different SPS microstructures is very complex due to their interdependencies among each other. Therefore in practice and different studies, most of the process parameters are frozen (kept constant) while only one (or very few) is modified in order to see the effect of that parameter on coatings’ microstructure and porosity. The discussion on achieving different

porosity content and pore size distribution is performed in more detail in all the appended papers. However, a brief summary of such a porosity tailoring strategy as discussed in great details in Paper H is provided in this section.

A first step towards tailoring of porosity in such complex SPS microstructures would be to identify the major microstructural features that can contribute to coatings’ total porosity. For example, a simplified flow chart of such different microstructural features typically noticed in SPS coatings studied in this work is provided in Figure 24. It is very important to identify different features in the microstructures as they vary vastly in their size and accordingly they can influence not only the total porosity but also the coating properties and performance.

Figure 24: Summary of major microstructural features present in SPS coatings

The second step of porosity tailoring would be to select the pore classification, where the total porosity can be smartly distributed in different categories/groups based on the size or structure of these features. For example, porosity in SPS coatings in this work (see Paper G & H) is classified based on the size range of various microstructural features and shown in Figure 25. The figure shows the classification of porosity based on the equivalent pore size diameter (see equation 9) where porosity is categorized mainly in two major groups which are coarse porosity (>1 µm), and fine porosity (<1 µm).

퐸푞푢𝑖푣푎푙푒푛푡 푝표푟푒 푠𝑖푧푒 푑𝑖푎푚푒푡푒푟 = √(4/휋 (푎푐푡푢푎푙 푝표푟푒 푎푟푒푎)) (9)

These two major groups are further subcategorized in four different subgroups i.e. large (>10 µm), micron (10 µm - 1 µm), submicron (1 µm - 100 nm), and nano (<100 nm)) and are also referred as Class-1, Class-2, Class-3 and Class-4 respectively. Major contribution from different microstructural features towards

MICROSTRUCTURE SELECTION AND COATING DESIGN FOR TBC APPLICATIONS USING SPS each pore class is also shown in Figure 25. Such a breakdown of porosity in different classes for SPS coatings can be very helpful while systematically tailoring the porosity as per the requirement for various TBC applications.

Figure 25: Flow chart showing the classification of total porosity and the contribution of various SPS microstructural features present at different length scales in different pore classes

In the third step, after classification of the porosity in different classes, a target for each porosity class should be set, for example negligible/low/optimum/high, based on the desired TBC response. In the fourth step, various sources/origins for each porosity class should be carefully identified from the different microstructural features as listed in step 1. For instance, as shown in Figure 26 the main sources of Class-1 porosity as noted in this study are the inter-columnar spacing and region in the vicinity of the inter-columnar spacing. In the fifth step, the target for each of these sources should be set in a way so that the desired target for the respective porosity class can be achieved. For instance, in order to achieve for example, an optimum Class-1 porosity, either an optimum inter- columnar spacing and a negligible vicinity of the inter-columnar spacing region can be targeted or vice-versa. After that, in the sixth step, all the major control factors (i.e. process parameters such as plasma power, feed rate, suspension viscosity, solid load etc.) which can significantly influence each of the source/origin should be listed. For instance, inter-columnar spacing can be affected by modifying the BC roughness as it can significantly alter the column density and hence the total number of inter-columnar spacing in the microstructure. After that, in the seventh step, with the help of systematic

experimental design (DOE), a “calibration” study to achieve the desired numbers for each of these control factors should be performed. After the calibration study, in the eighth and final step, it should be verified whether or not the desired target for each porosity class is achieved. If yes, then the tailoring of porosity is finished but if not then the “calibration” study (i.e. seventh step) should be repeated until the desired target for each porosity class is achieved.

Figure 26: Origin/source of Class-1 porosity (i.e. inter-columnar spacing & vicinity of the inter-columnar spacing) shown in case of columnar SPS microstructure [Ganvir et al. Materials and Design, 2017]

All these eight key steps involved in tailoring of porosity in SPS microstructures to achieve a desired TBC response are also schematically shown in the form of a flow chart in Figure 27. A detailed discussion regarding tailoring of porosity as in case of achieving higher TS life and high thermal insulation which is used as an example to describe the above mentioned tailoring strategy has been successfully demonstrated in detail in Paper H. A philosophy similar to that described in Paper H can also potentially be used to tailor the porosity for all other TBC applications.

MICROSTRUCTURE SELECTION AND COATING DESIGN FOR TBC APPLICATIONS USING SPS

Figure 27: A flow chart showing systematic steps to be followed for tailoring the porosity in SPS microstructures 6.2 Effect of SPS TC microstructure on thermal conductivity

As explained in Chapter 4 and the previous section, the presence of various microstructural features observed in SPS microstructures can significantly affect the porosity at different length scale and hence the total porosity. The total porosity content of the coating has a significant impact on thermal conductivity as explained in Chapter 4 and shown in Figure 28 (a) (more details can be found in appended Papers (A, B, C, D and E)). These features and their formation are mostly governed by the different deposition/spraying conditions which is investigated in more details in Papers (B, C, D and H).

Figure 28: A relationship between the porosity and the thermal conductivity in SPS TBCs (a) [Ganvir et al. JTST, 2015] and the comparison of total porosity content in SPS and APS TBCs (b) [Ganvir et al. IJACT, 2015]

The major reason for achieving extremely lower thermal conductivity in SPS TBC compared to the APS as well as EBPVD TBCs (see Figure 29) is not only the presence of extremely high total porosity content (see Figure 28 (b)), but also the presence of various fine features contributing to the relatively higher (about 50% of total porosity) fine porosity content (i.e. porosity <1 µm) as shown in Figure 31 (a).

Figure 29: Comparison of thermal conductivities of SPS and ASPS TBCs with conventional APS as well as EBPVD TBCs [Ganvir et al. IJACT, 2015 & Ganvir et al. JTST, 2015]

MICROSTRUCTURE SELECTION AND COATING DESIGN FOR TBC APPLICATIONS USING SPS

In addition to these fine features, coarse features such as vertical cracks, inter- columnar spacing, inter-pass porosity bands etc. also have a significant impact on the thermal conductivity. Presence of all these features at different length scales span the porosity over a wide size range in a SPS microstructure and depending on their distribution the resultant thermal conductivity can vary. This is the reason why the microstructures produced by SPS (more specifically by ASPS) show varying thermal conductivity (ranging from as low as 0.30 W/(m.K) to as high as 0.76 W/(m.K) or even higher) as shown in Figure 29.

The individual significance of some of the microstructural features in a typical SPS microstructure is shown in Figure 30.

Figure 30: Schematic of a trade-off between the density of various microstructural features present in the coating and the thermal conductivity of the coating

It can be seen from the figure that presence of higher fine porosity and interpass porosity bands or branching cracks give lower thermal conductivity (investigated in Papers (B and C)), whereas higher number of vertical cracks and inter-columnar spacing can increase the thermal conductivity (investigated in Paper C). Moreover, the effect of vertical cracks or inter-columnar spacing can be neglected if the total porosity content is increased by increasing the fine porosity in the microstructure (investigated in Paper D). More details about the influence of different

microstructural features in SPS TBCs on thermal conductivity can be found in the appended Papers (B, C, D, E and H).

As explained in earlier sections 4.2 and 4.3, vertical cracks, inter-columnar spacing, if very large or wide, may ease the passage of hot gases to flow through them, which can increase the thermal conductivity of the coating. However, inter- pass porosity bands are almost perpendicular to the direction of heat flow, so they can help in decreasing the thermal conductivity. Other fine scale pores also help in reducing the thermal conductivity due to the enhanced phonon scattering interfaces in the microstructure.

For a TBC application, lower thermal conductivity is needed which then can be achieved by tailoring the porosity as suggested in section 6.1. 6.3 Sintering behaviour of SPS TCs in TBCs

As presented in section 4.5, sintering is an important issue in TBCs and it can affect coating properties and hence performance. This becomes even more complicated in case of SPS coatings due to the varied pore size distribution. For this reason porosity at different length scale is investigated in a systematic way to understand their response individually when exposed to high temperature. As shown in Figure 31, after 200 hours of isothermal heat treatment at 1100 ᵒC the porosity at different length scale in a typical columnar SPS microstructure changes in a significantly different manner.

Figure 31: Relative average porosity distribution of different pore classes in a typical SPS TC microstructure (a) and change of average porosity at different length scale after 200 hours of heat treatment at 1100 ᵒC (b) [Ganvir et al. Materials and Design, 2017]

MICROSTRUCTURE SELECTION AND COATING DESIGN FOR TBC APPLICATIONS USING SPS

For instance, Class-1 and Class-4 porosity show a significant increase and decrease respectively after the heat treatment compared to the as-sprayed porosity suggesting both pore coarsening and densification occurring in SPS microstructures. More details regarding this study can be found in Paper H.

All possible sintering related microstructural changes in a typical SPS microstructure are summarized in Figure 32, which is based on the results explained in the appended Papers (D, G and H). Moreover, the corresponding pore classes that are affected by these changes are also listed in Figure 32.

Figure 32: Flow chart highlighting the major sintering related changes that are typically observed in SPS coatings and corresponding different pore classes affected by those changes

Figure 32 shows that Class-1 and Class-2 pores are most susceptible to the pore coarsening, especially Class-1 porosity which increases mainly due to the widening of inter-columnar spacing during heat treatment, whereas the very fine pores i.e. Class-4 porosity are more prone to undergo densification. As it can be seen from Figure 31, unlike Class-1 and Class-4 porosity, the changes in Class-2 and Class- 3 porosity are difficult to assess, i.e. whether the change is increasing or decreasing or no change. However, in general, based on the results as discussed in appended Papers (D, E, G and H) it can be said that the Class-3 pores are most stable showing a very little change after the heat treatment whereas the Class-2 pores show a mixed behaviour i.e. both increase as well as decrease but mostly increase.

Such a knowledge regarding the individual response of different pore classes at high temperature can assist in tailoring of porosity. This can ultimately help in

designing the optimized SPS coating microstructure for different TBC applications. 6.4 Designing SPS TBCs for superior TCF performance

Several factors summarized from the experimental work carried out during this study (see Paper G) which can help to design an improved TCF performance SPS TBC are listed in Figure 33. It should be borne in mind that there can be even more factors; however only those are listed which were studied in this work.

Figure 33: Summary of several important factors needed for designing high TCF performance columnar SPS TBC [Ganvir et al. Ceramics International, 2017]

Figure 33 suggests that all the three layers i.e. TC, BC and TGO are crucial and should be considered while designing the SPS TBC for superior TCF performance. Factors such as surface roughness and chemistry of the BC, the TGO composition, and the nature of TC microstructure, TC phases and TC sintering behaviour were identified as to be the most crucial. Figure 33 suggests that a BC with lower surface roughness and higher beta phase content; a TGO layer with lower spinels; and a TC having a columnar structure with high column density and high total porosity with higher submicron/Class-3 pores can result in a superior TCF performance SPS TBC. Based on such an understanding, a simplified schematic of a recommended along with a non-recommended SPS TBC microstructure design for achieving superior TCF performance is proposed in Figure 34. The detailed analysis of all these factors and the proposed design can be found in Paper G.

MICROSTRUCTURE SELECTION AND COATING DESIGN FOR TBC APPLICATIONS USING SPS

Figure 34: Simplified schematic of a design of two different SPS TBCs; (a) not recommended and (b) recommended for superior TCF performance [Ganvir et al. Ceramics International, 2017] 6.5 Designing SPS TBCs for superior TS performance

Similar to TCF, several factors which can help to design an improved thermal shock (TS) performance SPS TBC are listed in Figure 35.

Figure 35: Summary of important factors needed in tailoring of microstructure and porosity for designing high thermal shock performance SPS TBC [Ref: Ganvir et al. Materials and Design, 2017]

The figure in this case only lists factors related to the ceramic TC microstructure whereas the parameters related to the BC and TGO can be selected as suggested in Figure 33. It should be noted that there could be even more factors influencing the TS however only those are listed which were studied in this work (as discussed in Paper H). Figure 35 also outlines the microstructure tailoring methodology to achieve higher TS lifetime and optimum thermal conductivity in a columnar SPS TBC as pointed out in Figure 27 in section 6.1.2.

As it can be seen from Figure 35, for improved TS performance with an optimum thermal conductivity, the desired target for various sources/origins of porosity in a columnar microstructure to achieve a desired pore class distribution (i.e. optimum Class-1 and Class-3 porosity and negligible Class-2 and Class-4 porosity) can be as follows:

 Optimum column density  Low column gap width  Optimum inter-columnar spacing  Negligible pores in the vicinity of the inter-columnar spacing  Negligible partially unmolten or improper splats  Negligible inter-pass porosity bands  Optimum inter-splat porosity  Negligible spherical particles and  Negligible intra-splat porosity

Based on such an understanding based design philosophy, a simplified schematic of the SPS TC microstructure design, capable of achieving high TS lifetime without significantly compromising the thermal conductivity is proposed in Figure 36. The proposed microstructure design tends to be like EBPVD, except the presence of some porosity inside the columns in the form of Class-3 pores. Of course, more efforts in the SPS process optimization are certainly needed in order to achieve the proposed columnar SPS TBC design for superior TS performance.

MICROSTRUCTURE SELECTION AND COATING DESIGN FOR TBC APPLICATIONS USING SPS

Figure 36: A simplified schematic of the design of SPS sprayed TC microstructure to achieve higher thermal shock performance [Ref: Ganvir et al. Materials and Design, 2017]

7 Conclusions

This research work has explored the relationships between process parameters, coating microstructure, thermal conductivity, thermal cyclic fatigue and thermal shock lifetime in suspension plasma sprayed TBCs. The following conclusions are drawn from this work correspondingly with the research questions stated earlier in Chapter 1.

The liquid feedstock technology was shown to provide large opportunities in creating different types of microstructures by varying different spray techniques and process parameters. Predominantly, three different types of microstructures were obtained using the various liquid feedstock sprayed techniques. The types were categorized as highly porous, vertically cracked and columnar. The three types of microstructures were shown to give significantly different thermal properties. Axial suspension plasma spraying was shown as a promising technique to produce various microstructures including all the three types, as well as low thermal conductivity coatings.

Examples of microstructural features that were observed in the coatings are columns, inter-columnar spacing, spherical particles, well molten splats (dense zones), vertical cracks, fine cracks, branching cracks, inter-pass porosity bands and pores (micron, submicron and nano sized). It was found that all of these features influence the total porosity content in the coating, which in turn was shown to have a significant influence on the thermal and mechanical properties as well as the thermal cyclic and thermal shock lifetime of the coatings.

Porosity was categorized as open porosity, i.e. pores which are accessible to an external fluid, and closed porosity. Image analysis on SEM micrographs has shown to be capable of estimating both closed and open porosity and to provide a fair estimate of the total porosity if two different magnifications are used. Image analysis seems thus to provide the best estimate of the total porosity compared to the two other techniques investigated, i.e. water impregnation and MIP, which only could estimate the open porosity.

However, image analysis has also shown some limitations especially while measuring the fine porosity. For this reason the combination of two techniques i.e. image analysis and MIP for measuring coarse (>1 µm) and fine (<1 µm) porosity respectively was utilized. The results showed excellent correlation between the total porosity measured in such a way with the coating properties as well as performance. This has also helped in classifying the porosity based on the

equivalent pore size in four different classes that is Class-1 (>10 µm), Class-2 (10 µm – 1 µm), Class-3 (1 µm – 100 nm) and Class-4 (<100 nm). Such classification of pores has helped in tailoring the porosity in order to optimize coatings’ performance.

It was found that suspension plasma spraying enables an increase of the total porosity content in the coating compared to solid feedstock plasma spraying. Higher total porosity content and presence of fine pores and grains seem as the most probable reason to the significantly reduced thermal conductivity. Moreover, it was found that suspension plasma spraying enables a highly porous as well as dense columnar structured coatings which are shown to be promising candidates to provide higher thermal cyclic lifetime and higher thermal shock lifetime respectively.

The individual influence of different porosity classes on thermal shock lifetime was found to be related to their respective stress intensity factors. Class-1 porosity was found to have the most severe effect on thermal shock lifetime than other three pore classes due to its significantly higher stress-intensity factor. In terms of coatings’ thermo-mechanical properties, fracture toughness was found to be the most prominent thermal shock life determining factor. The thermal shock lifetime was found to be increased with the increase in fracture toughness.

Overall this work concludes that suspension plasma spraying and specifically axial suspension plasma spraying is a very promising technique to produce high performance coatings for gas turbine applications.

8 Future work

This work was dedicated towards developing high performance TBCs using SPS. However, no comparison was done with the industry standard conventional TBCs produced by APS as well as EBPVD. Hence, as a first step in continuation of this work can be a systematic comparative study between the best performing SPS TBC resulted from this work and the industry standard APS and EBPVD TBCs.

High temperature sintering related changes in the microstructures were investigated in this work but not in great details. More fundamental investigations of high temperature sintering related changes in the microstructures such as densification, pore coarsening and grain growth could be the next step in extending this work.

Modifying water based suspension properties using several surfactants to reduce their surface tension without significantly modifying the viscosity can lead to columnar type coatings, as noticed in this work. However, due to the lack of time no more experimental trials were conducted. This could be of specific interest especially for the industry as water based suspensions sprayed columnar coatings can replace the ethanol based columnar coatings reducing the safety issue possessed by ethanol handling.

Mechanical properties were measured using indentation technique but more techniques can be explored in order to strengthen the claims made in this work. In addition, properties such as erosion and corrosion resistance is also important characteristics of a TBC, which could also be investigated in future. Moreover, more fundamental assessment of failure mechanisms associated with the failure of TBCs studied in this work during thermal cyclic fatigue as well as thermal shock testing need to be investigated in future.

The work also showed that, axial suspension plasma spraying can obtain extremely porous as well as dense coatings, which can be utilized for solid oxide fuel cell applications; where a porous anode and dense electrolyte is needed. In addition, the possibility to deposit extremely dense coatings can also be explored for wear applications. Moreover, this work also showed that a very thin and dense layer can be deposited using ASPS which makes this technique an exciting possibility in the field of bio-coating application as well. Hence, these can also be the possible routes to extend this research work in future.

9 Summary of appended publications

Paper A- Characterization of thermal barrier coatings produced by various thermal spray techniques using solid powder, suspension and solution precursor feedstock material The paper was a first attempt in the extensive screening study, which was performed in this thesis work to achieve a microstructure for the TBC application with a lower thermal conductivity using liquid feedstock. Existing liquid feedstocks (Suspension and Solution) were chosen, which were sprayed using plasma spraying along with conventional solid powder feedstock. A comparative study was performed on SPS, SPPS and APS sprayed TBCs; which were sprayed using 8YSZ suspension, solution and powder feedstocks respectively. It was found that a TBC with a microstructure resembling columnar type (inherently strain tolerant structure) can be produced using SPS, with a much reduced thermal conductivity, about half of that of the conventional APS TBC. Apart from this, it was also found that vertically cracked microstructures can be produced using SPPS. However, both SPS and SPPS coatings had a shorter lifetime than a conventional APS coating.

Paper B- Comparative study of suspension plasma sprayed and suspension high velocity oxy-fuel sprayed YSZ thermal barrier coatings The second paper was a continuation of the first study where suspension was found to be the most feasible feedstock. Hence, this paper focused on achieving different microstructures using suspension as a feedstock, and different spraying techniques. A comparative study was performed on 8YSZ suspension sprayed TBCs, with S-HVOF and SPS. Suspension was injected axially and radially in S- HVOF and SPS respectively. It was found that two different types of microstructures namely, vertically cracked and columnar type resembling structure can be predominantly achieved using S-HVOF and SPS techniques respectively. SPS resulted in highly porous coatings with extremely low thermal conductivity and relatively high thermal cyclic fatigue lifetime; whereas S-HVOF resulted in comparatively dense coatings with higher thermal conductivity and lower TCF lifetime. Apart from this, significant influence of initial powder size in the suspension (D50) was also noticed on coating porosity and its thermal properties. Reduction in D50 resulted in higher porosity and lower thermal conductivity in both SPS and S-HVOF coatings.

Paper C- Characterization of microstructure and thermal properties of YSZ coatings obtained by axial suspension plasma spraying (ASPS) Although both techniques i.e. S-HVOF and SPS, explored in the study presented in Paper B showed several advantages, they also showed some limitations. S- HVOF had the limitation of achieving the desired high porosity due to the high particle velocity involved, and radial injection SPS showed overspray problems due to the radial injection. In an attempt to overcome these problems, it was decided to use ASPS. A screening study was performed in order to explore various possible microstructures which can be obtained by ASPS. It was found that ASPS can generate predominantly three different types of microstructures which are highly porous, vertically cracked and columnar type. These types were discussed in this paper in detail.

A further aim of this study was to investigate the relationships between the different microstructures and the thermal properties of the obtained coatings. It was found that ASPS can be a potential technique to produce low thermal conductivity TBCs, lower than the conventional APS and EBPVD TBCs. All three different types (highly porous, columnar (feathery columnar) and vertically cracked) discussed in this paper, showed varied scale porosity in a micron, submicron and nano-sized scale. Higher porosity, lower column density and higher inter-porosity bands present in the coating resulted in lower thermal diffusivity and conductivity. Image analysis was also shown as a promising technique to estimate the total porosity in the coatings.

Paper D- Influence of microstructure on thermal properties of axial suspension plasma sprayed YSZ thermal barrier coatings This paper continued the study presented in Paper C where the ASPS technique was explored to achieve various microstructures. This paper focused on discussing only columnar type microstructures and their influence on thermal properties. It was shown that the ASPS technique can produce a columnar structured TBC similar to EBPVD coatings; but with lower thermal conductivity due to the high amount of fine scaled microstructural features present in the coating. Higher total porosity content resulted in lower thermal diffusivity and thermal conductivity. No clear relationship between the column density and the thermal properties was found. Only the total porosity was shown to have a significant effect. Higher spray distance and surface speed along with other process parameters resulted in lower crystallite size, higher total porosity and higher column density which resulted in lower thermal diffusivity and thermal conductivity of the coating.

SUMMARY OF APPENDED PUBLICATIONS

Paper E- Influence of isothermal heat treatment on porosity and crystallite size in axial suspension plasma sprayed thermal barrier coatings for gas turbine applications During the first half of this work (investigated in Paper A, B, C and D) the thermal conductivity was found to be affected by the sintering related changes in the microstructure. Hence, as first step towards the second half of this PhD work, all the columnar coatings studied in paper D were isothermally heat treated at 1150 ᵒC in an inert atmosphere. The effect of heat treatment on columnar microstructures was studied in this work. Several sintering related microstructural changes such as pore densification, crystallite size growth and pore coarsening were noticed. The effect of these sintering related changes along with the effect of total porosity on thermal conductivity was also studied. In addition, the two magnification approach (as suggested in Paper C but not fully used) was successfully utilized to measure the total porosity.

Total porosity in 8YSZ ASPS sprayed ceramic top coats was found to be not only decreasing for some coatings as expected in a conventional APS TBCs but also increasing in some other cases after heat treatment at 1150 ᵒC for 200 h in Argon. The increase in porosity was attributed to pore coarsening, i.e., of coarser pores (at about larger than 1 µm^2), more specifically due to the widening of the inter- columnar spacing and coalescence of pores. It was also found that it is a trade-off between the three microstructural changes (coarsening, densification and crystallite size growth) that can decide the overall trend in thermal conductivity of the respective coating. If the pore coarsening dominated the densification and crystallite size growth then the overall coating thermal conductivity was observed to be decreased significantly.

Paper F- Experimental visualization of microstructure evolution during suspension plasma spraying of thermal barrier coatings In all the papers related to the SPS coatings during this PhD work, different coatings were sprayed using varied plasma spray parameters while using the same suspension (no change in suspension properties). However, this is the only paper where several coatings were sprayed with different microstructures using different suspensions but using same plasma spray parameters. This was an attempt to systematically identify various crucial suspension properties and their role in coating formation and microstructure evolution in SPS. Six different suspensions were prepared with different solid load content, solvent types and particle size distribution. The effect of all these three variables (solid load, solvent type and particle size distribution) was systematically studied on coating microstructure. All the ethanol based suspensions were found to produce columnar coatings whereas water based and mixed (water and ethanol) suspensions were found to produce

vertically cracked structures. Such a difference in coating formation was explained by studying the suspension droplet momentum and the theories available in the literature. The droplet size was calculated using several suspension properties measured experimentally along with some simplified assumptions. Similarly, the droplet speed was also experimentally measured in order to calculate the droplet momentum. A clear and strong correlation between the droplet momentum and the coating formation in SPS was shown. Additionally, a systematic progressive evolution of columnar as well as lamellar/vertically cracked type microstructures was also successfully shown. Such a progressive microstructure evolution study is by far the first systematic experimental proof to the much sought after shadowing effect theory for column formation used and widely accepted in the field of suspension plasma spraying.

Paper G- Failure analysis of thermally cycled columnar thermal barrier coatings produced by high- velocity-air fuel and axial-suspension-plasma spraying: A design perspective So far (in all the previous Paper A to Paper E) the focus was mainly on establishing the relationships between coating microstructure and thermal conductivity in SPS TBCs. In this work, the influence of coating microstructure on thermal cyclic fatigue lifetime was studied in detail. Eight different TBCs with four different top coats were sprayed using the same suspension but different plasma spray parameters on each of the two different bond coats, which were sprayed using two different BC chemistry with the same high velocity air fuel spray parameters.

An in-depth experimental investigation was performed to understand the failure behaviour of columnar TBCs subjected to thermal cyclic fatigue (TCF) test at 1100 ᵒC. The study revealed that the TCF performance was influenced by the TC microstructure to an extent, but was primarily affected by the severity of thermally grown oxide (TGO) growth at the bond coat-TC interface. Mixed failure modes comprising crack propagation through the bond coat-TGO interface, through TGO and within the TC were identified. Based on the analysis of the experimental results and thorough discussion a novel design of microstructure for the high TCF performance columnar TBC was also proposed.

Paper H- Tailoring columnar microstructure of axial suspension plasma sprayed TBCs for superior thermal shock performance This work was focused on analysing the thermal shock behaviour of the columnar SPS TBCs. In fact, this study was an effort towards understanding the basis for tailoring the columnar microstructures produced by a relatively modern high power ASPS technique to achieve high thermal shock lifetime for TBCs. Failure

SUMMARY OF APPENDED PUBLICATIONS analysis of columnar TBCs subjected to thermal shock test revealed a strong correlation between the total porosity and lifetime where the lifetime was found to be progressively decreasing with the increase in total porosity. Fracture toughness was found to be the most prominent life-determining factor and have shown a strong correlation with the lifetime of various ASPS TBCs. The porosity analysis in this work was done by combining the two techniques namely image analysis and MIP which were studied in earlier papers. Such a combination of two different techniques was found to be more beneficial as image analysis was used to measure only the coarse porosity (due to its sensitivity to measure fine pores) and MIP was used to measure only the fine pores (as MIP can over predict the coarse porosity due to the influence of the high surface roughness of SPS TCs).

In addition, a methodology towards tailoring the columnar microstructure by classifying the porosity in four different classes based on the pore size range was utilized. The porosity was categorized as Class-1(>10 µm), Class-2(10 µm-1 µm), Class-3(1 µm-100 nm), and Class-4(<100 nm) and their individual role towards improving the thermal shock lifetime by improving the toughness was successfully done. It was found that a tailored ASPS columnar microstructure to achieve higher thermal shock lifetime without significantly compromising coatings’ thermal insulation could be a microstructure without Class-4 and Class- 2 porosity, lower Class-1 porosity and optimum Class-3 porosity.

References

[1] R. A. Miller, “Current status of thermal barrier coatings — An overview”, Surf. Coat. Technol., vol. 30, no. 1, pp. 1–11, Jan. 1987. [2] R. A. Miller, “Thermal barrier coatings for aircraft engines: history and directions”, J. Therm. Spray Technol., vol. 6, no. 1, pp. 35–42, Mar. 1997. [3] W. A. Sirignano and F. Liu, “Performance Increases for Gas-Turbine Engines through Combustion Inside the Turbine”, J. Propuls. Power, vol. 15, no. 1, pp. 111–118, 1999. [4] N. P. Padture, M. Gell, and E. H. Jordan, “Thermal Barrier Coatings for Gas-Turbine Engine Applications”, Science, vol. 296, no. 5566, pp. 280–284, Apr. 2002. [5] R. W. Trice et al., “Effect of heat treatment on phase stability, microstructure, and thermal conductivity of plasma-sprayed YSZ”, J. Mater. Sci., vol. 37, no. 11, pp. 2359–2365, Jun. 2002. [6] H. B. Guo, R. Vassen, and D. Stover, “Atmospheric plasma sprayed thick thermal barrier coatings with high segmentation crack density”, Surf. Coat. Technol., vol. 186, no. 3, pp. 353–363, Sep. 2004. [7] X. Q. Cao, R. Vassen, and D. Stover, “Ceramic materials for thermal barrier coatings”, J. Eur. Ceram. Soc., vol. 24, no. 1, pp. 1–10, Jan. 2004. [8] S. Stecura, “Effects of compositional changes on the performance of a thermal barrier coating system for aircraft gas turbine engines”, presented at the American Ceramic Society, Annual Conference on Composites and Advanced Materials, Jan. 21-24, 1979, Merritt Island, FL, 1979. [9] A. D. Jadhav, N. P. Padture, E. H. Jordan, M. Gell, P. Miranzo, and E. R. Fuller Jr., “Low-thermal-conductivity plasma-sprayed thermal barrier coatings with engineered microstructures”, Acta Mater., vol. 54, no. 12, pp. 3343–3349, Jul. 2006. [10] F. Cernuschi, P. Bianchi, M. Leoni, and P. Scardi, “Thermal diffusivity/microstructure relationship in Y-PSZ thermal barrier coatings”, J. Therm. Spray Technol., vol. 8, no. 1, pp. 102–109, Mar. 1999. [11] Y. H. Sohn, E. Y. Lee, B. A. Nagaraj, R. R. Biederman, and R. D. Sisson Jr., “Microstructural characterization of thermal barrier coatings on high pressure turbine blades”, Surf. Coat. Technol., vol. 146–147, pp. 132–139, Sep. 2001. [12] J. Ilavsky, “Characterization of Complex Thermal Barrier Deposits Pore Microstructures by a Combination of Imaging, Scattering, and Intrusion Techniques”, J. Therm. Spray Technol., vol. 19, no. 1–2, pp. 178–189, Jul. 2009. [13] H. Kassner, A. Stuke, M. Rodig, R. Vassen, and D. Stover, “Influence of Porosity on Thermal Conductivity and Sintering in Suspension Plasma

Sprayed Thermal Barrier Coatings”, Adv. Ceram. Coat. Interfaces III Ceram. Eng. Sci. Proc, vol. 46, p. 147, Apr. 2009. [14] A. Kulkarni, A. Vaidya, A. Goland, S. Sampath, and H. Herman, “Processing effects on porosity-property correlations in plasma sprayed yttria-stabilized zirconia coatings”, Mater. Sci. Eng. A, vol. 359, no. 1–2, pp. 100–111, Oct. 2003. [15] N. Curry, N. Markocsan, X.-H. Li, A. Tricoire, and M. Dorfman, “Next Generation Thermal Barrier Coatings for the Gas Turbine Industry”, J. Therm. Spray Technol., vol. 20, no. 1–2, pp. 108–115, Jan. 2011. [16] K. W. Schlichting, N. P. Padture, E. H. Jordan, and M. Gell, “Failure modes in plasma-sprayed thermal barrier coatings”, Mater. Sci. Eng. A, vol. 342, no. 1–2, pp. 120–130, Feb. 2003. [17] A. Rabiei and A. G. Evans, “Failure mechanisms associated with the thermally grown oxide in plasma-sprayed thermal barrier coatings”, Acta Mater., vol. 48, no. 15, pp. 3963–3976, Sep. 2000. [18] P. L. Fauchais, J. V. R. Heberlein, and M. Boulos, Thermal Spray Fundamentals: From Powder to Part. Springer Science & Business Media, 2014. [19] G. W. Goward and L. W. Cannon, “Pack Cementation Coatings for Superalloys: A Review of History, Theory, and Practice”, J. Eng. Gas Turbines Power, vol. 110, no. 1, pp. 150–154, Jan. 1988. [20] J. R. Nicholls, N. J. Simms, W. Y. Chan, and H. E. Evans, “Smart overlay coatings — concept and practice”, Surf. Coat. Technol., vol. 149, no. 2–3, pp. 236–244, Jan. 2002. [21] F. H. Stott and G. C. Wood, “Growth and adhesion of oxide scales on Al2O3-forming alloys and coatings”, Mater. Sci. Eng., vol. 87, pp. 267–274, Mar. 1987. [22] V. V. Sizov, M. J. Lampinen, and A. Laaksonen, “Molecular dynamics simulation of oxygen diffusion in cubic yttria-stabilized zirconia: Effects of temperature and composition”, Solid State Ion., vol. 266, pp. 29–35, Nov. 2014. [23] W. R. Chen, X. Wu, B. R. Marple, D. R. Nagy, and P. C. Patnaik, “TGO growth behaviour in TBCs with APS and HVOF bond coats”, Surf. Coat. Technol., vol. 202, no. 12, pp. 2677–2683, Mar. 2008. [24] C. Zhou, N. Wang, Z. Wang, S. Gong, and H. Xu, “Thermal cycling life and thermal diffusivity of a plasma-sprayed nanostructured thermal barrier coating”, Scr. Mater., vol. 51, no. 10, pp. 945–948, Nov. 2004. [25] N. Curry, N. Markocsan, L. A–stergren, X.-H. Li, and M. Dorfman, “Evaluation of the Lifetime and Thermal Conductivity of Dysprosia- Stabilized Thermal Barrier Coating Systems”, J. Therm. Spray Technol., vol. 22, no. 6, pp. 864–872, Aug. 2013. [26] A. G. Evans, D. R. Mumm, J. W. Hutchinson, G. H. Meier, and F. S. Pettit, “Mechanisms controlling the durability of thermal barrier coatings”, Prog. Mater. Sci., vol. 46, no. 5, pp. 505–553, 2001.

REFERENCES

[27] C.-L. Wang et al., “Phase transformation and crystalline growth of 4 mol% yttria partially stabilized zirconia”, J. Sol-Gel Sci. Technol., vol. 70, no. 3, pp. 428–440, Apr. 2014. [28] C. G. Levi, “Emerging materials and processes for thermal barrier systems”, Curr. Opin. Solid State Mater. Sci., vol. 8, no. 1, pp. 77–91, Jan. 2004. [29] R. A. Miller, J. L. Smialek, and R. G. Garlick, “Phase stability in plasma- sprayed, partially stabilized zirconia-yttria”, in Science and Technology of Zirconia, Advances in Ceramics, vol. 3, Westerville, OH: The American Ceramic Society, 1981, pp. 241–253. [30] J. R. Davis, Handbook of Thermal Spray Technology. ASM International, 2004. [31] K. von Niessen and M. Gindrat, “Plasma Spray-PVD: A New Thermal Spray Process to Deposit Out of the Vapor Phase”, J. Therm. Spray Technol., vol. 20, no. 4, pp. 736–743, Apr. 2011. [32] G.-J. Yang, X.-D. Xiang, L.-K. Xing, D.-J. Li, C.-J. Li, and C.-X. Li, “Isothermal Oxidation Behavior of NiCoCrAlTaY Coating Deposited by High Velocity Air-Fuel Spraying”, J. Therm. Spray Technol., vol. 21, no. 3–4, pp. 391–399, Jan. 2012. [33] E. Bertilsson, “Identification of business cases for HVAF technology”, Master of Science Thesis, KTH, Stockholm, Sweden, 2012. [34] H. Panchal and S. Amin, “A Review on Thermal Spray Coating Processes”, Int. J. Curr. Trends Eng. Res. IJCTER, vol. 2, pp. 556–563, Apr. 2016. [35] A. Killinger, M. Kuhn, and R. Gadow, “High-Velocity Suspension Flame Spraying (HVSFS), a new approach for spraying nanoparticles with hypersonic speed”, Surf. Coat. Technol., vol. 201, no. 5, pp. 1922–1929, Oct. 2006. [36] Z. Zeng, N. Sakoda, T. Tajiri, and S. Kuroda, “Structure and corrosion behavior of 316L stainless steel coatings formed by HVAF spraying with and without sealing”, Surf. Coat. Technol., vol. 203, no. 3, pp. 284–290, Nov. 2008. [37] E. Sadeghimeresht, N. Markocsan, and P. Nylen, “Microstructural and electrochemical characterization of Ni-based bi-layer coatings produced by the HVAF process”, Surf. Coat. Technol., vol. 304, pp. 606–619, Oct. 2016. [38] N. Curry, Z. Tang, N. Markocsan, and P. Nylen, “Influence of bond coat surface roughness on the structure of axial suspension plasma spray thermal barrier coatings — Thermal and lifetime performance”, Surf. Coat. Technol., vol. 268, pp. 15–23, Apr. 2015. [39] Y. G. Shan, Y. L. Wang, and T. Coyle, “Analysis of deposits formation in plasma spraying with liquid precursors”, Appl. Therm. Eng., vol. 51, no. 1–2, pp. 690–697, Mar. 2013. [40] A. A. Kulkarni et al., “Advanced microstructural characterization of plasma- sprayed zirconia coatings over extended length scales”, J. Therm. Spray Technol., vol. 14, no. 2, pp. 239–250, Jun. 2005.

[41] A. Kulkarni et al., “Comprehensive microstructural characterization and predictive property modeling of plasma-sprayed zirconia coatings”, Acta Mater., vol. 51, no. 9, pp. 2457–2475, May 2003. [42] S. C. Tjong and H. Chen, “Nanocrystalline materials and coatings”, Mater. Sci. Eng. R Rep., vol. 45, no. 1–2, pp. 1–88, Sep. 2004. [43] H. Chen, X. Zhou, and C. Ding, “Investigation of the thermomechanical properties of a plasma-sprayed nanostructured zirconia coating”, J. Eur. Ceram. Soc., vol. 23, no. 9, pp. 1449–1455, Aug. 2003. [44] R. E. Taylor, X. Wang, and X. Xu, “Thermophysical properties of thermal barrier coatings”, Surf. Coat. Technol., vol. 120–121, pp. 89–95, Nov. 1999. [45] O. Tingaud, A. Bacciochini, G. Montavon, A. Denoirjean, and P. Fauchais, “Suspension DC plasma spraying of thick finely-structured ceramic coatings: Process manufacturing mechanisms”, Surf. Coat. Technol., vol. 203, no. 15, pp. 2157–2161, May 2009. [46] B. Liang and C. Ding, “Thermal shock resistances of nanostructured and conventional zirconia coatings deposited by atmospheric plasma spraying”, Surf. Coat. Technol., vol. 197, no. 2–3, pp. 185–192, Jul. 2005. [47] Z. Tang, H. Kim, I. Yaroslavski, G. Masindo, Z. Celler, and D. Ellsworth, presented at the S2TS, Tours, France, 10-Apr-2011. [48] H. Tabbara and S. Gu, “A study of liquid droplet disintegration for the development of nanostructured coatings”, AIChE J., vol. 58, no. 11, pp. 3533–3544, Nov. 2012. [49] L. Pawlowski, “Suspension and solution thermal spray coatings”, Surf. Coat. Technol., vol. 203, no. 19, pp. 2807–2829, Jun. 2009. [50] P. Fauchais and G. Montavon, “Latest Developments in Suspension and Liquid Precursor Thermal Spraying”, J. Therm. Spray Technol., vol. 19, no. 1– 2, pp. 226–239, Jan. 2010. [51] H. Kassner, R. Siegert, D. Hathiramani, R. Vassen, and D. Stover, “Application of Suspension Plasma Spraying (SPS) for Manufacture of Ceramic Coatings”, J. Therm. Spray Technol., vol. 17, no. 1, pp. 115–123, Mar. 2008. [52] F.-L. Toma, A. Potthoff, L.-M. Berger, and C. Leyens, “Demands, Potentials, and Economic Aspects of Thermal Spraying with Suspensions: A Critical Review”, J. Therm. Spray Technol., pp. 1–10, Jul. 2015. [53] R. Vassen, H. Kassner, G. Mauer, and D. Stover, “Suspension Plasma Spraying: Process Characteristics and Applications”, J. Therm. Spray Technol., vol. 19, no. 1–2, pp. 219–225, Jan. 2010. [54] P. Fauchais, V. Rat, J.-F. Couder, R. Etchart-Salas, and G. Montavon, “Operating parameters for suspension and solution plasma-spray coatings”, vol. 202, no. 18, pp. 4309–4317, Apr. 2008. [55] T. Bhatia et al., “Mechanisms of ceramic coating deposition in solution- precursor plasma spray”, Mater. Res. Soc., vol. 17, no. 9, Sep. 2002. [56] A. Killinger, R. Gadow, G. Mauer, A. Guignard, R. Vassen, and D. Stover, “Review of New Developments in Suspension and Solution Precursor

REFERENCES

Thermal Spray Processes”, J. Therm. Spray Technol., vol. 20, no. 4, pp. 677– 695, Jun. 2011. [57] L. Latka et al., “Thermal diffusivity and conductivity of yttria stabilized zirconia coatings obtained by suspension plasma spraying”, Surf. Coat. Technol., vol. 208, pp. 87–91, Sep. 2012. [58] Z. Chen, R. W. Trice, M. Besser, X. Yang, and D. Sordelet, “Air-plasma spraying colloidal solutions of nanosized ceramic powders”, J. Mater. Sci., vol. 39, no. 13, pp. 4171–4178, Jul. 2004. [59] R. Siegert, “A Novel Process for the Liquid Feedstock Plasma Spray of Ceramic Coatings with Nanostructural Features”, PhD Thesis, Ruhr University Bochum, Bochum, Germany, 2006. [60] R. Rampon, O. Marchand, C. Filiatre, and G. Bertrand, “Influence of suspension characteristics on coatings microstructure obtained by suspension plasma spraying”, Surf. Coat. Technol., vol. 202, no. 18, pp. 4337– 4342, Jun. 2008. [61] B. Pateyron, N. Calve, and L. Pawlowski, “Influence of water and ethanol on transport properties of the jets used in suspension plasma spraying”, Surf. Coat. Technol., vol. 220, pp. 257–260, Apr. 2013. [62] N. Curry, K. VanEvery, T. Snyder, J. Susnjar, and S. Bjorklund, “Performance Testing of Suspension Plasma Sprayed Thermal Barrier Coatings Produced with Varied Suspension Parameters”, Coatings, vol. 5, no. 3, pp. 338–356, Jul. 2015. [63] D. Waldbillig and O. Kesler, “The effect of solids and dispersant loadings on the suspension viscosities and deposition rates of suspension plasma sprayed YSZ coatings”, Surf. Coat. Technol., vol. 203, no. 15, pp. 2098–2101, May 2009. [64] B. M. Cetegen and S. Basu, “Review of Modeling of Liquid Precursor Droplets and Particles Injected into Plasmas and High-Velocity Oxy-Fuel (HVOF) Flame Jets for Thermal Spray Deposition Applications”, J. Therm. Spray Technol., vol. 18, no. 5–6, pp. 769–793, Sep. 2009. [65] J. Fazilleau, C. Delbos, V. Rat, J. F. Coudert, P. Fauchais, and B. Pateyron, “Phenomena Involved in Suspension Plasma Spraying Part 1: Suspension Injection and Behavior”, Plasma Chem. Plasma Process., vol. 26, no. 4, pp. 371– 391, Apr. 2006. [66] H. E. Wolfe and W. H. Andersen, “Kinetics, mechanism, and resultant droplet sizes of the aerodynamic breakup of liquid drops”, Aerojet-general corp downey ca, 0395-04-18-SP, Apr. 1964. [67] H. Kassner, R. Vassen, and D. Stover, “Study on instant droplet and particle stages during suspension plasma spraying (SPS)”, Surf. Coat. Technol., vol. 202, no. 18, pp. 4355–4361, Jun. 2008. [68] J. Oberste-Berghaus, S. Bouaricha, J.-G. Legoux, and C. Moreau, “Injection conditions and in-flight particle states in suspension plasma spraying of alumina and zirconia nano-ceramics”, in Proceedings of the International Thermal Spray Conference, 2005, Basel, Switzerland, 2005.

[69] K. VanEvery et al., “Column Formation in Suspension Plasma-Sprayed Coatings and Resultant Thermal Properties”, J. Therm. Spray Technol., vol. 20, no. 4, pp. 817–828, Jun. 2011. [70] H. Guo, S. Kuroda, and H. Murakami, “Microstructures and Properties of Plasma-Sprayed Segmented Thermal Barrier Coatings”, J. Am. Ceram. Soc., vol. 89, no. 4, pp. 1432–1439, 2006. [71] P. Carpio et al., “Correlation of thermal conductivity of suspension plasma sprayed yttria stabilized zirconia coatings with some microstructural effects”, Mater. Lett., vol. 107, pp. 370–373, Sep. 2013. [72] D. Schwingel, R. Taylor, T. Haubold, J. Wigren, and C. Gualco, “Mechanical and thermophysical properties of thick PYSZ thermal barrier coatings: correlation with microstructure and spraying parameters”, Surf. Coat. Technol., vol. 108–109, pp. 99–106, Oct. 1998. [73] I. O. Golosnoy, S. A. Tsipas, and T. W. Clyne, “An analytical model for simulation of heat flow in plasma-sprayed thermal barrier coatings”, J. Therm. Spray Technol., vol. 14, no. 2, pp. 205–214, Jun. 2005. [74] I. O. Golosnoy, A. Cipitria, and T. W. Clyne, “Heat Transfer Through Plasma-Sprayed Thermal Barrier Coatings in Gas Turbines: A Review of Recent Work”, J. Therm. Spray Technol., vol. 18, no. 5–6, pp. 809–821, Dec. 2009. [75] A. Bacciochini, G. Montavon, J. Ilavsky, A. Denoirjean, and P. Fauchais, “Porous Architecture of SPS Thick YSZ Coatings Structured at the Nanometer Scale (~50 nm)”, J. Therm. Spray Technol., vol. 19, no. 1–2, pp. 198–206, Jan. 2010. [76] F.-L. Toma, L.-M. Berger, T. Naumann, and S. Langner, “Microstructures of nanostructured ceramic coatings obtained by suspension thermal spraying”, Surf. Coat. Technol., vol. 202, no. 18, pp. 4343–4348, Jun. 2008. [77] C. C. Li, T. Wang, X. J. Liu, Z. H. Zheng, and Q. Li, “Evolution of mechanical properties of thermal barrier coatings subjected to thermal exposure by instrumented indentation testing”, Ceram. Int., vol. 42, no. 8, pp. 10242–10250, Jun. 2016. [78] J. Matejicek and S. Sampath, “In situ measurement of residual stresses and elastic moduli in thermal sprayed coatings: Part 1: apparatus and analysis”, Acta Mater., vol. 51, no. 3, pp. 863–872, Feb. 2003. [79] S. Guo and Y. Kagawa, “Young’s moduli of zirconia top-coat and thermally grown oxide in a plasma-sprayed thermal barrier coating system”, Scr. Mater., vol. 50, no. 11, pp. 1401–1406, Jun. 2004. [80] F. Kroupa, “Nonlinear Behavior in Compression and Tension of Thermally Sprayed Ceramic Coatings”, J. Therm. Spray Technol., vol. 16, no. 1, pp. 84– 95, Mar. 2007. [81] P. Carpio et al., “Microstructure and indentation mechanical properties of YSZ nanostructured coatings obtained by suspension plasma spraying”, Surf. Coat. Technol., vol. 220, no. Supplement C, pp. 237–243, Apr. 2013.

REFERENCES

[82] G. Dwivedi, V. Viswanathan, S. Sampath, A. Shyam, and E. Lara-Curzio, “Fracture Toughness of Plasma-Sprayed Thermal Barrier Ceramics: Influence of Processing, Microstructure, and Thermal Aging”, J. Am. Ceram. Soc., vol. 97, no. 9, pp. 2736–2744, Sep. 2014. [83] A. Feuerstein, J. Knapp, T. Taylor, A. Ashary, A. Bolcavage, and N. Hitchman, “Technical and Economical Aspects of Current Thermal Barrier Coating Systems for Gas Turbine Engines by Thermal Spray and EBPVD: A Review”, J. Therm. Spray Technol., vol. 17, pp. 199–213, Jun. 2008. [84] X. Chen et al., “Columnar and DVC-structured thermal barrier coatings deposited by suspension plasma spray: high-temperature stability and their corrosion resistance to the molten salt”, Ceram. Int., vol. 42, no. 15, pp. 16822–16832, Nov. 2016. [85] P. Carpio, E. Bannier, M. D. Salvador, A. Borrell, R. Moreno, and E. Sánchez, “Effect of particle size distribution of suspension feedstock on the microstructure and mechanical properties of suspension plasma spraying YSZ coatings”, Surf. Coat. Technol., vol. 268, pp. 293–297, Apr. 2015. [86] D. Zhou, O. Guillon, and R. Vassen, “Development of YSZ Thermal Barrier Coatings Using Axial Suspension Plasma Spraying”, Coatings, vol. 7, no. 8, p. 120, Aug. 2017. [87] M. Ahrens, R. Vassen, D. Stover, and S. Lampenscherf, “Sintering and creep processes in plasma-sprayed thermal barrier coatings”, J. Therm. Spray Technol., vol. 13, no. 3, pp. 432–442, Sep. 2004. [88] B. Cheng et al., “Sintering induced the failure behavior of dense vertically crack and lamellar structured TBCs with equivalent thermal insulation performance”, Ceram. Int., Aug. 2017. [89] R. Krishnamurthy and D. J. Srolovitz, “Sintering and microstructure evolution in columnar thermal barrier coatings”, Acta Mater., vol. 57, no. 4, pp. 1035–1048, Feb. 2009. [90] S. Kumar and A. C. F. Cocks, “Sintering and mud cracking in EB-PVD thermal barrier coatings”, J. Mech. Phys. Solids, vol. 60, no. 4, pp. 723–749, Apr. 2012. [91] S. Deshpande, “High Temperature Sintering and Oxidation Behavior in Plasma Sprayed TBCs [Single Splat Studies] Paper 2—Relevance of Variation in Materials Systems of TBC Components”, J. Surf. Eng. Mater. Adv. Technol., vol. 03, no. 01, pp. 116–132, 2013. [92] X. Zhao, X. Wang, and P. Xiao, “Sintering and failure behaviour of EB- PVD thermal barrier coating after isothermal treatment”, Surf. Coat. Technol., vol. 200, no. 20–21, pp. 5946–5955, May 2006. [93] M. N. Rahaman, Ceramic Processing and Sintering. CRC Press, 2003. [94] A. Cipitria, I. O. Golosnoy, and T. W. Clyne, “A sintering model for plasma- sprayed zirconia TBCs. Part I: Free-standing coatings”, Acta Mater., vol. 57, no. 4, pp. 980–992, Feb. 2009.

[95] A. Cipitria, I. O. Golosnoy, and T. W. Clyne, “A sintering model for plasma- sprayed zirconia thermal barrier coatings. Part II: Coatings bonded to a rigid substrate”, Acta Mater., vol. 57, no. 4, pp. 993–1003, Feb. 2009. [96] G.-R. Li, G.-J. Yang, C.-X. Li, and C.-J. Li, “A comprehensive mechanism for the sintering of plasma-sprayed nanostructured thermal barrier coatings”, Ceram. Int., vol. 43, no. 13, pp. 9600–9615, Sep. 2017. [97] B. Siebert, C. Funke, R. Vassen, and D. Stover, “Changes in porosity and Young’s Modulus due to sintering of plasma sprayed thermal barrier coatings”, J. Mater. Process. Technol., vol. 92–93, pp. 217–223, Aug. 1999. [98] K. Wang, H. Peng, H. Guo, and S. Gong, “Effect of Sintering on Thermal Conductivity and Thermal Barrier Effects of Thermal Barrier Coatings”, Chin. J. Aeronaut., vol. 25, no. 5, pp. 811–816, Oct. 2012. [99] N. Zotov, M. Bartsch, L. Chernova, D. A. Schmidt, M. Havenith, and G. Eggeler, “Effects of annealing on the microstructure and the mechanical properties of EB-PVD thermal barrier coatings”, Surf. Coat. Technol., vol. 205, no. 2, pp. 452–464, Oct. 2010. [100] U. Schulz, K. Fritscher, C. Leyens, and M. Peters, “High-Temperature Aging of EB-PVD Thermal Barrier Coatings”, in Ceramic Engineering and Science Proceedings, vol. 22, M. Singh and T. Jessen, Eds. Hoboken, NJ, USA: John Wiley & Sons, Inc., 2001, pp. 347–356. [101] H. E. Exner and C. Müller, “Particle Rearrangement and Pore Space Coarsening During Solid-State Sintering”, J. Am. Ceram. Soc., vol. 92, no. 7, pp. 1384–1390, Jul. 2009. [102] V. Lughi, V. K. Tolpygo, and D. R. Clarke, “Microstructural aspects of the sintering of thermal barrier coatings”, Mater. Sci. Eng. A, vol. 368, no. 1–2, pp. 212–221, Mar. 2004. [103] Z. Wu, L. Ni, Q. Yu, and C. Zhou, “Effect of Thermal Exposure on Mechanical Properties of a Plasma-Sprayed Nanostructured Thermal Barrier Coating”, J. Therm. Spray Technol., vol. 21, no. 1, pp. 169–175, Jan. 2012. [104] R. B. Dinwiddie, S. C. Beecher, W. D. Porter, and B. A. Nagaraj, “The Effect of Thermal Aging on the Thermal Conductivity of Plasma Sprayed and EB-PVD Thermal Barrier Coatings”, p. V005T13A006, Jun. 1996. [105] T. R. Kakuda, A. M. Limarga, T. D. Bennett, and D. R. Clarke, “Evolution of thermal properties of EB-PVD 7YSZ thermal barrier coatings with thermal cycling”, Acta Mater., vol. 57, no. 8, pp. 2583–2591, May 2009. [106] J. A. Thompson and T. W. Clyne, “The effect of heat treatment on the stiffness of zirconia top coats in plasma-sprayed TBCs”, Acta Mater., vol. 49, no. 9, pp. 1565–1575, May 2001. [107] Y. Zhao et al., “Thermal Aging Behavior of Axial Suspension Plasma- Sprayed Yttria-Stabilized Zirconia (YSZ) Thermal Barrier Coatings”, J. Therm. Spray Technol., vol. 24, no. 3, pp. 338–347, 2014. [108] A. Guignard, G. Mauer, R. Vassen, and D. Stover, “Deposition and Characteristics of Submicrometer-Structured Thermal Barrier Coatings by

REFERENCES

Suspension Plasma Spraying”, J. Therm. Spray Technol., vol. 21, no. 3–4, pp. 416–424, Jun. 2012. [109] D. Naumenko, V. Shemet, L. Singheiser, and W. J. Quadakkers, “Failure mechanisms of thermal barrier coatings on MCrAlY-type bond coats associated with the formation of the thermally grown oxide”, J. Mater. Sci., vol. 44, no. 7, pp. 1687–1703, Apr. 2009. [110] W. R. Chen, X. Wu, B. R. Marple, and P. C. Patnaik, “Oxidation and crack nucleation/growth in an air-plasma-sprayed thermal barrier coating with NiCrAlY bond coat”, Surf. Coat. Technol., vol. 197, no. 1, pp. 109–115, Jul. 2005. [111] X. Li, X. Huang, Q. Yang, and Z. Tang, “Effects of substrate material and TBC structure on the cyclic oxidation resistance of TBC systems”, Surf. Coat. Technol., vol. 258, pp. 49–61, Nov. 2014. [112] B. C. Wu, E. Chang, S. F. Chang, and C. H. Chao, “Thermal cyclic response of yttria-stabilized zirconia/CoNiCrAlY thermal barrier coatings”, Thin Solid Films, vol. 172, no. 2, pp. 185–196, May 1989. [113] E. A. G. Shillington and D. R. Clarke, “Spalling failure of a thermal barrier coating associated with aluminum depletion in the bond-coat”, Acta Mater., vol. 47, no. 4, pp. 1297–1305, Mar. 1999. [114] E. P. Busso, J. Lin, S. Sakurai, and M. Nakayama, “A mechanistic study of oxidation-induced degradation in a plasma-sprayed thermal barrier coating system.: Part I: model formulation”, Acta Mater., vol. 49, no. 9, pp. 1515– 1528, May 2001. [115] K. Yuan, “Oxidation and Corrosion of New MCrAlX Coatings : Modelling and Experiments”, 2014. [116] A. G. Evans, D. R. Clarke, and C. G. Levi, “The influence of oxides on the performance of advanced gas turbines”, J. Eur. Ceram. Soc., vol. 28, no. 7, pp. 1405–1419, Jan. 2008. [117] W. R. Chen, X. Wu, B. R. Marple, R. S. Lima, and P. C. Patnaik, “Pre- oxidation and TGO growth behaviour of an air-plasma-sprayed thermal barrier coating”, Surf. Coat. Technol., vol. 202, no. 16, pp. 3787–3796, May 2008. [118] H. E. Evans, “Oxidation failure of TBC systems: An assessment of mechanisms”, Surf. Coat. Technol., vol. 206, no. 7, pp. 1512–1521, Dec. 2011. [119] W. R. Chen, X. Wu, D. Dudzinski, and P. C. Patnaik, “Modification of oxide layer in plasma-sprayed thermal barrier coatings”, Surf. Coat. Technol., vol. 200, no. 20–21, pp. 5863–5868, May 2006. [120] T. S. Hille, S. Turteltaub, and A. S. J. Suiker, “Oxide growth and damage evolution in thermal barrier coatings”, Eng. Fract. Mech., vol. 78, no. 10, pp. 2139–2152, Jul. 2011. [121] D. R. Clarke and W. Pompe, “Critical radius for interface separation of a compressively stressed film from a rough surface”, Acta Mater., vol. 47, no. 6, pp. 1749–1756, Apr. 1999. [122] N. Curry, “Design of Thermal Barrier Coating Systems”, 2014.

[123] R. Vassen, S. Giesen, and D. Stover, “Lifetime of Plasma-Sprayed Thermal Barrier Coatings: Comparison of Numerical and Experimental Results”, J. Therm. Spray Technol., vol. 18, no. 5–6, p. 835, Dec. 2009. [124] V. Viswanathan, G. Dwivedi, and S. Sampath, “Engineered Multilayer Thermal Barrier Coatings for Enhanced Durability and Functional Performance”, J. Am. Ceram. Soc., vol. 97, no. 9, pp. 2770–2778, Sep. 2014. [125] M. Gupta, C. Kumara, and P. Nylén, “Bilayer Suspension Plasma-Sprayed Thermal Barrier Coatings with Enhanced Thermal Cyclic Lifetime: Experiments and Modeling”, J. Therm. Spray Technol., vol. 26, no. 6, pp. 1038–1051, Aug. 2017. [126] A. Bolcavage, A. Feuerstein, J. Foster, and P. Moore, “Thermal shock testing of thermal barrier coating/bondcoat systems”, J. Mater. Eng. Perform., vol. 13, no. 4, pp. 389–397, Aug. 2004. [127] M. Ranjbar-Far, J. Absi, S. Shahidi, and G. Mariaux, “Impact of the non- homogenous temperature distribution and the coatings process modeling on the thermal barrier coatings system”, Mater. Des., vol. 32, no. 2, pp. 728– 735, Feb. 2011. [128] R. Vassen, A. Stuke, and D. Stöver, “Recent Developments in the Field of Thermal Barrier Coatings”, J. Therm. Spray Technol., vol. 18, no. 2, pp. 181– 186, Jun. 2009. [129] W. Q. Wang, C. K. Sha, D. Q. Sun, and X. Y. Gu, “Microstructural feature, thermal shock resistance and isothermal oxidation resistance of nanostructured zirconia coating”, Mater. Sci. Eng. A, vol. 424, no. 1, pp. 1– 5, May 2006. [130] R. Ahmadi-Pidani, R. Shoja-Razavi, R. Mozafarinia, and H. Jamali, “Improving the thermal shock resistance of plasma sprayed CYSZ thermal barrier coatings by laser surface modification”, Opt. Lasers Eng., vol. 50, no. 5, pp. 780–786, May 2012. [131] H. Jamali, R. Mozafarinia, R. Shoja Razavi, and R. Ahmadi-Pidani, “Comparison of thermal shock resistances of plasma-sprayed nanostructured and conventional yttria stabilized zirconia thermal barrier coatings”, Ceram. Int., vol. 38, no. 8, pp. 6705–6712, Dec. 2012. [132] R. W. Steinbrech, V. Postolenko, J. Monch, J. Malzbender, and L. Singheiser, “Testing method to assess lifetime of EB-PVD thermal barrier coatings on tubular specimens in static and cyclic oxidation tests”, Ceram. Int., vol. 37, no. 1, pp. 363–368, Jan. 2011. [133] M. Karger, R. Vassen, and D. Stover, “Atmospheric plasma sprayed thermal barrier coatings with high segmentation crack densities: Spraying process, microstructure and thermal cycling behavior”, Surf. Coat. Technol., vol. 206, no. 1, pp. 16–23, Oct. 2011. [134] G. Di Girolamo, F. Marra, C. Blasi, E. Serra, and T. Valente, “Microstructure, mechanical properties and thermal shock resistance of plasma sprayed nanostructured zirconia coatings”, Ceram. Int., vol. 37, no. 7, pp. 2711–2717, Sep. 2011.

REFERENCES

[135] R. Vassen, M. O. Jarligo, T. Steinke, D. E. Mack, and D. Stover, “Overview on advanced thermal barrier coatings”, Surf. Coat. Technol., vol. 205, no. 4, pp. 938–942, Nov. 2010. [136] W. J. Brindley and T. A. Leonhardt, “Metallographic techniques for evaluation of thermal barrier coatings”, Mater. Charact., vol. 24, no. 2, pp. 93–101, Mar. 1990. [137] Y. Chen, R. C. Reed, and E. A. Marquis, “Microstructural Characterization of a Thermal Barrier Coating System Using SEM, TEM and APT Techniques”, Microsc. Microanal., vol. 19, no. Supplement S2, pp. 1866–1867, Aug. 2013. [138] C. Eberl, D. S. Gianola, and K. J. Hemker, “Mechanical Characterization of Coatings Using Microbeam Bending and Digital Image Correlation Techniques”, Exp. Mech., vol. 50, no. 1, pp. 85–97, Oct. 2008. [139] Y. Sheiretov, L. Evans, D. Schlicker, V. Zilberstein, N. Goldfine, and R. Sikorski, “TBC Characterization Using Magnetic and Electric Field Sensors”, presented at ASME Turbo Expo, pp. 97–103, Jan. 2007. [140] A. B. Abell, K. L. Willis, and D. A. Lange, “Mercury Intrusion Porosimetry and Image Analysis of Cement-Based Materials”, J. Colloid Interface Sci., vol. 211, no. 1, pp. 39–44, Mar. 1999. [141] W. J. Parker, R. J. Jenkins, C. P. Butler, and G. L. Abbott, “Flash Method of Determining Thermal Diffusivity, Heat Capacity, and Thermal Conductivity”, J. Appl. Phys., vol. 32, no. 9, pp. 1679–1684, Sep. 1961. [142] R. D. Cowan, “Pulse Method of Measuring Thermal Diffusivity at High Temperatures”, J. Appl. Phys., vol. 34, no. 4, pp. 926–927, Apr. 1963. [143] J. A. Cape and G. W. Lehman, “Temperature and Finite Pulse‐Time Effects in the Flash Method for Measuring Thermal Diffusivity”, J. Appl. Phys., vol. 34, no. 7, pp. 1909–1913, Jul. 1963. [144] R. Taylor, “Thermal Conductivity Determinations of Thermal Barrier Coatings”, Mater. Sci. Eng. A, vol. 245, no. 2, pp. 160–167, 1998. [145] J. I. Eldridge, C. M. Spuckler, K. W. Street, and J. R. Markham, “Infrared Radiative Properties of Yttria–Stabilized Zirconia Thermal Barrier Coatings”, in 26th Annual Conference on Composites, Advanced Ceramics, Materials, and Structures: B: Ceramic Engineering and Science Proceedings, H.-T. Lin and M. Singh, Eds. John Wiley & Sons, Inc., 2002, pp. 417–430. [146] A. Kucuk, C. C. Berndt, U. Senturk, R. S. Lima, and C. R. C. Lima, “Influence of plasma spray parameters on mechanical properties of yttria stabilized zirconia coatings. I: Four point bend test”, Mater. Sci. Eng. A, vol. 284, no. 1, pp. 29–40, May 2000. [147] X. Wang, S. Tint, M. Chiu, and A. Atkinson, “Stiffness of free-standing thermal barrier coating top coats measured by bending tests”, Acta Mater., vol. 60, no. 8, pp. 3247–3258, May 2012. [148] E. M. Donohue, N. R. Philips, M. R. Begley, and C. G. Levi, “Thermal barrier coating toughness: Measurement and identification of a bridging

mechanism enabled by segmented microstructure”, Mater. Sci. Eng. A, vol. 564, no. Supplement C, pp. 324–330, Mar. 2013. [149] B. A. Cottom and M. J. Mayo, “Fracture toughness of nanocrystalline ZrO2-3mol% Y2O3 determined by vickers indentation”, Scr. Mater., vol. 34, no. 5, pp. 809–814, Mar. 1996. [150] N. Margadant et al., “Impact of probing volume from different mechanical measurement methods on elastic properties of thermally sprayed Ni-based coatings on a mesoscopic scale”, Surf. Coat. Technol., vol. 200, no. 8, pp. 2805–2820, Jan. 2006. [151] R. Vassen, X. Cao, F. Tietz, D. Basu, and D. Stover, “Zirconates as New Materials for Thermal Barrier Coatings”, J. Am. Ceram. Soc., vol. 83, no. 8, pp. 2023–2028, Aug. 2000. [152] K. Niihara, R. Morena, and D. P. H. Hasselman, “Evaluation of KIc of brittle solids by the indentation method with low crack-to-indent ratios”, J. Mater. Sci. Lett., vol. 1, no. 1, pp. 13–16, Jan. 1982. [153] R. Vassen et al., “Recent Activities in the Field of Thermal Barrier Coatings Including Burner Rig Testing in the European Union”, Adv. Eng. Mater., vol. 10, no. 10, pp. 907–921, Oct. 2008.

APPENDED PUBLICATIONS

Paper A: - Characterization of thermal barrier coatings produced by various thermal spray techniques using solid powder, suspension and solution precursor feedstock material

Paper B: - Comparative study of suspension plasma sprayed and suspension high velocity oxy-fuel sprayed YSZ thermal barrier coatings

Paper C: - Characterization of microstructure and thermal properties of YSZ coatings obtained by axial suspension plasma spraying (ASPS)

Paper D: - Influence of microstructure on thermal properties of axial suspension plasma sprayed YSZ thermal barrier coatings

Paper E: - Influence of isothermal heat treatment on porosity and crystallite size in axial suspension plasma sprayed thermal barrier coatings for gas turbine applications

Paper F: - Experimental visualization of microstructure evolution during suspension plasma spraying of thermal barrier coatings

Paper G: - Failure analysis of thermally cycled columnar thermal barrier coatings produced by high-velocity-air fuel and axial-suspension-plasma spraying: A design perspective

Paper H: - Tailoring columnar microstructure of axial suspension plasma sprayed TBCs for superior thermal shock performance

Tidigare avhandlingar – Produktionsteknik

PEIGANG LI Cold lap formation in Gas Metal Arc Welding of steel An experimental study of micro-lack of fusion defects, 2013:2. NICHOLAS CURRY Design of Thermal Barrier Coatings, 2014:3. JEROEN DE BACKER Feedback Control of Robotic Friction Stir Welding, 2014:4. MOHIT KUMAR GUPTA Design of Thermal Barrier Coatings A modelling approach, 2014:5. PER LINDSTRÖM Improved CWM platform for modelling welding procedures and their effects on structural behavior, 2015:6. ERIK ÅSTRAND A Framework for optimised welding of fatigue loaded structures Applied to gas metal arc welding of fillet welds, 2016:7. EMILE GLORIEUX Multi-Robot Motion Planning Optimisation for Handling Sheet Metal Parts, 2017:10. EBRAHIM HARATI Improving fatigue properties of welded high strength steels, 2017:11. ANDREAS SEGERSTARK Laser Metal Deposition using Alloy 718 Powder Influence of Process Parameters on Material Characteristics, 2017:12. ANA ESTHER BONILLA HERNÁNDES On Cutting Tool Resource Management, 2018:16. [Submitted]

SATYAPAL MAHADE Functional Performance of Gadolinium Zirconate/YSZ Multi-layered Thermal Barrier Coatings, 2018:18. [Submitted] AMIR PARSIAN Regenerative Chatter Vibrations in Indexable Drills: Modeling and Simulation, 2018:21. [Submitted] ESMAEIL SADEGHIMERESHT High Temperature Corrosion of Ni-based Coatings, 2018:23. [Submitted].

ASHISH GANVIR PhD Thesis Production Technology 2018 No. 20

ISBN 978-91-87531-92-7 (Printed version) ISBN 978-91-87531-91-0 (Electronic version)